Calculus (Bradley & Smith) - Instructor's Edition

~r
INSTRUCTOR’S EDITION
iERALD L. BRADLEY & KARL J. SMITH

SUMMARIES - Where To Look For Help
PREREQUISITES
Your algebra, trigonometry, or precalculus textbooks, if you kept them, are good places to look because you are used to the
notation and presentation. If you do not have these books for reference, we have provided the necessary review in the
Student Mathematics Handbook. Also see:
Properties of Absolute Value, see Table 1.1, p. 10

Exact Values of Trigonometric Functions, see Table 1.2, p. 17
Directory of Curves, see Table 1.3, p. 35
LIMITS, DIFFERENTIATION, AND GRAPHING

Holes, Poles, Jumps, and Continuity, see Table 1.4, p. 58
A Summary of First-Derivative and Second-Derivative Tests, see Table 3.1, p. 200
Graphing Strategy, see Table 3.2, p. 215
Derivative Formulas, see page facing inside back cover
INTEGRATION
Definition of Inverse Trigonometric Functions, see Table 5.1, p. 365
Volumes of Common Solids, see Table 6.1, p. 384
Differential Equations: Growth Models, see Table 6.3, p. 420
Modeling Applications Involving Integration, see Table 6.4, ρ. 427-428
Integration Strategy, see Table 7.1, p. 465
Integration Formulas, see inside back cover
Short Integration Table, see Appendix C
Integration Table, see Student Mathematics Handbook
INFINITE SERIES
Guidelines for Determining Convergence of Series, see Table 8.1, p. 527
Power Series for Elementary Functions, see Table 8.2, p. 561
ANALYTIC GEOMETRY, VECTORS, AND THREE DIMENSIONS

Volume Formulas, see Table 6.1, p. 384
Summary of Volumes of Revolution, see Table 6.2, p. 392
Directory of Polar-Form Curves, see Table 9.1, p. 583
Drawing Lesson on Sketching a Prism, p. 386
Drawing Lesson on Three-dimensional Coordinate Systems, p. 640
Drawing Lesson on Sketching a Surface, see p. 644
Quadric Surfaces, see Table 10.1, p. 645
Drawing Lesson on Lines in Space, p. 666
Drawing Lesson on Planes, p. 640
Data on Planetary Orbits, see Table 11.1, p. 708
Curvature Formulas, see Table 11.2, p. 719
Summary of Velocity, Acceleration, and Curvature, see Table 11.3, p. 729
FUNCTIONS OF TWO VARIABLES

Drawing Lesson on Drawing Level Curves, see p. 747
Comparison of Important Integral Theorems, see Table 14.1, p. 960-961
DIFFERENTIAL EQUATIONS
Introduction to Differential Equations, see Section 4.5, p. 287
Growth and Decay, see Table 6.3, p. 420
CALCULUS
Northeast dome ceiling, Masjid-i-Jami, Isfahan, Iran. This structure was built entirely of
brick in the year 1088 and has stood firmly in an earthquake zone for more than 900 years. It
was designed with a sophisticated knowledge of geometry (by solving the problem of mounting
a nearly perfect hemisphere on a square base) and engineered through trial and error (some
earlier buildings simply collapsed). Architecture critic Eric Schroeder has written: “ European
dome-builders never approached their skill. How ingeniously the Western builder compensated
his ignorance of the mechanics of dome construction is attested by the ten chains round the
base of St. Peter’s, and the concealed cone which fastens the haunch of St. Paul’s. But engineers
could not hope to prescribe an ideally light dome of plain masonry before Newton’s work on the
calculus (late in the seventeenth century).” Read on, and upon completion of this text you will
be qualified, at least mathematically, to do more than design domes, (photo credit: Seherr-
Thoss)
CALCULUS
Gerald L. Bradley
Claremont McKenna College
Karl J. Smith
Santa Rosa Junior College
Prentice Hall
Englewood Cliffs, New Jersey 07632
Library o f Congress has cataloged another edition as follows:
Bradley, Gerald L., 1940-

Calculus / Gerald Bradley, Karl Smith.— 1st ed.
p. cm.
Includes index.
ISBN 0-13-178617-2
1. Calculus. I. Smith, Karl J. II. Title.
QA303.B88218 1995
515—dc20 94-27258
CIP
Acquisition Editor: George Lobell

Editor in Chief: Jerome Grant
Development Editor: Alan MacDonell
Production Editor: Edward Thomas
Marketing Manager: Frank Nicolazzo
Supplements Editor: Audra Walsh
Product Manager: Trudy Pisciotti
Design Director: Paula Maylahn
Text Designer: Lee Goldstein
Page layout: Lee Goldstein, Karen Noferi
Cover Designer: Jeanette Jacobs
Art Director: Amy Rosen
Photo Editor: Lorinda Morris-Nantz
Photo research: Melinda Reo, Page Poore-Kidder
Editorial Assistant: Mary DeLuca
Text composition: Black Dot Graphics
Art studio: Tech Graphics
Portrait artist: Bonnie J. Miljour
Copy Editor: Linda Thompson
Cover photo: Koji Horiuchi, courtesy Pei Cobb Freed and Partners
© 1995 by Prentice-Hall, Inc.

A Simon & Schuster Company
Englewood Cliffs, New Jersey 07632
All rights reserved. No part of this book may be
reproduced, in any form or by any means,
without permission in writing from the publisher.
Printed in the United States of America
10 9 8 7 6 5 4 3 2 1
ISBN □ -1 3 -3 3 □ 3 ≡ 5 -X
Prentice-Hall International (UK) Limited, London

Prentice-Hall of Australia Pty. Limited, Sydney
Prentice-Hall Canada Inc., Toronto
Prentice-Hall Hispanoamericana, S.A., Mexico
Prentice-Hall of India Private Limited, New Delhi
Prentice-Hall of Japan, Inc., Tokyo
Simon & Schuster Asia Pte, Ltd., Singapore
Editora Prentice-Hall do Brasil, Ltda., Rio de Janeiro
∞M≡n≡α≡M> I
Contents
About the Authors xii
Preface xiii
Preview of Calculus: Functions and

Limits 1
1.1 What is Calculus?, 2
1.2 Preliminaries, 9
1.3 Lines in the Plane, 20
1.4 Functions and Their Graphs, 26
1.5 The Limit of a Function, 41
1.6 Properties of Limits, 49
1.7 Continuity, 56
1.8 Introduction to the Theory of Limits, 67
Chapter 1 Review, 75
GUEST ESSAY: “Calculus was Inevitable,” by John I..
Troutman, 80
BOOK REPORT: Ethnomathematics, 82
v
vi Contents
2 Techniques of Differentiation with Selected

Applications 83
2.1 An Introduction to the Derivative: Tangents, 84
2.2 Techniques of Differentiation, 95
2.3 Derivatives of the Trigonometric Functions, 104
2.4 Rates of Change: Rectilinear Motion, 112
2.5 The Chain Rule, 120
2.6 Implicit Differentiation, 129
2.7 Related Rates, 136
2.8 Differentials and Tangent Line Approximations, 144
2.9 The Newton-Raphson Method for Approximating Roots, 154
« 1 GROUP RESEARCH PROJECT: “Ups and Downs,” 164
□ Additional Applications of the Derivative 165

Extreme Values of a Continuous Function, 166
The Mean Value Theorem, 176
First-Derivative Test, 183
Concavity and the Second-Derivative Test, 192
Infinite Limits and Asymptotes, 202
Summary of Curve Sketching, 213
Optimization in the Physical Sciences and Engineering, 219
Optimization in Business, Economics, and the Life Sciences, 230
ΓH6pitaΓs Rule, 239
Antiderivatives, 246
⅛ 7 GROUP RESEARCH PROJECT: “Wine Barrel Capacity,” 256
257
Area as the Limit of a Sum; Summation Notation, 258
Riemann Sums and the Definite Integral, 266
The Fundamental Theorem of Calculus; Integration by
Substitution, 276
Introduction to Differential Equations, 287
The Mean Value Theorem for Integrals; Average Value, 295
Numerical Integration: The Trapezoidal Rule and Simpson’s
Rule, 302
Area Between Two Curves, 310
GUEST ESSAY: “Kinematics of Jogging,” by Ralph Boas, 322
BOOK REPORT: To Infinity and Beyond, A Cultural History to the
Infinite, 324
Contents vii
5 Exponential, Logarithmic, and Inverse

Trigonometric Functions 325
5.1 Exponential Functions; The Number e, 326
5.2 Inverse Functions; Logarithms, 334
5.3 Derivatives Involving ex and In x, 344
V ιB e ^-.0 Θ 0 4 2 5 X 5.4 Applications; Derivatives of ex and In x, 351
Y sB.5
5.5 Integrals Involving ex and In x, 358
X nin= 0
Xnax=500θ 5.6 The Inverse Trigonometric Functions, 363
X sc1=500 5.7 An Alternative Approach: The Logarithm as an Integral, 373
V ∩ in = -.2 5
Y roax=l. 5 Chapter 5 Review, 378
V s c l= . 1 GROUP RESEARCH PROJECT: “Quality Control,” 382
OAdditional Applications of the Integral 383

6.1 Volume: Disks, Washers, and Shells, 383
6.2 Arc Length and Surface Area, 396
6.3 Physical Applications: Work, Liquid Force, and Centroids, 404
6.4 Growth, Decay, and First-Order Linear Differential Equations, 416
GROUP RESEARCH PROJECT: “Houdini’s Escape,” 434
Cumulative Review, Chapters 1-6, 435
7 Methods of Integration 436

7.1 Review of Substitution and Integration by Table, 437
7.2 Integration by Parts, 448
7.3 The Method of Partial Fractions, 454
7.4 Summary of Integration Techniques, 465
7.5 Improper Integrals, 469
7.6 The Hyperbolic and Inverse Hyperbolic Functions, 478
GROUP RESEARCH PROJECT: “Buoy Design,” 488
viii Contents
8 Infinite Series 489

8.1 Sequences and Their Limits, 490
8.2 Introduction to Infinite Series; Geometric Series, 500
8.3 The Integral Test; ^-series, 509
8.4 Comparison Tests, 516
8.5 The Ratio Test and the Root Test, 522
8.6 Alternating Series; Absolute and Conditional Convergence, 529
8.7 Power Series, 539
8.8 Taylor and Maclaurin Series, 548
■■ GROUP RESEARCH PROJECT: “Elastic Tightrope Project,” 568
9 Polar Coordinates and Parametric Forms 569

9.1 The Polar Coordinate System, 570
9.2 Graphing in Polar Coordinates, 576
9.3 Area and Tangent Lines in Polar Coordinates, 585
9.4 Parametric Representation of Curves, 594
9.5 Conic Sections: the Parabola, 604
9.6 Conic Sections: the Ellipse and Hyperbola, 612
GROUP RESEARCH PROJECT: “Security System Project,” 629
10 Vectors in the Plane and in Space 630

10.1 Vectors in the Plane, 631
10.2 Quadric Surfaces and Graphing in Three Dimensions, 638
10.3 The Dot Product, 648
10.4 The Cross Product, 657
10.5 Lines and Planes in Space, 665
10.6 Vector Methods for Measuring Distance in R3, 675
GROUP RESEARCH PROJECT: “Star Trek Project,” 683
Contents ix
zι Tangent
11 Vector Calculus 684
vector 11. 1 Introduction to Vector Functions, 685
11. 2 Differentiation and Integration of Vector Functions, 692
11. 3 Modeling Ballistics and Planetary Motion, 701
11. 4 Tangent and Normal Vectors; Curvature, 710
11. 5 Tangential and Normal Components of Acceleration, 721
GUEST ESSAY: “The Stimulation of Science,” by Howard
Eves, 733
≡≡ BOOK REPORT: Hypatia's Heritage, 736
12 Partial Differentiation 838

12.1 Functions of Several Variables, 839
Functional values along 12.2 Limits and Continuity, 847
Λ-axis or y-axis are 0.
12.3 Partial Derivatives, 756
12.4 Tangent Planes, Approximations, and Differentiability, 764
12.5 Chain Rules for Functions of Several Variables, 774
12.6 Directional Derivatives and the Gradient, 780
12.7 Extrema of Functions of Two Variables, 793
Functional values along 12.8 Lagrange Multipliers, 805
y = x are 1. Chapter 12 Review, 814
GROUP RESEARCH PROJECT: “Desertification,” 819
13 Multiple Integration 821

13.1 Double Integration Over Rectangular Regions, 822
13.2 Double Integration Over Non-rectangular Regions, 830
13.3 Double Integration in Polar Coordinates, 838
13.4 Surface Area, 847
13.5 Triple Integrals, 852
13.6 Mass, Moments, and Probability Density Functions, 865
13.7 Cylindrical and Spherical Coordinates, 875
13.8 Jacobians: Change of Variables, 885
M GROUP RESEARCH PROJECT: “Space-Capsule Design,” 896
X Contents
14 Vector Analysis 897

Properties of a Vector Field: Divergence and Curl, 898
Line Integrals, 906
Independence of Path, 916
Green’s Theorem, 924
Surface Integrals, 933
Stokes’ Theorem, 940
Divergence Theorem, 948
GUEST ESSAY: “Continuous vs Discrete Mathematics,” by
William F. Lucas, 966
BOOK REPORT: The Mathematical Experience, 968
Appendices, A -l
A Theorems by Chapter, A -l
B Selected Proofs, A-26
C Significant Digits, A-36
D Short Table of Integrals, A-41
E Answers to Odd-Numbered Problems, A-49
F Credits list, A-104
G Index, A-107
The rotunda of City Hall, in Manhattan, New York City. Designed by John McComb, Jr.,
reputedly the first American-born architect, and built in the early nineteenth century in the
neoclassical style, (photo credit: Behrenholtz)
xi
About the Authors
Gerald L. Bradley Karl J. Smith
Gerald L. Bradley received his M.S. in mathemat Karl J. Smith received his B.A. and M.A. (in
ics from Harvey Mudd College in 1962 and his 1967) degrees in mathematics from UCLA. Then
Ph.D. from the California Institute of Technology he moved in 1968 to northern California to teach
in 1966. A former NSF fellow, Professor Bradley at Santa Rosa Junior College, where he has been
has taught at Claremont McKenna College since ever since. Along the way, he served as depart
1966 and has served as chairman of the mathe ment chair, and he received a Ph.D. in 1979 in
matics department for two tours of duty. His field mathematics education at Southeastern Universi
of research has been matrix theory. He is the ty. A past president of the American Mathemati
author of A Primer o f Linear Algebra and co cal Association of Two-Year Colleges, Professor
author of a top-selling business calculus text. Smith is very active nationally in mathematics
education. He was founding editor of Western
AM ATYC News, a chairperson of the committee
on Mathematics Excellence, and a NSF grant
reviewer. He was a recipient in 1979 of an Out
standing Young Men of America Award, in 1980
of an Outstanding Educator Award, and in 1989
of an Outstanding Teacher Award. Professor
Smith is the author of several successful text
books. In fact, over one million students have
learned mathematics from his textbooks.
Preface
Calculus teaching is undergoing great changes, most of which will be of lasting

benefit to students. This text was written to blend many of the best aspects of
calculus reform with the reasonable goals and methodology of traditional calculus.
In incorporating so much of calculus reform, we made a deliberate effort not to
“ throw the baby out with the bath water.” Calculus should not be a terminal
course but rather one that prepares students of engineering, science, and math to
move on to more advanced and necessary professional courses. This text presents
most traditional topics, such as continuity, the mean value theorem, ΓHδpitaΓs
rule, parametric equations, polar coordinates, sequences, and series. In short, this
text is an attempt at Reform with Reason.
The major issue driving calculus reform is the poor performance of students
trying to master the concepts o f calculus. Much of this failure can be attributed to
ways in which students study and learn mathematics in high school. The high
school mathematics textbooks published over the last twenty years have increas
ingly focused on reducing mathematics to a series of small repeatable steps.
Process is stressed over insight and understanding. In these same books we often
find problems matched to worked-out sample problems, which encourages stu
dents to memorize a series o f problem-solving algorithms. Unfortunately, this
reduces mathematics to taxonomy; students show up for calculus generally willing
to work hard, but for many, this translates into working hard at rote memoriza
tion! The task of getting students to think conceptually falls to the teachers of
calculus, and we have provided you with a calculus book that will help you focus
on learning by providing you with sound development of concepts, challenging
problems, and a well-developed pedagogy.
Besides developing some minimal skills in algebraic manipulation and problem

Conceptual solving, a calculus text today should require students to develop verbal skills in a
Understanding mathematical setting. This is not just because real mathematics wields its words
precisely and compactly, but because verbalization should help students think
Through Verbalization conceptually.
Cooperative Learning (Group Research Projects) There is a great deal of interest

in encouraging students to work in small groups as an alternative mode of teaching
from the traditional lecture method. In July 1991, the National Science Founda
tion funded a group of instructors who met on the campus of New Mexico State
University to discuss the topic “ Discovering Calculus Through Student Projects.”
We believe that students should be encouraged to think independently and in
small groups. Students need to be put into classroom situations that help to
xiii
xiv Preface
prepare them for a working environment. They need to be able to read and write
mathematics and to ask whether an answer is reasonable without looking in the
back o f the book for an answer. To that end, we provide several examples of group
research projects. We are indebted to those who have constructed and class tested
these projects, and we thank them for their contribution to our book. Each group
research project is loosely tied to the material contained in the chapter in which it
is presented. Also note that the complexity of the projects increases as we progress
through the book and the mathematical maturity of the student develops. In
addition to doing the group research projects, the students could be encouraged to
collaborate on the Putnam and journal problems.
M athem atical Communication Students should be encouraged to communicate

mathematically, and in addition to the group research projects we have included
other opportunities for mathematical communication in terms that can be
understood by nonprofessionals. Shorter problems encouraging written communi
cation are included in nearly all of the problem sets. The guest essays provide
alternative viewpoints. The questions that follow are called MATHEMATICAL
ESSAYS and are included to encourage individual writing assignments and mathe
matical exposition. We believe that students will benefit from individual writing
and research in mathematics. Theorems and definitions are frequently followed
by short, intuitive explanations, and students are encouraged to provide their own
verbal explanations of mathematics in the problem sets, where they are asked to
summarize procedures and processes or to describe a mathematical result in
everyday terms. Mathematics is more than working problems and getting answers,
and mathematics education must include the communication of mathematical
ideas.
Computational Windows Reform is driven partly by the need to maximize the

Integration of benefits technology brings to the learning of mathematics. Simply adding a lab
Technology course to the traditional calculus is probably pointless and may lead to unaccepta
ble workloads for all involved. Instead, we choose to include technology as an aid
to the understanding of calculus, rather than to write a calculus course developed
around the technology. Although we have included over 60 pages devoted to the
use o f technology, we have kept such references “ platform-neutral” because
specific calculators and computer programs frequently change and are better
considered in separate technology manuals. The technology in the book is
organized under the title COMPUTATIONAL WINDOW to give insight into how
technological advances can be used to help understand calculus, COM PUTATIONAL
WINDOWS also appear in the student exercises: These WINDOWS contain problems
requiring a graphing calculator or software and computer. Many of the problems
were prepared by Jack Cohen and Frank Hagin, both of the Colorado School of
Mines.
Technology Laboratory M anuals For those students with access to a computer or

graphing calculator, companion Technology Manuals are available and can be
wrapped with the text at a small charge. The manuals differ from each other by
being “ platform specific” — that is, the key stroke instruction is specific to that
technology. The five different platforms (manuals) are: Texas Instruments graph
ing calculators, Hewlett-Packard graphing calculators, MATLAB, Maple, and
Mathematica. Each manual picks up the same technology examples from this
textbook and provides the specific keystrokes needed. Also, each manual has
additional technology questions (which are the same for each manual), some of
Preface XV
which are given with the keystrokes. Thus, the instructor need not be involved in
teaching the students basic key strokes or syntax in a given technology: the
appropriate m anual will do so. All of these manuals were prepared by Jack Cohen
and Frank Hagin, both of the Colorado School of Mines.
G reater Text Visualization Related to (but not exclusively driven by) the use of
technology is the greater use of graphs and other m athematical pictures through
out this text. Over 1,700 graphs appear— more than in nearly any other calculus
text. This increased visualization is intended to help develop greater student
intuition. Much of this visualization appears in the wide margins to accompany
the text. Its purpose is to provide explanation to supplement that of the text prose.
Additional graphs are related to the student problems, including answer art.
S ign ifican t D igitsWith the availability of calculators and computers, we were

necessarily faced with the question of how to treat significant digits in this book.
Even though our experience tells us that a thorough treatm ent of that topic is
better left to chemistry and physics classes, we have included a brief treatm ent in
Appendix C. On occasion, we show the entire calculator or computer output of 12
digits for clarity, even though such a display may exceed the requisite num ber of
significant digits. Our thanks to Neil Berger, of the University of Illinois at Chicago,
for his contribution to our treatm ent of significant digits and to Appendix C.
Problem s We believe that students learn mathematics by doing mathematics.

Problem Solving Therefore, the problems and applications are perhaps the most im portant feature
of any calculus book, and you will find that the problems in this book extend from
routine practice to challenging. The problem sets are divided into A Problems
(routine), B Problems (requiring independent thought), and C Problems (theory
problems). In this book we also include past Putnam examination problems as
well as problems found in current mathematical journals. The worked out
examples in the text are A and B problems. We believe very strongly that success in
understanding calculus depends on having students struggle with a healthy
num ber of challenging problems, which includes the journal and Putnam prob
lems as well as the C problems.
You will find the scope and depth of the problems in this book to be
extraordinary. Even though engineering and physics examples and problems play
a prom inent role, applications from a wide variety of fields— such as economics,
ecology, psychology, and sociology— are included. The problems have been in the
developmental stages for more than ten years, and most of them have been class
tested. In addition, the chapter summaries provide not only topical review, but
also many miscellaneous exercises. Although the chapter reviews are typical of
examinations, the miscellaneous problems are presented not as graded problems,
but rather as a random list of problems loosely tied to the ideas of the chapter. In
addition, the cumulative reviews occur at natural breaking points in the text: at
the ends of Chapters 6, 11, and 14.
Journal Problem s In an effort to show that m athematicians work problems too,

we have reprinted problems from leading m athematics journals. We have chosen
problems that are within reach of the intended audience of this book. If students
need help or hints for these problems, they can search out the original presentation
and solution in the cited journal. In addition, we have included problems from
various Putnam Examinations. These problems, which are more challenging, are
offered in the miscellaneous problems at the end of various chapters and are
provided to give insight into the type of problems asked in mathematical
xvi Preface
competitions. The Putnam Examination is a national annual examination given

under the auspices of the Mathematical Association of America and is designed to
recognize mathematically talented college and university students.
Think Tank Problems It has been said that mathematical discovery is directed
toward two major goals: the formulation of proofs and the construction of
counterexamples. Most calculus books focus only on the first goal (the body of
proofs and true statements), but we think that some attention should be paid to
the formulation of counterexamples for false statements. Throughout this book we
ask the student to formulate examples satisfying certain conditions. We have
designated this type of problem as a think tank problem.
Student Mathematics Handbook We begin Calculus with a minimum of review.

Additional Features The prerequisite material that is often included in a calculus textbook has been
bound separately in a companion book, Student Mathematics Handbook. We
think this is an important supplement to the textbook because we have found that
most errors our students make in a calculus class are not errors in calculus but
errors in basic algebra and trigonometry. A unified and complete treatment of this
prerequisite material, easily referenced and keyed to the textbook, has been a
valuable tool for our students taking calculus. Those portions of the text that
benefit from an appropriate precalculus review are marked by the symbol
l∖S M~HJ) . Thus, the review portion of the handbook functions as a series of “ help
screens” that appear as the student reads the text. Our handbook is offered free of
charge with every new copy of the textbook. The handbook not only includes the
necessary review material and formulas but also contains a catalog of curves and a
complete integral table.
Differential equations We also include differential equations early (in Section

4.4) and then develop methods of solutions of differential equations in a spiral
manner throughout the book. The reasons are twofold. First, differential equa
tions make some of the best applications for calculus. Second, learning the
differential equations is easier when they are seen in their most appropriate
context. The multivariable paperback version of the text has an additional chapter
on differential equations for those instructors who choose to teach a more
substantial course.
Sequence o f topics Some colleges and universities prefer reversing Chapters 5 and
6. For this reason, we have written these chapters so that they may be interchanged
without causing any difficulties. We resisted the temptation to label certain
sections as optional, because that is a prerogative of individual instructors and
schools. However, the following sections could be skipped without any difficulty:
1.8, 3.8, 5.7, 6.3 (delay until Sec. 13.6), 7.6, 12.8, and 13.8. Differential equations
(4.4 and 6.4) could be delayed for a subsequent course. To assist instructors with
the pacing of the course, we have written the material so each section can
reasonably be covered in one classroom day.
Proofs All the theorems are repeated in Appendix B for easy reference, but it is
important to note that we do not pretend to prove every theorem in this book. In
fact, we often refer the reader to Appendix C for some longer proofs, or sometimes
to an advanced calculus text. Why then do we include the heading PROOF after
each theorem? It is because we want the student to know that for a result to be a
theorem there must be a proof. We use the heading not necessarily to give a
complete proof in the text, but to give some direction to where a proof can be
found or an indication of how it can be constructed.
Preface xvii
B A Student Survival Manual by Ken Seydel offers a running commentary of

Supplementary hints and suggestions to help assure the students’ success in calculus. In addition,
Materials this manual contains detailed solutions to most odd-numbered problems in the
book.
I Student Mathematics Handbook by Karl J. Smith offers a review of prerequi
site material, a catalog of curves, and a complete integral table. This handbook is
presented free of charge along with the purchase of a new book.
■ Technology Manuals by Jack Cohen and Frank Hagin offer technology-based
applications keyed to the sections in the book. These manuals are identical except
for the specific keystrokes. The following manuals are available at low cost when
wrapped with the text:
Calculus Explorations with TI Calculators
Calculus Explorations with HP Calculators
Calculus Explorations with MATLAB
Calculus Explorations with Maple
Calculus Explorations with Mathematica
Each of these manuals is closely keyed to the text. Selected computational
windows from each text chapter are repeated in the manual with the appropriate
key strokes provided. Additional technology problem sets and projects are also
provided. Students could pick the technology manual that matches their technolo
gy choice and all students would still receive the same calculus course.
■ Instructor’s Technology Manual by Jack Cohen and Frank Hagin. Solutions to
the Technology Manuals problem sets. Available free to instructors upon request.
■ A Complete Solutions Manual by Henri Feiner, available free to instructors
upon text adoption, contains a brief solution for every problem in the book.
■ Instructor’s Guide offers sample tests and reviews for each chapter in the
book. This guide also includes sample transparency masters.
■ Computerized Testing Program is available in both IBM and Macintosh
formats.
■ Visual Calculus, by David Schneider (University of Maryland), is a superb
piece of software that offers a highly student-friendly approach to graphics and is
closely keyed to the text.
Bl Resources for Calculus, Volumes 1-5, A. Wayne Roberts (Project Director) is
available from the Mathematical Association of America, 1993:
Dudley, Underwood (ed), Readings for Calculus, MAA Notes, No. 31
Fraga, Robert (ed), Calculus Problems for a New Century, MAA Notes, No. 28
Jackson, Michael B., and Ramsay, John, (eds) Problems for Student Investigation,
MAA Notes, No. 30
Snow, Anita E. (ed), Learning by Discovery: A Lab Manual for Calculus, MAA
Notes No. 27
Straffin, Philip (ed), Applications of Calculus, MAA Notes, No. 29
This is a valuable collection of resource materials for calculus instructors. All
material may be reproduced for classroom use.
The writing and publishing of a calculus book is a tremendous undertaking. We

Acknowledgements take this responsibility very seriously, because a calculus book is instrumental in
transmitting knowledge from one generation to the next. We would like to thank
the many people who helped us in the preparation of this book. First, we thank our
editor George Lobell, who led us masterfully through the development and
xviii Preface
publication of this book. Our sincere appreciation to development editor Alan

MacDonell, who read and critiqued each word of the manuscript, pored over each
and every figure, often several times, and worked tirelessly on the project (except
when he went to Paris!). We also appreciate the work of production editor Ed
Thomas, who kept us all on track. Ed’s meticulous attention to detail was
extraordinary. We would also like to thank Priscilla McGeehon, who helped a
great deal in the early stages, and Ray Mullaney, director of development for
engineering, math, and science textbooks.
Of primary concern is the accuracy of the book. We had the assistance of
many: Henri Feiner, who worked all of the problems (sometimes more than once),
Ken Seydel, who offered us many valuable suggestions, Diana Gerardi, who read
every word and checked the accuracy of the problems, and Kurt Norlin and Terri
Bittner at Laurel Technical and Mary Toscano at Toscano, who provided us a
double back-up check of all of the problems.
Members o f the focus group: We would like to acknowledge and thank each
member of the focus group who met with us one weekend in San Francisco to
critique the manuscript on a page-by-page basis. Wc are convinced that the book is
better because of their contributions.
Neil Berger, University of Illinois at Chicago
Dan Chiddix, Ricks College
Lawrence J. Kratz, Idaho State University
John C. Michels, Chemeketa Community College
Reviewers o f this edition:

Neil Berger, University of Illinois at Chicago
Michael L. Berry, West Virginia Wesleyan College
Barbara H. Briggs, Tennessee Technical University
Robert Broschat, South Dakota State University
Robert D. Brown, University of Kansas
Dan Chiddix, Ricks College
Philip Crooke, Vanderbilt University
Ken Dunn, Dalhousie University
John H. Ellison, Grove City College
William P. Francis, Michigan Technological University
Harvey Greenwald, California Polytechnic San Luis Obispo
Richard Hitt, University of South Alabama
Joel W. Irish, University of Southern Maine
Clement T. Jeske, University of Wisconsin-Platteville
Lawrence Kratz, Idaho State University
Sam Lessing, Northeast Missouri University
Estela S. Llinas, University of Pittsburgh at Greensburg
Pauline Lowman, Western Kentucky University
William E. Mastrocola, Colgate University
Philip W. McCartney, Northern Kentucky University
E.D. McCune, Stephen F. Austin State University
John C. Michels, Chemeketa Community College
Pamela B. Pierce, College of Wooster
Connie Schrock, Emporia State University
Tatiana Shubin, San Jose State University
Tingxiu Wang, Oakton Community College
Gerald L. Bradley
Karl J. Smith
CALCULUS
Preview of Calculus:
Functions and Limits
PREVIEW
■ CONTENTS This chapter introduces you to those concepts that are essential for the
1.1 What Is Calculus?
The limit: Zeno’s paradox; The study of calculus. We begin by asking the question, “What is Calculus?”
derivative: the tangent problem; We shall see that calculus is used to model many aspects of the world
The integral: the area problem;
Mathematical modeling about us, so we spend a little time discussing what is meant by
1.2 Preliminaries
mathematical modeling. The necessary prerequisites for this course are
Distance on a number line; reviewed in a companion book, Student Mathematics Handbook and
Absolute value; Distance in the Integration Table for CALCULUS, and are briefly discussed in this
plane; Trigonometry
chapter.
1.3 Lines in the Plane
Slope of a line; Forms for equa Most ideas in calculus have useful geometric interpretations and
tions of lines; Parallel and per can be visualized in terms of functions. Therefore, we complete this
pendicular lines
preliminary chapter by discussing the basic properties of functions
1.4 Functions and Their Graphs
Definition of a function; and by showing how functions can be represented geometrically. An
Functional notation; Domain essential feature of calculus involves making “infinitesimally small”
and range of a function; changes in a quantity, and we give precise meaning to this notion by
Composition of functions;
Graph of a function; introducing and exploring the limit of a function and a related concept
Transformation of functions; known as continuity.
Classification of functions
1.5 The Limit of a Function
Intuitive notion of a limit;
Limits by graphing; Limits by
table; Limits that do not exist PERSPECTIVE
1.6 Properties of Limits
Computation with limits; Using Change is a fact of our daily lives. Physical scientists use mathematical models
algebra to find limits; Limits of to investigate phenomena such as the motion of planets, the decay of radioac
piecewise-defined functions tive substances, chemical reaction rates, ocean currents, and weather patterns.
1.7 Continuity Economists and business managers examine consumer trends; psychologists
Intuitive notion of continuity; study learning tendencies; and ecologists explore patterns of pollution and
Definition of continuity;
Continuity theorems; population changes involving complex relationships among species. Even areas
Continuity on an interval; The such as political science and medicine use mathematical models in which
intermediate value theorem; change is the key ingredient.
Approximation by the bisection
method Although modern science requires the use of many different skills and
1.8 Introduction to the Theory
procedures, calculus is the primary mathematical tool for dealing with change.
of Limits Sir Isaac Newton, one of the discoverers of calculus, once remarked that to
Limit of a function (formal def accomplish his results, he “stood on the shoulders of giants.” Indeed, calculus
inition); The believer/doubter was not born in a moment of divine inspiration, but developed gradually, as a
format; Epsilon-delta proofs;
Selected theorems with formal variety of apparently different ideas and methods merged into a coherent
proofs pattern. The purpose of this initial chapter is to lay the foundation for the
Chapter 1 Review development of calculus.
Guest Essay, John Troutman
1
2 Ch. 1 Preview of Calculus: Functions and Limits
1.1 WHAT IS CALCULUS?
IN THIS SECTION The limit: Zeno’s paradox, the derivative: the tangent
problem, the integral: the area problem, mathematical modeling
We informally introduce you to the three main topics of calculus: the concepts
of limit, derivative, and integral.
ELEMENTARY In the article “The Calculus According to Newton and Leibniz,” C. H. Edwards
MATHEMATICS writes
If there is an event that marked the coming of age of mathematics in Western culture, it
must surely be the essentially simultaneous development of the calculus by Newton and
Leibniz in the seventeenth century.* Before this remarkable synthesis, mathematics had
often been viewed as merely a strange but harmless pursuit, indulged in by those with an
excess of leisure time. After the calculus, mathematics became virtually the only accept
able languagefor describing the physical universe. This view of mathematics and its asso
ciation with the scientific method has come to dominate the Western view ofhow the world
ought to be explained. This domination is so complete that it is virtually impossible for
us to understand how earlier cultures explained what happened around them.↑
What distinguishes calculus from your previous mathematics courses of algebra,
geometry, and trigonometry is the transition from static or discrete applications
(see Figure 1.1) to those that are dynamic or continuous (see Figure 1.2). For
example, in elementary mathematics you considered the slope of a line, but in
calculus we define the (nonconstant) slope of a nonlinear curve. In elementary
mathematics you found average changes in quantities such as the position and
velocity of a moving object, but in calculus we can find instantaneous changes in
the same quantities. In elementary mathematics you found the average of a finite
collection of numbers, but in calculus we can find the average value of a function
3. Area of a region bounded by with infinitely many values over an interval.
line segments You might think of calculus as the culmination of all of your mathematical
studies. To a certain extent that is true, but it is also the beginning of your study of
mathematics as it applies to the real world around us. Calculus is a three-semester
or four-quarter course that begins your college work in mathematics. All your
prior work in mathematics is considered elementary mathematics, with calculus
the dividing line between elementary mathematics and mathematics as it is used
in a variety of theoretical and applied topics. It is the mathematics of motion and
change.
The development of calculus in the seventeenth century by Newton and
Leibniz was the result of their attempt to answer some fundamental questions
4. Average changes in position about the world and the way things work. These investigations led to two
and velocity fundamental concepts of calculus—namely, the idea of a derivative and that of an
integral. The breakthrough in the development of these concepts was the formula
5. Average of a finite collection
tion of a mathematical tool called a limit.
of numbers
1. Limit: The limit is a mathematical tool for studying the tendency of a
Figure 1.1 Topics from elementa function as its variable approaches some value. Calculus is based on the
ry mathematics concept of limit. We introduce the limit of a function informally in Section
1.5 and then examine the concept more formally in Section 1.8.
2. Derivative: The derivative is defined as a limit, and it is used initially to
*For a brief history of calculus see the Guest Essay at the end of this chapter.
↑C. H. Edwards, Jr., “The Calculus According to Newton and Leibniz,” in The Historical
Development o f the Calculus (New York: Springer-Verlag, 1979).
Sec. 1.1 What Is Calculus? 3
compute rates of change and slopes of tangent lines to curves. The study of
derivatives is called differential calculus. Derivatives can be used in
sketching graphs and in finding the extreme (largest and smallest) values of
functions. The derivative is introduced and developed in Chapter 2, and its
applications are examined in Chapter 3.
3. Integral: The integral is found by taking a special limit of a sum of terms,
and the study of this process is called integral calculus. Area, volume, arc
length, work, and hydrostatic force are a few of the many quantities that
can be expressed as integrals. Integrals and their applications are studied in
CALCULUS Chapters 4 and 6.
1. Slope of a curve We will briefly describe each of these concepts later in this section. First, let us
return to Edwards’s description about the foundation of calculus:
One might naturally suppose that an event so momentous must involve ideas so profound
that average mortals can hardly hope to comprehend them. In fact, nothing could be
further from the truth. The essential ideas of calculus—the derivative and the
integral—are quite straightforward and had been known prior to either Newton or
Leibniz. The contribution of Newton and Leibniz was to recognize that the idea of
finding tangents (the derivative) and the idea offinding areas (the integral) are related
and that this relation can be used to give a simple and unified description of both
processes. When calculus is described in this way one might wonder what all the fuss is
about.
2. Tangent line to a general
curve Let us begin by taking an intuitive look at each of these three essential ideas of
calculus.
■ THE LIMIT: ZENO'S PARADOX

In the guest essay at the end of this chapter, John Troutman mentions Zeno’s
paradoxes, which are concerned with infinite processes. Zeno (ca. 500 B.C.) was a
Greek philosopher who is known primarily for his famous paradoxes. One of these
3. Area of a region bounded by concerns a race between Achilles, a legendary Greek hero, and a tortoise. When
curves
4. Instantaneous changes in po

sition and velocity α 2 al
5. Average of an infinite collec Figure 1.3 Achilles and the tortoise

tion of numbers
Is it possible for Achilles to overtake the tortoise? Zeno pointed out that by the
Figure 1.2 Topics from calculus time Achilles reaches the tortoise’s starting point, a l = t0, the tortoise will have
moved ahead to a new point Z1. When Achilles gets to this next point, a2, the
tortoise will be at a new point t2 . The tortoise, even though much slower than
Achilles, keeps moving forward. Although the distance between Achilles and the
tortoise is getting smaller and smaller the tortoise will apparently always be ahead.
Of course, common sense tells us that Achilles will overtake the slow tortoise,
but where is the error in reasoning? The error is in the assumption that an infinite
Historical Note amount of time is required to cover a distance divided into an infinite number of
The numeration system we use
segments. This discussion is getting at an essential idea in calculus, the notion of a
evolved over a long period of
time. It is often called the limit.
Hindu-Arabic system because Consider the successive positions for both Achilles and the tortoise:
its origins can be traced back to Starting position
the Hindus in Bactria (now Af
ghanistan). Later, in A.D. 700,
Achilles-. a0 , a }, a2, a3, a4 , . . .
India was invaded by the Ar
abs, who used and modified the Tortoise: t0, t l , t2, t3, t4, . . .
Hindu numeration system and After the start, the positions for Achilles, as well as those for the tortoise, form sets
in turn introduced it to Western of positions that are ordered with positive integers. Such ordered listings are called
civilization. The Hindu
sequences.
Brahmagupta stated the rules
for operations with positive and
For Achilles and the tortoise we have two sequences {α1, a2, a3, . . . , an, . . .}
negative numbers in the sev and {Z1, t2, t3, . . . , tn, . . .} where an < tn for all values of n. We will see in Chapter 8
enth century A.D. There are that both the sequence for Achilles’ position and the sequence for the tortoise’s
some indications that the Chi position have limits, and it is precisely at that limit point that Achilles overtakes
nese had some knowledge of the tortoise. The idea of limit is introduced in this chapter and is used to define
negative numbers as early as the other two basic concepts of calculus: the derivative and the integral. Even if the
200 B.C. On the other hand, solution to Zeno’s paradox using limits seems unnatural at first, do not be
the Western mathematician discouraged. It took over 2000 years to refine the ideas of Zeno and provide
Girolamo Cardan (1501-1576) conclusive answers to those questions about limits that will be introduced in this
was calling numbers such as chapter. We revisit Zeno’s paradox in Problem Set 8.2.
( - 1) absurd as late as 1545.
EXAM PLE 1 An intuitive preview of a limit
The sequence
1 2 3 4
2 ’ 3’ 4 ’ 5’
can be described by writing a general term: —qyγ where n = 1, 2, 3, 4, . . . .

Can you guess the limit, L, of this sequence?
Solution
PREVIEW: The limit is an important idea in calculus, and we discuss this
concept extensively later in this chapter. We will say that L is the number that
the sequence with general term n tends toward as n becomes large without
bound. We will define a notation to summarize this idea:
At each point P on a circle, there
is only one line that intersects the
circle exactly once.
As you consider larger and larger values for n you find a sequence of
fractions:
1 2 3 1,000 1,001 9,999,999
2 3’ 4 ' ’ 1,001’ 1,002’ ' ' ' ’ 10,000,000’
It is reasonable to guess that the sequence of fractions is approaching the

number 1. M M
■ THE DERIVATIVE: THE TANGENT PROBLEM

At a point P on a curve, there
may be several lines that intersect A tangent line (or, if the context is clear, simply tangent) to a circle at a given point
the curve only once. P is a line that intersects the circle at P and only at P. This characterization does
Figure 1.4 Tangent line not apply for curves in general, as you can see by looking at Figure 1.4.
In order to find a tangent line, begin by considering a line that passes through two
points on the curve, as shown in Figure 1.5a. This line is called a secant line.
The coordinates of the two points P and Q are P(α,∕(α)) and Q(a + h,f(a + h)).
The slope of the secant line is
f(a + h )-f(a )
h
Recall from algebra that the slope of any line is defined to be rise/run. Can you see
how this formula for the slope of the secant line fits the model of “rise/run”? Now
imagine that Q moves along the curve toward P, as shown in Figure 1.5b. You can
see that the secant line approaches a limiting position as h approaches zero. In
Section 2.1 we define this limiting position to be the tangent line. The slope of the
tangent line is defined as a limit of the sequence of slopes of a set of secant lines.
PREVIEW: Once again, we can use limit notation to summarize this idea: We
say that the slopes of the secant lines as h becomes small, tend toward a
number which we call the slope of the tangent line. We will define the
following notation to summarize this idea:
lim ∕(Q + A) ~ /(a)

Λ→O h
■ THE INTEGRAL: THE AREA PROBLEM
You probably know the formula for the area of a circle with radius r:
A = πr2
The Egyptians were the first to use this formula over 5,000 years ago, but the
Greek Archimedes (2877-212 B.C.) showed how to derive the formula for the area
of a circle by using a limiting process. Consider the area of inscribed polygons, as
shown in Figure 1.6.
Figure 1.6 Approximating the area of a circle

Even though Archimedes did not use the following notation, here is the essence of
what he did, using a method called “ exhaustion” :
Let .-43 be the area of the inscribed equilateral triangle;

A 4 be the area of the inscribed square; and
√15 be the area of the inscribed regular pentagon.
How can we find the area of this circle? As you can see from Figure 1.6, if we
consider the area of A y then A 4 , then A y . . . we should have a sequence of areas
such that each successive area more closely approximates that of the circle. Later
in this book, we will write this idea as a limit statement:
A = lim A n
n→∞
In this course we will use limits in yet a different way to find the area of
regions enclosed by curves. For example, consider the area shown in color in
Figure 1.7. We can approximate the area by using rectangles. If A n is the area of the
nth rectangle, then the total area can be approxim ated by finding the sum
A i + A + A + ∙∙∙ + A-ι + A
2 3
Figure 1.7 Area under a curve
Figure 1.8 Approximating the area using circumscribed rectangles
The area problem leads to a process called integration, and the study of
integration forms what is called integral calculus. Similar reasoning allows us to
calculate such things as volumes, the length of a curve, the average value of a
function, or the amount of work required for a particular task.
■ MATHEMATICAL MODELING
A real-life situation is usually far too complicated to be precisely and m athem ati
cally defined. When confronted with a problem in the real world, therefore, it is
usually necessary to develop a mathematical framework based on certain assump
tions about the real world. This framework can then be used to find a solution to
the real-world problem. The process of developing this body of mathematics is
referred to as mathematical modeling.
Some mathematical models are quite accurate, particularly those used in the
physical sciences. For example, one of the first models we will consider in calculus
is a model for the path of a projectile. Other rather precise models predict such
things as the time of sunrise and sunset, or the speed at which an object falls in a
vacuum. Some mathematical models, however, are less accurate, especially those
that involve examples from the life sciences and social sciences. Only recently has
modeling in these disciplines become precise enough to be expressed in terms of
calculus.
What, precisely, is a mathematical model? Sometimes, mathematical model
ing can mean nothing more than a textbook word problem. But mathematical
modeling can also mean choosing appropriate mathematics to solve a problem
that has previously been unsolved. In this book, we use the term mathematical
modeling to mean something between these two extremes. That is, it is a process
we will apply to some real-life problem that does not have an obvious solution. It
usually cannot be solved by applying a single formula.
The first step of what we call mathematical modeling involves abstraction.
How Global Climate Is
Modeled Real-world Mathematical
Abstraction
We find a good example of problem model
mathematical modeling by
looking at the work being done
with weather prediction. In the
ory, if the correct assumptions With the method of abstraction, certain assumptions about the real world are
could be programmed into a made, variables are defined, and appropriate mathematics is developed. The next
computer, along with appropri step is to simplify the mathematics or derive related mathematical facts from the
ate mathematical statements of mathematical model.
the ways global climate condi
tions operate, we would have a
model to predict the weather Real-world
Abstraction
throughout the world. In the problem
global climate model, a system
of equations calculates time
dependent changes in wind as
well as temperature and mois
ture changes in the atmosphere
and on the land. The model
may also predict alterations in
the temperature of the ocean’s
surface. At the National Center
for Atmospheric Research, they
use a CRAY supercomputer to
do this modeling.
The results derived from the mathematical model should lead us to some
predictions about the real world. The next step is to gather data from the situation
being modeled and then to compare those data with the predictions. If the two do
From Scientific American, March

1991.
not agree, then the gathered data are used to modify the assumptions used in the
model.
Mathematical modeling is an ongoing process. As long as the predictions
match the real world, the assumptions made about the real world are regarded as
correct, as are the defined variables. On the other hand, as discrepancies are
noticed, it is necessary to construct a closer and more dependable mathematical
model. You might wish to read the article from Scientific American quoted in the
margin.
PROBLEM SET 1 .1 See Student Survival Manual for hints to odd-numbered problems.
3H 2
θ 1. ■ What Does This Say?* What is a mathematical model? 11. lim 12. lim 3n 2 + 1
n2 + 2
∕7→ ∞ n→∞ 2n — 1
2. ■ What Does This Say? Why are mathematical models
13. Copy the following figures on your paper. Draw what you
necessary or useful?
think is an appropriate tangent line for each curve at the
3. ■ What Does This Say? An analogy to Zeno’s tortoise point P.
paradox can be made as follows. A woman standing in a
room cannot walk to a wall. In order to do so, she would
first have to go half the distance, then half the remaining
distance, and then again half of what still remains. This
process can always be continued and thus never ends. Draw
an appropriate figure for this problem and then present
some sort of an argument using sequences to show that the
woman will, indeed, reach the wall.
4. ■ What Does This Say? Zeno’s paradoxes remind us of
an argument that might lead to an absurd conclusion:
Suppose I am playing baseball and decide to steal second
base. To run from first to second base, I must first go half
the distance, then half the remaining distance, and then
again half o f what remains. This process is continued so
that I never reach second base. Therefore it is pointless to
steal second base. Draw an appropriate figure for this
problem, and then present a mathematical argument
using sequences to show that the conclusion is absurd.
5. Consider the sequence 0.3, 0.33, 0.333, 0.3333, . . . .
What do you think is the appropriate limit of this
sequence?
6. Consider the sequence 0.2, 0.27, 0.272, 0.2727, . . . .
What do you think is the appropriate limit of this
sequence?
7. Consider the sequence 3, 3.1, 3.14, 3.141, 3.1415,
3.14159, 3.141592, . . . . What do you think is the appro
priate limit of this sequence? θ 14. Suppose the circle in Figure 1.6 has radius 1. We know
8. Consider the sequence 1, 1.4, 1.41, 1.412, 1.41421, from the area formula that the area of the circle is
1.414213, 1.4142135, . . . . What do you think is the A = τr(l)2 = τr. Find the sequence of areas for the in
appropriate limit of this sequence? scribed polygons A y A v A y . . . and show that these areas
form a sequence of numbers that seems to have a limit π.
In Problems 9 - 12, guess the requested limits.
15. Repeat Problem 14 for a circle with radius 2. What is the
9. lim
2n 10. lim
2n
n+4 n→∞ ^3n + 1
apparent value of the new limit?
16. Calculate the sum of the areas of the rectangles shown in
Figure 1.8a.
17. Calculate the sum of the areas of the rectangles shown in
* Many problems in this book are labeled What Does This Say?
Figure 1.8b.
Following the question will be a question for you to answer in your
own words, or a statement for you to rephrase in your own words. 18. Use the results of Problems 16 and 17 to see if you can
These problems are intended to be similar to the “What This Says” make a guess about the shaded area under the curve
boxes that appear throughout the book. y = x 2 from x = 0 to x = 1.
Sec. 1.2 Preliminaries 9
1 .2 PRELIMINARIES
IN THIS SECTION Distance on a number line, absolute value, distance in

the plane, trigonometry
We begin by reviewing one- and two-dimensional coordinate systems, absolute
value, absolute value equations and inequalities, and trigonometric equations.
These topics are needed for our work in calculus.
Every mathematical book that is This section provides a quick review of some fundamental concepts and tech
worth reading must be read “back niques from precalculus mathematics. If you have recently had a precalculus
wards and forwards, ” i f I may use course you could skip over this section.
the expression. I would modify Algebra, geometry, and trigonometry are important ingredients of calculus.
Lagrange’s advice a little and say, Even though we will review many ideas from algebra, geometry, and trigonome
“Go on, but often return to
try, we will not be able to develop every idea from these courses before we use it in
strengthen your faith. ” When you
come on a hard and dreary pas calculus. For example, the law of cosines from trigonometry may be needed to
sage, pass it over; come back to it solve a problem in a section that never mentions trigonometry in the exposition.
after you have seen its importance For this reason, we have made available a separate reference manual Student
or found its importance or found Mathematics Handbook and Integration Tablefor CALCULUS, which includes the
the need for it further on. background material you will need for this course. We suggest that you keep it
George Chrystal, Algebra, Part 2 close at hand. References to this handbook are indicated by the logo ∙
(Edinburgh, 1889) S H
■ DISTANCE ON A NUMBER LINE
You are probably familiar with the set of real numbers as well as with various of its
subsets, including the counting or natural numbers, the integers, the rational
numbers, and the irrational numbers.
The real numbers can most easily be visualized by using a one-dimensional
coordinate system called a real number line, as shown in Figure 1.9.
■»------- 1------- 1------- 1--------1--------H Notice that a number a is less than a number b if it is to the left of b on a real
- 2 - 1 0 1 2
number line, as shown in Figure 1.10.
Figure 1.9 Real number line
a < b '- a b
*------- H i H— H-------1 ∙H--------*
- 2 - 1 0 1 x 2
distance of x to the origin is lxl
0 Watch for these symbols. We Figure 1.10 Geometric definition of less than
use them to warn you of com
Similar definitions can be given for a > b, a ≤ b, and a ≥ b.
mon mistakes as well as those
things you should remem The location of the number 0 is chosen arbitrarily, and a unit distance is
ber. 0 picked (meters, feet, inches, . . . ). Numbers are ordered on the real number line
according to the following order properties.
Order Properties For all real numbers a, b, and c:

Trichotomy law: Exactly one of the following is true:
a < b, a > b, or a = b
Transitive law of inequality: If a 0 and ac > be if c < 0
■ A BSO LUTE VALUE
Absolute Value The absolute value of a real number a, denoted by ∣α∣, is

a if a ≥ 0
0 ∣α∣ is N O T the number a a if a < 0

without its sign. 0
The number x is located ∣x∣ units away from 0 — to the right if x > 0 and to the
left if x < 0.
Absolute value is used to describe the distance between points on a number
line.
Distance Between Two Points The distance between the numbers x 1 and x 2 on a number line is
on a Number Line
∣χ2 - χ 1∣
0 Recall that For example, the distance between 2 and - 3 is ∣2 - (-3)∣ = 5 units.
∣x2 - x 1∣ = ∣x1 - x 2l 0 Several properties of absolute value that you will need in this course are
summarized in Table 1.1.
TA B LE 1.1 Properties of Absolute Value

Let a and b be any real numbers:
Property Comment
ι . a∖ ≥ 0 1. Absolute value is nonnegative.

2∙ ~ a ∖= ∣α∣ 2. The absolute value of a number and the absolute value of its
opposite are equal.
3. α∣2 = α2 3. If an absolute value is squared, the absolute value can be dropped
because both squares are nonnegative.
4. ab∖= ∖a∖∖b∖ 4. The absolute value o f a product is the product of the absolute values.
aL H , ba . ∩
5. b∖ ∖b∖ * ° 5. The absolute value o f a quotient is the quotient o f the absolute
values.
6. -∣α∣ ≤ α ≤ ∣α∣ 6. Any number a is between the absolute value of that number and its
opposite, inclusive.
7. l>et ⅛ ≥ 0; 7. This property is useful in solving absolute value equations.
«| = b if and
only if a = ± b
0 “p if and only if q', is used to 8. j et b > 0; 8 and 9. These are the main properties used in solving absolute value
mean that both a statement and α∣ 0;
For example, Property 8 has a∖ > b if and
two parts: only if a > b
1. If ∣α∣ < b, then or a < - b
- b < a < b. 10. a + b∖ ≤ ∣α∣ + ∣⅛∣ 10. This property is called the triangle inequality. It is used in both
theory and numerical computations involving inequalities.
2. If - b < a < b, then
∣α∣ < b. 0 Property 7 is sometimes stated as ∣α∣ = ∣b∣ if and only if a = ± b. Since ∖b∖= ± b, it
follows that this property is equivalent to Property 7. Also, Properties 8 and 9 are
true for ≤ and ≥ inequalities. Specifically, if b > 0, then
∣α∣ ≤ b if and only if - b ≤ a ≤ b
and
∣a∣ ≥ b if and only i f α ≥ ⅛ o r α ≤ - 6
A convenient notation for representing intervals on a number line is called

interval notation and is summarized in the accompanying table. Note that a solid
dot (∙) at an endpoint of an interval indicates that the endpoint is included in the
interval, while an open dot (°) indicates that the endpoint is excluded. An interval
is bounded if both its endpoints are real numbers. A bounded interval is open if it
includes neither endpoint, half-open if it includes only one endpoint, and closed if
it includes both endpoints. The symbol "∞', (pronounced infinity) is used for
intervals that are not limited in one direction or another. In particular, (-∞ , ∞)
denotes the entire number line. If this notation is new to you, please check the
cj^ Handbook for further examples. We will use interval notation to write the
$ H solutions of absolute value equations and inequalities.
Name of Interval Inequality Notation Interval Notation Graph
Closed interval a ≤ x ≤ b [α> b] a b
a ≤ x [a, ∞) *" a
x ≤ b (-∞, ⅛] ) ■*— a b
Real number line All real numbers ----------------------------------------------- ►
Absolute Value Equations Absolute value property 7 allows us to solve

absolute value equations easily. Here is an example.
E X A M P L E 1 Solving an equation with an absolute value on one side
Solve ∣2x - 6∣ = x.
Solution If 2x - 6 ≥ 0, then 2x - 6 = x or x = 6
If 2x - 6 < 0, then -( 2 x - 6) = x
-3 x = - 6
x = 2
The solutions are x = 6 and x = 2.
f— 1 ' 1
Computational Window— Graphing Calculators
You can solve equations to any reasonable degree of find X = 2 , Y = 2 and X = 6 , Y = 6 as possible points of
accuracy by graphing on a calculator or computer. intersection. Checking x = 2 and x = 6 in the original
Look at Example 1. You can graph two functions, equation, we can verify the solution. There is also a
y 1 = ∣2x —6∣ and y 2 = x , and then find their intersec key called (ZOOM) that allows you to enlarge (zoom in)
tion. For the absolute value function, look for a key or shrink (zoom out). Finally, you can press (RANGE)
labeled (ABS). Enter Y l= A B S ( 2 X - 6 ) and Y 2 = X . to reset the scale. Example 1, with a scale with
When we show calculator graphs in this book we will 0 ≤ x ≤ 10 and 0 ≤ y ≤ 10, is shown at the right. We
show them just as you will see them on a calculator, will frequently show you a graph similar to the one
without scale and without labeling. There are many you would obtain on your calculator but with some
considerations of scale, limitations on the x-axis and scale and other designations to help you relate the
the y-axis, which we will discuss from time to time in geometric solution to the algebraic one. We will do
these special computational window boxes. this without using a special computational window
The (TRACE) key allows us to find approximate format, as shown after the following example.
coordinates of points on the curves. For Example 1 we
Y ιB a bs <2X -6)
V≥BX
X n in = - 1Θ Yr∣in = - 10 X n ιrι= 0 Y∩in=0
l√nax= 10
Xnax=10 Xnax=10 Vnax=1θ
X s c l= l V s c l= l X s c l= l Y s c l= l
If you are solving an absolute value equation with absolute value on both sides
of the equation, then you still use Property 7 and proceed as shown in Example 1.
EXAM PLE 2 Solving an equation with absolute value on both sides
Solve ∣x + 8∣ = ∣3x - 4∣
Solution x + 8 = 3 χ -4 or x + 8 = -( 3 x - 4) Property 1
-2 x = -1 2 4 x = -4
x = 6 x = - 1
A graphical solution for Example 2 is shown in the margin. Notice that the
points of intersection of the graphs of the functions on the left and the right
sides represents the solution. ∣≡≡≡
The absolute value expression ∣x - a∖ can be interpreted as the distance

between x and a on a number line. An equation of the form
y = ∣x + 8∣ and y = ∣3x —4∣ are
at x = -1 and x = 6. ∣x - a∖= b
is satisfied by two values of x that are a given distance b from a when represented
on a number line. For example, ∣x - 5∣ = 3 states that x is 3 units from 5 on a
number line. Thus, x is either 2 or 8.
Geometric representation Algebraic representation

3 units 3 units
∣x —5∣ = 3
x - 5= ± 3
→ --------------- Φ----------------♦—
2 5 8 x = 5± 3
I I = 8 or 2
two values that are 3 units from 5
Absolute Value Inequalities Because ∣x - 5∣ = 3 states that the distance from x

to 5 is 3 units, the inequality ∣x - 5∣ < 3 states that the distance from x to 5 is less
than 3 units, while ∣x - 5∣ > 3 states that the distance from x to 5 is greater than 3
units.
less than 3 units less than 3 units greater than 3 units greater than 3 units
.--*-------->,--------λ--------, .-------> ,—^—.
------Φ---------------- Φ---------------- ⅛------ *■ -<--------Φ---------------- ⅛---------------- ⅞------ *
2 5 8 2 5 8
midpoint midpoint
This number-line solution is a one-dimensional interpretation of an absolute

value inequality. For a two-dimensional interpretation with a graphing calculator,
graph y l = ∣x - 5∣ and y 2 = 3, and then look at the x-values for which y l = y 2,
T 1 < y 1 ∙>or y l > y 2 , as shown in Figure 1.11.
Figure 1.11 Two-dimensional graphs for absolute value inequalities
EXAM PLE 3 Solving an absolute value inequality
Solve ∣2x - 3∣ ≤ 4.
Solution Algebraic solution: —4 ≤ 2x —3 ≤ 4 Property 8
-4 + 3 ≤ 2 x - 3 + 3 ≤ 4 + 3
-1 ≤ 2x ≤ 7
1 ≤ 2x < 7
2 “ 2 “ 2
.1 < xx < 1
2 ~ ~ 2
The solution is the interval [-∣> £|.

Geometric solution: Graph y 1 = ∣

2x - 3∣and y 2 = 4. Because we are look
ing for ∣
2x - 3∣≤ 4, we note those x-values on the real num ber line for which
the graph of y l is below the graph of y 2 . We see that the interval is [-0 .5 , 3.5]
or Γ -⅛ f l. ’ —
When absolute value is applied to measurement it is called tolerance.

Tolerance is an allowable deviation from a standard. An example of a cement bag
with a weight of w lb is “ 90 lb plus or minus 2 lb” might be described as having a
weight given by ∣ w - 90∣≤ 2. When considered as a tolerance, the expression
∣
χ - a∖ ≤ b may be interpreted as x being compared to a having an absolute error
of measurement of b units. Consider the following example.
EXAM PLE 4 Absolute value as a tolerance
Suppose a 90-lb bag of cement is purchased. It will not be exactly 90 lb. The
material must be measured, and the measurement is approximate. Some bags
will be a little over 90 lb, and some will be a little under 90 lb, perhaps by a
couple of pounds. If so, the bag could weigh as much as 92 lb and as little as 88
lb. State this as an absolute value inequality.
Solution Let w = weight of the bag of cement in pounds. Then
88 ≤ ≤ 92
88 - 90 ≤ w - 90 ≤ 92 - 90 Subtract 90 from each
member, because this is the
-2 ≤ w -9 0 ≤ 2 average o f the extremes.
Equivalently, ∣
w - 90∣≤ 2.
■ DISTANCE IN THE PLANE
Absolute value is used to find the distance between two points on a num ber line.
In order to find the distance between two points in a coordinate plane, we use the
distance form ula, which is derived by using the Pythagorean theorem.
THEOREM 1.1 Distance between two points in the plane
The distance d between the points P l (x 1, y 1) and P 2 (x 2 , y 2) in the plane is given by
d = V(∆x) 2 + (∆y)2 = V (x 2 - x ,) 2 + (y 2 - y 1) 2
where ∆x (read “ delta x ” ) is the horizontal change defined by the distance x 2 - x 1
and ∆y (read “ delta y”) is the vertical change defined by y 2 - y l .
Proof: Using the two points, form a right triangle by drawing lines through the
given points parallel to the coordinate axes, as shown in Figure 1.12. The length of
the horizontal side of the triangle is ∣
x 2 —x 1∣= ∣∆x∣and the length of the vertical
side is ∣
y 2 - y 1∣
= ∣
∆y∣ . Then
d2 = ∣ 2 + ∣
∆x∣ ∆y∣2 Pythagorean theorem
,, ,, ,, ,, ,, Absolute value property 3
d 2 = (∆x)2 + (∆y)2
Figure 1.12 Distance formula d = V(∆x) 2 + (∆y)2
Midpoint Formula Related to the distance between two points is the formula for
finding the midpoint of a line segment, as shown in Figure 1.12.
Midpoint Formula The midpoint, M, of the segment with endpoints P 1(x1,y 1) and P 2(x2,y2) has
coordinates
x 1 + x2 y l + y 2
2 ’ 2
Notice that the coordinates of the midpoint of a segment are found by averaging
the first and second components of the coordinates of the endpoints, respectively.
You are asked to derive this formula in Problem 74.
Relationship Between an Equation and a Graph Analytic geometry is that

branch of geometry that ties together the geometric concept of position with an
algebraic representation, namely coordinates. For example, you remember from
algebra that a line can be represented by an equation. Precisely what does this
mean? Can we make a statement that is true for any curve, not just for lines? We
answer in the affirmative with the following definition.
Graph of an Equation The graph of an equation, in two variables x and y is the collection of all points
P(x, y) whose coordinates (x, y) satisfy the equation.
There are two frequently asked questions in analytic geometry:
1. Given a graph (a geometrical representation), find the corresponding

equation.
2. Given an equation (an algebraic representation), find the corresponding
graph.
In Example 5 we use the distance formula to derive the equation of a circle. This
means that if x and y are numbers that satisfy the equation, then the point (x, y)
will lie on the circle. Conversely, the coordinates of any point on the circle will
satisfy the equation.
EXAMPLE 5 Using the distance formula to derive an equation of a graph.
Find the equation of a circle with center (∕z, k) and radius r.

Solution Let (x, y) be any point on a circle. Recall that a circle is the set of all
points in the plane a given distance from a given point. The given point, called
the center, is (∕z, ∕c) and the given distance, the radius, is r.
r= DISTANCE FROM (A, k) TO (x, y)
r = V(x - h)2 + (y - k)2
r2 = (x - h)2 + (y - k)2, or (x - h)2 + (y - k)2 = r2
Note that if (A, k) = (0, 0) and r = 1, then the circle is called a unit circle and
the equation is x 2 + y 2 = 1.
EXAM PLE 6 Finding the equation of a circle
Find the equation of the circle with center (3, -5 ) that passes through the
point (1, 8).
Solution See Figure 1.13. The radius is the distance from the center to the given
point:
r = V (l - 3 ) 2 + [8 -( -5 )]2
= √ 4 + 169
= √F73
Thus, the equation of the circle is
Figure 1.13 Circle with center (x - 3)2 + [y - (-5 )] 2 = (VT73)2
(3, -5) passing through (1,8)
(x - 3)2 + (y + 5)2 = 173
EXAM PLE 7 Finding the graph of a circle given its equation
Sketch the graph of the circle whose equation is
SM H 4x 2 + 4y2 - 4x + 8y - 5 = 0
Solution We need to convert this equation into standard form. To do this we use
a process called completing the square (see Student Mathematics Handbook).
4x 2 + 4y2 - 4x + 8y - 5 = 0
x 2 + y 2 - x + 2y = j Coefficients of squared terms should
be 1.
) + (y 2 + 2y ) = ¾ Associate x-terms and y-terms.
I x 2 —x + (—∣ ) j + (y 2 + 2y + 12 ) = ξ + ∣ + 1 Complete the squares by

adding ∣and 1 to both
( χ - ∣ ) 2 + ( y + l) 2 = ^ = j sides.
This is a circle with center at (⅛, - 1) and radius λ∕f∙ The graph is shown in
Figure 1.14 Sketch of
4x2 + 4y2 - 4x + 8y - 5 = 0 Figure 1.14. MHB
■ TRIGONOMETRY
One of the prerequisites for this book is trigonometry. If you need some review,
consult a text on trigonometry or the Student Mathematics Handbook that
SM H accompanies this book. Angles are commonly measured in degrees and radians. A
degree is defined to be β∣θ revolution and a radian revolution. Thus, to convert
between degree and radian measure use the following formula:
θ measured in degrees _ θ measured in radians
360 2ττ
EXAM PLE 8 Converting degree measure to radian measure
Convert 255o to radian measure.

0 When angles are measured in Solution 255 .. θ

360 2τr
calculus, radian measure is gen
erally preferable to degree mea * = (⅜)(255)
sure. For example, in Chapter 2
we will use the formula for the ≈ 4.450589593
area of a sector, which requires
that the angle be measured in
radians. You may assume that EXAM PLE 9 Converting radian measure to degree measure
when we write expressions such
Express 1 radian in terms of degrees.
as sinx, cosx, and tanx, the
angle x is in radians unless oth θ 1
Solution 360 = ⅛
erwise specified by a degree
symbol. 0 e = m (D θ in degrees
≈ 57.29577951°
Solving Trigonometric Equations There will be many times in calculus when

you will need to solve a trigonometric equation. As you may remember from
trigonometry, solving a trigonometric equation is equivalent to evaluating an
inverse trigonometric relation. Because these inverse functions will not be
introduced in this book until Section 5.6, we will, for now, solve trigonometric
equations whose solutions involve the values in Table 1.2 (called a table of exact
values). These exact values from trigonometry are reviewed in Student Mathemat
ics Handbook.
TABLE 1.2 Exact Trigonometric Values
Angle θ 0 7Γ ττ ττ π
6 4 3 2
Function
√3 √2 1
COS# 1 2 2 2 0
1 √2 √3
sin# 0 2 2 2 1
√3 √3
tan# 0 3 1 undefined
It is customary to use the values from Table 1.2 whenever possible, and the
approximate calculator values will be given only when necessary.
E X A M P L E 10 Evaluating sine, cosine, tangent, secant, cosecant, and

cotangent
Evaluate cosy, sin-^rj tan(-¾ r)j seel.2; csc(-4.5); and cotl80°.
Solution cosf = ∣ Exact value-, Quadrant I
Exact value-, Quadrant II
tan(-¾ r) = -1 Exact value-, Quadrant II

seel.2 ≈ 2.759703601 Approximate calculator value
csc(-4.5) ≈ 1.022986384 Approximate calculator value
cot 180° is not defined.
EXAM PLE 11 Solving a trigonometric equation by factoring
Solve 2 cos0 sin θ = sin θ on [0, 2τr)

Solution 2 cos θ sin θ - sin θ = 0
sin 0(2 cos0 —1 )= 0 0 Do not divide both sides by sin0,
because you might lose a solution. No
sin0 = 0 2 co s0 - 1 = 0 tice that if 0 = 0 or 77, then sin 0 = 0 . You
0 = 0, τr cos0 = I cannot divide by 0. 0
n — ιτ 5ττ
“ 3’ 3
Computational Window— Graphing Calculators
The procedure for solving trigonometric equations using calculators is quite

similar to solving equations with absolute values. For Example 11, graph
Yl = 2 COS(X) SIN(X) and Y2 = SIN(X). We see there are four points of
intersection of the curves on the interval [0, 277): By inspection we can see the
curves intersect at 0 and π. (Note that the curves also intersect at x = 277, but
that value is not in the domain.)
By using the (ZOOM) we can find answers with better accuracy. We see that the
approximate solution is X ≈ 0, 3.1, 1.0, 5.2. Y ιB 2 c o s X s in X
V≡Bs-in X
X n in = 0
X n a x = 6 .2831853...
X ≥ c l= .5
Y n in = '1 . 5
V n a x = l. 5
V s c l= .5
EXAM PLE 12 Solving a trigonometric equation using identities
Solve tanx + V3 = secx on [0, 277).
Solution tanx + V3 = secx

tan 2x + 2V3 tanx + 3 = sec2x Square both sides.
tan 2x + 2V3 tanx + 3 = 1 + tan 2x Use the identity
1 + tan 2x = sec2x to write
the equation using tan x alone.
2Vz3 tanx = - 2
tanx =
However, since we squared both sides we need to check for extraneous roots
by substituting into the original equation. Checking, we see that x = -⅛r is
There is one intersection point: x ≈ 5.8. extraneous, and the only solution is x = ,l.'7. M M
PROBLEM SET 1 .2
O 1. Fill in the missing parts in the following table. Solve each inequality in Problems 27-38, and give your answer
using interval notation.
Inequality Interval 27. 3x + 7 < 2 28. 5(3 - x) > 3x - 1
Notation Notation 29. - 5 < 3x < 0 30. 3 ≤ - y < 8
-3 < x < 4 a. 31. - 3 < y - 5 ≤ 2 32. - 5 ≤ 3 - 2x < 18
33. t2 - 2 t ≤ 3 34. s 2 + 3s - 4 > 0
b. [3, 5] x(2x - ( g 36. x(x + 5)(x - 3)
35. 0
c. [ - 2 , 1) 5 -7 r > (x + 3)(x - 4) -
37. ∣ x - 8∣ ≤ 0.001 38. ∣ x - 5∣ < 0.01
2 < x ≤ 7 d.
Convert each angle in Problems 39-42 to degrees i f it is given
in radians or to radians i f it is given in degrees.
2. Fill in the missing parts in the following table. 39. (exact) a. 30o b. 450
40. (nearest hundredth of a radian) a. 670 b. 27 50
Inequality Interval
Notation
41. (exact) a. j b.
Notation
a.
42. (nearest degree) a. 4 b. -1
(-∞ , - 2 )
43. Find the length o f the arc subtended (cut off) by an angle
b. [_≡4’ Vv ?2J of 2 radians in a circle of radius 2 cm.
44. Find the arc length (measured to the nearest inch) of the
x > -3 c. arc subtended by an angle of 30o in a circle of radius 10 in.
-1 ≤ x ≤ 5 d. Evaluate the trigonometric functions in Problems 45-48. Use

exact values when possible; otherwise round to four decimal
places.
3. Represent each of the following on a number line. 45. a. co sf b. sec≈3^ττ∙ c. ta n y r
a. - 1 ≤ x < 4 b. - 1 ≤ ∕ ≤ 3 46. a. sin(-2) b. cos ( - 1.5) c. tan 0.5
c. (0, 2) d. [ - 2 , 1) 47. a. tan 1 b. sin 3.2 c. cos 6.02
4. Represent each of the following on a number line. 48. a. c o s ( -f) b. tan^y c. cot (-¾ r)
a. (-∞ , 2) U(2, ∞) b. x > —2
In Problems 49-52, fin d an equation o f a circle with given
c. z ≤ 4 d. - 2 < x ≤ 3 or x ≥ 5 center C and radius r.
In Problems 5-8 , plot the given points P and Q on a 49. C ( - 1, 2); r = 3 50. C(3, 0); r = 2
Cartesian plane, fin d the distance between them, and 51. C(0, 1.5); r = 0.25 52. C ( - l , -5 ); r = 4.1
fin d the coordinates o f the midpoint o f the line segment PQ.
5. a. P ( l,0 ) , β(5, 0) b. P(2, 3), 2 ( - 2 , 5) Graph the circles in Problems 53-56.
53. x 2 - 2x + y 2 + 2y + 1 = 0 54. 4x 2 + 4y2 + 4y - 15 = 0
6. a. P ( - 2 , 3), 2(4, 1) b. P (-5 , 3), 2 ( - 5 , - 7 )
55. x 2 + y 2 + 2x - 10y + 25 = 0
7. a. P ( - l , 1), 2 ( 3 ,-2 ) b. P ( - 2 , - l ) , 2 ( - U - 2 )
56. 2x 2 + 2y2 + 2x - 6y - 9 = 0
8. a. P ( - 4 , 3), 2(3, -4 ) b. P ( - 5 ,- 2 ) , 2 ( - 6 , - 8 )
Use the sum and difference formulas from trigonometry to fin d
IB S o lv e each equation in Problems 9-26. Assume that a, b, and c
are known constants.
9. x 2 - x = 0 10. 2y 2 + y - 3 = 0
the exact values o f the expressions in Problems 57-60.
57. sin (-∙⅛) 58. cos⅛ 59. tan-Q 60. sin l65 o
2
11. y - 5y + 3 = 17 12. x 2 + 5x + a = 0 © 6 1 . ■ What Does This Say? Describe a process for solving a
13. 3x 2 - bx = c 14. 4x 2 + 20x + 25 = 0 quadratic equation.
15. ∣2x + 4∣ = 16 16. ∣5y + 2∣ = 12 62. ■ What Does This Say? Describe a process for solving
absolute value equations.
17. ∣3 - 2w∣ = 7 18. ∣5 - 3t∣ = 14
63. ■ What Does This Say? Describe a process for solving
19. ∣3x + 1∣ = - 4 20. ∣1 - 5x∣ = - 2
absolute value inequalities.
21. sinx = —j on [0, 2τr)
64. ■ What Does This Say? Describe a process for solving
22. (sinx)(cosx) = 0 on [0, 2ττ) trigonometric equations.
23. (2 cosx + Λ√Z2)(2 C O SX — 1) = 0 on [0, 2π) © 65. ■ What Does This Say? If ax + b = 0, what effect does
24. (3 tanx + V3)(3 tanx - V3) = 0 on [0, 2τr) changing b have on the solution (a ≠ 0)?
25. cotx + V 3 = cscx on [0, 2ττ) 66. ■ What Does This Say? If a x 2 + bx + c = 0, what effect
26. sec2 x — 1 = V 3 tanx on [0, 2ττ) does changing c have on the solution (a ≠ 0)?
67. ■ What Does This Say? If sin ax = b, what effect does 73. Show that ∣∣x∣ - ∣.r∣∣ ≤ ∣x - y∣ for all x and y.
changing a have on the solution (α ≠ 0)? 74. Derive the midpoint formula
68. If c ≥ 0, show that ∣x∣ ≤ c if and only if - c ≤ x ≤ c.
69. Show that —∣x∣ ≤ x ≤ ∣x∣ for any number x. .√ * 1 + x ι τ l + >,2j
∖ 2 2 /
70. Prove that ∣α∣ = ∣⅛∣ if and only if a = b or a = - b .
71. Prove that if ∣α∣ 0, then - b < a < b. for the midpoint of a segment with endpoints P (x 1, y 1)
72. Prove the triangle inequality: ∣x + y∣ ≤ ∣x∣ + ∣y∣ and P(x 2, y 2)-
1 .3 LINES IN THE PLANE
Historical Note IN THIS SECTION Slope of a line, forms for equations of lines, parallel
RENE DESCARTES(1596-1650) and perpendicular lines
A line is one of the simplest geometric concepts yet one of the most useful from
the standpoint of calculus. In this section we review some of the basic facts
about lines in a plane.
■ SLOPE OF A LINE
A distinguishing feature of a line is the fact that its inclination with respect to the
horizontal is constant. It is common practice to specify inclination by means of a
concept called slope. A carpenter might describe a roofline that rises 1 ft for every
3 ft of horizontal “run” as having a slope or pitch of 1 to 3.
Let ∆x and ∆y represent, as before, the amount of change in the variables x
and y, respectively. Then a nonvertical line € that rises (or falls) ∆y units measured
The Cartesian coordinate sys from bottom to top for every ∆x units of run (measured from left to right) is said
tem is named in honor of the to have a slope of m = ∆y∕∆x. {Recall'. If ∆y is negative, then the “rise” is actually a
French mathematician Rene fall; and if ∆x is negative then the run is actually right to left.) In particular, if
Descartes. Tradition tells us F(x 1, j 1) and Q{x2, y,) are two distinct points on f, then the changes in the
that he thought of his coordi variables x and y are given by ∆x = x 2 - x l and ∆y = y 2 - y l , and the slope of € is
nate system while lying in bed
watching a fly crawl around on
the ceiling of his bedroom. He m = ∆ x
ΔJk√Λ'
= χ2 - χ1 for δ x ≠ 0
A - A ∣
See Figure 1.15.
noticed that the path o f the fly
could be described if he knew
the relation connecting the fly’s
distance from each of the walls.
This idea o f a coordinate sys
tem ties together the two great
branches of mathematics, alge
bra and geometry, to form what
is known as analytic geometry.
Plane analytic geometry is that
subject devoted primarily to the
graphing of equations in two
variables and finding the equa
tions o f points in the plane.
Figure 1.15 The slope of a line

Sec. 1.3 Lines in the Plane 21
Slope of a Line A nonvertical line that contains the points P(xl , y i ) and Q(x7, y 2) has slope
m _ ∆p _ y 2 - ,v1
∆x x 2 - X1
We say that a line with slope m is rising (when viewed from left to right) if
m > 0, falling if m < 0, and horizontal if m = 0.
There is a useful trigonometric formulation of slope. The angle of inclination
of a line € is defined to be the nonnegative angle ≠(0 ≤ φ < π) formed between £
and the positively directed x-axis.
Angle of Inclination The angle of inclination of a line € is the angle φ (0 ≤ φ < ττ) between line f
and the positive x-axis. Then the slope of line € with inclination φ is
m = tan φ
We see that the line £ is rising if 0 < φ < f and is falling if f < φ < τr. The
line is horizontal if φ = 0 and is vertical if φ = f. Notice that if φ = J, tan φ is not
defined; therefore m is not defined for a vertical line. A vertical line is said to have
no slope.
0 Sometimes we say a vertical line has infinite slope. Pay special attention
to the fact that if we say a line has no slope (vertical line), that is NOT the
same as saying the line has 0 slope (horizontal line). 0
To derive the trigonometric representation for slope we need to find the slope
of the line through P(x 1,y 1) and β(x 2 , y 2), where φ is the angle of inclination. From
the definition of the tangent, we have
Figure 1.16 Trigonometric form ,
for the slope of a line tan φ = xy ι -~ Pi = ΔF
x 1 ∆x See Figure 1.16.
2
Lines with various slopes are shown in Figure 1.17.
Figure 1.17 Examples of slope
■ FORMS FOR EQUATIONS OF LINES
In algebra you studied several forms of the equation of a line. The derivations of
some of these are reviewed in the problems, and the rest are reviewed in the
Student Mathematics Handbook. Here is a summary of the forms most frequently
used in calculus.
Forms of a Linear Equation STANDARD FORM: Ax + By + C = 0 A, B, C constants

(A and B not both 0)
SLOPE-INTERCEPT FORM: y = mx + b Slope m, y-intercept (0, b)
POINT-SLOPE FORM: y - k = m(x - Λ) Slope m, through point (h, k)
HORIZONTAL LINE: y =k Slope 0
VERTICAL LINE: x =h No slope (slope is not

defined for vertical lines)
EXAM PLE 1 Deriving the two-intercept form of the equation of a line
Derive the equation of the line with intercepts (a, 0) and (0, b), a ≠ 0, b ≠ 0.
Solution The slope of the equation passing through the given points is
Use the point-slope form with h = 0, k = b. (You can use either of the given
points.)
y - b = - ∣(χ - 0)
ay - ab = -b x
bx + ay = ab
γ ^y
- +j = 1 Divide both sides by ab. ∣
= -1
Two quantities x and y that satisfy a linear equation Ax + By + C = 0 (A and

B not both 0) are said to be linearly related. This terminology is illustrated in
Example 2.
EXAM PLE 2 Linearly related variables
When a weight is attached to a helical spring, it causes the spring to lengthen.

According to Hooke’s law, the length € of the spring is linearly related to the
weight w.* If £ = 4 cm when w = 3 g and t = 6 cm when w = 6 g, what is the
original length of the spring, and what weight will cause the spring to lengthen
to 5 cm?
Solution Because € is linearly related to w, we know that points (w, €) lie on a
line, and the given information tells us that two such points are (3, 4) and
(6, 6) as shown in Figure 1.18 on page 23.
We first find the slope of the line and then use the point-slope form to
derive its equation.
m . 66 -- 34 . 23
*Hooke, s law is useful for small displacements, but for larger displacements, it may not be a good
model.
Next, substitute into the point-slope form with h = 3 and k = 4:

£ - 4 = 2(w _ 3)
£ = ∣w + 2
The original length of the spring is found for w = 0:
e = ∣(0) + 2 = 2
The original length was 2 cm. To find the weight that corresponds to € = 5, we
solve the equation
Figure 1.18 The length £ of the 5 = jw + 2

spring is linearly related to the
weight w of the attached object. 3 = jw
2= w
Therefore, the weight that corresponds to a length of 5 cm is 4.5 g.
■ PARALLEL AND PERPENDICULAR LINES
It is often useful to know whether two given lines are either parallel or
perpendicular. A vertical line can be parallel only to other vertical lines and
perpendicular only to horizontal lines. Cases involving nonvertical lines may be
handled by the criteria given in the following theorem.
a. Parallel lines have T H E O R E M 1.2 Slope criteria for parallel and perpendicular lines
equal slope: m χ = m.,
If f 1 and are nonvertical lines with slopes m χ and m 2, then
f 1 and f 2 are parallel if and only if m χ = m,;

f 1 and f 2 are perpendicular if and only if m l m 2 = - 1 , or m i = ~ ~ ∙
In other words, lines are parallel if and only if their slopes are equal and are
perpendicular if and only if their slopes are negative reciprocals of each other.
Proof: The key ideas behind these two slope criteria are displayed in Figure
1.19. The criterion for parallel lines follows from the fact that lines are parallel
whenever their angles of inclination are equal (see Figure 1.19a). On the other
hand, if f and are perpendicular (see Figure 1.19b), one angle of inclination,
b. Perpendicular lines have negative say φ,, must be an acute angle and must satisfy
reciprocal slopes: m j m2 = - 1 .
Figure 1.19 ≠1 = ≠2 + 5
Then
m l = tan <⅛l
= tan (≠ 2 + f )
= tan ⅛ - ( - ≠ 2)]
= c o t ( - ≠ 2)
= -c o t≠ 2
= -1
tanψ 2
= Ξ4
m.
Thus, m 1∕n 2 = - 1 . We leave it to you to show that f 1 and f 2 are perpendicular if
m l m2 = - 1 . ■
EXAM PLE 3 Finding equations for parallel and perpendicular lines
Let f be the line 3x + 2y = 5.

a. Find an equation of the line that is parallel to f and passes through P(4, 7).
b. Find an equation of the line that is perpendicular to € and passes through
P(4, 7).
Solution By rewriting the equation of ( as y = - ∣ x + j , we see that the slope of f

is m = - j .
a. Any line that is parallel to € must also have slope m χ = —1. The required
line contains the point P(4, 7). Use the point-slope form to find the
equation and write your answer in standard form:
Figure 1.20 Graph of lines paral y - 7 = -¾(χ - 4)

lel and perpendicular to the given
line 3x + 2y = 5 2i' - 14 = - 3x + 12
3x + 2y - 26 = 0
b. Any line perpendicular to € must have slope m 2 = j (negative reciprocal of
the slope of €). Once again, the required line contains the point P(4, 7), and
we find _ 2/
y - 7 = ∣(x 4-λ 4)
3y - 21 = 2x - 8
l x - 3y + 13 = 0
These lines are shown in Figure 1.20.
PROBLEM SET 1 .3
θ 1. ■ What Does This Say? Outline a procedure for graph 3. passing through ( -1 , 7) and ( -2 , 9)
ing a linear equation. 4. horizontal line through (—2, -5 )
In Problems 2-15, fin d the equation in standard form for the 5. passing through the point (1, ∣) with slope 0
line that satisfies the given requirements.
2. passing through (1,4) and (3, 6) 6. slope 2 and y-intercept (0, 5)
7. vertical line through (-2 , -5 ) b. If the trend continues, what will the average SAT score
8. slope —3, x-intercept (5, 0) of incoming students be in the year 2000?
9. x-intercept (7, 0) and >,-intercept (0, -8 ) c. When will the average SAT score be 527?
10. x-intercept (4.5, 0) and ^-intercept (0, -5.4) 29. A manufacturer’s total cost consists of a fixed overhead of
11. passing through (-1 , 8) parallel to 3x + y = 7 $5,000 plus production costs of $60 per unit. Assuming
the cost varies linearly with the level of production,
12. passing through (3, —2) perpendicular to 4x - 3ρ + 2 = 0
express the total cost in terms of the number of units
13. passing through (4, 5) parallel to the line passing through produced and draw the graph.
(2, 1) and (5, 9)
30. A certain car rental agency charges $40 per day with 100
14. passing through (—1, 6) perpendicular to the line through free miles plus 34Φ per mile after the first 100 miles. First
the origin with slope 0.5 express the cost of renting a car from this agency for one
15. perpendicular to the line whose equation is day in terms of the number of miles driven. Then draw the
x - 4y + 5 = 0 where it intersects the line whose equa graph and use it to check your answers to these questions,
tion is 2x + 3y - 1 = 0 a. How much does it cost to rent a car for a 1-day trip of
50 mi?
In Problems 16-23, find, i f possible, the slope, the y-intercept, b. How many miles were driven if the daily rental cost
and the x-intercept o f the line whose equation is given. Sketch was $92.36?
the graph o f each equation. 31. A manufacturer buys $200,000 worth of machinery that
16. 5x + 3y - 15 = 0 17. 3x + 5y + 15 = 0 depreciates linearly so that its trade-in value after 10 years
18. ∣+ ∣= 2 will be $ 10,000. Express the value of the machinery as a
19 1
2 3 function of its age and draw the graph. What is the value
20. y = 2x 21. x = 5y of the machinery after 4 years?
22. y - 5=0 23. x + 3 = 0 32. Show that if the point P(x, y) is equidistant from .4(1, 3)
24. Find an equation for a vertical line f such that a region and 5 (-1 , 2), its coordinates must satisfy the equation
bounded by the x-axis, and the line 2y —3x = 6 has 4x + 2y —5 = 0. Sketch the graph of this equation.
area 3. 33. Let P l (2, 6), P 2( - 1, 3), P f0 , —2), and P fa. b) be points in
25. Three vertices of a parallelogram are (1, 3), (4, 11), and the plane that are located so that P γ P2P fif is a parallelo
(3, —2). If (1, 3) and (3, -2 ) lie on the same side, what is gram.
the fourth vertex? What are the equations of the lines that a. There are three possible choices for P 4 . One is P 4 (3, 1).
form the sides of the parallelogram? What are the others?
26. A life insurance table indicates that a woman who is now b. Show that for P 4(3, 1), the diagonals bisect each other.
A years old can expect to live E years longer. Suppose that 34. Since the beginning of the month, a local reservoir has
A and E are linearly related and that E = 50 when A = 24 been losing water at a constant rate (that is, the amount of
and E = 20 when A = 60. water in the reservoir is a linear function of time). On the
a. At what age may a woman expect to live 30 years 12th of the month, the reservoir held 200 million gallons
longer? of water; on the 21st, it held only 164 million gallons.
b. What is the life expectancy of a newborn female child? How much water was in the reservoir on the 8th of the
month?
c. At what age is the life expectancy zero?
35. To encourage motorists to form car pools, the transit
27. On the Fahrenheit temperature scale, water freezes at 32o
authority in a certain metropolitan area has been offering
and boils at 212°; the corresponding temperatures on the a special reduced rate at toll bridges for vehicles contain
Celsius scale are 0o and 100o . Given that the Fahrenheit ing four or more persons. When the program began 30
and Celsius temperatures are linearly related, first find days ago, 157 vehicles qualified for the reduced rate
numbers r and s so that F = rC + .s, and then answer these during the morning rush hour. Since then, the number of
questions. vehicles qualifying has increased at a constant rate (that
a. Mercury freezes at —39oC. What is the corresponding is, the number is a linear function of time), and 247
Fahrenheit temperature? vehicles qualified today. If the trend continues, how many
b. For what value of C is F = 0? vehicles will qualify during the morning rush hour 14
c. What temperature is the same in both scales? days from now?
28. The average SAT mathematics scores of incoming stu 36. The value of a certain rare book doubles every 10 years.
dents at an eastern liberal arts college have been declining The book was originally worth $3.
in recent years. In 1987 the average SAT score was 575, a. How much is the book worth when it is 30 years old?
and in 1992 it was 545. Assuming the SAT score varies When it is 40 years old?
linearly with time, answer these questions. b. Is the relationship between the value of the book and
a. Express the average SAT score in terms of time. its age linear? Explain.
θ 37. Show that, in general, a line passing through P(h, k) with

slope m has the equation y - k = m ix — h).
38. If A (x i , j>1) and B(x2, y 2) with x 1 ≠ x 2 are two points on the
graph of the line y = m x + b, show that
y-ι - >’i
m = ---------
Use this to show that the graph of y = m x + b is a line

with slope m; show that the line has j-intercept (0, b~).
39. Show that the distance 5 from the point (x0 , y 0) to the line
A x + By + C = 0 is given by the formula
Ax 0 + By 0 + C
5=
fA 2 + B2 Figure 1.21 Angle φ between given lines
40. Let L χ and L 2 have slopes m l and m 2, respectively, and let

φ be the angle from L i and L 2 , as shown in Figure 1.21.
Show that
41. J OURNAL PROBLEM*: Ontario Secondary School Mathe
∕∙n, - m. matics Bulletin. Show that there is just one line in the
tan φ = -i- =-------- family y - 8 = m(x + 1) that is five units from the (2, 4).
1 + m l m2
1.4 FUNCTIONS AND THEIR GRAPHS
IN THIS SECTION Definition of a function, functional notation, domain

and range of a function, composition of functions, graph of a function,
transformation of functions, classification of functions
The concept of function is the backbone of a calculus course. A thorough
understanding of this section is essential for your work in calculus.
Scientists, economists, and other researchers study relationships between quanti

ties. For example, an engineer may need to know how the illumination from a light
source on an object is related to the distance between the object and the source; a
biologist may wish to investigate how the population of a bacterial colony varies
with time in the presence of a toxin; an economist may wish to determine the
relationship between demand for a certain commodity and its market price. The
mathematical study of such relationships involves the concept of a function.
■ DEFINITION OF A FUNCTION
Function A function/ i s a rule that assigns to each element x of a set X a unique element
y of a set Y. The element y is called the image of x under f and is denoted by
f(x) (read as “/ of x”). The set X is called the domain of f and the set of all
images of elements of X is called the range of the function.
A function whose name is/can be thought of as the set of ordered pairs (x, y) for
which each member x of the domain is associated with exactly one member
y = f(x). The function can also be regarded as a rule that assigns a unique
“output” in the set Y to each “input” from the set X.
*Most mathematics journals have problem sections that solicit interesting problems and solutions
for publication. From time to time we will reprint a problem from a mathematics journal. If you have
difficulty solving a journal problem, you may wish to use a library to find the problem and solution as
printed in the journal. The title of the journal is included as part of the problem, and we will generally
give you a reference as a footnote. This problem is found in Volume 18, 1982, issue 2, p. 7.
Sec. 1.4 Functions and Their Graphs 27
What this says: To be called a function, the rule must have the property
that it assigns one and only one output to each input.
A visual representation of a function is shown in Figure 1.22. Note that it is quite

possible for two different elements in the domain X to map into the same element
in the range, and that it is possible for Y to include elements not in the range off. If
the range of/does, however, consist of all of Y, then/is said to map X onto Y.
Furthermore, if each element in the range is the image of one and only one
element in the domain, then / is said to be a one-to-one function. We shall have
more to say about these terms in Chapter 5. A function is called a real-valued
function if the range is a subset of the set of real numbers.** In this book, we work
with real-valued functions of a real variable which restrict both the domain and
range to the set of real numbers. A function/is said to be bounded on [a, b∖ if there
exists a number B so that ∣∕(x)∣ ≤ B for all x in [a, b].
Figure 1.22 A function as a mapping
■ FUNCTIONAL NOTATION
Functions can be represented in several ways, but usually they are defined by using
a mathematical formula. It is traditional to let x denote the input and y the
corresponding output, and to write an equation relating x and y. The letters x and
y that appear in such an equation are called variables. Because the value of the
variable y is determined by that of the variable x , we call y the dependent variable
and x the independent variable.
In this book, when we define functions by expressions such as
∕( x ) = 2x + 3 or g(x) = x 2 + 4x + 5
we mean the functions/ and g are the sets of all ordered pairs (x, y), satisfying the
equations y = 2x + 3 and y = x 2 + 4x + 5, respectively. To evaluate a function /
means to find the value of/for some particular value in the domain. For example,
to evaluate/at x = 2 is to find ∕(2 ).
EXAM PLE 1 Using functional notation
Suppose f(x ) = 2x 2 - x . Find / ( - 1), /(0), /(2), f ( x + ∕z), and —^λ- -

where x and h are real numbers and h ≠ 0.
Solution In this case, the defined function / tells us to subtract the independent
*The functions that appear in this book belong to a very special class called elementary functions,
defined by Joseph Liouville (1809-1882). You will study some nonelementary functions in complex
analysis and in advanced calculus. Many functions that appear in physics and higher mathematics are
not elementary functions.
variable x from twice its square. Thus, we have

Historical Note
/ (-1 ) = 2 ( - 1)2 —(-1 ) = 3
/(0) = 2(0)2 - (0) = 0
/(2) = 2(2)2 - 2 = 6
To find f ( x + h) we begin by writing the formula fo r/in more neutral terms,
say as
∕( □ ) = 2 ( □ ) 2 - ( □ )
Then we insert the expression x + h inside each box, obtaining
∕ ( ∣v + ⅞∣) = 2(∣x + ⅞[)2 - (∣x + ⅛∣)
= 2(x2 + 2xh + ∕ι 2) - (x + h)
= 2x 2 + 4xh + 2Λ2 - x - h
SOPHIE G ERMAIN (1776-1831) Finally, if h ≠ 0,
Sophie Germain was one of the
first women recorded in the f ( x + ∕z) - ∕( x ) _ [2x2 + 4xh + 2h 2 - χ - h∖- [2x2 - x]
history of mathematics as doing h ~ h
original mathematical research.
= *xh + 2h 2 - h = 4x + 2 h _ [
In her time, women were not
admitted to first-rate universi
ties and were not, for the most
part, taken seriously, so she The expression —------ --------— is called a difference quotient and is used in
wrote at first under the pseudo
nym LeBlanc. Even though Chapter 2 to compute the derivative.
Germain’s most important re Sometimes functions need to be defined in pieces because they have a split
search was in number theory, domain. These functions require more than one formula to define the function,
she was awarded the prize of and therefore are called piecewise-defined functions.
the French Academy for a pa
per entitled “Memoir on the
Vibrations of Elastic Plates.” E X A M P L E 2 Evaluating a piecewise-defined function
As we progress through this
book we will profile many
mathematicians in the history ' x sinx if x < 2
of mathematics, and you will i f ∕ω = ■
notice that most of them are J x 2 + 1 ifx ≥ 2
white males. Why? This issue is
addressed in the questions at find/(-0 .5 ), ∕( f ) , and∕(2 ).
the end of this chapter.
Solution To find / (-0 .5 ) we use the first line of the formula because -0 .5 < 2:
/ (-0 .5 ) = —0.5 sin (—0.5) = 0.5 sin 0.5 This is the exact value.
≈ 0.2397 This is the approximate value.
To find ∕( f ) we use the first line of the formula because f ≈ 1.57 < 2 :
∕( f ) = f sin7 = j This is the exact value.
Finally, because 2 ≥ 2, we use the second line of the formula to find ∕(2 ).
∕( 2 ) = 3(2)2 + 1 = 13
Functional notation can be used in a wide variety of applied problems, as

shown by Example 3 and again in the problem set.
EXAMPLE 3 Applying functional notation

It is known that an object dropped from a height in a vacuum will fall a
distance of s ft in t seconds according to the formula
s(l) = 16t2, t ≥ 0
a. How far will the object fall in the first second? In the next 2 seconds?
b. How far will it fall during the time interval t = 1 to t = 1 + h seconds?
c. What is the average rate of change of distance (in feet per second) during
the time t = 1 sec to t = 3 sec?
d. What is the average rate of change of distance during the time t = x
seconds to t = x + h seconds?
Solution
a. 5(1) = 16(1)2 = 16
In the first second the object will fall 16 ft. In the next two seconds the
object will fall
5(1 + 2 ) - 5(1) = 5(3) - 5(1) = 16(3)2 - 16(1)2 = 128
The object will fall 128 ft in the next 2 sec.
b. 5(1 + h) - 5(1) = 16(1 + h)2 - 16(1)2
= 16 + 32Λ + 16Λ2 - 16 = 32h + 16∕z2
CHANGE IN DISTANCE = ∙S-(3) ~ -S,( 1) = 128
C. AVERAGE RATE =
CHANGE IN TIME 3 -1 2 = 64
The average rate of change is 64 ft∕sec.
s(x + h) - 5(x) _ s(x + h) - s(x)
d. (x + h) - x ~ h Does this look familiar? See Example 1.
= 16(x + ⅛)2 - 16x2 = 16Λ 2 + 32xh + 16⅛2 - 16x2 = 3 2 γ + 16∕z

h h
■ DOMAIN AND RANGE OF A FUNCTION
0 Notice this agreement about In this book, unless otherwise specified, the domain of a function is the set of real
the domain of a function, which numbers for which the function is defined. We call this the domain convention. If a
will be used throughout the function f is undefined at x, it means that x is not in the domain of f The most
book. 0 frequent exclusions from the domain are those values that cause division by 0 and
negative values under a square root. In applications, the domain is often specified
by the context. For example, if x is the number of people on an elevator, the
context requires that negative numbers and nonintegers be excluded from the
domain; therefore, x must be an integer such that 0 ≤ x ≤ c where c is the
maximum capacity of the elevator.
EXAMPLE 4 Domain of a function

Find the domain for the given functions.
a. ∕(x ) = 2x - 1 b. g(v) = 2x - 1, x ≠ -3
c. h(x) = (2X~+(3+ 3) d. F(x) = V T + 2 e. G(x) = 5 ,*
Solution
a. All real numbers; D = (-∞, ∞)
b. All real numbers except - 3

c. Because the expression is meaningful for all x ≠ - 3 , the domain is all real
numbers except - 3 .
d. F h as meaning if and only if x + 2 is nonnegative; therefore the domain is
x ≥ - 2 , or D = [ - 2 , ∞).
e. G is defined whenever 5 —cosx ≠ 0. This imposes no restriction on x since
∣cos.v∣ ≤ 1. Thus, the domain of G is the set of all real numbers;
D = (—∞, ∞). ^bh
Equality of Functions Two functions/and g are equal if and only if

1. f and g have the same domain.
2. ∕( x ) = g(x) for all x in the domain.
∕( x ) = 2x - 1 In Example 4 (repeated in the margin here) the functions g and h are equal. A
<g(x) = 2x - 1, x ≠ - 3 common mistake is to “ reduce” the function h to the function /:
∕z(x) = (2x ~ l)(x + 3)
x + 3 0 W R O N G : ∕z(x) =
(2* ~ 1j⅛'+ 3) = 2x - 1 = ∕( x ) 0
R IG H T : ∕z(x) = (
l∕ j - = 2x - 1, χ ≠ -3 ; therefore, /?(x) = g(x).
JC ∣ J
The following concept can be used to reduce (by half) the amount of work
necessary on many problems.
Even and Odd Functions A function/is called

even i f ∕( - x ) = ∕( x ) , and
odd i f ∕( - x ) = - ∕( x )
for all x in the domain of /
Just as not every real number is even or odd (2 is even, 3 is odd, but 2.5 is neither),
not every function is even or odd. For example:
∕( x ) = x 2 is even, because/ (- x ) = ( - χ ) 2 = x 2 = ∕( x )
g(x) = x 3 is odd, because g ( -x ) = ( - x ) 3 = ~x 3 = ~g(x~)
∕z(x) = x 2 + x is neither because ∕z(-χ ) = ( - x ) 2 + ( - x ) = x 2 - x
Note that ∕z(-χ ) ≠ ∕z(x) and ∕z(-x) ≠ -∕z(x).
■ COMPOSITION OF FUNCTIONS
There are many situations in which a quantity is given as a function of one
variable that, in turn, can be written as a function of a second variable. Suppose,
for example, that your job is to ship x packages of a product via Federal Express to
a variety of addresses. Let x be the number of packages to ship, and let / be the
weight of the x objects and g be the cost of shipping. Then
The weight is a function of the number of objects: /(x);
The cost is a function of the weight: <g[∕(x)].
This process of evaluating a function of a function illustrates the idea of
composition o f functions.
Composition of Functions The composite function fa g is defined by

(∕°∙g )W =/[£(*)]
for each x in the domain of g for which g(x) is in the domain of f
What this says: To visualize how functional composition works, think of
fo g in terms of an “assembly line” in which g and f are arranged in series, with
output g(x) becoming the input o f f as illustrated in Figure 1.23.
Assembly line interpretation of composite functions Range o ff
Figure 1.23 Composition of functions
EXAM PLE 5 Finding the composition of functions
If f(x ) = 3x + 5 and g(x) = V x, find the composite functions ∕□ g and g o f

0 Example 5 illustrates that Solution The function ∕'o g is defined by ∕[g(x)].
functional composition is not
(/ ° <?)(*) = ∕I⅛ W ] = ∕(V x ) = 3√x + 5
commutative. 0 T h a tis ,∕o g is
not, in general, the same asgof. The function g o /is defined by ⅞[∕(x)].
(<g <√)W = g[∕(x)] = g(3x + 5) = √ 3x + 5
In addition to using function machines to visualize Initial value Output value from “some formula” in cell Al
the composition of functions, we can also visualize
composition in terms of “boxes” or “spreadsheet c e ll Al A2
cells”:
A spreadsheet is a computer program to manipulate
data and carry out calculations or chains of calcula
tions (as in composition of functions). If you have B
Output from “second ,,,, ,-.....
access to a computer and software such as Excel, formula, A2” is the
Lotus 1-2-3, or Quattro-Pro, you might wish to look composition of the
formulas. composite value
at the Technology Manual accompanying this book.
EXAM PLE 6 An application of composite functions
Air pollution is a problem for many metropolitan areas. Suppose that carbon
monoxide is measured as a function of the number of people according to the
following information:
Daily Carbon Monoxide Level

Num ber of People (in parts per million)
100,000 1.41
200,000 1.83
300,000 2.43
400,000 3.05
500,000 3.72
Further studies show that a refined formula for the average daily level of
carbon monoxide in the air is
L (p) = 0 .7 √ p 2 + 3
Further assume that the population of a given m etropolitan area is growing
according to the formula p(t) = 1 + 0 .0 2 /, where t is the tim e from now (in
years) and p is the population (in hundred thousands). Based on these
assumptions, what level of air pollution should be expected in 4 years?
Solution The level of pollution is L(p) = 0.7Vp 2 + 3, where p(t) = 1 + 0.02t 3 .
Thus, the pollution level at tim e t is given by the composite function
(L o jp)(Z) = L[ jp(t)] = £(1 + 0 .0 2 /) = 0 .7 √ (l + 0 .0 2 ∕) 2 + 3
In particular, when t = 4, we have
(£ o jp)(4) = 0 .7 √ [l + 0.02(4) 3] 2 + 3 ≈ 2.0 ppm —
In calculus, it is frequently necessary to express a function as the composite of

two simpler functions.
EXAM PLE 7 Expressing a given function as the composite of two functions
Express each of the following functions as the composite of two functions u

and g so that f( x ) = g[w(x)].
a. ∕( x ) = (x 2 + 5x + 1)5 b. f( x ) = cos 3x
c. ∕( x ) = sin x 3 d. f ( x ) = V 5x 2 - x
Solution Given Function Inner Function Outer Function

∕ W = ■?[«(*)] M(X) SM *)]
a. ∕( x ) = (x 2 + 5x + 1)5 M(X) = x 2 + 5x + 1 ^[ M(X )] = [w(x)]5
b. ∕( x ) = cos 3x u(x) = cosx g[w(x)] = [w(x)]3
c. ∕( x ) = sin x 3 u(x) = x3 g[w(x)] = sin [u(x)]
d. ∕( x ) = V 5x 2 - x w(x) = 5x 2 - x g[w(x)] = V M(X )
0 Consider Example 7 and this There are often other ways to express a composite function, but the most
paragraph carefully. 0 common procedure is to choose the function u to be the “ inside” portion of the
given fu n c tio n / Notice that in part a the “inside” portion is the portion inside the
parentheses. Compare parts b and c: the inside portion of b is cosx because ∕( x )
= cos 3x = (cosx) 3 . In part c, on the other hand, M(X ) = x 3 is the inside portion
because sin x 3 = sin (x 3). Finally, notice in part d that M(X ) is the part “inside” the
radical sign.
■ GRAPH OF A FUNCTION
Graphs have visual impact. They also reveal information that may not be evident
from verbal or algebraic descriptions. Two graphs depicting practical relationships
are shown in Figure 1.24.
a. A production function b. Bounded population growth

This graph describes the variation in total industrial This graph represents the growth of a population
production in a certain country over a five year when environmental factors impose an upper bound
time span. The fact that the graph has a peak suggests on the possible size of the population. It indicates
that production is greatest at the corresponding time. that the rate of population growth increases at first and
then decreases as the size of the population gets closer
and closer to the upper bound.
Figure 1.24 Examples of functions
To represent a function y = f(x~) geometrically as a graph, it is traditional to

use a Cartesian coordinate system on which units for the independent variable x
are marked on the horizontal axis and units for the dependent variable y on the
vertical axis.
Graph of a Function The graph of a function f consists of all points whose coordinates (x, y) satisfy
y = ∕(x), for all x in the domain o ff.
In Chapter 3 we will discuss efficient techniques involving calculus that you

can use to draw accurate graphs of functions. In beginning algebra you began
sketching lines by plotting points, but you quickly found out that this is not a very
efficient way to draw more complicated graphs, especially without the aid of a
graphing calculator or computer. Table 1.3 includes a few common graphs you
have probably encountered in previous courses. We will assume that you are
familiar with their general shape and know how to sketch each of them.
Intercepts The points where a graph intersects the coordinate axes are called
intercepts. Here is a definition.
Intercepts If the number zero is in the domain o f f and /(0) = b, then the point (0, b~) is
called the y-intercept of the graph of f If <7is a real number in the domain of f
such that ∕(α ) = 0, then (a, 0) is an x-intercept of f
What this says: To find the x-intercepts, set y equal to 0, and solve for x. To
find the y-intercept, set x equal to 0 and solve for y.
TABLE 1.3 Directory of Curves
Identity Function Standard Quadratic Function Standard Cubic Function
y ~X y = x2 y = x3
Standard Reciprocal Standard Reciprocal Squared Standard Square Root Reciprocal
Secant Function Cosecant Function Cotangent Function

V = SβCΛ' V = CSCΛ' V = COtΛ' _
EXAM PLE 8 Finding the intercepts
Find all intercepts of the function ∕(x ) = - x 2 + x + 2.

Solution The ^-intercept is (0, ∕(0)) = (0, 2). To find the x-intercepts, solve the
equation ∕(x ) = 0. Factoring, we find that
-x 2 + x + 2 = 0
x 2 —x - 2 = 0
(x + l)(x - 2) = 0
x = -1 or x = 2
Thus, the intercepts are (0, 2), (-1 , 0) and (2, 0).
Vertical Line Test Functions can have several x-intercepts but can have at most
one j-'-intercept. Thus follows a very useful technique for deciding if a given curve
is the graph of a function. The graph of a typical function is shown in Figure 1.25.
Notice that for each a in the domain there is only one point (a,f(a)) on the graph.
Figure 1.25 Graph of a function This means that a given vertical line passes through the graph of a function in at
most one point. This is the vertical line test for the graph of a function, as shown in
Figure 1.26.
The Vertical Line Test A curve in the plane is the graph of a function if and only if it intersects a
vertical line no more than once.
a. The graph of a function: No vertical line intersects the curve more than once. b. Not the graph of a function: The curve intersects
at least one vertical line more than once.
Figure 1.26 The vertical line test
■ TRANSFORMATION OF FUNCTIONS
Sometimes the graph of a function can be sketched by translating or reflecting the

graph of a related function. We call these translations and reflections transforma
tions of a function. This procedure is illustrated in Figure 1.27, in which we have
sketched the graph of y = x 2 and then translated and reflected that graph.
Functional Transformations The graph defined by the equation

y -k = f(x -h )
is said to be a translation of the graph defined by y = ∕(x).
The translation (shift, as shown in Figure 1.27) is
to the right if h > 0 up if k > 0
to the left if h < 0 down if k < 0
A reflection in the x-axis of the graph of y = ∕(x ) is the graph of

y = -f(χ )
A reflection in the j-axis of the graph of y = ∕(x ) is the graph of
T = ∕( - x ) I
What this says∙. If we replace x by x - h and y by y - k, the graph is

translated so the origin (0, 0) is moved to the point (h, k). Also, the graph is
reflected in the x-axis if we replace y by - y in its equation, and it is reflected in
the j>-axis if we replace x by —x.
Figure 1.27 Transformations of y = x 2
EXAM PLE 9 Graphing with a translation
Graph y + 2 = sin (x - 1).

Solution By inspection, the desired graph may
be obtained by translating the standard sine
curve y = sin x to the right by h = 1 unit and
vertically by k = - 2 units—that is, the ori
gin (0, 0) is moved to the point (1, -2). The
graph is shown in Figure 1.28.
Figure 1.28 Translation of y = sinx from (0, 0) to (Λ, k) = (1, -2 )
EXAM PLE 10 Graphing with a reflection
Graph y = -V x .
Solution The graph of this function is a reflection in the x-axis of the graph of
Figure 1.29 Reflection in the y = Vx, as shown in Figure 1.29. M M
x-axis of y = Vx
38 Ch. 1 Preview o f Calculus: Functions and Limits
■ CLASSIFICATION OF FUNCTIONS
The functions shown in Table 1.3 are examples of elementary functions. We will
now describe some of the common types of functions we use in this course.
Polynomial Function A polynomial function is a function of the form

∕( x ) = anx n + a n _ r x n ~ 1 + + a2x 2 + a l x + a0
where n is a nonnegative integer and an, . . . , a2 , a l , a0 are constants. If an ≠ 0,
the integer n is called the degree of the polynomial. The constant an is called
the leading coefficient and the constant rz0 is called the constant term of the
Examples of Polynomial polynomial function. In particular,
Functions:
∕W = 5 A constant function is zero degree: f(x ) = a
f{x) = l x - √2 A linear function is first degree: f(x ) = ax + b
∕'(x) = 3x 2 + 5x - 1 A quadratic function is second degree: ∕( x ) = a x 2 + bx + c
∕( x ) = V∑X 3 - 77X A cubic function is third degree: ∕( x ) = ax i + b x 2 + ex + d
A quartic function is fourth degree: ∕( x ) = αx 4 + bx 3 + cx 2 + dx + e
The identity function, standard quadratic function, and standard cubic function
in Table 1.3 on page 35 are examples of polynomial functions.
A second important algebraic function is a rational function.
Rational Function A rational function is the quotient of two polynomial functions, p(x) and d(x)∙.
ΦO≠° ∣
Examples of Rational
Functions:
∕( x ) = x~'
When we write d(x) ≠ 0 we mean that all values c for which r∕(c) = 0 are excluded
-
from the domain of d. The standard reciprocal and standard reciprocal squared
x2 + lx - 3 functions in Table 1.3 are examples of rational functions.
f(x) = x~ 3 + V l x If r is any nonzero real number, the function ∕( x ) = x r is called a power
function with exponent r. You should be familiar with the following cases:
Examples of Power Functions:
∕( x ) = x 6 Integral powers (r = n, a positive integer): f( x ) = x n = x ■x ......... x j
n factors
f(x) = x^ 4
Reciprocal powers (r is a negative integer): f(x ) = x~ n = — for x + 0
-A
∕( x ) = x 3/4
Roots (r = is a rational number): ∕( x ) = x m ,n = ∖zx"' = (Azx) m for
,∕'(x) = V x 2 x ≥ 0 if n even, n ≠ 0 I — is reduced )
Power functions can also have irrational exponents (such as V 2 or τr), but such
functions must be defined in a special way and are introduced in Chapter 5.
A function is called algebraic if it can be constructed using algebraic
operations (such as adding, subtracting, multiplying, dividing, or taking roots)
starting with polynomials. Any rational function is an algebraic function.
Functions that are not algebraic are called transcendental. The following
functions are transcendental functions:
Trigonometric functions are the functions sine, cosine, tangent, secant, cose-
CMJA c a n t ’ a n d c o t a n ge n t ∙ These
functions are reviewed in the Student Mathematics
< 1j ∕ Handbook. You can also review these functions by consulting a trigonometry
or precalculus textbook.
Exponential functions are functions of the form fix ) = bx , where b is a positive
constant. We will study these functions in Chapter 5.
Logarithmic functions are functions of the form fix ) = logbx, where b is a
positive constant. We will also study these functions in Chapter 5.
PROBLEM SET 1 .4
22. a. fix ) = 2 χ2 ~- x 2 ~ 6 ; g(x) = 2x + 3, x ≠ 2
In Problems 1-12, find the domain o f f and compute the
indicated values or state that the corresponding x-value is not in
the domain. b∙ fix ) =
3%\;_7 3~6; g(x) = 3x + 2, x ≠ 3
O l . ∕( x ) = 2 x + 3;/(-2),/(1),/(0)
2. ∕( x ) = - x 2 + 2x + 3;/(0),/(1),/(-2) 23. a. fix ) =
3χ2~2¾~2⅞S M = 3x + 1
3. ∕( x ) = 3x2 + 5x - 2;/(1),/(0),/(-2)
b∙ ∕ ( x ) ≈ ^ ÷ 1 ) ⅞ - 2 ∖ x ≠ 6 i
4. ∕( x ) = x + i j /(-1), /(1), /(2)
5∙ ∕W = (%+J+ Λ3~2); /(2), /(0), /(-3)

Classify the functions defined in Problems 24-27 as even, odd,
6 . ∕( x ) = ( 2 χ - l ) → 2∙,∕( l) ,∕⅛ ) ,∕( 1 3 )
or neither.
7. ∕( x ) = ∕( - l ) , ∕(⅜), /(1) 24. a. ∕( x ) = x 2 + 1 b. ∕ 2(x) = V x 2
8. ∕( x ) = V x 2 + 5x + 6;/(0),/(1),/(-2) 25. a. ∕ 3(x) = j - 1

4 b. ∕ 4(x) = x 3 + x
9. ∕( x ) = sin(l —2x);/(—1 ) ,∕( ∣) ,∕( 1 )
26. a. ∕ 5(x) = ( v3 3)2 b. ∕ 6(x) = (x 3 ∣ x)2
lθ. ∕( x ) = sinx - cosx;/(0), ∕( - 5 ) ,∕( π )
27. a. ∕ 7(x) = ∣x∣ b. ∕ 8(x) = ∣x∣ + 3
,, /-/.A _ f - 2 x + 4 i f x ≤ 1
ifx > 1 In Problems 28-33, find the composite functions f o g and g o f.
/(3),/(1),/(0) 28. fix ) = x 2 + 1 and g(x) = 2x
'3 i f x < —5 29. ∕( x ) = sinx and g(x) = 1 - x 2
12. ∕( x ) = ■x + 1 i f - 5 ≤ x ≤ 5
30. / ( Z) = fit and g(t) = t2
. Vx if x > 5
/(-6),/(-5),/(16) 31. fin ) = ''V ‘ and g(u) =
In Problems 13-20, evaluate the difference quotient 32. fix ) = sinx and g(x) = 2x + 3
f i x + h) - fix ) .
----------------7 — --------- jor the given Junction j.
33. fix ) = and g(x) = tanx
13. fix ) = 9x + 3 14. fix ) = 5 - 2x
15. fix ) = 5x2 16. fix ) = 3x2 + 2x In Problems 34-37, express f as the composition o f two
17. fix ) = ∣x∣ if x < -1 and 0 < h < 1 functions u and g such that fix ) = g[w(x)]∙
18. fix ) = ∣x∣ if x > 1 and 0 < h < 1 34. a. fix ) = (2x2 - 1)4 b. ∕( x ) = V 5x - 1
35. a. ∕( x ) = tan 2x b. ∕( x ) = tanx 2
19∙ ∙<W = X jo- ∕ ω - f⅛ τ
36. a. ∕( x ) = sin V x b. ∕( x ) = V sinx
37. a. ∕( x ) - sin (* _ *) b. fix ) - tan Q 3∖.)
State whether the functions f and g in Problems 21-23 are
equal.
Use the directory o f curves (Table 1.3) and the ideas o f
21. a. fix ) = fix ) = 2x + 1 translation and reflection to sketch the graphs o f the functions
given in Problems 38-51.
b. fix ) = 2 " Λ fix ) = 2x + 1, x ≠ 0 38. fix ) = x 2 + 4 39. fix ) = (x + 4)2
40. y = - χ 3 41. y = -∣x∣
43. y + 1 = 4 a. Express the illuminance on the floor as a composite

function of t for 0 < t < 6.
44. y = ⅜ + 2 45. y = Ψ x + 2
b. What is the illuminance when t = 1? When t = 4?
46. y = cos (x — 1) 47. y = cosx - 1
58. It is estimated that t years from now, the population of a
48. y = sin(x + 2) 49. y = sinx + 2
50. y = tan(x + 1) 51. y = tanx + 1
certain suburban community will be P(t) = 20 -
thousand people.
a. What will the population be nine years from now?
b. By how much will the population increase during the
ninth year?
c. What will happen to the size of the population in the
“long run”?
59. To study the rate at which animals learn, a psychology
student performs an experiment in which a rat is sent
repeatedly through a laboratory maze. Suppose that the
time (in minutes) required for the rat to traverse the maze
12
on the nth trial is approximately ∕(zι) = 3 + — .
a. What is the domain of the function f l
Figure 1.30 b. For what values of n does ∕'(zz) have meaning in the
context of the psychology experiment?
53. If point B in Figure 1.30 has coordinates (3, ^(3)), what c. How long does it take the rat to traverse the maze on
are the coordinates of R and S? the third trial?
54. A ball is thrown directly upward from the edge of a cliff d. On which trial does the rat first traverse the maze in 4
in such a way that t seconds later, it is s = —16i2 + minutes or less?
96z + 144 feet above the ground at the base of the cliff.
e. According to the function f, what will happen to the
Sketch the graph of this equation (making the Z-axis the
time required for the rat to traverse the maze as the
horizontal axis) and then answer these questions:
number of trials increases? Will the rat ever be able to
a. How high is the cliff?
traverse the maze in less than three minutes?
b. When (to the nearest tenth of a second) does the ball
60. Biologists have found that the speed of blood in an artery
hit the ground at the base of the cliff?
is a function of the distance of the blood from the artery’s
c. Estimate the time it takes for the ball to reach its central axis. According to Poiseuille, s law, the speed
maximum height. What is the maximum height? (cm∕sec) of blood that is r cm from the central axis of an
55. Suppose the total cost (in dollars) of manufacturing q artery is given by the function .S'(r) = C(R 2 —r 2), where C
units of a certain commodity is given by is a constant and R is the radius of the artery.* Suppose
that for a certain artery, C = 1.76 × 105 cm ∕s 2 and
C(q) = q 3 - 30√2 + 400√ + 500 for 0 ≤ q ≤ 30
R = 1.2 × 10^2 cm.
a. Compute the cost of manufacturing 20 units. a. Compute the speed of the blood at the central axis of
b. Compute the cost of manufacturing the twentieth unit. this artery.
56. An efficiency study of the morning shift at a certain b. Compute the speed of the blood midway between the
factory indicates that an average worker who arrives on artery’s wall and central axis.
the job at 8:00 A.M. will have assembled ∕( x ) = - x 3 + 61. At a certain factory, the total cost of manufacturing q
6x + 15x 2 CD players x hours later (0 ≤ x ≤ 8). units during the daily production run is C∖ q) = q2 +
a. How many players will such a worker have assembled q + 900 dollars. On a typical workday, the numbers of
by 10:00 A.M.? units manufactured during the first t hours of a produc
b. How many players will such a worker assemble be tion run can be modeled by the function q(t) = 25t.
tween 9:00 A.M. and 10:00 A.M.? a. Express the total manufacturing cost as a function of t.
57. In physics, a light source of luminous intensity K candles b. How much will have been spent on production by the
is said to have illuminance I = K ∕s 2 on a flat surface s ft end of the third hour?
away. Suppose a small, unshaded lamp of luminous c. When will the total manufacturing cost reach $ 11,000?
intensity 30 candles is connected to a rope that allows it to
be raised and lowered between the floor and the top of a
10-ft-high ceiling. Assume that the lamp is being raised
and lowered in such a way that at time t (in min) it is *The law and the unit poise, a unit of viscosity, are both named
s = 6t - t 2 ft above the floor. for the French physician Jean Louis Poiseuille (1799-1869).
Sec. 1.5 The Limit of a Function 41
Computational Window 65. Imagine a sphere of unknown radius r with a meter

stick sitting upright at the “north pole.” A wire is
62. Define the functions f(x ) = x 2 + 1 and g(x) = x 3 - strung from the top of the stick to a point of tangency
x 2 - 9x + 9 and plot them on the same graph. From of the sphere; and when paced off along the surface,
your graph, estimate the three values of x where the this distance is 25 meters.
plots cross. a. Show that one gets the following equation for r:
P^(χ]
63. Form the rational function r(x) = and plot it on
∞ s - -----£-j- = 0
[-2 0 , 20]. Make a hand sketch of what you see. The r r+ 1
graph of ∕'(x) looks linear for “large” x. Can you b. Graphically estimate, say to one decimal place
figure out what linear function approximates r for accuracy, the meaningful value of r. Note', there
large x? are many solutions, but the one that you want is
64. Define the function G(x) = x 5 + 2x 4 - 9x - 18 and greater than 25; why?
plot G on several x-values to get a good idea of its
behavior.
a. Zoom in on the largest x-value for which G = 0. © 66. JOURNAL PROBLEM;* The Mathematics Student Journal.
Does the function become “almost linear” as you Given that /(11) = 11 and
zoom in?
b. One way or another, factor G to find the exact
x-values at which G = 0. Hint'. G(2) = 0.
c. Define a new rational function, R = GIF, where for all x, find /(2000).
F(x) = x 3 + 3. Decide if R is “almost linear” for x
large (as was the case with the r discussed in
Problem 63). Discuss briefly.
1 . 5 THE LIMIT OF A FUNCTION
IN TH IS SECTIO N Intuitive notion of limits, limits by graphing, limits by

table, limits that do not exist
This section introduces you to the limit of a function, a concept that gives
calculus its power and distinguishes it from other areas of mathematics, such as
algebra.
Our goal in this section is to give an intuitive introduction to the limit o f a

function. A more rigorous treatment will be given in Section 1.8, and variations of
the limit concept will be introduced later in the text.
The development of the limit concept was a major mathematical break
through in the history of mathematics, and it is unrealistic for you to expect to
understand everything about this concept immediately. Have patience, read the
examples carefully, and work as many problems as possible, and eventually, the
limit concept will become a useful part of your mathematical toolkit.
■ INTUITIVE NOTION OF A LIMIT
The limit of a function/is a tool for investigating the behavior of ∕(x ) as x gets
closer and closer to a particular number c. To visualize this concept we begin with
an example.
t Volume 28, 1980, issue 3, p. 2; note the journal problem requests/(1979), which, no doubt, was
related to the publication date. We have taken the liberty of updating the requested value.
EXAM PLE 1 Velocity as a limit
A freely falling body experiencing no air resistance falls s(t) = 16∕2 feet in t
seconds. Express the body’s velocity at time t = 2 as a limit.
Solution We need to define some sort of “mathematical speedometer” for
measuring the instantaneous velocity of the body at time t = 2. Toward this
end, we first compute the average velocity v(Z) of the body between time t = 2
and any other time t by the formula
- (t\ = DISTANCE TRAVELED = -Ψ ) ~ ∙y ( 2 )
ELAPSED TIME t- 2
_ 16Z2 - 16(2)2 _ 16Z2 - 6 4
t- 2 t- 2
As t gets closer and closer to 2, it is reasonable to expect the average velocity
v(Z) to approach the value of the required instantaneous velocity at time t = 2.
lim v(∕) = lim^-~—
t→2 v ’ ∣→2 t - 2
This is the instantaneous velocity at t = 2.
Notice that we cannot find the instantaneous velocity at time t = 2 by simply
substituting t = 2 into the average velocity formula because this would yield
the meaningless form 0/0. ■■■
We now devote the remainder of this section to an intuitive introduction of how

we can find the value of limits such as the one that appears in Example 1.
Limit of a Function (Informal The notation

Definition)
lim ∕(x ) = L
x →c
is read “the limit of∕(x ) as x approaches c is L” and means that the functional
values∕(x ) can be made arbitrarily close to L by choosing x sufficiently close to
c (but not equal to c).
-'⅛ What this says: If∕(x ) becomes arbitrarily close to a single number L as x
approaches c from either side, then we say that L is the limit of ∕(x ) as x
approaches c. The limit lim ∕(x ) exists if and only if the limiting value from the
left equals the limiting value from the right.
This informal definition of limit cannot be used in proofs until we give precise
meaning to terms such as “arbitrarily close to L ” and “sufficiently close to c.”
This will be done in Section 1.8. For now, we shall use this informal definition to
gain a working knowledge of limits.
■ LIMITS BY GRAPHING
Figure 1.31 shows the graph of a function f and the number c = 3. The arrowheads
are used to illustrate possible sequences of numbers along the x-axis, approaching
from both the left and the right. As x approaches c = 3,∕(x) gets closer and closer
Figure 1.31 Limit as x to 5. We write this as .
approaches c Iim ∕W = 5
As x approaches 3 from the left we write x → 3^, and as x approaches 3 from the
right we write x → 3 + . We say that the limit at x = 3 exists only if the value
approached from the left is the same as the value approached from the right.
EXAMPLE 2 Estimating limits by graphing
Given the function/defined by the graph in Figure 1.32, find the following
limits by inspection, if they exist:
a. lim ∕(x ) b. lim .∕(x ) c. lim ∕(x )
Solution Take a good look at the given graph; notice the open circles on the graph
at x = 0 and x = - 2 and also notice that /(0) = 5.
a. lim ∕(x ) is the value that / approaches as x tends toward 3 from the left.
From Figure 1.32 we see that this value is -2 . We write lim_/(x) = -2 .
b. lim + ∕(x ) is the value that / approaches as x tends toward - 2 from the
right. We see that this value is 4, so
Figure 1.32 Limits from a graph γ lim + ∕(* ) = 4
c. To find lim ∕(x ) we need to look at both the left and right limits. Look at
Figure 1.32 to find
lim ∕(x ) = 1 and limt ∕(x ) = 1
so lim ∕(χ-) exists and lim ∕(x ) = 1. Notice here that the value of the limit as
x → 0 is not the same as the value o f the function at x = 0. ≡≡
EXAMPLE 3 Finding the limit from Example 1 by graphing
Find lim _ 2 ⅛ by graphing.
Solution
16∕2 - 6 4 . 16(∕2 - 4 ) 16(/-2)(/ + 2)
v ∕) - z _ 2 - t _ 2 - ------- r r 2 ------ -1 6 (/ + 2), ∕≠ 2
The graph of v(∕) is a line with a deleted point, as shown in Figure 1.33. If you
have a graphing calculator, compare this with the graph shown on your
calculator (see the computational window for Example 1).
The limit can now be seen:
lim v(∕) = 64
That is, the instantaneous velocity of the falling body in Example 1 is 64

ft∕sec. E≡a
Figure 1.33 lim
16f 64 = 64
t →2 t — 20 Notice from the preceding examples that when we write
lim ∕(x ) = L
0 The limit of a function as the we do not require c itself to be in the domain of / nor do we require ∕(c), if it is
defined, to be equal to the limit. Functions with the special property that
independent variable approach
es a point does not depend on lim ∕(x) = ∕(c)
the value of the function at that
point. 0 are said to be continuous at x = c. This idea is considered in Section 1.7.
■ LIMITS BY TABLE
It is not always convenient (or even possible) to first draw a graph in order to find
limits. You can also use a calculator or a computer to construct a table of values for
/ a s x → c.
EXAM PLE 4 Finding a limit with a table
Find lim * by using a table.
Solution You will recognize this limit from Examples 1 and 3. We need to begin
by selecting sequences of numbers for t → 2 “ and t → 2 + :
t approaches from the left; t → 2 . —> <—t approaches from the right; t → 2 + .
t 1.950 1.995 1.999 2 2.001 2.015 2.100
v(t) 63.200 63.920 63.984 Undefined 64.016 64.240 65.600
v(() approaches 64 from the left. —> <→,(0 approaches 64 from the right.
That is, the pattern of numbers suggests

lim 1 6 ^ 6 4 = 64
(→2 t - 2
as we found using a graphical approach in Example 3.
EXAM PLE 5 Finding limits of trigonometric functions
Evaluate lim sinx and lim cosx.

x →0 x →0
Solution We can evaluate these limits by table or by graph.
By table:
X 1 0.5 0.1 0.01 -0.5 -0.1 -0.01

sinx 0.8415 0.4794 0.0998 0.0099998 -0.4794 -0.0998 -0.0099998
cosx 0.5413 0.8776 0.9950 0.9985 0.8776 0.9950 0.99995
The pattern of numbers in the table suggests that

lim sin x = 0 and lim cosx = 1
x→0 x→0
EXAM PLE 6 Evaluating a trigonometric limit using a table
Evaluate lim
x→0 x
Solution ∕( x ) = s ιn ^
x is an even function because
This means that we only need to find the right-hand limit because the limiting
behavior from the left will be the same as that of the right-hand limit. These
values are shown in the following table.
x approaches 0 from the right. —>
X 0.1 0.05 0.01 0.001 0
∕ω 0.998334 0.999583 0.9999833 0.99999983 undefined
f( χ ) approaches 1 from below. —>
The table suggests that lim 51^ = l j therefore, l i m ^ ^ = l . We shall

consider this limit more completely in Section 2.3. M B
Computational Window
Numerical Approach Graphical Approach

1 X
sinx If you have a graphing calculator (or wish to work
X through the graph by plotting points) you can
verify the result of Example 6. The graph is shown
10 0.100 0.998334 in Figure 1.34. Notice that the function is not
20 0.050 0.999583
30 0.033 0.999815 defined at x = 0. This limit will be established
40 0.025 0.999896 formally in the next chapter.
50 0.020 0.999933
60 0.017 0.999954 The table of values at the left shows more detail
70 0.014 0.999966 than the table of values in Example 6 because we
80 0.013 0.999974 found those values using a spreadsheet program, Y tB ≥ in X×X
90 0.011 0.999979 but if you do not have a computer, you can use a ∣
Xr-,
in = ^ 4 '√nin= - , 5
100 0.010 0.999983 calculator. The limit being evaluated is Xrha×=4 '∣
∙T∣
a x = l. 5
120 0.008 0.999988 X s c 1=1 Y s c l= . 1
From the table, we con- lim ⅛ Figure 1.34 l i m ^ = l
elude the limit is 1. Λ→ 0 + X
'.,1^.2
We should note that when we use a table (using either calculator or computer
values) we may be misled. All we can say is that if a limit exists, then it can be
calculated using a table. That is why when we find a limit from a table we must be
cautious about the possibility of erroneous results. For example, if the table
method is used with Example 8 (given at the end of this section), an erroneous
conclusion is possible.
■ LIMITS THAT DO NOT EXIST
It may happen that a function f does not have a (finite) limit as x →c. When
lim ∕(x ) fails to exist, we say that ∕(x ) diverges as x approaches c. The following
examples illustrate how divergence may occur.
EXAMPLE 7 A function that diverges
Evaluate lim -⅛.

.<→0 x 1
Solution As x → 0, the corresponding functional values of ∕(x ) = grow

arbitrarily large, as indicated in the following table.
x approaches 0 from the left; x → 0 —> <—x approaches 0 from the right; x → 0 +
X - 0 .l -0.05 -0.001 0 0.001 0.005 0.01
f(×l = ⅛ 100 400 1 × 106 undefined 1 × 106 4 × 104 1 × 104
The graph of / i s shown in Figure 1.35.

(x →0-) →0 <- (x→0+ )
Geometrically, the graph of y = ∕(x ) rises without bound as x → 0. Thus,
Figure 1.35 lim does not exist,
.r→0 X 1 lim does not exist, so we say / diverges as x → 0.
and the graph illustrates that f
rises without bound.
A Function Diverges to Infinity A function/that increases or decreases without bound as x approaches c is said
to diverge to infinity (∞) at c. We indicate this behavior by writing
lim ∕(x ) = +∞ if /increases without bound and by
x →c
lim ∕(x ) = -∞ if it decreases without bound.
0 It is important to remember Using this notation, we can rewrite the answer to Example 7 as
that ∞ is not a number, but is
merely a symbol denoting unre lim -⅛ = +∞z
V— V
stricted growth in the magni

tude of the function. 0
EXAMPLE 8 A function that diverges by oscillation
Evaluate lim sin-.

.t→0 x
Solution Note this is not the same as lim The values of∕(x ) = sinoscillate
infinitely often between 1 and -1 as x approaches 0. For exam pie, ∕(x ) = 1 for
x = 2∕π,2∕(5π),2∕(9π), . . . and∕(x) = - 1 forx = 2∕(3π), 2/(7 π), 2/(1 lτr), . . . .
Graph of y = sin
Figure 1.36 lim sin - diverges by oscillation.
Because the values of∕(x ) do not approach a unique number L as x → 0, the

limit does not exist. This kind of function limiting behavior is called
divergence by oscillation. t= ι
In the next section, we will introduce some properties of limits that will help
us evaluate limits efficiently. In the following problem set remember that the
emphasis is on an intuitive understanding of limits, including their evaluation by
graphing and by table.
PROBLEM SET 1 .5
Given the functions defined by the graphs in Figure 1.37, find Thus, we simply need to check for x → 0 + . Find the limit
the limits in Problems 1-12. by completing the following table.
X 1 0.5 0.1 0.01 0.001 0.0001
f(χ ) 15.33
1. lim ∕(x ) 2. lim ∕'(x) 3. lim f(χ )

4. lim g(x) 5. lim g(x) 6. lim+ g(x)
7. lim+ g(x) 8. lim t(x) 9. lim+ t(x)
10. lim t(x) 11. lim t(x) 12. lim t(χ)
x →2 χ →4 χ →-4
Find the limits by filling in the appropriate values in the tables
in Problems 13-15.
13. lim ∕( x ) , where ∕( x ) = (4x - 5)
X 2 3 4 4.5 4.9 4.99
∕( v ) 3
r 3 _ o
14. χ →
lim
ι.
g(x),
O where Sg(x)
∖ > =Χ 2—+ 2—Χ -—
+ 4
X 1 1.5 1.9 1.99 1.999 1.9999
g(χ) -1
15. lim A(x), where h(x) = — ---- ⅛ — - 23. ■ What Does This Say? Explain a process for finding a
x →2 X ~ λ limit.
X 1 1.9 1.99 1.999 3 2.5 2.1 2.001

Evaluate the limits in Problems 24-58 to two decimal places by
h(x) 7 graphing or by using a table o f values. I f the limit does not
exist, explain why.
'
16. Find lim
∙ tclΓl V
using the following procedure based on 24. lim+ x 4 25. lim cosx
x →0 Λ'→0÷
the fact that f( x ) = tanx is an odd function: 26. lim (x 2 - 4) 27. lim (x 2 - 4)
If f(x) = ⅛h2¾ then 28. lim - j -7 29. lim — — τ∙
j v ’ tan 3x x → ι+ x - 3 x →- 3 + x - 3
7 v 7κ 30. lim —⅛ 31. lim tanx

’ ta n (-3 x ) -ta n 3 x ’ .γ →J X —3 x →τr∕2
32. a. Compute the limit

.............. X *) - s(t)
v(∕) = lim
33.
to find the instantaneous velocity of the ball at time t.
b. What is the ball’s initial velocity?
c. When does the ball hit the ground, and what is its
impact velocity?
d. When does the ball have velocity 0? What physical
34. lim ∣x∣ s in - interpretation should be given to this time?
x → 0.4 ' X 60. Tom and Sue are driving along a straight, level road in a
35. lim ∣x∣ s in - car whose speedometer needle is broken but which has a
x→0 x trip odometer that can measure the distance traveled from
an arbitrary starting point in tenths of a mile. At 2:50
P.M., Tom says he would like to know how fast they are
traveling at 3:00 P.M., so Sue takes down the odometer
readings listed in the table below, makes a few calcula
tions, and announces the desired velocity. What is her
1 ~ cos*
36. lim result?
Λ∙→ 0 x
1 cosλ
37. lim ~ time t 2:50 2:55 2:59 3:00 3:01 3:03 3:06
A'→ 7Γ X
odometer
reading 33.9 38.2 41.5 42.4 43.2 44.9 47.4
2 x 2 + 3x - 1 0
38. lim x + 3 χ - 1 0 39 lim
Λ →3 X- 2 x→ 3 X - 3
40.
42.
v v
44. lim * q ^∙ 45. lim * -/
x→9 x - 9 x→ 9 x - 3
lim Vx^+^2 —2 47 . ljm ⅛ x→
46.
x→2 x -2 x→ ∣ V χ - l
I-------3 +
V x x
1 _1
48. lim -~ 49. lim V x 2
„ s. 3- x - π
50. l i m ⅛ ⅛ 51. I im ⅛ ⅛
x→0 X λ -→ 0 χ
i-_L_ ι-l θ 65. The tabular approach is a convenient device for discuss
52. hm x 1 + 53. lim χ ing limits informally, but if it is not used very carefully, it
v ^*0 x x ^, l χ - l
can be misleading. For example, for x ≠ 0, let
54. l i m ( l + x ) 1'x 55. lim(l + x ) ιz* ∕( x ) = sin ∣
2 ____ 2x ∖
56. 1imlr
" 2,000/
tanx - x
5 7 ∙ 1≡ — - i ------ a. Construct a table showing the values o f ∕( x ) for x =
—2 —2 —2 2 2 2
^X ~ , TS —, ^τx~, ^=r~∙ Based on this table, what
58. lim cos — 77 9τr 13π 19ττ ∕τ r 3τr
x→0 X would you say about lim /(x)?
59. A ball is thrown directly upward from the edge o f a cliff
b. Construct a second table, this time showing the values
and travels in such a way that I seconds later, its height
o f∕( x ) for x = 1
above the ground at the base of the cliff is τ ≡ , 5 =J-, ∕ r , J-.
2π llτr 20π 50ττ 30ττ 5π
.s(∕) = - 1 6 ∕ 2 + 40/ + 24 ft c. What conclusions can you make about lim ∕(x)7
Sec. 1.6 Properties of Limits 49
1 .6 PROPERTIES OF LIMITS
IN THIS SECTION Computations with limits, using algebra to find limits,

limits of piecewise-defined functions
It is important to be able to find limits easily and efficiently, but the methods of
graphing and table construction of the previous section are not always easy or
efficient. In this section we state some limit rules and then show how to use
these rules along with algebra to evaluate limits.
■ COMPUTATIONS WITH LIMITS
Here is a list of properties that can be used to evaluate a variety of limits.
Basic Properties and Rules for For any real number c, suppose the functions f and g both have limits at x = c.
Limits Constant rule lim k = k for any constant k
x→c
Limit of x rule lim x = c

x→c
Multiple rule lim [s∕-(x)] = s lim f(x ) for any constant 5

x→c x→c
The limit of a constant times a function is the constant times the
limit of the function.
Sum rule lim [ ∕( x ) + g(x)] = lim ∕( x ) + lim g(x)

x→c x→c x→c
The limit of a sum is the sum of the limits.
Difference rule lim [∕( x ) - g(x)] = lim ∕( x ) - lim g(x)

x→c x→c x→c
The limit of a difference is the difference of the limits.
Product rule lim [∕(x)g(x)] = Γlim ∕(x)lΓlim g(x)l

X→C ]~X→C JL*- >C J
The limit of a product is the product of the limits.
f(χ} l i m ∕W
Quotient rule lim = — ■ ■■, if lim g(x) ≠ 0
X→c g(x) hm g ( x ) χ→c tδ v ’
The limit of a quotient is the quotient of the limits, as long as the
limit of the denominator is not zero.
Power rule lim ∣∕(x ) 1” = 1lim ∕( x ) l" n is a rational number and the
limit on the right exists.
The limit of a power is the power of the limit.
Another property that will be especially important to us in our future work is

given in the following box.
Squeeze Rule If on some interval about c,

g(x) ≤ f ix ) ≤ h(x) and lim g(x) = lim h(x) = L,
x →c x →c
then lim f ( x ) = L.
x →c
What this says: If a function can be squeezed between two functions with
equal limits, then that function must also have that same limit.
These limit rules can be used to evaluate a variety of limits. We shall not prove any
of these results until Section 1.8, but it is fairly easy graphically to justify the rules
for the limit of a constant and the limit of x, as shown in Figure 1.38.
EXAM PLE 1 Finding the limit of a polynomial function
Evaluate lim (2x 5 - 9x 3 + 3x 2 - 11).

Solution lim (2x 5 - 9x 3 + 3x 2 - 1 1 ) = lim (2x 5) - lim (9x 3) + lim (3x 2) - lim (l 1)
x→2 x →2 x →2 x →2 x →2
Sum and difference rides
= 2[lim x 5] - 9[lim x 3] + 3[lim x 2 ] — 11

x →2 x →2 x →2
Multiple and constant rules
= 2 [lim x ] 5 - 9[lim x ] 3 + 3[lim x ] 2 - 11

Λ^→ 2 X →2 X →2
Power rule
= 2(2) 5 - 9(2) 3 + 3(2)2 - 1 1 = - 7
Limit o f x rule ≡
C OM M ENT : If you consider Example 1 carefully, it is easy to see that i f / i s any

polynomial, then the limit at x = c can be found by substituting x = c into the
Figure 1.38 Two basic limits formula for ∕( x ) .
Limit of a Polynomial Function If P is a polynomial function, then

lim P(x) = P(c)
EXAM PLE 2 Finding the limit of a rational function
Evaluate lim - ~ —- / + 7
z→-ι 5z 2 + 9z + 6
lim^ (z 3 - 3z + 7)
Solution lim ⅛ — 3∕ + 7 Quotient rule
2
z →-ι 5z + 9z + 6 lim (5z 2 + 9z + 6)
z→- 1
0 You must be careful about
when you write the word “lim ( - l ) 3 - 3 (-l) + 7
it” and when you do not; pay 5 ( - l ) 2 + 9 (-1 ) + 6
particular attention to this when
looking at the examples in this 9 Both numerator and denominator are polynomial
2
section. 0 functions. ⊂=
Notice that if the denominator of the rational function is not zero, the limit can be
found by substitution.
p( x ∖
Limit of a Rational Function If Q is a rational function defined by Q(x) = ^7<7 then
1⅛2« - ≡⅛)
provided Z)(c) ≠ 0.
EXAM PLE 3 Finding the limit of a power (or root) function
Evaluate lim V'x 2 - 3x - 2.

x → -2
Solution lim ∖∙zx 2 - 3x - v 2 = lim (x 2 - 3x - 2)1/3

x → -2 x → -2 ’
= [ lin~ς (x 2 —3x —2)11/3 Power rule
= [(-2 ) 2 - 3(-2) - 2 ] 1/3 = 8 1/3 = 2
Once again, for values of the function for which ∕( c ) is defined, the limit can be
found by substitution.
In the last section we found that lim sinx = 0 and lim cosx = 1 using a table.
x→0 x→ 0
In the following example we use this information, along with the properties of
limits to find other trigonometric limits.
EXAM PLE 4 Finding trigonometric limits algebraically
Given that lim sinx = 0 and lim cosx = 1, evaluate:

X→ 0 X→ 0
a. lim sin 2 x b. lim (1 - cosx)
X→ 0 X→ 0 ’
Solution a. lim sin 2 x = [lim sinx∣ 2 Power rule

x→o LΛ-→O J
= 02 lim sin x = 0
x→0
= 0
b. lim (1 - cosx) = lim 1 - lim cosx Difference rule
x→ 0 x→ 0 x→0
Constant rule and
lim cosx = 1
x →0
The following theorem states that we can find limits of trigonometric functions by
direct substitution, as long as the number that x is approaching is in the domain of
the given function. The proof of the theorem makes use of the limit formulas
lim sinx = 0 and lim cosx = 1.
.t→ 0 x→0
THEOREM 1.3 Limits of trigonometric functions
If c is any number in the domain of the given trigonometric function, then

lim cosx = COSC lim seex = secc
lim sinx = sine lim esex = cscc
x—>c x—>c
lim tanx = tanc lim cot x = cot c
x →c
Proof: We shall show that lim sinx = sine. The other five limit formulas may
χ →c
be proved in a similar fashion (see Problems 71-72). Let h = x - c. Then
x = h + c, and as h → 0, x → c. Thus,
lim sinx = lim sin(Λ + c)
x →c h→0
Using the trigonometric identity sin (A + B) = sin√4 cosB + cosA sinB and the
limit formulas for sums and products, we find that
lim sinx = lim sin(Λ + c)
x →c h→0
= lim [sinZz cose + cosh sine]
= lim sin A ∙ lim cose + lim cosh ■lim sine
h→0 h→0 lι→0 h→0
= 0 ■cose + 1 ∙ sine lim sinΛ = 0 and
h→0
lim cosh = 1
= sinc h^*0
Note that sin c and cos c do not change as h → 0 because these are constants with
respect to h. ■
■ USING ALGEBRA TO FIND LIMITS
Sometimes the limit of∕( x ) a s x → c cannot be evaluated by direct substitution. In

such a case, we look for another function that agrees with f for all values of x
except at the troublesome value x = c. We illustrate with some examples.
EXAM PLE 5 Evaluating a limit using fraction reduction
2
Evaluate lim x x+ -x 2- 6
x →2
Computational Window B ≡ Solution If you try substitution on this limit, you will obtain:
-- If x = 2, then x 2 + x - 6 = 0
lim x + x ~ 6
x→ι x z ^ 1f x = 2, t h e n x - 2 = 0
The form 0/0 is called an indeterminate form because the form does not help us
determine the limit of the expression. In other words, we cannot evaluate a
limit for which direct substitution yields 0/0.
If the expression is a rational expression, the next step is to simplify the
function by factoring and simplifying to see if the reduced form is a
YιB<X2+X-6)z(X-2) polynomial.
Xnin= - 1Θ V∩in=-1Θ
X∩ax=19 V∩ax=lθ lim x ~ + x ~ 6 = lim + 3 ) ( * ~ 2 ) = lim (x + 3)
Xscl=l Vscl=l .r→2 X - 2 .v→2 X - 2 χ →2 v ’
This simplification is valid only if x ≠ 2. Now complete the evaluation of the

reduced function by direct substitution. This is not a problem, because lim is

concerned with values as x approaches 2, not the value where x = 2. λ - *^
lim 6 = lim (x + 3) = 5
x→ 2 X ~ 2 χ→ 2
Another algebraic technique for finding limits is to rationalize either the

numerator or the denominator to obtain an algebraic form that is not indetermi
nate.
EXAM PLE 6 Evaluating a limit by rationalizing
Evaluate lim ,2
x →4 x - 4
0 This method will work only Solution Once again, notice that both the numerator and denominator of this
if the resulting numerator al rational expression are 0 when x = 4, so that we cannot evaluate the limit by
lows the fraction to be direct substitution. Instead, we multiply by 1, its form chosen so that the
simplified. 0 numerator is rationalized.
lim 2 = lim ∖f x - 2 V x + 2 Multiply by 1.

x→ 4 X - 4 X→ 4 x - 4 Vx + 2
x →4 (χ—-4)(√χ + 2)
lim —FJ— -
√X + 2
-v→ 4
i = 1
√4 + 2 4
E X A M P L E 7 Evaluating a trigonometric function by multiplying

2 4 6 δ 10
by the conjugate
Evaluate lim -— ∞ sA
%→o x
Solution Both the numerator and the denominator of this quotient are 0 when
This was Problem 36 of the pre x = 0. However, if we multiply both the numerator and denominator of ∕( x )
vious section. by 1 + cosx (the conjugate), we can rewrite the quotient 1 ~ ^.0 s -- in a form in
which its limit can be found.
lim lim(-l - c o s -4
1 + cosx
x→ 0 X x→ 0∖ X 1 + cosx,
0.8
= lim 1 - cos2 x
.r→ () x (l + cosx)
= lim sin 2 x 1 - cos2 x = sin2 x

Λ→ 0 x( 1 + cos x)
= lim Γ/sinx∖( sin x ∖1

x → 0 L∖ x /∖1 + cosx/J
= l i m p M lim sinx ∖ Remember lim s l n ^λ = [ and

x → 0∖ X / x→ 0 1 + cosx/ I →0 x
evaluate the other limit by direct
= 1-0 substitution o f sin 0 = 0 and
=0 cos 0 = 1. ___
■ LIMITS OF PIECEWISE-DEFINED FUNCTIONS
In Section 1.4 we defined a piecewise-defined function. To evaluate

lim fix')
x→c
where the domain o f/ is divided into pieces, we first look to see if c is a critical
value separating two of the pieces. If so, we need to consider one-sided limits, as
illustrated by the following examples.
EXAMPLE 8 Limit of a piecewise-defined function
x + 5 if x > 0
Find lim ∕( x ) where f(x ) =
x if x < 0
Solution Notice that /(0) is not defined, and that it is necessary to consider left-
and right-hand limits.
lim_ ∕( x ) = lim x ∕(χ ) = χ to the left of0.
= 0
lim+ ∕( x ) = lim+ (x + 5) ∕(x ) = x + 5 to the right of0.
= 5
Because the left- and right-hand limits are not the same, we conclude that
lim ∕( x ) does not exist. ⊂z≡ι
EXAM PLE 9 Limit of a piecewise-defined function
f x + 1 if x :> 0
Find lim g(x) where g(x) = (
r
∙→ 0 (x 2 + 1 if x <; 0
lim g(x) = lim (x 2 + 1) = 1
ι →0 x →0
λ⅛ + g(x) = Jim + (x + 1 ) = 1
Because the left- and right-hand limits are equal, we conclude that lim g(x) = 1.
x→0
PROBLEM SET 1 .6
© In Problems 1-30, evaluate each limit. /x 2 - 3x + 2∖ 2
15. lim 16. lim V ∕x -x 2x
- 3
- 3
1. lim (x 2 + 3x - 7) 2. lim -(Z3 5z2
+ 4) x→ l ∖x2 + x - 2 / x→3
∕ →0 . .
lim f l + —=~ ) fix - 1 Vy + 2 - 2
3. lim (x + 5)(2x - 7) 4. 17. lim 18. lim
x →4 ∖x x - 5/ .t→ l x - 1 r→2 + -2
2 1 — sinx 20. lim z 2 cosx
1 -
6. hm x + 3 χ -1 0
z2 + 3 19. lim
5. lim ∕ /
x→0 cos2 x x→0 V 3 - 2sinx
χ→l Z + 1 .v→ 3 3 χ- + 5x - 7
21. lim sin 2x 22. lim sin 3x
1 + t a n -'
7. lim secx 8. lim X→ 0 X x→0 2x
x → π∕3 x→π∕4 cscx + 2 tanx 2
,. tan(∙7τ∙∕x) 23. lim 24. lim sin 2 x
xsinτrx X→ 0 X x→0 x
9. lim 10. h m -------- r -
x → ι∕3 1 + cos τrx x→6 x - 1 1 1
25. lim x + 3 3 26. lim x____ 3_
x 4x + 4 x→0 X x→3 x - 3
11. lim 12. lim ∖~
u -> -2 2+ U x→2 Xλ - X - 2 27. lim secx — 1 28. hm . ~
1 ta n x
x → 0 x secx χ→π∕4 sin x - cosx
1 - 1 (x + I) 2 - 1 2
13. hm χ____ _ 14. rlim ----------------
x -0 ∣ X 29. lim tan x 30. lim tanx
*→1 x - 1 x→0 X 1 + secx
O 31. ■ What Does This Say? How do you find the limit of a 60. T H IN K T A N K P ROBLEM Evaluate
polynomial function?
COS X
32. ■ What Does This Say? How do you find the limit of a lim x 2
x→oL 1,000,000,000
rational function?
Explain why a calculator solution might lead you to an
33. ■ What Does This Say? How do you find lim -s ι n αλ∕
.≈→ 0 X incorrect conclusion about the limit.
In Problems 34-41, compute the one-sided limit. In the next chapter we will formally define the quantity
34. lim (x 2 —2x) 35. l im ∕-l÷Y ∆ ∕~ ∕( x + ∆x) - f i x )

∆x ∆x
l*l
36. lim 37. rlim+ — to be the difference quotient o f a function f. The number ∆ x is
x→0^ x x→0 x
3 - 2x if x ≤ 2 an arbitrary number (usually assumed to be very small). In
38. lim fix ) where fix ) = < 2 each o f Problems 61-64, first fin d ∆f∕∆xι then compute*
x→2~ I x - 5 if x > 2
∫3 - 2x if x ≤ 2
39. lim+ ∕( x ) where f ix ) = l x 2 - 5 if x > 2 Δx→ 0 ∆ X
if S < 1 61. ∕( x ) = 3x - 5 62. ∕( x ) = x 2

40. lim g(s) where g(s) = ∙
V l - S if s ≥ 1 63. f ix ) = ∣ 64. f ix ) = f i x
R ⅛ ifs < 1
41. lim+ g(s) where g(s) = s - 1 Verify the limit statements, in Problems 65-66. You may use
s→1
wVl - J if .s ≥ 1 lim s *n u = 1 and lim . u = 1 without rproof
■ What Does This Say? In Problems 42-49, explain why the u→o u u →o sin u
given limit does not exist. 65. lim s ⅜ n = τ for constants a, b, with b ≠ 0
x → o sin bx b
42. lim — l - γ 43. lim+ , - - - -
x→l x ~ 1 x→2 √ x - 2 66. lim *a n a x = τ for constants a, b, with b Ψ 0
x→o tan bx b
∕2 -4 45. l im ^ θ 67. Let ∕( x ) = Λ j with x ≠ 0, and let L be any fixed positive
44. lim
t→ 2 t2 - 4t + 4 x→0 X
integer. Show that
46. lim ∕( x ) where ∕( x ) = 5 < J
f ix ) > 100L if ∣x∣ <
47. firn g(t) where g(t) = ^ + 1
2
What does this imply about lim ∕(x )7
48. lim tan x 49. lim cscτrx
x → τr∕2 x→ l 68. T H IN K T A N K P ROBLEM Give an example for which
We can use algebra, graphs, or tables to fin d limits. Find the neither lim f ix ) nor lim g(x) exists, but lim [∕(x) + g(x)]
x →c x→c x→c
limits in Problems 50-53. does exist.
50. lim 51. lim x ~ s∕ n x
69. T H IN K T A N K P RO BLEM It is not necessarily true that
x→3 X - 3 1 →0 χ 3
lim ∕(x ) and limg(x) exist whenever lim [∕(x) ∙ ⅛ , (x)] or
x r c f ix ) . w
52. lim 53. lim ~-,-1-
χ→l X - 1 χ->l X - 1 lim —τ-τ exists.
X→c g(x)
In Problems 54-59, either evaluate the limit or explain why it a. Find functions/and g such that lim [fix) ■ ^(x)] exists
does not exist. and lim g(x) = 0, but lim ∕(x ) does not exist,
x→c x→c .
54. lim χ ~ 1 55. lim ( - - 4 5 ) b. Find functions f and gs such that lim —f i exists, but
*→' √ χ - 1 χ →0 ' λ' x , x → 0 g(x)
neither lim f ix ) nor lim g(x) exists.
56. lim ∕( x ) where ∕( x ) = + 3 ∙f * ≠ 5
70. Use the sum rule to show that if lim [∕(x) + g(x)] and
57. lim g(t) where g(∕) = _ ~J . ≤ ' < 2 lim f ix ) both exist, then so does limg(x).
2
71. Show that lim cosx = cosx 0 .
f 2(x + 1) if x < 3
Hint’. You will need to use the trigonometric identity
58. lim ∕( x ) where ∕( x ) = <4 if x = 3
lx 2 - 1 if x > 3 cos¢4 + B) = cosΛ cosB — sin√4 sinB.
72. Show that lim tanx = tanx 0 whenever cosx 0 ≠ 0.
f 2(x + 1) if x < 3 Λ'→X0
59. lim ∕( x ) where ∕( x ) = <j 4 if x = 3 *This kind of limit will be used to define an idea called a
2
lx - l ifx > 3 derivative in Section 2.1.
1 . 7 CONTINUITY
IN THIS SECTION Intuitive notion of continuity, definition of continuity,

continuity theorems, continuity on an interval, the intermediate value theorem,
approximation by the bisection method
Informally, a continuous function is one whose graph has no “breaks” or
“jumps.” However, do not be misled by this simple geometric characterization.
Continuity is an important and complex concept that plays a central role in the
further development of calculus.
■ INTUITIVE NOTION OF CONTINUITY
The idea of continuity may be thought of informally as the quality of having parts
that are in immediate connection with one another. The idea evolved from the
vague or intuitive notion of a curve “without breaks or jumps” to a rigorous
definition first given toward the end of the 19th century (see the historical note on
page 59).
We begin with a discussion of continuity at a point. It may seem strange to talk
about continuity at a point, but it should seem natural to talk about a curve being
“discontinuous at a point,” as illustrated by Figure 1.39.
a. Continuous at x = 1
e. Discontinuous at x = 1
(Sometimes a pole is called
a vertical asymptote.)
Figure 1.39 The curves in b, c, d, e, and f have a discontinuity at the point x = 1.

Figure 1.39 shows some curves that illustrate the idea of continuity from an
intuitive standpoint, but we need to define this idea more precisely.
Sec. 1.7 Continuity 57
■ DEFINITION OF CONTINUITY
Let us consider the conditions that must be satisfied for a function f to be

continuous at a point c. First, ∕( c ) must be defined. For example, the curves in
Figures 1.39b and 1.39e are not continuous at x = 1 because they are not defined
for x = 1. (An open dot indicates an excluded point). A second condition for
continuity at a point x = cis that the function makes no jumps there. This means
that if “ x is close to c,” then “/(x) must be close to/(c).” This condition is satisfied
if lim ∕(x ) exists and is equal to∕( c ) . Looking at Figure 1.39, we see that the graphs
in 'parts c, d, and f have a jump at the point x = 1.
Continuity of a Function at a A function/is continuous at a point x = c if
1. ∕( c ) is defined;
2. lim ∕( x ) exists;
3. lim ∕( x ) = ∕( c ) .
x →c
A function that is not continuous at c is said to have a discontinuity at that

point.
■ What this says: Step 1 refers to the domain of the function and ignores
what happens at points x ≠ c, whereas step 2 refers to points close to c, but
ignores the point x = c. Continuity looks at the whole picture — at x = c and
at points close to x = c — and checks to see if they are somehow “ alike.”
If /is continuous at c, the

points (x,∕(x)) converge to If / is continuous at x = c, the difference between ∕( x ) and ∕( c ) is small
(c,∕(c)) as x→c.
whenever x is close to c because lim∕( x ) = ∕( c ) . Geometrically, this means that the
Figure 1.40 The geometric inter
points (x ,∕(x )) on the graph of/converge to the point (c,∕(c)) as x → c, and this is
pretation of continuity
what guarantees that the graph is unbroken at (c,∕(c)) with no “ gap” or “ hole,” as
There are three common ways for a function to be discontinuous at x = c. (A
fourth way a function can be discontinuous is by oscillation.) These can be
remembered as holes, poles, and jumps and are summarized in Table 1.4.
EXAM PLE 1 Testing the definition of continuity with a given function
Test the continuity of each of the following functions at x = 1. If it is not

continuous at x = 1, explain.
*∙ ∕W = *2 χ-x i 3
b. g(x) = λ + 2-λ ]~ 3 if χ ≠ ι and β if _
= χ 1
c. ∕z(x) = - - " ¾ ~ 3 if x ≠ 1and ∕ι(x) = 4 if x = 1
d. F(x) = * * ∣ if x ≠ 1 and ∕ , '(x) = 4 if x = 1
e. G(x) = 7x 3 + 3x 2 - 2
f. H(x) = 2 sin x - tanx
Solution We verify that the three criteria for continuity are satisfied for c = 1.
Historical Note
The Navajo are a Native Amer
a. The function/is not continuous at x = 1 (hole;/(c) not defined) because it
ican people who, despite con
siderable interchange and as is not defined at this point.
similation with the surrounding b. 1. g(l) is defined; g(l) = 6.
dominant culture, maintain a 2. lim g(x) = lim x " + 2 x 1~ 3
world view that remains vital
and distinctive. The Navajo be
lieve in a dynamic universe.
Rather than consisting of ob ,v→i x- 1
jects and situations, the uni = lim(x + 3) = 4
verse is made up of processes.
Central to our Western mode of 3. lim g(x) ≠ g(l), so g is not continuous at x = 1 (hole; g(c) defined).
thought is the idea that things c. Compare h with g of part b. We see that all three conditions of continuity
are separable entities that can are satisfied, so h is continuous at x = 1.
be subdivided into smaller dis
crete units. For us, things that
change through time do so by TABLE 1.4 Holes, Poles, Jumps, and Continuity
going from one specific state to
another specific state. While we ∕( c j Defined fic ) Not Defined
believe time to be continuous,
Not continuous at x = c: Hole Not continuous at x = c: Hole
we often even break it into dis
crete units or freeze it and talk lim f(.xj exists and is not equal to fie). lim fix ) exists.
about an instant or point in
time. Among the Navajo, where
the focus is on process, change
is ever present; interrelation
ship and motion are of primary
significance. These incorporate
and subsume space and time.
From Ethnomathematics by Mar

cia Ascher, pp. 128-129.
Not continuous at x = c: Pole Not continuous at x = c: Pole
lim fix ) does not exist. lim fix ) does not exist.
Continuous at .v = c:
lim ∕(x) exists and is equal to fic ).
Not continuous at x = c: Jump Not continuous at x = c: Jump

lim fix ) does not exist. lim fix ) does not exist.
lim∕'(x) = L and
lim ∕(x) = R
d. 1. F(l) is defined; F(l) = 4.

Historical Note
2. lim F(x) = lim y _ y ; the limit does not exist.
The function F is not continuous at x = 1 (pole).
e. 1. G(l) is defined; G(l) = S.
2. lim G(x) = 7(lim x) 3 + 3(lim x) 2 - lim 2 = 8.
x→ l x→ l x→ l .r→ l
3. lim G(x) = G(l).

Because the three conditions of continuity are satisfied, G is continuous at
x = 1.
f. 1. 77(1) is defined; //(1) = 2 sin 1 —tan 1 ≈ 0.1255 (approximate value).
2. lim H(x) = 2 lim sinx - lim tanx = 2 sin 1 - tan 1 (exact value).
3. lim ∕∕(x ) = 77(1).
x→ l
K ARL W EIERSTRASS Because the three conditions of continuity are satisfied, H is continuous at
(1815-1897) x = 1. MB
The idea of continuity evolved
from the notion of a curve ■ CONTINUITY THEOREMS
“without breaks or jumps” to a
rigorous definition given by It is often difficult to determine whether a given function is continuous at a
Karl Weierstrass. Our definition specified number. However, many common functions are continuous wherever
of continuity is a refinement of they are defined.
a definition first given by
Bernhard Bolzano (1781-1848). THEOREM 1.4 Continuity theorem
Galileo and Leibniz had
thought of continuity in terms I f /is a polynomial or a rational function, a power function, or a trigonometric
of the density of points on a function, th e n /is continuous at any number x = c for which ∕(c ) is defined.
curve, but using today’s stan
dards we would say they were Proof: The proof of the continuity theorem is based on the limit properties
in error because the rational stated in the last section. That is, a polynomial is a function of the form
numbers have this property, yet
do not form a continuous ∕ , (x) = anx n + αn -1 x"^ 1 + . . . + α l x + α0
curve. However, this was a dif where a0 , a l , . . . an are constants. We know that limα0 = a0 and that limxw = cm
ficult concept, which evolved
over a long period of time. An for m = 1, 2, . . . , n. This is precisely the statement that the function g(x) = ax m
other mathematician, J.W.R. is continuous at any number x = c, or simply continuous. Because P is a sum of
Dedekind (1831 -1916), took functions of this form and a constant function, it follows from the limit properties
an entirely different approach that P is continuous.
to conclude that continuity is The proofs of the other parts (rational, power, and trigonometric functions)
due to the division of a seg follow similarly. ■
ment into two parts by a point
on the segment. As Dedekind The limit properties of the last section can also be used to prove a second
wrote, “By this commonplace continuity theorem. This theorem tells us that continuous functions may be
remark, the secret of continuity combined in various ways without creating a discontinuity.
is to be revealed.”
From Carl Boyer, A History o f THEOREM 1.5 Properties of continuous functions

Mathematics (New York, John Wiley &
Sons, Inc., 1968), p. 607. If 5 is a real number (called a scalar) and/ and g are functions that are continuous
at x = c, then the following functions are also continuous at x = c.
Scalar multiple sf
Sum and difference f + g and f - g
Product fg
Quotient provided g(c) ≠ 0

Composition f og provided/is continuous at g(c)
Proof: The first four properties in this theorem follow directly from the basic
limit rules given in Section 1.6. For instance, to prove the product property, note
that since f and g are given to be continuous at x = c, we have
lim ∕(x ) = ∕(c ) and Inn g(x) = g(c)
If P(x) = f(x)g(x), then
lim P(x) = lim ∕(x)g(x) Product rule for limits
Continuity off and g at x = c

= f(c)g(c) = P(c)
so P(x) is continuous at x = c, as required.
The continuous composition property is proved in a similar fashion but
requires the following limit rule, whose proof is outlined in Problem 25 of Section
1.8.
Composition Limit Rule: If f is continuous at L and lim g(x) = L, then

lim ∕(g(x)) = /(£ ) = ∕(lim g(x))
x →c x →c
Now we can prove the continuous composition property of Theorem 1.5. Let
∕z(x) = (∕og)(x). Then we have
lim ∕z(x) = lim (∕□g)(x) = lim ∕[g(x)] Definition of composition
= /[lim g(x)] Composition limit rule
~ ∕L ^( c )] S is continuous at x = c.
= U°gKc) = h(c) ■
What this says: Roughly speaking, the composition limit rule says that the
limit of a continuous function is the function of the limiting value. The
continuous composition property says that a continuous function of a continu
ous function is continuous.
We need to talk about a function being continuous on some interval. In order

to do so we must first know how to handle continuity at the endpoints of the
interval, which leads to the following definition.
One-Sided Continuity The function/is continuous from the right at a if and only if
lim ∕(x ) = f{d)
x→a
and it is continuous from the left at b if and only if
lim ∕(x ) =∕(Z>)
X→D
■ CONTINUITY ON AN INTERVAL
The function /is said to be continuous on the open interval (a, b) if it is continuous
at each number in this interval. If/is also continuous from the right at a, we say it
is continuous on the half-open interval [a, b). Similarly,/is continuous on the half
open interval (a, ⅛] if it is continuous at each number between a and b and is
continuous from the left at the endpoint b. Fin ally,/is continuous on the closed
interval [a, ∕>] if it is continuous at each number between a and b and is both
continuous from the right at a and continuous from the left at b.
E X A M P L E 2 Testing for continuity on an interval
Solution
a. Function ∕ 1 is not defined when x 2 - 4 = 0 or when x = 2 o r x = - 2 . The
curve is continuous on (—∞, —2) U ( - 2 , 2) U (2, ∞).
b. Function / is continuous on (-∞ , ∞).
c. The cosecant function is not defined at x = nττ, n an integer. At all other
points it is continuous. Thus ∕ 3 is continuous on
. . . U (-2-ττ, —-ττ) U (—-77, 0) U (0, -77) U (-77, 277) U . . .
d. Because l/x is continuous except at x = 0 and the sine function is
continuous everywhere, we need only check continuity at x = 0.
lim sin - does not exist.

Λ→0 x
Therefore, ∕( x ) = s i n i s continuous on (-∞ , 0) U (0, ∞).

e. It can be shown that
-∣x∣ ≤ x sin^ ≤ ∣x∣, x ≠ 0
(see Figure 1.41 to see the plausibility of this inequality). We can now use
the squeeze rule. Because lim ∣x∣ = 0 and lim(-∣x∣) = 0, it follows that
Figure 1.41 The graph o f
lim x s in - = 0. Because /(0)
5
= 0 we see that /is continuous at x = 0 and
y = sin - and y = —∣x∣ x→o x
therefore is continuous on (-∞ , ∞). M M
Because we do not always have the graph of a function readily available, as we

did in Example 2, and because the task of checking for continuity will focus on
certain values, we consider a procedure involving identifying and then checking
those values of concern. To help us describe the situation, we define a suspicious
point as a point having an x- value for which the definition of the function changes,
or a value that causes division by zero for the given function.
For Example 2 the suspicious points can be listed:
x *12 — 1
a. — - has suspicious points for division by zero when x = 2 and x = —2.
-X ι
b. ∣x - 4∣ = x 2 - 4 when x 2 - 4 ≥ 0 and ∣x2 - 4∣ = 4 - x 2 when x 2 - 4 < 0.
2
This means the definition of the function changes when x 2 - 4 = 0,

namely, when x = 2 and x = - 2 .
c. The points x = nττ are suspicious points because these are the values for
which the function is not defined.
d. sin-, has a suspicious point when x = 0 (division by 0).
e. EIere, the only suspicious point is when x = 0.
EXAM PLE 3 Checking continuity at suspicious points
τL te /■
/ ∫3 - x if - 5 ≤ x < 2 , .. 12 - x if - 5 ≤ x < 2
t »λ . [ x _ 2 i f 2 a λ , < 5 and g(x) = ( χ . 2 l f 2s x < 5
Find the intervals on which f and g are continuous.

Solution The domain for both functions is [ -5 , 5); the continuity theorem tells
us both functions are continuous everywhere on that interval except possibly
at the suspicious points. Exam ining/ we see
f( x ) = 3 - x on [ -5 , 2), which is a polynomial function and thus is
continuous
f(x ) = x - 2 on [2, 5), which is a polynomial function and thus is
continuous
The suspicious point on the real number line is the x -value for which the
definition of f changes— in this case, x = 2.
Although a graph is not part of the
derivation, it can often be helpful in For g we likewise see that the function is continuous except possibly at the
finding suspicious points. This is the suspicious point x = 2, the value where the definition of the function changes.
graph off.
1. Because /(2) = 2 - 2 = 0 and g(2) = 2 - 2 = 0, both functions are defined

at x = 2.
2. The second condition for continuity requires that we find the limits of/ and
g as x → 2. To find this limit we need to check both the left- and right-hand
limits because the value of the function is defined by a different equation if
we approach from the left or from the right.
lim ∕( x ) = lim (3 - x) = 1 and lim+ ∕( x ) = lim (x - 2) = 0
Thus lim ∕( x ) does not exist, because the left- and right-hand limits are not
x →2
the same.
This is the graph of g. Does the graph lim g{x) = lim (2 - x) = 0 and lirn g(x) = lim (x - 2) = 0
reinforce our conclusions?
Therefore, lim g(x) = 0.
3. The third condition of continuity cannot hold at x = 2 for the function/

because the required limit in step 2 does not exist. It is continuous on
[ -5 , 2) and on [2, 5), but not on the entire interval [ -5 , 5). The function g
also satisfies the third continuity property, namely, that lim g(x) = g(2); g
therefore is continuous on [ -5 , 5). A-> ∣gsαa
■ THE INTERMEDIATE VALUE THEOREM
Intuitively, i f / is continuous throughout an entire interval, its graph on that

interval may be drawn “ without the pencil leaving the paper.” That is, if ∕( x )
varies continuously from f(a) to f(b) as x increases from a to b, then it must hit
every number L between ∕(α ) and f(b ), as shown in Figure 1.42.
Here is an example illustrating the property shown in Figure 1.42. Suppose/is
a function defined as the weight of a person at age x. If we assume that weight
varies continuously with time, a person who weighs 50 pounds at age 6 and 120
pounds at age 15 must weigh 100 pounds at some time between ages 6 and 15.
This feature of continuous functions is known as the intermediate value property.
A formal statement of this property is contained in the following theorem.
Figure 1.42 If L lies between ∕(α )
and f(b), th e n ∕(c ) = L for some c
between a and b. T H E O R E M L6 The intermediate value theorem
I f / is a continuous function on the closed interval [a, b] and L is some number

strictly between f(a) and /(6), then there exists at least one number c on the open
interval (a, b) such that ∕( c ) = L.
Proof: This theorem is intuitively obvious, but it is not at all easy to prove. A
proof may be found in most advanced calculus textbooks. ■
I What this says: If/ is a continuous function (with emphasis on the word
continuous) on some closed interval, and if x takes on all values between a and
b, th en ∕(x) must take on all values between ∕(α ) and ∕(6 ). Γm reminded of my
calculus instructor who asked me my height, “ Six foot,” I answered. “ This
means,” said he, “ that at one time you were exactly 2 ft tall; why?” The
answer, of course, is because of the intermediate value theorem.
One important consequence of the intermediate value theorem is the fact that
it allows us to find, or estimate, roots of the equation f(x ) = 0. If, for example,
there is some positive value of a continuous function /(the graph of/at this point
is above the x-axis) and another value of /is negative (the graph of/at this point is
Figure 1.43 Because f(a ) > 0 and
below the x-axis), then there must be at least one number c between a and b for
/(6) < 0, then f(c ) = 0 for some c
which ∕( x ) = 0 (the graph crosses the x-axis). This is shown in Figure 1.43 and is
between a and b.
summarized in the following theorem.
T H E O R E M L7 Root location theorem
I f/ is continuous on the closed interval [a, b∖ and i f ∕( α ) and/(6) have opposite

algebraic signs (one positive and the other negative), then∕(c) = 0 for at least one
number c on the open interval (a, b).
Proof: This follows directly from the intermediate value theorem (see Figure
1.43). The details of this proof are left as a problem, ■
EXAM PLE 4 Using the root location theorem
Show that x 4 - 2Vx = 1 for at least one number c on [- 1, 1].

Solution Notice that the function f(x ) = x 4 - 2 ∖ ∕ - 1 is continuous on [ - 1 , 1]∙
Also notice that/ (-1 ) = 2 (it is positive) and/(1) = - 2 (it is negative), so the
conditions of the root location theorem apply.
Therefore, there is at least one number c on ( -1 , 1) for which /(<?) = 0.
This means that c 4 - 2Vzc - 1 = 0, or, equivalently, c4 - 2Vc = 1 as re
quired. MM
The intermediate value theorem and the root location theorem are called
existence theorems. That is, they tell us certain values exist, but they do not give us
a method for finding c. We conclude this section by giving one possible method for
finding c.
■ APPROXIMATION BY THE BISECTION METHOD
The intermediate value theorem is the basis for a procedure that can be used to
approximate roots of equations to any desired degree of accuracy. Suppose/is a
continuous function on the closed interval [a, ⅛] with, for example, f(a ) > 0 and
/(6) < 0. Then, by the intermediate value theorem, the equation ∕'(x) = 0 has a
root somewhere between x = a and x = b. To estimate the location of this root
with more precision, evaluate / at the midpoint
_ a+ b
xr ∖ 2
of the interval [a, b]. There are three possibilities, as shown in Figure 1.44.
Figure 1.44 Approximation of a root by bisection
Either x l is the root or you have found a new interval, one that is only half as wide
as the original interval and that must also contain the desired root. This is shown
in Figure 1.45.
As you consider Example 5, notice that at each stage of the procedure, the
midpoint x n of the interval approximates the root with a maximum error E n that is
no greater than the width of the (n + l)th interval.
E X A M P L E 5 Using the bisection method
The equation x 3 - x 2 - 1 = 0 has one real root. Find this root with an error no
Figure 1.45 The bisection method greater than γg.
Computational Window E ≡ Solution The goal is to find a root of the equation ∕( x ) = 0, where
f(x ) = x 3 - x^λ - 1
You might think that calcula
tors have eliminated the need In other words, you are to find the value(s) of x for which f(x ) = 0. Note that
for the algorithm described /(1) = 13 - 1 2 - 1 = -1 < 0 and/(2) = 2 3 - 2 2 - 1 = 3 > 0. Therefore, the
here, but that is not the case real root must lie on the interval [1, 2] = [a, b], with positive values of∕( x ) on
For example: the (TRACE) the right and negative values on the left.
gives you the approximate val
ues where the graph crosses the
Step 1: Apply the bisection method by assigning the midpoint of the interval
[1, 2] to x { : x l = 1 ξ ~= ∣ = 1.5. Because∕( x l ) = ∕( j ) = 1.53 - 1.52 - 1 =
X∩in=-4 Vnin="10
Xnax=4 Vnax=10 0.125, we see t h a t∕( x 1) > 0. Therefore, j replaces 2 as the right, positive
Xscl=l Vscl=l
8=1.4l ifiB0BE
V=-.0647155 yielding endpoint of the interval, and the root must be on [1, The first
8=1.531B1H9 Y=.2HBZ7B3≥
estimate is x 1 = 1.5 and the error is no more than ∣ (that is, the length of the
The root lies between
X = 1.4468085 and interval).
X = 1.5319149 Step 2: We know that at the left endpoint,
In order to obtain a bet a = 1,/(1) < 0 and at the right endpoint,
ter-approxim ation, use the x
(ZOOM). These extra zoom op
∖ = 2’ ∕( ∙ χ ∣) ~ 0.125 > 0. Therefore, we
erations can be viewed as will assign the midpoint of this interval to x 2:
equivalent to the iterations of 1+ ∣ 5
the bisection method. x 2 = —2— = = 1.25
Now calculate∕( x 2) = ∕( f ) = -0.609375.

Because this value is negative, ⅜ replaces 1 as the left, negative-yielding
endpoint of the interval, which is narrowed down to ∣J> j ] . The second
estimate is x 2 = 1.25, and the error is no more than ∣.
Step 3: Because the root must lie on the interval [∣> j l , that is where we will
seek the next estimate.
■J | J
x , = i - r - = ⅛ ∕( x 1) = -0.291015625 < 0
-2 Z o - *
The third estimate is x 3 = ⅛ = 1.375 and the error is no more than ∣.

11 | J
Step 4: On [⅛> j] we find x 4 = = γ ∣ l ∕( x 4) ≈ -0.095947265 < 0
The fourth estimate is x 4 = γ∣ = 1.4375 and the error is no more than -⅛, as
requested in the problem. We estimate the root to be 1.4375. ≡≡≡
Convergence o f the Bisection Method When you estimate a root such as the
one in Example 5, be sure you are not misled about the accuracy of the answers.
When you write 1.4375, the error could be as large as 0.0625. Here is what we
mean: If r represents the root, x n is the estimate of the root, and εn is the maximum
error for the wth estimate, then x n - ε < r < x n + ε. For this answer we see that
the root is contained in the interval (x4 ~ ⅛, * 4 + ⅛) = (1-375, 1.500).
The approximations x l , x 2, . . . generated by the bisection method converge
on the true value of the desired root. However, the rate of this convergence is often
fairly slow when compared to other methods. After four steps in Example 5, for
example, the maximum possible error is E4 = 0.0625. After n steps, the maximum
error would be 1/2", and more than 14 steps would be required to guarantee four-
decimal-place accuracy.
PROBLEM SET 1 .7
secx
θ Which o f the functions described in Problems 1-6 represent 19. hfx) = cscx cotx 20. f ix ) =
sinx tanx
continuous functions? State the domain, i f possible, for each
example.
1. the temperature on a specific day at a given location
considered as a function o f time
2. the humidity on a specific day at a given location consid
ered as a function of time
3. the selling price of ATT stock on a specific day considered
as a function of time
4. the number of unemployed people in the United States In Problems 21-26, the given function is defined everywhere
during the month o f January 1995 considered as a func except at x = 2. In each case, fin d what value should be
tion o f time assigned to /(2), i f any, in order to guarantee that f will be
continuous at 2.
5. the charges for a taxi ride across town considered as a
function of mileage 21 ∕(x) - Λ
2 ∕ J 2^ 22. √(.v) - f f f ∆
6. the charges to mail a package considered as a function of
its weight 2 3 .∕ ( x ) - ⅛ ⅛ > 24,∕W - ⅛
Identify all suspicious points and determine all points o f ~- , l2 x + 5 if x > 2 „. .. , ⅜- 1
discontinuity in Problems 7-20. 2 5..∕f l W = ( ∣
5 - jt 2 2 26,∕W - J - 2
7. f ix ) = x 3 - 7x + 3 8. f ix ) = ∣ ∣ ⅛ 7
In Problems 27-34, determine whether or not the given func
3x tion is continuous on the prescribed interval.
Q
Jf (W√ i 1U. J(t) 3 P + ZQ
x2 - x 2 7 . ∕( x ) = ∣ o n [ l , 2 ] 28. fix ) = ∣ on [0, 1]
A A
11. Λ(x) = √ x + ∣ 12. flu ) =Ψ u2 - 1
29. ∕( t ) = ∣ on [ - 3 , 0) 30. g(x) = τ 4 ^ o n [ 0 ,7 )
13. fit) = ] ______ 3 f(y∖ = [χ2 if 0 ≤ χ < 2
t÷ 1
sJI I V1
' 7 w ι,3x + 1 if 2 ≤ x < 5
15. fix ) =
⅛1 IΛ 1 ,,
3 - r, ∣ f ∖ - I-t if 0 < t ≤ 3
’ 115- t 2 i f -3 ≤ t ≤ 0
^3t2 - 1 if t > 3
16. g f) = ( 5 ifl < t ≤ 3 33. f ix ) = x sin x on (0, ττ) 34. ∕( x ) = ON ∣-Q , 7Γ J
.31 + 2 if t ≤ 1
φ 35.w ω ={j
Show that/ + g is continuous at x = 0 even though/ and g
are both discontinuous there.
36. Show that iff/g and g are both continuous at x = c, then /
must also be continuous there.
37. T HINK . T A N K P ROBLEM Find functions/ and g such that
/ is discontinuous at x = 1 but ∕ g is continuous there.
38. T H IN K T A N K P RO BLEM Give an example of a function
defined for all real numbers which is continuous at only
one point.
Sec. 1.8 Introduction to the Theory of Limits 67
In Problems 39-44, show that the given equation has at least ax 2 - 3

if x > - 5
one solution on the indicated interval. 52. g(x) =
bx + 2 if x ≤ - 5
39. Λ√X = x 2 + 2x - 1 on [0, 1] 40. χ = x 2 - x - 1 on [1, 2]
' sin ax
if x < 0
41. ∕ x - 8 + 9x 2z3 = 29 on [0, 8] 42. tan x = 2x 2 - 1 on [ - J , 0^j
53. g(x) = if x = 0
43. cosx - sin x = x on [θ, ,x + b if x > 0
tan ax
44. cosx = x 2 - 1 on [0, τr] ifx < 0
tan bx
45 Let f(χ} = ( χ 2 if x > 2 54. ∕( x ) = 4 if x = 0
45. L e t / W ^ + 1 i f χ s 2
ax +2 b i f x > 0
56. the root of x 2 - 3x + 21 = 0 between 0 and 1
Show that/is continuous from the left at 2, but not from 57. the root
55. of x root
negative —xof—x1 =- 02xbetween
- 1 = 01 and 2
the right. : 3
58. theroot of x + x - 1 = 02
46. Find constants a and b such that /(2) = /(0) and / is Computational Window
59. the zero o f ∕( x ) = x 3 + 2x 2 + x - 5
continuous at x = 1. ■ 60. the zero o f ∕( x ) = x 3 + 2x 2 - x + 1
ra x + 3 if x > 1 In Problems 55-60, use the bisection method, a graphing
∕( x ) = ] 4 if x = 1 j calculator,
The functions or a computer
in Problems to approximate a solution tof(x)
61-62 are undefined at an
I
lx 2 + b ifx < 1 j = 0. I f you use the bisection method, estimate
obvious point. I f possible, define the function there the root
so with
it is
j error no greater
continuous for allthan
x. I pg;
f noti fpossible,
you use aexplain
computer,
why.estimate
47. ■ What Does This Say? Use the intermediate value the root with error no greater than γ⅛Q.
[ (t χ sin x - x
theorem to explain why the hands of a clock coincide at
x3
least once every hour.
■ 6i--,2 x 4 - 2x 3 + 3x 2 - 5x - 2
48. The population (in thousands) of a colony of bacteria I ' ------------ 1⅛^2∣-------------
minutes after the introduction of a toxin is given by the
function
Λ0 = { / + 1 if 0 ≤ t < 5
-8/ + 66 if t ≥ 5
a. When does the colony die out?

b. Show that at some time between t = 2 and t = 7, the
population is 9,000.
In Problems 49-54, fin d constants a and b so that the given

function will be continuous fo r all x. You may use the result o f
Problem 62.
'a x + 3 if x > 5
49. ∕( x ) = 8 if x = 5
2 if x < 5
i x + bx + 1 θ 63. Prove the root location theorem.
' at - 4 if t ≠ 2 64. Show that/is continuous at c if and only if it is both
50. /(0 = t - 2 continuous from the right and continuous from the
fi if t = 2
r V x —a
left at c.
if x > 0, x ≠ 1 if x ≠ 0
51∙ / « = x - 1 65. Let u(x) = x and ∕( x ) =
l⅛ if x = 1 i f x = 0'
Show that lim ∕[ιz(x)] ≠ /pin t w(x)
1.8 INTRODUCTION TO THE THEORY OF LIMITS*
IN THIS SECTION Limit of a function (formal definition), the

believer/doubter format, epsilon-delta proofs, selected theorems with formal
proofs
This section gives a firm footing for calculus by making precise the notion of a
limit. The techniques of this section will be important if you take future courses
in advanced mathematics, but they are not required for the material that
follows in this text.
This section can be omitted without disrupting the development of the remainder of the text.
In Section 1.5 we defined the limit of a function as follows:

Historical Note
The notation
lim f(x ) = L
is read “ the limit o ff(x ) as x approaches c is L , ' and means that the function
values∕(x) can be made arbitrarily close to L by choosing x sufficiently close to
c but not equal to c.
This informal definition was valuable because it gave you an intuitive feeling for
the limit of a function and allowed you to develop a working knowledge of this
fundamental concept. For theoretical work, however, this definition will not
suffice, because it gives no precise, quantifiable meaning to the terms “ arbitrarily
close to £ ” and “ sufficiently close to c.” The following definition, derived from
the work of Cauchy and Weierstrass, gives precision to the limit definition.
AUGUSTIN-LOUIS CAUCHY ■ LIMIT OF A FUNCTION (FORMAL DEFINITION)

(1789-1857)
In the nineteenth century, lead
The limit statement
ing mathematicians, including
Augustin-Louis Cauchy and lim f(x ) = L
Karl Weierstrass (1815-1897) x →c
sought to make the concept of means that for each ε > 0, there corresponds a number δ > 0 with the
limit more precise. The ε-δ def property that
inition given in this section is
∣∕(λ') - L ∖ < ε whenever 0 < ∣x - c∣ < δ
the result of those efforts.
Cauchy is described by the
historian Howard Eves not only Because the Greek letters ε (epsilon) and δ (delta) are traditionally used in this
as a first-rate mathematician context, the formal definition of limit is sometimes called the epsilon-delta
with tremendous mathematical definition of the limit. The goal of this section is to show how this formal
productivity, but also as a law definition embodies our intuitive understanding of the limit process and how it
yer, a mountain climber, and a
can be used rigorously to establish a variety of results.
painter.
Do not be discouraged if this material seems difficult— it is. Probably your
Cauchy wrote a treatise on
integrals in 1814 that was con best course of action is to read this section carefully and examine a few examples
sidered a classic, and in 1816 closely. Then, using the examples as models, try some of the exercises. This
his paper on wave propagation material often takes several attempts, but if you persevere, you should come away
in liquids won a prize from the with an appreciation of the epsilon-delta process— and a better understanding of
French Academy. It has been calculus. We begin by restating the formal definition graphically.
said that with his work the Behind the formal language is a fairly straightforward idea. In particular, to
modern era of analysis began. establish a specific limit, say lim f( x ) = L , a number ε > 0 is chosen first to
In all, he wrote over 700 pa
pers, which are today consid
establish a desired degree of proximity to L , and then a number δ > 0 is found
ered no less than brilliant. that determines how close x must be to c to ensure that f(x ) is within ε units of L .
The situation is illustrated in Figure 1.46, which shows a function that

satisfies the conditions of the definition. Notice that whenever x is within δ units
of c (but not equal to c), the point (x, ∕( x ) ) on the graph of f must lie in the
rectangle (purple region) formed by the intersection of the horizontal band of
width 2ε (blue screen) centered at L and the vertical band of width 2δ (red region)
centered at c. The smaller the ε-interval around the proposed limit L, generally the
smaller the δ-interval will need to be in order for L to lie in the ε-interval. If such a
δ can be found, no m atter how small ε is, then L must be the limit.
■ THE BELIEVER/DOUBTER FORMAT

Figure 1.46 The epsilon-delta def The limit process can be thought of as a “ contest” between a “believer” who
inition of limit claims that lim ∕( x ) = L and a “ doubter” who disputes this claim. The contest
begins with the doubter choosing a positive num ber ε so that whenever x is a
num ber (other than c) within δ units of c (the red region), the corresponding
function value f(x ) is within ε units of L (the blue region). Naturally, the doubter
Computational Window ∞ tries to choose ε so small that no m atter what δ the believer chooses, it will not be
possible to satisfy the accuracy requirement. When will the doubter win and when
The tolerance setting on a cal will the believer win the argument? As you can see from Figure 1.46, if the believer
culator often uses the δ-ε nota has the “correct lim it” L, it will be in the intersection (the purple portion) no
tion. For example, the TI-85 m atter what ε the doubter chooses. On the other hand, if the believer has an
allows you to set the tolerance “ incorrect lim it,” as shown in Figure 1.47, the limit may be in the purple portions
for maximum and minimum for some choices of ε, but for other choices of ε the doubter can force the believer
of a function by inputting a δ-
to be the loser of this contest by prohibiting the believer from making a choice
value.
within the purple portion.
To avoid an endless series of ε-δ challenges, the believer usually tries to settle
the issue by producing a formula relating the choice of δ to the doubter’s ε that will
satisfy the requirement no m atter what ε is chosen. The believer/doubter format is
used in Example 1 to verify a given lim it statement (believer “ wins” ) and in
Example 2 to show that a certain limit does not exist (doubter “wins” ).
EXAM PLE 1 Verifying a limit claim (believer wins)
Show that lim (2x + 1) = 5.

x →2
Solution To verify the given limit statement, we begin by having the doubter
choose a positive num ber ε. Before picking δ, the believer might entertain the
thought process shown in the following screen.
Write ∣∕(x ) - L ∖in terms of ∣x - c∣ (where c = 2) as follows:

∣∕(x) - L∣ = ∣(2x + 1) - 5∣ = ∣2x - 4∣ = 2∣x - 2∣
Thus, if 0 < ∣x - c∣ < δ or 0 < ∣x - 2∣ < δ, then ∣∕(x ) - L ∖ < 2δ
The believer wants ∣∕(x ) - L ∖ < ε, so we see that the believer should choose
2δ = ε, or δ = ∣.*
With the information shown in this screen, we can now make the following
argument, which uses the formal definition of limit:
Let ε > 0 be given. Choose δ = f > 0. It follows that if
Figure 1.48 lim (2x + 1) = 5 0 < ∣x - 2∣ < δ = ∣ , then

∣(2x+ 1) - 5 ∣ = 2 ∣x -2 ∣ < 2(∣) = ε
Thus, the conditions of the definition of limit are satisfied, and we have
lim (2x + 1) = 5
This is shown graphically in Figure 1.48a.

Notice that no matter what ε is chosen the doubter, by choosing a number
that is one-half of that ε, the believer will force the function to stay within the
purple portion of the graph, as shown in Figure 1.48b. ⊂=ι
EXAM PLE 2 Disproving a limit claim (doubter wins)
O -γ-2 _ ^2 -γ- _ O
Determine whether lim —-------- ≡----- = 6.
x→2 X — Z.
Solution Once again, the doubter will choose a positive number ε and the
believer must respond with a δ. As before, the believer does some preliminary
work with ∕( x ) - L:
∣2x 2 - 3x - 2 _ A∣ = ∣2x 2 - 3x - 2 - 6x + 12
I x - 2 0l I x - 2
_ ∣2x 2 - 9x + 10∣
x - 2
= ∣(2x - 5)(x - 2)1

I x - 2
= ∣2x - 5∣
The believer wants to write this expression in terms of x - 2. This example

does not seem to “ fall into place” as did Example 1. Suppose the doubter
chooses ε = 1.5 (not a very small ε). Then we must find a δ so that whenever x
is within this δ distance of 2 ,∕( x ) will be within 1.5 units of 6. This is shown in
Figure 1.49a. However, if the doubter chooses ε = 0.5 (still not a very small ε),
then the believer is defeated, because no δ can be found, as shown in Figure
1.49b. H
Figure 1.49 Example of an incor *In fact, there are other choices that would work; for example, δ = ~ where a ≥ 2.
rectly chosen limit
The believer/doubter format is a useful device for dramatizing the way certain
choices are made in epsilon-delta arguments, but it is customary to be less
“ chatty” in formal mathematical proofs.
■ EPSILON-DELTA PROOFS
The following examples illustrate epsilon-delta proofs, two in which the function
has a limit and one in which it does not.
EXAM PLE 3 An epsilon-delta proof of a limit of a linear function
Show that lim (4x - 3) = 5.

Solution We have ∖f( x ) - L ∖= ∣4x - 3 - 5∣
= ∣4x - 8∣
= 4∣x - 2∣
This must be less than ε

whenever ∣x —2∣ < δ.
Choose δ = then
∖f(x ) - L ∖= 4∣x - 2∣ 4δ = 4(∣) = ε
EXAM PLE 4 An epsilon-delta proof of a limit of a rational function
Show that lim — ¾ +2 = _ j

x →2 X — 4
Solution We have ∣∕'(x) - L ∖ = ⅛---~ ⅛+-2 _
- ∣χ 2 - χ - 2∣
x -4
This must be less than the given ε

whenever x is near 2.
Certainly ∣x - 2∣ is small if x is near 2, and the factor I* * is not large (it
is close to ∣). Note that if ∣x - 2∣ is small it is reasonable to assume
∣x - 2∣ < 1 so that 1 < x < 3

v + 1
The largest value of the fraction χ _ is
1+ 1 _ 2
1 -4 -3
Jv- _1_ 1
and the smallest value of the fraction χ _ is
Thus, if ∣x - 2∣ < 1, we have

so that
—4 < ^γ + ! < 4 Because - S < 4
x - 4 3
and
* + 1 < 4
lx -4 ∣ 4
Now let ε > 0 be given. If simultaneously

∣x -2 ∣< ⅞ and ⅛⅛⅛∣ < 4
then ∣v , -1
∣∕(x ) - L ∖= ∣x - 2∣ ∣⅞⅛⅛∣ < ∣(4) = ε
Thus, we have only to take δ to be the smaller of the two numbers 1 and f
in order to guarantee that
∣∕(x ) - L ∖ < ε
That is, given ε > 0, choose δ to be the smaller of the numbers 1 and f. We
write this as δ = min( 1, ∣). nm
EXAM PLE 5 An epsilon-delta proof that a limit does not exist
Show that lim - does not exist.

x→ 0 X
Solution Let f(x ) = - and L be any number. Suppose that lim ∕( x ) = L . Look at
the graph o f f as shown in Figure 1.50. It would seem that no matter what
value of ε is chosen, it would be impossible to find a corresponding δ.
Consider the absolute value expression required by the definition of limit: If
∣∕(x ) - L ∖ < ε, or, for this example, - L∣ < ε
then —ε < - —L < ε Property of absolute value (Table 1.1, p. 10)
and Z, — ε < — < L + ε

X
If ε = 1 (not a particularly small ε), then

Figure 1.50 lim
⅛l < ∣l ∣ + 1
w > ∣Z ∣+ T
which proves (because L was chosen arbitrarily) that lim does not exist. In
other words, because ε can be chosen very small, ∣x∣ will be very large, and it
will be impossible to squeeze between L - ε and L + ε. βββ
X
■ SELECTED THEOREMS WITH FORMAL PROOFS
Next, we shall prove several theoretical results using the formal definition of the
I f ∕( x ) > 0 for all x near c, then lim ∕( x ) ≥ 0. limit.
x→c
The next two theorems are useful tools in the development of calculus. The
Figure 1.51 Limit limitation first states that the points on a graph that are on or above the x-axis cannot
theorem possibly “ tend toward” a point below the axis, as shown in Figure 1.51.
THEOREM 1.8 Limit limitation theorem
Suppose lim ∕( x ) exists and ∕( x ) ≥ 0 throughout an open interval containing the

number c, except possibly at c itself. Then
lim ∕( x ) ≥ 0
x →c
Proof: Let L = lim ∕( x ) . To show that L ≥ 0, assume, on the contrary, that

L < 0. According to the definition of limit (with ε = - L ) , there is a number
δ > 0 such that
∣∕(x ) - L ∖ < - L whenever 0 < ∣x - c∣ < δ
In particular,
∕( x ) - L < - L
or
∕( x ) < 0
Thus
fi x ) < 0 whenever 0 < ∣x - c∣ < δ
However, this contradicts the hypothesis that fi x ) ≥ 0 throughout an open
interval containing c. Therefore, we reject the assumption that L < 0 and
conclude that L ≥ 0, as required. ■
0 It may seem reasonable to conjecture that if f i x ) > 0 throughout an

open interval containing c, then lim f i x ) > 0. This is not necessarily true,
and the most that can be said in this situation is that lim ∕( x ) ≥ 0, if the
limit exists. For example, if
fι ∖ _ f χ 2 for x ≠ 0
/ W ’ ll fo r χ = 0
then fix ) > 0 for all x , but lim fi x ) = 0.

Λ^→ 0
0
Useful information about the limit of a given function/can often be obtained
by examining other functions that bound /from above and below. For example, in
Section 1.5 we found
lim ⅛ = 1
x →0 X
by using a table. This will be proved in Section 2.3 by first showing that
cosx ≤ ≤ 1
for all x near 0 and then noting that because cosx and 1 both tend toward 1 as x
approaches 0, the function
sinx
x
which is “ squeezed” between them, must converge to 1 as well. Theorem 1.9
provides the theoretical basis for this method of proof.
THEOREM 1.9 The squeeze theorem
If g(χ) ≤ ∕( χ ) ≤ ∕z(x) for all x in an open interval containing c (except possibly at

c itself) and if
lim g(x) = lim h(x) = L
then lim ∕( x ) = L .
Proof: Let ε > 0 be given. Since lim g(x) = L and lim h(x) = L , there are
positive numbers δ1 and δ, such that '~''
∣g(x) - L ∖ < ε and ∖lι(x) - L ∖ < ε
whenever 0 < ∣x - c∣ < δ1 and 0 < ∣x - c∣ < δ,
respectively. Let δ be the smaller of the numbers δl and δ1 . Then, if x is a number
that satisfies 0 < ∣x - c∣ < δ, we have
- ε < g(x) - L ≤ ∕( x ) - L ≤ ∕z(x) - L < ε
and it follows that ∣∕(x ) - L ∖ < ε. Thus, lim ∕( x ) = L , as claimed. ■
Figure 1.52 The squeeze theorem The geometric interpretation of the squeeze theorem is shown in Figure 1.52.
Notice that because g(x) ≤ f(x) ≤ Λ(.x), the graph of f is “ squeezed” between
those of g and h in the neighborhood of c. Thus, if the bounding graphs converge
to a common point P as x approaches c, then the graph o f f must also converge to P
as well.
PROBLEM SET 1 .8
Q i n Problems 1-6, use the believer/doubter format to prove or lim f(x ) ≥ lim g(x). Hint: Apply the limit limitation the
disprove the given limit statement. orem (Theorem 1.8) to the function h(x) = f(x) —g(x).
1. lim (2x - 5) = - 3 2. lim (3.v + 7) = 1
Problems 20-23 lead to a proof o f the product rule for limits.
3. lim (3x + 1 ) = 5 4. lim (x 2 - 2) = 5
20. If lim f(x ) = L , show that lim ∣∕(x)∣ = |£|. Hint: Note that
5. lim (x 2 + 2) = 6 6. lim (x 2 — 3x + 2) = 0 ∣∣∕(x)∣ - ∣L∣∣ ≤ '∣∕(x) - L ∖.
A →2 .V → 1
In Problems 7-12, use the form al definition o f the limit to 21. If lim f( x ) = L , L 4= 0, show that there exists a δ > 0
prove or disprove thegiven limit statement. such that ∣∣L∣ < ∣∕(x)∣ < j∣L∣ for all x for which
7. lim (x + 3) = 5 8. lim (ft - 1 ) = 0 0 < ∣x - C∣ < δ.
9. lim (3x + 7) = 1 10. lim (2x — 5) = - 3 Hint: Use ε = ∣∣Z,∣ in Problem 20.
11. lim (x 2 +2) = 6 12. lim - = ⅛ 22. If lim f( x ) = L and £ ≠ 0, show that lim [∕( x ) ] 2 = £ 2 by
.v→2 x→2 X 2 x →c x →c
completing these steps:
<-> γ> .ι . .∙v > I s in - if x Ψ 0
13. Prove that j( x ) = t x a. Use Problem 21 to show that there exists a δl > 0 so
lθ if x = 0
that ∖f( x ) + L ∖ < ∣∣L∣ whenever 0 < ∣x - c∣ < δ 1.
is not continuous at x = 0. Hint: To show that
lim ∕( x ) ≠ 0, choose ε = 0.5 and note that for any δ > 0, b. Given ε > 0, show that there exists a δ7 > 0 such that
∣ [ ∕W J 2 ^^ T 2 ∣ < j∣L∣ε whenever 0 < ∣x - c∣ < δ2.
there exists an x of the form x = -, ^ . , with n a
ττz(2n + 1) c. Complete the proof that lim [∕( x ) ] 2 = L 2 .
natural number for which 0 < ∣x∣ < δ.
23. Prove the product rule for limits: If lim ∕( x ) = L and
Q In Problems 14-19, construct a form al ε-δ proof to show that lim g(x) = M , then lim f(x)g(x) = L M . Hint: Use the
the given limit statement is valid fo r any number c.
result of Problem 22 along with the identity
14. If lim f(x') exists and k is a constant, then lim kf'(x) =
k lim f(x ). f g = ∣ κ ∕+ g ? - σ - g)2]
15. If lim f(x ) and lim g(x) both exist, then lim [f(x) —g(x)] 24. Show that if f is continuous at c and ∕( c ) > 0, then
1⅛ 1≡ . f(x } > 0 throughout an open interval containing c. Hint:
= χ ∕W ~ g(χ f
Note that lim f(x ) = f(c) and use ε = ⅛f(c) in the defini
16. If lim f(x ) and lim g(x) both exist and a, b are constants,
Λ'~*C X— *C tion of limit.
then lim [af(x} + ⅛(x)] = a lim f(x ) + b lim g(x).
x →c x →c x →c 25. Show that if/ is continuous at L and lim g(x) = L , then
17. If lim f(x ) = 0 and lim g(x) = 0, then lim ∕(x)g(x) = 0.
lim f(g(x)) = f(L ) by completing the following steps:
x →c
18. If f(x ) ≥ g(x) ≥ 0 for all x and lim f(x ) = 0, then a. Explain why there exists a δ l > 0 such that
lim g(x) = 0. ∣∕(n , ) - ∕(T )∣ < ε whenever ∣n, - L ∖ > δ l .
x →c
19. Iff(x ) ≥ g(x) for all x in an open interval containing the b. Complete the proof by setting w = g(x) and using part
number c and lim f(x) and lim g(x) both exist, then (a).
Chapter 1 Review 75
Computational Window continuous at x = 3, but determine graphically how you

would pick δ to accommodate a given ε. (For example,
26. For linear functions, the relation between ε and δ is δ = Kε.)
clear; but for complicated functions, it is not. Howev
er, the correct graph can illustrate this relationship. x 4 - 2x 3 + 3x 2 - 5x + 2
For the following function, illustrate not only that /is x - 2
C H A P T E R 1 R E V IE W
IN T H IS R E V I E W Proficiency examination, supplementary problems

Problems 1-21 of the proficiency examination test your knowledge of the
concepts of this chapter, and Problems 22-30 test your understanding of those
concepts by asking you to work sample practice problems. The supplementary
problems present questions relating to the material of this chapter in a wide
variety of settings.
P R O F I C I E N C Y E X A M IN A T IO N
Concept Problems 14. State the informal definition o f a limit of a function.

1. Characterize the following sets of numbers: natural Discuss this informal definition.
numbers, whole numbers, integers, rational numbers, 15. State the following basic formulas and rules for limits:
irrational numbers, and real numbers. a. limit o f a constant b. multiple rule
2. Define absolute value. c. sum rule d. difference rule
3. State the triangle inequality. e. product rule f. quotient rule
4. State the distance formula for points JP (X 1, J∙, ∣) and g. power rule
β (⅞ J , 2)∙ 16. Define the continuity o f a function at a point and
5. Define slope; include a discussion of angle o f inclina discuss.
tion. 17. State the continuity theorem as well as the properties of
6. List the following forms of the equation o f a line: continuous functions.
a. standard form b. slope-intercept form 18. State the intermediate value theorem.
c. point-slope form d. horizontal line 19. What is the bisection method?
e. vertical line 20. State the formal definition of a limit.
7. State the slope criteria for parallel and for perpendicular 21. State the squeeze theorem for limits in formal terms.
lines.
8. Define function.
Practice Problems
9. Define the composition o f functions.
22. Find an equation for the lines satisfying the following
conditions:
10. What is meant by the graph o f a function?
a. through (— 5) with slope m = - ∣
11. Draw a quick sketch of an example o f each of the
following functions. b. through ( -3 , 5) and (7, 2)
a. identity function b. standard quadratic c. with x-intercept (4, 0) and j>-intercept (θ, —
function
c. standard cubic function d. absolute value function d. through (-∣> 5) and parallel to the line
e. cube root function f. standard reciprocal 2x + 5y - 11 = 0
g. standard reciprocal squared h. cosine function e.the perpendicular bisector of the line segment joining
i. sine function j. tangent function P ( - 3 , 7) and β(5, 1)
k. secant function 1. cosecant function 23. Sketch the graph of each o f the following equations:
m. cotangent function a. 3x + 2y - 12 = 0 b. y = x 2 - 4x - 10
12. What is a polynomial function? c. y - 3 = ∣x + 1∣ d. y = 2 cos(x - 1)
13. What is a rational function? e. y + 1 = tan(2x + 3)
24. If ∕( x ) = . + ∣ , what values of x satisfy 28. Evaluate each of the following limits.
⅛ ÷ τ ) ⅛ i > v x 2
b. lim
Λ - 4
∙→4 X
√7 „ x 2 ~ 5x + 6 1 - cosx
c. lim 2 A d. lim
25. and g(x) = _ ∙ , does f = g rl Why λ-→2 XL - 4 x → 0 2 tanx
or why not? 29. Find constants A and B such that/is continuous for all
26. If ∕( x ) = sinx and g(x) - V l - x 2 , find the composite
functions f o g and g o f
'A x + 3 if x < 1
27. An efficiency expert has found that when X Y Z Corpora
∕ω = • 2 if x = 1
tion employs N workers (N > 3), it takes . x 2 + B if x > 1
3#+ 4 30. Show that the equation
2N- 5
hours to complete a certain job. If each worker earns x + sin x = —τ=J—-

√x + 3
$25 per hour, express the total labor cost C as a function
o f# has at least one solution on the interval [0, ττ].
SUPPLEMENTARY PROBLEMS*
In Problems 1-3, fin d the perimeter and area o f the given 10. Let∕( x ) = x 2 + 5x - 9. For what values of x is it true that
figure. ∕(2 x ) = ∕(3x)7
1. the right triangle with vertices ( - 1, 3), ( -1 , 8), and (11,8) 11. If an object is shot up from the ground with an initial
2. the triangle bounded by the lines y = 5, 3y —4x = 11, velocity o f 256 ft∕sec, its distance in feet above the ground
and 12x + 5 j = 25 at the end of t seconds is given by s{f) = 256/ - 16z2
3. the trapezoid with vertices (—3, 0), (5, 0), (0, 8), and (2, 8) (neglecting air resistance). What is the highest point for
this projectile?
In Problems 4-7, fin d an equation for the indicated line or 12. It is estimated that t years from now, the population of a
circle. certain suburban community will be
4. the vertical line through the point where the line
y = 2x - 7 is tangent to the parabola y = x 2 + 6x - 3
5. the circle that is tangent to the x-axis and is centered at
(5, 4) thousand people. What is the current population o f the
6. the circle that contains the origin and is tangent to the line community? What will the population be in 6 years?
3x + 4y - 40 = 0 When will there be 5,000 people in the community?
7. the line through the two points where the circles
x 2 + j>2 - 5x + 1y = 3 and x 2 + y 2 + 4-y = 0 intersect In Problems 13-16, evaluate the given limit.
8. In each of the following cases, find constants A and B so 13. lim λ ∕ - 2 ~ 1x 14. lim
that the given equation will be true whenever both sides λ∙→r V x - 1 x→ι χ 2 - 1
are defined: 15. lim χ 2 2 3 x ,+ 2 1 + cos⅛
16. lim
χ→ι x1 - 1 χ → ∖ ∕π τrx - 1
b. sin 3 x = A sin 3x + B sinx ∕YX ^⅛^ ∆ j t ) — / ( -^)

Evaluate Jim — ------— — - for the function f in Problems
9. In a triangle, the perpendicular segment drawn from a
given vertex to the opposite side is called the altitude on 17-20.
that side. Consider the triangle with vertices (—2, 1), 17. ∕( x ) = 3x + 5 18. ∕( x ) = √∑v
(5, 6), and (3, -2 ). 1 9 .∕( x ) = x ( x + 1) 2 0 .∕( x ) = ⅛
a. Find an equation for each line containing an altitude
of this triangle. 21. Find constants A and B so that the following function∕(x)
b. Show that the three lines found in part a intersect at will be continuous for all x:
the same point. Find the coordinates of this point. x 2 —A x —6
if x > 2
*Tħe supplementary problems are presented in a somewhat ∕ω = • x - 2
2
.x + B
random order, not necessarily in order of difficulty. if x ≤ 2
Chapter 1 Review 77
22. A bus charter company offers a travel club the following e. What percentage of the households in the community
arrangements: If no more than 100 people go on a certain had received new telephone books by the time 150
tour, the cost will be $500 per person, but the cost per worker-hours had been expended?
person will be reduced by $4 for each person in excess of 26. First sketch the graph of v = —x 2 + 5x - 6. Next, use
100 who takes the tour. your graph to obtain the graphs of the related functions:
a. Express the total revenue R obtained by the charter a. y = - x 2 + 5x
company as a function of the number of people who go b. y = x 2 - 5x + 6
on the tour.
c. y = (x + 1)2 + 5(x + 1 ) - 6
b. Sketch the graph of R. Estimate the number of people
that results in the greatest total revenue for the charter 27. Find the area of each of the following plane figures:
company. a. the circle x 2 + y 2 = 9
23. Many materials, such as brick, steel, aluminum, and b. the rectangle with vertices (0, 0), (3, 0), (3, 2), and (0, 2)
concrete, expand due to increases in temperature. This is 28. Find the volume and the surface area of each of the
why spaces are placed between the cement slabs in side following solid figures:
walks. Suppose you have a 100-ft length of material a. a sphere with radius 4
securely fastened at both ends, and assume that the b. a rectangular parallelepiped (box) with sides of length
buckle is linear. (It is not, but this assumption will serve 2, 3, and 5
as a worthwhile approximation.) If the height of the
29. Find the volume and the surface area of each of these
buckle is x ft, and the percentage of swelling is y, then x
solid figures:
a n d y are related as shown in Figure 1.53.
a. the right circular cylinder with height 4 and radius 2
b. the inverted cone with height 5 and top radius 3
30. Show that the tangent line to the circle x 2 + y 2 = r 2 at the
point (x0 , y0 ) has the equation y 0y + xθx = r2 . Hint. Recall
that the tangent line at a point P on a circle with center C
50 ft 50 ft is the line that passes through P and is perpendicular to
the line segment CP.
Figure 1.53 Buckling of a given material
31. a. Consider the triangle with vertices Λ ( - l, 4), B{î, 2),
and C(3, -6 ). Determine the midpoints M and M-, of
Find the amount of buckling (to the nearest inch) for the sides AB and AC, respectively, and show that the line
following materials: segment M i M 2 is parallel to side BC with half its
a. brick; y = 0.03 b. steel; y = 0.06 length.
c. aluminum; y = 0.12 d. concrete; y = 0.05 b. Generalize the procedure of part a to show that the line
24. A ball has been dropped from the top of a building. Its segment joining the midpoints of any two sides of a
height (in feet) after t seconds is given by the function given triangle is parallel to the third side and has half
its length.
Λ(0 = - 1 6 t 2 + 256
32. Let ABCD be a quadrilateral in the plane, and let P, Q, R,
a. How high will the ball be after 2 sec? and .S' be the midpoints of sides AB, BC, CD, and DA, as
b. How far will the ball travel during the third second? shown in Figure 1.54. Show that PQRS is a parallelogram.
c. How tall is the building?
d. When will the ball hit the ground?
25. Suppose the number of worker-hours required to distrib
ute new telephone books to x percent of the households in
a certain rural community is given by the function
f( )
J ( Xx > =
600x
300 - x
a. What is the domain of the function / ?

b. For what values of x does f(x~) have a practical inter
pretation in this context? Figure 1.54 Problem 32
c. How many worker-hours were required to distribute 33. Find a constant c that guarantees that the graph of the
new telephone books to the first 50% of the house equation
holds? x 2 + xy + cy = 4
d. How many worker-hours were required to distribute will have a y-intercept of (0, -5 ). What are the x-
new telephone books to the entire community? intercepts of the graph?
34. A manufacturer estimates that when the price for each 37. In Figure 1.56, ship A is at point P at noon and sails due
unit is p dollars, the profit will be N = - p 2 + 14/7 —48 east at 9 km∕hr. Ship B arrives at point P at 1:00 P .M . and
thousand dollars. Sketch the graph of the profit formula sails at 7 km/hr along a course that forms an angle of 60o
and answer these questions. with the course of ship A. Find a formula for the distance
a. For what values of p is this a profitable operation? .s,(∕) separating the ships t hours after noon. Approximate
(That is, when is N > 0?) ly how far apart (to the nearest kilometer) are the ships at
b. What price results in maximum profit? What is the
4:00 P.M.?
maximum profit?
35. When analyzing experimental data involving two varia
bles, a useful procedure is to pass a smooth curve through
a number of plotted data points and then perform compu
tations as if the curve were the graph of an equation
relating the variables. Suppose the following data are
gathered as a result of a physiological experiment in
which skin tissue is subjected to external heat for t
seconds, and a measurement is made of the change in
temperature ΔT required to cause a change of 2.5oC at a
depth of 0.5 mm in the skin:
I (sec) 1 2 3 4 10 20 25 30 Figure 1.56 Problem 37

Δ T ( 0C) 12.5 7.5 5.8 5 4.5 3.5 2.9 2.8
38. Under the provisions of a proposed property tax bill, a
homeowner will pay $ 100 plus 8% of the assessed value of
Plot the data points in a coordinate plane and then answer the house. Under the provisions of a competing bill, the
these questions. homeowner will pay $1,900 plus 2% of the assessed value.
a. What temperature difference Δ7' would be expected if If only financial considerations are taken into account,
the exposure time is 2.2 sec? how should a homeowner decide which bill to support?
b. Approximately what exposure time t corresponds to a 39. SPY PROBLEM The hero of a popular spy story has
temperature difference ΔT = 8°C? escaped from the headquarters of an international
36. A mural 7 feet high is hung on a wall in such a way so that diamond-smuggling ring in the tiny Mediterranean coun
its lower edge is 5 feet higher than the eye of an observer try of Azusa. Our hero, driving a stolen milk truck at 72
standing 12 feet from the wall. (See Figure 1.55.) Find tan km/hr, has a 40-minute head start on his pursuers, who
θ, where θ is the angle subtended by the mural at the are chasing him in a Ferrari going 168 km/hr. The
observer’s eye. distance from the smugglers’ headquarters in Azusa to the
border, and freedom, is 83.8 km. Will he make it?
40. Two jets bound for Los Angeles leave New York 30
minutes apart. The first travels 550 mph, while the second
goes 650 mph. How long will it take the second plane to
pass the first?
41. To raise money, a service club has been collecting used
bottles that it plans to deliver to a local glass company for
recycling. Since the project began 8 days ago, the club has
collected 2,400 pounds of glass, for which the glass
company currently offers 154 per pound. However, since
bottles are accumulating faster than they can be recycled,
the company plans to reduce the price they pay by 14 per
pound each day until the price reaches 04 fifteen days
from now. Assuming that the club can continue to collect
bottles at the same rate and that transportation costs
make more than one trip to the glass company unfeasible,
express the club’s revenue from its recycling project as a
function of the number of additional days the project
runs. Draw the graph and estimate when the club should
conclude the project and deliver the bottles in order to
Figure 1.55 Problem 36 maximize its revenue.
Chapter 1 Review 79
42. An open box with a square base is to be built for $48. The 46. A function /is said to satisfy a Lipschitz condition (named
sides o f the box will cost $3 per square ft and the base will for the 19th century mathematician, Rudolf Lipschitz) on
cost $4 per square ft. Express the volume of the box as a a given interval if there is a positive constant M such that
function of the length of its base.
l∕( χ ) - ∕( τ ) l < M ∖χ - y ∖
43. A regular polygon of n sides is inscribed in a circle of
radius R as shown in Figure 1.57. for all x and y in the interval (with x ≠ y). Suppose f
satisfies a Lipschitz condition on an interval and let c be
a fixed number chosen arbitrarily from the interval. Use
the formal definition o f limit to prove that/is continuous
at c.
47. In this problem, you are to find the tangent line to a
curve using the idea that the tangent line is the limit
of secant lines. As a nontrivial example, consider the
function defined by
Figure 1.57 A regular polygon of seven sides with central
angle θ (Problem 43)
and graphically find the tangent line at the specific
point (2,/(2)) = (2, 1.8). To this end, consider two
points on the graph, namely (2, ∕(2)) and {b, f(b)),
a. Show that the perimeter of the polygon is given by P(θ) where b is close to, but not equal to 2. The slope of
∕j q -I? ■ f) r) '7T the secant line (the line that connects the two points)
= sin τy where θ = is the central angle subtend
is ∕(⅛) ~ ∕(2 )
ed by one side o f the polygon.
b - 2.0
b. Use the formula in part a together with the fact that
To find the secant line for a given b, the slope is
lim s m x = J to show that a circle of radius R has
x→0 x ,, » -/ (2 )
m b -2
circumference 2∙πR.
c. Modify the approach suggested in parts a and b to and hence the secant is expressed as
show that a circle of radius R has area π R 2 . Hint: First y = /(2.0) + m(x - 2.0). Experimentally find b so
express the area o f the shaded polygon in Figure 1.58 that the secant is effectively tangent to the curve at (2,
as a function of θ. ∕(2 )). Record your choice of b and m, and if possible,
44. A cylindrical tank containing 50 f of water is drained into hand in a graph.
an empty rectangular trough that can hold 75 £. Explain 48. Here you investigate a function whose tangent at a
why there must be a time when the height of water in the particular point is problematic. Consider ∕( x ) =
tank is the same as that in the trough. (x — 2)2z3 + 2x 3 near x = 2. As in Problem 49, chose
45. The famous author John Uptight must decide between b close to, but not equal to 2 (try both b < 2 and
two publishers who are vying for the rights to his new b > 2). Report on your findings. Next try to explain
book, Zen and the Art o f Taxidermy. Publisher A offers what is happening by zooming the graph o f∕( x ) near
royalties of 12% of net proceeds on the first 30,000 copies (2,/(2)).
sold and 17% on all copies in excess of that figure, and
expects to net $5 on each copy sold. Publisher B will pay
no royalties on the first 4,000 copies sold but will pay 15% *49. P U T N A M E X A M IN A T IO N P ROBLEM If / and g are real-
on the net proceeds o f all copies sold in excess of 4,000 valued functions of one real variable, show that there
copies, and expects to net $6 on each copy sold. exist numbers x , y such that
a. With whom should John sign if he expects to sell 0 ≤ x ≤ 1, 0 ≤ y ≤ 1, and ∣x j - ∕( x ) - g( v)∣ ≥ ∣
30,000 copies?
b. What if he expects to sell 100,000 copies? *50. P U T N A M E X A M IN A T IO N P ROBLEM Consider a polyno
mial ∕( x ) with real coefficients having the property
c. Derive a formula for determining which publisher
John should choose if he expects to sell N copies. ∕(g (x )) = <⅞^(∕(x )) f° r every polynomial g(x) with real
coefficients. What is /(x)?
d. For what value of N are the two publishers’ offers
equivalent? This is called the break-even point. *Putnam examination problems should be considered optional.
GUEST ESSAY
Calculus Was Inevitable

by John L. Troutman, Professor of Mathematics at Syracuse University
The invention of calculus is now credited jointly to Isaac Newton (1642-1727)

Historical Note
and Gottfried W. Leibniz (1646-1716); when they produced their separate
publications, however (around 1685), there was a bitter controversy about who
created it. Part of the explanation for this is the fact that each had done his actual
work earlier (Newton in 1669; Leibniz, a little later); some can be attributed to the
rivalry between scientists in England who championed Newton and those in
Europe who supported Leibniz; but much simply reflects how ripe the intellectual
climate was for the blooming of calculus. Indeed, even if neither of these great
mathematicians had existed, it seems almost certain that the principles of
calculus—including the fundatmenta theorem (introduced in Chapter 4)—would
have been announced by the end of the seventeenth century.
Emergence of calculus was effectively demanded by the philosophical spirit of the

times. Natural philosophers had long believed that the universe was constructed
according to understandable mathematical principles, although they disagreed
about just what these principles were and how they might be formulated. For
Isaac Newton (1642-1727) example, the Pythagorean school (ca. 600 B.c.) maintained that everything
Sir Isaac Newton was one of consisted of (whole) numbers and their ratios; hence their consternation upon
the greatest mathematicians of
all time. He was a genius of the
discovering that different entities such as V2 could be constructed. Next, the early
highest order but was often astronomers announced that heavenly bodies move in circular orbits around the
absent-minded. One story Earth as center, and later that the Earth itself must be a perfect sphere to reflect
about Newton is that when he the divine hand of its Creator. Both of these assertions are now known to be false,
was a boy, he was sent to cut a and by 1612, Kepler had already explained why. Galileo (1564-1642) announced
hole in the bottom of the barn that the distance traveled by a heavy body falling from rest is proportional to the
door for the cats to go in and square of the elapsed time, and Fermat asserted (in 1657) that light moves along
out. He cut two holes— a large those paths that minimize the time of travel. The question was whether such laws
one for the cat and a small one could be formulated and justified mathematically, and what kind of mathematics
for the kittens. Newton consid would be appropriate to describe these phenomena.
ered himself a theologian rather
than a mathematician or a
F or already from antiquity there was the warning of Zeno (ca. 450 B.C.), in the
physicist. He spent years
searching for clues about the form of paradoxes, against unwise speculation about phenomena whose analysis
end of the world and the geog involved infinite processes. In particular, he “proved” that motion was impossible
raphy of hell. One of Newton’s if time consisted of indivisible instants, and conversely, that covering a given
quotations about himself is, “I distance was impossible if length was capable of infinite subdivision (see Chapter
seem to have been only like a 8). Thus, although 17th-century scientists and philosophers might propose such
boy playing on the seashore and principles, it was evident that there must be underlying subtleties in the mathe
diverting myself in now and matics required to support them (that, in fact, required another two centuries for
then finding a smoother pebble satisfactory clarification).
or prettier shell than ordinary,
whilst the great ocean of truth
To understand better how the mathematics developed, we must examine some of
lay all undiscovered before
me.”
the previous attempts to solve the twin problems of classical origin that motivated
the emergence of the calculus—those of finding the tangent line to a given planar
geometrical curve and finding the area under the curve. Newton himself acknowl
edged that his greater vision resulted from his “standing on the shoulders of
Chapter 1 Review 81
giants.” Who were these giants, and what had they contributed? First, there were
the efforts of the Greek mathematicians, principally Eudoxus, Euclid, and
Archimedes, who had originated the geometric concepts of tangency and area
between 400 B.c. and 200 B.C., together with examples of each, such as the
construction of tangent lines to a circle and the area under a parabolic curve. The
Hindu and Arabic mathematicians had extended the number system and the
formal language of algebra with certain of its laws by A.D. 1300, but it was not
until Newton’s own era that the methods of algebra and geometry were combined
Historical Note satisfactorily in the analytic geometry of Rene Descartes (1596-1650) to produce
the recognition that a geometrical curve could be regarded as the locus of points
whose coordinates satisfied an algebraic equation. This provided potential
numerical exactitude to geometric constructions as well as the possibility of giving
geometric proofs for limiting algebraic arguments. Since Euclidian geometry was
then generally regarded as the only reliable mathematics, it was the latter direction
that was most frequently taken. And it was this direction that was taken in the
seventeenth century by de Roberval, Fermat, Cavalieri, Huygens, Wallis, and
others, not the least of whom was Newton’s own teacher at Cambridge, Isaac
Barrow (1630-1677). These giants, as Newton called them, obtained equations
for tangent lines to, and the (correct) areas under, polynomial curves with
equations y = x n for n = 1, 2, 3, . . . , 9, and certain other geometrically defined
curves such as the spiral and the cycloid.
In his analysis of tangency, Barrow incorporated the approximating infinitesimal

Gottfried Leibniz (1646-1716) triangle with sides ∆x, ∆y, ∆.s, that is now standard in expositions of calculus, and
At the age of 14, Leibniz at Cavalieri attempted to “count” an indefinite number of parallel equidistant lines
tempted to reform Aristotelian to obtain areas. It was known that in specific cases these problems were related,
logic. He wanted to create a
and that they were equivalent, respectively, to the kinematic problems of
general method of reasoning by
calculation. At the age of 20, he characterizing velocity and distance traveled during a motion, problems which
applied for his doctorate at the directly confronted the paradoxes of Zeno.
university in Leipzig and was
refused (because they said he All that remained was for some mathematicians to sense the generality underly
was too young). He received his ing these specific constructions and to devise a usable notation for presenting the
doctorate the next year at the results. This was accomplished essentially independently by Newton (who, justly
University of Altdorf, where he mistrusting the required limiting arguments, suppressed his own contributions
made such a favorable impres until he could validate them geometrically) and by the only slightly less cautious
sion that he was offered a pro Leibniz. However, as we have argued, by this time (about 1670), it was almost
fessorship, which he declined, inevitable that someone should do so.
saying he had very different
things in view. Leibniz went on
W hat calculus has provided is a mathematical language that, by means of the
to invent calculus, but not with
out a bitter controversy devel derivative, can describe the rates of change used to characterize various physical
oping between Leibniz and processes (such as velocity) and, by means of the integral, can show how
Newton. Most historians agree macroscopic entities (such as area or distance) can emerge from properly
that the bitterness over who in assembled microscopic elements. Moreover, the fundamental theorem, which
vented calculus materially af states that these are inverse operations, supplies an exact method for passing
fected the history of mathemat between these types of description. Finally, the ability to relate the results of
ics. J. S. Mill characterized limiting arguments by simple algebraic formulas permits the correct use of
Leibniz by saying, “It would be calculus while retaining skepticism regarding its foundations. This has enabled
difficult to name a man more applications to go forward while mathematicians have sought an appropriate
remarkable for the greatness axiomatic basis.
and universality of his intellec
tual powers than Leibniz.”
O ur present technological age attests to the success of this endeavor, and to the
value of calculus.
MATHEMATICAL ESSAYS
Use a library or references other than this textbook to research 10. “There are, in every culture, groups or individuals who
the information necessary to answer the questions in Problems think more about some ideas than do others. For other
1-11. cultures, we know about the ideas of some professional
groups or some ideas of the culture at large. We know little,
1. Write an essay about Isaac Newton and his discovery of however, about the mathematical thoughts of individuals
calculus. This essay should be at least 500 words. in those cultures who are specially inclined toward mathe
2. Write a 500-word essay about Gottfried Leibniz and his matical ideas. In Western culture, on the other hand, we
discovery of calculus. focus on, and record much about, those special individuals
3. Write a 500-word essay about the controversy surrounding while including little about everyone else. Realization of
the discovery of calculus by Newton and Leibniz. this difference should make us particularly wary of any
4. The Greek mathematicians mentioned in this guest essay comparisons across cultures. Even more important, it
include Eudoxus, Euclid, and Archimedes. Write a short should encourage finding out more about the ideas of
paper about contributions they made that might have been mathematically oriented innovators in other cultures and,
used by Newton. simultaneously, encourage expanding the scope of Western
history to recognize and include mathematical ideas held
5. Write a paper about the contributions of Rene Descartes
by different groups within our culture or by our culture as
that are needed for the development of calculus.
a whole.” From Ethnomathematics by Marcia Ascher
6. What is the definition of elementary functions as given by (Pacific Grove: Brooks/Cole, 1991), pp. 188-189. Write a
Joseph Liouville? paper discussing this quotation.
7. The guest essay mentions the contributions of de
Book report “Ethnomathematics, as it is being ad
Roberval, Fermat, Cavalieri, Huygens, Wallis, and Barrow
toward the invention of calculus. Write a short paper about dressed here, has the goal of broadening the history of
mathematics to one that has a multicultural, global per
these contributions.
spective.” Read the book Ethnomathematics by Marcia
8. In this guest essay, Troutman argues that the invention of Ascher (Pacific Grove: Brooks/Cole, 1991), and prepare a
calculus was inevitable, and even if Newton and Leibniz book report.
had not invented it, someone else would have. Write a 500-
word essay either defending or refuting this thesis. 12. Make up a word problem involving a limit or continuity.
Send your problems to:
9. For years women were prohibited from serious study in
mathematics. Write a paper on the history of women M athematics Editor,
mathematicians and their achievements. Your paper re: Bradley, Smith: CALCULUS
should include a list of many prominent women mathema Prentice Hall
ticians and their primary contribution. It should also 113 Sylvan Ave., Route 9W
include a lengthy profile of at least one woman mathemati Englewood Cliffs, NJ 07632
cian.
The best ones submitted will appear in the next edition
(along with credit to the problem poser).
Techniques of
■ CONTENTS
Differentiation with
2.1 An Introduction to the
Derivative: Tangents
Tangent lines; Slope of a
Selected Applications
tangent; The derivative;
Existence of derivatives;
Continuity and differentiability;
Derivative notation
2.2 Techniques of Differentiation
Derivative of a constant PREVIEW
function; Derivative of a power
function; Procedural rules for
finding derivatives; Higher de In this chapter, we develop the main ideas of differential calculus. We
rivatives begin by defining the derivative, which is the central concept of
2.3 Derivatives of the Trigonomet differential calculus. Then we develop a list of rules and formulas for
ric Functions finding the derivative of a variety of expressions, including polynomial
Two special trigonometric limits;
Derivatives of the sine and the functions, rational functions, root functions, and trigonometric func
cosine; Differentiation of the tions. We then see how the derivative can be interpreted as a rate of
other trigonometric functions
change, and we develop a method for making approximations using the
2.4 Rates of Change: Rectilinear
Motion derivative.
Rate of change—geometric
interpretation; Average and in
stantaneous rate of change;
Application to physics—rectilin
ear motion; Falling body prob PERSPECTIVE
lem
2.5 The Chain Rule
In this chapter we focus on the concept of a derivative and are introduced to
Introduction to the chain rule; efficient ways of finding the derivative. The derivative is one of the fundamen
Differentiation of composite tal ideas in all of calculus and is the cornerstone of more advanced mathemat
functions; Generalized deriva ics.
tive formulas; Justification of
the chain rule If you have been introduced to the mechanics of finding a derivative in
2.6 Implicit Differentiation another course, you may feel that you already know the content of this chapter;
General procedure for implicit but there is much more to the idea of a derivative than finding the derivative of a
differentiation; Higher implicit polynomial function. W. Dale Compton of the National Academy of Engineer
differentiation ing stated at the National Colloquium on Calculus for a New Century (1987):
2.7 Related Rates Engineering, in the words of the Accreditation Board for Engineering and
Related rates and applications
Technology, is “the profession in which knowledge of the mathematical and
2.8 Differentials and Tangent Line natural sciences gained by study, experience, and practice is applied with
Approximation
Tangent line approximation; judgment to develop ways to utilize, economically, the materials and forces of
The differential; Error nature for the benefit o f mankind.” The operative words in this statement are
propagation; Marginal analysis knowledge and judgment. Whereas mathematics has most often been consid
in economics
ered as a requirement for knowledge, it is time for us to begin to consider its role
2.9 The Newton-Raphson Method in judgment. In addition to finding derivatives, you need to understand the
for Approximating Roots
The Newton-Raphson method; concept of a derivative if you ever hope to be able to use mathematics in making
When the Newton-Raphson judgments in scientific investigations. Your success in a large portion of the
method fails material of this book rests on your success in this chapter.
Chapter 2 Review
84 Ch. 2 Techniques of Differentiation with Selected Applications
2 .1 AN INTRODUCTION TO THE DERIVATIVE: TANGENTS
IN THIS SECTION Tangent lines, slope of a tangent, the derivative,

existence of derivatives, continuity and differentiability, and derivative notation.
Ancient Greek mathematicians were able to construct tangents to circles and a
few other simple curves, but before the 17th century there was no general
procedure for finding the equation of a tangent to the graph of a function at a
given point. To solve this “tangent problem,” Sir Isaac Newton (see Historical
Note on page 80) used an approach originally developed by the French
m athem atician Pierre de Fermat (see Historical Note on page 483). This
approach leads to the definition of the derivative.
■ TANGENT LINES
In elementary mathematics a tangent to a circle is defined as a line in the plane of

the circle that intersects the circle in exactly one point. However, this is much too
narrow a view for our purposes in calculus. We will now consider the concept of a
line tangent to a given curve (not necessarily a circle) at a given point. In general, it
is not a simple m atter to find the slope of a tangent line at a given point P 0(x 0 , y 0 ).
This is because the formula
, ∆y y - J'
slope = m = χ - = —l — o
o
requires knowledge not only of the point of tangency (xθ, y 0), but of at least one
other point (x 1, y 1) on the line as well.
■ SLOPE OF A TANGENT
The limit procedure for finding the slope of a tangent was originally developed by
Pierre de Fermat and was later used by Isaac Newton. The brilliance of the
Ferm at-N ew ton approach was the use of the “dynamic” limit process to attack
the “ static” problem of finding tangents.
Suppose we wish to find the slope of the tangent to y = ∕( x ) at the point
P(x 0 , ∕( x 0)). The strategy is to approximate the tangent by other lines whose slopes
can be computed directly. In particular, consider the line joining the given point P
to the neighboring point Q on the graph o ff, as shown in Figure 2.1. This line is
called a secant (a line that intersects, but is not tangent to, a curve). Compare the
secants shown in Figure 2.1.
0 ∆x is a single symbol and

does not mean delta times x. 0
Figure 2.1 The secant PQ

Sec. 2.1 An Introduction to the Derivative: Tangents 85
Notice that a secant is a good approximation to the tangent at point P as long as Q

is close to P.
To compute the slope of a secant, first label the coordinates of the neighboring
point Q, as indicated in Figure 2.1. In particular, let ∆x denote the change in the x-
coordinate between the given point P(x0, f( x 0)) and the neighboring point
O(x0 + ∆ x ,∕(x o + ∆x)). The slope of this secant, m sfχ, is easy to calculate:
∆v =
∕( ⅞ + δ ^ ) - ∕( ⅞ )
777 — —~ ------- - --------------------------^^^
sec ∆ λ' ∆ x
To bring the secant closer to the tangent, let Q approach P on the graph of /b y
letting Ax approach 0. As this happens, the slope of the secant should approach the
slope of the tangent at P. We denote the slope of the tangent by mtan to distinguish
it from the slope of a secant. These observations suggest the following definition.
Slope of a Line Tangent to a At the point P(x0 ,f(x 0)), the tangent line to the graph of/ has slope given by the
Graph at a Point formula
∕( x 0 + Ax) ~ ∕( x 0)
m tan = lim ∆x
ΔΛ∙→O
provided that this limit exists.

≡
EXAMPLE 1 Slope of a tangent at a particular point
Find the slope of the tangent to the graph o f∕(x ) = x 2 at the point P ( - 1, 1).
Solution Figure 2.2a shows the tangent t o /a t x = - 1.
The slope of the tangent is given by
/( - 1 ÷ ∆ χ ) - ∕( - i ) B ecauseR x^ ^
m, = Δ,r→
ta n
lim0 x
ax / ( —1 + ∆x) = (—1 + ∆x)2.
(—1 + ∆x)2 —( - l ) 2
= lim ∆x
ΔΛ →0
1 - 2∆x + (∆x)2 - 1
= lim ∆x
Δ,V→ 0
-2∆ x + (∆x)2
= lim ∆x
ΔΛ-→0
(-2 + ∆xj∆x
= lim ------ Ax
τ-------- Factor out A x ana reduce.
∆Λ→ 0
= lim (-2 + ∆x) = - 2 —≡
In Example 1 we found the slope of the tangent y = x 2 at the point ( - 1, 1). In

b . T an g en t line at (x, x 2 ) Example 2, we perform the same calculation again, this time representing the
given point algebraically as (x, x 2). This is the situation shown in Figure 2.2b for
Figure 2.2 Tangent lines to the
the slope of the tangent to y = x 2 at any point (x, x 2).
graph of y = x 2
EXAMPLE 2 Slope of a tangent at an arbitrary point
Derive a formula for the slope of the tangent to the graph of ∕(x ) = x 2, and
then use the formula to compute the slope at (4, 16).
Solution Figure 2.2b shows a tangent at an arbitrary point P(x, x 2) on the curve.
C3 Computational Window B H H From the definition of slope of the tangent,

∕( x + ∆x) - ∕( x )
—Tangent lines using a graph m tan = lim Because f(x ) = x 2,
Δ.V→ 0 ∆x
ing calculator ∕( x + ∆x) = (x + ∆x)2 .
We can use a calculator to (x + ∆x)2 - x 2

= lim ∆x
∕( ⅞ + δ *) ~ ∕(⅞ ) e o
graph ---------- x------------ for a
x 2 + 2x∆x + (∆x)2 - x 2
small value of ∆x. The result = lim
, ∆x
ΔΛ →0
ing graph can be used to ap
proximate the slope of the tan 2x∆x + (∆x)2
gent line at a point xθ. Using = lim ∆x
∆Λ→0
f(x ) = x 2 from Example 2, we
created the calculator graph by (2x + ∆x)∆x
= lim ∆x Factor and reduce.
graphing ΔΛ-→0
Yl = ((X + 0.01)2 - X 2)∕0.01 = lim (2x + ∆x) = 2x

Δx→0
Note this is At the point (4, 16), x = 4, so mtan = 2(4) = 8.
∕( x o ÷O.Ol) - ∕ ( x 0 )
0.01 The result of Example 2 gives a general formula for the slope of a line tangent
to the graph of∕(x ) = x 2, namely, mtan = 2x. The answer from Example 1 can now
be verified using this formula; if x = -1 , then mtan = 2(-1) = —2.
PREVIEW
/.0 1
Xmin= ~10 Y n in = -10
Xnax=10 9∩ax=10
X s c l= l V s c l= l
Use the (TRACE) to find the

which gives a formula for the slope of a secant to the Figure 2.3 Tangent lines
approximate slope of f at vari
ous values of x. graph of a function f is called the difference quotient to a curve indicate
of f. The limit of the difference quotient the shape of the curve.
Compare the results with those
obtained in Example 2.
lim f(× + ∆x) ~ ∕(x )
Hint-. Try resetting
(RANGE) value Xmin =
and note the trace values. which gives a formula for the slope of the tangent to the graph of f at the point
(x,∕(x)) is called the derivative o f/an d is frequently denoted by the symbol ∕'( x )
(read “eff prime of x”). To differentiate a function f at x means to find its
derivative at the point (x, ∕(x)).
Derivative The derivative of / at x is given by

∕( x + ∆x) —∕(x )
= lim ∆x
ΔΛ-→0
provided this limit exists.

The derivative is one of the fundamental concepts in calculus, and it is im portant

to make some observations regarding this definition.
1. If the limit for the difference quotient exists, then we say that the function f
is differentiable at x.
2. The value of a derivative depends only on the limit process and not on the
symbols used in that process. In Example 2, we found that if ∕( x ) = x 2 , then
∕'( x ) = 2x. This means that we also know
If g(f) = t 2, then g'(t) = 2t.
If h(u) = u 2 , then h'(u) = 2u.
3. Notice that the derivative of a function is itself a function.
Computational Window— Derivatives by difference quotient and graphing
A graphing calculator can help us “guess” the appro X = 1.0212766, Y = 3.159756. This means the slope
priate derivative, which can then be found using the of the secant line passing through the points where
definition (or derivative theorems to be developed x = 1 and x = 1 + 0.01 = 1.01 is about 3.2; this could
later in this chapter). For example, suppose we wish to be used as an approxim ation for the slope of the
find the derivative o f∕( x ) = x 3 at x = 1. tangent line. Later, we will show that the derivative of
∕( x ) = x 3 is ∕'( x ) = 3x 2 . We might “guess” this by
Graph Y l = ((X + 0.01) 3 - X 3)∕0.01.
looking at the graph of the difference quotient. Can
you im prove this approximation? Also, if you use
different (RANGE) values you can see a different-sized
or differently centered graph. Try once again by
setting Xmin = - 9 , Xmax = 10, and note the differ-
ent appearance of the graph. Also note that the
(TRACE) gives X = 1, Y = 3.0301.
Another way of using a graphing calculator to
approximate the derivative is to consider ∆x as the
3 ) / . Θl variable in the difference quotient. For this example,
∕( x ) = x 3 at x = 1, so you look at
X n in = - 3 V n in = 9
X∣ηax=3 V∩ax=8 ∕( x + ∆x) - ∕( x ) _ (1 + ∆x) 3 - 13
X s c l= .5 YS Q 1=1 ∆x ∆x
Recognize this as the difference quotient Trace values of ∆x as ∆x → 0. You can use the table or
∕( x + ∆x)— / ( x ) fθ r γ -(x ) = 3 a n ( j ∆x = 0.01. Re graph functions of your calculator. Graphically, we
χ
say we are looking at
member, this approximates the derivative and not the (1 + ∆x) 3 - 1
graph of ∕( x ) = x 3 . We use the (TRACE) to find in the neighborhood o f ∆x = 0.
∆x
Besides giving the slope of the tangent at a point, the derivative has a variety of
other im portant applications that we will see in subsequent sections. For example,
the derivative can be used to study rates of change, rectilinear motion, marginal
analysis in economics, maximum and m inimum values and concavity of a
function, to m ention just a few applications.
EXAM PLE 3 Derivative using the definition
0 Notice that f(f) = V t is de

fined for all t ≥ 0, whereas its
derivative f∖ t) = is defined
for all t > 0. This shows that a
function need not be differenti
able throughout its entire
domain. 0
THEOREM 2.1 Equation of a line tangent to a curve at a point
If/is a differentiable function at x0 , the graph of y = f(x) has a tangent line at the
point P(x0,f ( x 0)) with slope f '( x 0) and equation
y = Γ ( ⅞ ) ( χ - * 0) + ∕(⅞ )
Proof: To find the equation of the tangent to the curve y = f(x) at the point
P(x0 , J o), we use the fact that the slope of the tangent is the derivative f '( x 0) and
apply the point-slope formula for the equation of a line:
y —k = m(x — /?) Point-slope formula
y - y 0 = m ( x - x 0) Given point (xθ, y 0 )
T0 = ∕( ⅞ ) and m van = f '( x 0 )
y ~ f ( x 0) = ∕'(⅞ )(* - ⅞)
τ = ∕'( ⅞ ) ( * - ⅞ ) + ∕(⅞ ) Add f ( x 0) to both sides.
EXAM PLE 4 Equation of a tangent
Find an equation of the tangent to the graph off(x) = ⅛ at the point where x = 2.
Solution The graph of the function y = ⅛, the point where x = 2, and the tangent
at the point are shown in Figure 2.4. First, find f'(x).
f( x + ∆x) - f ( x )
f'(x ) = lim ∆x Definition o f derivative
ΔΛ^→ 0
1 _1
= lim X + Δ..V X ∕( x ) = ^ ∕( x + ∆ x ) = 7 ⅛
Δ,r→0 ∆x
x - (x + ∆x)
- lim Simplify the fraction.
ΔΛ → (o x∆x(x + ∆x)
= lim -1 Cancel factor o f ∆x.

Δ∆'→(o x(x + ∆x)
F i g u r e z .⅛ ι a ιι g c n ι ιυ y — - a ι
= -1
v2
Next, find the slope of the tangent at x = 2: mtan = ∕'( 2 ) = - ∣ . Finally, find
the point of tangency (2,/(2)) = (2, ∣). The equation of the tangent can now be
found by using Theorem 2.1:
y = - ∣( * ~ 2) + I or, in standard form, x + 4y - 4 = 0
0 The slope of f at x = x 0 is not the derivative ∕ ' but the value of the
derivative at x = xθ. In Example 4 the function is defined by∕( x ) = l ∕x , the
derivative is ∕'( x ) = - l ∕x 2, and the slope at x = 2 is the number
∕'( 2 ) = -⅛∙ 0
EXAM PLE 5 A line that is perpendicular to a tangent
Find the equation of the line that is perpendicular to the tangent of∕( x ) = - at
x = 2 and intersects it at the point of tangency.
Solution From Example 4, we found that the slope of the tangent is ∕'( 2 ) = - ∣
and that the point of tangency is (2, ∣). In Section 1.3, we saw that two lines are
perpendicular if and only if their slopes are negative reciprocals of each other.
Thus, the perpendicular line we seek has slope 4 (the negative reciprocal of
- ∣ ) , as shown in Figure 2.5. The desired equation is
Figure 2.5 Graph for Example 5

y —2 —4(x —2) Point-slope formula
The line we found in Example 5 has a name:
Normal Line to a Graph The normal line to the graph of/at the point P is the line that is perpendicular
to the tangent to the graph at P.
■ EXISTENCE OF DERIVATIVES
We observed that a function is differentiable only if the limit in the definition of

derivative exists. At points where a function /is not differentiable, we say that the
derivative o f f does not exist. The three common ways for a derivative to fail to
exist are shown in Figure 2.6.
Figure 2.6 Common examples where a derivative does not exist

EXAM PLE 6 A function that does not have a derivative because of a corner
Show that the absolute value function f(x ) = ∣x∣ is not differentiable at x = 0.
Solution The graph o ff(x ) = ∣x∣ is shown in Figure 2.7. Note that because the
slope “ from the left” at x = 0 is - 1 while the slope “ from the right” is + 1, the
graph has a corner at the origin, which prevents a unique tangent from being
drawn there.
We can show this algebraically by using the definition of derivative:
/(0 + ∆x) -/(0 )
∕'( 0 ) = Δ.ι-<0
Hmι Δ.V
∕( ∆ x ) -/ (0 )
Figure 2.7 f( x ) = ∣x∣ is not differ = lim
Δx→ 0 ∆x
entiable at x = 0, because the
slope from the left does not equal ∣∆χ∣
= lim
the slope from the right. Δx→ 0 ∆x
We must now consider one-sided limits, because
∣∆x∣ = ∆x when ∆x > 0, and ∣∆x∣ = - ∆ x when ∆x < 0.

lim ⅛ ~ = lim .^ x = —1 Derivative from the left
∆X Δx→ 0" ∆x
ΔΛ∙→ 0 - j j
Jim —Jim = 1 Derivative from the right
The left- and right-hand limits are not the same; therefore, the limit does not
exist. This means that the derivative does not exist at x = 0.
The continuous function ∕( x ) = ∣x∣ in Example 6 failed to be differentiable at

Figure 2.8 Vertical tangent line at x = 0 because the one-sided limits of its difference quotients were unequal. A
x = c continuous function may also fail to be differentiable at x = c if its difference
quotient diverges to infinity. In this case, the function is said to have a vertical
tangent at x = c, as illustrated in Figure 2.8.
Vertical Tangent Line The graph o f/has a vertical tangent at x = c if/ is continuous at c and
∆lxim
→c I ∕'W ∣ = ∞
(that is, the derivative function increases or decreases without bound).
We will have more to say about such functions when we discuss curve sketching
with derivatives in Chapter 3.
■ CONTINUITY AND DIFFERENTIABILITY
If the graph of a function has a tangent at a point, we would expect the function to
be continuous at that point. The following theorem confirms this expectation, for
nonvertical tangents.
T H E O R E M 2.2 Differentiability implies continuity
If a function/is differentiable at c, then it is also continuous at c.

Proof: There are three conditions for establishing continuity: (1) ∕( c ) is

defined; (2) lim ∕( x ) exists; and (3) lim f(x ) = f(c). Because f(x ) is differentiable at
c ,f(c ) must be defined. Therefore, to establish continuity, it suffices to show that
lim f(x ) = f(c), because doing so also implies that lim f(x ) exists. This is
X→ c t X→ c
equivalent to showing
b.m ∕( c + ∆x) = ∕( c ) or firn [ ∕( c + ∆x) - ∕( c ) ] = 0
Because/is a differentiable function at x = c ,∕'( c ) exists and

∕( c + ∆ x ) - ∕( c ) , c
Λh m n ---------- ∆ r =/
ΔΛ' → 0 ∆∆Λ
Therefore, by applying the product rule for limits, we find that
lim [ ∕( c + ∆x) - ∕( c ) ] = lim + ■∆x^l

Δx→ 0 Δx→ θL ∆x J
∕( c + ∆x) - ∕( c )
lim Γ lim ∆x^∣
Δ.t → 0 ∆x L∆x→o j
Thus lim ∕( x ) = ∕( c ) , and the conditions for continuity are satisfied.
0 Be sure you understand what we have just shown with Example 6 and
Theorem 2.2: If a function is differentiable at x = c, then it must be
continuous at that point. The converse is not true: If a function is continuous
at x = c, then it may or may not be differentiable at that point. Finally, if a
function is discontinuous at x = c, then it cannot possibly have a derivative
at that point. (See Figure 2.9c.) 0
Figure 2.9 A function continuous at x = 2 may or may not be differentiable at x = 2. A

function discontinuous at x = 2 cannot be differentiable at x = 2.
■ DERIVATIVE NOTATION
Symbols other than∕'( x ) are often used to denote the derivative. In particular, if y
rather th an ∕(x) is used to denote the function itself, the symbol (suggesting the
slope ∆y∕∆x) is frequently used instead of∕'( x ) or y '. Because the notation was
originally used by Leibniz, it is known as the Leibniz notation for the derivative.
dy
For example, if y = x 2, the derivative is y = 2x and the Leibniz notation is =
2x. The symbol is read “ the derivative of y with respect to x .” When we wish to
denote the value of the derivative at c in the Leibniz notation, we shall write
⅛d
t∕x∣x=c
For instance, we would evaluate = 4x 2 at x = 3 by writing
4x‰ 3 = 4(3)2 = 36
By omitting reference to y and f altogether, we can combine these two
statements and write
≠ ( vx 2) = 2 x
βX ’
which is read “ the derivative of x 2 with respect to x is 2x.”
0 Despite its appearance, is a single symbol and is not a fraction. In

Section 2.8, we introduce a concept called a differential that will provide
independent meaning to symbols like dy and dx, but for now, these symbols
have meaning only in connection with the Leibniz derivative symbol 0
EXAM PLE 7 Derivative at a point with Leibniz notation
Fmd IL -i
⅛ = ⅛ 3)
Solution dx dx[x ,
(x + ∆x) 3 - x 3
= lim
∆Λ-→ 0 ∆x
[x 3 + 3x 2∆x + 3x(∆x)2 + (∆x)3] - x 3
= lim
ΔΛ → 0 ∆x
= lim [3x2 + 3x(∆x) + (∆x)2]
ΔΛ∙→O
= 3x 2
At x = - 1,
⅛∣ = 3x 2 ∣x=-ι = 3
αxlx=-ι 1
EXAM PLE 8 Estimating a derivative using a table
Estimate the derivative of ∕( x ) = cosx at x = by evaluating the difference

quotient Δ T . ∕( x + ∆x) - ∕( x )
7
∆x ∆x
near the point x = f .

Computational Window ~ π ∆y cos + ∆x) - cosf

Solution
∆x- ∆x
Many calculators will find nu
merical values for derivatives. Choose a sequence of values for ∆x → 0: say 1, y y y . . . and use a
Check your owner’s manual. calculator (or a computer) to estimate the difference quotient by table. We
Typically, you will need to in
show selected elements from this sequence of calculations.
put {expression, variable, val
ue). Forf( x ) = cos (x) at x =
input
(COS X , X , π∕6) ∆y
∆x τ~ + Δ A' Difference Quotient
0
The output is - .5 .
1 1.523598776 -0.81885
0.5 1.023598776 -0.69146
0.125 0.648598776 -0.55276
0.0625 0.586098776 -0.52673
0.015625 0.539223776 -0.50675
0.00195313 0.525551906 -0.50085
0.00012207 0.523720846 -0.50005
0.00000763 0.523606405 -0.50000
I
0 f ≈ 0.5235987756 2
From the table, we would guess that ∕'( f ) = -0.5.
PROBLEM SET 2 .1
θ 1. ■ What Does This Say?* Describe the process of find 19. f{x ) = √ 5 x 20. ∕( x ) = V x + 1
ing a derivative using the definition.
Find an equation for the tangent to the graph o f the function at
2. ■ What Does This Say? What is the definition of a
the specified point in Problems 21-26.
derivative?
21. f(x ) = 3x - 7 at (3, 2) 22. g(t) = 2>t1 at ( -2 , 12)
Find the difference quotient for the functions given in Problems 23. f(s) = s 3 at s = - ∣ 24. ∕z(x) = 9 - x 2 at x = 0
3-8 , and then use the limit definition o f derivative to fin d f'(c)
for the given number c. 25. ∕(
j xv )7 = x
—~ at x = 2 26. g{t) = V t - 5 at t = 9
+ 3
3. f{x ) = 3 at c = - 5 4. ∕( x ) = x at c = 2
5. ∕( x ) = 2x at c = 1 6. ∕( x ) = 2 x 2 at c = 1 Find an equation o f the normal line to the graph o f the function
2
7. ∕( x ) = 2 - x at c = 0 8. ∕( x ) = - χ 2 at c = 2 at the specified point in Problems 27-30.
27. ∕( x ) = 5x - 2 at (3, 13) 28. g(t) = 2t - 3 at (0, -3 )
Use the definition to differentiate the functions given in Prob
lems 9-20, and then describe the set o f all numbers for which 29. g(t) = ÷ g at t = 1 30. ∕( x ) = V 4 x at x = 1
the function is differentiable.
9. f(x ) = 5 10. g(x) = 3x
F∕7ι<√ for the functions and values o f c given in Problems
11. f{t) = 3 t - 1 12. f{u) = 4 - 5w
31-34.
13. g{r) = 3r2 14. h(x) = 2 x 2
31. y = 2x, c = -1 32. y = 4 - x , c = 2
2
15. f{x ) = x - x 16. g(l) - 4 - l 1 , 4
33. y = 1 - x 2, c = 0 34. y = - , c = 1
17. f(s) = (s - 1)2 is . / « - ⅛
© 35. Suppose f(x ) = x 2 .
a. Compute the slope of the secant joining the points on
*Many problems in this book are labeled What Does This Say?
the graph o f f whose x-coordinates are —2 and -1 .9 .
Following the question will be a question for you to answer in your
own words, or a statement for you to rephrase in your own words. b. Use calculus to compute the slope of the line that is
These problems are intended to be similar to the “What This Says” tangent to the graph when x = - 2 and compare this
boxes that appear throughout the book. slope with your answer in part a.
36. Suppose fix') = x 3. c. In general, if∕(x ) = g(x) + A(x), what would you guess
a. Compute the slope of the secant joining the points on is the relationship between the derivative of f and the
the graph of f whose x-coordinates are 1 and 1.1. derivatives of g and A?
b. Use calculus to compute the slope of the line that is 50. Find the point on the graph of ∕(x ) = - χ 2 such that the
tangent to the graph when x = 1 and compare this tangent at that point passes through the point (0, -9).
slope to your answer from part a.
37. Sketch the graph of the function y = x 2 —x. Determine
the value of x for which the derivative is 0. What happens Computational Window
to the graph at the corresponding point(s)?
38. a. Find the derivative of f{x) = x 2 — 3x. It is often helpful to draw the graph o f a function and its de
b. Recall that a horizontal line has slope 0. Show that the rivatives. Graphing complicated functions has been made
parabola whose equation is y = x 2 - 3x has one hori relatively easy by calculators and computers. Use a graph
zontal tangent and find the equation of this line. ing utility to graph the functions given in Problems 51-54.
c. Find a point on the parabola whose equation is 51. ∕( x ) = x + sinx on [0, 6τr]
y = x 2 —3x where the tangent is parallel to the line 52. ∕(x ) = x sinx on ^0, ⅞]
3x + y = 11.
53. ∕(x ) = x 3 - 4x 2 + 5x - 2 on [ - 1, 4]
d. Sketch the graph of the parabola whose equation is
54. ∕(x ) = sin(2π cosx) on [0, π]
y = x 2 —3x. Display the horizontal tangent and the
tangent found in part c. 55. For the functions given in Problems 51-54, use the
39. a. Find the derivative of∕(x ) = 4 —2x 2. differentiation program of your calculus software to
draw the graph of the first derivative of each func
b. The parabola whose equation is y = 4 - 2x 2 has one
tion. Print out the graphs, if possible. For each graph,
horizontal tangent. What is its equation? note the following:
c. At what point on the parabola whose equation is
a. On the graph, plot the points where the derivative
y = 4 - 2x2 is the tangent parallel to the line
is 0 and note where the derivative is positive and
8x + 3y = 4? r where it is negative.
∕(rx∕ )∖ = } ~2
χ2
4,0n . τLet t .ife x < .0
Ix if x ≥ 0 b. Compare the graphs of∕(x ) and∕'(x ). Can you see
Does ∕'( 0 ) exist? Hint: Find the difference quotient and a correlation between the sign of f ' and the
take the limit as ∆x → 0 from the left and from the right. interval on which f is increasing or decreasing?
41. Let . fV x -3 if x ≥ 1 56. Return to a function seen earlier (Problem 47, page
≠w - t - 2 x ifx < l 79)
(x3 - 5)(x2 - 1)
a. Sketch the graph of f ∕(x ) = x2 + 1
b. Show th a t/is continuous but not differentiable at 1.
42. T HINK TANK PROBLEM Give an example of a function With the help of a computing device, you can com
continuous on (-∞, ∞) but not differentiable at x = 5. pute ∕'( 2 ) by, in it’s definition, making ∆x sufficient
ly small. Do so; then find the equation of the tangent
Estimate the derivative f'(x) in Problems 43-46 by evaluating
to the curve (2, ∕(2)).
the difference quotient
∆y = f(c + ∆xj -f(c ) 57. Return to another function seen earlier (Problem 48,
∆x ∆x page 79)
at a succession of numbers near c. ∕(x ) = (x - 2)2z3 + 2x3
43. ∕(x ) = (2x - 1)2 for c = 1 44. ∕(x ) = for c = 2 which gives trouble in seeking the tangent at (2,/(2)).
Attempt to compute∕'(2 ), either “by hand” or using
45. ∕(x ) = Vx for c = 4 46. ∕(x ) = ΛVZX for c = 8 a computer. Describe what happens; in particular, do
47. Show that the tangent to the parabola y = Ax 2 (for A ≠ 0) you see why the tangent there is meaningless?
at the point where x = c will intersect the x-axis at the
point (c∕2, 0). Where does it intersect the y-axis?
48. a. Find the derivatives of the functions y = x 2 and © 58. If∕'(c) ≠ 0, find the equation of the normal line toy = ∕(x )
y = x 2 - 3 and account geometrically for their simi at point P(c,∕(c))2 What is the equation if∕'( c ) = 0?
larity. 59. Suppose a parabola is given in the plane along with its axis
b. Without further computation, find the derivative of of symmetry. Explain how you could construct the tan
y = x 2 + 5. gent at a given point P on the parabola using only compass
49. a. Find the derivative of y = x 2 + 3x. and straightedge methods. Hint: Use Problem 47. You
b. Find the derivatives of the functions y - x 2 and y = 3x may assume that the parabola has an equation of the form
separately. How are these derivatives related to the y = A x 2 in which the y-axis is the axis of symmetry and
derivative in part a? the vertex of the parabola is at the origin.
Sec. 2.2 Techniques of Differentiation 95
2 .2 TECHNIQUES OF DIFFERENTIATION
IN THIS SECTION Derivative of a constant function, derivative of a power

function, procedural rules for finding derivatives, higher derivatives
In Section 2.1, you learned how to find the derivative of a function / b y
computing the limit of its difference quotient
f ( x + ∆x) - ∕( x )
lim
lt →0 ∆x
For even the simplest functions, this process is tedious and time-consuming. In
this section, we will describe some rules that simplify the process of differentia
tion.
■ DERIVATIVE OF A CONSTANT FUNCTION
We begin by proving that the derivative of any constant function is zero. Notice
that this is plausible because the graph of the constant function ∕( x ) = k is a
horizontal line, and its slope is zero. Thus, for example, if∕( x ) = 5, then f '( x ) = 0.
THEOREM 2.3 Constant rule
A constant function ∕( x ) = k has a derivative ∕'( x ) = 0; in Leibniz notation,
=0
Proof: Note that if ∕( x ) = k, then ∕( x + ∆x) = k for all ∆x. Therefore, the
difference quotient is
∕( x + ∆ x ) - ∕( x ) k - k , . c.
∆x ∆x υ
and
,, X 1∙ /( ^ + δ ∙x ) - ∕( ∙ x ) 1∙ n n
r
j
f (v x)' = Δhm
x→ 0
—------τ-∆ 2X— - = hm
Δx→ 0
0=0
as claimed. ■
■ DERIVATIVE OF A POWER FUNCTION
Recall that a power function is a function of the form ∕( x ) = x n where n is a real

number. For example, ∕( x ) = x 2 , g(x) = x^^3 , h(x) = x 1/2 are all power functions.
So are 1
F(x) = -^ = x~ 2 and G(x) = V x 2 = x 2z3
Here is a simple rule for finding the derivative of any power function.
THEOREM 2.4 Power rule
For any real num ber n, the power function ∕( x ) = x n has the derivative
∕'( x ) = n x n~'∙, in Leibniz notation,
⅛ (x ") = nx"~ l
Proof: If the exponent n is a positive integer, we can prove the power rule by
using the binomial theorem with the definition of derivative. Begin with the
difference quotient:
∕( x + ∆x) - ∕( x ) _ (x + ∆x)'l - x n
∆x - ∆x
= [x" + nx n ~' ∆x + "f'l , ~ 1V ~2(Δx)2 + ∙∙∙ + (∆x)', ] - x',

∆x
= nx"~ l ∆x + " f'n 2 1) x"^ 2(∆x)2 + + (∆x)',
∆x
= nx n~' + 2 '~ χ n ~2 ∆x + ∙∙∙ + (∆x)"^l
Note that ∆x is a factor of every term in this expression except the first. Hence, as
∆x → 0, we have
ι∙ f ( x + δ -y ) - ∕(∙x )
jf v(x)’ = hm —------ v - —
λ
Xv→0 ∆x
= Jimθ [ MX"^, + n^
~n 2 1 ) χ n ~2 δ -x + ” ■+ (Δx)"- l J
= nx n~ l
If n = 0, then ∕( x ) = x 0 = 1, so ∕' ( λ ') = θ∙ We shall prove the power rule for
negative integer exponents later in this section, and we shall deal with the case in
which the exponent is a rational number in Section 2.6. Note, however, that we
have already verified the power rule for the rational exponent ∣ in Example 3 of
Section 2.1, when we showed that the derivative of f(t) = ∖Γt = t' ,2 is ∕'( t ) =
⅛t~l '2 = -~y for t > 0. We shall introduce power functions with irrational
exponents (for example, x v 2 or x π) in Chapter 5, and at the same time we shall
derive the power rule for such functions. For now, however, we will assume that
the power rule holds when the exponent is a real number. ■
EXAM PLE 1 Using the power rule to find a derivative
Differentiate each of the following functions.
a. ∕( x ) = x 8 b. g(x) = x 3/2 c. ∕ι(x) =
Solution
a. Applying the power rule with n = 8, we find that
^ ( x 8) = 8x 8^ 1 = 8x 7
b. Applying the power rule with n = j , we get

_d _(y-3 /2 \ _ 3 (3/2)-1 _ 3 v - l∕2 _ 3
√ χ lλ ) - 2x ~ 2 x - 2
c. For this part you need to recognize that ∕z(x) = i y = x -5/3 . Applying the
power rule with n = —j , we find that 'λ
∆ t ∕ v -5 ∕3 ∖ _ _ 5 „-5 / 3 -1 _ _ 5 „-8 / 3
√A ΛΛ )- 3 x - 3 λ
■ PROCEDURAL RULES FOR FINDING DERIVATIVES
The next theorem expands the class of functions that we can differentiate easily by
giving rules for differentiating certain combinations of functions, such as sums,
differences, products, and quotients. We shall see that the derivative of a sum
(difference) is the sum (difference) of derivatives, but the derivative of a product
or a quotient does not have such a simple form.
For example, to convince yourself that the derivative of a product is not the
product of the separate derivatives, consider the power functions
f(x ) = x and g(x) = x 1
0 Note the product and quo
and their product
tient rules do not behave as you
p(x) =∕(x)g(x) = x 3
might expect. 0
Because∕'( x ) = 1 and g'(x) = 2x, the product of the derivatives is
f∖ x)g∖ x) = (l)(2λ') = 2x
whereas the actual derivative o fp(x) = x 3 is p'(x) = 3x2. The product rule tells us
how to find the derivative of a product.
THEOREM 2.5 Basic rules for combining derivatives—procedural forms
Iff and g are differentiable functions at all x, and a, b, and c are any real numbers,
then the functions cf, f + g,fg, and fig (for g(x) ≠ 0) are also differentiable, and
their derivatives satisfy the following formulas:
N ame of Rule Prime Notation Function N otation Leibniz Notation
∕ ( c ∕ ) = c⅛-
Constant multiple (<∕)' = cf' [c∕(x)]' = c∕'(v ) dxκ j ’ dx
, 4 - ( f + g) = ⅛ → 7Γ
Sum rule ( / + g)' = f ' + g' [ ∕W + S(*)Γ = f (×) + g'(χ) dxj si dx dx
τL f f - < Λ = V ~ d g
Difference rule ( f ~ g)' = f ' ~ g ' [f(χ) - £(*)]' = f'(χ ) - ∕ W dxj 6 dx dx
The constant multiple, sum, and difference rules can be combined into a single rule, which is called the linearity rule.
4 -(a f+ bg) = a%- + bfi -
Linearity rule (a f + bg)' = a f' + bg' [af(x) + ⅛ (x)]' = α ∕'(x ) + ⅛ '(x) dxκ j 6, dx dx
d_( f s } = f ⅛ + s f f
Product rule (fg)' = fg' + f ' g [∕(x)g(x)]' = f(x)g'(x) + f'(x)g(x) dx ' !g> ' dχ " dx
(f ∖ ' g f' ~ fg' Γ ∕(χ)T g(χ)f'(χ) - f(χ)g'(χ) ,z f ∖ d f_ dg

Quotient rule ∖ g)
=
g2 L^WJ
=
[g(x)]2 ≈tf) - ⅛N*
Proof of the Product Rule: We shall prove the product rule in detail, leaving
the other rules as problems. We begin by rearranging the difference quotient of the
product function p = fg so that it appears in a form involving the difference
quotients of f and g. We do this by considering areas, and to help with the
geometric representation, we assume that ∆x > 0 and that f and g are both
increasing functions. It should be noted that the algebraic steps stand alone
without considering area.
Begin with the difference quotient for the
function p.
gω-----------
p(X) = ⅛ 2 ¢ )
Area of rectangle Length Width /?lA)
The product of p can be represented as the area of a rectangle: x

p(χ + ∆x)
∖_________ J
= ∖∕( x + ∆x) ∖g(x + ∆x)
__________________ √ _________sz________ √
Area of large rectangle Length Width

Area of L-shaped region
p(x + ∆x) - p(x) = ∕( x + ∆x)g(x + ∆x) - ∕( x k ( x )
This difference is the “ outside (larger)” rectangle minus the “ inside (smaller)”
rectangle, giving the L-shaped region shown in color.
⅞(x + ∆ x ) - -
I Area of I: [g(% + ∆x) - g(x)]∕(x + Δ Λ )
sW-*
p(x + ∆.v Area of II: [ ∕( x + ∆x) -f(x)]g(x')
II
p(.v) Area of ∕Λ.v + ∆.v) - p(χ) = Area of I + Area of II
∕( x ) ∕ ( X + Δ Λ )
The key to the proof of the product rule is to rewrite the difference. We can
write this as the sum of regions I and II.
area of I area of II
p(x + ∆x) - p(x) = [g(x + ∆x) - g (x )]∕(x + ∆x) + [ ∕( x + ∆x) -∕(x)]<g(x)
Divide both sides by ∆x (where ∆x ≠ 0):
p(x + ∆x) - p(x) = [g(x + ∆x) - g(x)] [ ∕( x + ∆x) ~ ∕(x)]

∆x ∆x ∙λ λ + δ λj + ∆x
The last step in deriving the product rule is to take the limit as ∆x → 0.
dp ,. p(x + ∆x) - p(x)

- ~ r = hm —------ τ------ *i -
dx ∆%→o ∆x
,. ∫ ∕- ∕ , ΔΛ + Δ -T∙ - "(χ )l + , / J ∕ ( x + ∆x) -∕( x ) ^ ll
= ∆1v→o *T ------ Av— J ------ ∆ X JJ
ι∙ √-∕ , A ∖ 1∙ Γ<g(X + ∆x) - ,g(x)l Γ∕( x + ∆x) -∕(x)^ l

= hm ∕( x + ∆ x ) hm —------ τ - — ≡-l -z lim s w ⅛ [ -------- A x J
Δ x→ 0 Δ.Y→ 0 L ∆JV J ∆Λ→0 ∖______________ v______________ /
This is the derivative of g This is the derivative of f
= f(x )g , (x) + g (x )∕'(x ) lim f ( x + ∆x) = ∕( x ) because f is continuous ■
ΔΛ → 0
We might point out that in proving the previous theorem we are both proving
that the derivative of the sum, difference, product, and quotient exist and finding
appropriate formulas. For example, if we assume q(x) = f(x)∕g(x) where g(x) ≠ 0,
along with the existence o f∕'( x ) and g , (x), we need to prove two parts:
, '(x) exists
1. ⅛
g (χ )∕'(x ) - ∕( x ) √ ( x )
2. √(x) =
te(x)]2
EXAM PLE 2 Using the basic rules to find a derivative
a. ∕( x ) = 2x 2 - 5 V x b. p(x) = (3x 2 -1 )(7 + 2x 3)

c∙ tfW = ⅛ ⅛ 2 d ∙ = (4 * + 3 )2
3 .A
e .F ( x ) = ⅛ - j + f + ⅛ l
Solution
a. Apply the linearity rule (constant multiple, sum, and difference) and power
Historical Note
rules:
When working with rational
forms, we need to be careful ∕'( x ) = 2(x2)' - 5(x ιz2)' = 2(2x) - 5(∣)(x-172 ) = 4x - j x -ιz 2
about division by zero and the
possibility of obtaining an inde b. Apply the product rule; then apply the linearity and power rules:
terminate form. One of the ear
liest recorded treatments of in p'(x) = (3x2 -1 )(7 + 2x 3)' + (3x 2 - 1)'(7 + 2x 3)
determinate equations is
attributed to the Hindu mathe = (3x2 - l)[0 + 2(3x2)] + [3(2x) - 0](7 + 2x 3)
matician Aryabhaja (476-550?). = (3x2 - l)(6x 2) + (6x)(7 + 2x 3)
He gave rules for approxima
tions of square roots and sums = 6x(5x 3 - x + 7)
of arithmetic progressions as
well as rules for basic algebraic c. Apply the quotient rule, then the linearity and power rules:
manipulations. One example of _ (3 - x 2)(4x - 7)' - (4x - 7)(3 - x 2)'
his work is the following calcu q W (3 - x 2)2
lation for τr.∙ “Add four to one
hundred, multiply by eight and (3 - x 2)(4 - 0 ) - (4x - 7)(Q - 2x)
add again sixty-two thousand; (3 - x 2)2
the result is the approximate
value of the circumference of a : 12 - 4x 2 + 8x 2 - 14x _ 4x 2 - 14x + 12
circle whose diameter is twenty (3 - x 2)2 (3 - x 2)2
thousand.” It is remarkable
that today, most of us would d. Apply the product rule:
not attempt such a verification g'(x) = (4x + 3)(4x + 3)' + (4x + 3)'(4x + 3)
of the approximation 3.1416
without using a calculator. The = (4x + 3)(4) + (4)(4x + 3) = 8(4x + 3)
first Indian satellite was named
Sometimes when the exponent is 2, it is easier to expand before differenti
ARYABHAT in his honor.
ating:
ιg(x) = (4x + 3)2 = 16x2 + 24x + 9

g'(x) = 32x + 24
e. Write the function using negative exponents for rational expressions:
F(x) = jx ^ 2 - ∣x + 5 + 1 + x ^ 1
Then apply the power rule term by term to obtain
F'(x) = j( - 2 x ^ 3) - ∣ + 0 + 0 + ( - l) x ^ 2
In applying the power rule term by term in Example 2e, we really used the
following generalization of the linearity rule.
10 0 Ch. 2 Techniques of Differentiation with Selected Applications
COROLLARY TO THEOREM 2.5 The extended linearity rule
Iff x, f 2 , . . . , f n are differentiable functions and a x, a 2 , . . . , a n are constants, then

√ df df 7 df
S [α 1∕ 1 + a 2f 2 + ∙∙∙ + α π∕ π] = α 1^ + α ⅛E + + a ^dχ
2 n
Proof: The proof is a straightforward extension of the proof of the linearity

rule of Theorem 2.5. ■
Example 3 illustrates how the extended linearity rule can be used to differenti
ate a polynomial.
EXAM PLE 3 Derivative of a polynomial function
Differentiate the polynomial function p(x) = 2x 5 - 3x 2 + 8x - 5.
Solution p '(x) = ~j^∖ fχ 5 ~ 3χ 2 + 8x - 5]
- 2 ≡+>~ 3 ⅛+>÷8 ⅛ w-⅛ ( 5> Extended

linearity rule
= 2(5x 4 ) - 3 (2 x ) + 8 ( 1 ) - 0 Power rule',
constant rule
= 10x 4 - 6x + 8
EXAM PLE 4 Derivative of a product of polynomials
Differentiate p(x) = (x 3 - 4x + 7)(3x 5 - x 2 + 6x).

Solution We could expand the product function p(x) as a polynomial and
proceed as in Example 3, but it is easier to use the product rule.
p '(x) = (x 3 - 4x + 7)(3x 5 - x 2 + 6x)' + (x 3 - 4x + 7)'(3x 5 - x 2 + 6x)
= (x 3 - 4x + 7)(15x 4 - 2x + 6) + (3x2 - 4)(3x 5 - x 2 + 6x)
EXAM PLE 5 Equation of a tangent line
Find the standard form equation for the line tangent to the graph of
2
∕ ω = 2 x3x + 5
2 + x - 3
at the point where x = - 1 .

Solution Evaluating ∕( x ) at x = - 1 , we find th a t/ ( - 1) = - 4 (verify); therefore,
the point of tangency is ( - 1, -4 ). The slope of the tangent line at ( - 1 , - 4 ) is
∕ ' ( - 1). Find f '( x ) by applying the quotient rule:
, = (2x 2 + x - 3)(3x 2 + 5)' - (3x 2 + 5)(2x 2 + x - 3)'
J W (2X 2 + x - 3) 2
V ιB < 3 X *+ 5 )x< 2 X *+
X -3 ) (2x 2 + x - 3)(6x) - (3X 2 + 5X4x + 1)
X n in = ^5 V n in = "lΘ (2x 2 + x - 3)2
X∩ax=5 Yrnax=lΘ
X s c l= l Y s c l= l _ 12x 3 + 6x 2 - 18x - 12x 3 —3x 2 —20x —5
(2x 2 + x - 3)2
_ 3x 2 - 38x - 5
(2x 2 + x - 3)2
,e 2
τTh sιlope off tthκe ttangenτt line i■s r/ ,(-11)x = 3 ( - l) 2- 3 8 ( - 1 ) - 52 = 9.
n
[Z ( 1J ^r ( JJ
From the formula (Theorem 2.1) y = ∕'( x 0)(x - x 0) + ∕( x 0), we find that an
equation for the tangent line is
j> = 9(x + 1) + (-4)
or in standard form 9x - y + 5 = 0. The graphs of both f and its tangent line at
( - 1, -4 ) are shown in Figure 2.10. hbb
EXAMPLE 6 Finding horizontal tangent lines
Figure 2.10 Graph of f and the Let y = (x - 2)(x2 + 4x - 7). Find all points on this curve where the tangent
tangent line at the point (—1, —4) line is horizontal.
Solution The tangent line will be horizontal when dy/dx = 0, because the
derivative dy/dx is the slope and a horizontal line has slope 0. Applying the
product rule, we find
= (x - 2)(x2 + 4x - 7)' + (x - 2)'(x 2 + 4x - 7)
= (x - 2)(2x + 4) + (l)(x 2 + 4x - 7) = 2x 2 - 8 + x 2 + 4x - 7
= 3x2 + 4x - 15 = (3x - 5)(x + 3)
Thus, = 0 when x = j o r x = -3 . The corresponding points (j, and

(-3 , 50) are the points on the curve at which the tangent is horizontal, M B
In the following example, we use the quotient rule to extend the proof of the
power rule to the case in which the exponent n is a negative integer.
EXAMPLE 7 Proof of the power rule for negative exponents
Show that ^-,(x") = «x"-1 if n = -fli, where m is a positive integer.
Solution We have∕(x ) = x n = x~ m = l∕x m, so apply the quotient rule:

⅛ λ = J √ J j = * m( 1 )' ~ ( W m)'
= x m(θ) ~ m x m ~'
dx' x ' dx∖ xm ) (x ">y χ 2m
= -rn x i∙m ~r,~2m = -m x ~ m ~l = nx n ~' Substitute -m = n.
■ HIGHER DERIVATIVES
Occasionally, it is useful to differentiate the derivative of a function. In this

context, we shall refer to f as the first derivative of f and to the derivative off as
the second derivative off We could denote the second derivative by (f')', but for
simplicity we write f " . Other higher derivatives are defined and denoted
similarly. Thus, the third derivative o f/is the derivative off " and is denoted by
f ''. In general, for n > 3, the nth derivative off is denoted by f (n\ for example,
/ (4) or / (5). In Leibniz notation, higher derivatives for y = f(x) are denoted as
follows:
Computational Window Leibniz Notation
Many graphing calculators can dy

First derivative: y' ∕'(χ) dx Tχ-f'^
find numerical values for high
er derivatives. The TI-85, for II d2y
Second derivative: y ∕"(*) dx 2
example, has keys labeled
HI d 3y
Third derivative: y ∕ " ,w dx 3
di y
Fourth derivative: y>W ∕ ,4>ω dx 3
d"y
Λth derivative y(") ∕<">(x)
dx"
What this says: Because the derivative of a function is a function, differen

tiation can be applied over and over, as long as the derivative itself is a
differentiable function.
Notice also that for derivatives higher than the third, the parentheses
distinguish a derivative from a power. For example, ∕ 4 ≠ / (4).
You should also note that all higher derivatives of a polynomial p(x) will also be
polynomials, and if p has degree n, then p ,-k ∖ x) = 0 for k ≥ n + 1.
EXAM PLE 8 Higher derivatives for a polynomial function
Find the first and all higher derivatives of

p(x) = -2 x 4 + 9x3 - 5x2 + 7
Solution
p'(x) = —8x3 + 27x2 - lOx; p"(x) = -2 4 x 2 + 54x - 10;
p"'(χ) = -48x + 54; √ 4>(x) = -48; p<5>(x) = 0; . . .
p t-n∖ x) = 0 (« ≥ 5) ≡≡
PROBLEM SET 2 .2
θ In order to demonstrate the power o f the theorems o f this 7. a. ∕( x ) = x 3 + C b. g(x) = C 2 + x
section, Problems 1-4 ask you to go back and rework some
problems in Section 2.1, using the material o f this section 8. a. / ( 0 = 10 r 1 b. g(f) = 7
instead o f the definition o f derivative.
9- r(0 = f- - i + 10. ∕( x ) = ττ3 —3τr2
1. Find the derivatives of the functions given in Problems
3-8 of Problem Set 2.1.
11. ∕( x ) = - ^ + x 2z3 + C 12∙ ' w +τ + c
2. Find the derivatives of the functions given in Problems =⅛
9-14 of Problem Set 2.1. 2χ,- +11
13. ∕( x ) = -γ 3 + x [+ x ~ 7
14. s (.v) . y
15-20 of Problem Set 2.1. 15. ∕( x ) = (2x+ 1)(1 - 4 x 3) 16. g(x) = (x + 2)(2Vx + x 2)
17. ∕ W =⅜ ⅛ 18∙^)=⅛÷1
21-26 of Problem Set 2.1.
Differentiate the functions given in Problems 5-20. Assume 19. g(x) = x 2(x + 2)2 20. ∕(x ) = x 2(2x + 1)2
that C is a constant.
5. a. ∕( x ) = 3x 4 - 9 b. g(x) = 3(9)4 - x In Problems 21-24, fin d f ' , f " , f" ', and f w .
6. a. ∕( x ) = 5x 2 + x b. g(x) = τr3 21. ∕( x ) = x 5 —5x 3 + x + 12
22. ∕( x ) = ∣ x 8 - ∣ x 6 - x 2 + 2 44. Find equations for two tangent lines to the graph of
23. ∕( x ) = ∕W = ⅜ ⅛ ⅛
24 ∙ Λ v ) - ⅛ that are perpendicular to the line 2x - y - 1 = 0.

d 1v 45. Let f ix ) = (x 3 - 2x 2)(x + 2).
25. Find -τ-⅛, where y = 3x 3 - 7x 2 + 2x - 3. a. Find an equation for the tangent to the graph of/at the
aχ-
point where x = 1.
d 2v
26. Find where y = (x 2 + 4)(1 - 3x 3). b. Find an equation for the normal line to the graph o f/
at the point where x = 0.
In Problems 27-32, fin d the standard form equation for the 46. Find an equation for a normal line to the graph of
tangent line to y = f(x) at the specified point. f ix ) = (x 3 - 2x 2)(x + 2) that is parallel to the line
27. f ix ) = x 2 - 3x - 5, where x = - 2 x — 16y + 17 = 0.
28. ∕( x ) = x 5 - 3x 3 - 5x + 2, where x = 1 47. Find all points (x, ι∙,) on the graph o f y = 4x 2 with the
29. ∕( x ) = (x 2 + 1)(1 —x 3), where x = 1 property that the tangent at (x, y) passes through the point
(2, 0).
30. ∕( x ) = *1∣, where x = 0 48. Find the equations of all the tangents to the graph o f the
function
31. f ix ) = - - jf I , where x = 1
j v ’ x + 5 f ix ) = x 2 - 4x + 25
32. f ix ) = 1 - - + -⅜=, where x = 4 that pass through the origin.
7v ’ x fix
Determine which (if any) o f the functions y = fix ) given in
Find the coordinates o f every point o f the graph o f the functions
Problems 49-52 satisfy the equation y " ' + y " + y' = x + 1
given in Problems 33-39 where there is a horizontal tangent
49. f ix ) = x 2 + 2x - 3 50. f ix ) = x 3 + x 2 + x
line.
33. f ix ) = 2x 3 - 7x 2 + 8x - 3 34. g(x) = f i x - 5)(x - 8) 51. f ix ) = ∣ x 2 + 3 52. f ix ) = 2x 2 + x
35. fit ) = γ 2 ~ γ 3 36. f f ) = t* + 4 ∕ 3 - 8t2 + 3 θ 53. What is the relationship between the degree of a polyno
mial function P and the value of k for which P (fix ) is first
37. f ix ) = V x (x - 3) 38. A(x) = x ~~ 2 x + 1 equal to 0?
54. Prove the constant multiple rule ( c ∕) ' = c f'.
3>. ⅜ ) - ¾ ÷ i 55. Prove the sum rule (/ + g)' = f ' + g'.
56. Prove the difference rule I f - g)' = f ' - g '.
θ 40. a. Differentiate the function fix ) = 2 x 2 — 5x — 3.
57. Use the definition of derivative to find the derivative of
b. Factor the function in part a and differentiate by using ∕ 2, given that/is a differentiable function.
the product rule. Show that the two answers are the
58. Prove the product rule by using the result of Problem 57
same.
and the identity
41. a. Use the quotient rule to differentiate
fg = ^ f + g ) 1 - f 1 ~ g -]
∕w =⅛ i
59. If qfx) = fix)∕gfx), where g(x) ≠ 0, prove that q'fx) exists
i f ∕'( x ) and g'(x) exist.
b. Rewrite the function in part a as
60. Prove the quotient rule
f ix ) = x^ 3(2x - 3) (fX g f' ~ fg'
=
and differentiate by using the product rule. ∖g J g2
c. Rewrite the function in part a as where g(x) ≠ 0. Hint'. First show that the difference
quotient for f ig can be expressed as
f ix ) = 2x^ 2 - 3x^3
G X + A a) (g _ f i x + ∆x)g(x) - fix )g lx + ∆x)
and differentiate.
d. Show the answers to parts a, b, and c are all the same. ∆x (∆x)g(x + ∆x)g(x)
42. Find numbers a, b, and c that guarantee that the graph of
the function ∕( x ) = ax 2 + bx + c will have x-intercepts at and then subtract and add the term g(x)∕(x) in the
(0, 0) and (5, 0) and a tangent with slope 1 where x = 2. numerator.
43. Find the equation for the tangent to the curve with 61. Show that the reciprocal function r(x) = l ∕∕( x ) has the
equation y = x 4 - 2x + 1 that is parallel to the line derivative r'(x) = - ∕' ( x ) ∕[ ∕( x ) ] 2 at each point x where
2x - y - 3 = 0. / is differentiable and f ix ) ≠ 0.
62. If f g, and h are differentiable functions, show that the 64. Find constants A, B, and C so that
p ro d u c t// is also differentiable and
y = A x 3 + B x+ C
( / / ) ' = fgh' +fg'h + f'gh satisfies the equation
63. L et/be a function that is differentiable at x. y"' +2y" - 3 y ' + y = x
a. If g(x) = [∕(x)] 3, show that
65. J OURNAL P ROBLEM: The Amatyc Review* by Michael W.
g'(x) = 3[∕(x)] 2∕'( x ) Ecker. Let f(x) be a polynomial of degree n with the
property that ∕(x ) ≥ 0 for all x. Prove that
Hinf. Write g(x) = [∕(x)] 2∕(x ) and use the product rule.
b. Show that p(x) = [∕(x)] 4 has the derivative ∕(x ) + ∕'( x ) + ∕" (x ) + - + ∕< ''V ) ≥ θ
p'(x) = 4[∕(x)] 3∕'( x ) for all x.
2.3 DERIVATIVES OF THE TRIGONOMETRIC FUNCTIONS
IN THIS SECTION Two special trigonometric limits, derivatives of the sine

and the cosine, differentiation of the other trigonometric functions
In Section 2.2, we developed general rules for differentiation and applied those
rules to various algebraic functions involving power functions, polynomials,
and rational functions. In this section, we develop differentiation formulas for
the trigonometric functions.
■ TWO SPECIAL TRIGONOMETRIC LIMITS
We will need the following two basic limit properties (from Theorem 1.3)
involving the cosine and sine functions.
lim cosx = cost and lim sinx = sine
x →c x →c
In Example 6 of Section 1.5, we found lim m" λ by using a table. We will now
prove this same result by using a geometric argument.
THEOREM 2.6 Special limits involving sine and cosine
lim ⅛ ⅛ = l l i m cos⅛ - l = θ
A→O h Λ→O n
Proof: To prove the sine limit theorem requires some principles that are not
entirely obvious (see Problem 68). However, we can demonstrate its plausibility
by considering Figure 2.11 in which A O C is a sector of a circle of radius 1. (See
Student Mathematics Handbook for a review of the terminology about sectors.)
This derivation requires that Obe measured in radians. The line segments HZ) and
B C are drawn perpendicular to segment O C .
Assume 0 < ∕z < f ; that is, h is in Quadrant I.
∣ΛC∣ = h
∣ΛZ)∣ = sin A
∣5C∣ = tan/z = ⅛ ⅛
1 1 cos n
Figure 2.11 Trigonometric rela
tionships in the proof that ∣OZ)∣ = cos/z
lim ⅛ ⅛ = I
h→0 h *Volume 6, 1984, Issue 1, p. 55.
Sec. 2.3 Derivatives of the Trigonometric Functions 105
Now, compare the area of the sector A O C with those of ΔAOD and Δ B O C . In
particular, since the area of the circular sector of radius r and central angle θ is
SM H ⅛2 A (see page 2 of the Handbook), sector A O C must have area
∣ ( l) 2∕z = ∣A
We also find that ΔA O D has area
2∣OZ>∣∣∕1Z>∣ = J cos A sin A

and Δ B O C has area
>2 C ∣∣O C ∣1 = I2 = √ ιn z7
" cos A 2 cos A
By comparing areas (see Figure 2.11) we have:
AREA OF ∆AOD AREA OF SECTOR AOC AREA OF ∆BOC
⅜cosh sinh ≤ ∣ A ≤ sin A

2 cos A
Divide all parts of these inequalities by ∣ sin A:
cos A ≤ - Λ χ ≤ —¼-
sιn A cos A
Take reciprocals and reverse the inequality symbols:
cos A ≤ ⅛ 1
A cos A
This same inequality holds in the interval - J < A < 0. This can be shown by
using the trigonometric identities cos ( - A) = cos A and sin ( - A) = -s in A. Finally,
we take the limit of all parts as A → 0 to find
lim cos A ≤ lim ≤ lim —i-π

∕ι→o h→o n ∕ i →ocosA
Byj Theorem 1.3, lim cos A = cos0 = 1. Thus,
h→0
1 sin A 1
i
From the squeeze theorem (Theorem 1.9, Section 1.8), we conclude that
lim ⅛ = l.
h→0 A
<. , ,- 1 — cos A The second part of this theorem is left as a problem, but consider**
Graph of y = ----- -------
1 ~fos/z? ∖J
0 <
∖ 1 - cosA ∖
< smA D∙ill f,l ∖ , D∙ill f,l
What is lim
⅛→o h
for some constant k for which < k.
Examples 1 and 2 show how the two special limits in Theorem 2.6 can be used
to compute trigonometric limits.
EXAM PLE 1 Evaluation of a trigonometric limit
Find lim
.v→o 5x
*This string of inequalities was suggested by Leonard Gillman in the American Mathematical
Monthly, April 1991.
Solution We prepare the limit for evaluation by Theorem 2.6 by writing

sin 3x _ 3∣ sin 3x'j
5x 5∖ 3x )
Because 3x → 0 as x → 0, we can set h = 3x in Theorem 2.6 to obtain
i
lim ⅛ ⅛ = l i m ∣ ( ⅜ Y = ∣ l i m ⅞ M ( vl ) = 35
x→0 5x χ →o 5∖, 3x /
5 ∕∣→o Λ 5
E X A M P L E 2 Evaluation of a trigonometric limit
Find lim ≡ ≠
1 →o sm 7x
Solution As in Example 1, we First prepare the limit for evaluation by Theorem

2.6:
tan 2x _ / sin 2x∖/ 1 ∖ _ 2 /sin 2x∖/ 7x ∖/ 1 ∖
sin7x ∖cos2x∕∖sinlx ) 7∖ 2x ∕∖ s i n 7 x ∕∖ ∞ s 2 x ∕
Because 2x → 0 and 7x → 0 as x → 0, we have
lim t a n ∣ x = lim Γ2 /sin 2x∖/ 7x ∖/ 1 ∖^∣

x →o sm7x x →o
|_7\ 2X ∕∖ s in 7 x ∕∖ c o s 2 x ∕J
= ⅛flim s ⅛ 2 x Y lim ∙ 7 ⅞ -Y lim — U - )
7∖χ →o 2x ∕∖x→o sιn 7 x∕∖ x →o co s2 x ∕
i ( lim s⅛ 2x∖ /_____ 1------∖ Λ∙ ___ L__\
∕∖ x →0 2x ' lim s *n 7 λ ' )∖*→θ co s2 x ∕
∖ i→o l x /
= ¾ l)(τ)(l)= 7
■ DERIVATIVES OF THE SINE AND THE COSINE
By combining the limits in Theorem 2.6 with some elementary trigonometric

identities (see Student Mathematics Handbook for a review of the elementary
trigonometric identities), we obtain the following differentiation formulas for the
cosine and sine functions. In calculus we assume that the trigonometric functions
0 The formulas we state and are functions of real numbers or of angles measured in radians. We make this
derive assume radian measure. 0 assumption because the trigonometric differentiation formulas become more
complicated if degree measurement is used instead of radians.
T H E O R E M 2.7 Derivatives of the sine and cosine functions

The functions sinx and cosx are differentiable for all x and
Y - s in x = cosx cosx = - s in x
Proof: The proofs of these two formulas are similar. We shall prove the first
using the trigonometric identity
sin (α + β) = sin a cos β + COS a sin β
and leave the proof of the second formula as a problem. From the definition of the
derivative
_d_ <,ιl l Av _ 11111

,j _ _ o ln
ιι rτι sin (% +∖ ∆x) - sinx
ax ΔΛ-→O ΔX
_ ι1∙rn sinx cos∆x + cosx sin∆x - sinx

Δv→0 ∆x
= lim ∣^sinx + cosx ( ⅛ ^ ) - ⅛ ^ 1
Δ,v→0 L ∖ ∆x ) ∖ ∆x / ∆x J
= (sinx) Jim ( c 0 s ⅛ ~ 1 ) + (cosx) Jim
= sinx (0) + cosx (1)

= cosx
≡
Before stating the theorem that derives the derivative of sine and cosine,
suppose we look at the graph of the difference quotient. Consider
∕( x ) = sinx. Then
sin(x + ∆x) - sinx _ sin(x + 0.01) - sinx
∆x 0.01
is the difference quotient. The graph shown at the right is
Y l = (sin(X + 0 .0 1 )- sin(X))∕0.01 V ιΞ < sin <X+.Θ1)-
By looking at the graph of this difference quotient, it looks like the derivative s in X)×.Θ1
of∕( x ) = sinx is∕'( x ) = cosx. We now verify this with the following theorem. X n in = -9. 424777...
Xna×=9.4247779...
Xsc 1=1.5707963...
V∩in= - 1
Ynax=l
Y ≥ c l= .1
EXAM PLE 3 Derivative involving a trigonometric function
Differentiate∕( x ) = 2x 4 + 3 cosx + sinα, for constant a
Solution ∕'( x ) = (2x4 + 3 cosx + sinα)
= 2 ^ (x 4) ÷ 3 (cosx) + (sinα) Linearity rule
= 2(4x3) + 3( —sinx) + 0 Power rule, derivative of cosine, and

. , . . derivative o f a constant
= 8x j - 3 sinx r—
EXAM PLE 4 Derivative of a trigonometric function with product rule
Differentiate ∕( x ) = x 2sinx.
Solution ∕'( x ) = (x 2sinx)
= x2 (sinx) + sinx (x 2) Product rule
= x 2cosx + 2x sinx Power rule and derivative of sine

EXAM PLE 5 Derivative of a trigonometric function with quotient rule
Differentiate ∕z(∕) = - ^ - ∙
v ’ cos t
Solution Write V t as ∕ ιz2.
iτ'(t(∖) - dt ∣^∕ ιz21

d Lcos∕]
c o s/⅛ v( ∕ ιz2), —∕ ιz2 ⅛ cos/ z~, .. . ,

= dt _____ dt Quotient rule
cos2/
= ¼ ιz2 cos / - ^ ∕ ιz2( - s in /) p o w er ru le a n d

COS2/ derivative of sine
= ^ →z2(cos∕ + 2 ∕s in ∕) c o m m o n fa c to r ⅛-m
cos2/
_ cos / + 2/ sin t
2∖Γt cos2/
■ DIFFERENTIATION OF THE OTHER TRIGONOMETRIC

FUNCTIONS
You will need to be able to differentiate not only the sine and cosine functions, but
also the other trigonometric functions. In order to find the derivatives of these
functions you must remember the following identities, which are given in the
Student Mathematics Handbook.
tanx = ^ ≡ c o tx = cosx
cosx sinx
secx = —-— cscx = - J —

cosx sinx
You will also need the following identity:
cos2x + sin2x = 1
THEOREM 2.8 Derivatives of the trigonometric functions
The six basic trigonometric functions sinx, cosx, tanx, cscx, secx, and cotx are
all differentiable wherever they are defined, and
sinx = cosx cosx = -sin x
d tanx = sec^2>x cotx = —csc2x
secx = secx tanx CSCx = -cscx cotx
Proof: The derivatives for sine and cosine were given in Theorem 2.7. All the
other derivatives in this theorem are proved by using the appropriate quotient
rules along with formulas for the derivatives of the sine and cosine. We will obtain
the derivative of the tangent function and leave the rest as problems.
_/ tt a n r
_ dL sinx Trigonometric identity
dx dx cos x
_ cosx sin x - sin x cosx

Quotient rule
cos zx
_ cosx(cosx) - sin x ( - sinx)
Derivative of sinx and
cos 2x cosx
_ cos 2x + sin 2x
cos 2x
1
cos2x + sin2x = 1
cos 2x
= sec 2x l∕cos 2x = sec2x
EXAM PLE 6 Derivative of a trigonometric function with the product rule
D ifferentiate/(0 ) = 30 sec0.
Solution ∕'( 0 ) = ⅛ (3 0 s e c 0 )
= 30 sec 0 + sec 0 ^ ( 3 0 ) Product rule
= 30 sec 0 tan 0 + 3 sec 0
EXAM PLE 7 Derivative of a product of trigonometric functions
Differentiate ∕( x ) = secx tanx.
Solution ∕'( x ) = (secx tanx)
= secx ta n x + ta n x secx Product rule
= secx (sec 2x) + ta n x (secx tanx)

= sec 3x + secx tan 2x
EXAM PLE 8 Equation of a tangent line involving a trigonometric function
Find the equation of the tangent to the curve y = co tx - 2 cscx at the point
where x =
Solution The slope of the tangent is the derivative of y with respect to x at the
point x = f.
dy d , x 2 c
Tx = T x ^ ~ ^
= co tx —2 cscx Linearity rule
= - c s c 2x - 2(- c s c x cotx)
= 2 cscx co tx - csc 2x
Writing this expression in terms of sine and cosine, we find

- . 2 π[ 1 V COS X ) 1
2 cscx cotx - csc2x =
sin2x
_ 2 cosx - 1
sin 2x
dy∖ = 2 cos⅞ - 1 = 2⅛) - 1 = 0 =
√X∣x=π∕3 sin2 f (√3)2 2
Thus, the tangent line at the point where x = y has slope 0, which means that
the tangent line is horizontal. When x = j , we have
y = c o tf - 2 c s c f = ^ - 2(⅛≡) = ≡⅜5 = - √ 3
so the point of tangency is (j, - V z3). The desired equation is
J , + √3 = 0
If you are using a software program to calculate derivatives, such as Mathematica, Derive, or Maple, the form of
the derivative may vary. For example, you might obtain
⅛ (tanx) = tan 2 x + 1 ⅛ (∞ tx) = - c o t 2 x - 1 ⅛ (secx) = ⅛ cscx) = - g g
Although these are different forms than those shown in this section, you should notice that they are equivalent by
recalling some of the fundamental identities from trigonometry.
PROBLEM SET 2 .3
© Evaluate the limits in Problems 1-12. 15. g(f) = t 2 + cos / + cosf
1. lim sin 2x 2. lim
sin4x 16. g(f) = 2 sec / + 3 tan / - ta n j
Λ-→0 X x →0 9x 17. p(x) = x 2cosx
3. lim sin 3x 4. lim sin 3/ 18. p(t) = ( ∕ 2 + 2)sin∕
x →0 sin 2x !→0 cos/
19. fit) = sin 2 /. Hint: Use the product rule on (sin∕)(sin∕).
5. l i m tan 5/ 6. l i m cot 3x 20. gl×) = cos2 x. Hint: Use the product rule on (cosx)(cosx).
f→0 tan 2/ x →0 cotx
2 21. fix ) = Vx cosx + x cotx
7. l i m 1 - cosx 8. lim sin x
Λ→0 sinx x →0 2x 22. fix ) = 2x 3sinx - 3x cosx
sin (cosx)
9. l i m 10. lim tan2x cotx 23. qlx) = ~l - v , = -Λ -
24. r(x)
Λ→0 secx Λ→0 x sinx
11. lim sin 2x 2

12. lim x cos 2x 25. h f) = 26. flθ ) =
.γ→0 x2 x→0 1 - cosx I v
27. fix ) = -ii g s f - 28. g f ) = 1 + *i n r
j v 7 1 - 2x 6 v 7 yz
Differentiate the functions given in Problems 13-38.
13. fix ) = sinx + cosx 14. ∕( x ) = 2 sinx + tanx 29. fit) = 2 ÷ s i ° f 30. flθ ) = ω
1
a
j v 7 t+2 j v 7 2 + cos θ
sinx
31. ∕( x ) = 32. ∕( x ) = 64. Complete the proof o f Theorem 2.7 by showing that
1 — cosx 1 — sinx
cosx = - s in x . Hint: You will need to use the identity
33. ∕( x ) =
1 + sinx COSX
34. g(x) = cos (a + β) = cos a cos β - sin a sin β.
2 - cosx 1 + cosx
sinx + cosx x 2 + tanx Prove the requested parts o f Theorem 2.8 in Problems 65-67.
35. ∕( x ) 36. ∕( x )
sin x - cosx 3x + 2 tanx
65. 4 - cotx = - c s c 2 x
37. g(x) = sec2x - tan 2x + cosx dx
38. g(x) = cos 2x + sin 2x + sinx 66. 4 - secx = secx tanx
dx
Find the second derivative o f each function given in Problems
39-48. 67. 4~ escx = - c o t x cscx
39. f(θ) = sin θ
dx
40. /(0) = cos θ
68. J O U R N A L P RO BLEM The theorem
41. f(θ) = tan 0 42. /(0) = cot 0
43. f(θ) = sec θ 44. /(0) = esc 0
45. ∕( x ) = sinx + cosx 46. ∕( x ) = x sinx
47. g(y) = esc y - coty 48. g(f) = sec t - tan t stated in this section has been the subject of much
discussion in mathematical journals:
W. B. Gearhart, and H. S. Shultz, “ The Function

Computational Window sinx/x,” College Mathematics Journal (1990): 90-99.
L. Gillman, “ TT and the Limit of ( s in α ) ∕α ∕ American
49. Graph the difference quotient for∕( x ) = sinx, where Mathematical Monthly (1991): 345-348.
∆x = 0.1, and the cosine function on the same axis
F. Richman, “ A Circular Argument,” College Mathemat
with Xm in = 0 and Xm ax = 1π. That is, let
ics Journal (1993): 160-162.
Y l = (sin(X + .1) - sin (X ))∕.l D . A. Rose, “ The Differentiability o f sin x,” College
Mathematics Journal (1991): 139-142.
Y2 = cos X
P. Ungar, “ Reviews,” American Mathematical Monthly
Use (TRACE) to compare the values of the difference (1986): 221-230.
quotient and the cosine function. What do you
notice? Some of these articles argue that the demonstration
50. Repeat Problem 49 for ∆x = 0.01 and then for shown in the text is circular, since we use the fact that the
∆x = 0.001. What do you notice? area of a circle is ττr2 . How do we know the area of a
circle? The answer, o f course, is that we learned it in
elementary school, but that does not constitute a proof.
Find the equation for the tangent line at the prescribed point for On the other hand, some of these articles argue that the
each function in Problems 51-57. reasoning is not necessarily circular. Read one or more of
51. /(0) = tan 0 at (jf, 1) 52. /(0) = sec 0 at (j, 2) these journal articles and write a report.
69. a. Show that sin x < x if 0 < x < f . Refer to Fig
53. ∕( x ) = sinx, where x = f 54. ∕( x ) = cosx, where x = j
ure 2.12. Hint: Compare the area o f an appropriate
55. y = x + sin x, where x = 0 56. x = t sec t, where t = 0
triangle and sector.
57. y = x cosx, where x = ?
b. Show that ∣sinx∣ < ∣x∣ if 0 < ∣x∣ < f .
d 2v
θ 58. If y = 2 sinx + 3 cosx, what is -p½2 + y? c. Use the definition of continuity to show that sinx is
dx
continuous at x = 0.
d 2y
59. If v = 4 cosx - 2 sinx, what is ~τs + y? d. Use the formula sin (α + β) = sin a cos β + sin β cos a
dx2
to show that sinx is continuous for all real x.
60. For what values of A and B does y = A cosx + B sinx
satisfy y " + 2y' + 3y = 2 sinx?
61. For what values of A and B does
y = A x cosx + B x sinx
satisfy y " + y = - 3 cosx?

θ 62. T H IN K T A N K P ROBLEM Give an example of a differenti
able function with a discontinuous derivative.
63. Prove that „„,
„cos ,
1 Figure 2.12 Figure for Problems 69 and 70
lim
Λ→O h
l~= 0
70. Follow the procedure outlined in Problem 69 to show that Hint: Start with Figure 2.11 (the same figure as the one
cosx is continuous for all x. Hint. After showing that used in the proof of Theorem 2.6).
72. Use the limit of a difference quotient to prove that
lim cosx = 1
Λ'→ 0
⅛ -y-tanx = sec2x
you may need the identity ax
73. J OURNAL P ROBLEM Write a short paper on difficulties of
cos(x + h) = cosx cos A - sinx sin A differentiating trignometric functions measured in de
grees.*
71. Use the squeeze theorem to prove that
74. J OURNAL T HINK T ANK P ROBLEM (From the American
lim ⅛ = 1 Mathematical Monthly). Suppose f and g are differentia
h→0 h ble real-valued functions defined on (-∞, +∞). Must there
exist a differentiable real-valued function h defined on
(-∞, +∞) such that h' = f'g "l
2 . 4 RATES OF CHANGE: RECTILINEAR MOTION
IN THIS SECTION Rate of change (geometric interpretation), average and

instantaneous rate of change, rectilinear motion (physics application), falling
body problem.
The derivative can be interpreted as a rate of change, which leads to a wide
variety of applications. Viewed as rates of change, derivatives may represent
such quantities as the speed of a moving object, the rate at which a population
grows, a manufacturer’s marginal cost, the rate of inflation, or the rate at which
natural resources are being depleted.
a. Linear function: rate of ■ RATE OF CHANGE — GEOMETRIC INTERPRETATION

change ∆y∕∆.v is constant.
Let us begin by thinking of rate of change in geometric terms. Intuitively, a rate of
change means how fast one variable changes with respect to another variable.
The rate of change of a linear function ∕(x ) = ax + b with respect to its
independent variable x is measured by the steepness or slope of its straight-line
graph, as shown in Figure 2.l3a. For the graph of a linear function, the slope or
rate of change is constant. However, if the function under consideration is not
linear, its rate of change varies from point to point, as shown in Figure 2.13b.
Because the slope of the tangent is given by the derivative of the function, the
preceding geometric observations suggest that the rate of change of a function is
measured by its derivative. This connection will be made more precise in the
following discussion.
b. Nonlinear function: rate of change
∆y∕∆.r depends on chosen points. ■ AVERAGE AND INSTANTANEOUS RATE OF CHANGE
Figure 2.13 Rate of change is
measured by the slope of a tangent Suppose that y is a function of x, say, y = f(x). Corresponding to a change from x
line. to x + ∆x, the variable y changes from ∕(x ) to ∕( x + ∆x). Thus the change in y is
∆p = ∕( x + ∆x) —f(x), and then the average rate of change of j with respect to x is
change in y _ ∆y _ /( x + ∆x) - ∕( x )
AVERAGE RATE OF CHANGE =
change in x ∆x ∆x
0 The formula for ∆v is impor As the interval over which we are averaging becomes shorter (that is, as
tant. Note the change in y in ∆x → 0), the average rate of change approaches what we would intuitively call the
Figure 2.14. 0
*See, for example, “Fallacies, Flaws, and Flimflam,” The College Mathematics Journal, Vol. 23,
No. 3, May 1992 and Vol. 24, No. 4, September 1993.
Sec. 2.4 Rates of Change: Rectilinear Motion 113
instantaneous rate of change of y with respect to x, and the difference quotient

approaches the derivative f'(x), or Thus, we have
INSTANTANEOUS .. ∆y ,. f( x + ∆x) - f(x) .,,

RATE OF CHANGE * ⅛ ∆t ^ ⅛ ---------- ∆X---------- = < « x
To summarize:
Instantaneous Rate of Change
-I------------------ 1----- If y = f(x), the instantaneous rate of change of y with respect to x, is given by
X ___ V+ ∆.v the derivative of f
..,, , dy
INSTANTANEOUS RATE OF CHANGE OF y = / (x) = -j~
Figure 2.14 A change in ∆y corre
sponding to a change ∆x
EXAMPLE 1 Instantaneous rate of change
Find the rate at which the function y = x 2sinx is changing with respect to x
when x = 77.
Solution For any x, the instantaneous rate of change is the derivative,
dy . . -,
= 2x stnx + x^cosx
Thus, the rate when x = ττ is
_ = 2ττ sin 77 + 772COS τr = 2τ7(0) + τr2( - 1) = —τr2
The negative sign indicates that when x = 77, the function is decreasing at the
rate of τr2 ≈ 9.9 units of y for each one-unit increase in x. MM
Let us consider an example comparing the average rate of change and the
instantaneous rate of change.
EXAM PLE 2 Comparison between average rate and instantaneous rate of

change
Let f(x) = x 2 - 4x + 7.
a. Find the instantaneous rate of change of f a t x = 3.

b. Find the average rate of change of f with respect to x between x = 3 and 5.
Solution
a. The instantaneous rate of change of the function with respect to x is the
derivative of the function; that is,
∕'( x ) = 2x - 4
The instantaneous rate of change o f /’at x = 3 is
∕'( 3 ) = 2(3) - 4 = 2
The tangent line at x = 3 has slope 2, as shown in Figure 2.15.
b. The (average) rate of change from x = 3 to x = 5 is found by dividing the
Figure 2.15 Comparison of in change in f by the change in x. That is, the change in/from x = 3 to x = 5
stantaneous and average rates of
is
change from x 1 to x 2
/(5) - /( 3 ) = [52 - 4(5) + 7] - [32 - 4(3) + 7] = 8
The average rate of change is

/(5) -/ (3 ) . 8 _
5 -3 2 ■
The slope of the secant line is 4, as shown in Figure 2.14.
■ AN APPLICATION TO PHYSICS— RECTILINEAR MOTION
Rectilinear motion is motion along a straight line. For example, the up and down
motion of a yo-yo may be regarded as rectilinear, as may the motion of a rocket
early in its flight.
When studying rectilinear motion, we may assume that the object is moving
along a coordinate line. The position or displacement of the object from the origin
in relation to the line is a function of time t and is often expressed as s(t). The rate
of change of the displacement with respect to time is the object’s velocity v(Z), and
the rate of change of the velocity with respect to t is its acceleration α(Z).
Interpreting these rates as derivatives, we see that
. ∙ //c
Velocity is v(z) =
Acceleration is a(f) = yp = ⅛ f.
wt cιr
If v(Z) > 0, we say that the object is advancing and if v(Z) < 0, the object is
retreating. If v(Z) = 0, the object is neither advancing nor retreating, and we say it
is stationary. The object is accelerating when a{t) > 0 and is decelerating when
α(Z) < 0. The significance of the acceleration is that it gives the rate at which the
velocity is changing. These ideas are summarized in the following box.
Rectilinear Motion An object that moves along a straight line with displacement s(t) has velocity
v(t) = ⅛ a n d a c c e ê r a don a(f) ~ ~ft = when these derivatives exist. The
speed of the object is ∣v(Z)∣.
0 Rectilinear motion involves displacement, velocity, and acceleration.

Sometimes there is confusion between the words speed and velocity. Because
speed is the absolute value of the velocity, it indicates how fast an object is
moving, whereas velocity indicates both speed and direction (relative to a
given coordinate system). 0
N OTATIONAL C O M M E N T : If distance is measured in meters and time in
seconds, velocity is measured in meters per second (m∕s) and acceleration in
meters per second per second (m∕s∕s). The notation m/s/s is awkward, so m ∕s 2 is
more commonly used. Similarly, if distance is measured in feet, velocity is
measured in feet per second (ft∕s) and acceleration in feet per second per second
(ft∕s 2).
When you are riding in a car,
moving along a straight road,
you do not feel the velocity, but E X A M P L E 3 The position, velocity, and acceleration of a moving object
you do feel the acceleration.
Assume that the displacement at time Z of an object moving along a line is
That is, you feel changes in the
given by
velocity. s(t) = 3Z3 - 40.5Z2 + 162z
for Z on [0, 8]. Find the position, velocity, and acceleration.
.ττ.¾g>ς n ⅞ ∙ " ' r , π t " . t ∙ t √ . M ⅛ . κ > r ' n υ ⅛ v ∙,,^ ι

[ Computational Window B Solution The position at tim e t is given by the function s. The starting position is
found at tim e t = 0, so
You can use a graphing calcula
tor or computer to look at the s∙(0) = 0 The object starts at the origin.
graphs of the position, veloci The velocity is determ ined by finding the derivative o f the position function.
ty, and acceleration functions.
v(Z) = s'(f) = 9 t 2 — 8 It + 162 The starting velocity is
Y l = 3X^3 - 40.5X^2 + 162X — o( f 2 — Ot + 1st w ^ c ^l f o u n d by
* ' evaluating v(0).
Y2 = 9X ^2-81X + 162
= 9(t - 3)(t - 6)
Y3 = 18x - 81
W hen t = 3 and w hen t = 6, v the velocity is 0, which m eans the object is
stationary at those tim es. F urtherm ore,
v(t) :> 0 on [0, 3) Object is advancing.
v(7) <; 0 on (3, 6) Object is retreating.
v(t) 2> 0 on (6, 8] Object is advancing.
a(t) = √'(Z) = v'(z) = 1 8 Z - 81 = 18(Z - 4.5)
decreasing We see th at a(t) < 0 on [0, 4.5) Velocity is decreasing-, that is,
decreasing increasing the object is decelerating.
a: increasing a(f) = 0 at t = 4.5 Velocity is not changing.

Notice that the velocity is in a(t) > 0 on (4.5, 8] Velocity is increasing; that is,
creasing when both the posi the object is accelerating.
tion and acceleration functions
have the same sign. T he table (com putational window) shown in the m argin gives values for s, v,
and a. We use these values to plot a few points, as shown in Figure 2.16. The
actual path o f the object is back and forth on the axis and the figure is for
clarification only.
t = 4.5
v = -20 .2 5
r= 6 r= 8
a = 27 a = 63
Computational Window H Q !
r = 4.5
The table shows representative
values of 5, v, and a. α = -2 7
t s V a
0 0 162 -81
1 124.5 90 -6 3
2 186 36 -4 5
3 202.5 0 -2 7
4 192 -1 8 -9 a. Position of the object: when v > 0, the b. Velocity of the object: when a > 0, the
4.5 182.25 -20.25 0 object advances, and when v < 0, the velocity increases, and when a < 0, the
object retreats. velocity decreases.
5 172.5 -1 8 9
6 162 0 27 Figure 2.16 Analysis of rectilinear motion
7 178.5 36 45
8 240 90 63 Recall that the speed o f the particle is the absolute value o f its velocity.
The speed decreases from 162 to 0 between t = 0 and t = 3, and it increases
from 0 to 20.25 as the velocity becom es negative between t = 3 and t = 4.5.
T hen, for 4.5 < t < 6, the particle slows down again, from 20.25 to 0, after
which it speeds up. B B
A com m on m istake is to th in k th at a particle m oving on a straight line speeds

up when its acceleration is positive and slows down when the acceleration is
negative, but this is not quite correct. Instead, the following is generally true:
• The speed increases (particle speeds up) when the velocity and acceleration
have the same signs.
, The speed decreases (particle slows down) when the velocity and accelera
tion have opposite signs.
■ FALLING BODY PROBLEM
As a second example of rectilinear motion, we shall consider a falling body

problem. In such a problem, it is assumed that an object is projected (that is,
thrown, fired, dropped, etc.) vertically in such a way that the only acceleration
acting on the object is the constant downward acceleration g due to gravity, which
on the earth near sea level is approximately 32 ft∕s 2 or 9.8 m ∕s 2. At time t, the
height of the object is given by the following formula:
Formula for the Height of an Object

v0 is the initial velocity.
h(t) = -⅛gt2 + v0 t + 50
g is the acceleration due to gravity.
where s0 and vθ are the object’s initial height and velocity, respectively. We shall
derive this formula in Chapter 4.
EXAM PLE 4 Position, velocity, and acceleration of a falling object

r= 3
h = 320
v=0
Suppose a person standing at the top of the Tower of Pisa (176 ft high) throws
a ball directly upward with an initial speed of 96 ft/s.
a. Find the ball’s height, its velocity, and acceleration at time t.
b. When does the ball hit the ground, and what is its impact velocity?
c. How far does the ball travel during its flight?
Solution First, draw a picture such as the one shown in Figure 2.17 to help you
understand the problem.
a. Substitute the known values into the height of an object formula:
v0 = 96 ft/s: Initial velocity
I
h(t) = -∣( 3 2 ) ∕ 2 + 96/ + — A = 176 ft: Height o f tower
↑
g = 32 ft/s2 : Constant downward gravitational acceleration
h(f) = - 1 6 ∕ 2 + 96/ + 176 This is the displacement, or position,
function.
It gives the height of the ball.
The velocity at time / is the derivative:
Figure 2.17 The motion o f a ball
thrown upward from the Tower of v(0 = ⅞ = -32/ + 96
Pisa
The acceleration is the derivative of the velocity function:
a (f
()
∖= dv = d l *1h = - 3 2
dt dt
This means that the velocity of the ball is always decreasing at the rate of
32 ft∕s 276*3.
b. The ball hits the ground when ∕z(Z) = 0. Solve the equation
-16z 2 + 96z + 176 = 0
to find that this occurs when Z ≈ -1 .4 7 (Z = 3 - 2Vz5) and t ≈ rlA1
(t = 3 + 2Vz5). Disregarding the negative value we see that impact occurs at
t = 3 + 2Vz5 sec. The impact velocity is
v(3 + 2√5) = -32(3 + 2√5) + 96 ≈ - 143 ft/s
The negative sign here means the ball is coming down at the moment of
impact.
c. The ball travels upward for some time and then falls downward to the
ground, as shown in Figure 2.17. We need to find the distance it travels
upward plus the distance it falls to the ground. The turning point at the top
(the highest point) occurs when the velocity is zero. Solve the equation
-32z + 96 = 0
to find that this occurs when t = 3. For Z < 3, the velocity is positive and
the ball is rising, and for Z > 3, the ball is falling. It follows that the ball is
at the highest point when Z= 3. The ball starts at Λ(0) = 176 ft and rises to
a maximum height when Z= 3:
Λ(3) = - 16(3)2 + 96(3) + 176 = 320
Thus, the total distance traveled is
Upward distance Downward distance
(320 - 176) + 320 = 464
↑
Initial height
The total distance traveled is 464 ft.
PROBLEM SET 2 .4
θ For each function f given in Problems 1-14, fin d the rate o f 8. ∕( x ) = (x + 1) sinx when x 0 = f
change with respect to x at x = x 0 .
9. ∕( x ) = x + _ 4χ. when x 0 = 0
1. ∕( x ) = x 2 —3x + 5 when x 0 = 2 2
2. fix') = 14 + x - x 2 when x 0 = 1 10. fix') = — ⅞-r ------ L γ when x∩ = 3

j v ’ x + 1 x - 1 0
3. ∕( x ) = - 2 x 2 + x + 4 for x 0 = 1
11. fix') = sinx cosx when x 0 = ?
4 ∙ Λ x ') = f ~ f f fo r x = 1 2
0 r
12. jf ivx ) = ----- when x n0= 1
-T;
x2 + 1
5∙ = 3x + 5 w h e n x °= ~1
13. f( x ) = (x - ∣ ) when x 0 = 1
6. ∕( x ) = (x 2 + 2)(x + V x) when x 0 = 4
7. ∕( x ) = x co sx when x () = π 14. fix') = sin 2x when x 0 = f
Give both exact and approximate values.
The function s(t) in Problems 15-22 gives the displacement o f a. At what speed (to the nearest tenth m∕min) is the
an object moving along a line. In each case: object moving at the end of 4 min?
a. Find the velocity at time t. b. How far (to the nearest tenth m) does the object
b. Find the acceleration at time t. actually travel during the 5th minute?
c. Describe the motion o f the object; that is, tell where is it 28. A bucket containing 5 gal of water has a leak. After Z
advancing and where it is retreating. Compute the total seconds, there are
distance traveled by the object during the indicated time Q(Z) = 5(1 - ⅛ ) 2
interval.
d. Tell where the object is accelerating and where it is decelerat gallons of water in the bucket.
ing.
a. At what rate (to the nearest hundredth gal) is water
⅞W For a review o f solving equations see Sections 2.5, 2.6, and 3.7 leaking from the bucket after 2 sec?
Ji j o f the Student Mathematics Handbook. b. How long does it take for all the water to leak out of the
15. s(f) = t 2 - 2t + 6 on [0, 2] bucket?
16. s(t) = 3t2 + 2t - 5 on [0, 1] c. At what rate is the water leaking when the last drop
17. s(t) = t 3 - 9z2 + 1St + 25 on [0, 6] leaks out?
18. s(t) =t4 - 4z3 + 8t on [0, 4] 29. A person standing at the edge of a cliff throws a rock
directly upward. It is observed that 2 sec later the rock is
19. s(f) = for 1≤ t ≤ 3 at its maximum height (in feet) and that 5 sec after that, it
hits the ground at the base of the cliff.
20. s(f) = l C i for 1≤ t ≤ 2 a. What is the initial velocity of the rock?
b. How high is the cliff?
21. s(t) =3 cosZ for 0 ≤ Z ≤ 2π
c. What is the velocity of the rock at time Z?
22. s(t) = 1 + sec Zfor 0 ≤ t ≤ % d. With what velocity does the rock hit the ground?
© 23. It is estimated that x years from now, 0 ≤ x ≤ 10, the 30. A projectile is shot upward from the earth with an initial
average SAT mathematics score of the incoming students velocity of 320 ft/s.
at a certain eastern liberal arts college will be a. What is its velocity after 5 sec?
f( x ) = - 6 x + 582 b. What is its acceleration after 3 sec?
Derive an expression for the rate at which the average
a. 31. A rock is dropped from a height of 90 ft. One second later
SAT score will be changing with respect to time x years another rock is dropped from height H. What is H (to the
from now. nearest foot) if the two rocks hit the ground at the same
time?
b. What is the significance of the fact that the expression
in part a is a negative constant? 32. A ball is thrown vertically upward from the ground with
24. A particle moving on the x-axis has displacement
an initial velocity of 160 ft/s.
a. When will the ball hit the ground?
x(Z) = 2z3 + 3z2 - 36Z + 40
b. With what speed will the ball hit the ground?
after an elapsed time of t seconds. c. When will the ball reach its maximum height?
a. Find the velocity of the particle at time Z. d. What is the speed of the ball as it hits the ground?
b. Find the acceleration at time I. 33. An object is dropped (initial velocity v0 = 0) from the top
c. What is the total distance traveled by the particle of a building and falls 3 seconds before hitting the
during the first 3 sec? pavement below. Find the height of the building in feet.
25. An object moving on the x-axis has displacement 34. An astronaut standing at the edge of a cliff on the moon
x(z) = z3 - 9z2 + 24z + 20 throws a rock directly upward and observes that it passes
her on the way down exactly 4 sec later. Three seconds
after t seconds. What is the total distance traveled by the after that, the rock hits the ground at the base of the cliff.
object during the first 8 sec? Use this information to determine the initial velocity v0
26. A car has velocity and the height of the cliff. Note; g = 5.5 ft∕s 2 on the moon.
35. Answer the questions in Problem 34 assuming the astro
naut is on Mars, where g = 12 ft∕s 2.
ft/s after t seconds of motion. What is its acceleration (to 36. A car is traveling at 88 ft/s (60 mph) when the driver
the nearest hundredth ft∕s 2) after 10 sec? applies the brakes to avoid hitting a child. After Zseconds,
27. An object moves along a straight line so that after Z the car is s(f) = 88Z - 8 ∕ 2 feet from the point where the
minutes, its distance from its starting point (in meters) is brakes were first applied. How long does it take for the car
to come to a stop, and how far does it travel before
χz) = 10Z + F ^ - Γ stopping?
37. It is estimated that t years from now, the circulation of a a. At what rate was the G D P changing in 1992?
local newspaper will be b. At what percentage rate was the G D P changing in
1992?
C(∕) = 1OOZ2 + 400/ + 5,000
43. It is projected that x months from now, the population of
a. Find an expression for the rate at which the circulation a certain town will be P{x) = 2x + 4x 3z2 + 5,000.
will be changing with respect to time / years from now. a. At what rate will the population be changing with
b. At what rate will the circulation be changing with respect to time 9 months from now?
respect to time 5 years from now? b. At what percentage rate will the population be chang
c. By how much will the circulation actually change ing with respect to time 9 months from now?
during the 6th year? 44. Assume that your starting salary is $30,000 and you get a
38. An efficiency study of the morning shift at a certain raise of $3,000 each year.
factory indicates that an average worker who arrives on a. Express the percentage rate of change of your salary as
the job at 8:00 A . M . will have assembled a function of time and draw the graph.
b. At what percentage rate will your salary be increasing
∕( x ) = —j x 3 + ¼ 2 + 50x
after one year?
units x hours later. c. What will happen to the percentage rate of change of
a. Find a formula for the rate at which the worker will be your salary in the long run?
assembling the units after x hours. 45. The gross domestic product (GDP) o f a certain country is
b. At what rate will the worker be assembling units at growing at a constant rate. In 1990 the G D P was 125
9:00 A.M.? billion dollars, and in 1992 it was 155 billion dollars. At
what percentage rate will the G D P be growing in 1995?
c. How many units will the worker actually assemble
between 9:00 A.M. and 10:00 A.M.? 46. If y is a linear function of x , what will happen to the
percentage rate of change of y with respect to x as x
39. An environmental study o f a suburban community sug
increases without bound?
gests that t years from now, the average level of carbon
monoxide in the air will be q{t) = 0.05∕2 + 0.1/ + 3.4
parts per million.
a. At what rate will the carbon monoxide level be chang Computational Window
ing with respect to time one year from now? ———-------------------------------------------------------
47. A disease is spreading in such a way that after /weeks
b. By how much will the carbon monoxide level change in
the first year? for 0 ≤ / ≤ 6, it has affected
c. By how much will the carbon monoxide level change N(t) = 5 - ∕ 2( ∕ - 6)
over the next (second) year?
hundred people. Health officials declare that this i
40. According to Newton's law o f universal gravitation, if an
disease will reach epidemic proportions when the j
object o f mass M is separated by a distance r from a
percentage rate of increase o f N(f) at the start o f a ∣
second object of mass m, then the two objects are attract
particular week is at least 30% per week. The epidem
ed to one another by a force that acts along the line
ic designation level is dropped when the percentage j
joining them and has magnitude
rate falls below this level.
c, Gm M a. Find the percentage rate of change of Λr(∕) at time i
rz t.
b. Between what weeks is the disease at the epidemic
where G is a positive constant. Show the rate of change of
level?
F with respect to r is inversely proportional to r 3.
41. The population of a bacterial colony is approximately
P(t) = P o + 61/ + 3 ∕ 2 thousand / hours after observation
begins, where P o is the initial population. Find the rate at 48. An object attached to a hel
which the colony is growing after 5 hours. ical spring is pulled down
from its equilibrium posi
∕∕T =f i x i- the percentage rate o f change o f y with respect to x tion and then released, as
is defined by the expression shown in Figure 2.18.
Use this definition in Problems 42-47.

42. The gross domestic product (GDP) of a country was
g(∕) = ∕ 2 + 5/ + 106 billion dollars / years after 1990. Figure 2.18 Helical spring
Suppose that t seconds later, its displacement (in centime and the values of g for various heavenly bodies. There
ters measured in relation to the equilibrium position) is fore, to deduce his whereabouts, he throws a rock directly
given by upward (from ground level) and notes that it reaches a
maximum height of 37.5 ft and hits the ground 5 sec after
s(t) = 7 cos t leaving his hand. Where is he? Note: You will need to
a. Find the velocity and acceleration of the object at know that g is 32 ft∕s 2 on earth, 5.5 ft∕s 2 on the moon, 12
time t. ft∕s 2 on Mars, and 28 ft∕s 2 on Venus.
b. Find the length of time required for one complete θ 51. Find the rate of change of the volume of a cube with
oscillation. This is called the period of the motion. respect to the length of one of its edges. How is this rate
related to the surface area of the cube?
c. What is the distance between the highest point reached
by the object and the lowest point? Half of this 52. Show that the rate of change of the volume of a sphere
distance is called the amplitude of the motion. with respect to its radius is equal to its surface area.
49. Two cars leave a town at the same time and travel at 53. Van der Waal’s equation states that a gas that occupies a
constant speeds along straight roads that meet at an angle volume V at temperature T (Kelvin) exerts pressure P,
of 60° in the town. If one car travels twice as fast as the where
other and the distance between them increases at the rate
(P + - ^ ) ( Γ - B ) = kT
of 45 mi∕hr, how fast is the slower car traveling?
50. SPY P ROBLEM Our friend the spy, who escaped from the
diamond smugglers in Problem 39 of Chapter 1 Supple and A, B, and k are physical constants. Find the rate of
mentary Problems, is on a secret mission in space. An change of pressure with respect to volume, assuming fixed
encounter with an enemy agent leaves him with a mild temperature.
concussion that causes him to forget where he is. Fortu 54. According to Debye’s formula in physical chemistry, the
nately, he remembers the formula for the height of a orientation polarization P of a gas satisfies
projectile,
p
⅛(0 = - ∣ gt 2 + V + s0
where μ, k, and N are constants and T is the temperature
of the gas. Find the rate of change of P with respect to T.
2 . 5 THE CH AIN RULE
IN TH IS SECTIO N Introduction to the chain rule, differentiation of

composite functions, generalized derivative formulas, justification of the chain
rule
In many practical situations, a quantity is given as a function of one variable,
which, in turn, can be thought of as a function of a second variable. In such
cases, we shall find that the rate of change of the quantity with respect to the
second variable is equal to a product of rates.
■ INTRODUCTION TO TH E CHAIN RULE
Suppose it is known that the carbon monoxide pollution in the air is changing at
the rate of 0.02 ppm (parts per million) for each person in a town whose
population is growing at the rate of 1,000 people per year. To find the rate at which
the level of pollution is increasing with respect to time, we form the product
(0.02 ppm ∕person)( 1,000 people/year) = 20 ppm/year
In this example, the level of pollution L is a function of the population p, which is
itself a function of time t. Thus, L is a function of t, and:
ΓRATE OF CHANGE OF Z,l _ ΓRATE OF CHANGE OF L∖ ΓRATE OF CHANGE OF p l
L WITH RESPECT TO t J L WITH RESPECT TO p J L WITH RESPECT TO t J
Sec. 2.5 The Chain Rule 121
Expressing each of these rates in terms of an appropriate derivative in Leibniz

form, we obtain the following equation:
dL = dLdp
dt dp dt
These observations anticipate the following important theorem.
THEOREM 2.9 Chain rule
Ify = ∕(w) is a differentiable function of u and u = g(x) is a differentiable function

of x, then y = ∕[g(x)] is a differentiable function of x and
dy = dydu
dx du dx
∖ What this says: The derivative of y with respect to x is the derivative of y

with respect to u times the derivative of u with respect to x. This is one of the
most important derivative rules in calculus, and you must make sure you
understand it thoroughly. We have noted that is not a fraction, but in
Section 2.8 we will define dy and dx. Thus, one way to remember the chain rule
is to pretend that the derivatives and are quotients and “cancel” du,
reducing the expression on the right-hand side of the equation to the
expression on the left-hand side. Remember that this is only a mnemonic
device, and no actual reducing of fractions occurs.
Proof: A proof of the chain rule is included in Appendix B. We give a

justification at the end of this section. ■
EXAM PLE 1 The chain rule
Find if y = u 3 - 3u2 + 1 and u = x 2 + 2.
Solution Because = 3w2 - 6ιι and = 2x
it follows from the chain rule that

dy dy du -. . . _ . _ _ .. _
d x = d u d x = ^ , u ~ 6M)(2X) = (2x)(3w)(w - 2 ) = 6xu(u - 2)
Notice that this derivative is expressed in terms of the variables x and u. To
express dy/dx in terms of x alone, we substitute u = x 2 + 2 as follows:
= 6XM(M - 2) = 6Λ^(Λ' 2 + 2)[(x2 + 2) - 2] = 6x(x2 + 2)(x2)
= 6x3(x2 + 2) t≡m
■ DIFFERENTIATION OF COMPOSITE FUNCTIONS
The chain rule is actually a rule for differentiating composite functions. In

particular, if y = flu) and u = u(x), then y is the composite function y = ( ∕o w)(x)
= ∕ [ M(X )], and the chain rule can be rewritten as follows:
T H E O R E M 2.9a The chain rule (alternate form)
If u is differentiable at x and f is differentiable at u(x), then the composite

function f o u is differentiable at x and
( / ° u)'(x) = f'[u(x)]u∖x) or ⅛ f[u(x)] = ⅛ > ) ⅛
■ What this says: In Section 1.4 when we introduced composite functions,

we talked about the “ inner” and “ outer” functions. Using this terminology,
the chain rule says that the derivative of the composite function ∕[u (x)] is
equal to the derivative of the inner function u times the derivative of the outer
function f evaluated at the inner function. You might wish to review Section
1.4 where this idea was introduced.
EXAM PLE 2 The chain rule applied to a power
Differentiate y = (3x4 - l x + 5)3.

Solution y' = [3~∣(3x4 - 7x + 5)1 1 l(3x4 - 7x + 5)'
This step is usually done mentally. Note that we are thinking o f

u(x) = 3x 4 - 7x + 5. What you write down is:
y' = 3(3x4 - 7x + 5)2(12x3 - 7) —
You could, with a lot of work, have found the derivative in Example 2 without
using the chain rule, either by expanding the polynomial or by using the product
rule. The answer would be the same but would involve much more algebra. In
order to compare these methods, however, consider the following problem with a
simpler function.
EXAM PLE 3 Comparison of differentiation with and without the chain rule
Differentiate y = (3x + 2)2.
a. by expansion b. by the product rule c. by the chain rule.
Solution a. y = (3x + 2)2 = 9x 2 + 12x + 4

y' = 18x + 12 = 6(3x + 2)
b. y = (3x + 2)2 = (3x + 2)(3x + 2)
y' = (3x + 2)(3x + 2)' + (3x + 2)'(3x + 2)
= (3x + 2)(3) + (3)(3x + 2) = 6(3x + 2)
c. y = (3x + 2)2
y ' = 2(3x + 2)(3) = 6(3x + 2) —
It is worthwhile to interpret the chain rule using slopes. Let

u = g(x) = x 3 - 3; if x - 2, then u = g(2) = 5
y = f(μ) = V u ; if u = 5, then y = /(5) = V5
We are interested in the slope of the line tangent to y = ∕[g (x )] at x = 2. The
graphs of these functions, along with the slope of their tangent lines, are shown in
Figure 2.19. The slope of the composite function y = ∕[g (x )] at the point x = 2 is
the product of the slopes of the functions f and g∙.
dy = dy ■ du
dx^ du dx
Slope of y = ∕[g (x )] Slope of y = f(u) Slope of u = g(x)
= ∣ ( x 3 - 3 ) - 1/2 (3x2)
‘ Slope o f / " Slope""ofg
The chain rule allows us to find derivatives that would otherwise be very difficult
to handle.
EXAM PLE 4 Differentiation with quotient rule inside the chain rule
Differentiate g(x) = x
1 - 3x -
( x V z4
Solution Write g(x) = I ∣ _ gχ I ∙
, 1∕ χ ∖-3M∕ V
g (x) = 4 ( 1 1 g γ 1 1 j —3v ) The inside function is u = ∣ 2 gχ.
and the outside function is u l,4 .
3z4Γ(l
= 4I f< ∏x ¾ )V I---------3(%Γ)(~
1) W
- x--(----3--)Jl Q„ uotient rule
= 1 ∕1 - 3 X Y Z4 Γ 1 ^∣ = 1
4∖ x ) [ ( l - 3 x ) 2J 4x 3z4(l - 3x)5z4
■ GENERALIZED DERIVATIVE FORMULAS
The chain rule can be used to derive the generalizations of the power rule and the
c. Graph of y = ∕⅛ (x)]
Slope at the point (2, √5) product of rules for differentiating the trigonometric functions, as summarized in the
slopes of g and/: 12 = following box:
Figure 2.19 Graphical representa dL 1u,n — nr∣

ιu∣
n-l du
tion of chain rule dx dx
d ■ _ du d du
d x s ι n u - cosu d χ d xcosu ^ s ιn " dx
Generalized Derivative
Formulas i ta n M = sec2 ^ 2 du
⅛cotw = " csc u Tx
sec u = sec u tan u ⅛ CSCM = - C S C M COtM

sf ⅛v ln
¾
K⅝½⅛⅛⅞⅛w⅛⅛⅛⅛S5V⅛∕⅛J ¾L‰=⅛, "⅞⅛'⅛⅛8⅛J⅛⅛⅛Λ ⅛,>⅛x⅛⅛⅛S⅛M⅛∙⅛⅛ ⅛⅛⅛⅞⅛⅛w( w⅛ ⅛'s⅛⅛r_r⅛⅛¾⅞wvX ⅛⅞w⅛⅛⅛j<⅛⅛⅛⅜⅛x⅛⅛⅛⅛⅛s¾'⅛'l⅞w⅛⅛⅛⅛iS v . S K w s ⅛⅛su⅛i<l¾V⅛x⅛3⅛s⅛κ3β∞>wz1⅛'Λ⅛⅛-⅞⅞⅛i⅛¾,'⅞s⅛⅛⅛¾**⅛, xs⅛5⅛f⅛⅛*1Λ1XXX ,⅛
The chain rule is so important that you need to see several examples, with a
few variations.
EXAM PLE 5 Chain rule with a trigonometric function
Differentiate ∕'(x) = sin(3x 2 + 5x - 7).

Solution Think of this as ∕ ( M) = sinw, where u = 3x 2 + 5x - 7, and apply the
chain rule:
∕'( x ) = cos(3x 2 + 5x - 7) ■(3x2 + 5x - 7)'
= (6x + 5) cos(3x 2 + 5x - 7) E ®
EXAM PLE 6 Chain rule with other rules

/o ∖66
Differentiate g(x) = cosx 2 + 51 + 4 1 .
* - ⅛ c∞ x + 5 ⅛ ( 3 * - ' + 4 )6
Solution
= - s i n x 2 ⅛ (x 2) + 5[β(3x" 1 + 4)5 ⅛ (3x~1 + 4)]
= ( - s in x 2)(2x) + 30(3x 1 + 4)5( - 3 x 2)
= - 2 x sinx 2 - 90x - 2 (3x~1 + 4)5
E X A M P L E 7 Power of a trigonometric function using the chain rule
Differentiate^ = cos4 (3x + 1)2 .
Solution = 4 cos3 (3x + 1)2 ^ c o s ( 3 x + 1)2 Power rule for cosine function
= 4 cos3 (3x + 1)2 ∙ [ - sin(3x + 1)2] ∙ + 1)2

Derivative of cosine function
= -4 cos3 (3x + 2)2 sin(3x + 1)2 ∙ 2(3x + 1)(3)
Derivative of (3x + 1)2
= -24(3x + l)cos 3 (3 x + l) 2 sin(3 x+ 1)2
E X A M P L E 8 Differentiation with chain rule inside quotient rule
Differentiate p(x) = . 1t a n J⅜ .
rv ’ (1 - 4x)5
β l t . dp ( l-4 x ) ⅛ a n 7 x -ta n 7 x ∣μ i-4 x ) 3

s°lut,0n s =----- [ ( l - 4 x ) ψ --------------------
= (1 - 4x)5(sec2 7x) (7x) - (tan 7x)[5(l - 4x)4 (1 - 4x)]

(1 - 4x) l0
(1 - 4x)5(sec2 7x)(7) - (tan7x)(5)(l - 4x)4(-4 )
(1 - 4 x ) l0
= (1 - 4x)4 [7(l - 4x)sec27x + 20 tan 7x]

(1 - 4x) 10
_ 7(1 - 4x)sec27x + 20 tan 7x
^^ (1 - 4x)6
EXAM PLE 9 Chain rule with a radical
Differentiate ∕z(x) = V s e c x 3.
Solution Write h(x) = (secx 3) ιz2.
= ∣(secx 3)- ι z 2 ^ (secx 3) = ∣(secx 3)- ιz 2 (secx 3 tan x 3) (x 3)
= j(secx 3)~ιz2(sccx 3 tan x 3)(3x2) = ∣ x 2V s e c x 3 tan x 3
E X A M P L E 10 Finding a horizontal tangent
Find the x-coordinate of each point on the curve

y = (x + l) 3(2x + 3)2
where the tangent is horizontal.
Solution Because horizontal tangents have zero slope, we need to solve the
equation ∕'( x ) = 0. First use the product rule and the extended power rule to
find the derivative.
∕'( x ) = (x + 1)3 ⅛ (2x + 3)2 + (2x + 3)2 ⅛ (x + 1)3
= (x + l) 3(2)(2x + 3)1 ⅛ (2x + 3) + (2x + 3)2(3)(x + 1)2 (x + 1)
= (x + l) 3(2)(2x + 3)(2) + (2x + 3)2(3)(x + 1)2(1)

= (x + l) 2(2x + 3)[4(x + 1) + 3(2x + 3)]
= ( x + l) 2(2x + 3)(10x+ 13)
From the final factored form, we see that ∕'( x ) = 0 when x = - 1 , - j , or - ∣ g ,

and these are the x-coordinates of all the points on the graph where horizontal
tangents occur. B M
It is worth taking a moment to discuss graphing the function in Example 10 to

verify the answer graphically. If you input Y = ( X + l)^3(2X+3)^2 and graph with
a standard scale, the graph is not satisfactory.
This might be a good place to experiment with

the (ZOOM) feature of your calculator. What scale is V ιB <X +1 )Λ 3<2X+3)Λ 2
necessary to obtain a graph similar to the follow
ing? X∩in=~10 V n in = -1 0
X n a x= lθ Vnax=10
X s c l= l VS G 1=1
,√1 B<X+1 )Λ 3C2X+3)Λ 2

EXAMPLE 11 An application of the chain rule

An environmental study of a certain suburban community suggests that the
average daily level of carbon monoxide in the air is described by the formula
C(p) = Vθ.5p2 + 17
parts per million when the population is p thousand. It is estimated that t
years from now, the population of the community will be
p(f) = 3.1 + 0.1t2
thousand. At what rate will the carbon monoxide level be changing with
respect to time 3 years from now?
Solution f - φ ⅛ = ⅛ 0 ∙⅛2 + 17)"' 2<
0 ∙5(⅛>)][°x2 ')i
When t = 3, p(t) = 3.1 + O.l(3)2 = 4, so

⅛ I ,- J ^ ⅛ 0 '5 ' 42 + 1 7 ) ' l ' 2<0 ∙5 <2 ■4))][0.1(2 ∙ 3)] - 0.24
The carbon monoxide level will be changing at the rate of 0.24 parts per
million. B M
■ JUSTIFICATION OF THE CHAIN RULE
To get a better feel for why the chain rule is true, suppose x is changed by a small
amount ∆x This will cause u to change by an amount ∆w, which, in turn, will
cause y to change by an amount ∆y. I f ∆u is not zero, we can write
∆y _ ∆y ΔM
ΔX ^^ ΔM ΔX
By letting ∆x → 0, we force ΔM to approach zero as well, because
ΔM = (⅛^)∆x so lim ∆w = (Φ )(0) = 0
∖ΔΛ ∕ ΔΛ∙→O ∖ d xf ’
It follows that
Historical Note rlim -δτ>-, =
A calculus book written by the Δx→0 ∆x (⅛ s) lim
ΔΛ →0 Δ X
famous mathematician G. H.
Hardy (1877-1947) contained or, equivalently,
essentially this “incorrect” dy _ dy du
proof rather than the one given dx du dx
in Appendix B. It is even more Unfortunately, there is a flaw in this “proof’ of the chain rule. At the beginning we
remarkable that the error was
assumed that ΔM ≠ 0. However, it is theoretically possible for a small change in x
not noticed until the fourth edi
tion. to produce no change in u so that ΔM = 0. This is the case that we consider in the
proof in Appendix B.
PROBLEM SET 2 .5
0 1. ■ What Does This Say? What is the chain rule? In Problems 3-8, use the chain rule to compute the derivative
2. ■ What Does This Say? When do you need to use the dy/dx and write your answer in terms o f x only.
chain rule? 3. y = u2 + 1; u = 3x - 2 4. y = 2 M2 - u + 5; u = 1 - x 2
M÷ 3 .
5 ∙T = ⅞ 2 -9 6∙ y = 2u + 5’ u = 5x- 3 44. The graphs of u = g(x) and y = /(w) are shown in Figure
u~ 1 - 2x
2.20.
7. y = cos M; U = x2 + 7 8. y = u tan u; u = 3x + -
Differentiate the given function in Problems 9-32 with respect

to the given variable o f the function.
9. a. g{u) = ιι5 b. M(X ) = 3x - 1
c. ∕( x ) = (3x - 1)5
10. a. g(ιι) = ιii b. w(x) = x 2 + 1
c. ∕( x ) = (x 2 + 1)3
11. a. g(ιi) = κ ' 5 b. w(x) = 3x 2 + 5x - 7 b. y = ∕(u )
2
c. ∕( x ) = (3x + 5x - 7) 15
Figure 2.20 Find slope of y = ∕[g (x )]
12. a. g(w) = M7 b. M(X ) = 5 - 8x - 12x2
c. ∕( x ) = (5 - 8 x - 12x2)7 a. Find the approximate value o f y at x = 2. What is the
13. a. ∕( x ) = sin(2x) slope of the tangent line at that point?
b. ,y(x) = sin (2 cosx) b. Find the approximate value o f y at x = 5. What is the
slope of the tangent line at that point?
14. a. f(θ) = sin (sin 0) b. /(0) = sin (cos 0)
c. Find the slope o f y = ∕[g ( x ) ] at x = 2.
15. s(θ) = sin (40 + 2) 16. c(0) = cos (5 - 30)
45. Assume that a snowball melts in such a way that its radius
17. ∕( x ) = x 2ta n x 2 18. h(x) = x 2(2x - 5)3
decreases at a constant rate (that is, the radius is a linear
19. p(x) = sin x 2cosx 2 20. ∕( x ) = csc 2 (Vx) function o f time). Suppose it begins with a radius of 10
21. ∕( x ) = (2x 2 + l) 4(x 2 - 2)5 22. ∕( x ) = (x 3 + l) 5(2x 3 - 1)6 cm and takes 2 hours to disappear. (Assume the snowball
23. f(t) = (1 - t2)sint 2 24. ∕( x ) = sin 2 (V x + 3) is a sphere, and recall that V = ∣ττr3, S = 4 w 2 .)
a. What is the rate of change of its volume after 1 hour?
25. / ( / ) = ( ∕ 2 - l) 3cos(3t + 2) “ ■∕ ω - √
b. At what rate is its surface area changing after 1 hour?
27∙ ^ - ∖ S ≡ 46. It is estimated that t years from now, the population of a
certain suburban community will be
28. a. ∕( x ) = V s in x 2 b. ∕( x ) = sin2 (Vx)
29. a. ∕( x ) = J — . b. ∕( x ) = λp z />«>- 2o- r π
’ V cosx2 ycos-χ
thousand persons. An environmental study indicates that
30. ∕( x ) = yx + ∖[ x 31. ∕( x ) = yX 2 + 2√ x the average daily level of carbon monoxide in the air will
32. ∕( x ) = sin (sin (sin x)) be
£(/?) = 0 .5√ p 2 + p + 58
For each function in Problems 33-38, fin d an equation for the
tangent to the graph at the prescribed point. ppm when the population is p thousand. Find the rate at
33. ∕( x ) = V x 2 + 5 at (2, 3) which the level of carbon monoxide will be changing with
34. ∕( x ) = (5x + 4)3 at ( -1 , -1 ) respect to time two years from now.
35. ∕( x ) = x 2(x - 1)2 where X = ∣ 47. At a certain factory, the total cost of manufacturing q
units during the daily production run is
36. ∕( x ) = sin(3x - π) where x = y C(q) = 0.2<∕2 + q + 900
37. ∕( x ) = x 2tan
(4 - 3x) where x = 0
dollars. From experience, it has been determined that
∕ s _ r 2V approximately
38. ∕( x ) = I βχ _ ∣ ) where x = 0
q(f) = ∕ 2 + 100/
Find the x-coordinate o f each point in Problems 39-42 where
the graph o f the given function has a horizontal tangent line. units are manufactured during the first t hours of a
39. ∕( x ) = x V l - 3x 40. g(x) = x 2(2x + 3)2 production run. Compute the rate at which the total
manufacturing cost is changing with respect to time one
41. <7(x) = 42. ∕( x ) = (2x2 - 7)3 hour after production begins.
θ 43. Let ∕( x ) = (x + 3)2(x - 2)2 . 48. An importer of Brazilian coffee estimates that local
consumers will buy approximately
a. Compute∕'( x ) and determine all values o f x for which
∕'( x ) = 0. 4,374
D(p) =
b. Compute f ' '(x) and find all solutions of f ' '(x) = 0. p-
pounds of the coffee per week when the price is p dollars 54. To form a simple pendulum, a weight is attached to a rod
per pound. It is estimated that t weeks from now, the price which is then suspended by one end in such a way that it
of Brazilian coffee will be can swing freely in a vertical plane. Let θ be the angular
displacement of the rod from the vertical as shown in
p(0 = 0.02∕2 + 0.1t + 6 Figure 2.21.
dollars per pound. At what rate will the weekly demand
for the coffee be changing with respect to time 10 weeks
from now? Will the demand be increasing or decreasing?
49. When electric blenders are sold for p dollars apiece, local
consumers will buy
blenders per month. It is estimated that t months from

now, the price of the blenders will be Figure 2.21 Problem 54
p(Z) = 0.04z3z2 + 15 It can be shown that as long as the maximum displace

dollars. Compute the rate at which the monthly demand ment ΘM is small, it is reasonable to assume that θ =
for the blenders will be changing with respect to time 25 ΘM sin kt at time t, where k is a constant that depends on
months from now. Will the demand be increasing or the length of the rod. Show that
decreasing?
√ 2 ..
⅛ + k 2θ = 0
50. The equation + ks = 0 is called a differential equation df-
of simple harmonic motion. Let A be any number. Show 55. A lighthouse is located 2 km directly across the sea from a
that the function s(t) = A sin 2z satisfies the equation point O on the shoreline, as shown in Figure 2.22.
⅛⅛ + 4s = 0
dt~
51. Suppose L(x) is a function with the property that
L'(x) = x ~ l . Use the chain rule to find the derivatives of
the following functions.
a. ∕( x ) = L (x 2) b. ∕( x ) =
CM ∙ - L (⅛) '
D ZW - L (T ⅛)
Figure 2.22 Problem 55
When a point source o f light o f luminous intensity K (candles)
shines directly on a point on a surface s meters away, the
A beacon in the lighthouse makes 3 complete revolutions
illuminance on the surface is given by the formula I = Ks~2 .
(that is, 6π radians) each minute, and during part of each
Use this formula to answer the questions in Problems 52-53.
(Note that 1 lux = 1 candle∕m 2.) revolution, the light sweeps across the face of a row of
cliffs lining the shore.
52. Suppose a person carrying a 20-candlepower light walks
toward a wall in such a way that at time t (seconds) the a. Show that t minutes after it passes point O, the beam of
person is s(t) = 28 - t 2 meters from the wall. light is at a point P located s(f) = 2 tan(6τrt) km from
O.
a. How fast is the illuminance on the wall increasing
when the person is 19 m from the wall? b. How fast is the beam of light moving at the time it
b. How fast is the illuminance at a point on the wall
passes a point on the cliff that is 4 km from the
lighthouse?
increasing when the person is 3 m from the wall? Your
answer should be in lux/sec rounded to two decimal θ 56. Using only the formula
places. d ■ du
~ r sin zz = cos u -r
53. A lamp of luminous intensity 40 candles is 10 m above dx dx
the floor of a room and is being lowered at the constant cosx = sin (y —Λ') and
and the identities sinx =
rate of 2 m∕s. At what rate will the illuminance at the
point on the floor directly under the lamp be increasing cos - x), show that
when the lamp is 15 m above the floor? Your answer -γ- cosu = -sinw Φ
should be in lux/sec rounded to the nearest hundredth. dx dx
Sec. 2.6 Implicit Differentiation 129
57. T HINK T ANK P ROBLEM Let g(x) = ∕ [ M(X)], where 60. If = UTA and u(x) = cotx, what is - ¾
w(-3) = 5, w'(-3) = 2,/(5) = 3, a n d ∕'(5 ) = -3 . Find an ax x v ' ’ du
equation for the tangent to the graph of g at the point 61. Suppose that / is twice differentiable for all x (that is,
where x = -3 . ∕ ' '(x) exists for all x). Use the chain rule to find
58. L e t/b e a function for which
⅛ ∕V M 1 and ⅛ Λ Λ r ) ] .
a. If g(x) = ∕( 3 x - 1), what is g'(x)7 62. Show that if a particle moves along a straight line with
displacement s{t) and velocity v(∕), then its acceleration
b. If ⅛(x) = /(£), what is h '(x)7
satisfies
59. L e t/b e a function for which ∕ -(2) = - 3 and <≈(<) - V«) ⅞
∕'( x ) = V x 2 + 5. If
2
g w = x ∕ ( ^ τ ) Use this formula to find Φ in the case where
ds
what is g'(2)7 s(t) = - 2 ∕ 3 + 4 / + t - 3.
2 . 6 IMPLICIT DIFFERENTIATION
IN THIS SECTION General procedure for implicit differentiation, higher

implicit differentiation
In this section we introduce implicitly defined functions and show how such
functions can be differentiated when their derivatives exist, without first
solving for y explicitly.
■ GENERAL PROCEDURE FOR IMPLICIT DIFFERENTIATION
An equation of the form y = f(x) is said to define y explicitly as a function of x. On

the other hand, if an equation involves both x and y, as in x = y 1 + 1, and if y is
thought of as the dependent variable, then we say that y is defined implicitly as a
function of x. Suppose we have an equation that defines y implicitly as a function
of x, and we want to find the derivative dy∕dx. For example, we may be interested
in the slope of a line that is tangent to the graph of the equation at a particular
point. One approach might be to solve the equation for y explicitly and then
differentiate using the techniques we have already developed. For functions that
are defined implicitly, it may not be possible to solve for y. For example, there is
no obvious way to solve for y in equations such as
x 1y + 2 j 3 = 3x + 2y
Moreover, even when we can solve for y explicitly, the resulting formula is often
complicated and unpleasant to differentiate. For example, the equation
x 2y 3 - 6 = 5v3 + x
can be solved for y to yield
The computation of dy/dx for this function in explicit form would be routine but
tedious, involving both the chain rule and the quotient rule.
Fortunately, there is a simple technique based on the chain rule that allows us
to find dy/dx without first solving for y explicitly. This technique is known as
implicit differentiation. It consists of differentiating both sides of the (unsolved)
equation with respect to x and then solving algebraically for dy/dx.
E X A M P L E 1 Slope of a tangent using implicit differentiation
Find the slope of the tangent to the circle x 2 + y 2 = 10 at the point P ( - 1, 3).
Solution We recognize the graph of x 2 + y 2 = 10 as not being the graph of a
function. However, if we look at a small neighborhood around the point
( - 1, 3), as shown in Figure 2.23, we see that this part of the graph does pass the
vertical line test for functions. Thus, the required slope can be found by
evaluating the derivative dy/dx at x = - 1 . Instead of solving for y and finding
the derivative, we take the derivative o f both sides o f the equation'.
x 2 + y 2 = 10 Given equation
Figure 2.23 Graph of x 2 + y 2 = 10
⅛ [x≈ + ∕l - ⅛ ( 1 0 )
showing a neighborhood about
P ( - l,3 )
⅛ ≈ l⅛ > ) -o Derivative o f a sum∖ derivative o f a constant is
zero.
2 x + 2 > 'd
i = 0
Because y is a function o f x , apply the general
ized power rule to fin d the derivative o f jA
dy _2 x dy
= Solve the equation for
dx 2y
The slope of the tangent at P ( - 1, 3) is
√x∣*= - 1 (-1 ,3 )
Here is a general description of the procedure for implicit differentiation:
Procedure for Implicit Suppose an equation defines y implicitly as a differentiable function of x.

Differentiation
To find
Step 1. Differentiate both sides of the equation with respect to x. Remember
that y is really a function of x for part of the curve and use the chain rule
when differentiating terms containing y.
Step 2. Solve the differentiated equation algebraically for
E X A M P L E 2 Implicit differentiation
Find if x 2y + 2y3 = 3x + 2y.
Solution The process is to differentiate both sides of the given equation with
respect to x. To help you remember that y is a function of x , we will replace y
with the symbol y(x).
x 2y(x) + 2[p(x)]3 = 3x + 2p(x)
Differentiate both sides of this equation term by term with respect to x:

⅛ { χ 2 TW + 2[y(x)]3} = ⅛ 3 x + 2y(x)]
⅛ U W ) 1 + 2 ⅛ M X) ] ≈ - 3 ⅛ X + 2 ⅛ ^ )
^ ⅜ > 'W t j W ⅜ 1 ÷ 2 { 3 ⅛ W ] ⅛ w } - 3 + 2 ⅛ ) √ Λ-)

j
Product rule Generalized power rule

χ2 i c -v ( x ) + 2χy ^ + 6 [>O)J2 ⅛ y( χ ) = 3 + 2 ⅛ y( χ )
Now replace y(x) by y and y(x) by and rewrite the equation:
j ≡ +2x-l, + 6^⅛ = 3 + 2 ⅛dx

Finally, solve this equation for
.2 ⅛L + f i l ,2 dy _ dy
d x + ■ dx 2? dx = 3 - 2xy
(x2 + 6y2 - 2) = 3 - 2xy
dy _ 3 - 2xy
dx ~ x 2 + 6y2 - 2
Notice that the formula for dy!dx contains both the independent variable x
and the dependent variable y. This is usual when derivatives are computed
implicitly. β≡
In Example 2 it was suggested that you temporarily replace y by y(x), so you

would not forget to use the chain rule when first learning implicit differentiation.
In the following example, we eliminate this unnecessary step and differentiate the
given equation directly. Just keep in mind that y is really a function of x and
remember to use the chain rule when it is appropriate.
EXAM PLE 3 Implicit differentiation; simplified notation
Find if y is a differentiable function of x that satisfies
sin(x 2 + y) = y 2(3x + 1)
Solution There is no obvious way to solve the given equation explicitly for y.
Differentiate implicitly to obtain
^ [ s i n ( x 2 + y)] = ^ [ y 2(3x+ 1)]
cos(x 2 + y) ⅛ x 2 + y) = y 2 ^ ( 3 x + 1) + (3x + 1) ^ y 2
Chain rule Product rule
cos(x 2 + y) ∣
2x + = y 2(3) + (3x + l)^2y
Finally, solve for
2x cos(x 2 + y) + ∞ s(x 2 + y) = 3y2 + 2y(3x + 1)
[cos(x 2 + y) - 2y(3x + 1)] = 3y2 - 2x cos(x 2 + y)
dy _ 3y2 - 2x cos (x 2 + y)
dx cos(x 2 + y) - 2y(3x + 1) B H
EXAM PLE 4 Slope of a line tangent to a given curve
Find the slope of a line tangent to the circle x 2 + y 2 = 5x + 4y at the point

P(5, 4).
Solution The slope of a curve y = f(x ) is ⅛ which we find implicitly.
x 2 + y 2 = 5x + 4y
⅛ * 2 + Λ - ⅛ ( 5 * + ⅛')
dy dy
2y dx 4A dx 5 2x
( 2 ,- 4 ) g - 5 - 2 x
Note that the expression is un

dy = 5 -2 x
defined at y = 2. This makes dx 2y - 4
sense when you see that the tan
gent is vertical there. Look at At (5, 4), the slope of the tangent line is
the graph and see if you should
dy∖ = 5 - 2x∣ = 5 -2 (5 ) = -5
exclude any other values. 4Zx∣(5,4) 2y —41(5,4) 2 ( 4 ) - 4 4
■ HIGHER IMPLICIT DIFFERENTIATION
The use of implicit differentiation to find higher derivatives is illustrated in the

following example.
EXAM PLE 5 Second derivative by implicit differentiation
Find -yj ζ if x 2 + y 2 = 10.

dx~
Solution In Example 1 we found (implicitly) that = - 2 ∙ Thus
d*y= ± ( - χ ∖ = y ⅛ -⅛ -< ~ x ⅛ y = - JZ + X ⅜
dx 1 dx∖ y ) y2 y2
Quotient rule
Note that the expression for the second derivative contains the first derivative
dy∕dx. To simplify the answer, substitute the algebraic expression previously
found for dy∕dx∖
<r-y ~r + x Tx ~ ,' + -⅛ r)
„ , dy x
Substitute -j- - -----
dx 2 y2 y2 dx y
= -y 2 - x2
T3
_ - ( x 2 + y 2)
y3
— 35- Substitute x 2 + y 2 = 10.

y
Implicit differentiation is a valuable theoretical tool. For example, in Section

2.2 we proved the power rule for the case in which the exponent is an integer. We
are now able to prove this result for rational exponents by using implicit
differentiation.
THEOREM 2.10 Power rule for rational exponents
If n = ~ is a rational number, then = ∏xn 1∙
Proof: lξy = χ n ~ χP⅛ is differentiable at x, then y q = χt( pl<b = χ p , Differenti

ate both sides of this equation with respect to x:
w q~ 'T x = Pχ p ~γ Power rule for integer exponents
dy _ p x p 1 c 1 for
f dy
Solve
dx ~ qy q~l
px p ~l
Substitute y = x p,q .
q (χ P , ^ - '
= P χ p- ∖ -p+ ptq
— P y-~^ + P∕Q
q
- nx n ~l Substitute n = ~.
q
0 It is important to realize that implicit differentiation is a technique for
finding dy/dx that is valid only if y is a differentiable function of x, and
careless application of the technique can lead to errors. For example, there is
clearly no real-valued function y that satisfies the equation x 2 + y 2 = -1 ,
yet formal application of implicit differentiation yields the “derivative”
dy!dx = -x ∕y . To be able to evaluate this “derivative,” we must find some
values for which x 2 + y 2 = -1 . Because no such values exist, the derivative
does not exist. 0
PROBLEM SET 2 .6
θ Find dy/dx by implicit di ,entiation in Problems 1-14.
1. x 2 + y 2 = 25 2. x2 + y = x3 + y 3
3. xy = 25 4. xy(2x + 3y) = 2
5. x 2 + 3xy + y 2 = 15 6. x 3 + y3 = x + y
7. (x + y) 3 + 3y = 3 8. x 3 + xy + y 3 = x
9. 1 + - = 1 10. (2x + 3y)2 = 10
y x
11. sin(x + y) = x - y 12. tan - = y
y
13. cosxy = 1 - x 2 14. (x 2 + 3y2) 5 = 2xy
In Problems 15-18, find in two ways:
a. By implicit differentiation o f the equation 26. Cissoid of Dioceles:
b. Bv differentiating an explicit formula for y y 2(6 - x) = x 3 at (3, 3)
15. x 2 + y 3 = 1 2 16. xy + 2y = x 2
17. x +- =5 18. x y - x = y + 2
y
In Problems 19-22, find an equation o f the tangent to the
graph o f each equation al the prescribed point.
19. x 2 + y 2 = 1 3 a t( - 2 , 3)
20. x 3 + y 3 = y + 21 at (3, -2 )
21. sin(x - y) = xy at (0, τr)
22. sec* = V x + 1 at (3, π)
27. Find an equation of the normal line to the curve
Find the slope o f the tangent to the graph at the points indicated x 2 + 2xy = y 3 at (1, -1 ).
in Problems 23-26. 28. Find an equation of the normal line to the curve
23. Bifolium: (x 2 + y 2) 2 = 4x 2y at (1, 1) x 2Vy - 2 = y 2 - 3x - 1 at (1, 2).
Use implicit differentiation to find the second derivative y " o f
the functions given in Problems 29-32.
29. l x + 5y 2 = 1 30. x 2 + 2y3 = 4
2 2
31. 4x - 3y = 9 32. x 2 + 5xy - y 2 = 1
..2 2
33. Let ¾
2
+ v÷z = 1, where a and b are nonzero constants.
a b~
Find
34. Let (a — b)ιι3 - (α + b)v2 = c, where a, b, and c are con

24. Lemniscate of Bernoulli: stants. Find
(x 2 + y 2) 2 = y ( x 2 - y 2) at (2, 1)
θ 35. Show that the tangent at the point (a, b) on the curve
whose equation is 2x 2 + 3xy + y 2 = —2 is horizontal if
4α + 3⅛ = 0. Find two such points on the curve.
36. Find two points on the curve whose equation is
x 2 — 3xy + 2y 2 = - 2 where the tangent is vertical.
37. Let g be a differentiable function of x that satisfies
g(x) < 0 and x 2 + g 2(x~) = 10 for all x.
a. Use implicit differentiation to show that
dg _ - χ
dx g(x)
b. Show that 42. The tangent to the curve

g(x) = - V 1 0 - x2 χ
2∕3 + y
2∕3 = 8
satisfies the given requirements. Then use the chain at the point (8, 8) and the coordinate axes form a triangle,
rule to verify that ^∣o as shown in Figure 2.26. What is the area of this triangle?
dx ~ g(x)
V, ■
38. Find the equation of the tangent and the normal line to
the curve ∙> . j 3 -i .
x + y = 2Axy
at the point (A, A), where A is a constant.
39. T H IN K . T A N K P R O B L E M
a. If x 2 + y 2 = 6v - 10 and ⅞ exists, show that
dx
∖∖ i 11ι ∣ 1
x ∖ ( 8 . 8)
x
dx 3 - y
b. Show that there are no real numbers x , y that satisfy
the equation
X 2 + y 2 = 6y - 10
What can you conclude from the result found in part a

c.
43. T H IN K T A N K P R O B L E MFind two differentiable func
in light of this observation in part b? tions f that satisfy the equation
40. Find all points on the lemniscate x - l> ) ]2 = 9
(x 2 + y 2)2 = 4(x 2 - y 2) Give the explicit form of each function, and sketch its
graph.
where the tangent is horizontal. (See Figure 2.24.)
© 44. Show that the tangent to the ellipse
Figure 2.24 Lemniscate: v∣ ⅛ = 1
(x 2 + ∕) 2 = 4(x 2 - y 2) 2"
a2 b2
1-- at the point (x0 , y 0) is
-A
—( — ∣κv ∖ι k1-
-i / x
⅞*
a2
W
b2
= 1
-1-- (See Figure 2.27.)
-2--
41. Find all points on the cardioid

(x 2 + y 2)3z2 = V x 2 + y 2 + x
- ( — ∣------ 1
— ÷ →
where the tangent is vertical. (See Figure 2.25.) ∖

X .
-1
-1 - -
1 √
Figure 2.25 Cardioid: y ..

(x 2 + JI2)372 = --------
V x2 + y2 + x 7^ ∣∖^
( J
r2 y2
Figure 2.27 Ellipse: ^ + ^ = 1
45. Find an equation of the tangent to the hyperbola

x i „ Z ! = 1
a2 b2
at the point (x0, y0). (See Figure 2.28.)
Figure 2.29 The angle θ from C 1 to C2 at P
48. If θ is the acute angle from curve C 1 to curve C, at P and

the tangent lines to C l and C, at P, have slopes m χ and m 2,
respectively, show that
46. Use implicit differentiation to find the second derivative

y " , where y is a differentiable function of x that satisfies 49. Find the angle from the circle
ax~ + by 2 =c x2 + y 2 = 1
(a, b, and c are constants). to the circle
47. Show that the sum of the x-intercept and the ^-intercept x 2 + ( y - 1)2 = 1
of any tangent to the curve at each of the two points of intersection.
V x+ Vy = C 50. JOURNAL PROBLEM: The Pi Mu Epsilon Journal* by Bruce
W. King. When a professor asked the calculus class to
is equal to C 2 . find the derivative of y 2 with respect to x 2 for the function
y = x 2 - x, one student found
The angle between curves C l and C2 at the point o f intersection dy y
-
P is defined as the angle θ ≤ j between the tangent dx x
lines. Specifically, the angle from C 1 to C2 is the angle from the
tangent line to C 1 at P, to the tangent line to C2 at P, as shown and obtained the correct answer. Help the professor to
in Figure 2.29. Use this information for Problems 48-49. enlighten the student about taking derivatives.
2.7 RELATED RATES
IN THIS SECTION Related rates and applications

In many practical applications of implicit differentiation, the derivatives
represent rates of change (often with respect to time). In a typical situation, two
or more quantities that vary with time may be related to each other, and
implicit differentiation allows us to obtain corresponding relationships among
their respective rates without first finding explicit formulas for the quantities as
functions of time. Techniques for solving such problems are discussed in this
section.
When working a related rate problem you must distinguish between the
general situation and the specific situation. The general situation comprises
properties that are true at every instant of time, whereas the specific situation
Volume 7, 1981, p. 346.
Sec. 2.7 Related Rates 137
refers to those properties that are guaranteed to be true only at the particular
instant of time that the problem investigates. Here is an example.
E X A M P L E 1 An application involving related rates
A spherical balloon is being filled with a gas in such a way that when the radius
is 2 ft, the radius is increasing at the rate of ∣ ft/min. How fast is the volume
changing at this time?
Solution
The general situation: Let V denote the volume and r the radius, both of
which are functions of time t (minutes). Since the container is a sphere, its
volume is given by
L = ≤ π r3
Differentiating both sides implicitly with respect to time t yields
dt d t ∖3 rr
= 4ττr2
dr
Do not forget to use the chain rule
dt
because r is also a function o f
time.
The specific situation: Our goal is to find at the time when r = 2 and
dr _ ι
dt ~ 6∙
⅛ = 4-7τ(2)2 f⅛) = ⅛ ≈ 8.37758041
at x ' ∖o∕ 3
This means that the volume of the container is increasing with respect to time
at the rate of about 8.38 ft 3∕m in. ■■■■
Although each related rate problem has its own “ personality,” many can be
handled by the following summary:
Procedure for Solving Related The General Situation

Rate Problems Step 1. Draw a figure, i f appropriate, and assign variables to the quantities that
vary. Because rates involve variables and not constants, be careful not to
label a quantity with a number unless it never changes in the problem.
Step 2. Find a formula or equation that relates the variables. Eliminate
unnecessary variables; some of these “ extra” variables may be constants,
but others may be eliminated because of given relationships among the
variables.
Step 3. Differentiate the equations. Often this involves differentiating implic
itly with respect to time.
The Specific Situation
Step 4. Substitute specific numerical values and solve algebraically for any
required rate. List the known quantities; list as unknown the quantity you
wish to find. Substitute all values into the formula. The only remaining
variable should be the unknown, which may be a variable or a rate. Solve for
the unknown.
EXAM PLE 2 Related rates
B A person 6 ft tall is walking away from a streetlight 20 ft high at the rate of 7

ft∕s. At what rate is the length of the person’s shadow increasing?
Solution
The general situation: Step 1: Let x denote the length (in feet) of the
person’s shadow and y, the distance between the person and the streetlight, as
shown in Figure 2.30. Let t denote the time (in seconds).
Step 2: Because ΔABC and ΔDEC are similar, we have
χ +y _ x
20 6
Step 3: Write this equation as x + y = 2^x, or y = ⅛x, and differentiate both
sides with respect to t.
A y D x
dy = ld x
Figure 2.30 A person walking dt 3 dt
away from a street lamp.
The specific situation: Step 4: List the known quantities. We know that
= 7. Our goal is to find Substitute and then solve for the unknown
value:
dy = 7 d x
dt 3 dt
7 ∕7v
7= 3 Substitute.
3= Multiply both sides by η.
The length of the person’s shadow is increasing at the rate of 3 ft∕s.
A bag is tied to the top of a 5-m ladder resting against a vertical wall. Suppose
the ladder begins sliding down the wall in such a way that the foot of the
T ladder is moving away from the wall. How fast is the bag descending at the
instant the foot of the ladder is 4 m from the wall and the foot is moving away
at the rate of 2 m/s?
Solution
The general situation: Let x and y be the distances from the base of the wall
x
to the foot and top of the ladder, respectively, as shown in Figure 2.31.
O B
Notice that ΔTOB is a right triangle, so a relevant formula is the
Figure 2.31 A ladder sliding down Pythagorean theorem:
a wall x~ + / = 25
Differentiate both sides of this equation with respect to f.
2x ⅜' + 2y ⅛ = 0
dt j dt
The specific situation: At the particular instant in question, x = 4 and

y = V25 - 4 2 = 3. We also know that = 2, and the goal is to find at this
instant. We have
2(4)(2) + 2 (3 )^ = 0
dy_ _8
dt ~ 3
This tells us that at the instant in question, the bag is descending (because
dy/dt is negative) at the rate of f ≈ 2.7 m∕s.
EXAMPLE 4 A physical application involving related rates
When air expands adiabatically (that is, with no change in heat), the pressure
P and the volume Jz satisfy the relationship
PF 1-4 = c
where C is a constant. At a certain instant, the pressure is 20 lb∕in.2 and the
volume is 280 in 3. If the volume is decreasing at the rate of 5 in.3∕s at this
instant, what is the rate of change of the pressure?
Solution
The general situation: The required equation was given, so we begin by

differentiating both sides with respect to t. Remember, because C is a
constant, its derivative with respect to t is zero.
1.4 ∕V 0∙4 + K1∙4 ^ = 0 Product rule
The specific situation: At the instant in question, P = 20, V = 280, and

dV/dt = - 5 (negative because the volume is decreasing). The goal is to find
dP∕dt. First substitute.
(20)( 1.4)(28O)o∙4( - 5) + (280)1∙4 ⅛y = 0
z√P
Now, solve for
√p 5(20)( l,4)(28O)o∙4
dt (280)1∙4
¢ = 0.5
dt
Thus, at the instant in question, the pressure is increasing (because its
derivative is positive) at the rate of 0.5 lb∕in.2 per second.
A tank filled with water is in the shape of an inverted cone 20 ft high with a
circular base (on top) whose radius is 5 ft. Water is running out of the bottom
of the tank at the constant rate of 2 ft3∕min. How fast is the water level falling
when the water is 8 ft deep?
Solution
The general situation: Consider a conical tank with height 20 ft and circular
base of radius 5 ft, as shown in Figure 2.32. Suppose that the water level is h ft
and that the radius of the surface of the water is r.
Figure 2.32 A conical water tank
Let V denote the volume of water in the tank after t minutes. We know that
Jz = ⅛πr2h
(See Student Mathematics Handbook for appropriate formulas.) Once again,

we use similar triangles (see Figure 2.32) to write = o r r = 4' We
substitute this into the formula to obtain

v = ⅛r (4) h = ⅛7γ^ 3
0 A common error in solving and then differentiate both sides of this equation with respect to t.
related-rate problems is to sub
d V = iL z,2 dh
stitute numerical values too dt 16 dt
soon or, equivalently, to use re
lationships that apply only at a The specific situation: Begin with the known quantities: we know that
particular moment in time. This dV/dt = - 2 (negative, because the volume is decreasing). The goal is to find
explains why we have separated dh∕dt. At the particular instant in question, h = 8 and we substitute to find:
related-rate problems into two
distinct parts. Be careful to -)_ 7τ ∕o∖7 dh
work with general relationships
among the variables and substi - 1 _ dh
tute specific numerical values 2π dt
only after you have found gener
al rate relationships by differen At the instant when the water is 8 ft deep, the water level is falling (because
tiation. 0 dh/dt is negative) at a rate o f ⅛ ≈ 0.16 ft/min ≈ 1.9 in./min.
Every day, a Tight from Los Angeles to New York flies directly over my home
at a constant altitude of 4 mi. If I assume that the plane is flying at a constant
speed of 400 m i∕hr, at what rate is the angle of elevation of my line of sight
changing with respect to time when the horizontal distance between the
approaching plane and my location is exactly 3 mi?
Solution
The general situation: Let x denote the horizontal distance between the
plane and the observer, as shown in Figure 2.33a. The height of the observer is
insignificant when compared to the height of the plane.
A general relationship is
cot θ = ∣x
Differentiate both sides of this equation with respect to t to obtain

- c
c ss cc
2n≤⅛= ∖ dx
' dt 4 dt
a. Observer of an approaching plane The specific situation:. Solving for we obtain
de_ ^ θ Tt
dt
dx
At the instant when x = 3, we are given that = -4 0 0 (negative because the
distance is decreasing), and from Figure 2.33b we see sin θ = 4/5. Thus, the
angle of elevation is changing at the rate of
⅛ = -⅛(f)2( - 4 °°) = 64 rad/hr

b. Triangle for computing sin θ when x = 3
Figure 2.33 Angle of elevation The angle of elevation is increasing at the rate of 64 radians per hour or,
problem equivalently,
(64 rad/hr)^-360 deg' f 1 hr
27τ rad ∖36OO sec/ ≈ 1.02 deg/sec
PROBLEM SET 2 .7
θ Find the indicated rate in Problems 1-9, given the other 11. If a spring is stretched at the constant rate of a ∣ in .∕s, how
information. Assume x > 0 and y > 0. fast is the force ∕∙'(x) changing when x = 3 in.?
1. Find ⅞ where x 2 + y 2 = 25 and ⅛ = 4, when x = 3. 12. A particle moves along the parabolic path given by
dt z dt y 2 = 4x in such a way that when it is at the point
dχ 1 , dv (1, —2), its horizontal velocity (in the direction o f the x-
2. Find ⅛7 where χ - + y - = 25 and -∏ = 2, when x = 4.
dt ■ dt axis) is 3 ft∕s. What is its vertical velocity (in the direction
Find ⅞∙ where 5x2 - y = 100 and ⅛ = 10, when x = 10. o f the y-axis) at this instant?
3.
dt -, dt 13. A particle moves along the elliptical path given by
Find ⅛ where 4x 2 - y = 100 and ⅞ = - 6 , when x = 1. 4x 2 + y 2 = 4 in such a way that when it is at the point
4.
dt ■ dt (V 3 ∕2, 1) its x-coordinate is increasing at the rate of 5
5. Find ⅛ where y = 2 V x - 9 and -τ- = 5, when x = 9.
units per second. How fast is the y-coordinate changing at
dt ■ dt that instant?
6. Find where y = 5 V x + 9 and ⅛ = 2, when x = 7. 14. A rock is dropped into a lake and an expanding circular
dt ■ dt ripple results. When the radius of the ripple is 8 in., the
dv dx radius is increasing at a rate of 3 in.∕s. At what rate is the
7. Find ∙÷ where xy = 10 and t ⅛- = —2, when x = 5. area enclosed by the ripple changing at this time?
dt ■ dt
dy 15. A pebble dropped into a pond causes a circular ripple.
8. Find where 5xy = 10 and ⅛ = —2, when x = Find the rate at which the radius of the ripple is changing
dt
at a time when the radius is one foot and the area enclosed
dx -> ι dv
9. Find -÷ where x 2 + xy — v~ = 11 and - r = 5, when x = 4 by the ripple is increasing at the rate of 4 ft 2∕s .
dt ∙ dt
and p > 0. 16. An environmental study of a certain community indicates
that there will be Q lp) = p 2 + 2>p + 1,200 units of a
In physics, Hooke’s law says that when a spring is stretched x harmful pollutant in the air when the population is p
units beyond its natural length, the elastic force F(x) exerted by thousand. The population is currently 30,000 and is
the spring is F(x) = - k x , where k is a constant that depends on increasing at a rate of 2,000 per year. At what rate is the
the spring. Assume k = 12 in Problems 10 and 11. level of air pollution increasing?
10. If a spring is stretched at the constant rate of a ∣ in .∕s, how 17. It is estimated that the annual advertising revenue re
fast is the force F(x) changing when x = 2 in.? ceived by a newspaper will be R(x) = 0.5x 2 + 3x + 160
thousand dollars when its circulation is x thousand. The rate of 80 ft 3∕m in. How fast is the water level rising when
circulation of the paper is currently 10,000 and is increas the water is 12 ft deep?
ing at a rate of 2,000 per year. At what rate will the annual 30. Suppose that in Problem 29, water is also flowing out the
advertising revenue be increasing with respect to time 2 bottom of the tank. At what rate should the water be
years from now? allowed to flow out so that the water level will be rising at
18. Hospital officials estimate that approximately N(p) = a rate of only 0.05 ft/min when the water is 12 ft deep?
p 1 + 5p + 900 people will seek treatment in the emergen Give your answer to the nearest hundredth of a cubic foot
cy room each year if the population of the community is p per minute.
thousand. The population is currently 20,000 and is 31. A person is standing at the end of a pier 12 ft above the
growing at the rate of 1,200 per year. At what rate is the water and is pulling in a rope attached to a rowboat at the
number of people seeking emergency room treatment water line at the rate of 6 ft of rope per minute, as shown
increasing? in Figure 2.34. How fast is the boat moving in the water
19. Boyle’s law states that when gas is compressed at constant when it is 16 ft from the pier?
temperature, the pressure P of a given sample satisfies the
equation PV = C, where V is the volume of the sample
and C is a constant. Suppose that at a certain time the
volume is 30 in. 3, the pressure is 90 lb∕in. 2, and the
volume is increasing at the rate of 10 in.3∕s. How fast is
the pressure changing at this instant? Is it increasing or
decreasing?
θ 20. ■ What Does This Say? What do we mean by a related-
rate problem?
21. ■ What Does This Say? Outline a procedure for solving
related-rate problems.
∣
C ∖ 22. The volume of a spherical balloon is increasing at a Figure 2.34 Problem 31
⅛∏j constant rate of 3 in.3∕s. At what rate is the radius of the 32. One end of a rope is fastened to a boat and the other end
balloon increasing when the radius is 2 in.? is wound around a windlass located on a dock at a point 4
23. Can you stretch your imagination to consider a piece of m above the level of the boat. If the boat is drifting away
ice in the shape of a sphere? Suppose such a piece of ice is from the dock at the rate of 2 m∕min, how fast is the rope
melting at the rate of 5 in. 3∕min. How fast is the radius unwinding at the instant when the length of the rope is 5
changing at the instant when the radius is 4 in.? How fast m?
is the surface area of the sphere changing at the same 33. A ball is dropped from a height of 160 ft. A light is located
instant? Assume the ice retains its spherical shape at all at the same level, 10 ft away from the initial position of
times. the ball. The height of the ball after t seconds is
(fT 24. The surface area of a sphere is decreasing at the constant //(/) = - 1 6 ∕ 2 + 160 ft. How fast is the ball’s shadow
⅛~J rate of 3ττ cm 2∕s. At what rate is the volume of the sphere moving along the ground one second after the ball is
decreasing at the instant its radius is 2 cm? dropped?
25. A certain medical procedure requires that a spherical 34. A point P(x, y) moves along the curve xy = 100 in the
balloon be inserted into the stomach and then be inflated. first quadrant of the xv-plane. Let A be the area of
If the radius of the balloon is increasing at the rate of 0.3 the triangle with vertices P, Q(l, 0), and A(0, 2). If the x-
cm∕min, how fast is the volume changing when the radius coordinate of P is increasing at the rate of 3 units per
is 4 cm? second, how fast is the area changing at the time the
26. A person 6 ft tall walks away from a street light at the rate ^-coordinate of P is four times the y-coordinate? Is the
of 5 ft∕s. If the light is 18 ft above ground level, how fast is area increasing or decreasing at this time?
the person’s shadow lengthening? 35. A person 6 ft tall stands 10 ft from point P directly
27. A ladder 13 ft long rests against a vertical wall and is beneath a lantern hanging 30 ft above the ground, as
sliding down the wall at the rate of 3 ft/s at the instant the shown in Figure 2.35.
foot of the ladder is 5 ft from the base of the wall. At this
instant, how fast is the foot of the ladder moving away
from the wall?
28. A car traveling north at 40 mi/hr and a truck traveling east
at 30 mi/hr leave an intersection at the same time. At what
rate will the distance between them be changing 3 hr
later?
29. A very peculiar water tank is built in the shape of a cone
40 ft high with a circular base of radius 20 ft with the
point on top. Water is flowing into the tank at a constant Figure 2.35 Problem 35
The lantern starts to fall, thus causing the person’s shad At what rate is the angle of elevation of the person’s line of
ow to lengthen. Given that the lantern falls 16t2 ft in t sec, sight changing with respect to time when the person is 9 ft
how fast will the shadow be lengthening when t = 1? from the base of the light?
36. A race official is watching a race car approach the finish 41. A revolving searchlight in a lighthouse 2 mi off shore is
line at the rate of 200 km∕hr. Suppose the official is sitting following a beachcomber along the shore, as shown in
at the finish line, 20 m from the point where the car will Figure 2.38.
cross, and let (9be the angle between the finish line and the
official’s line of sight to the car, as shown as Figure 2.36.
At what rate is θ changing when the car crosses the finish
line?
When the beachcomber is 1 mi from the point on the

shore that is closest to the lighthouse, the searchlight is
turning at the rate of 0.25 rev∕hr. How fast is the beach
comber walking at that moment? Hint: Note that 0.25
rev/hr is the same as ? rad∕hr.
37. At noon on a certain day, a truck is 250 mi due east of a 42. A water trough is 2 ft deep and 10 ft long and has a
car. The truck is traveling west at a constant speed of 25 trapezoidal cross section with base lengths 2 ft and 5 ft, as
mi∕hr, whereas the car is traveling north at 50 mi∕hr. shown in Figure 2.39.
a. At what rate is the distance between them changing at
time fl
b. At what time is the distance between the car and the
truck neither increasing nor decreasing?
c. What is the minimal distance between the car and the
truck? Hint: This distance must occur at the time
found in part b. Do you see why?
38. A weather balloon is rising vertically at the rate of 10 ft∕s. Figure 2.39 Problem 42
An observer is standing on the ground 300 ft horizontally
from the point where the balloon was released. At what
a. Find a relationship between the volume of water in the
rate is the distance between the observer and the balloon
trough at any given time and the depth of the water at
changing when the balloon is 400 ft high?
that time.
39. An observer watches a plane approach at a speed of 500
b. If the water enters the trough at the rate of 10 ft 3∕min,
mi/hr and an altitude of 3 mi. At what rate is the angle of
how fast is the water level rising (to the nearest ⅛
elevation of the observer’s line of sight changing with
in./min) when the trough is 1 ft deep?
respect to time when the horizontal distance between the
plane and the observer is 4 mi? Give your answer in 43. At noon, a ship sails due north from a point P at 8 knots
radians per hour or in degrees per minute. (nautical miles per hour). Another ship, sailing at 12
knots, leaves the same point 1 h later on a course 60o east
40. A person 6 ft tall is watching a streetlight 18 ft high while
of north. How fast is the distance between the ships
walking toward it at a speed of 5 ft∕s, as shown in Figure
increasing at 2 P .M .? At 5 P .M .? Hint: Use the law of
2.37.
cosines.
44. A swimming pool is 60 ft long and 25 ft wide. Its depth
varies uniformly from 3 ft at the shallow end to 15 ft at
the deep end, as shown in Figure 2.40.
Figure 2.37 Problem 40 Figure 2.40 Problem 44

Suppose the pool is being filled with water at the rate of a. Show that t minutes after it passes point S, the beam of
800 ft 3∕min. At what rate is the depth of water increasing light is at a point P located s(Z) = 2 tan 6τrt km from S.
at the deep end when it is 5 ft deep at that end? b. How fast is the beam of light moving at the time it
45. A water bucket has the shape of the frustum of a cone with passes a point on the cliff located 4 km from the
height 1 ft and upper and lower radii of 1 ft and 9 in., lighthouse?
respectively, as shown in Figure 2.41. 47. A car is traveling at the rate of 40 ft/s along a straight,
level road that parallels the seashore. A rock with a family
of seals is located 50 yd offshore.
a. Find an expression for the rate of change of the angle θ
between the road and the driver’s line of sight to the
seals.
b. As the distance x in Figure 2.42 approaches 0, what
happens to dθ∕dtl
If water is leaking from the bottom of the bucket at the

rate of 8 in. 3∕m in, at what rate is the water level falling
when the depth of water in the bucket is 6 in.? Hint: The
volume of the frustum of a cone with height h and base
radii r and R is
V = y - (R 2 + rR + r2)
θ 46. A lighthouse is located 2 km directly across the sea from a
point 5 on the shoreline. A beacon in the lighthouse c. Suppose the car is traveling at v ft∕sec. Now what
makes 3 complete revolutions (6ττ radians) each minute, happens to dθ∕dt as x → 0? What effect does this have
and during part of each revolution, the light sweeps across on a passenger looking at the seals if the car is traveling
the face of a row of cliffs lining the shore. at a high rate of speed?
2 .8 DIFFERENTIALS AND TANGENT LINE APPROXIMATIONS
IN THIS SECTION Tangent line approximation, the differential, error

propagation, marginal analysis in economics
The material of this section was historically motivated by the desire to make
certain approximations by using the derivative. However, because fast, accurate
calculators and computers are now commonplace and because computations
can usually be made without the use of calculus approximations, we now study
approximations to provide insight into the nature of the derivative. We also
introduce a new symbol, dx, called the differential, which is used to estimate
changes in functional values. A special case of the approximation procedure is
the concept of marginal analysis, which plays an important role in business and
economics.
■ TANGENT LINE APPROXIMATION
Consider the graph of a function that is differentiable at a point x = a. (Three

examples are shown in Figure 2.43.) Next, look at successive enlargements, and
notice that the curve eventually looks linear. In fact, if we draw that tangent line at
some point (α,∕( α)) on the curve, we see that for a small interval on the x-axis, the
ι,-values along the tangent line give good approximations to the y-values on the
curve. It is for this reason that we call the equation of the tangent line to a function
f at a point a linearization of the function at a point x = a.
Sec. 2.8 Differentials and Tangent Line Approximations 145
a. y = x 3 - 2x + 5
b. y = "√15 + x
-0 .2 0.2 0.4 0.6 0.8 1 1.2 1.4 -0 .9 0.9 0.95 I 1.05 1.1 1.15 1.2
Figure 2.43 Comparison of graphs and tangent lines for functions passing through (1,4)
In order to find the equation of the tangent line at the point (<2,∕(<2)), we note
that the slope is / '( « ) and u s e the point-slope form of the equation of a line:
y ~ f(β ) = ∕'( α ) ( x - a)
y = f(a ) + f'(a )(x - a)
Figure 2.44 Equation of the tan
gent line is We can use this line as an approxim ation of f as long as the line remains close to
y = f(μ) + ∕'(α )(x - a). the graph of f as shown in Figure 2.44.
Linearization of a Function If y = ∕( x ) is differentiable at x = a, then

L(x) = ∕( α ) + ∕'( α ) ( x - a)
is the linearization o f / a t a.
EXAM PLE 1 Finding a linearization of a given function
Find the linearization of the function defined by f( x ) = x 3 - 2x + 5 at a = 1

(see Figure 2.43a).
Solution The derivative o f∕ i s ∕ ' ( x ) = 3x 2 - 2. We also find
/(1 ) = P - 2 ( 1 ) + 5 = 4 ∕'( 1 ) = 3(1)2 - 2 = 1
so that the linearization o f/ a t a = 1 is

L(x) = /(1) + ∕'( l ) ( x - 1) = 4 + l(x - 1) = x + 3
We might ask about the accuracy of the linearization in Example 1. The

variable ∆x is called the increment of x. The accuracy of the approximation
∕( x ) ≈ L(x) is found by considering
x 3 - 2x + 5 ≈ x + 3
If ∆x = 1, then a + ∆x = 1 + 1 = 2. Also,
∕( α + ∆x) = /(2) = 2 3 - 2(2) + 5 = 9 and
L(« + ∆x) = L(2) = 2 + 3 = 5
These (along with other) values are summarized in the following table:
a ∆x a + ∆x /(« + ∆x) L(a + ∆x) ∣Z,(α + ∆x) - ∕( α + ∆x)∣

1 1 2 9 5 4
1 0.1 1.1 4.131 4.1 0.031
1 0.01 1.01 4.010301 4.01 0.000301
Let us consider a geometric interpretation. Let / b e a differentiable function at

x = Λ'o . You can see from Figure 2.45a that as ∆x becomes small, Q becomes a
good approximation of P. To find an expression for the y-value of O, look at Figure
2.45b. You can see that the y-value of Q is equal to∕( x 0) + ∆y. Because the slope of
the tangent at R is the derivative o f/ at x 0 , we have
∕'W -E E
∆p = ∕'( ⅞ ) ∆x
Thus, the y-value of Q is
∕(⅞ ) + ∕'( ⅜ ) ∆x
a. Tangent line t o /a t R
Because Q is a good approximation of P if ∆x is small,
∕(⅞ ) + ∆x) ≈ ∕( x 0) + ∕ ' ( x 0) ∆ x

V - J ∖_______ _______ /
y-value of P y-value of Q
If we wish to find the change in the function / that corresponds to a change in

x, we use the following notation:
∆x = increment of x (that is, change in x)
Δ ∕ = increment of / (that is, change in / caused by a change ∆x in x)
Then,
Δ ∕= ∕( x 0 + ∆x) - ∕( x 0)
b. Tangent line approximation so that
Figure 2.45 Tangent line approxi
mation A ∕≈ ∕'(% 0) λ x
We call this the incremental approximation formula.
Incremental Approximation I f / i s differentiable at x 0 and ∆x is small, then the corresponding change (in
Formula functional notation) in f satisfies
Δ ∕= ∕( ⅞ + δ x ) ~ ∕ ( x o) ≈ ∕'( ⅞ ) δ ∙x
W hat this says: At a point where the function is differentiable, the change in
the function is approximately the derivative of the function at that point times
the change in the independent variable.
EXAM PLE 2 Incremental approximation
Show that if∕( x ) = sinx, the function ^(x) = ∆f∕∆,x approximates the function
∕'( x ) = cosx for small values of ∆x.
Solution The approxim ation formula Δ ∕ = ∕(x θ + ∆x) - ∕(xθ ) ≈ ∕' ( x o )∆x im
plies that
Δ ∕ = ∕( ⅞ + ∆x) - ∕( x 0 )
∆x ∆x ■' W
Because∕( x ) = s in x ,∕'( x ) = cosx, and
sin(x + ∆x) - sinx
≈ cosx
∆x
Figure 2.46 shows the graphs of f and g for three different choices of ∆x.
Notice as ∆x becomes smaller, it is more difficult to see the difference between
f and g. In fact, for very small ∆x, the graphs are indistinguishable.
■ THE DIFFERENTIAL
We have already observed that writing the derivative of ∕( x ) in the Leibniz

notation d f/d x suggests that the derivative may be incorrectly regarded as a
quotient of “d f" by “ <7x.” It is a tribute to the genius of Leibniz that this erroneous
interpretation of his notation often turns out to make good sense.
To give dx and dy meaning as separate quantities, let x be fixed and define dx
to be an independent variable equal to ∆x, the change in x. That is, define dx,
called the differential of x, to be an independent variable equal to ∆x, the change in
x. Then, i f / i s differentiable at x, we define dy, called the differential of y, by the
formula
dy = ∕'( x ) dx or, equivalently, d f = ∕'( x ) dx
If we relate differentials to Figure 2.47, we see that dx = ∆x and that ∆y is the rise
off that occurs for a change of ∆x, whereas dy is the rise of a tangent relative to the
same change in x. It is important to note that ∆y and ∆x are generally not the same
(see Figure 2.47).
Figure 2.47 Geometrical definition of dx and dy
We now restate the standard rules and formulas for differentiation in terms of
differentials. Remember that a and b are constants, while f and g are functions.
Differential Rules Linearity ride d (af + bg) = a d f + b dg

Product rule d(fg) = f dg + g d f
Quotient rule d (~∖ = -- (g ≠ 0)
∖⅛'/ g^
Power rule d (x n) = nx n ~l dx
Trigonometric rules
<7(sinx) = cosx dx d(cosx) = -s in x dx
√(tanx) = sec2x dx c∕(cotx) = -csc 2x dx
√(secx) = secx tanx dx d(cscx) =-c scx cotx dx
EXAM PLE 3 Differential involving a product and a trigonometric function
Find <7(x2sinx).
Solution <√(x2sinx) = x 2<7(sinx) + sinx <7(x2)
= x 2(cos x dx) + sin x(2x dx)
= (x2cosx + 2x sinx) dx ™
Next, we use differentials to make an approximation. Keep in mind that we

are using a numerical example to help us understand the differential. If all we
wanted to do was to evaluate the expression, we would use a calculator and not
differentials.
EXAM PLE 4 Using a differential to approximate a function
Use differentials to approximate V'8.02 and check by using a calculator.

Solution In order to use differentials, we need to write this problem using

functional notation. I f ∕( x ) = ^^x, then we want to approximate/(8.02).
/(8.02) = /(8 + 0.02)
= /(8) + Δ ∕
We now approximate Δ ∕ by using differentials:
Δ ∕ ≈ d f = ∕'( x ) dx
= ∣ x -2z3 dx
Let x = 8 and dx = 0.02. Then,
d f = ⅛(8)-2z3(0.02) = ≡ ≡ 1 ≈ 0.001666667
Therefore, since Δ ∕ =/(8.02) -/ (8 ), we have

/(8.02) = /(8) + Δ ∕
≈ /(8) + d f
≈ 2 +0.001666667
≈ 2.0016667
3,-----
Our calculator tells us that V8.02 = 2.0016653 to 7 digits, so the error in the
approximation is less than 0.00001. ■■■
■ ERROR PROPAGATION
In the next example, the approximation formula is used to study propagation of

error, which is the term used to describe an error that accumulates from other
errors in an approximation. In particular, in the next example, the derivative is
used to estimate the maximum error in a calculation that is based on figures
obtained by imperfect measurement.
EXAM PLE 5 Propagation of error in a volume measurement
You measure the side of a cube and find it to be 10 cm long. From this you
conclude that the volume of the cube is 103 = 1,000 cm 3. I f your original
measurement of the side is accurate to within 2%, approximately how
accurate is your calculation of volume?
Solution The volume of the cube is F(x) = x 3, where x is the length of a side. If
you take the length of a side to be 10 when it is really 10 + ∆x, your error is
Ax; and your corresponding error when computing the volume will be Δ F,
given by
Δ K = Jz (10 + ∆x) - K(10) ≈ K'(10)Δx
Now, Jz '(x) = 3x 2, so K'(10) = 300. Also, your measurement of the side can
be off by as much as 2%— that is, by as much as 0.02(10) = 0.2 cm in either
direction. Substituting ∆x = ±0.2 in the incremental approximation formula
for Δ V, we get
Δ K ≈ 3(10)2(±0.2) = ±60
K'(10) ∆x
Thus the propagated error in computing the volume is approximately ±60
cm 3. M M
Error Propagation If x 0 represents the measured value of a variable and x 0 + ∆x represents the
exact value, then ∆x is the error in measurement. The difference between
∕( x + ∆x) a n d ∕(x ) is called the propagated error and is defined by
Δ ∕ = f ( x + ∆x) - ∕( x )
The relative error is y " ≈ y .
Relative Error
Δ∕
Percentage Error The percentage error is 10 0 y %.
In Example 5, the approximate propagated error in measuring the volume is ±60,

the approximate relative error is
Δ F ~ ±60 _ ∩ ∩
V ~ ( T θ j 5 - i+ ° ∙ 0 6a
and the approximate percentage error is ±6%
EXAM PLE 6 Relative error and percentage error
The height of a certain right circular cylinder is twice the radius of the base.
The radius is measured to be 17.3 cm, with a maximum measurement error of
0.02 cm. Estimate the corresponding propagated error, the relative error, and
the percentage error when calculating the surface area.
Solution We have (Figure 2.48)
Height
S = 2ττr 2r + πr2 + τrr2 = 6τrr2
Circumference Top Bottom
Lateral side
where r is the radius of the cylinder’s base. Then the propagated error is
Δ S ≈ 5 '( r ) Δ r = 12πr∆r = 12π(17.3)(±0.02) ≈ ±13.0438927
Figure 2.48 Surface area of a right

Thus, the maximum error in the measurement of the surface area is about
circular cylinder 13.04 cm 2 . Is this a large or a small error? The relative error is found by
computing the ratio
= y x y F = 2r^ l ∆r = 2(17.3)^1(±0.02) ≈ ±0.0023121387
This tells us that the maximum error of approximately 13.04 is fairly small
relative to the surface area S. The corresponding percentage error is found by
100∣yr ≈ 100(±0.0023121) ±0.23121
This means that the percentage error is about ±0.23%.
■ MARGINAL ANALYSIS IN ECONOMICS
When the derivative is used in economics to study the rate of change of one
quantity with respect to another, the procedure is called marginal analysis.
Marginal cost is the instantaneous rate of change of production cost, C , with
respect to the number of units produced, x.
Marginal revenue is the instantaneous rate of change of revenue, R , with respect to
the number of units produced, x.
Sec. 2.8 Differentials and Tangent Line Approximations 15 1
In economics, we are often interested in knowing the change in cost or the change
in revenue caused by a unit change in production. That is, if ∆x = 1, then
Change in cost due to a unit change in x:
Δ C = C(x + ∆x) - C(x) = C(x + 1 ) - C(x)
Change in revenue due to a unit change in x:
ΔR = R(x + ∆x) - R(x) = R(x + 1 ) - R(x)
In practice, these functions are approximated using marginal cost and marginal
revenue, respectively. Because ∆x = 1 we have the following approximations:
Δ C = C(x + ∆x) - C(x) ≈ C'(x) ∆x = C'(x)
ΔR = R(x + ∆x) - R(x) ≈ R'(x) ∆x = R'(x)
■ What this says: If C(x) is the total cost of manufacturing x units and Λ(x) is
the total revenue from the sale of x units, then:
The marginal cost, C'(x), approximates the cost of producing the (x + l)st
unit. The marginal revenue, R'(x), approximates the revenue derived from the
sale of the (x + l)st unit.
E X A M P L E 7 Approximation by marginal cost and marginal revenue
A manufacturer estimates that when x units (0 ≤ x ≤ 50) of a particular

commodity are produced, the total cost will be
C(x) = ∣ x 2 + 3x + 98
dollars and that all x items will be sold if the price used by the manufacturer is
∕j(x) = j(75 - x)
dollars per unit.
a. Find the marginal cost and the marginal revenue.
b. Use marginal cost to estimate the cost of producing the 9th unit.
c. What is the actual cost of producing the 9th unit?
d. Use marginal revenue to estimate the revenue derived from producing the
9th unit.
e. What is the actual revenue derived from producing the 9th unit?
Solution
Computational Window O S
a. The marginal cost is
Compare the values on [0, 50]
C'(x) = ∣ x + 3
using a calculator.
To find the marginal revenue, we must first find the revenue function:
n, C, Total Cost/ p, Price
No. Avg Cost per Item R(x) = x∕t(x) = x(∣)(75 —x) = - ∣ x 2 + 25x
0 $98.00 $25.00 Thus the marginal revenue is
5 $116.13; $23.23 $23.33
10 $140.50; $14.05 $21.67
15 $171.13; $11.41 $20.00 R'(x) = - j x + 25
20 $208.00; $10.40 $18.33
25 $251.13; $10.05 $16.67 b. The cost of producing the 9th unit is the change in cost as x increases from
30 $300.50; $10.02 $15.00 8 to 9 and is estimated by
35 $356.13; $10.18 $13.33
40 $418.00; $10.45 $11.67
$486.13; $10.80 $10.00
C'(8) = ∣(8) + 3 = 5
45
50 $560.50; $11.21 $8.33
We estimate the cost of producing the 9th unit to be $5.
c. Actual cost is
C(8) = [∣(9)2 + 3(9) + 9δ] - [∣(8)2 + 3(8) + 9δ]
Δ C = C(9) -
= 5∣ = 5.125
The actual cost of producing the 9th unit (to the nearest cent) is $5.13.
Notice that this is approximated fairly well by the marginal cost in part b.
d. The revenue (to the nearest cent) obtained from the sale of the 9th unit is
approximated by the marginal revenue:
7?'(8) = -y(8) + 25 = y ≈ 19.67
e. The actual revenue (nearest cent) obtained from the sale of the 9th unit is
AT? = 7?(9) - 7?(8) = ^ ≈ 19.33
Remember that the revenue from the sale of the 9th item is not the price of
one item if 9 are sold. Rather, it is the additional revenue the company has
earned by selling the 9th item— that is, the total revenue of 9 items minus
the total revenue of 8 items. ⊂=m
PROBLEM SET 2 .8
θ ∕w Problems 1-14 25. r∕(x co sx ) 26. d lx sin 2x)
a. Use a calculator to approximate the value o f the given 27. r f ⅛ ) , j I x 2secx∖
c
quantity; ∖ 2x /
b. Choose an appropriate value for x and for ∆x so that you can
use the approximation formula 29. d { x V x 2 - 1) 30. d ∖ xr 5J
∖ V x + 4/
f ( x 0 + ∆x) ≈ f ( x 0 ) + f '( x ) ∆x θ 3 1 . ■ What Does This Say? What is a differential?
32. ■ What Does This Say? Discuss error propagation,
Suggested values are given in Problems 1 and 2.
including relative and percentage error.
1. V17; let f ix ) = V x where x 0 = 16, ∆x = 1. 33. Use differentials to approximate the value of
2. λ∕z28ζ let ∕( x ) = W where xθ = 27, ∆.r = 1. (3.01)5 - 2(3.01)3 + 3(3.01)2 —2 and determine the error
3 . 1vz l5 4. ^z26 as compared to the calculator value.
34. Use differentials to approximate the value of
5. - J = 6. 37k
V4.5 VT8 √74,096 +∖λt,096 + 3√4,096
Λ
7. ( l .l ) 4 8. 1.95
9. λ < 9 9 10. 0.99 10° and determine the error as compared to the calculator
value.
11. cos + 0.01) 12. sin -0 .0 1 )
Use calculus to obtain the required estimates in Problems
13. sin ( f - 0.001) 14. co s(j + 0.001) 35-49.
35. You measure the radius of a circle to be 12 cm and use the
In Problems 15-20, use the approximation formula to approxi formula A = πr2 to calculate the area. If your measure
mate ∆ f the change in f that occurs as x varies from a to b. ment of the radius is accurate to within 3%, approximate
15. fix') = x 1 + x - 2, a = 1, b = 1.001
ly how accurate (to the nearest percent) is your calculation
16. fix ) = 3x 2 - x + 2, a = 2, b = 2.01 of the area?
17. fix ) - x l ° + x 5, a = 1, b = 0.99 36. Experiment Suppose a 12-oz can of Coke has a height of
18. fix ) = 3x 7 + x 4 , a = 1, b = 0.999 4.5 in. Determine the radius with an accuracy to within
19. fix ) = V x + x 3,2 , a = 4, b = 4.001 1%. Check your answer by examining a Coke can.
20. fix ) = V x + x 2, a = 8, b = 7.99 37. You measure the radius of a sphere to be 6 in. and use the
formula V —⅜πr3 to calculate the volume. If your mea
Find the differentials indicated in Problems 21-30. surement o f the radius is accurate to within 1%, approxi
21. √(2x 3)' 22. r ∕( 3 - 5 x 2) mately how accurate (to the nearest percent) is your
23. d (2 V x) 24. d l x s + V x 2 + 5) calculation of the volume?
38. It is projected that t years from now the circulation of a 46. A drug is injected into a patient’s bloodstream. The
local newspaper will be concentration of the drug in the bloodstream Zhours after
the drug is injected is given by
C(Z) = 1OOZ2 + 400z + 5,000
Estimate the amount by which the circulation will in C(Z) = -θ∙- 1∙⅜-
v ’ t2 + t + 1
crease during the next 6 months.
39. An environmental study o f a certain community suggests milligrams per cubic centimeter. Estimate the change in
that t years from now, the average level o f carbon monox concentration over the time period from 30 to 35 minutes
ide in the air will be after injection.
2(Z) = 0.05Z2 ÷ 0.1Z + 3.4 47. According to Poiseuille’s law, the speed o f blood flowing
along the central axis of an artery o f radius R is
parts per million (ppm). By approximately how much will .SY/?) = cR 2 , where c is a constant. What percentage error
the carbon monoxide level change during the next 6 (rounded to the nearest percent) will you make in the
months? calculation of S(Λ) from this formula if you make a 1%
40. A manufacturer’s total cost (in dollars) is error in the measurement o f R r!
48. The period o f a pendulum is given by the formula
C(q) = O.l<73 - 0.5q 2 + 500<y + 200
r - 2 '√ l
when the level of production is q units. The current level
o f production is 4 units, and the manufacturer is planning
to decrease this to 3.9 units. Estimate how the total cost where L is the length of the pendulum in feet, g = 32 ft∕s 2
will change as a result. is the acceleration due to gravity, and T is time in
41. At a certain factory, the daily output is seconds. If the pendulum has been heated enough to
increase its length by 0.4%, what is the approximate
β(L) = 60,000L ιz3 percentage change in its period?
units, where L denotes the size of the labor force mea 49. A thin horizontal beam o f alpha particles strikes a thin
sured in worker-hours. Currently 1,000 worker-hours of vertical foil, and the alpha particles that are scattered
labor are used each day. Estimate the effect on output that through an angle θ will travel along a cone o f vertex angle
will be produced if the labor force is cut to 940 worker θ as shown in Figure 2.49.
hours.
Time 1: Stream is focused.
42. The speed of blood flowing along the central axis of a
certain artery is Time 2: Stream hits foil.
S(A) = 1.8 × 105Λ 2
cm ∕s, where R is the radius of the artery measured in cm.
A medical researcher measures the radius o f the artery to
Time 3: Stream is scattered after colliding
be 1.2 × 10"2 cm and makes an error o f 5 × 10^4 cm. with foil and disperses in the shape of a cone.
Estimate the amount by which the calculated value o f the
speed of the blood will differ from the true speed if the
incorrect value of the radius is used in the formula.
43. A soccer ball made o f leather ∣ in. thick has an inner
diameter of 8∣ in. Estimate the volume of its leather shell.
Figure 2.49 Paths o f alpha particles
44. A cubical box is to be constructed from three kinds of
building materials. The material used in the four sides of A vertical screen is placed at a fixed distance from the
the box costs 2Φ∕in.2 , the material in the bottom costs point o f scattering. Physical theory predicts that the
3Φ∕in.2, and the material used for the lid costs 4Φ∕in.2 number N of alpha particles falling on a unit area of the
Estimate the additional total cost of all the building screen is proportional to the reciprocal of sin4 (j). Sup
materials if the length of a side is increased from 20 in. to
pose
21 in.
45. In a healthy person of height x in., the average pulse rate
in beats per minute is approximated by
596 Estimate the change in the number of alpha particles per

P(x) = 2 7≤ 30 ≤ x ≤ 100 unit area of the screen if θ changes from 1 to 1. 1.
Vx
50. Suppose the total cost of manufacturing q units is
Estimate the change in pulse rate that corresponds to a
height change from 59 to 60 in. C(q) = 3q2 + q + 500
dollars.
a.
Use marginal analysis to estimate the cost of manufac 54. At a certain factory, the daily output is
turing the 41st unit. 0(L) = 360T lz3
b. Compute the actual cost of manufacturing the 41 st unit.
units, where L is the size of the labor force measured in
51. A manufacturer’s total cost is worker-hours. Currently, 1,000 worker-hours of labor are
C⅛) = 0.1 q 3 - 5q2 + 500√ + 200 used each day. Use differentials to estimate the effect that
one additional worker-hour will have on the daily output.
dollars, where q is the number of units produced. 55. Suppose that when x units of a certain commodity are
a. Use marginal analysis to estimate the cost of manufac produced, the total cost is C(x) and the total revenue is
turing the 4th unit. R(x). Let ∕ j (x) denote the total profit, and let
b. Compute the actual cost o f manufacturing the 4th unit.
* ) = ≡
52. Suppose the total cost o f producing x units of a particular
commodity is
be the average cost.
C(x) = ∣ x 2 + 4x + 100
a. Show that P'(x) = 0 when marginal revenue equals
and that each unit o f the commodity can be sold for marginal cost.
b. Show that A '(x) = 0 when average cost equals margin
p(x) = ∣(80 - x) al cost.
dollars. 56. The radius R of a spherical ball is measured as 14 in.
a. What is the marginal cost? a. If the maximum error in measuring R is ∣ in., what is
b. What is the price when the marginal cost is $ 10?
the maximum propagated error in computing the
c. Estimate the cost o f producing the 11th unit. volume F of the ball?
d. Find the actual cost of producing the 11th unit. b. With what accuracy must the radius R be measured to
53. Suppose the total cost o f producing x units of a particular guarantee an error o f at most 2 in.3 in the calculated
commodity is volume?
C(x) = j x 2 + 3x + 10 θ 57. Show that if h is sufficiently small, then
and that each unit of the commodity can be sold for a. V l + A is approximately equal to 1 +
p(x) = ⅜(45 - x) b. is approximately equal to 1 — h.

dollars.
58. If h is sufficiently small, find an approximate value for
a. What is the marginal cost? ^∖, A n + h for constant A.
b. What is the price when the marginal cost is $23? 59. Tangent line approximations are useful only if ∆ x is small.
c. Estimate the cost o f producing the 11th unit. Illustrate this fact by trying to approximate V97 by
d. Find the actual cost of producing the 11th unit. regarding 97 as being near 81 (instead of 100).
2.9 THE NEWTON-RAPHSON METHOD FOR APPROXIMATING ROOTS
IN THIS SECTION The Newton-Raphson method, when the

Newton-Raphson method fails
Solving equations of the form ∕( x ) = 0 can sometimes be very difficult, if not
impossible. Often, the most we can hope for is to obtain approximate solutions
by using a computer, calculator, or by following a regular procedure, such as the
bisection method discussed in Section 1.7. Another such approxim ation
procedure is called the Newton-Raphson method, because it was first used by
Newton and then was refined and improved by Joseph Raphson (1648-1715).
This m ethod uses tangent lines to the graph of f to approximate roots of the
equation ∕( x ) = 0.
■ THE NEWTON-RAPHSON METHOD
The basic idea behind the N ew ton-R aphson method is illustrated in Figure 2.50.
In this figure, r is a root of the equation ∕( x ) = 0, x 0 is an approxim ation to r, and
Sec. 2.9 The Newton-Raphson Method for Approximating Roots 155
x l is a better approximation obtained by taking the x-intercept of the line that is

tangent to the graph of f at (x0 , ∕( x 0)).
a. Estimating a root, r, of y = f∖ x)
Figure 2.50 The Newton-Raphson method
THEOREM 2.11 The Newton-Raphson method

To approximate a root of the equation ∕'(x) = 0, start with a preliminary estimate
x 0 and generate a sequence of increasingly accurate approximations x 1, x 2, x,, . . .
using the formula
x x
* f '( x J ≠ 0
ll+ ι = , , ~ f ^ f j
Either this sequence of approximations will approach a limit that is a root of the
equations or else the sequence does not have a limit.
Proof: Rather than present a formal proof, we will present a geometric
description of the procedure to help you understand what is happening. Let xθ be
an initial approximation such that ∕'( x 0) ≠ 0. To find a formula for the improved
approximation x p recall that the slope of the tangent line through (x0 ,∕( x 0)) is the
derivative ∕'( x 0). Therefore (see Figure 2.50b),
Ay _ ∕( x 0) - 0
∕'( ⅞ ) = ∆x x 0 - X1
Slope of the tangent through (x0, ∕( x 0))
or, equivalently (by solving the equation for x 1),
X |= X °- 7 ^ ∕'( ⅞ ) ≠ θ
If this procedure is repeated using x 1 as the initial approximation, an even better

approximation may often be obtained (see Figure 2.50b). This approximation, x2,
is related to x 1 as x 1 was related to x 0 . That is,
∕(⅞ )
X- Λ , ∕ ' ( X 1)
If this process produces a limit, it can be continued until the desired degree of
accuracy is obtained. In general, the nth approximation x n is related to the
(n - l)th by the formula
_ _ ∕⅛ - , )
” ,
"- ∕ '⅛ - , ) ■
Computational Window S ≡ ( EXAM PLE 1 Estimating a root with the Newton-Raphson method
If you have a graphing calcula Approximate a real root of the equation x 3 + x + 1 = 0 on [ - 2 , 2].
tor, you can use a graph and
Solution Let ∕( x ) = x 3 + x + 1. Our goal is to find the root of the equation
the (TRACE) instead of the
Newton-Raphson method. ∕( x ) = 0. The derivative o f/ is ∕'( x ) = 3x2 + 1, and so
For Example 1 define f:
Y1=X^3+X+1 ∕( * ) _ _ x 3 + x + 1 = 2x 3 - 1
∕'( x ) 3x 2 + 1 3x 2 + 1
Thus, for n = 0, 1, 2, 3, . . . ,
x _ λ- M ) _ 2V - 1
"+ l " ∕'(X ,,) 3 x, 2 + l
A convenient choice for the preliminary estimate is x 0 = —1. Then
2⅞3 ~1
0.75 You will need a calculator or a spreadsheet
3xθ2 + 1 to help you with these calculations.
Xmin=-4 Vmin=-4
Xnax=4 Vnax=4 2x 13 - 1
Xscl=l Vscl=l -0.6860465
^^ 3 x , 2 + 1
Move the cursor to the place(s) 2x 3 - 1
where the curve crosses the x- x3 = τ Λ ---- r ≈ -0.6823396
3 x ,2 + 1
axis:
K=-.fiBOHEll Y=.00SE3HH6 In general, we shall stop finding new estimates when successive approximations x n
X=-.765557H Y=-.2153376 and x n+1 are within a desired tolerance of each other. Specifically, if we wish to
This tells us the root is between have the solutions to be within ε (ε > 0) of each other, we compute approxima
-0.68 and -0.77. (Trace val tions until ∣xn+1 - x J < ε is satisfied. Using the Newton-Raphson formula
ues will vary with different cal (Theorem 2.11), we see that this condition is equivalent to
culator models.) By using the
(ZOOM) we can find the root to
ε
any desired degree of accuracy.
Let us take another look at Example 1. The previous Explain what is happening when you carry out the
computational window shows that a calculator will process
give a faster approximation. You might ask, “ If I can (z l)(s fδ )(S T θ ) (ENTER) (2nd) (VARS) (ENTER) (ENTER)
do it on my calculator, why should I learn
Newton-Raphson?” The reason is twofold: (1) Un on a TI-81, as compared with
derstanding the procedure will give you insight into (τ i ) (s fδ ) (STO) (ENTER) (2nd) (VARS) (ENTER)
how to obtain results on a calculator or a computer, (s τδ ) (sτδ) (E N T E R ) ( E N T E R ) (E N T E R )
and (2) as you will see, the Newton-Raphson method
sometimes fails. To illustrate what we mean, consider The first sequence gives one value for the
(from Example 1) finding the roots of x 3 + x + 1 = 0 Newton-Raphson method, and the second reevalu
on [—2, 2]. We calculate ates the old Y 1 as the new x. Can you explain why this
works? Also, if you do not have a TI-81, can you
.. f(×) 2x 3 - 1 duplicate the steps on your calculator?
∕'( x ) 3x 2 + 1
Store this function as Y l ; that is,
Y1=(2X^3 - 1)∕(3X^2+1).
___ ΞΞ_ -.— ........
Now we can describe a step-by-step procedure for applying the

Newton-Raphson method. A flow chart for the method appears in Figure 2.51.
Procedure for Applying the Newton-Raphson Method to Solve the Equation

∕ (χ )= o
1. Choose a number ε > 0 that determines the allowable tolerance for
estimated solutions.
2. Compute ∕'( x ) and choose a number x 0 (with ∕' ( x 0) ≠ 0) “ close” to a
solution o f∕( x ) = 0 as an initial estimate.
3. Compute a new approximation with the formula
4. Repeat step 3 until ∣xπ+1 - x j < ε. The estimate x = x π then has the
required accuracy.
EXAM PLE 2 The Newton-Raphson method with error analysis
Estimate a solution of the equation sinx = Use the Newton-Raphson

method until successive approximations agree "to 3 decimal places.
Figure 2.51 Flowchart for the
Solution Geometrically, we are looking for the x-coordinate of the point where
Newton-Raphson method
the curve y = sinx intersects the liney = ∣x , as shown in Figure 2.52. If we set
∕( x ) = x - 2 sinx, the problem can be restated as “ solve ∕( x ) = 0.” We find
∕'( x ) = 1 - 2 cosx, and
γ _ v _ jf(x
' n ') — γ _ x_H___________
- 2 sin x n
«+> n f'( x , l ) n 1 - 2 cosx π
As a first approximation we choose x 0 = 2. Then,
x 1 = 2 - f — 2_sm2 sb 1 9009956. 1
∣ _ x l∣ ≈ 0.0990044
1 - 2 cos 2 ’ χ1 θ
ι O ∩∩O Q S A 1.9009956 - 2 sin 1.9009956 ~ <- . ,
x 2 - 1.9009956 -------- 1 - 2 cos 1.9009956------- ≈ k18 9s q5 5 1s 1 6a
_ ι 0955114 _ 1.89551 16 —2 sin 1.8955116 ∣ 095494-3.

x 3 - 1.89551 161 1 - 2 cos 1.8955116 ~ 1-8954943,
∣x3 - x 2 ∣ ≈ 0.0000174
The error for x 3 is less than the required three decimal places, so the
approximate solution is x ≈ 1.895.
■ WHEN THE NEWTON-RAPHSON METHOD FAILS

Figure 2.52 The Newton-
Raphson method applied to the We have already observed that for the Newton-Raphson method to work, we
equation sinx = y cannot have f ' ( x n) = 0 for any estimate of x n; and the method may also fail if the
initial estimate x 0 is too far from the desired solution of ∕( x ) = 0, as shown in
Figure 2.53b. Usually, these problems can be overcome by simply making another
0 Do not round yet; work with choice for the initial estimate, xθ. However, there are other more troublesome
all the accuracy you have avail ways for the method to fail. In particular, it can happen that no matter how we
able. choose x 0, the method will produce a sequence of estimates x 1, x 2, x 3, . . . such
that ∣x - x π∣ will get larger instead of smaller (see Figure 2.53b) as n increases.
∣x2 - x 1∣ ≈ 0.0054839 0 This case is illustrated by the following example.
Historical Note
root sought.
Figure 2.53 The Newton-Raphson method (part a) might fail (part b).
MARIA AGNESI (1718-1799)

EXAM PLE 3 Failure of the Newton-Raphson method
One of the more famous wom
en in the history of mathemat Try to find a zero for the function∕( x ) = x 1/3 by using the Newton-Raphson
ics is Maria Agnesi. She was method.
born in Milan, the first of 21
children. Her first publication, Solution According to the Newton-Raphson method,
at age 9, was a Latin discourse x 1/3
defending higher education for Λ x J _ .- xn- 3xn= ~2x,
Γ xT ⅛ =
women. Her most important ∕'⅛ ) ' 3 n
work is a now-classic calculus
We see that x κ+1 is always twice as big as x in absolute value. The graph of
textbook published in 1748.
Maria Agnesi is also remem ∕( x ) = x 1'3 is shown in Figure 2.53b. Notice how the shape of the graph causes
bered for a curve called the the estimates to disperse instead of converging. That is, the difference of
witch of Agnesi. This curve is successive estimates ∣xπ+1 - x j does not decrease as n increases.
defined by the equation
The sequence of estimates x 1, x 2, x 3, . . . produced by the Newton-Raphson
method is said to converge to a root of∕( x ) = 0 if for any choice of ε > 0, we have
∣xκ+1 - x j < ε for n sufficiently large. In a more advanced course, called
numerical analysis, it is shown that the sequence of estimates will converge to a
solution if
∕( x ) ∕" ( x )
The curve was named versoria [ ∕'( * ) ] 2
by Guido Grandi (1672-1742),
but Agnesi confused the word on an open interval containing the solution. Note that the function∕( x ) = x 1/3 fails
versoria with the word versiera, the test because we have
which means “devil’s grand 2/3 5/3
mother” in Latin. Later, when ∕'( x ) = ∣x and ∕" ( x ) = - j x
translated into English, versiera
was rendered as “witch.” The so that (for x ≠ 0)
curve has ever since been called ∕( x ) ∕" ( x )
the “witch of Agnesi.” [∕'( x ) ] 2
which is not less than 1.

PROBLEM SET 2 .9
θ Find estimate x 3 in the Newton-Raphson method for ∕(x ) = 0 5. ∕(x ) = x 4 - 16.2 for x 0 = 2 6. ∕(x ) = x 3 - 17 for x 0 = 2
with the initial estimate x 0 given in Problems 1-10. 7. ∕(x ) = sinxforx 0 = 1 8. ∕(x ) = tanxforx 0 = 0.2
1. ∕(x) = x 3 - 29 for x 0 = 3 2. ∕(x ) = x 3 - 29 for x0 = 3.1 9. ∕(x ) = x 3 + 2x - 1 for xθ = 0
3. ∕(x ) = x - 3 for x 0 = 1
3 4. ∕(x ) = x 5 - 33 for x 0 = 2 10. ∕(x ) = x 5 - x 4 - x 2 + 13 for x0 = 2
© 11. ■ What Does This Say? Discuss the Newton-Raphson 33. L e t∕( x ) = - 2 x 4 + 3x 2 + ⅛l ∙
method. Include a discussion of the shortcomings of this a. Show that the equation ∕( x ) = 0 has at least 2 solu
method. tions. Hint: Use the intermediate value theorem.
Use the Newton-Raphson method to estimate a solution o f the b. Show that the Newton-Raphson method fails if you
equations on the intervals given in Problems 12-19. In each choose xθ = J as the initial estimate. Hint: You should
case, continue the process until two successive approximations obtain x l = - x 0 ; x 2 = x 0 ; . . . .
differ by less than ε = 0.001. 34. The volume o f a spherical segment is given by
12. x 3 - 3x 2 + 3 = 0 on (2, 3)
V = ^ H 2f3R - H )
13. x 3 + 3x - 1 = 0 on (0, 1)
where R is the radius o f the sphere and H is the height of
14. x 3 + 2x — 1 = 0 on (0, 1)
the segment, as shown in Figure 2.54.
15. x 3 + 2x 2 - x + 1 = 0 on ( - 3 , —2)
16. Ψ x + 2 = x - 1 on [2, 4]
17. ^ ∕ x - 3 = x + 1 on [ - 3 , -2 ]
18. cosx = x - 1 on ^0, 19. sinx = x 2 on (0, π)
Show that the Newton-Raphson method fails in Problems
20-23 for the equation f(x) = 0 and the initial estimate xθ.
20. ∕( x ) = 9x 2 + 2x + 3 for x 0 = - ∣ Figure 2.54 Problem 34
21. ∕( x ) = 1 - ∙∣ for x 0 = 2 22. ∕( x ) = x I/6 for any x fl ≠ 0
a. If V = 8 and R = 2, use the Newton-Raphson method
23. ∕( x ) = 9x 4 - 16x3 — 36x2 + 96x —60 for x 0 = ∣ to estimate the corresponding H.
b. A celebrated problem of the Greek geometer Archime
24. Use the Newton-Raphson method to estimate a solution des (ca. 287-212 B.C.) asks where a sphere should be
o f the equation cut in order to divide it into two pieces whose volumes
have a given ratio. Show that if a plane at distance x c
sin (x 3 + 2) = for 0 < x < 2
from the center of a sphere with R = 1 divides the
25. Use the Newton-Raphson method to estimate a solution sphere into two parts, one with volume twice that of
to the equation the other, then 3xc 3 - 9xc 2 - 4 = 0. Use the
Newton-Raphson method to estimate x .
x 3 - 3x = cosx - sinx c. In part b, suppose the plane is located so that it divides
on the interval [θ, j ] . the sphere in the ratio of 1:3. Find an equation for x c ,
and estimate the value of x by the Newton-Raphson
method.
26. Consider the equation cosx =
Γ
a. Sketch the graphs of y = cosx and y = How many Computational Window
times do these graphs intersect?
b. Use the Newton-Raphson method to estimate each In Problems 35-36 we look at the speed at which Newton ,s
method converges. Let en = ∣xexact - x n∣ denote the error
root of the equation cosx = ∣ . after n applications o f Newton’s method. Typically, the
27. First show that there exists at least one solution of the error in each iteration tends to be the square o f the previous
equation x 6 —x 5 + x 3 = 3. Hint'. Recall the intermediate error: that is, e ,, = K e 2 where K is some constant. You
value theorem. Then use the Newton-Raphson method are to investigate this relationship.
to estimate a root of this equation. 35. Do 3 or 4 iterations of Newton’s method toward
finding the positive root o f x 2 - 2 = 0, starting with
28. A wire 10 inches long is to be cut into two pieces, one of
χ = 2. You should see the above convergence pat
which is bent to form a square and the other to form a
tern. What is K I
circle. If the circle is to have twice the area of the square,
36. Take several iterations of Newton’s method in look
where should the wire be cut? Use the Newton-Raphson
ing for the obvious solution of (x — 5)2 = 0, starting
method.
with x 0 = 4. Comment on the rate of convergence.
In Problems 29-30, use the Newton-Raphson method to
Why is this problem not “ typical” in its rate of
estimate at least one root o f the given equation.
convergence?
29. 4x 3 + 3x 2 - 12x + 7 = 0
30. 13x5 + x 4 - 15x3 + 8x 2 + 4x - 7 = 0
© 37. T H IN K T A N K P ROBLEM Let ∕( x ) = - x 4 + x 2 + A for
In Problems 31 and 32, s(t) is the displacement o f a moving constant A. What value of A should be chosen that
object. Use the Newton-Raphson method to estimate a value o f guarantees that if x 0 = ∣ is chosen as the initial estimate,
t where the object is stationary (that is, where s'(t) = 0). the Newton-Raphson method produces x l = —x 0;
31. s(f) = 3 ∕ 4 + 4 ∕ 3 - 30t2 + t + 4 32. s(t) = t 3 - 2t2 - 5t + 3 x 2 = x 0 ; x 3 = —x Q; . . . , as in Problem 33b?
38. Can you solve the equation Vx = 0 using the a. Apply the Newton-Raphson method to∕'(x) = x 2 - N
Newton-Raphson method with the initial estimate to justify this formula.
x 0 = 0.05? Does it make any difference if we choose b. Apply the formula to estimate V 1,265 correct to 5
another initial estimate (other than x 0 = 0)? decimal places.
39. Suppose that when we try to use the Newton-Raphson 41. Use the Newton-Raphson method to derive a formula
method to approximate a solution of ∕(x ) = 0, we find like the one in Problem 40 for the reciprocal function
th a t∕( x j = 0, but ∕'( x n) ≠ 0 for some x n. What does this f(x ) = l∕x. Then apply your formula to find the reciprocal
imply about
⅞ +1, x n+2, ■∙ ■? 1
Explain. 2.355673
40. The ancient Babylonians (ca. 1700 B.C.) approximated correct to 5 decimal places. This method permits you to
∖ rN by applying the formula do a division just by doing a multiplication.
x ,,÷ι = ⅛(⅞ + for «, = 1, 2, . . .
CHAPTER 2 REVIEW
IN THIS REVIEW Proficiency examination; supplementary problems

PROFICIENCY EXAMINATION
Concept Problems Practice Problems

1. What is the slope of a tangent? How does this compare dy
Find ÷ in Problems 15-19.
to the slope of a secant? dx
2. Define the derivative of a function. 15. y = x i + xVx + cos2x 16. y = (V3x + -⅜)
3. What is a normal line to a graph?
17. y = Vsin(3 - x 2) 18. xy + y 3 = 10
4. State the following derivative rules.
19. y = sin (x ° + Vx) + cos (x 10 + Vx)
2 1 2
a. constant rule b. power rule c. constant multiple
dv^
d. sum rule e. difference rule f. linearity rule 20. Find where x 2 + xy - 2y2 = 4.
g. product rule h. quotient rule i. cosine rule
21. Use the definition of derivative to find ^^(x - 3x2).
j. sine rule k. tangent rule I. secant rule
m. cosecant rule n. cotangent rule 22. Find the equation of the tangent to the graph of
5. What is a higher derivative? List some of the different
notations for higher derivatives. y = (x2 + 3x - 2)(7 - 3x) at (1, 8)
6. How do you find the velocity and the acceleration for an
23. Let ∕(x ) = sin2 (2*p). Find equations of the tangent and
object with displacement s(Z)?
the normal line to the graph o f f at x = 1.
7. State the chain rule.
24. A rock tossed into a stream causes a circular ripple of
8. Outline a procedure for implicit differentiation. water whose radius increases at a constant rate of 0.5
9. Outline a procedure for solving related rate problems. ft∕s. How fast is the area contained inside of the ripple
10. State the tangent line approximation formula. changing when the radius is 2 ft?
11. What is meant by marginal analysis? 25. Use the Newton-Raphson method to estimate a solu
12. Define the differential of x and the differential of y for a tion of the equation x 3 - x 2 - x = 13 on [0, 3]. Contin
function y = f(x). Draw a sketch showing ∆x, ∆y, dx, ue the process until two successive approximations
and dy. differ by less than 0.001.
13. Define the terms propagated error, relative error, and
percentage error.
14. What is the Newton-Raphson method?
Chapter 2 Review 161
SUPPLEMENTARY PROBLEMS
Find the derivative in Problems 1-24. 43. Let

dx
1. y = x4
+ 3x 2 - 7x + 5 2. y = x 5 + 3x 3 — 11
1 jts 1
M I
C Γ > 4. _ 1 - cos 3x ∙4 - 2'
⅜
J N1 x + sinx
2 Find ∕'( x ) by using implicit differentiation to differenti
5. 2x - xy + 2y = 5 6. y = xVsin(3 - x 2)
ate [ ∕x ) ] 3 .
7 .y = ( v ¾ + J ) ’ " 44. Find y ' and y " if x 3y 3 + x - y = 1. For y " , leave your
8. y = (x 2 + 3x - 5)7.
answer in terms of x, y, and y ' .
9. y = (x 3 + x ) 1° 10. y = √ x (x 2 + 5) 10 45. Find y ' and y " if x 2 + 4xy —y 2 = 8. Your answer may
11. y = 4vzx (x 3 + 1)5 12. y = (x 3 + 3)5(x 3 - 5) involve x and y, but not y ' .
13. y = (x 4 - l ) 10(2x 4 + 3)7 14. y = Vsin 5x 46. Find the derivative of
15. y = √cosVx 16. y = (sinx + cosx) 3
'x 2 + 5x + 4 for x ≤ 0
17. y = (Vx+λZx) 5 18. I/ f ( / = ■ 5x + 4 forO < x < 6
/ .x 2 - 2 for x ≥ 6
19. y = sin (sinx) 20. y = cos (sinx)
21. x ιz2 + y υ 2 = x 22. 4x 2 - 1 6 / = 64 47. Find equations of the tangent and normal lines to the
curve given by x 3 — y 3 = 2xy at the point ( - 1, 1).
23. sin xy = y + x
48. Use differentials to approximate (16.0 l) 3z2 + 2V16.01.
24. sin (x + / + cos (x - y) = xy
49. Use differentials to approximate cos-^yj1. Use the approx-
i-mations V 3 ≈ 1.7321 and π ≈ 3.1416.
dv 1
Find in Problems 25-28. 50. Use differentials to estimate the change in the volume of a
dχ-
25. y = x 5 - 5x 4 + 7x 3 - 3x 2 + 17
cone if the height of the cone is increased from 10 cm to
10.01 cm while the radius of the base stays fixed at 2 cm.
26∙J'-≡⅞⅞+<3* - 1>
2 51. The newspaper clipping below was taken from the Wall
Street Journal. Restate the headline, “Effective Marginal
27. x2 + / = 10
Tax Rates,” using calculus.
28. x 2 + sin y = 2
In Problems 29-36, fin d an equation o f the tangent to the curve

at the indicated point.
29. y = x 4 - 7x 3 + x 2 - 3 at (0, -3 )
30. y = (3x 2 + 5x —7)3 where x = 1
31. y = x cosx where x =
33. x y 2 + x 2y = 2 at (1, 1)
34. y = (x 3 - 3x 2 + 4)2 where x = 1
Qr —4
35. y = —-+------7 where x = 1
• 3x 2 + x - 5
36. x 2z3 + y 2z3 = 2 at (1, 1)
37. Let f( x ) = (x 3 —x 2 + 2x - 1)4 . Find equations of the *Senate Finance Committee
tangent and normal lines to the graph of f a t x = 1. Source: Harris Bank Economic Research Office
38. Find equations for the tangent and normal lines to the
graph of y = (fix + £)3 at the point (1, 27). 52. On New Year’s Eve, a network TV camera is focusing on
d d2 the descent of a lighted ball from the top of a building that
39. Use the chain rule to find +7 and -+y when y = x 3 - 7 x is 600 ft away. The ball is falling at the rate of 20 ft/min.
dt dt 2
and x = t sin t. At what rate is the angle of elevation of the camera’s line
of sight changing with respect to time when the ball is 800
40. Find ∕" ( x ) if ∕( x ) = x 2sinx 2 .
ft from the ground?
41. F i n d / ', / " , / " ' a n d ∕< 4>if ∕( x ) = x 4 - 53. Suppose / is a differentiable function whose derivative
42. Find∕ ' , ∕ " , an d ∕ ' " if∕( x ) = x(x 2 + l) 7z2. satisfies ∕'( x ) = 2x 2 + 3. Find j y f ( x 3 ~ 1)∙
54. Suppose/is a differentiable function such that (thousand dollars) for x tons (0 ≤ x ≤ 7) produced.
Estimate the change in the unit price that accompanies
∕'( x ) = x 2 + x . Find + x)∙ each change in sales:
55. Let ∕( x ) = 3x 2 + 1 for all x . Use the chain rule to find a. From 2 tons to 2.05 tons
⅛ (W )(⅛ b. From 2 tons to 2.1 tons
c. From 2 tons to 1.95 tons
56. Let f(x ) = sin2x + cos3x and g(x) = x 2 . Use the chain 67. A company sends out a truck to deliver its products.
rule to find ( ∕o g)(x). When traveling at a constant rate of x mi/hr (with x ≥ 5),
the truck consumes gasoline at the rate of
57. Let
1 1,500
x sin i if x Ψ 0 300
f( x ) =
0 ifx = 0 gal∕m i. The driver is paid $ 16 per hour to drive the truck
Use the definition of the derivative to find∕'( x ) if it exists. 300 mi. Gasoline costs $2 per gallon.
58. A car and a truck leave an intersection at the same time. a. Find an expression for the total cost C(x) of the trip.
The car travels north at 60 mi/hr and the truck travels east b. Use increments to estimate the additional cost if the
at 45 mi/hr. How fast is the distance between them truck is driven at 57 mi/hr instead of 55 mi/hr.
changing after 45 minutes? 68. A viewer standing at ground level and 30 ft from a
59. A truck leaves an intersection going north at 45 m i∕h. At platform watches an object rise from that platform at the
the same time a car leaves the intersection and is traveling constant rate o f 3 ft/s. (See Figure 2.55.) How fast is the
east at 60 mi/hr. How fast is the distance between the car angle o f sight between the viewer and the object changing
and the truck changing after 1 hour? at the instant when 0 = -J?
60. A spherical balloon is being filled with air in such a way
that its radius is increasing at a constant rate of 2 cm ∕s. At
what rate is the volume of the balloon increasing at the
instant when its surface has area 4π cm 2?
61. A block of ice in the shape of a cube originally having
volume 1,000 cm 3 is melting in such a way that the length
of each of its edges is decreasing at the rate of 1 cm ∕hr. At
what rate is its surface area decreasing at the time its
volume is 27 cm 3? Assume that the block o f ice maintains
its cubical shape.
62. Show that the rate of change of the area of a circle with
respect to its radius is equal to the circumference.
63. A charged particle is projected into a linear accelerator.
The particle undergoes a constant acceleration that 69. A particle of mass m moves along the x-axis. The velocity
changes its velocity from 1,200 m/s to 6,000 m/s in
v = dx
~
2 × 10- 3 sec. Find the acceleration o f the particle. 6∣ t and position x = x(t) satisfy the equation
64. A rocket is launched vertically from a point on the ground
that is 3,000 horizontal feet from an observer with m(v2 - v0 2) = k(x 0 2 - x 2)
binoculars. If the rocket is rising vertically at the rate of
where ∕<, x 0 , and v0 are positive constants. The force F
750 ft/s at the instant it is 4,000 ft above the ground, how
acting on the object is defined by F = ma, where a is the
fast must the observer change the angle of elevation of her
object’s acceleration. Show that F = - k x .
line of sight in order to keep the rocket in sight at that
instant? 70. Let f and g be differentiable functions such that
∕[g ( x ) ] = x a ∏d g[∕(X)] = χ for all x (such functions are
65. Assume that a certain artery in the body is a circular tube
said to be inverses). Show that
whose cross section has radius 1.2 mm. Fat deposits are
observed to build up uniformly on the inside wall o f the dg= J_
artery. Find the rate at which the cross-sectional area of dx df
the artery is decreasing relative to the thickness of the fat dx
deposit at the instant when the deposit is 0.3 mm thick.
66. A processor who sells a certain raw material has analyzed 71. A baseball player is stealing second base. He runs at 30
the market and determined that the unit price should be ft/s and when he is 25 ft from second base, the catcher,
while standing at home plate, throws the ball toward the
p(x) = 60 - x 2 base at a speed of 120 ft/s. At what rate is the distance
between the ball and the player changing at the time the
ball is thrown?
d. Pick x fl bigger (for example, X Q = 3, or larger). Do
several iterations and explain the results.
e. Try x 0 around 0.5, and explain what happens, and
why it happens.
f. Try x 0 around - 1.0 and tell what happens, and
why.
75. Chaos Here we investigate the notion o f chaos using
the Newton-Raphson method. Basically, in mathe
matical terms, this means that the outcome of a
problem is extremely sensitive to starting values.
Here we apply the Newton-Raphson method (unnec
essarily) to the function
72. A rod OA 2 m long is rotating counterclockwise in a plane
about O at the rate o f 3 rev/s as shown in Figure 2.56. ∕( x ) = x 3 - x = x(x - 1)(x + 1)
Investigate what happens as we change the starting
value, x 0 .
a. Two important x-values in our study are
1>= V5 a" d ⅛^√5

One would not want to pick x 0 as either ±s 3. Why
not? Also, what happens if one picks xθ = s5? That is,
what are x 1, x 7, . . . .
Figure 2.56 Problem 72 b. Generate a good plot o f∕( x ) on [ - 2 , 2]. Explain
from the plots why you would expect that an
The rod A B is attached to OA, and the end B slides along initial value x n > s., would lead to lim x = 1.
the x-axis. Suppose A B is 5 meters long. What are the (Also, by symmetry, x fl < j - 5 , leads to lim,1 x
velocity and the acceleration of the motion of the point B _ _ jj n→∞
along the x-axis? c. Explain why if ∣x0 ∣ < s s , you would expect
73. The ideal gas law states that for an ideal gas, P V = kT x ,, → 0 ∙
where P is the pressure, V is the volume, T is the d. Numerically verify your assertions of parts b and
temperature, and k is a constant. Suppose the tempera c.
ture is kept fixed at 100 °C, and the pressure decreases at
e. Now to see the “ chaos,” use the following xθ
the rate of 7 lb ∕in . 2 per min. At what rate is the volume
values and take 6 to 10 iterations, until conver
changing at the instant the pressure is 25 lb ∕in .2 and the
gence occurs, and report what happens: 0.448955;
volume is 30 in.3?
0.447503; 0.447262; 0.447222; 0.4472215;
0.4472213.
76. P U T N A M E X A M IN A T IO N P R O BLEM Suppose ∕( x ) = ax 2
74. This problem deals with the function + bx + c, where a, b, and c are real numbers and
∣∕(x)∣ ≤ 1 fo r x ≤ 1. Show that ∣∕'(x)∣ ≤ 4 for ∣x∣ ≤ 1.
∕∙ ∕ vλ _ x 3 - 2.1x 2 + A' - 2
x6 + 1 77. P U T N A M E X A M IN A T IO N P R O BLEM A point P is taken on
a. Prove that /h as at least one zero in the interval the curve y = x 3. The tangent at P meets the curve again
[-1 0 , 10]. at 0 . Prove that the slope o f the curve at Q is four times
the slope at P.
b. Graphically show that/has only one zero in this
interval. Note: This will require some careful 78. P U T N A M E X A M IN A T IO N P RO BLEM A particle o f unit
specification of domains and ranges. mass moves on a straight line under the action of a force
that is a function /(v) o f the velocity v of the particle, but
c. Pick a reasonable starting x 0-vakιe (for example,
the form of this function is not known. A motion is
the closest integer), and do enough iterations of
observed, and the distance x covered in time t is found to
the Newton-Raphson method to get at least six
be related to t by the formula x = at + bt1 + ct3, where a,
significant figure precision. Hand in computer
b, and c have numerical values determined by observation
outputs showing all the x n values.
of the motion. Find the function ∕(v ) for the range of v
covered by this experiment.
GROUP RESEARCH PROJECT*
“Ups and Downs"
T h is project is to be done in groups o f three or four students. Each group will submit
a single written report.
Y ou have been hired by Two Flags to help with the design of their new roller
coaster. You are to plan the path design of ups and downs for a straight stretch
with horizontal length of 185 ft. One possible design is shown below, and here is
the inform ation you will need:
a. There must be a support every 10 ft.
b. The descent can be no steeper than 80° at any point (angles refer to the
angle that the path makes with a horizontal).
c. The design must start with a 45° incline.
d. The amount of material needed for a support is the square of the height of
the support. For example, a support that is 20 ft high requires 20 2 = 400 ft
Mathematics is one o f the oldest of of material.
the sciences', it is also one o f the
most active, for its strength is the We also define the thrill of the coaster as the sum of the angle of steepest descent in
vigor o f perpetual youth. each fall in radians plus the num ber of tops.
A. R. Forsyth
Nature 84 (1910): 285
The whole o f mathematics consists Y our paper should include, but is not limited to, the following concerns: You
in the organization o f a series o f should design two roller coaster paths; one should be the most thrilling, and the
aids to the imagination in the proc other should use the least material for supports. Your paper should address where
ess o f reasoning. the path is increasing at an increasing rate, and decreasing at a decreasing rate.
A. N. Whitehead
You should address the safety criteria (a-d listed above) and show the graphs of the
Universal Algebra (Cambridge,
1898), p. 12 path, the slope, and the rate of change of the slope.
This group project is from Diane Schwartz from Ithaca College, Ithaca, N.Y.
Additional Applications of
the Derivative
■ CONTENTS
3.1 Extreme Values of a
Continuous Function
Extreme value theorem;
Relative extrema; Absolute
extrema; Optimization
3.2 The Mean Value Theorem
Rolle’s theorem and a proof of PREVIEW
the MVT; Some uses of the
mean value theorem In Chapter 2, we used the derivative to find tangent lines and to
3.3 First-Derivative Test compute rates of change. The primary goal of this chapter is to examine
Increasing and decreasing
functions; The first-derivative the use of calculus in curve sketching, optimization, and other applica
test; Curve sketching with the tions.
first derivative
3.4 Concavity and the Second-
Derivative Test
Concavity; Inflection points; PERSPECTIVE
Curve sketching with the second
derivative; The second-deriv Homing pigeons and certain other birds are known to avoid flying over large
ative test for relative extrema bodies of water whenever possible. The reason for this behavior is not entirely
3.5 Infinite Limits and Asymptotes known. However, it is reasonable to speculate that it may have something to do
Asymptotes; Limits involving with minimizing the energy expended in flight, because the air over a lake is
infinity; Graphs with asymp often “heavier” than that over land. Suppose a pigeon is released from a boat at
totes
point B on the lake shown in the accompanying figure.* It will fly to its loft at
3.6 Summary of Curve Sketching point L on the lakeshore by heading across water to a point P on the shore and
General curve sketching;
Graphing strategy then flying directly from P to L along the shore. If the pigeon expends ew units
3.7 Optimization in the Physical of energy per mile over water and e1 units over land, where should Pbe located
Sciences and Engineering in order to minimize the total energy expended in flight?1.
Optimization procedure;
Physics: Fermat’s principle of B
optics and Snell’s law
3.8 Optimization in Business, Eco
nomics, and the Life Sciences
Optimization of discrete
functions; Economics: two prin
ciples of marginal analysis;
Business management: an
inventory model; Physiology:
the optimal angle for vascular
branching Optimization (finding the maximum or minimum values) is one of the most
important applications we will study in calculus. From maximizing profit and
3.9 l'H6pital's Rule
ΓH∂pitaΓs rule; Indeterminate minimizing cost, to maximizing the strength of a structure or minimizing the
form 0/0; Indeterminate form distance traveled, it is easy to think of optimization problems that can be solved
∞∕∞j ΓH6pitaΓs rule for other with calculus. Optimization problems are considered in Sections 3.7 and 3.8.
forms
3.10 Antiderivatives
Reversing differentiation; *Edward Batschelet, Introduction to Mathematics for Life Scientists, 2nd ed. (New York: Springer-
Antiderivative notation; Verlag, 19), pp. 276-277.
Antidifferentiation formulas; t See Problem 25, Section 3.8.
Applications 165
Chapter 3 Review
166 Ch. 3 Additional Applications of the Derivative
3 .1 EXTREME VALUES OF A CONTINUOUS FUNCTION
IN THIS SECTION Extreme value theorem, relative extrema, absolute

extrema, optimization
It is often very important to know what the largest and smallest values of a
function are and where these extreme values occur. In this section, we shall see
how the derivative may be used to locate extreme values.
■ EXTREME VALUE THEOREM
One of the principal goals of calculus is to investigate the behavior of various

functions. As part of this investigation, we will be laying the groundwork for
solving a large class of problems that involve finding the maximum or minimum
value of a function. Such problems are called optimization problems. Let us begin
by introducing some useful terminology.
Absolute Maximum Let /b e a function defined on a set D that contains the number c. Then
∕(c ) is the absolute maximum of / o n D if
∕( t ,) ≥ f(χ) for all x in D
f(c) is the absolute minimum of /o n D if
∕(c ) ≤ ∕(x ) for all x in D
Together, the absolute maximum and minimum of/ on the interval / are called
the extreme values, or the absolute extrema, of / on I. A function does not
necessarily have extreme values on a given interval. For instance, the continuous
function g(x) = x has neither a maximum nor a minimum on the open interval (0,
1), as shown in Figure 3.la.
The discontinuous function defined by
a. The continuous function
g(x) = x has no extrema Λ(x) = { * 2 for x ≠ 0
on the open interval (0, 1).
for x = 0
has a maximum on the closed interval [-1 , 1] but no minimum, as shown in
Figure 3. lb. Incidentally, this graph also illustrates the fact that a function may
assume an absolute extremum at more than one point. In this case, the maximum
is at (-1 , 1), (0, 1), and (1, 1). If a function/is continuous and the interval / i s
closed and bounded, it can be shown that both an absolute maximum and an
absolute minimum must occur. This result, called the extreme value theorem, plays
an important role in our work.
THEOREM 3.1 Extreme value theorem
b. The discontinuous function h A continuous function / on a closed, bounded interval [a, b∖ has an absolute
has a maximum, but not a
minimum on the closed
maximum and an absolute minimum.
interval [-1, 1].
Figure 3.1 Functions that lack one Proof: Even though this result may seem quite natural (see Figure 3.2), its
or both extreme values proof requires concepts beyond the scope of this text and will be omitted.
Sec. 3.1 Extreme Values of a Continuous Function 167
Absolute maximum at x = c 1 and x = c.

Absolute minimum at.r = d
Figure 3.2 Extreme value theorem
This theorem does not apply if the function is not continuous and it also does
not apply if the interval is not closed. You will be asked for appropriate
counterexamples in the problem set.
If we compare a function with its graph, we note that the maximum of a
function occurs at the highest point on the graph and the minimum occurs at the
lowest point. These properties are illustrated in Example 1.
EXAM PLE 1 Extreme values of a continuous function
Using the graph in Figure 3.3, locate the extreme values of the function f
defined on the closed interval [a, b].
Figure 3.3 A continuous function on a closed interval [a, b]
Solution The highest point on the graph occurs at the right endpoint F, and the
lowest point occurs at C. Thus, the absolute maximum is f(b), and the
absolute minimum is ∕( c 2). M M
EXAM PLE 2 Conditions of the extreme value theorem
Explain why each of the following examples does not contradict the extreme
value theorem:
if 0 ≤ x < 1
a. fM = if 1 ≤ χ ≤ 2 b. g(x) = x 2 for 0 < x ≤ 2
Solution
a. The function f has no maximum. It takes
on all values arbitrarily close to 2, but it
never reaches the value 2. The extreme
value theorem is not violated because f is
not continuous on [0, 2].
f does not have a maximum value.
b. Although the functional values g(x) be

come arbitrarily small as x approaches 0,
it never reaches the value 0, so it has no
minimum. The function g is continuous
on the interval (0, 2], but the extreme
value theorem is not violated because the
interval is not closed.
g does not have a minimum value

(but it does have a maximum value).
a RELATIVE EXTREMA
Typically, the extrema of a continuous function occur either at endpoints of the

interval or at points where the graph has a “peak” or a “valley” (points where the
graph is higher or lower than all nearby points). For example, the function f in
Figure 3.3 has “peaks” at B and D and “valleys” at C and E. Peaks and valleys are
what we call relative extrema.
Relative Maximum The function/is said to have a relative maximum at the point c if∕(c ) ≥ ∕(x )
Relative Minimum for all x in an open interval containing c. Also, f is said to have a relative
minimum at d if∕( √ ) ≤ ∕(x ) for all x in an open interval containing d. Relative
maxima and relative minima are called, in general, relative extrema.
We would like to formulate a procedure for finding relative extrema. By

looking at Figure 3.3, we see that there are horizontal tangents at B, C, and E,
whereas a tangent does not exist at D. This suggests that the relative extrema off
occur either where the derivative is zero (horizontal tangent) or where the
derivative does not exist (no tangent). This notion leads us to the following
definition.
Critical Values and Critical Suppose/is defined at c and either∕'( c ) = 0 or ∕'( c ) does not exist. Then the
Points number c is called a critical value o f / and the point P(c,∕(c)) on the graph o f /
is called a critical point.
0 Note that if ∕(c ) is not de If there is a relative maximum at c, then the functional value, ∕(c), is that
fined, then c cannot be a critical maximum value. Similarly, if there is a relative minimum at c, then that minimum
value. 0 value is ∕(c).
EXAMPLE 3 Finding critical values

Find the critical values for the given functions.
a. ∕(x ) = 4x 3 - 5x2 - 8x + 20 b. f(x) = - ⅛ c. ∕(x ) = 12xιz2 - 2x 3z2
Solution
a. f'(x ) = 12x2 - 1Ox - 8 is defined for all values of x. Solve
12x2 - 1 0 x -8 = 0
2 (3 x -4 )(2 x + 1) = 0
X = j, —2 These are the critical values.

_ (x - 2)(2x) - x 2(l) = x(x - 4)
7 , (x - 2)2 (x - 2)2
The derivative is not defined at x = 2, but f is not defined at 2 either; so
x = 2 is not a critical value. The actual critical values are found by solving
∕'( x ) = 0.
*(* ~ 4 ) _ ∩
(x - 2)2
x = 0, 4 These are the critical values.
c. ∕'( x ) = 6x -ιz 2 - 3x ιz2
The derivative is not defined at x = 0. We have /(0) = 12(0)1 /2 —
2(0)3z2 = 0, so we see that f is defined at x = 0, which means that x = 0 is a
critical value. For other critical values, solve ∕'( x ) = 0:
6x" ιz2 - 3x ιz2 = 0
3x~ιz2( 2 - x ) = 0
x=2
The critical values are x = 0, 2. BBB ∣
EXAMPLE 4 Critical values and critical points

Find the critical values and the critical points for the function
∕(x ) = (x - l) 2(x + 2)
Solution Because the function/is a polynomial, we know that it is continuous
and that the derivative exists for all x. Thus, we find the critical values by
using the product rule and chain rule to solve the equation ∕'( x ) = 0.
∕'( x ) = ( x - l ) 2(l) + 2 (x - l)(l) (x + 2)
= ( x - l)[(x - 1) + 2(x + 2)]
= (x —l)(3x + 3)
= 3(x - l)(x + 1)
The critical values are x = ± 1. To find the critical points, we need to find the
p-component for each critical value.
/(1) = ( 1 - 1 ) 2(1 +2) = 0
/( - 1 ) = ( - 1 - 1)2(-1 + 2 ) = 4
Thus, the critical points are (1, 0) and (-1 , 4).
We will use the critical points to help us draw a graph in Section 3.3. However, if
we use a computer or calculator to graph f in Example 4, we can see that the
relative extrema are related to the critical points as shown in Figure 3.4.
Our observation that relative extrema occur only at points on a graph where
there is either a horizontal tangent line or no tangent at all is equivalent to the
following result.
THEOREM 3.2 Critical value theorem

If a continuous function /h a s a relative extremum at c, then c must be a critical
value o f /
Figure 3.4 The graph of
f(x) = (x - l) 2(x + 2) I What this says: If a point is a relative maximum or a relative minimum
value for a function, then either the derivative is 0 or it does not exist at that
point.
Proof: If∕'(c ) does not exist, then c is a critical value by definition (∕(c) must
exist). We shall show that if∕'( c ) exists and a relative maximum occurs at c, then
f'(c) = 0. Our approach will be to examine the difference quotient. (The case
where f'(c) exists and a relative minimum occurs at c is handled similarly in
Problem 56 of Problem Set 3.1.)
Because a relative maximum occurs at c, we have ∕(c ) ≥ ∕(x ) for every
number X in an open interval (a, b) containing c. Thus, if ∆x is small enough so c +
∆x is in (a, b), then
∕(c ) ≥ ∕( c + ∆x) Because a relative maximum occurs at c
f(c) - ∕( c + ∆x) ≥ 0
∕( c + ∆x) - ∕(c) ≤ 0 Multiply both sides by - 1.
{Remember to reverse the inequality.)
For the next step we want to divide both sides by ∆x (in order to write the left side
as a difference quotient). However, as this is an inequality, we need to consider
two possibilities:
1. Suppose ∆x > 0 (the inequality does not reverse):
≤ 0 Divide both sides by ∆x.
Now we take the limit of both sides as ∆x approaches from the right (because
∆x is positive).
∕( c + ∆x) - ∕(c)
lim lim 0
Δ,τ→0* ∆x ΔΛ → 0 +
∕'(C) ^^δ-
∕'(C) ≤ o
2. Next, suppose ∆x < 0 (the inequality reverses). Then
f(c + ∆x) - f(c)
-------- ≥ 0
-------------- T — Divide both sides by ∆x.
This time we take the limit of both sides as ∆x approaches 0 from the left
(because ∆x is negative).
.. ∕( c + ∆ x ) - ∕( c ) ,.
hm —----- τ - — ~ l* ≥ lim 0n
Δ,v→0^ ∆X Δx→0"
Because we have shown that ∕'( c ) ≤ 0 and ∕'( c ) ≥ 0, it follows that ∕'( c ) = 0.B
0 Theorem 3.2 tells us that a relative extremum of a continuous function f

can occur only at a critical value, but z7 does not say that a relative extremum
must occur at each critical value.
For example, if∕( x ) = x 3, then f ∖x) = 3x 2 and ∕'( 0 ) = 0, but there is no
relative extremum at c = 0 on the graph of /because the graph is rising for
x < 0 and also for x > 0, as shown in Figure 3.5a. It is also quite possible
for g ∖c) not to exist for a continuous function g that has no relative
extremum at c, as shown in Figure 3.5b. 0
EXAM PLE 5 Critical values where the derivative does not exist
Find the critical values fo r ∕( x ) = ∣x + 1∣ on [ -5 , 5].

a. The graph of∕( x ) = x 3
No relative extremum occurs Solution If x > - 1 , then∕( x ) = x + 1 a n d ∕'(x ) = 1. How
at c = 0 even though/ '(0) = 0
ever, if x < - 1 , then ∕( x ) = - x - 1 and ∕'( x ) = - 1 .
Because these are not the same,
Jim ∕'( x ) ≠ l i m ∕' ( x )
the derivative does not exist at x = c = - 1 . Because
/ ( - 1) is defined, —1 must be a critical value.
■ ABSOLUTE EXTREMA
Suppose we are looking for the absolute extrema of a continuous function / on the
closed, bounded interval [a, b]. We know that these extrema exist by Theorem 3.1.
Also, Theorem 3.2 enables us to narrow the list of “ candidates” for points where
b. Although g'( 1) does not exist, extrema can occur from the entire interval [a, b} to just the endpoints x = a, x = b,
no relative extremum occurs
at c = 1. and the critical values c between a and b. This suggests the following procedure.
Figure 3.5 A relative extremum
may not occur at each critical Procedure for Finding Absolute Extrema
value. To find the absolute extrema of a continuous function / on [a, b}∖
Step 1. Compute ∕'( x ) and find all critical values o f/ on [a, b].
Step 2. Evaluate / at the endpoints a and b and at each critical value c.
Step 3. Compare the values in Step 2.
The largest value is the absolute maximum of /o n [a, b].
The smallest value is the absolute minimum o f/ o n [a, b∖.
Figure 3.6 shows some of the possibilities in the application of this procedure.
f(x ) = x i - 4x 2 + 3Λ' + 3 on [0.5, 2] g(x) = sin x on [0, 2π] ⅛(x) = x 2 - 4 x + 1 on[0, 3]

a ./ i s continuous; both b. g is continuous; neither c. h is continuous; one
Figure 3.6 Absolute extrema extrema at endpoints extremum at an endpoint extremum at an endpoint
EXAM PLE 6 Absolute extrema of a polynomial function
Find the absolute extrema of the function defined by the equation

f(χ ) = x 4 - 2x 2 + 3 on the closed interval [ - 1, 2].
Solution Because f is a polynomial function, it is continuous on the closed
interval [ -1 , 2]. Theorem 3.1 tells us that there must bean absolute maximum
and an absolute minimum on the interval.
Step 1: ∕'( x ) = 4x 3 - 2(2x)
= 4x(x 2 - 1)
= 4x(x - 1)(x + 1)
The critical values are x = 0, 1, and - 1 .
Step 2: Values at endpoints: / (-1 ) = 2,/(2) = 11
Figure 3.7 The graph of Critical values: /(0) = 3,/(1) = 2
∕( x ) = x 4 - 2x 2 + 3 on [ - 1, 2] Step 3: The absolute maximum of / occurs at x = 2 and is /(2) = 11; the
absolute minima of/occur at x = 1 and x = - 1 and a r e ∕( l) = / (-1 ) = 2. The
graph o f/ is shown in Figure 3.7. ∣= ι
EXAM PLE 7 Absolute extrema when the derivative does not exist
Find the absolute extrema o f ∕( x ) = x 2z3(5 - 2x) on the interval [ -1 , 2].

Solution Step 1: In order to find the derivative, rewrite the given function as
∕( x ) = 5x 2z3 - 2x 5z3. Then
∕'( x ) = j fx ~ ιz3 - j f x 2z3 = j f χ - ιz3(l - x)
Critical values are found by solving∕'(x) = 0 and by locating the places where
the derivative does not exist. First,
∕'( x ) = 0 when x = 1
Even though /(0) exists, we note that ∕'( x ) does not exist at x = 0 (notice the
division by zero when x = 0). Thus, the critical values are x = 0 and x = 1.
Step 2: Values at endpoints: / (-1 ) = 7
/(2) = 2 2 z3 ≈ 1.587401052
Critical values: /(0) = 0
/(1) = 3
Step 3: The absolute maximum o f/ occurs at x = -1 and is / (-1 ) = 7; the
absolute minimum of/occurs at x = 0 and is/(0) = 0, as shown in Figure 3.8.
EXAM PLE 8 Absolute extrema for a trigonometric function
Find the absolute extrema of the following continuous function on [θ, f j :

T(x) = ∣(sin 2x + cosx) + 2 sinx - x.
Figure 3.8 Graph of
∕( x ) = 5x 2z3 - 2x 5z3 on [ - 1, 2] Solution Step 1: We have T'(x) = ⅛(2 sinx cosx - sinx) + 2(cosx) - 1
= ∣(2 sinx cosx — sinx + 4 cosx —2)
= ∣[2 cosx(sinx + 2 ) - (sinx + 2)]
= j[(sin x + 2)(2 cosx - 1)]
Set each factor equal to zero and solve to obtain x = y.
Step 2: Evaluate the function at the endpoints:

7,(0) = ∣(sin 2 0 + cos 0) + 2 sin 0 - 0 = ∣(0 + 1 ) + 2(0) - 0 = 0.5
T(y) = j(sin 2 J + co sf) + 2 s in f - J
= ∣(1 + 0) + 2(1) - ⅞= 2 - f ≈ 0.9292036732

Critical value:
T(j) = j(sin 2 j + cos + 2 sin j j
= ⅛(¾+ ⅛) + 2 (λ f ) - y = ∣ + V 3 - f ≈ 1-309853256
Step 3: The absolute maximum of T is approximately 1.31 at x = y and the

absolute minimum of T is 0.5 at x = 0. The graph is shown in Figure 3.9 as
part of the computational window. M ≡
■ OPTIMIZATION
In our next two examples, we examine two applications involving optimization.

Such problems are investigated in more depth in Sections 3.7 and 3.8.
EXAM PLE 9 Maximum and minimum velocity of a moving particle
A particle moves along the Z-axis with displacement

s(t) = Z4 - 8 Z 3 + 18Z2 + 60r —8
Find the largest and smallest values of its velocity for 1 ≤ t ≤ 5.
Solution The velocity is
v(Z) = s∖t) = 4 ∕ 3 - 24Z2 + 36t + 60
17 4 Ch. 3 Additional Applications of the Derivative
To find the largest value of v(t), we compute the derivative of v:

v'(t) = 12t2 - 48t + 36
= 1 2 ( Z - 3 ) ( t- 1)
Setting v'(t) = 0, we find that the critical values of v(t) are t = 3, 1. {Note-, vis a
polynomial function, so it has a derivative for all t.) Now we evaluate v at the
critical values and endpoints.
t = 1 is both a critical value and an endpoint:
v(l) = 4(i)3 _ 24(1)2 + 36(1) + 60 = 76
t = 5 is an endpoint:
v(5) = 4(5)3 _ 24(5)2 + 3 6 (5) + 60 = 140
t = 3 is a critical value:
v(3) = 4(3)3 _ 24(3)2 + 36(3) + 60 = 60
The largest value of the velocity is 140 at the endpoint where t = 5, and the
smallest value is 60 when t = 3. ■■■
EXAM PLE 10 An applied maximum value problem
A box with a square base is constructed so that the length of one side of the
base plus the height is 10 in. What is the largest possible volume of such a box?
Solution We let b be the length of one side of the base and h be the height of the
box, as shown in Figure 3.10. The volume, V, is
K = b 2h
Because our methods apply only to functions of one variable, it may seem that
we cannot deal with E as a function of two variables. However, we know that
b + h = 10; therefore, h = 10 - b, and we can now write V as a function of b
a lo n e : V{b) = 6 2( 1 0 - b)
The domain is not stated, but we must have b ≥ 0 and 10 - b = h ≥ 0, so

that 0 ≤ b ≤ 10. First, we find the critical values. Note that V is a
polynomial function, so the derivative exists everywhere in the domain. Write
V(b) = 1062 - b i and find K'(6) = 206 - 3 6 2 . Then
V'(b) = 0
Figure 3.10 Volume of a box
206 - 36 2 = 0
6(20 - 36) = 0
6 = 0 ,f
Checking the endpoints and the critical values, we have

K(0) = 0
K(10) = 102( 1 0 - 10) = 0
^ ) = m 2( 1 0 -⅛ )= fr
Thus the largest value for the volume Kis ≈ 148.1 in 3 . It occurs when the
square base has a side of length in. and the height is 6 = 1 0 - y = j y in . H
PROBLEM SET 3 .1
O In Problems 1-12, fin d the largest and smallest values o f each 20. g(x) 2x 3 - 3x 2 - 36x + 4 on [ - 4 , 4]
=
continuous function on the given closed, bounded interval. 21. g(x) x 3 + 3x 2 - 24x - 4 on [ - 4 , 4]
=
1. f(x ) = 5 + 10x - x 2 on [ - 3 , 3] 22. f ∖x) ∣ x 3 - 5X 2 + 8x - 5 on [ - 4 , 4]
=
2. fix ) = 10 + 6x - x 2 on [ - 4 , 4]
23. f ix )∣( x 3 - 6x 2 + 9x + 1) on [0, 2]
=
3. f{x ) = x i - 3x 2 on [ - 1, 3]
24. g fι) = 98M3 - 4 M2 + 72M on [0, 4]
4. fit) = t4 - St2 on [ - 3 , 3]
25. flw ) = V'⅛,(½, —5)lz3 on [0, 4]
5. fix ) = x 3 on [ - ∣ , 1] 26. A(x) = ^tyχ λ∕z(x —3)2 on [—1, 4]
6. glx) = x 3 - 3x on [ - 2 , 2] 27. ∕ι(x) = tan x + secx on [0, 2τr]
7. f ix ) = x 5 - x 4 on [ - 1, 1] 28. sit) = t cost - sin / on [0, 2ττ]
8. git) = 3/ - 2 0 t 3 on [ - 1 , 3] i2o9 , / r(∕vx-∖) - f 9 - 24x if x < 1 , .
n [r - 1 > 4 i1
l - x + 6x i f x ≥ 1 o
9. f ix ) = ∣x∣ on [ - 1 , 1]
10. f ix ) = ∣x - 3∣ on [ - 4 , 4] 3 0 . ∕( x ) = { 8 ^ 3^ ' f∕ < 2 on [ - 1 , 4]
l - x 2 + 3x if x ≥ 2 l ’ j
11. f(u) = sin 2 w + cos M on [0, 2]
12. glu) = sin u - cosw on [0, τr] Find the required extrema in Problems 31-35.
31. The smallest value o f∕( x ) = x 2 on [ - 1 , 1]
13. ■ What Does This Say? Outline a procedure for finding
the absolute extrema o f a function. Include in your 32. The largest value o f f ix ) = , * n on [-0 .5 , 0]
outline a discussion o f what is meant by critical values.
33. The smallest value of g(x) = $ + x — 3 on [ 1, 9]
34. The smallest value of g(x) = * 2 * on [ - 1 , 1]

Computational Window +
35. The largest value of

14. In Example 7 we t
found that the max f< t ∖ = f ~ t2 ~ t + 2 if < 1 0 n γ 2τ 3∏
∕ ( z) [3 _ t if t ≥ ι t > ]
imum value of
f ix ) = x 2z3(5 - 2x) θ Find a function that not only meets the conditions specified in
on [—1, 2] is 7 and Problems 36-39, but also meets each o f the following condi
occurs at x = - 1. If tions:
we graph this func a. a minimum but no maximum
tion on a leading b. a maximum but no minimum
⅛, 1≡X^<2×3) < 5-2X )
brand of graphing
X π in=^10 Y m in=-10 c. both a maximum and a minimum
calculator, we obtain
the graph at the X∩ax=10 Vnax=lΘ d. neither a maximum nor a minimum
right.
X s c l= l V ≥ c l= l That is, each o f these parts a -d may need a different example
to satisfy the conditions requested in Problems 36-39.
This graph is not correct. Can you explain the 36. T H IN K T A N K P RO BLEM Find a function that is discon
discrepancy? It is not our intent to “ make up” tinuous and defined on an open interval.
problems to use with a calculator, but whenever you
37. T H IN K T A N K P ROBLEM Find a function that is discon
use a calculator or computer to assist you with
tinuous and defined on a closed interval.
calculus, you must understand the nature o f the
functions with which you are working and not rely 38. T H IN K T A N K P RO BLEM Find a function that is continu
only on the calculator or computer output. ous on an open interval.
39. T H IN K T A N K P RO BLEM Find a function that is continu
ous on a closed interval.
40. T H IN K T A N K P ROBLEM Give a counterexample to show
In Problems 15-30, fin d the largest and smallest values o f each
that the extreme value theorem does not necessarily apply
continuous function on the closed, bounded interval. I f the
if one disregards the condition that/be continuous (that
function is not continuous on the interval, so state.
is,/need not be continuous).
15. flu ) = 1 - u 2' 3 on [ - 1 , 1]
41. T H IN K T A N K P RO BLEM Give a counterexample to show
16. git) = (50 + ∕) 2z3 on [-5 0 , 14]
that the extreme value theorem does not necessarily apply
17. fl&) = cos3 θ - 4 cos2 θ on [—0.1, π + 0.1] if one disregards the condition that / be defined on a
18. glθ) = θ sin θ on [—2, 2] closed interval (that is, / may be defined on an open
19. f ix ) = 20 sin(378πx) on [ - 1 , 1] interval).
42. An object moves along the Z-axis with displacement is minimized when the average cost is equal to C'(x).
(Recall from Chapter 2 that C'(x) is called the marginal
s(t) = t 3 - 6t 2 - 15/+ 11 cost.)
Find the largest value of its velocity on [0, 4]. θ 52. a. Show that ⅜is the number that exceeds its own square
43. An object moves along the Z-axis with displacement by the greatest amount.
b. Which nonnegative number exceeds its own cube by
s(f) = t4 - 2z3 - 12z2 + 60z - 10 the greatest amount?
Find the largest value of its velocity on [0, 3]. c. Which nonnegative number exceeds its nth power
(n > 0) by the greatest amount?
44. Find two nonnegative numbers whose sum is 8 and the
product of whose squares is as large as possible. 53. Given the constants α l , α2, . . . , an, find the value of x that
45. Find two nonnegative numbers such that the sum of one guarantees that the sum
and twice the other is 12 if it is required that their product S,(x) = (α 1 - x) 2 + (α2 - x) 2 + + {an - x) 2
be as large as possible.
46. Under the condition that 3x + y = 80, maximize the will be as small as possible.
product x y i when x ≥ 0, y ≥ 0. 54. Find the smallest value of m so that the line y = m x will
47. Under the condition that 3x + y = 126, maximize xy lie above the curve
when x ≥ 0 and y ≥ 0.
48. Under the condition that 2x —5y = 18, maximize x 2y
when x ≥ 0 and y ≤ 0. for all x > 0. Hint'. Begin by locating the value of x that
49. Show that if a rectangle with fixed perimeter P is to minimizes
enclose the greatest area, it must be a square. ∕( x ) = m x - 1 + i
50. Find all points on the circle x 2 + y 2 = a 2 (α ≥ 0) such
that the product of the ^-coordinate and y-coordinate is as 55. Explain why the function
large as possible.
51. A commodity has a cost function given by j , sιnx cosx
C(x) = 0.125x 2 + 20,000 must attain a minimum in the open interval (θ, ?). Show
where x is the number of units produced. Show that the that if the minimum is attained at x = θ. then
average cost,
tan 0 = 5
4«-®
56. Show that if ∕'( c ) exists and a relative minimum occurs at
c, then c must be a critical value of f.
3 .2 THE MEAN VALUE THEOREM
IN THIS SECTION Rolle’s theorem and a proof of the MVT, some uses of
the mean value theorem
The mean value theorem is an extremely important result in calculus, both in
theory and for practical applications. The goal of this section is to prove the
mean value theorem and to examine some of its consequences.
9 0 __ If the velocity is always above
60, then the average velocity
80 cannot be 60. If an automobile averages 60 mi/hr on a trip, it is reasonable to expect that the
7 0 --
speedometer must read exactly 60 at least once during the trip. More generally,
60-------------------------------------------------------
suppose an object moves along a straight line so that its distance is s(t) from its
If the velocity is always below starting point at time t. The average velocity of the object from time t = a to time
60, then the average velocity t = b is given by the ratio
cannot be 60.
s(b) - 5(a)
b- a
I I I I I I I I I ►
I 2 3 4 5 6 7 8 9 '
Just as in the case of the automobile, we expect that there should be at least one
Sec. 3.2 The Mean Value Theorem 177
time t0 between a and b when the instantaneous velocity equals this average
velocity. That is, there exists some t0 such that
∙Λ∕) = s(b) - s(a)
Instantaneous velocity ∖_b_- -aJ
Average velocity
This is an example of the mean value theorem for derivatives, which we shall
refer to simply as the mean value theorem.
Mean Value Theorem (MVT) I f /is continuous on the closed interval [a, 6] and differentiable on the open
interval (a, h), then there exists at least one number c such that
∕(⅛) - ∕(α )
— ^ _ a = f (c) for a < c < b
The proof of this theorem follows later in this section.
1 What this says: Under reasonable conditions, there must be at least one
number c between a and b such that the derivative ∕'( c ) actually equals the
difference quotient
f(,b) - Rd)
b- a
The equation in the MVT may or may not be easy to solve. Example 1 illustrates
one method of finding such a number c.
EXAMPLE 1 Finding the number c specified by the MVT
Show that the function ∕(x ) = x 3 + x 2 satisfies the hypotheses of the MVT on
the closed interval [1, 2], and find a number c between 1 and 2 such that
Solution Because / i s a polynomial function, it is differentiable and hence also

continuous on the entire interval [ 1, 2]. Thus, the hypotheses of the MVT are
satisfied.
By differentiating / we find that
f'(x ) = 3x2 + 2x
for all x. Therefore, we have ∕ , (c) = 3c2 + 2c, and the MVT equation
/(2 )-/(1 )
∕'( c ) 2 —1
is satisfied when
3c, +2c, Λ ^ a n , i2f i = 10
Solving the resulting equation 3c2 + 2c - 10 = 0 by the quadratic formula, we

obtain
c = -1 ± √31
3
The negative value is not in the open interval (1, 2), but the positive value
1 + √ 31 ≈ 1.522588121
satisfies the requirements of the MVT.
■ ROLLE'S THEOREM* AND A PROOF OF THE MVT
The key to the proof of the mean value theorem is the following result, which is
really just the MVT in the special case where f(b) = ∕(α).
THEOREM 3.3 Rolle’s theorem
Suppose/is continuous on the closed interval [a, b] and differentiable on the open
interval (a, b). Then if f(a) = fib), there exists at least one number c between a and
b such that ∕'( c ) = 0.
Proof: If f is constant on the closed interval [a, b], then f'(x ) = 0 for all x
between a and b. If/is not constant throughout the interval [a, b∖, the largest and
smallest values of/ on [a, b} cannot be the same. Because∕(α ) = ∕(⅛), at least one
extreme value of/ does not occur at an endpoint. According to the extreme value
theorem and the critical value theorem (Theorems 3.1 and 3.2 of Section 3.1),
such an extreme value must occur at a critical value between a and b. Recall that a
critical value c has the property that ∕'( c ) = 0 or else∕'( c ) does not exist. Because
f i x ) is assumed to exist throughout the interval (a, b), it follows that ∕'( c ) = 0 for
some number between a and b, as claimed. ■
The proof of Rolle’s theorem makes much more sense if it is considered

geometrically, as shown in Figure 3.11. Figure 3.1 la shows that if/is constant on
the closed interval [a, b], then ∕'( x ) = 0 for all x in the interval.
However, if/is not constant, as shown in Figure 3.1 lb, its graph must change
a. The special case where direction at least once in order to begin and end on the same level. [Remember
f(a) = f(b )∙Jis
constant for this example. th a t/is continuous and∕(α ) = f(b).] The graph must have a horizontal tangent
(with a derivative of zero) wherever the direction changes. Figure 3.1 lb shows that
it is quite possible for the graph to have several such transition points, even though
the theorem asserts only that at least one exists.
The mean value theorem has a similar geometric interpretation, as shown in
Figure 3.12. Rolle’s theorem can be viewed as saying that if∕(α ) = ∕(Z>) = d, then
for at least one number c between a and b, the tangent line at (c,∕(c)) is parallel to
the line y = d through the endpoints (a, d) and (h, d). It is reasonable to expect a
similar result to hold if the endpoints of the graph are not necessarily at the same
height. Note that in Figure 3.12b, the graph of Figure 3.12a has been “lifted and
tilted,” but the tangent line at (c, ∕(c)) is still parallel to the line L through the
endpoints P(a, f(a)) and Q(b, f(b)). Because the tangent line at (c, ∕(c)) has slope
b. The case w here/is

not constant on [a, ⅛]
*Rolle, s theorem is named for Michel Rolle (1652-1719), a French mathematician who
Figure 3.11 A geometric interpre investigated a special case of the mean value theorem in a book on the algebra of equations published
tation of Rolle’s theorem in 1690.
∕'( c ) , it follows that
b- a
Now we are ready to prove the mean value theorem.
THEOREM 3.4 The mean value theorem (MVT) for derivatives
If f is continuous on the closed interval [a, 6] and differentiable on the open

interval (a, 6), then there exists at least one number c such that
∕(⅛) - »
b —a for a < c < b
Proof: The proof of the M V T uses Rolle’s theorem. Suppose f is continuous

on [a, 6] and differentiable on (a, b). Define a new function g as follows:
g(×) = ∏ _ - a ) + ∕(α ) - ∕( * )
for a ≤ x ≤ b. Because f satisfies the hypotheses of the M V T , the function g is

also continuous on the closed interval [a, b] and differentiable on the open interval
(a, b). In addition, we find that
' ftb ) - f(a)'
g( a ) = b- a ( a - a ) + f(a) - j ∖a) = 0
'f(b) - f(a)
g(b) = b- a . (b - a) + ∕(α ) - /(6)
= lf(b) - ∕(α )] + f(a) - f(b) = 0
Thus, g satisfies the hypotheses of Rolle’s theorem, so there exists at least one
number c between a and b for which g'(c) = 0. Differentiating the function g, we
find that _ f( }
^'(X) =
2⅛7⅛
D - ∕'( x )
Because g'(c) = 0, we have
O ≈ g ,∕) = ≡ - Z{w - ∕' ( c )

This means that
' ,Λ _ ∕(⅛) ~ ∕(Q)
j f (l C > ~ b -a
Figure 3.12 A geometrical com as required. ■
parison of Rolle’s theorem and the
mean value theorem
■ SOME USES OF THE MEAN VALUE THEOREM
The M V T has many different uses. In Example 2, we use it to establish a

trigonometric inequality.
EXAM PLE 2 Using the MVT to prove a trigonometric inequality
Show that ∣sinx - siny∣ ≤ ∣x ~ y∣ for numbers x and y by applying the mean
value theorem.
Solution The inequality is true if x = y. Suppose x ≠ y; then /(0) = sin θ is
differentiable and hence continuous for all θ, with ∕'( 0 ) = cos θ. By applying
the M V T to f on the closed interval with endpoints x and y, we see that
A*) - M . jmv
x - y ’
for some c between x and y. Because f'(c ) = cosc and
∕(-v) - ∕(T ) = sin x - siny
we have s in x -s in y
-------------- - = cos c
x -y
Finally, we take the absolute value of the expression on each side, remember
ing that ∣cosc∣ ≤ 1 for any number c:
Isin x - siny∣ l .
-------------- -∖ = 1cosc 1 ≤ 1
I x - y I
Thus, ∣sinx - siny∣ ≤ ∣x - y∣, as claimed.
The M V T is usually used to prove certain key theoretical results of calculus.

For example, in Theorem 3.5 we will use the M V T to prove that a function whose
derivative is always zero on an interval must be constant on that interval. This
apparently simple result and its corollary (Theorem 3.6) turn out to be crucial to
our development of the integral in Chapter 4.
THEOREM 3.5 Zero derivative theorem
Suppose/is a continuous function on the closed interval [a, b] and is differentia

ble on the open interval (a, b). Then iff'( x ) = 0 for all x on (<2, ⅛), the function f is
constant on [a, b∖.
Proof: Let x 1 and x 2 be two distinct numbers (x l ≠ x 2) chosen arbitrarily from
the closed interval [a, b]. The function/satisfies the requirements of the M V T on
the interval with endpoints x l and x 2, which means that there exists a number c
between x , and x 2 such that
∕( x 2) - ∕( x 1) =
By hypothesis, ∕'( x ) = 0 throughout the open interval (a, ⅛), and because c lies
within this interval, we have ∕'( c ) = 0. Thus, by substitution we have
∕( x 2) ~ ∕( χ ∣)
x2 - x 1
∕( x 2) = ∕( x l )
Because x 1 and x 2 were chosen arbitrarily from [a, b], we conclude that∕( x ) = k, a
constant, for all x, as required. ■
THEOREM 3.6 Constant difference theorem
Suppose the functions / and g are continuous on the closed interval [a, 6] and
differentiable on the open interval (a, b). Then i f ∕'( x ) = g'(x) for all x on (a, b),
there exists a constant C such that
f(x ) = g(x) + C
for all x on [a, b].
Proof: Let ∕t(x) = f ( x ) - g(x); then
Λ'W = ∕ , w - √ w
= 0 Because f'( x ) = g'(x)
T h us, by Theorem 3.5, ∕z(x) = C for some constant C and all x on [a, ⅛], and
because ∕z(x) = ∕( x ) - g (x), it follows that
∕( x ) - g(x) = C
f i x ) = g(x) + C Add g(x) to both sides. ■
I W hat this says: Two functions with derivatives that are equal on an open
interval differ by a constant on that interval.
PROBLEM SET 3 .2
Q 1. I What Does This Say? What does Rolle’s theorem say, 19. f ix ) = Ψ x - 1 on [ - 8 , 8] 20. fix ) = on [ - 1 , 1]
and why is it important?
21. f ix ) = i on [1, 2] 22. f ix ) = 3x + secx on [-ττ, ττ]
2. ■ What Does This Say? Without looking, state the χ - ?
hypothesis for the M V T . How are the hypotheses used in
23. f ix ) = sin 2 x on [ —J,
the proof? Can the conclusion of the M V T be true if all or
part of the hypotheses are not satisfied? θ 24. Alternative form of the mean value theorem: If/is continu
ous on [a, b] and differentiable on (a, b), then there exists
In Problems 3-14, verify that the given function f satisfies the a number c in (a, b) such that
hypotheses o f the M V T on the given interval [a, b]. Then fin d fib ) = fid ) + lb - a)f'lc)
all numbers c between a and b for which
Derive this alternative form of the M V T .
W = rm t
25. Let u and v be any two numbers between —J and / Use
b- a
the M V T to show that
3. fix ) = 2x 2 + 1 on [0, 2] 4. f ix ) = - x 2
+ 4 on [ - 1, 0]
∣tan u - tan v∣ ≥ ∣M — v∣
5. f{x ) = x 3 + x on [1, 2] 6. f ix ) = 2x 3
- x 2 on [0, 2]
7. fix ) = x 4 + 2 on [ - 1, 2] 8. f ix ) = x 5 + 3 on [2, 4] 26. If f ix ) = on [-1 , 1], does the mean value theorem
9. fix ) = V x on [1,4] 1 0 .∕( x ) = ^ = o n [ l ,4 ] apply? Why or why not?
11. f ix ) = on [0, 2] 1 2 . ∕( x ) = l + ± o n [ l ,4 ] 27. If gl×) = ∣x∣ on [—2, 2], does the mean value theorem
apply? Why or why not?
13. f ix ) = cosx on [θ, 14. f ix ) = sinx + cosx on 28. T H IN K T A N K P RO BLEM I S it true that
[0, 2τr]
∣cosx - cosy∣ ≤ ∣x - y∣
Decide whether Rolle's theorem can be applied to f on the for all x and y? Either prove that the inequality is always
interval indicated in Problems 15-23. valid or find a counterexample.
16. f ix ) = tanx on [0, 2τr] 29. T H IN K T A N K P RO BLEM Consider
/ W = { -1 ifx ≥ 0
ifx < 0
f i x ) = 0 for all x in the domain, b u t/is not a constant.

Does this example contradict the zero derivative theor
em? Why or why not?
30. a. Let n be a positive integer. Show that there is a number
c between 0 and x for which
(1 + x)" - 1
= w(l + c)n 1
18. f ix ) = ∣x∣ - 2 on [0, 4]
b. Use part a to evaluate a < x < b, show that there is at least one level of
production between a and b where marginal cost equals
(1 + x)" - 1
lim marginal revenue—that is, C'(x) = R'(x).
x →0
38. Use Rolle’s theorem to show that the equation
31. a. Show that there is a number h between 0 and x for
tanx = 1 —x
which
has at least one solution for 0 < x < 1.
cosx - 1 = (-sin h)x
39. Use the MVT to show that
b. Use part a to evaluate
V l + x < 4 + λ' / 15 if x > 15
lim co s* - I 0
.r→0 X
Hint. Let ∕( x ) = V l + x and observe th a t/ ( 15) = 4.
32. Evaluate 40. Use the MVT to show that
lim COSX + 1
x - π —
2x -+
—1 > 5- + 2—5 l(2 - xj)
33. Let∕( x ) = 1 + i . If a and b are constants such that a < 0 ifθ ≤ x ≤ 2.

© 41. L e t∕(x ) = tanx. Note that
and b > 0, show that there is no number w between a and
b for which ∕(τr) = /(0) = 0
∕(⅛) - ∕(α ) = f'(w )(b - a) Show that there is no number w between 0 and τr for
which f'(w ) = 0. Why does this fact not contradict the
34. Show that for any x > 4, there is a number u-, between 4 MVT?'
and x such that
42. Use Rolle’s theorem or the MVT to show that there is no
V χ -2 = 1 number a for which the equation
x - 4 2√vv
x 3 - 3x + a = 0
Use this fact to show that if x > 4, then
has two distinct solutions in the interval [0, 1].
< 43. If a > 0 is a constant, show that the equation
2 + 4
x 3 + ax - 1 = 0
35. Show that if an object moves along a straight line in such a
way that its velocity is the same at two different times has exactly one real solution. Hint-. Let ∕( x ) = x 3 +
(that is, v(Zl ) = v(L) for t l ≠ t2), then there is some inter ax - 1 and use the intermediate value theorem to show
mediate time when the acceleration is zero. that there is at least one root. Then assume there are two
roots, and use Rolle’s theorem to obtain a contradiction.
36. Two radar patrol cars are located at fixed positions 6 mi
apart on a long, straight road where the speed limit is 55 44. If a > 0 and n is a positive integer, use Rolle’s theorem
mi∕hr. A sports car passes the first patrol car traveling at or the MVT to show that the polynomial
53 mi∕hr, and then 5 min later, it passes the second patrol
p(x) = x 2"+1 + ax + b
car going 48 mi∕hr. Use the MVT to show that at some
time between the two clockings, the sports car exceeded can have at most one real root for constants a and b. Hint:
the speed limit. See the hint for Problem 43.
45. Suppose/and g are differentiable functions on the closed
interval [a, b] and that
∕(α ) = g(a) and ∕(⅛) = g(b)
Prove that there is a number c between a and b for which
37. The total profit from sales of a commodity is given by ∕'( c ) = ∕( c )
P(x) = Λ(x) - C(x) 46. Show that if∕" ( x ) = 0 for all x, th e n /is a linear function.
where R(x) is the revenue and C(x) is the cost associated (That is ,∕(x ) = A x + B for constants A ≠ 0 and 5.)
with a production level of x units. Suppose that a and b 47. Show that if ∕'( x ) = A x + B for constants A ≠ 0 and B,
are equally profitable levels of production—that is, then ∕( x ) is a quadratic function. (That is, ∕( x ) = ax 2
P(α) = P{b). If R{x) and C(x) are both differentiable for + bx + c for constants a, b, and c, where a ≠ 0.)
Sec. 3.3 First-Derivative Test 183
3 .3 FIRST-DERIVATIVE TEST
IN THIS SECTION Increasing and decreasing functions, the first-derivative

test, curve sketching with the first derivative
The sign of the derivative of a function can be used to determine whether the
function is increasing or decreasing on a given interval. We shall use this
information to develop a procedure called the first-derivative test for classifying
a given critical point as a relative maximum, a relative minimum, or neither.
Finally, we shall use these ideas to begin the development of a procedure for
sketching the graph of a function; this procedure will be refined and extended in
the next three sections.
■ INCREASING AND DECREASING FUNCTIONS
Suppose an ecologist has determined the size of a population of a certain species

as a function f of time t (months). If it turns out that the population is increasing
until the end of the first year and is decreasing thereafter, it is reasonable to expect
the population to be maximized at time t = 12 and for the population curve to
have a high point at t = 12, as shown in Figure 3.13. If the graph of a function/
such as this population curve, is rising throughout the interval 0 < x < 12 (and
Figure 3.13 A population curve never flattens out on that interval), we say that f is strictly increasing on that
interval. Similarly, the graph of the function in Figure 3.13 is strictly decreasing on
the interval 12 < t < 20. These terms may be defined more formally as follows:
Strictly Increasing The function/is strictly increasing on an interval 7 if
Strictly Decreasing ∕( x 1) < J∖x fi whenever x l < x 2

for x 1 and x 2 on I. Likewise,/is strictly decreasing on I if
∕( x 1) > ∕ ( Λ^2) whenever x 1 < x 2
for x l and x 2 on I.
Figure 3.14 Increasing and decreasing functions
A function / is said to be (strictly) monotonic on an interval I if it is either

strictly increasing on all of I or strictly decreasing on I. Monotonic behavior is
closely related to the sign of the derivative ∕ ' . In particular, if the graph of a
function has tangent lines with only positive slopes on I, the graph will be tilted
upward and f will be increasing on I (see Figure 3.15). Because the slope of the
tangent at each point on the graph is measured by the derivative f ' , it is reasonable
to expect f to be increasing on intervals where f ' > 0. Similarly, it is reasonable
to expect f to be decreasing on an interval where f ' < 0. These observations are
established formally in Theorem 3.7.
Figure 3.15 The graph is rising where f ' > 0 and falling where f ' < 0.
Notice that the small flags indicate the slopes at various points on the
graph.
THEOREM 3.7 Increasing and decreasing function theorem

Let f be differentiable on the open interval (a, ⅛). Then the function f is strictly
increasing on (a, b) if
f'(x ) > 0 for a < x 0
throughout the interval. The strictly decreasing case is similar and is left as an
exercise for the reader.
Suppose f'(x ) > 0 throughout the interval (a, b). and let x j and x 2 be two
numbers chosen arbitrarily from this interval, with x i < x 2 . The MVT tells us
th a t ∕(⅞ ) - ∕(* ,) f ,. . f( χ f l ∖ f'( V A
—x — = f (c ) o r ∕ ( x >) ~ ∕ ( x ι) = f (c )(x 2 - x ι)
x
for some number c between x 1 and x 2. Because both∕'( c ) > 0 and x 2 - x l > 0, it
follows that ∕'(c)(x 2 - x l ) > 0, and therefore
∕(⅞ ) - ∕( * 1) > 0 or ∕( x 2) > ∕( x l )
That is, if x i and x 2 are any two numbers in (a, b) such that x l < x 2, then
∕( x 1) < ∕( x 2), which means th a t/is strictly increasing on (a, b). ■
To determine where a function f is increasing or decreasing, we begin by
finding the critical values (where the derivative is zero or does not exist). These
values divide the x-axis into intervals, and we test the sign of f ' in each of these
intervals. Finally, if∕'( x ) > 0 in an interval, th e n /is increasing in that entire
interval; and if∕'( x ) < 0 in an interval, then/is decreasing in that same interval.
This procedure is illustrated in the following example.
EXAM PLE 1 Finding intervals on which a function is increasing or
Determine where the function defined by ∕(x ) = x 3 - 3x2 - 9x + 1 is strictly

increasing and where it is strictly decreasing.
Solution First, we find the derivative:

∕'( x ) = 3χ 2 - 6x - 9 = 3(x + 1)(x - 3)
Next, we determine the critical values: ∕'( x ) exists for all x and ∕'( x ) = 0 at
x = -1 and x = 3 (by inspection). These critical values divide the x-axis into
three parts, as shown in Figure 3.16a, and we select a typical number from
each of these intervals. For example, we select —2, 0, and 4, evaluate the
derivative at these values, and mark each interval as increasing ( ↑ ) or
decreasing ( 1 ), according to whether the derivative is positive or negative,
respectively. This is shown in Figure 3.16b.
a. Intervals where/is increasing and where it is decreasing b. The graph of/along with the sign graphs fo r/a n d /'
Figure 3.16 ∕(x) = x 3 - 3x2 - 9x + I
The function / i s increasing for x < -1 and for x > 3 ; / i s decreasing for
-1 < x < 3. —
EXAMPLE 2 Comparison of the graphs of a function and its derivative
Graph ∕(x ) = x 3 - 3x2 - 9x + 1 and f'(x ) = 3x2 - 6x - 9 and compare.
a. When / ' is positive, what does that mean in terms of the graph of /?
b. When the graph o f / i s decreasing, what does that mean in terms of the
graph o f/'?
Solution The graphs of/ and ∕ '

are shown in Figure 3.17.
0 Notice that critical values o f /
are intercepts for the graph of
a. When/' is positive,/is increasing.
b. When/’is negative,/is decreasing.
Figure 3.17 Graphs of/and ∕ '

■ THE FIRST-DERIVATIVE TEST
Every relative extremum is a critical point. However, as we saw in Section 3.1, not
every critical point is necessarily a relative extremum. If the derivative is positive
to the left of a critical value and negative to its right, the graph changes from
increasing to decreasing and the critical point must be a relative maximum, as
shown in Figure 3.18a. If the derivative is negative to the left of a critical value and
positive to its right, the graph changes from decreasing to increasing and the
critical point is a relative minimum (Figure 3.18b). However, if the sign of the
derivative is the same on both sides of the critical value, then it is neither a relative
maximum nor a relative minimum (Figure 3 .18c). These observations are summa
rized in a procedure called the first-derivative test for relative extrema.
Figure 3.18 Three patterns of behavior near a critical value
The First-Derivative Test for Step 1. Find all critical values off. That is, find all numbers c such that ∕( c ) is
Relative Extrema defined and either∕'( c ) = 0 or∕'( c ) does not exist.
Step 2. Classify each critical point (c, ∕(c )) as follows:
↑ ∖ a. The point (c, f(c)) is a relative maximum if ∕'( x ) > 0 (rising)
/ ∖ for all x in an openinterval (a, c) to the left of c, and
∕'( x ) < 0 (falling) for all x in an open interval (c, b) to the
, A right of c.
∖ / b. The point (c, ∕(c )) is a relative minimum i f ∕'( x ) < 0 (falling)
for all x in an open interval (a, c) to the left of c, and
t t f'( x ) > 0 (rising) for all x in an open interval (c, ⅛) to the
/ I right of c.
∖ ∖ c. The point (c,∕(c)) is not an extremum if the derivative ∕'( x ) has
ξ the same sign in open intervals («, c) and (c, ⅛) on both
sides of c.
Suppose we apply this first-derivative test to the polynomial

∕( x ) = x 3 —3x 2 - 9x + 1
In Example 2 we found that this function has the critical values - 1 and 3 and that
f is increasing when x < - 1 and x > 3 and decreasing when -1 < x < 3 (see
the arrow pattern above). The first derivative test tells us there is a relative
maximum at -1 ( ∕ , ∖ pattern) and a relative minimum at 3 (∖ 7 pattern).
EXAM PLE 3 Relative extrema using the first-derivative test
Find all critical values of g(t) = t - 2 sin t for 0 ≤ t ≤ 2τr, and determine
whether each corresponds to a relative maximum, a relative minimum, or
neither.
Solution Because g'(t) = 1 - 2 cos t exists for all t, the only critical values occur
when g'(t) = 0; that is, when cost = ∣. Solving, we find that the critical values
for g(f) on the interval [0, 2τr] are f and ⅞p.
Next, we examine the sign of g ∖t). Because g'(t) is continuous, it is enough
to check the sign of g'{t) at convenient numbers on each side of the critical
values, as shown in Figure 3.19a. Notice that the arrows show the increasing
and decreasing pattern for g. According to the first-derivative test, there is a
relative minimum at y and a relative maximum at ⅞f. The graph of g is shown in
Figure 3.19b. Note that g(y) = j - V 3 ≈ -0 .6 8 and <g(¾r) = ⅝r + V3 ≈ 6.97.
Figure 3.19 The first-derivative a. Intervals where g(t) = t - 2 sin t b. The graph of g(t) = t - 2 sin t
test for g(t) = t - 2 sin t is increasing or decreasing for 0 < t < 2π
■ CURVE SKETCHING WITH THE FIRST DERIVATIVE
The ability to draw a quick and accurate sketch is one of the most important skills
Computational Window ∣
you can learn in mathematics. It has been said that Rene Descartes revolutionized
Do not make the mistake of thinking when he introduced the coordinate system. Today, nearly every aspect of
saying, “I have a graphing cal mathematics, engineering, physics, industry, education, and the social sciences is
culator, so I don’t need to learn enhanced by the use of graphs.
this material.” On the con In calculus, we use graphs to help us understand analysis, and we use analysis
trary, graphing calculators and to help us understand graphs. Plotting points, although useful in introducing a
computer software work hand
curve, is a very poor method for drawing a graph. You also cannot rely solely on
in hand with calculus to affirm
that the sketch is accurate and
graphing technology to do your graphing. As we have seen, there are many
complete. In Section 3.1 (Prob properties of curves that require analysis or calculus, and those properties could
lem 14) we saw calculator diffi easily be overlooked if we relied only on graphing technology. Software and
culties when using a calculator graphing calculators are helpful but incomplete ways of visualizing mathe
to graph y = x 2z3(5 - 2x) for matics. For example, the relatively simple function y = sin (5f⅛ is difficult to
negative values. sketch using most graphing calculators. We now consider some graphing tech
niques that should help you to sketch curves quickly and accurately.
The First-Derivative Step 1. Compute the derivative f ∖x).

Procedure for Sketching the Step 2. Find the critical values: values of x for which ∕ -(x) is defined but ∕
1(x) is
Graph of a Continuous zero or undefined. Substitute each critical value into f(x ) to find the y-
Function coordinate of the corresponding critical point. Plot these critical points on a
coordinate plane.
Step 3. Determine where the function is increasing or decreasing by checking
the sign of the derivative on the intervals whose endpoints are the critical
values found in step 2.
Step 4. Sketch the graph so that it increases on the intervals where f'( x ) > 0,
decreases on the intervals where f'( x ) < 0, passes through the critical
points, and has a horizontal tangent where f ∖x) = 0.
EXAM PLE 4 Sketching the graph of a polynomial using the first-derivative

procedure
Sketch the graph of∕(x ) = 2x 3 + 3x2 - 12% - 5.

Solution We begin by computing and factoring the derivative:
∕'( x ) = 6x2 + 6x - 12 = 6(x + 2)(x - 1)
We see that f'(x ) exists for all x, and from the factored form of the derivative
we see th a t/'(x) = 0 when x = - 2 and when x = 1. The corresponding critical
points are found as follows:
/( - 2 ) = 2(-2 ) 3 + 3(-2) 2 - 12(-2) - 5 = 15; critical point (-2 , 15)
/(1) = 2(1)3 + 3(1)2 - 12(1) - 5 = - 12; critical point (1, - 12)
Next, to find the intervals of increase and decrease of the function, we plot
the critical numbers on a number line and check the sign of the derivative at
values to the left and right o f - 2 and 1. We find that / i s increasing and
decreasing, as indicated in Figure 3.20.
The arrow pattern in Figure 3.20 suggests that the graph o f / has a relative
maximum at (-2 , 15) and a relative minimum at (1,-12). We begin the
sketch by plotting these points on a coordinate plane. We put a “cap” at the
relative maximum (—2, 15) and a “cup” at the relative minimum (1, -12), as
shown in Figure 3.2la. Finally, we note that because/(0) = -5 , the graph has
its y-intercept at (0, -5). We complete the sketch by drawing the curve so that
it increases for x < - 2 to the relative maximum at (-2 , 15), decreases for
- 2 < x < 1, passing through the j-intercept (0, -5 ) on its way to the relative
Figure 3.20 Intervals of increase minimum at (1, -12), and then increases again for x > 1. The completed
and decrease for graph is shown in Figure 3.2lb.
∕( x ) = 2x 3 + 3x 2 - 12x - 5
The graphs of most of the func

tions in this section can be
checked by drawing them with
a graphing calculator. Use
(GRAPH) and (ZOOM).
Figure 3.21 The graph of∕( x ) = 2x 3 + 3x 2 - 12% - 5
EXAM PLE 5 Sketching the graph of a trigonometric function using the

first-derivative procedure
X m in= - 3 V n in = ^2Θ Sketch the graph of∕(x ) = cos2x + cosx on [0, 2ττ].
Xnax=3 Vnax=20
X s c l= .5 Y s c l= 5 Solution The derivative o f/is
∕'( x ) = - 2 cosx sinx - sinx
Because ∕'( x ) exists for all x, we solve ∕'( x ) = 0 to find the critical values:
- 2 cosx sinx - sinx = 0
sinx(2 cosx + 1 ) = 0
sinx = 0 or cosx = -⅛
The critical values on the interior of the interval [0, 2π] are x = ττ, and
Because∕(⅛ r j = ~⅝∕(τr) = 0 and∕(⅛ ) = - ∣ , the corresponding critical points
are (¾r, - ∣ ) , (ττ, 0), and (⅛, - ∣ ) .
Next, we plot the critical values on a number line and determine the sign
of the derivative∕'( x ) in each interval, as indicated in Figure 3.22.
The arrow pattern in the diagram indicates that we have relative minima
at (⅞f-, - ∣ ) and (⅛∕ - - ) , and a relative maximum at (ττ, 0). We plot these
points on a coordinate plane and place “ cups” at the relative minima and a
“ cap” at the relative maximum. Finally, we sketch a smooth curve through
these points, obtaining the graph shown in Figure 3.23. If the curve is defined
over a particular interval, then you should also plot the endpoints.
∕( x ) = cos2x + cosx
Figure 3.23 The graph of

∕( x ) = cos2x + cosx
EXAM PLE 6 Sketching a graph using the first-derivative procedure
Sketch the graph o f∕( x ) = x 1z3(x - 4).

Solution Rewrite the given function as ∕( x ) = x 4z3 - 4x ιz3 and differentiate:
∕'( x ) = f x 1z3 - jx ^ 2z3 = jx ^ 2z3(x - 1)
We see that ∕'( x ) does not exist when x = 0 and that ∕'( x ) = 0 only when
x = 1. Because∕(0) = 0 and/(1) = - 3 , the critical points are (0, 0) a n d (l, -3 ).
We plot the critical values on a number line and determine the sign of
Figure 3.24 Intervals of increase and decrease for ∕( x ) = x lz3(x - 4)
We see that there is a relative minimum at (1, -3 ) and neither kind of

extremum at (0, 0). We plot the critical points on a coordinate plane with a
“ cup” at (1, -3 ) , as shown in Figure 3.25a. Finally, we pass a smooth curve
0 In Section 2.1, we showed through these points, obtaining the graph shown in Figure 3.25b. Note that at
that the derivative does not exist (0, 0) the curve looks like it has a vertical tangent.
at a vertical tangent, and you
might wish to verify that the
derivative does not exist at
x = 0. On the other hand, you
must be careful; just because the
derivative does not exist does
not mean that it must have a
vertical tangent. It may have a
corner or cusp, for example. 0
PROBLEM SET 3 .3
θ 1. ■ What Does This Say? What is the first-derivative In Problems 9-12 draw the graph o f a function whose deriva
test? tive matches the graph shown.
2. ■ What Does This Say? What is the relationship be
tween the graph of a function and the graph of its
derivative?
In Problems 3-4, identify which curve represents a function

y ~ f( χ ) a n d which curve represents its derivative y = f ' (x).
Draw a curve that represents the derivative o f the function

defined by the curves shown in Problems 5-8.
In Problems 13-34,
a. find the critical values;
b. show where the function is increasing and where it is
decreasing;
c. plot each critical point and label it as a relative maximum, a
relative minimum, or neither;
d. sketch the graph.
13. f(x ) = x 3 + 3x 2 + 1
14. f ix ) = ∣x 3 - 9x + 2
15. fix ) = Λ-3 + 35x 2 - 125x - 9,375
16. f ix ) = x 5 + 5x 4 - 550x 3 - 2,000.v2 + 60,000x
17. fix ) = x 5 - 5x 4 + 100
18. ∕ w =x + i ι 9 . ∕ ( λ- ) = ^ l 20. ∕ W =2⅛L

21. fit) = It + 1)2 (∕ - 5) 22. fit) = (2t - l) 2(t 2 - 9)
23. f( x ) = (x - 3)(x - 7)(2x + 1) 24. ∕( x ) = 16(x - 3)2(x - 7) 47. Use calculus to prove that the relative extremum of the
25. ∕( x ) = V x 2 + 1 26. ∕( x ) = V x 2 - 2x + 2 quadratic function
2z3
27. g(x) = x (2x - 5) 28. g(x) = x lz3V x + 15 ∕( x ) = (x - p)(x - q)
29. f(θ) = 2 cos θ - θ on [0, 2π]
occurs midway between its x-intercepts.
30. c(θ) = θ + cos2θ for 0 ≤ t ≤ π
48. The formula
31. ZL(X) = tan 2 x for —J ≤ x ≤ J J _ τ I sin g ⅝
32. ∕( x ) = 9 cosx —4 cos2 x on [0, π] 0∖ θ J
33. ∕( x ) = sin(x∕50π) on [0, 1000] 34. ∕( x ) = (x - 2)3
with I a positive constant, occurs in physics in connec
35. ■ What Does This Say? For n a counting number, tion with the study of Fraunhofer diffraction. (If θ = 0,
consider the graph of∕( x ) = (x - 2)''. What is the effect of then I = Z .) Assuming the graph of I is continuous, sketch
n on these graphs? the graph for [-3 π , 3τr].
In Problems 36-39, determine whether the given function has a 49. At a temperature of T (in degrees Celsius), the speed of
relative maximum, a relative minimum, or neither at the given sound in air is given by
critical value(s). v = v0 y 1 + 2⅛Γ
36. ∕( x ) = (x 3 - 3x + 1)7 at x = 1; x = -1
37. ∕( x ) = (x 4 - 4x + 2) 5 at x = 1 where v0 is the speed at 0 oC. Sketch the graph of v for
38. ∕( x ) = (x 2 - 4)4(x 2 - 1)3 at x = 1; x = 2 T > 0.
39. ∕( x ) = ΛVZX 3 - 48x at x = 4 50. Find constants a, b, and c that guarantee that the graph of
θ 40. S uppose/is a differentiable function with derivative
∕( x ) = x 3 + ax 2 + bx + c
∕'( x ) = (x - l) 2(x - 2)(x - 4)(x + 5)4
will have a relative maximum at (-3 , 18) and a relative
Find all critical values o f/ and determine whether each minimum at (1, —14).
corresponds to a relative maximum, a relative minimum, 51. Find constants A, B, C, and D that guarantee that the
or neither. graph of
41. S uppose/is a differentiable function with derivative
∕( x ) = 3x 4 + A x 3 + B x 2 + Cx + D
_ (2 χ -l)(x + 3 )
(χ -l)2 will have horizontal tangents at (2, 1) and (0, -15).
Classify each critical point as a relative maximum, a
Find all critical values of / a n d determine whether each
relative minimum, or neither.
corresponds to a relative maximum, a relative minimum,
or neither. 52. Let
∕( x ) = (x - ∕l)'"(x - /?)"
42. Sketch a graph of a function/that is differentiable on the
interval [—1, 4] and satisfies the following conditions: where A and B are real numbers and m and n are positive
i. The function / i s decreasing on (1, 3) and increasing integers with m > 1 and n > 1. Find the critical values
elsewhere on [ - 1, 4]. of / and classify each as a relative maximum, a relative
minimum, or neither.
ii. The largest value o f / i s 5 and the smallest is 0.
53. Let
Hi. The graph o f / has relative extrema at (1, 4) and (3, 1). ∕( x ) = x ιz3V.4x + B
43. Sketch a graph of a function f that satisfies the following
conditions: where A and B are positive constants. Find all critical
i. f'( x ) > 0 when x < —5 and when x > 1. values of /
ii. f'( x ) < 0 when - 5 < x < 1.
Hi. / ( - 5 ) = 4 and/(1 ) = -1 . Computational Window
44. Sketch a graph of a function / that satisfies the following
conditions: 54. Given th at∕( x ) = x 3 - cosx, fin d ∕'(x ) and then use
i. f'( x ) < 0 when x < —1. the bisection method, the Newton-Raphson method,
or a calculator to find all real critical values of/ cor
ii. f'( x ) > 0 when -1 < x < 3 and when x > 3.
rect to the nearest hundredth. Sketch the graph of /
/77. ∕ ' ( - l ) = 0 a n d ∕'(3 ) = 0 .
55. Given that ∕( x ) = x 5 + sinx 2, find ∕'( x ) and then use
45. Find constants a, b, and c such that the graph of the bisection method, the Newton-Raphson method,
∕(x ) = ax 2 + bx + c has a relative maximum at (5, 12) or a calculator to find all real critical values o f/c o r
and crosses the y-axis at (0, 3). rect to the nearest hundredth. Sketch the graph of /
46. Use calculus to prove that the relative extremum of the
quadratic function θ 56. Suppose ∕( x ) and g(x) are both continuous for all x and
y = ax 2 + bx + c (a ≠ 0) differentiable at x = c. Show that if c is a critical value for
occurs at x = -⅛∕(2α). b o th /a n d g, then it is also a critical value for the product
function 7g∙. If a relative maximum occurs at c for both f interval [a, b], then/is strictly decreasing on [a, b].
and g, is it true that a relative maximum o ffg occurs at c? T H IN K T A N K P ROBLEMS : Problems 61-67 are statements
57. Show that if f(x ) is strictly increasing on an open interval about continuous functions. In each case, either show that the
/where it is differentiable, then f'{ x ) ≥ 0 for all x in I. statement is generally true or fin d a counterexample.
Hint. You may use the fact that if ∕( x ) > 0 for every x in 61. If/and g both have a relative maximum at c, then so does
an interval except perhaps for c, and if lim ∕( x ) exists, f+ g -
then lim ∕( x ) ≥ 0. 62. I f/ and g are both differentiable and / and g both have a
58. Let/and g be strictly increasing functions on the interval relative extremum at c, then ∕'( c ) = g'(c) = 0.
[a, b]. 63. If/ is strictly increasing for x > c and strictly decreasing
a. If f(x ) > 0 and g(x) > 0 on [a, ⅛], show that the for x < c, then ∕'( c ) = 0.
product fg is also strictly increasing on [a, b]. 64. If ∕'( 0 ) = 0 , ∕ ' ( - ⅛) < 0, and ∕' ( ∣ ) > 0, then / has a
b. If ∕( x ) < 0 and g(x) < 0 on [a, ⅛], is fg strictly relative minimum at (0,/(0)).
increasing, strictly decreasing, or neither? Explain. 65. If f ' is nonnegative on an interval /, then / is strictly
59. a. Suppose f is strictly increasing for x > 0. Let n be a increasing on I.
positive integer and define g(x) = x"∕(x) for x > 0. If 66. I f / i s strictly decreasing on an interval /, then - / is
∕( x ) > 0 on an interval [a, ⅛] with a > 0, show that g strictly increasing on I.
is strictly increasing on the same interval. 67. If/and g are each strictly increasing on an interval /, then
b. Let g(x) = x n sin x for 0 < x < J , where n is a posi / - g is also strictly increasing on I.
tive integer. Explain why g is strictly increasing on this 68. J OU RN AL PROBLEM : Mathematics Magazine.* Give an
interval. elementary proof that
60. Complete the proof of the increasing and decreasing ∕( x ) = J — - i > 0 < x ≤ f
j v , sin x x 2
function theorem (Theorem 3.7). Specifically, show that if
/ is differentiable on (a, h) and ∕'( x ) < 0 throughout the is positive and increasing.
3 . 4 CONCAVITY AND THE SECOND-DERIVATIVE TEST
IN THIS SECTION Concavity, inflection points, curve sketching with the

second derivative, the second-derivative test for relative extrema
Knowing where a given graph is rising and falling gives only a partial picture of
the graph. For example, suppose we wish to sketch the graph of ∕(x ) = x 3 +
3x + 1. The derivative∕'(x) = 3x2 + 3 is positive for all x so the graph is always
rising. But in what way is it rising? Each of the graphs in Figure 3.26 is a
possible graph o f / but they are quite different from one another. Later in this
section we shall show that the correct graph is the one in Figure 3.26c. The
focus of our work is a characteristic of graphs called concavity that will allow us
to distinguish among graphs such as these. In addition, we shall develop a
second-derivative test that enables us to classify a critical point P of a function/
as a relative maximum or a relative minimum by examining the sign of the
second derivative ∕ " at P.
Figure 3.26 Which curve is

the graph of
∕( x ) = x 3 + 3x + 1?
*Volume 55, 1982, p. 300. “ Elementary proof’ in the question means that you should use only
techniques from beginning calculus. For our purposes, you simply need to give a reasonable argument
to justify the conclusion.
Sec. 3.4 Concavity and the Second-Derivative Test 193
■ CONCAVITY
A portion of graph that is cupped upward is called concave up, and a portion that
is cupped downward is concave down. Figure 3.27 shows a graph that is concave
up between A and C and concave down between C and E. At various points on the
graph, the slope is indicated by “flags,” and we observe that the slope increases
from A to C and decreases from C to E. This is no accident! The slope of a graph
increases on an interval where the graph is concave up and decreases where the
graph is concave down.
Conversely, a graph will be concave up on any interval where the slope is
increasing and concave down where the slope is decreasing. Because the slope is
found by computing the derivative, it is reasonable to expect the graph of a given
function/to be concave up where the derivative f ' is strictly increasing. According
to the increasing and decreasing function theorem (Theorem 3.7), this occurs
Figure 3.27 The slope of a graph when ( ∕') ' > 0, which means that the graph o f/is concave up where the second
increases or decreases, depending derivative f " satisfies f " > 0. Similarly, the graph is concave down where/" < 0.
on its concavity. We use this observation to define concavity.
Concavity The graph of a function / is concave upward on any open interval I where
∕" (x ) > 0, and it is concave downward where f" (x) < 0.
EXAMPLE 1 Concavity for a polynomial function
Find where the graph of ∕(x ) = x i + 3x + 1 is concave up and where it is

concave down.
Solution We find that ∕'( x ) = 3x2 + 3 and ∕" (x ) = 6x. Therefore, ∕" (x ) < 0 if
x < 0 and ∕" (x ) > 0 if x > 0, so the graph o f/is concave down for x < 0
and concave up for x > 0. Using this information, we can now answer the
question asked in Figure 3.26—the correct graph is c. c=ι
■ INFLECTION POINTS
In Figure 3.27, notice that the graph changes from concave up to concave
down at the point C. It will be convenient to give a name to such a transition
point.
Inflection Point Suppose the graph of a function / has a tangent line (possibly vertical) at the
point P(c,∕(c)) and that the graph is concave up on one side of P and concave
down on the other side. Then P is called an inflection point of the graph.
Returning to Example 1, notice that the graph of ∕(x ) = x 3 + 3x + 1 has

exactly one inflection point, at (0, 1), where the concavity changes from down
to up.
Various kinds of graphical behavior are illustrated in Figure 3.28. Note that
the graph is rising on the interval [a, c l ], falling on [c1, c2], rising on [c2, c3], falling
on [c3, c4], rising on [c4 , c5], and falling on [c5, &]. The concavity is up on [pl , p 7∖
and [p3, p 4] and down otherwise.
Figure 3.28 A graph of a function showing critical values and inflection points
The graph has relative maxima at c1, c3, and c5 and relative minima at c2 and
c4 . There are horizontal tangents ( ∕ ' = 0) at all of these points except c1 and c3,
where there are sharp points called cusps (∕'(c l ) and ∕'( c 3) do not exist). There is
also a horizontal tangent at p l ( f'( p l ) = 0) but no relative extrema appear there.
Instead, we have points of inflection at p l and p 2, because the concavity changes
direction at each of these points.
In general, the concavity of the graph of f will change only at points where
f " = 0 or f " does not exist; that is, at critical values of the derivative f ' . We shall
call the number c a second-order critical value if∕" (c ) = 0 or ∕" (c ) does not exist,
and in this context an “ordinary” critical value (where f ' = 0 or f ' does not exist)
will be referred to as a first-order critical value. Inflection points correspond to
second-order critical values and must actually be on the graph off Specifically, a
number c such that f"(c) is not defined and the concavity of f changes at c will
correspond to an inflection point only if∕(c) is defined.
0 A continuous function/does not have to have an inflection point where
f " = 0. For instance, if∕(x ) = x 4, we have f" (x ) = 12x2, so ∕''(0 ) = 0, but the
graph o f /is always concave up (see Figure 3.29). 0
EXAMPLE 2 Concavity and inflection points
x 2
Discuss the concavity off{x) = f~τ= + sinx on the interval [0, f] and find all
Figure 3.29 Th ci graph of f(x) = the inflection points of/
x 4 has no inflection point at Solution Find f ' and f ' :
(0, 0) even though ∕" (0 ) = 0.
f'(x ) = + cosx ∕" (x ) = 4 = - sinx
Both ∕'( x ) and ∕" (x ) are defined for all x on the interval [0, J] and f" (x) = 0
when sinx = Thus, the only possible location for an inflection point is
where x = f . Testing to the left and right o ff we see that the concavity changes
(from up to down) at x = f, so a point of inflection occurs there, and this is the
only such point on the interval [0, f]. The graph o f/is shown in Figure 3.30.
Here is an example of one way inflection points may appear in practical

applications.
EXAM PLE 3 Finding the maximum rate of production
An efficiency study of the morning shift at a factory indicates that an average

worker who arrives on the job at 8:00 A.M. will have turned out Q(t) =
- t 3 + 9z2 + 12? units t hours later. At what time in the morning is the worker
performing most efficiently?
Solution The worker’s rate of production is the derivative
2?(Z) = Q∖ t) = - 3 t 2 + 18/ + 12
Assuming the morning shift runs from 8:00 A.M. until noon, the goal is to
maximize the function R(t) on the closed interval [0, 4]. The derivative of R is
/?(/) = Q∖ t) = ~f>t +18
which is zero when t = 3; this is the critical value. Using the optimization
criterion of Section 3.1, we know that the extrema of R(f) on the closed
interval [0, 4] must occur at either the interior critical value 3 or at one (or
both) of the endpoints (which are 0 and 4). We find that
Λ(0) = 12 R(3) = 39
J Λ(4) = 36
so the rate of production R(t) is greatest and the worker is performing most
efficiently when t = 3, that is, at 11:00 A.M. The graphs of the output function
Q and its derivative, the rate-of-production function R, are shown in Figure
3.31. Notice that the production curve is steepest and the rate of production is
greatest when Z= 3. M M
In Example 3, note how the rate of assembly, as measured by the slope of the
graph of the average worker’s output, increases from 0 to the inflection point I and
then decreases from I to E, as shown in Figure 3.3la. Because the point I marks
the point where the rate of production “peaks out,” it is natural to refer to I as a
point of diminishing returns. It is also an inflection point on the graph of Q.
Knowing that this point occurs at 11:00 A.M., the manager of the factory might be
able to increase the overall output of the labor force by scheduling a break near
this time. (Remember that the domain for Example 3 is [0, 4].)
■ CURVE SKETCHING WITH THE SECOND DERIVATIVE
In Section 3.3, we learned how to find where a graph is rising and where it is
falling by examining the sign of the first derivative. Now, by examining the sign of
the second derivative, we can determine the concavity of the graph, which gives us
a more refined picture of the graph’s appearance.
EXAM PLE 4 Graph of a polynomial function
Determine where the function f(x ) = x 4 - 4x3 + 10 is increasing and decreas

ing and where its graph is concave up and concave down. Find the relative
extrema and inflection points and sketch the graph of f.
Solution The first derivative,
b. Rate of production
Figure 3.31 Graph of a produc

∕'( x ) = 4x3 - 12x2 = 4x2(x - 3)
tion curve showing the point of is zero when x = 0 and when x = 3. Because 4x2 > 0 for x ≠ 0, we have
diminishing returns ∕'( x ) < 0 for x < 3 (except for x = 0) an d ∕'(x ) > 0 for x > 3. The pattern
showing w here/is increasing and where it is decreasing is displayed in Figure
3.32a.
4x 2 ( x - 3 ) : Negative 0 Negative 3 Positive 12x(x- 2): Positive 0 Negative 2 Positive

a. First derivative shows intervals b. Second derivative shows concavity
of increase and decrease. pattern for the graph.
Figure 3.32 First and second derivatives o f∕(x ) = x 4 - 4x 3 + 10
Next, to determine the concavity of the graph we compute

∕''(x ) = 12x2 - 24x = 12x(x - 2)
If x < 0 or x > 2, th en ∕''(x ) > 0 and the graph is concave up. It is concave
down when 0 < x < 2, because∕" (x ) < 0 on this interval. The concavity of
the graph o f/is shown in Figure 3.32b.
The two diagrams in Figure 3.32 tell us that there is a relative minimum at
x = 3 and inflection points at x = 0 and x = 2 (because the second derivative
changes sign at these points).
To find the critical points and the inflection points, evaluate/at x = 0, 2,
and 3:
/(0) = (0)4 - 4(0)3 + 10= 10
/(2) = (2)4 - 4(2)3 + 10= - 6
/(3) = (3)4 - 4(3)3 + 1 0 = -1 7
Finally, to sketch the graph of/ we first place a “cup” (U) at the minimum
point (3,-1 7 ) and note that (0, 10) and (2 ,-6 ) are inflection points;
remember there is also a horizontal tangent at (0, 10). The preliminary graph
is shown in Figure 3.33a. Complete the sketch by passing a smooth curve
through these points, using the two diagrams in Figure 3.32 as a guide for
determining where the graph is rising and falling and where it is concave up
and down. The completed graph is shown in Figure 3.33b.
a. Preliminary sketch
Figure 3.33 Graphing∕(x) = x 4 - 4x 3 + 10
EXAM PLE 5 Using the first and second derivative to graph a function
The function/from Example 5 Sketch the graph o f∕( x ) = x 2z3(2x + 5).

may not be graphed correctly
Solution Find the first and second derivative and write them in factored form.
by using graphing software:
To do this, write ∕( x ) = 2x 5z3 + 5x 2z3.
∕'( x ) = 2(j)x 2z3 + 5(j)x^ ιz3 = -yx^ ιz3(x + 1)
∕" ( χ ) = τ ( i > ^ ιz3 + τ (-⅜ k ^ 4z3 = ^yx^4z3(2x - 1)
From these equations we see

∕'( x ) = 0 when x = - 1; ∕'( x ) does not exist at x = 0. The first-order critical
values are - 1 and 0.
X n in = -10 Vr*⅛in= - lθ ∕''( x ) = 0 when x = ∕''( x ) does not exist at x = 0. The second-order
X n a x= lθ Ynax=10 critical values are ∣ and 0.
Check the intervals of increase and decrease (Figure 3.34a) and of concavity
(Figure 3.34b).
Direction: Shape:
lp -1/3. ip -4/3
3 9a
jr+ 1: 2x- 1
f C 1Z3(Λ 1): 4 Z3 (2X
+ Positive -1 Negative 0 Positive f T - 1) Negative 0 Negative ∣ Positive
a. First derivative shows intervals b. Second derivative shows concavity
o f increase and decrease. pattern for the graph.
Figure 3.34 Preliminary work for graphing ∕(x ) = x 2z3(2x + 5)
The diagrams in Figure 3.34 suggest that the graph of /h a s a relative

maximum at x = - 1 . a relative minimum at x = 0, and an inflection point at
x = ⅛. We must find the ^-coordinates of these points by evaluating/:
/ (-1 ) = 3 / (0 )= 0 ∕Q ) = λ^ = ≈ 3.7 7 9 7 6 3 15
Plot and label the points (-1 , 3), (0, 0), and (0.5, 3.8) on the preliminary
graph, as shown in Figure 3.35a. Note that because the graph is concave down
on both sides of the origin and because the first derivative grows large without
bound as x → 0 from both the left and the right, there is a vertical tangent at
(0, 0). By plotting the relative extrema and the inflection point on a coordinate
plane and passing a smooth curve through these points, we obtain the graph
shown in Figure 3.35b.
EXAMPLE 6 Using the first and second derivatives to graph a

trigonometric function
Sketch the graph of T(x) = sinx + cosx on [0, 2τr].
Solution You probably graphed this function in trigonometry by adding ordi
nates. However, with this example we wish to illustrate the power of calculus
to complete the graph. Thus, we begin by finding the first and second
derivatives.
T'(x) = cosx - sinx
T''(x) = -s in x - cosx
From these equations we find the critical values (both T' and T" are defined
for all values of x):
Γ'(x) = 0 when cosx = sinx; thus, x = f and x = ∙¾r .
T"(x) = 0 when cosx = -sin x ; thus, x = ¾r and x = ⅞r .
We check the intervals of increase and decrease as well as the concavity
pattern, as shown in Figure 3.36.
a. Intervals of increase and decrease
Find the critical points, as well as the points of inflection.
Relative maximum: Γ(%) = V2; the critical point is (f, V2).
Relative minimum: τ(¾ r) = —V2; the critical point is (j¾r , -V z2).
Inflection: τ(¾ r) = 0; the inflection point is (¾r, θ).
b. Concavity pattern
Y(⅞r) = 0; Λe inflection point is (¾r, θ).
Figure 3.36 Preliminary work for
sketching T(x) = sin x + cosx Finally, find the value of T at the endpoints:
T(0) = 1 and T(2τr) = 1
This information is shown in Figure 3.37.
b. Completed sketch
Figure 3.37 Graph of T(x) = sin x + cosx on [0, 2π]
■ THE SECOND-DERIVATIVE TEST FOR RELATIVE EXTREMA
It is often possible to classify a critical point P(c, f(c)) on the graph of / b y

examining the sign of∕"(c). Specifically, suppose ∕'( c ) = 0 and ∕" (c ) > 0. Then
there is a horizontal tangent line at P and the graph of f is concave up in the
neighborhood of P. This means that the graph of / i s cupped upward from the
horizontal tangent at P, and it is reasonable to expect P to be a relative minimum,
as shown in Figure 3.38a. Similarly, we expect P to be a relative maximum if
∕'( c ) = 0 and ∕" (c ) < 0, because the graph is cupped down beneath the critical
point P, as shown in Figure 3.38b.
Figure 3.38 Second-derivative test for relative extrema
These observations lead to the second-derivative test for relative extrema.
The Second-Derivative Test for L e t/b e a function such that f'(c) = 0 and the second derivative exists on an
Relative Extrema open interval containing c.
If f'∖ c) > 0, there is a relative minimum at x = c.
If f"(c) < 0, there is a relative maximum at x = c.
Iff"(c) = 0, then the second-derivative test fails and gives no information.
EXAM PLE 7 Second-derivative test to find extrema
Use the second-derivative test to determine whether each critical value of the
function ∕(x ) = 3x5 - 5x3 + 2 corresponds to a relative maximum, a relative
minimum, or neither.
Solution Once again, we begin by finding the first and second derivatives:
∕'( x ) = 15x4 - 15x2 = 15x2(x - l)(x + 1)
∕" (x ) = 60x3 - 30x = 30x(2x2 - 1)
To apply the second-derivative test, evaluate the second derivative for the
critical values x = 0, 1, and -1 :
∕" ( 0 ) = 0; test fails at x = 0.
∕" ( 1 ) = 30; positive; test tells us that there is a relative minimum at x = 1.
∕" ( - l ) = -3 0 ; negative, so test tells us that there is a relative maximum
at x = - 1 .
When the second-derivative
test fails (in this case at x = 0),
revert to the first-derivative test:
The derivative is negative to the
right of 0 and negative to the left
of 0.
Neither kind of extremum occurs
at x = 0 ( ∖ ∖ pattern). Actually,
there is an inflection point at
x = 0 because the second deriva
tive is negative to the right of 0
and positive to the left of zero.
The graph is shown in Figure 3.39.
Example 7 demonstrates both the strength and the weakness of the second-
derivative test. In particular, when it is relatively easy to find the second derivative
(as with a polynomial) and if the zeros of this function are easy to find, then the
second-derivative test provides a quick means for classifying the critical points.
However, if it is difficult to compute∕" ( c ) or iff " ( c ) = 0, it may be easier, or even
necessary, to apply the first-derivative test.
TABLE 3.1 Summary of the First-Derivative and Second-Derivative Tests

Find the critical values c such that f(c ) is defined and either ∕'( c ) = 0 or f'(c ) does not exist.
Second-Derivative Test
If∕" ( c ) < 0, If∕" ( c ) > 0, If f" (c ) = 0, second-derivative test fails.
relative maximum at x = c. relative minimum at x = c.
f"(c) < 0 f"(c) > 0
First-Derivative Test
If the first derivative changes from If the first derivative changes from
positive to negative (left to right) at c, negative to positive (left to right) at c,
th e n /h as a relative maximum at x = c. th e n /h a s a relative minimum at x = c.
Inflection Points
There is an inflection point at c iff " (the concavity) changes sign there. This may occur if∕" ( c ) = 0 or if∕" (c ) is not defined.
PROBLEM SET 3 .4
θ 1. ■ What Does This Say? What is the second-derivative 30. T H IN K T A N K P ROBLEM Sketch the graph of a function
test? with the following properties:
2. ■ What Does This Say? What is the relationship be
∕'( x ) > 0 when x < 1
tween concavity, points of inflection, and the second
derivative? ∕' W < 0 when x > 1
3. ■ What Does This Say? The cartoon on page 196
exclaims, “ Our prices are increasing slower than any of ∕" ( x ) > 0 when x < 1
our competitors!” Restate using calculus. ∕" ( x ) > 0 when x > 1
Find all critical points o f the given function in Problems 4-27,
What can you say about the derivative o f /when x = 1?
and determine where the graph o f the function is rising, falling,
concave up, or concave down. Sketch the graph. 31. T H IN K T A N K P ROBLEM Sketch the graph of a function
4. ∕( x ) = x 2 + 5x - 3 with the following properties: There are relative extrema
at (—1, 7) and (3, 2). There is an inflection point at (1, 4).
5. ∕( x ) = 2(x + 20)2 - 8(x + 2 0 )+ 7
The graph is concave down only if x < 1. The intercepts
6. ∕( x ) = x 3 - 3x - 4 are ( -4 , 0) and (0, 5).
7. ∕( x ) = ⅜x3 - 9x + 2 32. ■ What Does This Say? Consider the graphs of
8. ∕( x ) = ( x - 12)4 - 2 ( χ - 12)3 a. ∕( x ) = (x + l ) ιz3
9. ∕( x ) = 1 + 2x + b∙
∕W = (λ^ + 1)2/3
10. f(ιi) = 3u 4 - l u 3 - 12M2 + 18M - 5
c. ∕( x ) = (x + l) 4z3
11. g(ιi) = ιι4 + 6M3 - 24ιz2 + 26
d. ∕( x ) = (x + l) 5z3
1 2 . ∕( x ) = ⅛ 13. ∕( x ) = ⅛ Generalize to make a statement about the effect of a
positive integer n on the graph o f∕( x ) = (x + l)" z3.
14. ∕( x ) = V x 2 + I 15. (0 = (z3 + 0 2
g 33. Use calculus to show that the graph of the quadratic
16. ∕(Z) = (Z3 + 3Z2) 17. ∕(Z) = Z3 - 3z4 function y = ∕l x 2 + B x + C is concave up if A > 0 and
18. /(0 = 4z5 - 5Z4 19. g(Z) = 5Z6 - 6z3 concave down if A < 0.
20. ∕W = ⅛ 21. ∕( x ) = ⅛ - 34. Find constants A , B, and C that guarantee that the
function∕( x ) = A x 3 + B x 2 + C will have a relative extre
22. ∕(Z) = 2z4z3 + 9z + 1 23. ∕( x ) = x 4z3(x - 27) mum at (2, 11) and an inflection point at (1, 5). Sketch the
24. Z(0) = sin θ - 2 cos θ for 0 ≤ θ ≤ 2π graph o f/
25. t(θ) = θ + cos 20 for 0 ≤ 0 ≤ π 35. ■ What Does This Say? Consider the graph of
y = x 3 + bx 2 + ex + d for constants b, c, and d. What
26. A
v (M
’
)= Τ fη--
2 + cos u
for 0 ≤ u ≤ 2τr happens to the graph as b changes?
27. / ( Z) = - 1 sin 2z + cos Z for 0 ≤ Z ≤ τr 36. At noon on a certain day, Frank sets out to assemble five
stereo sets. His rate of assembly increases steadily
θ 28. T H IN K T A N K P RO BLEM Sketch the graph of a function
throughout the afternoon until 4 P . M ., at which time he
with the following properties:
has completed 3 sets. After that, he assembles sets at a
∕'( x ) > 0 when x < -1 and when x > 3 slower and slower rate until he finally completes the fifth
set at 8 P . M . Sketch a rough graph of a function that
∕'( x ) < 0 when -1 < x < 3 represents the number of sets Frank has completed after Z
hours of work.
f" ( x ) < 0 when x < 2
37. The deflection o f a hardwood beam of length 6 is given by
∕" ( x ) > 0 when x > 2
Z>(x) = ∣ x 4 - 7 fx 3 + 5 f 2x 2
29. T H IN K T A N K P RO BLEM Sketch the graph of a function
with the following properties:
where x is the distance from one end of the beam. What
f'( x ) > 0 when x < 2 and when 2 < x < 5 value of x yields the maximum deflection?
38. A manufacturer estimates that if x units o f a particular
∕'( x ) < 0 when x > 5
commodity are produced, the total cost (in dollars) will be
∕'( 2 ) = 0
C(x) = x 3 - 24x 2 + 350x + 400
∕" ( x ) < 0 when x < 2 and when 4 < x < 7
At what level of production will the marginal cost C ' be
∕" ( x ) > 0 when 2 < x < 4 and when x > 7 minimized?
39. The total cost of producing x thousand units of a certain that the average worker will produce the maximum
commodity is number of blenders during the morning shift? How is
this optimum time related to the point of diminishing
C(x) = 2x 4 - 6x 3 - 12x2 - 2x + 1 returns?
Determine the largest and smallest values of the marginal @ 42. In a paper published in 1969,* C . J. Pennycuick provided
cost C ' for 0 ≤ x ≤ 3. experimental evidence to show that the power P required
40. Let/be a function that is differentiable at 0 and satisfies by a bird to maintain flight is given by the formula
/(0) = 0. If y = fix ) , suppose
dy _ 3y2 + x
dx~ y2 + 2 where v is the relative speed of the bird, w is its weight, p is
the density of air, and S and A are constants associated
a. Find an equation for the tangent line to the graph of /
with the bird’s size and shape. What speed will minimize
at the point where x = 0.
the power? You may assume that w, p, S , and A are all
b. Note that the origin is a critical point. What kind of positive.
relative extremum (if any) occurs at this point?
43. Show that if / is concave up on an interval I, then the
41. An industrial psychologist conducts two efficiency studies graph o f/lies above all its tangents on I. Hint: Use the
at the Chilco appliance factory. The first study indicates mean value theorem to show that if c is a point in I, then
the average worker who arrives on the job at 8 A.M. will
have assembled ∕( x ) > ∕( c ) + ∕'( c ) ( x - c)
—Z3 ÷ 6z2 + 13Z for all x in I. This property is sometimes used as the
definition o f concave up.
blenders in t hours (without a break), for 0 ≤ Z ≤ 4. The
44. ■ What Does This Say? Rephrase in your own words
second study suggests that after a 15-minute coffee break,
what Problem 43 is all about.
the average worker can assemble
- ∣ Z 3 + ∣Z2 + 25z
blenders in t hours after the break for 0 < t ≤ 4.
Note-. The 15-minute break is not part of the work time. In Problems 45-48, fin d f'( x ) and f"(x ), then use the
a. Verify that if the coffee break occurs at 10 A.M., bisection method, the Newton-Raphson method, or a
the average worker will assemble 42 blenders before calculator to fin d (correct to the nearest hundredth) all
the break and 49∣ blenders for the two hours after the critical values o f f and f ' . Find all inflection points and
break. sketch the graph o f f.
b. Suppose the coffee break is scheduled to begin x hours 45. ∕( x ) = x 4 - 3x 3 + x - 5
after 8 A.M. Find an expression for the total number of 46. ∕( x ) = x 3 + sinx on [-ττ, π]
blenders N(x) assembled by the average worker during 47. ∕( x ) = tanx - x 2 + 3 on [-2 π , 2-π]
the morning shift (8 A.M. to 12:15 P.M.). 48. ∕( x ) = x 5 + 2x 3 - x 2 + 11
c. At what time should the coffee break be scheduled so ....
3 .5 INFINITE LIMITS AND ASYMPTOTES
IN THIS SECTION Asymptotes, limits involving infinity, graphs with

asymptotes
In order to complete our discussion of curve sketching we need to introduce two
concepts, asymptotes and limits involving infinity.
■ ASYMPTOTES
Recall from Chapter 1 that lim∕(x ) does not exist when∕(x ) increases or decreases
without bound as x approaches c. This is one kind of limit involving infinity, and a
*“The Mechanics of Bird Migration,” Ibis III, pp. 525-556.
Sec. 3.5 Infinite Limits and Asymptotes 203
goal of this section is to examine such limits and to see how they may be used in
curve sketching.
As a preview of the ideas we plan to explore, let us examine the graph shown in
Figure 3.40.
Figure 3.40 A typical graph with asymptotes
Suppose the graph shown in Figure 3.40 is the graph of a function we will c a ll/
Notice that as x approaches 3 from either side, the corresponding functional
values of f get large without bound (in absolute value), and the graph of f
approaches the vertical line x = 3. This approach is through positive values (“up”)
as x approaches 3“ (from the left) and through negative values (“down”) as x
approaches 3 + (from the right). We indicate the behavior of f for x near 3 by
writing*
lim f(x) = +∞ and lim f(x ) = -∞
and we describe the corresponding geometric behavior by saying that the line
x = 3 is a vertical asymptote of the graph of f.
Note also that as x increases without bound (that is, as x moves toward the
right on the x-axis), the graph of/follows the line y = mx + b. For this reason, the
line y = mx + b (where m ≠ 0) is called an oblique (or slant) asymptote.
At the other end of the x-axis (as x decreases without bound), the graph
approaches the line y = 2. We write
lim ∕(x ) = 2
Λ'→ —∞
and say that the line y = 2 is a horizontal asymptote of the graph.

More specifically, an asymptote is a line having the property that the distance
from a point P on the curve to the line approaches zero as the distance from P to
the origin increases without bound, and P is on a suitable part o f the curve. This
last phrase (in italics) is best illustrated by considering Figure 3.41, where L is an
asymptote for the function /
Consider P and d, the distance from Pto the line L, as shown in Figure 3.4la.
Now, the distance from P to the origin can increase in two ways, depending on
whether P moves along the curve in direction 1 or direction 2. In direction 1, the
distance d increases without bound, but in direction 2 the distance d approaches
zero. Thus, if you consider the portion of the curve in the shaded region of Figure
3.4lb, you see that the distance decreases as P moves along the curve in a certain
direction.
There are three types of asymptotes that occur frequently enough when
sketching curves to merit our consideration: vertical, horizontal, and oblique
Figure 3.41 An asymptote to a *Many books simply write ∞ to mean +∞. For now, we use +∞ to remind you that +∞ is not the
curve same as -∞ .
(slant) asymptotes. An example of each is shown in Figure 3.42. A graph may pass
through a horizontal or oblique (slant) asymptote.
We will define the concepts formally later in this section, and we will also
examine examples of graphs with asymptotes. However, we first need to discuss
limits involving infinity.
■ LIMITS INVOLVING INFINITY
If the functional values of f get closer and closer to the number L as x increases
without bound, we say that ∕ ( Λ') approaches L as x approaches infinity and we write
lim f(x ) = L . If f(x ) approaches L as x decreases without bound, we write
Λ'→ ÷∞
lim f( x ) = L . These concepts may be defined formally as follows:
Limits to Infinity The limit statement lim f(x) - L means that for any number ε > 0, there
λ'→+∞
exists a number N l such that
∖f(x ) - L ∖ < ε whenever x > N i
for x in the domain of f . Similarly, lim f(x ) = M means that for any ε > 0,
there exists a number N 2 such that '~*
∣∕(x ) - M∖ < ε whenever x < N ,
Figure 3.43 Graphical representation of limits to infinity

With this formal definition, we can show that all the rules for limits established in
Chapter 1 also apply to lim f(x) and lim fix ) . We now state those limit rules for
JC→+∞ x→-∞
χ→+∞.
Limit Rules If lim fix ) exists, then for constants a and b:

x→+∞
Power rule: lim [∕(x )]" = Γ lim ∕( x ) l π
x→÷∞ Lx→+∞ J
Linearity rule: lim [af(x) + bg(x)] = a lim fix ) + b lim g(x)
X→+∞ χ→+∞ Λ'→+∞
Product rule: lim [∕(x)g(x)] = Γ lim ∕Yx)lΓ lim g(x)^l
x→+∞ G →+≈ Jfx→+∞ J
,., .. lim f(x )
Quotient rule: lim -τ~r = -~ -- if lim g(x) ≠ 0
x→+∞ g(x) hm g(x) x→+∞o '∙
The same results hold for lim fix ) , if it exists.
χ→— ∞
The following theorem will allow us to evaluate certain limits to infinity with
ease.
T H E O R E M 3.8 Limits to infinity
If n is a positive rational number, and A is any nonzero real number, then
lim j ⅜l = 0
X'
A → +∞
Furthermore, if x n is defined when x < 0, then
lim ^ = 0
.V→ - ∞ Λ
Proof We begin by proving that lim -. = 0. For ε > 0, let N = - . Then for
x > N we have ∖→ ≈ x
x > N = - so that - < ε
ε x
Historical Note This means that ⅛ - 0 < ε so that from the definition of limit we have
IA
BHASKARACHARYA
The possibility of division by lim ⅛= 0. Now let n be a rational number— say, n = - . Then
JC→+∞ X Cl
zero is an idea that causes spe
cial concern to mathematicians. lim 2⅜ = lim ~ 4 J - = A lim 1 P
n
One of the first recorded obser x →+ ≈>X x →+∞ χP∣<l .γ →+≈
vations of division by zero
comes from the twelfth-century
Hindu mathematician
Bhaskaracharya, who made the The second part of the proof follows similarly.
following observation: “The
fraction, whose denominator is
Example 1 illustrates how Theorem 3.8 can be used along with the other limit
zero, is termed an infinite
quantity.” Bhaskaracharya then properties to evaluate limits to infinity.
went on to give “a very beauti
ful conception of infinity” that E X A M P L E 1 Evaluating limits to infinity
involved his view of God and
creation. Evaluate
From Mathematics in India in the lim -J l χ ∑ι and lim f
Middle Ages by Chandra B. Sharma. x → +∞ y X Λ x→-∞ ∖ X /
Solution Notice that for x ≠ 0, , .

3 x -5 f o - i )s .3 - ⅞
x ^ 2 *(1~9
Also, according to Theorem 3.8, we know that
lim = lim ⅞= 0
X→ +∞ X χ → +∞ X
We now find the limits using the quotient rule, the power rule, and Theorem
3.8:
1/2 1/2
— —γ = lim 3x - 5 lil ∏
x - 2 χ-»+= -→+∞
L Λ Δ
1/2
= √3
Similarly,
lim 3x - 5 ∣3 = 33 = 27
x - 2
When evaluating a limit of the form

.. p(x) .. p(x)
hm -5τ-r
x → +∞
C l( X )
or hm -U s∖-Xx)
x → -∞
where p(x) and d(x) are polynomials, it is often useful to divide both p(x) and d(x)
by the highest power of x that occurs in either. The limit can then be found by
applying Theorem 3.8. This process is illustrated by the following examples.
E X A M P L E 2 Evaluating a limit to (positive) infinity
Evaluate ⅛ J j ,+gA97
Solution We may assume x ≠ 0, because we are interested only in very large
values of x. Dividing both the numerator and denominator of the given
expressions by x 3, we find
, , -L 7 _ A. -9
3x 3 - 5x + 9 _ 3x 3 - 5x + 9 . _ i v ++ x 3
3 2 3 2
5x + 2x - 7 5x + 2x - 7 _L3 s + 2 _ 13
x x x
Thus,
3
lim 3x3 - 5x2 + 9 lim
5x + 2x - 7
A'→ + ∞ A^→ + ∞ 7
5 + j - x3
Graph of y = ⅛ L ~ 5 -y , + 9 _ 3- 0+ 0 3
1 ∙ 5x 3 + 2x 2 - 7 5 + 0 -0 5
EXAM PLE 3 Evaluating a limit to negative infinity
Evaluate lim + 57x t 3 θ.

5
χ→-∞ xi - 1,000
Solution Dividing the numerator and the denominator by the highest power, x 5,
we find that
~ ZU I - ιu ∖j 1U zu
1
1∙ 95x 3 + 57Λ- + 30 1∙ 95x 3 + 57x + 30 x5
Λ∙→-χ x 5 - 1,000 c→-≈ x 5 - 1,000 1
9 5 ,5 7 ,3 0 „
+ +
lim iC F F -0 +0+ 0
Graph .95x3 + 57x + 30 c→-∞ [ _ 1.000 1- 0 = 0
o f
1 j x 5 - 1,000
EXAM PLE 4 A limit to infinity that does not exist

Q -γ-2 __ Q v
Evaluate xh→m
+∞ —X ,1
Solution We note that the degree of the numerator is greater than the degree of
the denominator. Intuitively, we expect that the limit does not exist because
the numerator is growing faster than the denominator. If we divide the
numerator by the denominator we find
lim — — = lim [ ⅛ + l ) + ⅛ ]
Λ-→+≈ x - 1 X→+∞
= lim (3x + 1 ) + lim . 2. ∣

X— >+∞ x —>÷∞ X 1
= +∞
Compare the graphs of f(x ) EXAM PLE 5 A limit where technology may fail
= 3 λ ' 2 ^ .2 x and y = 3x + 1.
x - 1 ,. „4 .3 5 9 , 1
Evaluate
c 1 t
hm - ----- Z7+-L
χ → +∞ v √19
Solution Consider three similar functions:

A'4359 + 1 _ x 4 359 + 1 χ 4.359 + 1
19 E? „4 .3 6 ”3 „4.359
If you begin with a calculator (table) or a graph you might obtain a faulty
conclusion. Look at the graph of these three functions in Figure 3.44. From the
graph you might be led to the incorrect conclusion that the limits as
x → +∞ for y 1, y 2 , and y 3 are all 1. Let us consider the situation more carefully.
We note √19 ≈ 4.358898944.
If we use the approximation V19 ≈ 4.36, we obtain y 2 .

I 2 3 4 5 6 7 8
„4 .3 5 9 4 . 1 / I 1 ∖
⅛ 3c 7436 = ¾ L (χ 0∙00 ∣ + √ ∙ 3V = 0 + 0 = 0
Figure 3.44 Graphs of y 1, y 2, and
Λ → +∞ A X—>^r∞ ∖A A /
y ∙ with the given scale, the three
graphs look the same. Let’s use a better approximation. With VT9 ≈ 4.359, we obtain y 3 .
„4 .3 5 9 ∣ ι . / 1 ∖
⅛ 3 „4 .3 5 9 - ⅛ } 00 ( ' + „ 4 .3 5 9 ) =
Λ → -Γ ∞ A Λ → -Γ ∞ ∖ A /
We note that our preliminary conclusion that these limits were the same was
incorrect. Let’s look at some better graphs (which we might not consider if we
relied only on the technology), as shown in Figure 3.45.
c. For large enough x, we can

Figure 3.45 Detail of Figure 3.4 make y j as large as we wish.
Instead of using a numerical approximation (which is what most calculators or

computers would use), we confirm what we begin to expect from Figure 3.45c.
r 4.359 1 ∕ v 4.359 1 ∖ ____
lim ----- -∕τ^— = lim " . TQ 3— v77∏5 = +∞ Because λ 19 < 4.359
Λ∙→ +∞ χV 19 x→ + ≈ ∖ χ V is Λ ∕
In mathematics, we use the symbol ∞ to describe the situation in Example 5 or

whenever it is necessary to indicate either the process of unrestricted growth or the
end result of such growth. With this understanding, we can speak of infinite
lim its— that is, limits that increase or decrease without bound. The limit
statement lim fix ) - +∞ may be defined formally as follows.
Infinite Limit
We write lim ∕'(x) = +∞ if for any num ber N > 0 (no m atter how large), it is
possible to find a num ber δ > 0 such that f( x ) > N whenever
0 < ∣x - c∣< δ.
What this does not say: An infinite limit does not exist in the sense
that limits were defined in Chapter 1. The symbolism lim f ( x ) = +∞ does not
mean that f ( x ) approaches a number +∞ as x approaches c. That is, ∞ is not a
number. Nevertheless, writing lim fix ) = +∞ or lim fix ) = - ∞ conveys more
x →c x →c
specific information than simply saying “the limit of f i x ) as x approaches c
does not exist.” 0
A similar definition holds for the infinite limit statement lim ∕( x ) = - ∞ (see
Problem 62, Problem Set 3.5). For example,
lim ——F = -∞ and lim — = +∞ or lim — ⅛ = + 0° and lim — ⅛ = -∞

x→ 5-X- 5 X- 5
χ → 5+ x→ -4 - X + 4 x +4
χ → -4 +
are infinite limits. The graphs of these functions are shown in Figure 3.46.
Sec. 3.5 Infinite Limits and Asymptotes 20 9
EXAMPLE 6 Infinite limit
Find lim —— and lim ——A

j →2' X - 2 χ→2+ X - 2
Solution Notice that £ increases without bound as x approaches 2 from the
r.ight A
and -Z _ 9 decreases without bound as x approaches 2 from the left. That
IS,
lim — = +∞ and lim — = —∞
z→2÷ x - 2 ,v→2- x - 2
We also have lim (3x - 5) = 1, and it follows that
lim2~ x -~x2 , = -∞
lim x -~~2 = +∞ and χ→
Λ → 2÷
■ GRAPHS WITH ASYMPTOTES
We shall now show how limits involving infinity can be used along with the curve
sketching techniques developed in the last two sections to sketch graphs with
asymptotes. We need the following definitions:
Vertical Asymptote The line x = c is a vertical asymptote of the graph o f/if either of the one-sided
limits
lim ∕(x) or lim f(x)
x→c x→c+
Horizontal Asymptote is infinite. The line y = L is a horizontal asymptote of the graph o f / i f
lim f(x) = L or lim ∕(x ) = L
X→+∞ X→-∞
Oblique Asymptote The line y = mx + b is an oblique asymptote of the graph o f / i f / i s a rational
function such that the numerator and denominator have no common factors
and
= τdτ(rx^v) = mx
j∕ ( vx )' + b + d-j(jx-)τ
w h e re ⅛ ⅛ = 0∙
EXAMPLE 7 Graphing a rational function using asymptotes
Sketch the graph of∕(x ) = y.
Solution
Vertical Asymptotes. First, make sure the given function is written in

simplified form. Because vertical asymptotes for∕(x ) = v _ 9 occur at values
of x for which/is not defined, we look for values that cause the denominator
to be zero; that is, we solve d(c) = 0 and then evaluate lim ∕(x )
and x→limr+ f(x) to ascertain the behavior of the function at x = c. For this
example, x = 2 is a value that causes division by zero, so we find

ι —5
lim 3x ~ 5 -∞ and lim ---r--- ~- = +∞
r→2^ x —2 x→2+ X —2.
(We found these limits in Example 6.) This means that x = 2 is a vertical
asymptote and that the graph is moving downward as x → 2 from the left and
upward as x → 2 from the right. This information is recorded on the
preliminary graph shown in Figure 3.47a by a dashed vertical line with
upward ( ↑ ) and downward ( J ,) arrows.
Horizontal Asymptotes. In order to find the horizontal asymptotes we

compute
3x 3*Sρ5 ■f = l i m ----- ⅛= τ∈^n = 3
jrl→
im
+∞ X ~ 22 =χ→l+im
∞x 2 1 χ→+∞ ι 2 1 (J
x x
ana1
lim _ ~ = 3 (The steps here are the same as for x → +∞.)

χ→- ∞ X Λ
This means that y = 3 is a horizontal asymptote. This information is recorded
on the preliminary graph shown in Figure 3.47a by a dashed horizontal line
with outbound arrows («—, → ).
Figure 3.47 Graph o f∕(x ) = 3 λ '_ ~
Oblique Asymptotes. This function does not have oblique asymptotes. (Do
you see why?)
The preliminary sketch gives us some valuable information about the
graph, but it does not present the entire picture. Next, we use calculus to find
where the function is increasing and decreasing (first derivative) and where it
is concave upward and concave downward (second derivative).
-1 2
∕'( x ) and ∕" ( x ) The details for finding
(x - 2) 2 (X - 2)3 these derivatives are
not shown.
Neither derivative is ever zero, and both are undefined at x = 2. Checking the
signs of the first and second derivatives, we find that this curve has no points
of inflection (the function is not defined at x = 2). This information is added
to the preliminary sketch shown in Figure 3.47a. The completed graph is
shown in Figure 3.47b. E≡⊂a
PROBLEM SET 3 .5
O Evaluate the limits in Problems 1-36. all critical points and points o f inflection. Finally, sketch the
7,000 graph.
1. lim 2,000 2. lim
χ → +∞ x + 1 VA + 1 37. ∕( x ) - 3* -+ 5
χ 38 ∙ ^ W -X +4
3. lim 3x + 5 4. lim A + 2
x - 2 x → +∞ 3x - 5 39. ∕( x ) = 4 + ξ ¾ 3 40. g(x) = x *
4 χ
5. lim 5 + / 6. lim 3^ + 4
∕ → -x 7 - t 5→ - ∞ 7 - 2s 41. ∕z(x) = x 2 + 42. ∕( x ) = χ 2 ÷ 2γ3
2
3x 2 - 7x + 5 9z5 + 50Z2
+ 800
7. lim 8. lim
A'→ + ∞ -2 x 2 + x - 9 t→ +∞ Z5 - 1,000 43∙ ^ ^ 3 -8 44. ∕( A ) - 2 x
χ 2
3x + 7
9
5
9. lim 17z + 800/
6
+ 1,000
+
t→ +∞ t - 1 45∙ ^ = χ - l÷ x + 4 * ∙ Λ > > - ∕ι ∕ 1
10. lim 100 x6 + x2 - x + 500

A'- > —∞ x7 + 2 Sketch the graphs o f the piecewise-defined functions in Prob
8x 3 - 9x + 5 lems 47-50.
11. lim . f x 2 — 3x if x ≤ 1
A'→ + ∞ x 2 + 300x 47. ∕( x ) - if χ > 1
4 3
12. lim 5,000x + 5300x + 1,000 48 = if? > 0
x + 3 ll+ Z 3 ifz ≤ 0
(2x + 5)(x - 3)
13. lim fcos Z - sin Z if Z ≤ : 0
x → -∞ ( 7 χ -2 ) ( 4 x + 1) 49. ∕(Z) = ■ 2
lcos Z ifZ > 0
(3x 2 - 10)(x2 + x)
14. lim Γx 2 - 3x ifx ≤ 1
(4x 3 —x 2 + l)(x - 7)
l(3x - 7) if 1 < X <: 3
(2x + 1)(x + 5)(x - 3) 50∙ ∕ ω =
15. lim 3x - 1
x4 + 9 if x ≥ 3
*∙ x + 5
x (x + l)(x + 2) f 8/ + 1 _ 1 ~ ∕ ∖
16. lim 17. lim © 51. T H IN K T A N K P ROBLEM Sketch the graph of a function f
X → +∞ x4 - 4 z→ +∞ ∖ 'll t ÷ 5/
with all the following properties:
l x 2x ) ∕δZ + 5 ?
18. lim 19. lim ∖3 - 2t) The graph has y = 1 and x = 3 as asymptotes;
∖x + 1 x - 1/ l →+∞
/is increasing for x < 3 and 3 < x < 5 and is decreas
/18/ + Z - 4 X
ing elsewhere;
20. lim 21. lim
t →+∞ ∖ 3t + 2z2 V x 2 + 1,000
The graph is concave up for x < 3 and for x > 7 and
3x x 5∙916 + 1 concave down for 3 < x < 7;
22. lim 23. lim
X→ -∞ √ 4 x 2 + 10 χ → +∞ x ^
/(0) = 4 = /(5) and/(7) = 2.
x 6-083 + 1 A 5'831 + 1
24. lim 25. lim 52. T H IN K T A N K P ROBLEM Sketch a graph of a function g
χ → ÷∞ Av37 x → +∞ X v34
with all of the following properties:
x 5∙744 + I A2 - 3A + 4
26. lim 27. lim The graph has no inflection points and only one critical
χ → +∞ Av33 x → -2 A+ 2
point (1, 1);
A- 1 A2 - 4x + 3
28. lim 29. lim+ lim g(x) = lim g(x) = 2; and
x→Γ x 2 - 1∣ x →3 x 2 —6x + 9 Λ '-* —c c X→ +∞
lim g(x) = lim g(x) = —∞∙,
1 x2- A+ 1
30. lim + ( o 31. lim+
x →3 ∖x —7 x - 3/ x →0 x - sinx g is increasing for x < - 1 and is decreasing only
sec x for —1 < x < 1.
32. lim 33. lim (x sin -θ
tanx - 1 ,γ→ +∞
λ ~*4 53. ■ What Does This Say? Frank Kornerkutter has put off
2
34. lim+ ~AX2 35. lim+ x cscx doing his math homework until the last minute, and he is
x →0 x - sinx x →0 1 — cosx now trying to evaluate
x - sinx
36. lim+
x →0 V sin x
At first he is stumped, but suddenly he has an idea:

Find all vertical and horizontal asymptotes o f the graphs o f the Because
functions given in Problems 37-46. Find where each graph is lim -⅛ = +∞ and lim - = +∞
rising and where it is falling, determine concavity, and locate χ→o+ X Λ x→o÷ x
it must surely be true that the limit in question has the 58. A town planning commission estimates that t years from
value +∞ - (+∞) = 0. Having thus “solved” his problem now, the population of the town will be p thousand people
he celebrates by taking a nap. Is he right, and if not, what given by the formula
is wrong with his argument? 571 + 8
54. Find constants a and b that guarantee that the graph of P(0 = 3/ + 4
the function defined by
a. What is the population of the town? What will it be in
f( ) = Q X ± 5
J ( Xx ) 3 - bx 10 years?
b. Sketch the graph of the population function p. What is
will have a vertical asymptote at x = 5 and a horizontal the “ultimate population” of the town—that is, what
asymptote at y = —3. is lim p(t)2
55. JOURNAL PROBLEM: College Mathematics Journal* by 59. a. Show that, in general, the graph of the function
Michael G. Murphy. Find the oblique asymptote of the
curve with equation f ( χ ) = a x ∖ + b x +- c-
j v 7 rx2 + sx + t
x 2 + 3x + 7
y = x+2 will have y = - as a horizontal asymptote and that
Solution 1. By division, when br ≠ as, the graph will cross this asymptote at
the point where
x 2 + 3x + 7 _ , , , 5 „ _ al ~ cr
x +2 x +2 br - as
Because the final term tends to zero as x grows, the b. What can be said about the case where br = as?
asymptote is the line of equation y = x + 1. c. Sketch the graph of each of the following functions:
Solution 2. Following a procedure frequently used in x 2 - 4x - 5
calculating limits at infinity, g(x) 2x2 + x - 10
x 2 + 3x + 7 = x + 3 + l
∕ι(x) = 3x2 - x - 7
x +2 1 + 2
X -1 2 x 2 + 4x+ 8
For large x, the value is approximately x + 3, so the © 60. Let ∕(x ) = p(x)∕d(x) be a rational function, and let c be a
asymptote should be the line of equation y = x + 3. number for which <7(c) = 0 but p(c) ≠ 0. Explain why
lim ∕(x ) and lim ∕( x ) must both be infinite.
Reconcile these solutions. What is wrong?
56. JOURNAL PROBLEM: ParabolaV Draw a careful sketch of Note that this means x = c is a vertical asymptote of the
the curve , graph of f.
61. Consider the rational function
= a,,x'i + αn~l x"-' + - + α , x + fl0

indicating clearly any vertical or horizontal asymptotes, b χ m + b . x m ~ '+ - + b x + b
m m l l 0
turning points, or points of inflection.
57. Let a. If m > n and b ≠ 0, show that the x-axis is the only
P(x) = anx" + απ. 1x"^' + ∙∙∙ + a γx + a0 horizontal asymptote of the graph of f
b. If m = n, show that the line y = an∣ bm is the only
be a polynomial with an ≠ 0 and let horizontal asymptote of the graph of f.
L = lim P(x) and M = lim P(x). c. If m < n, is it possible for the graph to have a
χ→~∞ X→+∞ horizontal asymptote? Explain.
Fill in the missing entries in the following table: 62. ■ What Does This Say? State what you think should be
the formal definition of each of the following limit
Sign of an n L M
statements:
+ even a. +∞ a. lim ∕(x ) = —∞ b. lim ∕(x ) = +∞
+ odd —∞ b.
— even c. d. 63. Prove that if
- odd e. —∞
lim ∕(x ) and lim g(x)
both exist, so does
*Volume 22, May 1991, p. 221.
t Volume 20, Issue 1, 1984. A turning point is what we have called
lim [∕(x) + g(x)]
χ - > +∞
a relative extrema.
Sec. 3.6 Summary of Curve Sketching 213
and 64. Prove the following limit rule. If lim ∕( x ) = +∞ and

lim [∕(x ) + g(χ)] = lim ∕( x ) + lim g(x)
λ'→+∞ X→+∞ Λ'→+∞ lim g(x) = A A > 0
Hint'. The key is to show that if
then lim [∕(x),r,r (-vj] = +∞
l∕(x ) - L ∖ < f
Hint: Notice that because lim g(x) = A , the function g(x)
for x > N, and
is near A when x is near c. Therefore, because lim ∕( x )
x→+∞
∣g(x) - M∖ < f = +∞, the product ∕(x)g(x) is large if x is near c. Forma
lize these observations for the proof.
for x > N 2, then 65. Prove the following limit rule. If lim ∕( x ) = +∞ and
∣ [ ∕W + g(*)] - (T + M)∣ < ε lim g(x) = A where A < 0, then '-n
whenever x > N for some number N. You should also lim 4 4 = - 0β

show that N relates to N i and N 2 . X→c g W
3 . 6 SUM M ARY OF CURVE SKETCHING
IN TH IS SECTIO N General curve sketching, graphing strategy

In this section we tie together your graphing skills from previous courses as well
as the calculus-based curve sketching developed in this chapter.
■ G EN E R A L CURVE SKETCHING
It is worthwhile to combine the techniques of curve sketching from calculus with

those techniques studied in precalculus courses. You may be familiar with extent
(finding the domain and the range of the function) or symmetry (with respect to
the x-axis, y-axis, or origin). These features are reviewed in the Student M athemat
ics Handbook and Integration Table fo r CALCULUS. We now have all the tools we
SM H need to describe a general procedure for curve sketching, and this procedure is
summarized in Table 3.2 on page 216.
EXAM PLE 1 Sketching a curve with an oblique asymptote
γ- _ γ _ y
Discuss and sketch the graph of y = ---- χ , — .
X J
Solution Performing the division, we write
Find the first and second derivatives, along with the critical values for y ' and
y"∙
Derivatives Critical values
y ' = 1 - 4(x - 3)" 2 1 - 4(x - 3)~2 = 0

1= 4(x - 3)" 2
(x - 3)2 = 4
x - 3= ± 2
x = 5, 1
y" = 8(x - 3)^ 3 8 ( x - 3 )- 3 = 0
No second-order
critical values
The second derivative test: ∕" ( 5 ) > 0; relative m inimum at x = 5.

∕''( 1 ) < 0; relative maximum at x = 1.
The intervals of increase and decrease, relative extrema, and concavity
are indicated the the preliminary sketch in Figure 3.48a. Note that there is
a vertical asymptote at x = 3 and an oblique asymptote y = x + 2,
because y = x + 2 H---- There are no horizontal asymptotes. Finally, plot a
X 3
few additional points and then draw a smooth curve, as shown in Figure
3.48b.
EXAM PLE 2 Sketching a trigonometric curve
Discuss and sketch the graph of y = 20 sin(378ττx).

Solution We recognize this as a standard sine curve, but we need to pay some
attention to the domain and range. We note that this sine curve has amplitude
20 and period 2τ7∙∕(378π) = These numbers help us set the scale on both
the axes. We find the critical values for y':
y ' = [20 cos(378πx)](37877) = 7560τr cos(378ττx)
y ' = 0 whenever COS(378 TTΛ') = 0 or 378τrx = + nπ for an integer n
Because of our knowledge of the shape of a sine curve, it is not necessary for us
to find the concavity or points of inflection. The graph is shown in Figure
3.49.
Figure 3.49 Graph of

y = 20 sin(378πx)
TABLE 3.2 Graphing Strategy for a Function Defined by y = ∕( x )

Step Procedure
Simplify. I f possible, simplify algebraically the function you wish to graph.

Find derivatives and Compute the first and second derivatives; factor, if possible, and
critical values. find the first- and second-order critical values.
Apply the second- Use the second-derivative test to find the relative maxima or
derivative test. minima: substitute the first-order critical values, c 1, with the
following test:
∕" ( c 1) > 0, relative minimum
∕" ( c l ) < 0, relative maximum
∕' ' ( c 1) = 0, test fails; use first-derivative test
Determine concavity Use the second-order critical values, c ,, to determine concavity:
and points of Substitute f ' ∖c-,) > 0, concave up; ∕' ' ( c 2) < 0, concave down.
inflection. Find the points of inflection. These are located where the
concavity changes from up to down or down to up. and are found
at the x-intercepts o f the second derivative.
Apply the first- Use the first-derivative test if the second-derivative test fails or is
derivative test. too complicated.
/ ∖ 1. Let c be a first-order critical value of f.
2. a. (c, ∕(c )) is a relative minimum if
Maximum f'( x ) < 0 for all x in an open interval (a, c) to the left ofc and
∖ / f'( x ) > 0 for all x in an open interval (c, b) to the right o f c.
I>. (c,∕(c)) is a relative maximum if
Minimum ∕'( x ) > 0 for all x in an open interval (a, c) to the left o f c and
∕'( x ) < 0 for all x in an open interval (c, b) to the right of c.
c. (c, ∕(c)) is not an extremum i f ∕'( x ) has the same sign in open
intervals («, c) and (c, ⅛) on both sides o f c.
Determine intervals Use the first-order critical values o f/ to determine intervals of
of increase and increase and decrease:
decrease. ∕'( x ) > 0, curve rising (indicate these regions by ↑ )
∕'( x ) < 0, curve falling (indicate these regions by ∣ )
Find asymptotes. 1. Vertical asymptote— The vertical asymptote, if one exists, will
occur at a value x = c for w hich/is not defined.
Use the limits lim ∕( x ) and lim ∕( x ) to determine the behavior
of the graph near x = c. Show this behavior with arrows ( ↑ ∣ ).
2. Horizontal asymptote— Compute lim ∕( x ) and lim ∕( x ) . If
x→+∞ x→-∞
either is finite, plot the associated horizontal asymptote.
3. Oblique asymptote— I f / is a rational function with no common
factors and with the degree of the numerator 1 more than the
degree of the denominator, then write∕( x ) = m x + b
+ r(x)∕rf(x) and plot the asymptote y = m x + b.
Plot points. 1. If c is a relative maximum, relative minimum, or inflection
point, plot (c, ∕(c)). Show the relative maximum points by using
a “ cap” (∩), and relative minimum points by using a “ cup” (U).
2. x-intercepts: Set y = 0 and if x = a is a solution, plot (a, 0).
y-intercepts: Set x = 0 and if y = b is a solution, plot (0, ⅛),. a
function will have, at most, one y-intercept.
3. Asymptote intercepts: Plot those points where/intersects
horizontal or oblique asymptotes.
Sketch the curve. Draw a smooth curve through the plotted points.
EXAM PLE 3 Sketching a curve with asymptotes

iI3 Computational Window
Discuss and graph ∕(x ) = 2 + 3(x 7- 2) 10
3(x + I)’
Solution We first simplify the algebraic representation of the given function.
2(3)(x + l)(x - 2 ) + 7(x + 1) - 10(Λ- - 2) 2

2 χ - 3x + 5
∕W = 3(x + l)(x - 2) (x + l)(x - 2)
Y ιB 2 + 7 ∕< 3 < X -2 )) - Find the first and second derivatives, along with the critical values fo r/a n d /':
10A3<X+1))
Xnin=-10 V∣ηin=-10 _ (x2 - x - 2)(4x - 3 ) - (2x2 - 3x + 5)(2x - 1) _ x 2 - 18x + 11
∕'( τ ) (x + l) 2(x - 2)2
Xr⅛ax=10 Yrcax=10 (x + l) 2(x - 2)2
X s c l= l VS G 1=1
To find the critical values for / solve
It is sometimes difficult to
choose the necessary informa x 2 - 18x + 11 _ ∩
tion without using calculus. On (x + l) 2(x - 2)2
the other hand, sometimes the
calculator graph will suffice. x 2 - 18x + 11 = 0
This example shows that the χ = 9 ± V70
calculus can sometimes be very
tedious; on the other hand, a The first derivative is not defined when x = -1 and x = 2, but these
calculator will not supply values are not in the domain of / Thus the first-order critical values are
9±√70. 9 + V70, and 9 - λ∕70. Next, find the second derivative (we will not show the
intermediate steps).
„ 2 ( - χ 3 + 27x2 - 33x + 29)
j {X’ = ( x + l ) 3( x - 2 ) 3
To find the second-order critical values, solve

2 ( - χ 3 + 27x2 - 33x + 29) z
(x + l) 3(x - 2)3
- x 3 + 27x2 - 33x + 29 = 0
You can approximate the roots of this cubic equation by using a spreadsheet,
the bisection method, or the Newton-Raphson method: there is one real root,
namely, x≈ 25.8. Additional second-order critical values are where the
second derivative is not defined, namely, x = -1 and x = 2, but as we have
seen, these are not in the domain of f.
The second-derivative test gives us the following information:
∕" ( 9 + √70) > 0; relative minimum at x = 9 + V70
∕" ( 9 - √70) < 0; relative maximum at x = 9 - λz70
The function has vertical asymptotes at x = -1 and x = 2:
li 2x 2 - 3x + 5 +∞ lim 2x2 - 3x + 5 —∞
→≡ι- (x + l) ( x - 2 ) x∏rn 1÷ (x+ l)( x -2 )
li 2x2 - 3x + 5 = lim 2x2 - 3x + 5 = +∞

™ (x + 1)(x - 2) t →r (x + l)(x - 2)
These asymptotes (along with the appropriate arrows) are shown in the
preliminary sketch in Figure 3.50a.
The function has a horizontal asymptote:
2 - 3x + 5
lim (2xx + l)(x -2 )
= ?
“
X —» +∞
lim 2x 2- 3x + 5 =
(x + l)(x - 2)
The horizontal asymptote is y = 2. This is shown in Figure 3.50a.
Plot a few important points:
Critical Values
x = 9 + √70 / ( 9 + √ 7 0 ) ≈ /(17.4) ≈ 1.97
Plot relative minimum point (17.4, 1.97) and label it with a “cup” (U).
x = 9 -√ 7 0 /(9 - √70) ≈ /(0.6) ≈ -1.7
Plot relative maximum point (0.6, -1.7) and label it with a “cap” (∩).
Intercepts
x=0 Z(θ) = - j Plot (o, - ∣).
y=0 2x2 - 3x + 5 _ ∩
( x + l ) ( x - 2 ) υ No real roots; thus there are no
x-intercepts.
asymptote y = 2 2x2 —3x + 5 _ 9 Root x = 9

(x + l)(x - 2)
/(9) = 2 The curve crosses the horizontal
asymptote at (9, 2).
∕( x ) = 2 + - 7 0 , 10 If necessary, you can plot a few additional points to draw a smooth curve, as
j v ’ 3(x - 2) 3(x + 1) shown in Figure 3.50b. ≡≡3
EXAM PLE 4 Sketching a trigonometric curve
Discuss and graph T(x) = 1 cs ∏J

1 - z- S in y on [0, ττ].
l l .∖
Solution The first and second derivatives are found to be
2 - sin x rr7,, zx x cosx(7 - 2 sinx)

∏ x) and
( 1 - 2 sinx) 2 T ( )= ( 1 - 2-, s∙m
/1 ---x- )√3Γ^
None of the functions T, T ', or T " is defined where (1 - 2 sinx) is 0. Solve

1 - 2 sinx = 0
sinx = J
The roots in [0, τr] are x = f and yy. We find that Γ'(x) > 0 for all other x in
the interval, so there are no other critical values for T. We also note from this
information that T must be increasing for all defined points.
The critical values of T " (other than those for which it is not defined) are
found by solving
7 - 2 sinx = 0 and cosx = 0
sinx = 5 x = J
Because sinx is never j , we see that the only second order critical value of t is
x = f.
a. Preliminary sketch
By checking the sign of T " on each side of f , we see that the concavity
changes, so it corresponds to a point of inflection.
The graph has a vertical asymptote when the denominator of T is 0;
namely, when sinx = ∣ or when x = f or x = We find that
lim Γ(x) = +∞ lim Γ(x) = -∞

A'→ π∕6 +
lim T(x) = +∞ lim T(x) = -∞
x → 5π∕6^
There are no horizontal or oblique asymptotes.
Finally, plot some points:
Γ(0) = 1; plot the point (0, 1). This is an endpoint.
T(ττ) = - 1 ; plot the point (π, -1 ). This is an endpoint.
T(y) = 0; plot the point (7, θ). This is an inflection point.
COS X The preliminary graph is shown in Figure 3.51a and the completed graph in
Γ(x) = on [0, TT]
1 - 2 sinx Figure 3.5 lb.
PROBLEM SET 3 .6 _________________________

θ 1. ■ What Does This Say? Outline a method for curve Sketch the graphs of the functions in Problems 5-22.
sketching. 5. ∕( x ) = 3x 4 - 8x 3 + 6x 2 + 2 6. ∕( x ) = 324x - 72x2 + 4x 3
2. ■ What Does This Say? Discuss the importance of
critical values in curve sketching. 7. ∕( x ) = 2x 3 + 6x 2 + 6x + 5
3. ■ What Does This Say? Discuss the importance of 8. ∕( x ) = 10x6 + 24x 5 + 15x4 + 3
concavity and points o f inflection in curve sketching.
9. /( /) = ( / + 1)2( ∕ - 5) 10. ∕( z ) = ( 1 - z)2(z + 2)
4. ■ What Does This Say? Discuss the importance of
asymptotes in curve sketching. 11. g(x) = 2x 3 - 3x 2 - 12x + 14 12. ∕z(x) = x 3 + 3x 2 - 6x + 2
Sec. 3.7 Optimization in the Physical Sciences and Engineering 21 9
13. g(∕) = ( ∕ 3 + t)2 14. ∕(x ) = xV16 - χ 2 where g is the constant acceleration due to gravity, r is the
15. ∕(x ) = 12x(x - 1)2 + (9x + 2)(x - 2)3 radius of the curve, and θ is the angle of the bank.
16. g(u) = u 4 + 6 M3 - 24 M2 —62w - 15 Assuming that r is constant, sketch the graph of v as a
17. fit) = 4 ∕ 5 + 5 ∕ 4 18. g(∕) = f lz2 + ⅜∕3z2 function of θ for 0 ≤ θ ≤ j.*
45. According to Einstein’s special theory of relativity, the
19∙ ∕w = -⅛
A i A
2 θ∙ ∕ω = ~ *j∖ +3I 2+ 6
t
mass of a body satisfies the equation
21. c(x) = 25 cos(450πx) 22. s(x) = 30 sin '- A ? )

. 180π∕
θ Find the first- and second-order critical values o f the functions
given in Problems 23-43, and classify each such value as a
relative maximum, a relative minimum, or neither. Sketch the where m 0 is the mass of the body at rest in relation to the
graph o f each function. observer, m is the mass of the body when it moves with
23. fix ) = x 3 + 3x 2 + 1 24. fix ) = x 4 - 2x 2 + 3 speed v in relation to the observer, and c is the speed of
2
25. fix ) = (x - 9) 2 26. g(x) = x(x 2 - 12) light. Sketch the graph of m as a function of v.
27. fix ) = 2x + 1 + 28. fix ) = 2x 3 - 3x 2 - 12x + 13

29. fix ) = (x + 1)2(5 - x) 30. fit) = ( / - 3)2(∕ - 4)2 Computational Window
31. fix ) = 3x 4 - 2x 3 - 12x2 + 18x
32. fix ) = x l,3 (x - 4) 33. fiu ) = W2Z3(M - 7) In Problems 46-47, use the bisection method, the Newton-
2 Raphson method, or a computer to find all asymptotes off,
34. g (u) = (w2 + u + I)’ 2 3 5 . ∕( X) = ^ ⅜ 2
correct to the nearest hundredth. Sketch the graph.
36. t(θ) = sin 0 - cos0 for 0 ≤ θ ≤ 2ττ 4A6r∙ r ∕. .∖ - P
X^^ T X 3
V 3FΠ
A7f ∙ rr.∖ - x
4A X53 X2+ 1
+x +2
37. fix ) = x - sin 2x for 0 ≤ x ≤ π
38. g(f) = tan / - sec/ for 0 ≤ t ≤ j
39. fix ) = sin 2 x - 2 sinx + 1 for 0 ≤ x ≤ π θ T HINK T ANK P ROBLEMS In Problems 48-51, either show that
40. g(t) = -t ± ½ t the statement is generally true or find a counterexample.
41. g(t) =
∖ ∕t2 + 1 48. Iff and g are concave up on the interval I, then so i s ∕+ g.
42. g(x) = sinx tan x on [ - π , τr] 49. I f /is concave up and g is concave down on an interval I,
n γ on [-2 π , 2π] then fg is neither concave up nor concave down on I.
43. go v(x) = ,f -
’ 1 + cosx l j 50. If fix ) < 0 and ∕" ( x ) > 0 for all x on I, then the
44. The ideal speed v for a banked curve on a highway satisfies function g = ∕ 2 is concave up on I.
the equation 51. If∕(x ) > 0 and g(x) > 0 for all x on I and if/a n d g are
v2 = gr tan θ concave up on /, then fg is also concave up on I.
3 . 7 OPTIMIZATION IN THE PHYSICAL SCIENCES AND ENGINEERING
IN THIS SECTION Optimization procedure, Fermat’s principle of optics

and Snell’s law
In Chapter 1 we introduced the processes involved with model-building. Now
that we have developed the necessary calculus skills, we shall use those skills in
solving practical optimization problems.
■ OPTIMIZATION PROCEDURE
Nothing takes place in the world It is common to ask for the best procedure, the greatest value, the least cost, or the
whose meaning is not that o f some shortest path. The process of developing something to the utmost extent is called
maximum or minimum. optimization. Entire courses in mathematics are devoted to this topic, and in this
Leonhard Euler section and the next we will develop procedures involving calculus to solve real-
*In physics it is shown that if one travels around the curve at the ideal speed, no frictional force is
required to prevent slipping. This greatly reduces wear on tires and contributes to safety.
life problems that seek the maximum or minimum value. We will model our
procedure after one attributed to George Polya.
Optimization Procedure Step 1: Understand the problem. Ask yourself if you can separate the given
quantities and those you must find. What is unknown? Draw a picture to
help you understand the problem.
Step 2: Choose the variables. Decide which quantity is to be optimized (that
is, maximized or minimized) and call it Q. Choose other variables for
unknown quantities and label your diagram using these symbols.
Step 3: Express Q in terms of the variables defined in Step 2. Use the
information given in the problem and the principle of substitution to
rewrite Q in terms of a single variable, say x. In other words, Q may begin as
a formula involving several variables, but by using given information or
known formulas, the goal is to write Q as a function of one variable so that
Q = /(*)•
Step 4: Find the domain for the function Q = f(x).
Step 5: Use calculus to find the absolute maximum or minimum value o f/ In
particular, if the domain o f/ is a closed interval [a, b∖, then the procedure
outlined in Section 3.1 can be used:
a. Compute ∕'( x ) and find all critical values o f/ o n [a, ⅛].
b. Evaluate / at the endpoints a, b, and at each critical value, c.
c. Compare the values in step b to find either the largest or smallest value.
Step 6: Convert the result obtained in Step 5 back into the context of the
original problem, making all appropriate interpretations. Be sure to answer
the question asked.
EXAM PLE 1 Maximizing a constrained area
You need to fence in a rectangular play zone for children, to fit into a right-
triangular plot with sides measuring 4 m and 12 m. What is the maximum
area for this play zone?
D E
*------ 12 m -------> Solution A picture of the play zone is shown in Figure 3.52. Let x and y denote
the length and width of the inscribed rectangle. The appropriate formula for
Figure 3.52 Children’s play zone the area is a = €w = χ y
We wish to maximize Q where 0 = A = x y. Now, write this as a function of a

single variable. To do this, note that
ΔABC ~ ΔADF
This means that corresponding sides of these triangles are proportional;
therefore
4 ~T . . χ
4 12
4 - y = ⅛x
y = 4 -⅛ x
We can now write O as a function of x alone:
Q(x) = x(4 - ∣x) = 4x - ∣ x 2

Sec. 3.7 Optimization in the Physical Sciences and Engineering 221
The dom ain for x is 0 ≤ x ≤ 12. The critical values for Q are values such
Computational Window SIES that <2'(x) = 0 (since there are no places where the derivative does not exist).
You can find the maximum Because
value by using the (TRACE) on <2'(Λ ) = 4 —∖ x
a graphing calculator. Your
first attempt at graphing the critical value is x = 6. Evaluate O(x) at the endpoints and the critical
y = 4x - ∣ x 2 will probably value:
look like: <2(6) = 4(6) - ⅛(6)2 = 12; β(0) = 0; 2 (1 2 )= 0
The maximum area occurs when x = 6. This means that

y = 4 - 1(6) = 2
The largest rectangular play zone that can be built in the triangular plot is a
rectangle 6 m long and 2 m wide, so the maximum area is 12 m 2 .
Y ιB 4 X - < l∕3 )X ≡ EXAM PLE 2 Maximizing profits

X n in = - 5 Yroin= - 5 A m anufacturer can produce a pair of earrings at a cost of $3. The earrings
Xnax=15 Yroax=20 have been selling for $5 per pair, and at this price, consumers have been
buying 4,000 per month. The m anufacturer is planning to raise the price of the
By using the (ZOOM) you can earrings and estimates that for each $ 1 increase in the price, 400 fewer pairs of
find the following (or similar) earrings will be sold each month. At what price should the m anufacturer sell
values: the earrings to maximize profit?
X = 5.833795 Y = l 1.990792
X =6.0110803 Y = l l .999959 Solution Let x denote the num ber of $ 1 price increases, and let P(x) represent the
X=6.1883657 Y= 11.988173 corresponding profit.
It looks like the maximum val PROFIT = REVENUE - COST
ue is 12. Note that calculus
= (NUMBER SOLD)(PR ICE PER PAIR) - (NUMBER SOLD)(COST PER PAIR)
may be easier than using a cal
culator. On the other hand, the = (NUMBER SOLD)(PRICE PER PAIR - COST PER PAIR)
calculator does well in check
Recall that 4,000 pairs of earrings are sold each m onth when the price is $5
ing an answer you have found
through calculus. per pair and 400 fewer pairs will be sold each m onth for each added dollar in
the price. Thus,
NUMBER OF PAIRS SOLD = 4 ,0 0 0 - 400(NUMBER OF $ 1 INCREASES)
= 4,000 - 400x
Knowing that the price per pair is 5 + x, write the profit as a function of x:
P(χ) = (NUMBER SOLD)(PRICE PER PAIR - COST PER PAIR)
= (4,000 - 400x)[(5 + x) - 3] = 400(10 - x)(2 + x)
To find the domain, we note that x ≥ 0. And 400(10 - x), the number of
pairs sold, should be nonnegative, so x ≤ 10. Thus, the domain is [0, 10].
The critical values are found when the derivative is 0 (P is a polynomial
function, so there are no values for which the derivative is not defined):
P '(x) = 400(10- X )(1) + 4 0 0 (-1 )(2 + X) = 400(10 - x - 2 - x)
= 400(8 - 2x) = 800(4 - x)
The critical value is x = 4 and the endpoints are x = 0 and x = 10. Checking
for the maximum profit:
jp(4) = 400( 1 0 - 4)(2 + 4) = 14,400; P(0) = 8,000; P(10) = 0
The maximum possible profit is $14,400, which will be generated if the
Figure 3.53 The profit function earrings are sold for $9.00 per pair. The graph of the profit function is shown
P(-<) in Figure 3.53. ≡≡
h------------ 45 - 2 x ------------- H EXAM PLE 3 Maximizing a volume
A carpenter wants to make an open-topped box out of a rectangular sheet of

tin 24 in. wide and 45 in. long. The carpenter plans to cut congruent squares
out of each corner of the sheet and then bend the edges of the sheet upward to
form the sides of the box, as shown in Figure 3.54. If the box is to have the
greatest possible volume, what should its dimensions be?
Solution If each square corner cut out has side x , the box will be x inches deep,
45 - 2x inches long, and 24 - 2x inches wide. The volume of the box shown in
Figure 3.54 is
F(x) = x(45 - 2x)(24 - 2x) = 4x 3 - 138x2 + l,080x
To find the domain, we note that the dimensions must all be nonnegative;
therefore, x ≥ 0, 45 - 2x ≥ 0 (or x ≤ 22.5), and 24 - 2x ≥ 0 (or x ≤ 12).
This implies that the domain is [0, 12]. To find the critical values (the
Figure 3.54 A box cut from a 24 derivative is defined everywhere in the domain), we find values for which the
in. by 45 in. piece of tin derivative is 0:
K'(x) = 12x2 - 276x + 1,080 = 12(x - 18)(x - 5)
Computational Window The critical values are x = 5 and x = 18, but x = 18 is not in the domain,
The necessity of knowing the so the only relevant critical value is x = 5. Evaluating K(x) at the critical value
endpoints of the interval is em x = 5 and the endpoints x = 0, x = 12, we find
phasized when you use a F(5) = 5(45 - 10)(24- 10) = 2,450; F(0) = 0; K(12)=0
graphing calculator. For exam
ple, the TI-85 finds the approx Thus, the box with the largest volume is found when x = 5. Such a box has
imate maximum value of a dimensions 5 in. × 14 in. × 35 in. M B
given function but requires
that you enter function, vari
able, lower limit, and upper
EXAM PLE 4 Maximizing an area using trigonometry
limit. For V'(x) in Example
3, we obtain the answers
Two sides of a triangle are 4 inches long. What should the angle between these
5.00231626223 with a toler
ance of 0.01 and δ = 0.001. sides be to make the area of the triangle as large as possible?
Solution The triangle is shown in Figure 3.55. In general, the area of a triangle is
given by
A = ⅛bh
In this case, the base b is 4 and, because sin θ = the height h is 4 sin θ. Thus,
A = ^(4)(4 sin θ) = 8 sin θ
To find the domain, we note that the largest angle in a triangle is TT; therefore,
0 ≤ θ ≤ ττ. Also, because A '(θ) is defined throughout this interval, the
critical values are found when the derivative is 0. We have
A'(θ) = 8 cosθ
which is zero only when θ = ∙j. Because
A(f) = 8 s in f = 8, A(0) = 8 sin 0 = 0, A(π) = 8 sin τr = 0
Figure 3.55 Triangle with two we see that the area is maximized when the angle θ is J — that is, when the
4-in. sides triangle is a right triangle. ⊂zz∑3
EXAMPLE 5 Minimizing time of travel
A dune buggy is on the desert at a point A located 40 km from a point B, which

lies on a long, straight road, as shown in Figure 3.56. The driver can travel at
45 km/hr on the desert and 75 km/hr on the road. The driver will win a prize if
she arrives at the finish line (point D) in less than 1 hr. If the distance from B
to D is 28 km, is it possible for her to choose a route so that she can collect the
prize?
Figure 3.56 A path traveled by a Solution Suppose the driver heads for a point C located x km down the road
dune buggy from B toward her destination, as shown in Figure 3.56. We want to minimize
the time. We will need to remember the formula d = rt, or in terms of time,
t = dlr.
TIME = TIME FROM A TO C + TIME FROM C TO D
DISTANCE FROM A TO C , DISTANCE FROM C TO D
Γ(%) = RATE FROM A TO C RATE FROM C TO D
∖, x 1 + 1,600 28 ~ x
45 75
The domain for x is [0,28]. Next, find the derivative of time T with respect to x.
Γ'(x) = ⅛ [⅛ χ 2 + 1 ^ 0 0 )- ,z2(2x)] + ⅛ (-l)
_____ x_____ _ 1
45√ χ 2 + 1,600 75
5x —3Vx2 + 1,600
225Vx 2 + 1,600
The derivative exists for all x and is zero when

5x - 3Vx2 + 1,600 = 0
Solving this equation, we find x = 30 (-30 is extraneous), which must be
rejected because it is not within the interval [0, 28]. Evaluating 7’(x) at the
endpoints, we find that
Γ(O) = ^ ≡ ≡ 2 8 -0
75 1.26
_ √ 2 8 2 + 1,600 28 - 28 ~
1 (2 δ ) - — JΞ + — 7∑ 1.09
The driver can minimize the total driving time by heading directly across the
desert from A to D. Even so, she takes a little more than an hour for the trip, so
she does not win the prize. ≡≡≡
EXAMPLE 6 Optimizing a constrained area

A wire of length L is to be cut into two pieces, one of which will be bent to
form a circle and the other, to form a square. How should the cut be made to
a. maximize the sum of the areas enclosed by the two pieces?

b. minimize the sum of the areas enclosed by the two pieces?
Solution In order to understand the problem we draw a sketch, as shown in

Figure 3.57, and label the radius of the circle r and the side of the square s.
Wire, length L
Figure 3.57 Forming a circle and a square from a wire with length L
We find that the combined area is

AREA = AREA OF CIRCLE + AREA OF SQUARE
= πr 2 + s 2
We need to write the radius r and the side 5 in terms of the length of wire, L .
L = CIRCUMFERENCE OF CIRCLE + PERIMETER OF SQUARE
= 2πr + 4.s,
Thus, 5 = ∣(L - 2ττr). Remember, L i s a given constant, so the variable for the
area function is r (although the problem could just as easily have been done in
terms of A). By substitution,
/1(r) = πr2 + [∣(L - 2ττr)]^ = ττr2 + ⅛(L —2ττr)2
To find the domain, we note that r ≥ 0 and that L - 2τrr ≥ 0, so the domain
is 0 ≤ r ≤ The derivative of A(r) is
A'(r) = 2πr + ∣(L - 2τrr)(-2π)

= 2πr - f (L - 2ττr) = %(8r - L + 2τrr)
We solve A '(vr) = 0 to find r = ~-, . ..

' 2{π + 4)
Thus, the extreme values of the area function on 0, must occur either
L 2τrj
at the endpoints or at the critical value g∙ Evaluating Λ(r) at each of these
numbers, we find
Λ(0) = TT∙(O)2 + ⅛[L - 2ττ(0)]2 = f ∣ = 0.0625L 2
∕t(⅛ 2 = + ⅛ [ L - 2 √ ^ Γ = ^ ≈ 0 .0 7 9 577 471L 2

∖2ττ∕ ∖2ττ∕ 10 L ∖2ττ∕J 4τr
δ)∖~ + T6∣∕ T ∖∖l

7 ∖= / τ ι 7 2
2τr+ 8/ π (,2π+ - ~ 2 ∖2τr+ 8y)j = 4(πT+ 4)
O.O35OO6 197L2
Comparing these values, we see that the smallest area occurs at x = ∏

Z7Γ + o
0
and the largest area occurs at x = To summarize:
1. To maximize the sum of the areas, do not cut the wire at all. Bend the wire
to form a circle of radius r = τA .
2τr
Sec. 3.7 Optimization in the Physical Sciences and Engineering 2 25
1rπL
2. To minimize the sum of the areas, cut the wire at the point located 2π +8
units from one end, and form the circular part with radius r = Zπ 77 + 0o
.
If you use a graphing calculator or a computer, you can verify this result (for
L = l)bylookingatthegraphoftheareafunction,asshowninFigure3.58. ⅛≡≡
■ PHYSICS: FERMAT’S PRINCIPLE OF OPTICS AND SNELL’S LAW

Light travels at different rates in different media; the more optically dense the
Figure 3.58 Graph of medium, the slower the speed of transit. Consider the situation shown in Figure
A(r) = τrr2 + ⅛[1 - 2πr] 2 3.59, in which a beam of light originates at a point A in one medium, then strikes
the upper surface of a second, denser medium at a point P, and is refracted to a
point B in the second medium.
Figure 3.59 The path of a light beam through two media of different density
Suppose light travels with speed v1 in the first medium and speed v2 in the
second. What can be said about the path followed by the beam of light?
Our method of investigating this question is based on the following optical
property.
Fermat’s Principle of Optics Light travels between two points in such a way as to minimize the time of
transit.
This problem is very similar to the dune buggy problem of Example 5: Light
minimizes the time, and minimum time was the goal of Example 5. In Figure 3.59,
let a = AC, b = CD, and c = DB, and let x denote the distance from C to P.
Because A and B are fixed points, the path APB is determined by the location of P,
which, in turn, is determined by x. Because v1 is a constant, the time required for
the light to travel from A to P is given by
τ _ K∣ _ VFKP
1 vι vι
and the time required for the light to go from P to B is

∖PB∖ ∖ ∕b2 + (c - x)2
2 V2 V2
Therefore, the total time of transit is

τ = τ , τ = √ Ξ K √ , V⅛2 + (c ~ x)2
1 1 ∖ ~r 1 2 vκl vv 2
where it is clear that 0 ≤ x ≤ c. According to Fermat’s principle, the path

followed by the beam of light is the one that corresponds to the smallest possible
value of T; that is, we want to minimize T as a function of x .
irτι
Toward this end, we begin by finding
dT _ x _ c —x
dx v1V α 2 + x 2 V2V⅛ 2 + (c - x) 2
Note from Figure 3.59 that if a is the angle of incidence of the beam of light and β
is the angle of refraction, then (from the definition of sine)
x ∙ c —X
sin a = , , -- and sin β = ,,. , .
V α 2 + x~ h V⅛2 + (c - x)v 2
Therefore (by substitution), we see the derivative of T can be expressed as
√ 7 ∖ .sin α sin y8
dx v1 v2
and it follows that the only critical value occurs when
sin a _ v ι
sin β v2
By using the second-derivative test in the interval [0, c], it can be shown that this
critical value is an absolute minimum. The corresponding value of x enables us to
locate P and hence to determine the path followed by the beam of light. We have
established the following law of optics.
Snell’s Law of Refraction If a beam of light strikes the boundary between two media with angle of
incidence a and is refracted through an angle β, then
sin a _ v ι
sin β v2
where v1 and v2 are the rates at which light travels through the first and second
medium, respectively. The constant ratio
∕t = ¾ ≡
sin β
is called the relative index of refraction of the two media.
PROBLEM SET 3 .7
O 1. ■ What Does This Say? Describe an optimization pro 5. A farmer wishes to enclose a rectangular pasture with 320
cedure. ft of fence. Find the dimensions that give the maximum
2. ■ What Does This Say? Why is it important to check area in these situations:
end values when finding an optimum value? a. The fence is on all four sides of the pasture.
© 3. A woman plans to fence off a rectangular garden whose b. The fence is on three sides of the pasture and the
area is 64 ft 2 . What should be the dimensions of the fourth side is bounded by a wall.
garden if she wants to minimize the amount of fencing 6. E X P E R IM E N T Pull out a sheet of 8∣-in. by 11-in. binder
used? paper. Cut squares from the corners and fold the sides up
4. The highway department is planning to build a rectangu to form a container. Show that the maximum volume of
lar picnic area for motorists along a major highway. It is such a container is about 1 liter.
to have an area of 5,000 yd 2 and is to be fenced off on the 7. JOURNAL P R O B LEM : Parabola* Farmer Jones has to
three sides not adjacent to the highway. What is the least build a fence to enclose a 1200 m 2 rectangular area
amount of fencing that will be needed to complete the
job? * Volume 19, Issue 1, 1983, p. 22.
A B C D . Fencing costs $3 per meter, but Farmer Smith has 19. S P Y PROBLEM It is noon. The spy has returned from
agreed to pay half the cost of fencing C D , which borders space (Problem 50 of Section 2.4) and is driving a jeep
the property. Given x is the length of side B C , what is the through the sandy desert in the tiny principality o f Alta
maximum amount (to the nearest cent) Jones has to pay? Loma. He is 32 km from the nearest point on a straight,
8. Find the rectangle of largest area that can be inscribed in a paved road. Down the road 16 km is a power plant in
semicircle of radius R , assuming that one side of the which a band of international terrorists has placed a time
rectangle lies on the diameter of the semicircle. bomb set to explode at 12:50 P . M . The jeep can travel at
9. A tinsmith wants to make an open-topped box out o f a 48 km/hr in the sand and at 80 km/hr on the paved road.
rectangular sheet of tin 24 in. wide and 45 in. long. The I f he arrives at the power plant in the shortest possible
tinsmith plans to cut congruent squares out of each corner time, how long will our hero have to defuse the bomb?
o f the sheet and then bend the edges of the sheet upward
to form the sides of the box. What are the dimensions of
the largest box that can be made in this fashion?
10. Find the dimensions of the right circular cylinder of
largest volume that can be inscribed in a sphere of radius
R.
11. Given a sphere of radius R , find the radius r and altitude h
of the right circular cylinder with largest lateral surface 20. Missy Smith is at a point A on the north bank of a long,
area that can be inscribed in the sphere. Hint'. The lateral straight river 6 mi wide. Directly across from her on the
surface area is 5 = 2πrh. south bank is a point B, and she wishes to reach a cabin C
12. Each edge o f a square has length L . Determine —⅛ located s mi down the river from B. Given that Missy can
the edge o f the square of largest area that can ∖ ∕ Γ∖∖ row at 6 mi/hr (including the effect of the current) and run
be circumscribed about the given square. ∑K Z ∕∖ at 10 mi/hr, what is the minimum time (to the nearest
13. Find the dimensions of the right circular '— minute) required for her to travel from A to C in each
cylinder o f largest volume that can be case? a. 5 = 4 b. 5 = 6
inscribed in a right circular cone of radius ⅛⅛
R and altitude H . ⅛Si '*∕
14. A truck is 250 mi due east o f a sports car HH
and is traveling west at a constant speed of . ■
60 mi/hr. Meanwhile, the sports car is
going north at 80 m i∕hr. When will the
truck and the car be closest to each other?
What is the minimum distance between 21. Two towns A and B are 12 mi apart and are located 5 mi
them? Hint'. Minimize the square of the and 3 mi, respectively, from a long, straight highway. A
distance. *________ L construction company has a contract to build a road from
15. Show that of all rectangles with a given perimeter, the A to the highway and then to B. How long is the shortest
square has the largest area. (to the nearest mile) road that meets these requirements?
16. Show that of all rectangles with a given area, the square
has the smallest perimeter.
17. A closed box with square base and vertical sides is to be
built to house an ant colony. The bottom of the box and
all four sides are to be made of material costing $ l ∕f t 2,
and the top is to be constructed of glass costing $5/ft2 .
What are the dimensions o f the box of greatest volume
that can be constructed for $72?
18. According to postal regulations, the girth plus the length 22. A poster is to contain 108 cm 2 of printed matter, with
of a parcel sent by fourth class mail may not exceed 108 margins o f 6 cm each at top and bottom and 2 cm on the
in. What is the largest possible volume o f a rectangular sides. What is the minimum cost of the poster if it is to be
parcel with two square sides that can be sent by fourth made of material costing 204/cm2?
class mail? 23. An isosceles trapezoid has a base of 14 cm and slant sides
of 6 cm, as shown in Figure 3.60. What is the largest area
of such a trapezoid?
Length = y
Figure 3.60 Area of a trapezoid
24. EXPERIMENT Use the fact that 12 fl oz is approximately The distance f from the center 0 of the lens to the point
6.89ττ in. 3 to find the dimensions of the 12-oz Coke® can labeled F is called the focal length of the lens.
that can be constructed using the least amount of metal. a. Using similar triangles, show that
Compare these dimensions with a Coke from your refrig
i + i =7
erator. What do you think accounts for the difference?
An interesting article that discusses a similar ques p q f
tion regarding tuna fish cans is “What manufacturers say b. Suppose a lens maker wished to have p + q = 24. What
about a max/min application” by Robert F. Cunningham, is the largest value o f/fo r which this condition can be
Trenton State College, The Mathematics Teacher, March satisfied?
1994, pp. 172-175.
29. When a mechanical system is at rest in an equilibrium
25. A cylindrical container with no top is to be constructed to position, its potential energy is minimized with respect to
hold a fixed volume of liquid. The cost of the material any small change in its position. Figure 3.64 shows a
used for the bottom is 50<t∕in.2, and the cost of the system involving a pulley, two small weights of mass m
material used for the curved side is 30Φ∕in2 . Use calculus and a larger weight of mass M.
to find the radius (in terms of the volume) of the least
expensive container.
26. Light emanating from a source A is reflected by a mirror
to a point B, as shown in Figure 3.61. Use Fermat’s
principle of optics to show that the angle of incidence a
equals the angle of reflection β.
It can be shown that the total potential energy of the

Figure 3.61 Angle of incidence is equal to the angle of system is given by
reflection.
E = -M g (x + h~) - 2mg(L —V x 2 + d 1)
27. A stained glass window in the form of an equilateral where x, h, and d are the distances shown in Figure 3.64,
triangle is built on top of a rectangular L is the length of the cord from A around one pulley to B,
window, as shown in Figure 3.62. The and g is the constant acceleration due to gravity. All the
window is of clear glass and transmits other symbols represent constants. Use this information
twice as much light per ft2 as the trian to find the value of x for which E is minimized.
gular window, which is made of stained
30. One end of a cantilever beam of length L is built into a
glass. If the entire window has a perime
wall, but the other end is supported by a single post. The
ter of 20 ft, find the dimensions (to the
deflection, or “sag,” of the beam at a point located x units
nearest ft) of the rectangular window
from the built-in end is given by the formula
that will admit the most light.
D = fc(2x4 - 5Lx 3 + 3L 2x 2)
Figure 3.62 Maximizing the amount where A: is a positive constant. Where does the maximum
of light deflection occur on the beam?
31. A resistor of R ohms is connected across a battery of E
28. Figure 3.63 shows a thin lens located p cm from an object volts whose internal resistance is r ohms. According to the
AB and q cm from the image RS. principles of electricity, a current of
E
R +r
amperes will flow in the circuit and will generate
P = I 2R
watts of power in the external resistor. Assuming that E
and r are constant, what value of R will result in maxi
Figure 3.63 Image from a lens mum power in the external resistor?
32. A lamp with adjustable height hangs directly above the express cos θ and cos 20 in terms of x and then use the
center of a circular kitchen table that is 8 ft in diameter. trigonometric identity
The illumination at the edge of the table is directly
proportional to the cosine of the angle θ and inversely cos 20 = 2 cos2 0 - 1
proportional to the square of the distance d, given θ and d to eliminate 0 and express L in terms of x alone.
as shown in Figure 3.65
35. Find the length of the longest pipe that can be carried
horizontally around a corner joining two corridors that
are 2∖ z2 ft wide, as shown in Figure 3.67.
Figure 3.65 Illumination on a kitchen table

a. Let I denote the illumination at the edge of the table.
Show that
7(0) = γg cos θ sin 2 θ
Find Z'(6>).
b. Show that ∕'(0 ) = 0 when tan θ = V z2. What are sin θ
and cos 0?
c. How close to the table (to the nearest ft) should the
lamp be pulled to maximize the illumination at the
36. Congruent triangles are cut out of a square piece of paper
edge of the table?
20 in. on a side, leaving a star-like figure that can be
33. In Problem 32, suppose the illumination is directly pro folded to form a pyramid, as indicated in Figure 3.68.
portional to the sine of the angle φ and inversely propor
tional to the square of the distance d, where φ is the angle
at which the ray of light meets the table. How close to the
table (to the nearest ft) should the lamp be pulled to
maximize the illumination at the edge of the table? Is
your answer the same as in Problem 32? Explain why or
why not.
© 34. The lower right-hand corner of a piece of paper is folded
over to reach the leftmost edge, as shown in Figure 3.66.
Figure 3.68 Constructing a pyramid
What is the largest volume of the pyramid that can be

formed in this manner?
37. It is known that water expands and contracts according to

its temperature. Physical experiments suggest that an
amount of water that occupies 1 liter at 0 oC will occupy
V(T) = 1 - 6.42 ∙ 10~5 T + 8.51 ■10^6 Γ 2 - 6.79 × 1O^8 T 3
liters when the temperature is T 0C. At what temperature
is F(T) minimized? This seems to suggest that the maxi
Figure 3.66 Paper folding problem mum density o f water occurs at a temperature above its
If the page is 20 cm wide and 30 cm long, what is the freezing point (0 °), and explains why ice forms only at the
length L of the shortest possible crease? Hint: First upper levels of a lake during winter.
3 .8 OPTIMIZATION IN BUSINESS, ECONOMICS, AND THE LIFE SCIENCES
IN THIS SECTION Optimization of discrete functions, economics: two

principles of business management; marginal analysis; an inventory model;
physiology: the optimal angle for vascular branching
Optim ization problems appear in many different areas that use calculus as a
m ethod of analysis. We have already examined a few of these applications in
Section 3.7, and this section deals with more specialized applications.
■ OPTIMIZATION OF DISCRETE FUNCTIONS
Sometimes the function to be maximized has practical meaning only when its
independent variable is a positive integer. This can lead to certain problems when
applying the processes of calculus, because the theorems we have developed
require continuous functions. If we model a function whose variable is defined as
a positive integer, but as part of the modeling process we assume the function is
defined for all real values (so that it is continuous), it may happen that the
optim ization procedure leads to a nonintegral value of the independent variable,
and additional analysis is needed to obtain a meaningful solution.
EXAM PLE 1 Maximizing a discrete revenue function
X No. Fare Revenue A bus company will charter a bus that holds 50 people to groups of 35 or more.
∩ ? ιoo If a group contains exactly 35 people, each person pays $60. In larger groups,
75 6∩
everybody’s fare is reduced by $1 for each person in excess of 35. Determ ine
JO
the size of the group for which the bus company’s revenue will be the greatest.
2 37 58 2,146
3 38 57 2,166 Solution We wish to maximize the revenue:
4 39 56 2,184
5 40 55 2,200 REVENUE = (NUMBER OF PEOPLE IN THE GROUP)(FARE PER PERSON)
6 41 54 2,214
7 42 53 2,226 Let x be the num ber of people in excess of 35 who take the trip. Then,
8 43 52 2,236 NUMBER OF PEOPLE IN THE GROUP = 35 + X
9 44 51 2,244 FARE PER PERSON = 6 0 - X
10 45 50 2,250
Let JR(X) be the revenue for the bus company:
11 46 49 2,254
12 47 48 2,256. R(x) = (35 + x)(60 - x) = 2,100 + 25x - x 2
13 48 47 2,256 Next, find the domain: We note that there must be at least 35 people
14 49 46 2,254 (x = 0) and at most 50 people (x = 15); thus 0 ≤ x ≤ 15, but because x
15 50/ 45 2,250 represents a number o f people, it must also be an integer.
The critical values are found by looking at the derivative.
R '(x) = 25 - 2x
Because the derivative exists throughout the interval, the only critical value is
x = 12.5. But x must be an integer, so x = 12.5 is not in the domain. To find
the optimal integer solution, observe that R is increasing on (0, 12.5) and
decreasing on (12.5, 15), as shown in Figure 3.69. If follows that the opti-
Sec. 3.8 Optimization in Business, Economics, and the Life Sciences 231
mal integer value of x is either x = 12 or x = 13. Because

Λ(12) = 2,256 and Λ(13) = 2,256
we conclude that the bus company’s revenue will be greatest when the group
contains either 12 or 13 people in excess of 35—that is, for groups of 47 or 48.
In either case, the revenue will be $2,256.
The graph of revenue as a function of x is a collection of discrete points
corresponding to the integer values of x, as indicated in Figure 3.69b.
Technically, calculus cannot be used to study such a function, so instead, we
worked with the differentiable function R(x) = - x 2 + 25x + 2,100, which is
defined for all values of x and whose graph “connects” the points in the
discrete graph. After applying calculus to this continuous model, we obtained
a mathematical solution that was not the solution of the discrete practical
problem, but that did suggest where to look for the required solution. To verify
that this integer solution is correct, we can look at the margin table, p. 230,
which shows all possibilities from x = 0 (35 people) to x = 15 (50 people). For
each number of travelers, the total revenue is calculated, and we see that the
maximum value 2,256 is obtained when x = 12 and again when x = 13.
■ ECONOMICS: TWO GENERAL PRINCIPLES OF MARGINAL

ANALYSIS
In Chapter 2, we described marginal analysis as that branch of economics that is
concerned with the way quantities such as price, cost, revenue, and profit vary
with small changes in the level of production. Specifically, recall that if C(x) is the
total cost and p(x) is the market price per unit when .x units of a particular
commodity are produced, then R(x) = xp(x) is the total revenue and P(x) =
R(x) - C(x) is the profit function. These functions are shown in Figure 3.70.
Figure 3.70 Cost, revenue,

and profit functions
In Section 2.8, we worked with marginal quantities and their role as rates of
change, namely, C'(x), the marginal cost, and R ∖x), the marginal revenue*
We will now consider these functions in optimization problems. For example,
the manufacturer certainly would like to know what level of production results in
maximum profit. To solve this problem, we want to maximize the profit function
P(x) = R(x) - C(x)
We differentiate with respect to x and find that
P'(x) = R , (x) - C'(x)
Thus, P'(x) = 0 when R '(x) = C'(x), and by using economic arguments we can
show that a maximum occurs at the corresponding critical point.
Maximum Profit Profit is maximized when marginal revenue equals marginal cost.
EXAM PLE 2 Maximum profit
A manufacturer estimates that when x units of a particular commodity are

produced each month, the total cost (in dollars) will be
C(x) = ∣ x 2 + 4x + 200
and all units can be sold at a price of p(x) = 49 - x dollars per unit.
Determine the price that corresponds to the maximum profit.
Solution The marginal cost is C'(x) = ∣x + 4. The revenue is
R(x) = xp(x) = x(49 - x) = 49x - x 2

The marginal revenue is R'(x) = 49 - 2x. The profit is maximized when
R ∖x) = C ∖x)t
R '(x) = C'(x)
49 - 2x = ∣ x + 4
x = 20
Thus, the price that corresponds to the maximum profit is
jp(20) = 49 - 20 = 29 dollars H
A second general principle of economics involves the following relationship

between marginal cost and the average cost A(x) = C(x)∕x.
Minimum Average Cost Average cost is minimized at the level of production where the marginal cost
equals the average cost.
To justify this second business principle, find the derivative of the average cost
function:
xC'(x) - C(x) = C'(x) - ⅜ j = C'(x) - A(x)
1 ( ’ χ2 x x
*Recall (from Section 2.8) that marginal cost is the instantaneous rate of change of production
cost C with respect to the number of units x produced, and the marginal revenue is the instantaneous
rate of change of revenue with respect to the number of units x produced.
Sec. 3.8 Optimization in Business, Economics, and The Life Sciences 23 3
Thus, A '(x) = 0 when C'(x) = A(x), and once again, economic theory justifies this
result as a minimum (rather than a maximum, which would seem possible using
only calculus).
EXAM PLE 3 Minimizing average cost
A manufacturer estimates that when x units of a particular commodity are

produced each month, the total cost (in dollars) will be
C(x) = ∣ x 2 + 4x + 200
and they can all be sold at a price o fp(x) = 49 - x dollars per unit. Determine
the level of production where the average cost is minimized.
Solution The average cost is A(x) = = ⅛x + 4 + 200x~1 and A(x) is mini

mized when C'(x) = A(x). Thus,
⅛x + 4 = ∣x + 4 + 200x~l
x 2 = 1600 Multiply both sides by 8x and simplify.

If we disregard the negative solution, it follows that the minimal average cost
occurs when x = 40 units. You might have noticed that we did not use the
information p(x) = 49 - x in arriving at the solution. It is included, because
when doing real-world modeling it is often necessary to make a choice about
which parts of the available information are necessary to solve the problem.
The average cost, marginal cost, marginal revenue, and total profit graphs for
Figure 3.71 Average cost, marginal cost, marginal revenue, and profit functions
Notice that the marginal revenue and marginal cost graphs intersect below the
peak of the profit graph and that the marginal cost graph intersects the average
cost graph at the lowest point of the average cost graph, as predicted by the theory.
■ BUSINESS MANAGEMENT: AN INVENTORY MODEL
Sometimes mathematical methods can be used to assist managers in making

certain business decisions. As an illustration, we shall show how calculus can be
applied to a problem involving inventory control.
For each shipment of raw materials, a manufacturer must pay an ordering fee
to cover handling and transportation. When the raw materials arrive, they must be
stored until needed and storage costs result. If each shipment of raw materials is
large, few shipments will be needed and ordering costs will be low, but storage
costs will be high. If each shipment is small, ordering costs will be high because
many shipments will be needed, but storage cost will be low. A manufacturer
would like to determine the shipment size that will minimize the total cost. Here is
an example of how the problem may be solved using calculus.
EXAM PLE 4 Placing orders to minimize cost
A retailer buys 6,000 calculator batteries a year from a distributor and is

trying to decide how often to order the batteries. The ordering fee is $20 per
shipment, the storage cost is $0.96 per battery per year, and each battery costs
the retailer $0.25. Suppose that the batteries are sold at a constant rate
throughout the year, and that each shipment arrives just as the preceding
shipment has been used up. How many batteries should the retailer order each
time to minimize the total cost?
Solution We begin by writing the cost function:
TOTAL COST = STORAGE COST + ORDERING COST + COST OF BATTERIES
We need to find an expression for each of these unknowns. Assume that the
same number of batteries must be ordered each time an order is placed;
denote this number by x so that C(x) is the corresponding total cost.
AVERAGE NUMBER ∖ ∕C O S T OF STORING 1∖ _ ∕y ∖ .θ og∖ _ θ 40
STORAGE COST = IN STORAGE PER Y R ∕∖ BATTERY FOR 1 YR 7 \ 2 / U ' ' ' X
The average number of batteries in storage during the year is half of a given
order, that is, xΓ2. Thus, the total yearly storage cost is the same as if x/2
batteries were kept in storage for the entire year. This situation is shown in
Figure 3.72. In order to find the total ordering cost, we can multiply the
ordering cost per shipment by the number of shipments. We also note that
Figure 3.72 Inventory graph because 6,000 batteries are ordered during the year and because each
shipment contains x batteries, the number of shipments is 6,000/x
/ORDERING COST ) NUMBER OF 120,000
ORDERING COST =
∖ PER SHIPMENT / SHIPMENTS
^ r- , /TOTAL NUMBER) /COST PER) , cnn
COST OF BATTERIESr = = 6,000(0.25)7 = 1,500
∖ OF BATTERIES / ∖ BATTERY / ’ ’
Thus the total cost function is
,o 120,000 , c n n
C(x)λ = n0.48x H------ —------1- 1,500
The goal is to minimize C(x) on (0, 6,000]. To find the critical values, we
differentiate C'(x)
C'(x) = 0.48 - 120,000χ-2
and then solve C'(x) = 0 to find
0.48x 2 = 120,000
x = ±500
The root x = -5 0 0 does not lie in the interval (0, 6000]. It is easy to check that
C is decreasing on (0, 500) and increasing on (500, 6000], as shown in Figure
Figure 3.73 The total cost 3.73. Thus, the absolute minimum of C on the interval (0, 6000] occurs when
function x = 500, and we conclude that to minimize cost, the manufacturer should
C(x) = 0.48x + 120,000x~1 + 1,500 order the batteries in lots of 500.
The simple inventory model analyzed in Example 4 allows no shortages', that

is, inventory is never allowed to become depleted during the order cycle. An
inventory model that allows shortages is more realistic, but it is also more difficult
to analyze.
■ PHYSIOLOGY: THE OPTIMAL ANGLE FOR VASCULAR

BRANCHING
The blood vascular system operates in such a way that the circulation of
blood— from the heart, through the organs of the body, and back to the heart— is
accomplished with as little expenditure of energy as possible. Thus, it is reason
able to expect that when an artery branches, the angle between the “ parent” artery
and its “ daughter” should minimize the total resistance to the flow of blood.
B D Figure 3.74 shows a small artery of radius r branching from a larger artery of
Figure 3.74 Vascular branching
radius R . Blood flows in the direction of the arrows from point A to the branch at
B and then to points C and D. For simplicity, we assume that C and D are located
in such a way that C B is perpendicular to the main line through A , B, and D. We
wish to find the value of the branching angle θ that minimizes the total resistance
to the flow of blood as it moves from A to B and then to point C , which is located a
fixed perpendicular distance h from the line through A and B.*
Poiseuille’s Law The resistance to the flow of blood in an artery is directly proportional to the
artery’s length and inversely proportional to the fourth power of its radius.
According to Poiseuille’s law, the resistance to flow from A to B is

ks.
•A = £4
and the resistance from B to C is

∕cs,
c ∕ 2 = 7 T
where k is a viscosity constant, s 1 = ∣AB∣, and s 2 = ∣-BC∣. Thus, the total resistance
to flow is
f = f + r - _k_s.1-∣___k__s^2 = > /_.s,i _ , _s,∖
21
A B D J J ∖ J2 τ
r 4 \7? *4 Γ4 ∕
1------- -t------- 1 The next task is to write/as a function of θ. In order to do this, reconsider Figure
Figure 3.75 Minimizing the resist 3.74 by labeling s 1, s 2 , h, and f as shown in Figure 3.75. We want to find equations
ance to blood flow for s 1 and s2 in terms of h and 0; to this end we notice that
sin θ = ⅛- tan θ =
52 = iE l I - sl = t a ∏0
= Λcsc0 s t= i ~^Γ θ
= f - h cot θ
*The key to solving this problem is a result due to work of the 19th-century French physiologist,
Jean Louis Poiseuille (1789-1869). Our discussion of vascular branching is adapted from Introduction
to Mathematics for Life Scientists, 2nd ed., by Edward Batschelet (New York: Springer-Verlag, 1976,
pp. 278-280). In this excellent little book, Batschelet develops a number of interesting applications of
calculus, several of which appear in the problem set.
We can now w r ite / a s a function o f θ'.
/(0) = k i{ s- ±. + -μ
f ∖) = ∖J I - A cot 0 + h esc 0∖
R4 r4 /
To m in im iz e / w e need to find the critical values (remember that h, ( , and k are
constants):
df Γ - ∕z ( - c s c 2 0) Λ ( -c s c 0 c o t0 )' esc 0 cot 0*∖
Tθ ~ k [ R4 r4 = kh esc 0
R4 r4 /
Because esc 0, k , and h are never zero, we see that the derivative is zero when
/esc 0 _ cot 0∖ _
∖ Ri r4 )
CSC 0 _ cot 0
/?4 r4
1 _ cos 0
R 4 sin 0 r 4 sin 0
r4
= COS 0
By finding the second derivative and noting that r and R are both positive, we can
show that any value θm that satisfies this equation yields a value f ( β n ) that is a
m inim um .
PROBLEM SET 3 .8
© In Problems 1-3, we give the cost C o f producing x units o f a and that the company receives twice as much for selling a
particular commodity and the selling price p when x units are Mopsy doll as for selling a Flopsy doll. Find the level of
produced. In each case, fin d the profit function P and determine production for both x and y for which the total revenue
the level o f production that maximizes profit. derived from selling these dolls is maximized. You may
1. C(x) = ∣ x 2 + 5x + 98 and p(x) = ∣(75 - x) assume that the company sells every doll it produces.
7. A manufacturer knows that when p dollars are charged for
2. C(x) = j x 2 + 3x + 10 and jp(x) = ∣(45 - x) every unit of a product, the sales will be x = 380 - 20p
3. C(x) = ∣(x + 30) and p(x) = ∣ ⅛ units. At this level o f production, the average cost is
4. Suppose the total cost o f producing x units of a certain T(x) = 5 + ^

commodity is
a. Find the total revenue and total cost functions, and
C(x) = 2x 4 - 10x 3 — 18x 2 + x + 5 express the profit as a function o f x.
Determine the largest and smallest values of the marginal b. What price should the manufacturer charge to maxi
cost for 0 < x < 5. mize profit? What is the maximum profit?
5. Suppose the total cost o f manufacturing x units of a © 8. The owner o f the Pill Boxx drug store expects to sell 600
certain commodity is bottles of hair spray each year. Each bottle costs $4, and
the ordering fee is $30 per shipment. In addition, it cost
C(x) = 3x 2 + x + 48 90d per year to store each bottle. Assuming that the hair
dollars. Determine the minimum average cost. spray sells at a uniform rate throughout the year and that
each shipment arrives just as the last bottle from the
6. A toy manufacturer produces an inexpensive doll (Flopsy)
previous shipment is sold, how frequently should ship
and an expensive doll (Mopsy) in units of x hundreds and
ments o f hair spray be ordered to minimize the total cost?
y hundreds, respectively. Suppose it is possible to produce
the dolls in such a way that 9. An electronics firm uses 18,000 cases of connectors each
year. The cost of storing one case for a year is $4.50, and
_ 82 - 10x the ordering fee is $20 per shipment. Assume that the
y 10- x 0 ≤ x ≤ 8
connectors are used at a constant rate throughout the year
and that each shipment arrives just as the preceding source of the pollution according to the principle that for
shipment has been used up. How many cases should the distances greater than or equal to 1 mi, the concentration
firm order each time to keep total cost to a minimum? of particulate matter (in parts per million, ppm) decreases
10. Suppose the total cost (in dollars) of manufacturing x as the reciprocal of the distance from the source. This
units of a certain commodity is means that if you live 3 mi from a plant emitting 60 ppm,
C(x) = 3x 2 + 5x + 75
a. At what level of production is the average cost per unit
the smallest?
b. At what level of production is the average cost per unit
equal to the marginal cost?
c. Graph the average cost and the marginal cost functions
on the same set of axes, for x > 0.
11. Suppose the total revenue (in dollars) from the sale of x
units of a certain commodity is
R(x) = - 2 x 2 + 68x - 128
a. At what level of sales is the average revenue per unit the pollution at your home is y = 20 ppm. On the other
equal to the marginal revenue? hand, if you live 10 miles from the plant, the pollution at
your home is = 6 ppm. Suppose that two plants 10 mi
b. Verify that the average revenue is increasing if the level
apart are releasing 60 and 240 ppm, respectively. At what
of sales is less than the level in part a and decreasing if
point between the plants is the pollution a minimum?
the level of sales is greater than the level in part a.
17. A tour agency is booking a tour and has 100 people signed
c. On the same set of axes, graph the relevant portions of
the average and marginal revenue functions. up. The price of a ticket is $2,000 per person. The agency
has booked a plane seating 150 people at a cost of
12. A manufacturer finds that the demand function for a $125,000. Additional costs to the agency are incidental
certain product is fees of $500 per person. For each $5 that the price is
i <", ■ ¾
lowered, a new person will sign up. How much should the
price be lowered for all participants to maximize the
Should the price p be raised or lowered in order to profit to the tour agency?
increase consumer expenditure? Explain your answer. 18. A bookstore can obtain the best seller 20,000 Leagues
13. A store owner expects to sell 0 units of a certain commod Under the Majors from the publisher at a cost of $6 per
ity each year. The cost of each unit is n dollars; it costs s book. The store has been offering the book at a price of
dollars to order each new shipment of N units; and it costs $30 per copy and has been selling 200 copies a month at
t dollars to store each unit for a year. Assuming that the this price. The bookstore is planning to lower its price to
commodity is used at a constant rate throughout the year stimulate sales and estimates that for each $2 reduction in
and that each shipment arrives just as the preceding the price, 20 more books will be sold per month. At what
shipment has been used up, show that the total cost of price should the bookstore sell the book to generate the
maintaining inventory is minimized when the ordering greatest possible profit?
cost equals the storage cost. 19. A Florida citrus grower estimates that if 60 orange trees
14. A commuter train carries 600 passengers each day from a are planted, the average yield per tree will be 400 oranges.
suburb to a city. It now costs $5 per person to ride the The average yield will decrease by 4 oranges for each
train. A study shows that 50 additional people will ride additional tree planted on the same acreage. How many
the train for each 25¢ reduction in fare. What fare should trees should the grower plant to maximize the total yield?
be charged in order to maximize total revenue? 20. Farmers can get $2 per bushel for their potatoes on July 1,
15. A store has been selling skateboards at the price of $40 per and after that the price drops 24 per bushel per day. On
board, and at this price skaters have been buying 50 July 1, a farmer has 80 bu of potatoes in the field and
boards a month. The owner of the store wishes to raise the estimates that the crop is increasing at the rate of 1 bu per
price and estimates that for each $1 increase in price, 3 day. When should the farmer harvest the potatoes to
fewer boards will be sold each month. If each board costs maximize revenue?
the store $25, at which price should the store sell the 21. A viticulturist estimates that if 50 grapevines are planted
boards in order to maximize profit? per acre, each grapevine will product 150 lb of grapes.
16. As more and more industrial areas are constructed, there Each additional grapevine planted per acre (up to 20),
is a growing need for standards ensuring control of the reduces the average yield per vine by 2 lb. How many
pollutants released into the air. Suppose that the pollution grapevines should be planted to maximize the yield per
at a particular location is based on the distance from the acre?
22. To raise money, a service club has been collecting used 26. In an experiment, a fish swims 5 meters upstream at a
bottles, which it plans to deliver to a local glass company constant velocity v m/sec relative to the water, which itself
for recycling. Because the project began 80 days ago, the has velocity v1 relative to the ground. The results o f the
club has collected 24,000 pounds of glass, for which the experiment suggest that if the fish takes t seconds to reach
glass company offers 1Φ per pound. However, because its goal (that is, to swim 5 meters), the energy it expends is
bottles are accumulating faster than they can be recycled, given by
the company plans to reduce the price it will pay by 1Φper
E = cv k t
100 pounds of used glass. What is the most advantageous
time for the club to conclude its project and deliver all the where c > 0 and k > 2 are physical constants. Assum
bottles? (Assume that the club can continue to collect ing v is known, what velocity v minimizes the energy?
bottles at the same rate and that they plan on making only 27. In certain tissues, cells exist in the shape of circular
one trip to the glass company.) cylinders. Suppose such a cylinder has radius r and height
23. Suppose that the demand function for a certain commodi h. If the volume is fixed (say, at v0), find the value of r that
ty is expressed as minimizes the total surface area (5, = 2πrh + 2τrr2) o f the
cell.
∕Xχ ) = ∕ ⅞ P O ≤X ≤12O
28. In a learning model, two responses (A and 2?) are possible
for each of a series of observations. If there is a probabili
a. Find the total revenue function explicitly and use its ty p of getting response A in any observation, the probabil
first derivative to determine its intervals of increase ity of getting a response A exactly n times in a series of M
and decrease and the price at which revenue is maxi observations is
mized.
F(p) = p«(l - P)M ~"
b. Graph the relevant portions of the demand and reve
nue functions. The maximum likelihood estimate is the value of p that
24. Suppose the demand equation for a certain commodity is maximizes F(p) on [0, 1]. For what value of p does this
linear— that is, occur?
29. During coughing, the diameter of the trachea decreases.
p(x) = l, --- for 0 ≤ x ≤ b The velocity v o f air in the trachea during a cough is
related to the radius r of the trachea during the cough by
where a and b are positive constants. the equation
a. Find the total revenue function explicitly and use its v = A r ∖ r0 - r)
first derivative to determine its intervals of increase
and decrease. where A is a constant and r0 is the radius o f the trachea in
b. Graph the relevant portions of the demand and reve a relaxed state. Find the radius of the trachea when the
nue functions. velocity is greatest, and find the maximum velocity of air.
25. Homing pigeons will rarely fly over large bodies o f water (Notice 0 ≤ r ≤ rθ.)
unless forced to do so, presumably because it requires 30. The work o f V . A. Tucker and K . Schmidt-Koenig*
more energy to maintain altitude in flight over the cool indicates that the energy expended in flight by a certain
water. Suppose a pigeon is released from a boat floating kind of bird is given by
on a lake 3 mi from a point A on the shore and 10 mi away
from the pigeon’s loft, as shown in Figure 3.76. E = ∣[α(v - b)2 + c]
where a, b, and c are positive constants, v is the velocity of

the bird, and the domain of v is [16, 60]. What value o f v
will minimize the energy expenditure in the case where
a = 0.04, b = 36, and c = 9?
31. A plastics firm has received an order from the city
recreation department to manufacture 8,000 special sty
rofoam kickboards for its summer swimming program.
Figure 3.76 Flight path for a pigeon The firm owns 10 machines, each o f which can produce
50 kickboards per hour. The cost of setting up the
machines to produce the kickboards is $800 per machine.
If the pigeon requires twice as much energy to fly over Once the machines have been set up, the operation is fully
water as over land and it follows a path that minimizes
total energy, find the angle θ of its heading as it leaves the *Tucker, V. A. and Schmidt-Koening, K., “Flight of Birds in
boat. This is the situation in the perspective section of the Relation to Energetics and Wind Directions,” The Auk 88 (1971):
chapter introduction. 97-107.
Sec. 3.9 l'H6pital's Rule 239
automated and can be overseen by a single production should be used to minimize the total cost of filling the
supervisor earning $35 per hour. order.
a. How many machines should be used to minimize the b. Show that when the total cost is minimal, the setup
cost of production? Remember, the answer should be a cost is equal to the cost of operating the machines.
positive integer. 33. An important quantity in economic analysis is elasticity
b. How much will the supervisor earn during the produc
o f demand, defined by E(x) = where x is the num
tion run if the optimal number of machines is used?
© 32. Generalize Problem 31. Specifically, a manufacturing ber of units of a commodity demanded when the price is p
firm received an order for Q units of a certain commodity. dollars. Show that
Each of the firm’s machines can produce n units per hour. dR = R∣". 11
The setup cost is .s, dollars per machine, and the operating dx xL £(/?)J
cost is p dollars per hour. That is, show that the marginal revenue is [1 - l∕7Γ(p)]
a. Derive a formula for the number of machines that times average revenue.
3.9 L'H ∂ PITAL 'S RULE

Historical Note IN THIS SECTION ΓH6pitaΓs rule, indeterminate form 0/0, indeterminate
form ∞ ∕∞ , ΓH6pitaΓs rule for other forms
We can use limit theorems for sums, differences, products, and quotients (with
nonzero division) provided certain meaningless expressions are not involved.
For example, lim χ results in the meaningless form g when the quotient limit
theorem is applied. In such cases, we have had to use other means to evaluate
the limit (see Sections 1.6 and 2.3). Johann Bernoulli (1667-1748) developed
an easier method, involving derivatives. We shall introduce this procedure,
called l, H6pital, s rule, and illustrate its use in a variety of problems.
In curve sketching, optimization, and other applications, it is often necessary to

evaluate a limit of the general form
GUILLAUME FRANCOIS
ANTOINE DE L’HOPITAL L = lim 4 4
(1661-1704) x→e g(*)
The French mathematician where c is either a finite number or +∞ or -∞. If lim g(x) ≠ 0, then the quotient
Marquis de l’Hopital published
the rule in the late 17th century rule can be used. However, if both the numerator and denominator approach zero,
in what is often regarded as the limit can be any real number or it can diverge to +∞ or -∞. For this reason, a
the first calculus text. Actually, limit of this kind is said to be a indeterminate form. Similarly, if lim ∕(x) and
the rule was discovered by
ΓHδpitaΓs teacher, Johann Ber lim g(x) are both infinite, lim g(x) is called an indeterminate form. For example:
noulli. Not only did l’Hopital
neglect to cite his sources in his Qin w 11
lim -—- is a X indeterminate form.
book, but there is also evidence .γ →o x U
that he paid Bernoulli for his x2—
results and for keeping their ar h1∙m √2⅝z
+ -1 i∙s a„n —indeterminate form,
- 3÷x----
.γ →=c 3 χ + 5x — 2 ∞
rangements for payment confi
dential. In a letter dated March
17, 1694, he asked Bernoulli
■ L’HOPITAL’S RULE
“to communicate to me your
discoveries . . .” — with the re
quest not to mention them to Earlier in this text, we evaluated certain indeterminate forms such as lim ⅜-- ~j~
others—“ . . . it would not please using algebraic procedures, but not all indeterminate forms can be handled in this
me if they were made public.”
manner. Fortunately, there is a procedure, called l, Hδpital, s rule, that allows us to
∕7x) ∙ ∙ ∙
relate an indeterm inate form lim —4 4 to the limit of the ratio of derivatives,
f'(χ } x →c 8 ( ’
lim 7 ' Here is a formal statement of the rule.
THEOREM 3.9 ΓHδpital, s rule
Let f and g be functions that are differentiable on an open interval containing c

Computational Window a n
(except possibly at c itself). If
An interesting article on lim t⅛
ΓH6pital, s rule with the com
Λ→<∙ g W
puter program Mathematica
can be found in The AMATYC 0 θθ
produces an indeterm inate form or —>then
Review, Fall 1991, p. 40.
lt a t M = lim
x→e g(x) x→c
provided the limit on the right-hand side of the equation exists (or is infinite).
L’Hopital’s rule also applies in the case where x → + ∞ (or x → - ∞) instead
of x → c.
_ f(x)
■ W hat this says: If the lim ~ τ χ yields one of the indeterm inate forms of the
type θ or then you can evaluate the limit by replacing f with f ' and g with
g '. This does not mean you use the quotient rule for derivatives on f( x ) ∣ g(x).
Moreover, you can use l, H δpital , s rule on a reduced rational expression as long
as you still get one of the forms
0 or
+∞ oγ
+∞ or
—∞ or
(∩J 3S∞7
+ ~∞
—T +3∞
Γ7
at x = c. Because ΓHδpitaΓs rule can be applied repeatedly, you can keep

applying it until you get something other than these forms (for either the
num erator or the denominator).
Proof: This proof follows from a more general result called the extended mean
value theorem, which is stated in Appendix B. The proofs of ΓH6pitaΓs rule for
the θ case, as well as the extended mean value theorem are found in Appendix B. ■
■ INDETERMINATE FORM 0 /0
EXAM PLE 1 ΓHδpitaΓs rule of the type 0/0
Evaluate
Solution It is easy to use the algebraic method of Chapter 1 to show that this
limit has the value 2, but as a first example, we will verify that we obtain the
same result using ΓHδpitaΓs rule instead of algebraic simplification. We first
note that the conditions for ΓHδpitaΓs rule apply; this is of the form 0/0. Then,
Sec. 3.9 l'HδpitaΓs Rule 241
lim ~ — i = lim ⅜ = 2
x→l X — 1 χ→l 1
EXAM PLE 2 l , Hδpital , s rule of the type 0/0
Evaluate
lim ⅛ ≡
.v→2 X 3 - 8
Solution For this example,
∕( x ) = x 7 - 128 and g(x) = x 3 - 8
and the form is 0/0. Notice that l, HδpitaΓs rule tells us that
I, m ⅛ J J > . l i m ⅛ _ lim 7 X . Z l ∣ r _ 112
χ→2 χ i - 8 χ→2 3x 2 X →2 3 3 3
EXAM PLE 3 ΓH6pitaΓs rule with trigonometric functions
Evaluate
lim 1 ~.c o s 3 χ
Λ→O sιn-χ
Solution This is the indeterminate form 0/0; by applying ΓHδpitaΓs rule, we
obtain
ι∙ 1 - cos3x .■ - 3 cos2x (-s ιn x ) ,. 3 COSY 3
hm 1 . , λ = hm — ~ -. --------- = hm j u T x = ⅛
A-→O sιn⅛ ι →o 2 s ιn x c o s x .v →o 2 2
You can use a graphing calculator to help find many

of these limits. For example, the graph of
_ 1— cos x ∙s s ]1 0 w n a t t jι e r ight. Note that as
sιn 2 x b
x → 0 from either the left or the right, the limit is the

same, 1.5. The graphs of some of the functions for the
remainder of the examples of this section are shown
so that you can note the relationship between the
requested limit, the function, and ΓHδpitaΓs rule. l√∙∣ Ξ< l-√ co ≥ X ) Λ 3 )
Such analysis using a graphing calculator is not essen z < s in X)^2
tial when working these problems, especially because X n in = - 10 Y m ∏ = J
graphing many of these functions is more difficult Xnax=10 ⅛nax=5
than finding the limits. However, if you have access to X s c l= l V s c l= l
EXAM PLE 4 l , H6pital , s rule applied more than one time
Evaluate lim —— i
Λ-→O x
Solution This is a 0/0 indeterminate form, and we find that
lιm i z p < lιm Lz-ces∆-

x →o x i χ→o 3x 2
This is still the indeterminate form 0/0, so ΓHδpitaΓs rule can be applied once
again:
lim T ^C 2SX = lim - ( - ≡ ) = 1 lim sm x = 1( 1 ) = 1
By looking at the graph, you Λ-→0 3χ2 x →0 OX O χ →0 X 0 v 0
can reinforce the result obtained
using ΓH6pitaΓs rule.
EXAM PLE 5 Limit is not an indeterminate form
Evaluate lim 1
x →o secx
Solution
0 If you apply ΓH6pitaΓs rule blindly, you obtain the W R O N G answer:
lim -t— = lim — ' ,' l— T h i s is N O T c o r r e c t.

A -→O secx χ→o secx tanx
=1 0
ΓHδpitaΓs rule does not apply here, because this limit is n o t an indeterminate
form. In fact,
lim (1 - cosx) n
lim 1 - cosx s N L __________ - = θ = 0
Λ∙→0 secx hm secx 1
.v→0
EXAM PLE 6 Substitution with ΓH6pitaΓs rule
Evaluate lim x t a n - .
χ→+∞ X
Solution This is an ∞ ■0 indeterminate form, but we can convert it into a 0/0
form by using the substitution u = 1/x and noting that u → 0 as x → +∞.
lim x ta n - = lim - tan u = lim t a n t∙∙ T h i s is a 80 f7o r m .

x → +≈ χ u →0 u u →o M
= lim ⅛ = 1
ll→0 1 MM
■ INDETERMINATE FORM ∞ ∕∞
EXAM PLE 7 ΓH6pital, s rule of the type ∞ ∕∞
Evaluate lim ⅛ — ∖x +
χ → +∞ 3x 2 + 5x - 2
Sec. 3.9 l'H6pital's Rule 243
Solution Using the methods of Section 3.5, we can compute this limit by
multiplying by (1∕x 2)∕( 1∕x 2). Instead, let us note that this is of the form ∞∕∞
and apply ΓH6pitaΓs rule.
li 2x 2 - 3x + 1 lim 4 x - 3
Λ + ≈ 3X 2 + 5x - 2 Apply ΓH∂pitaΓs rule again.
X → +o 6x + 5
= lim 4 = ⅜
x—>+∞ OJ
0 You must be careful when applying ΓHδpitaΓs rule. Pay particular

attention to the following points to avoid common mistakes;
1. ΓH6pitaΓs rule involves differentiation of the numerator and the denomi
nator separately. A common mistake is to differentiate the quotient as a
whole by using the quotient rule.
2. ΓHδpitaΓs rule applies only to the indeterminate forms 0/0 and ∞∕∞.
Indiscriminate use of ΓH6pitaΓs rule often leads to erroneous results. For
y _j_ 4Γ
example, note that lim ^y + ∣ = 5, but if you try to apply l, HδpitaΓs rule to
this limit, you will obtain an incorrect result. 0
EXAM PLE 8 l, Hδpital, s rule with a one-sided limit
Show that lim ? j^fec x

→(π∕2)~b + tanx
= 1 for anyj constants a and b.
Solution The limit is indeterminate of the form ∞∕∞. By applying l, HδpitaΓs rule,
lim a + secx_ lim secx ta n x lim ta n ^

x →(ιτ∕2)- D + tanx x →( π ∕2)^ sec2 x x →(π∕2)^ S 6C X
If you apply ΓHδpitaΓs rule again, you will obtain
lim hm
secfx lim
→(π∕2)- secx x→(π∕2)- secx tanx x→(π∕2)-tanx
The curve has a deleted point at Finally, applying l , H6pital, s rule once again;
x = τr∕2. However, if you evalu
ate the function for x = 1.57, lim = lim secx tanx lim ⅛
→<π∕2)-tanx (π∕2)- sec2 x x →(π∕2) - S eC X
you will find the function has a
value very close to 1. Note that You see that for this example, ΓHδpitaΓs rule leads us in circles. Instead, use
left- and right-hand limits to trigonometric identities to write
τr∕2 are the same. Is this true for S1 Π Λ '
x → 2? tan x _ cos.v = sin x
secx ι
C O S Λ'
Thus, lim ia∏Λ' = lim sinx = 1

secx
x →(τr∕2)^ x →(W2)-
■ L'HOPITAL'S RULE FOR OTHER FORMS
Certain limits that do not have either of the standard forms of 0/0 or ∞∕∞ do not
allow direct use of ΓHδpital, s rule. However, the expression can sometimes be
manipulated algebraically into a standard indeterminate form and then can be
evaluated by applying ΓHδpitaΓs rule.
EXAM PLE 9 ΓH 6pitaΓs rule with the form ∞ - ∞
Evaluate lim ( - -
χ → 0 '∖ X
∙J, ,).
v
SHIA./
Solution As it stands, this has the form ∞ - ∞, because
i → +∞ and -4 ----- > +∞ as x → 0 from the right.

x sιnx
However, using a little algebra we find
lim f - - 4 - ) = lim ¾ ⅛
x→o+ ∖x s m x ∕ x→o+ x s ιn x
This limit is now of the form 0/0, so the hypotheses of ΓH6pitaΓs rule are
satisfied. Thus
lim s*n T ~x = lim . cos ,γ ~ 1

— Again, the form
-→O+ x sιnx
Λ χ→o∙ sm x + x cosx ⅛ , √ o
= l im+ ------- -u Γ s i n * ------- = ⅞= 0
X→O COSΛ' + x ( -s ι n x ) + c o s x 2
E X A M P L E 10 ΓH 6pitaΓs rule with the form 0 ∙ ∞
Evaluate lim (x - 5)tanx.

x → π∕2~∖ 2-/
Solution This limit has the form 0 ■∞, because
lim (x - 5) = 0 and lim tanx = +∞

x → π∕2"∖ 2/ x → π∕2^
Write tanx = —τ - to obtain

cot x
lim (x - 5)tanx = lim Ξ___ 2 Form 0/0

x → τr∕2∖ 2/ x → π∕2 cotx
Look at the graph as = lim ---- ¼2 - ΓH∂pitaΓs rule

x → π∕2~ - C S C X
x → ~ from both the left
2
and the right. = lim ( - sin2x) = - 1
x → π∕2
It may happen that even when ΓH6pitaΓs rule applies to a limit, it is not the best
way to proceed, as illustrated by Example 11.
E X A M P L E 11 U sing l , H 6pitaΓs rule with other limit properties
„ . ,. (1 - cosx) sιn4x
Evaluate lim -------- 35-------------.
Λ-→O x cosx
Solution This limit has the form 0/0, but direct application of ΓH6pitaΓs rule
leads to a real mess (try it!). Instead, we compute the given limit by using the
product rule for limits along with two simple applications of ΓH6pitaΓs rule.
Specifically, using the product rule for limits (assuming the limits exist), we have
,. (1 cosx) sin 4x Γ . ι —cosx~l Γ,∙ sιn4x^l Γ.- 1
hm --------,3 -------- = hm -— 2 hm hm ——
x→0 X COSX L Λ-→ 0 X J Lx→ 0 X J Lx→ 0 c o s x j
= him ⅛ 1 him 4 cp s4 xl L J 1 = /n ( 4 ) ( 1 ) = 2
Lx→o 2x J Lx→o 1 J Lx→o cosxj '2 λ λ , ≡≡
ΓH6pitaΓs rule
Sec. 3.9 l'Hδpital's Rule 24 5
EXAMPLE 12 Hypotheses of ΓH6pitaΓs rule are not satisfied
Evaluate lim -x + s *n x .
%→+∞ x - cosx
Solution This limit has the indeterminate form ∞∕∞. If you try to apply
ΓH6pitaΓs rule, you find
lim x + sin x = lim I d: cosx

x→+∞x —cosx x→+∞ 1 + sm x
The limit on the right does not exist, because both sinx and cosx oscillate
between —1 and 1 as x → +∞. Recall that ΓHδpital, s rule applies only if this
limit exists. This does not mean that the limit of the original expression does
not exist or that we cannot find it; it simply means that we cannot apply
ΓH6pitaΓs rule. To find this limit, factor out an x from the numerator and
denominator and proceed as follows:
ι1 +I sinx
x + sm x ~ 1 +0
lim = lim lim : 1
x→+∞X - cosx x → +∞ X→+∞ *I _ COS.Y
v
1 —0
PROBLEM SET 3 .9
θ 1. ■ What Does This Say? An incorrect use of ΓHδpitaΓs
23. lim x 3z2s in - 24. lim (1 - x)cotx
rule is illustrated in the following limit computations. In Λ'→ + ∞ X x→ I
each case, explain what is wrong and find the correct
value of the limit. 2
25. lim ( * - 1 )s ⅛ ( * - 1 ) 26. lim 0 - 12 + cos θ
Λ-→∣ 1 - cos(x - 1) (→ 0 θ + 5θ
a. l i m ^ ^ = l i m ⅜ ^ = 0
.r→ π X x→ π 1
__ __ x + sin(x 2 + x)
27. lim — z— - ÷ ------- 28. lim sec 3x cos 9x
b. lim = lim ∞ = 0 .γ →o∣ 3x + sm x x → π∕2~
x → ττ∕2 X 1x → π∕2
2. ■ What Does This Say? Sometimes ΓH6pitaΓs rule Q 29. lim 'ιλ 1 30. lim ( - U 5 - - J - )
Λ→ 0 . ' x sin x ∕ .v →o ∖sm2x l x )
leads to inconclusive computation. For example, observe
what happens when the rule is applied to
/ 1 1
31. lim 32. lim I cotx - -
x Λ-→ 0 ∖sin2 x x, .v→ 0 ∖ x
lim 2
Vx - 1
2 co sx
Use any method you wish to evaluate this limit. 33. lim 34. lim -=5- r
ι → 0+ sin 2 x γ →-χ∖x - 1
Use ΓH∂pitaΓs rule to evaluate the limits given in Problems .3 3

3-37. 35. lim
4 ∙3 - 27
3. lim ∙ 2 - 16 4. lim ≤ ---- ∣∙ 5. lim 2
c - 4 v - 9 sin 3x sin 2x 37. lim s i n 2 x ∞ s x
36.
∙3 - 1 10 - 1 10 Λ→O x sin 4x χ→π∕2 x sin 4x
6. lim 2 7. lim 8. lim x - 1
∙ - 1
θ 38. Find constants a and b so that
1 - cos2 x 2
10. lim 1 - cos x 11. lim cos 2
9. lim sin2x , a_
sin 3 x Λ'→ 0 3 sinx lim
>,-→ (! v 3 v 2
sin a x
12. lim 1 ~ C22Q
-SΛ' 13. lim ----- y -, ab ≠ 0 14. l i m ⅛ φ
x →o cos ox A-→O sin 5x
x→0 X 39. A weight hanging by a spring is made to vibrate by
1 c s2 γ applying a sinusoidal force, and the displacement at time t
16. lim " , ° -
2¼
15. lim ~ x s ιn x
17. lim ⅛
x →o tanx - x
.r →o x t a n x t→ιτ∕2 2 + tan t is given by
x s
18. lim ÷2 ≡
3χ
19. lim
3 x + s in 3 χ
20. lim 2f - sι∏gx /(0 = —---- X (si∏ at - sin βt)
.v →o x + sm bx or — β~
A-→O x + 2x x→0 x c o s x
where C , a, and β are constants such that a ≠ β. What
21. l i m ⅛ ⅛i 22. lim x 2 sin -
x→0 X + X happens to f(t) as a → β1 You may assume that β is fixed.
3 . 1 0 ANTIDERIVATIVES
IN THIS SECTION Reversing differentiation, antiderivative notation,

antidifferentiation formulas, applications
Our goal in this section is to study a process called antidifferentiation, which
reverses differentiation in much the same way that division reverses m ultiplica
tion. We shall examine algebraic properties of the process and several useful
applications.
■ REVERSING DIFFERENTIATION
A physicist who knows the acceleration of a particle may want to determine its
velocity or its position at a particular time. An ecologist who knows the rate at
which a certain pollutant is being absorbed by a particular species of fish might
want to know the actual amount of pollutant in the fish’s system at a given time. In
each of these cases, a derivative f ' is given and the problem is that of finding the
corresponding function f. Toward this end, we make the following definition.
Antiderivative An antiderivative of a fu n c tio n /is a function F th a t satisfies
Suppose we know ∕( x ) = 3x 2 . We want to find a function F(xj so that F '(x) = 3x 2 .

It is not difficult to use the power rule in reverse to discover that F(x) = x 3 is such
a function. However, that is not the only possibility:
Given: F(x) = x 3 G(x) = x 3 - 5 H(x) = x 3 + τr2
Find: F '(x ) = 3x 2 G∖ x) = 3x 2 H '(x) = 3x 2
Are there other functions that have a derivative equal to / ? Clearly, there are
infinitely many, and they all seem to differ by a constant. If F is an antiderivative
o f/ then so is F + C for any constant C, because
[F(x) + Q ' = F '(x) + 0 = ∕( x )
Conversely, the following theorem shows that any antiderivative of f can be
expressed in this form. This is the constant difference theorem we proved in
Section 3.2.
THEOREM 3.10 Antiderivatives differ by a constant

If F is an antiderivative of the continuous fu n c tio n / then any other antiderivative
of Jf must have the form , , , „
G(X ) = F(x) + C
I W hat this says: Two antiderivatives of the same function differ by a

constant (which may be 0).
Proof: If F and G are both antiderivatives of f then F ' = / and G' = f and
Theorem 3.5 (the constant difference theorem) tells us that G(x) — F(x) = C, so
G(x) = F(x) + C. ■
Sec. 3.10 Antiderivatives 247
Also, keep in mind the difference between a particular antiderivative and a general
antiderivative. For the function defined by∕(x) = 3x2 we note that F(x) = x 3 is a
particular antiderivative and (7(x) = x 3 + C is a general antiderivative.
EXAM PLE 1 Finding antiderivatives
Find general antiderivatives for the given functions.
a. ∕(x ) = x 5 b. s(x) = sin x
Solution
a. If F(x) = x 6, then F'(x) = 6x5, so we see that a particular antiderivative of/’
is F(x) = -g- in order to obtain F'(x) = ⅛ - = x 5. By Theorem 3.10, the
• ∙ ∙ Λ *6
most general antiderivative is G(x) = =3⅛- + C.
b. If 5,(x) = -cosx, then 5'(x) = sinx, so G(x) = -co sx + C. e™
■ ANTIDERIVATIVE NOTATION
It is worthwhile to define a notation to indicate the operation of antidifferentia

tion.
Indefinite Integral The notation

I ∕(x ) dx = F(x) + C
where C is an arbitrary constant means that F is an antiderivative of f It is ⅛
called the indefinite integral of f and satisfies the condition that F'(x) = ∕(x )
for all x in the domain of f.
What This Says: This is nothing more than the definition of antiderivative,
along with a convenient notation. We also agree that in the context of
antidifferentiation, C is an arbitrary constant.
The graph of F(x) + C for different values of C is called a family of functions (see
Figure 3.77). Because each member of the family y = F(x) + C has the same
derivative at x, the slope of each graph at x is the same. This means that the graph
Figure 3.77 Several members of of all functions of the form y = F(x) + C is a collection of curves, as shown in
the family of curves y = F(x) + C Figure 3.77.
<2) It is important to remember The process of finding indefinite integrals is called indefinite integration.
that [ f ( x ) dx represents a fami Notice that this process amounts to finding an antiderivative o f f and adding on an
arbitrary constant C, which is called the constant of integration.
ly of functions. 0
EXAM PLE 2 Antidifferentiation
Find
a. j 5x3 dx b. J sec2 x dx c. j secx tanx √x
Solution a. Because ^m(x)4 = 4x 3, it follows that = 5x3. Thus,

f Sv4 aX
5x3 dx = + C.
b. Because ^ (ta n x ) = sec2 x, we have sec2 x dx = tanx + C
c. Since ^ (se cx ) = secx tanx, we have I secx tanx dx = secx + C

■ ANTIDIFFERENTIATION FORMULAS
Example 2 leads us to state formulas for antidifferentiation. Theorem 3.11

summarizes several fundamental properties of indefinite integrals, each of which
can be derived by reversing an appropriate differentiation formula.
THEOREM 3.11 Basic integration rules
P R O C E D U R A L RU LES Differentiation Formulas Integration Formulas
Constant multiple: ~fd(cf ) = cf f

u
cf(u) du = c J f(ιι) du
Sum rule: -τ -(∕'+ <g) = ∣p + ~Γ L [/(») + g(ιι)∖du = ι f(u)du + I g(u)du

d ι ι j 6 ' du du
Difference rule: 4 - ( f - g) = - ⅛ ∣ giιι)du
duj 6, du du [∕(w) - g(u)]du = J f(u)du - 1
_, ∙ . d , r , , d f , dg [af(u) + bg(u)]du = a J f(u)dιι + b j^ g(u)du
Linearity rule: ^ (« /+ ⅛) = ⅛ + ⅛
Constant rule: ~dfl (c) = θ 0 du = c

ι>n+∖
Power rule: ^ ( w ") ~ nu n ^' u n du = , + C; n Ψ 1
n + I1
Trigonometric rules: ^(cosw ) = -sinw sinw du = - COSM + C
^^(sin u) = cos u cos u du = sin u + C
^^(tan u) = sec2w sec24w

* du = tan u + C
Proof: Each of these parts can be derived by reversing the accompanying

derivative formula. For example, to obtain the power rule, note that if n is any
number other than - 1 , then
4 - I — Γ ι M"+1 1 = — ττ[(fl + 1)M"] = u n
du tn + 1 J n + ll , i
so that — ⅞-rw"+1 is an antiderivative of u'1 and

n+ I
f u n du = — 7—r M"+1 + C for n ≠ -1 _
i n + 1 ■
Now we shall use these rules to compute a number of indefinite integrals.
EXAM PLE 3 Indefinite integral of a polynomial function
Evaluate J (x 5 - 3x 2 - 7) dx.
Solution The first two steps are usually done mentally:
j^ (x 5 - 3x 2 - 7) dx = J x 5 dx - 3 J x 2 dx - 7 J dx
-y-5÷ 1 γ2 + I 1
= g+ 1 - 3 5+ ι- l x + C = ∣ x 6 - x i - 7x + C
Sec. 3.10 Antiderivatives 249
EXAMPLE 4 Indefinite integral with a mixture of forms

Evaluate J (5Vx + 4 sinx) dx.
Solution J (5Vx + 4 sinx) dx = 5J x ιz2 dx + 4 J sinx dx
x 3z2
= 5- Γ^ + 4(-cosx) + C
= -yX3z2 - 4 COSX + C
This concept of antiderivative will be used extensively in integration, and in

the next chapter it will be related to a marvelous result called the fundamental
theorem of calculus.
■ APPLICATIONS
In Chapter 2, we used differentiation to compute the slope at each point on the

graph of a function. Example 5 shows how this procedure can be reversed.
EXAMPLE 5 Given the slope, find the function

The graph of a certain function F has slope 4x 3 - 5 at each point (x, y) and
contains the point (1, 2). Find the function F.
Solution Because the slope of the tangent at each point (x, y) is given by F'(x), we
have
F'(x) = 4x 3 - 5
and it follows that
J F'(x) dx = J (4x3 - 5) dx
F(x) = 4 0 j^ - 5x + C
= x 4 - 5x + C
The family of curves is y = x 4 - 5x + C. To find the one that passes through
(1, 2), substitute:
2 = 1 4 -5 (1 ) + C
6= C
The curve is y = x 4 - 5x + 6. h b
In Section 2.4, we observed that an object moving along a straight line with
displacement s(t) has velocity v(t) = and acceleration α(t) = Thus, we have
v(Z) = J a(f) dt and s(t) = J v(∕) dt
These formulas are used in Example 6.
EXAMPLE 6 Stopping distance for an automobile

The brakes of a certain automobile produce a constant deceleration of 22 ft∕s 2.
If the car is traveling at 60 mi/h (88 ft∕s) when the brakes are applied, how far
will it travel before coming to a complete stop?
Solution Let a(t), v(∕), and s(f) denote the acceleration, velocity, and displace
ment of the car t seconds after the brakes are applied. We shall assume that s is
measured from the point where the brakes are applied, so that s(0) = 0.
v(∕) = J a(t) dt
= J (—22) dt Negative because the car is decelerating
= -2 2 t + Cl v(0) = -22(0) + C1 = 88, so that C1 = 88.

Starting
s
velocity is 88.
= - 2 2 / + 88
Similarly,
s(f) = j^ v(∕) dt
— j ( - 2 2 t + 88) dt
= —111~ ÷ 88/ ÷ C 9
5(0) = - 1 l(0) 2 + 88(0) + c 2 = 0 so that C2 = 0
= —11∕2 + 88/ Starting distance is 0.
Finally, the car comes to rest when its velocity is 0, so we need to solve v(∕) = 0
for /:
- 2 2 / + 88 = 0
/=4
This means that the car decelerates for 4 s before coming to rest, and in that
time it travels
s(4) = -1 1 (4 ) 2 + 88(4) = 176 ft β
Indefinite integration also has applications in business and economics. Recall

that the dem and function for a particular commodity is the function p(x), which
gives the price p that consumers will pay for each unit of the commodity when x
units are brought to market. Then the total revenue is R(x) = xp(x), and the
marginal revenue is R '(x). Our final example of this section shows how the
demand function can be determined from the marginal revenue.
EXAM PLE 7 Finding the demand function given the marginal revenue
A m anufacturer estimates that the marginal revenue of a commodity is

R '(x) = 240 + 0.1x when x units are produced. Find the demand function
p(x).
Solution R(x) = J R'(x) dx
= j^ (240 + 0. lx) dx = 240x + 0.1(¾ 2) + C = 240x + 0.05x 2 + C
Because R(x) = xp(x) where p(x) is the demand function, we must have
R(0) = 0 so that
240(0) + O.O5(O)2 + C = 0 or C = 0
Thus, R(x) = 240x + 0.05x 2 . It follows that the demand function is
P M . ⅛ ) _ 240Λ- ÷ 0.0⅞√ _ 2 4 0 + 0 05 γ _
PROBLEM SET 3 . 1 0
© Find the indefinite integral in Problems 1-22. 29. The marginal cost of a certain commodity is C'(x) =
6x 2 - 2x + 5, where x is the level of production. If it costs
1. ∣ 2 dx 2. I - 4 dx
$5 to produce 1 unit, what is the total cost of producing 5
units?
3. I∣ (2x + 3) dx 4. 1 (4 - 5x) dx
30. The marginal revenue o f a certain commodity is
F '(x) = —3x 2 + 4x + 32, where x is the level o f produc
5. I (4P + 3t2) dt 6. I∣ ( -8 t 3 + 15t5) dt tion (in thousands). Assume 7?(x) = 0 when x = 0.
3 a. Find the demand function p(x).
7. 1 (6u2 - 3 cos ιi) du 8. 1∣ (5t - V t) dt
b. Find the level of production that results in maximum
revenue. What is the market price per unit at this level
9. 1' sec2 θ dθ 10. 1 sec θ tan θ dθ
of production?
31. It is estimated that t months from now the population o f a
11. ∣
' 2 sin θ dθ 12. 1I 3 cos θ dθ
certain town will be changing at the rate of 4 + 5/2/3
people per month. If the current population is 10,000,
13. I (w3z2 - M1Z2 + u^ 10) du 14. 1I (x 3 - 3x + V x - 5) dx what will the population be 8 months from now?
32. A particle travels along the x-axis so that its acceleration
15. 1 x(x + V x) dx 2
16. ' y ( y - 3y) dy at time t is α(t) = V t + t2 . If it starts at the origin (that is,
r z
5(0) = 0) with an initial velocity o f 2 (that is, v(0) = 2),
i7 ∙ 1 (⅛ - A + τ ¼ 18∙ 1U ( l - l ∖ d t
2
determine its position and velocity when t = 4.
∖t2 t3 J t ∖t t3)
33. An automobile starts from rest (that is, v(0) = 0) and
ι9 ∙ 1 (2x2 + 5)2 dx 20. 1' (3 - 4x 3)2 dx travels with constant acceleration a(f) = k in such a way
that 6 s after it begins to move, it has traveled 360 ft from
r -y2 + 1
2 i∙ 1f"(*2 + 3f ~1) dx 22. I
^ ÷ √x its starting point. What is k?
∖ X J J x2 34. The price of bacon is currently S1.80∕lb pound in Styx-
ville. A consumer service has conducted a study which
The slope F'(x) at each point on a graph is given in Problems
predicts that t months from now, the price will be
23-26 along with one point (x0, y 0) on the graph. Use this
changing at the rate of 0.984 + 0.012∖Γt cents per month.
information to fin d F.
How much will a pound of bacon cost 4 mo from now?
23. slope x 2 + 3x with point (0, 0)
35. An airplane has a constant acceleration while moving
24. slope ( V x + 3)2 with point (4, 36)
down the runway from rest. What is the acceleration of
25. slope (2x - 1)2 with point (1,3) the plane at liftoff if the plane requires 100 mi/h and 1.2
26. slope 3 —2 sin x with point (0, 1) mi of runway before lifting off?
θ 27. a. If F(x) = J dx find F so that F(l) = 0. 36. The brakes of a certain automobile produce a constant
deceleration of k ft∕s 2 . The car is traveling at 60 m/h (88
b. Sketch the graphs o f y = F(x), y = F(x) + 3, and ft∕s) when the driver is forced to hit the brakes, and it
comes to rest at a point 121 ft from the point where the
y = F W - i∙
brakes were applied. What is k l
c. Find a constant C o so that the largest value of
θ 37. If a, b, and c are constants, use the linearity rule twice to
G(x) = F(x) + C o is 0.
show that
28. A ball is thrown directly upward from ground level with
an initial velocity o f 96 ft∕s. Assuming that the ball’s only [α∕(x) + ⅛(x) + ch(x)]dx
acceleration is that due to gravity (that is, a(l) = —32
ft∕s 2), determine the maximum height reached by the ball = a I fix ) dx + b g(x) dx + c h(x) dx
and the time it takes to return to ground level.
C H A P T E R 3 R E V IE W
IN T H IS R E V IE W Proficiency examination, supplementary problems

C o n c e p t P r o b le m s 16. lim ( - - 4 = ) 17. lim

x →+≈ ∖x v x ∕ *→+ ∞ ∖x - 2 x + 2/
1. What is the difference between absolute and relative
extrema o f a function? Sketch the graph o f the function in Problems 18-21. Include
2. State the extreme value theorem. all key features, such as relative extrema, inflection points,
3. What are the critical values o f a function? W hat is the asymptotes, and intercepts.
difference between critical values and critical points? 18. ∕( x ) = x 3 + 3x 2 - 9x + 2 19. ∕( x ) = x ιz3 (27 - x)
4. Outline a procedure for finding the absolute extrema o f 20. ∕( x ) = sin 2 x - 2 co s x for 0 ≤ x ≤ 2π∙
a function on a closed interval [a, ⅛]. 21∙ /(*) = J ⅛
5. State both Rolle’s theorem and the mean value theorem,
and discuss the relationship between the two. 22. If ∕' ( x ) = 3 x 2 + 1, use antidifferentiation to find f .
6. State the first-derivative test. Graph both f and f ' on the same coordinate grid.
7. State the second-derivative test. Suppose /(0) = 10.
8. What is an asymptote? 23. Determine the largest and smallest values o f
9. Define lim f( x ) = L and lim ∕( x ) = +∞. ∕( x ) = x 4 - 2 x 5 + 5

10. Outline a graphing strategy for a function defined by on the closed interval [0, 1].
y = ∕(*)∙ 24. A box is to have a square base, an open top, and volume
11. W hat do we mean by optimization? Outline an optimi o f 2 ft 3. Find the dimensions (to the nearest inch) o f the
zation procedure. box that uses the least amount o f material.
12. State ΓH 6pital , s rule. 25. It is projected that t years from now, the population o f a
13. What do we mean by antidifferentiation? certain suburban community will be
P r a c tic e P r o b le m s m = 2o - τ f j
Evaluate the lim its in Problems 14-17.

lim ⅛⅛⅛ 15. lim -— ∖ λ - thousand people. Use increments to estimate how much
14.
A'→ TT∕2 COSX x→ι x - 1 the population will change during the next quarter-year.
Sketch the graph o f each function in Problems 1-14. Show all 19. lim ( V x 2 - x - x) 20. lim [ V x ( x + ⅛) - x j
Λ'→+∞''
o f the key features, such as relative extrema, inflection points,
asymptotes, and intercepts. 21. lim ^ 111⅞j - 22. lim √γ s ι n * j
λ-→o x + sm x
.γ →o χ - - sιn x
1. ∕( x ) = x 3 + 6 x 2 + 9x — 1 2. ∕( x ) = x 4 + 4 x 3 + 4 x 2 + 1 s i n .2 *
23. lim * Vt ⅝ g
a n x~
4 3 24. lim
3. ∕( x ) = 3x - 4 x + 1 4. ∕( x ) = 3x 4 - 4 x 2 + 1 x→o χ - — sιn 2 x Λ→O
5. ∕( x ) = 6 x 5 - 15x4 + 10x3 lim ⅛ ⅜ 1 1

25. 26. lim
6. ∕( x ) = 3x 5 - 10x3 + 15x+ 1 s ιn r
Λ-→ O x →0 ' vx 2 x 2s e c x ∕
7 ∙ π>-) - ⅛ ⅞ s . ∕(.v ) = ⅛ lim - - ⅛ 28. lim - ( 1 - sin x)tan x
27.
X— >7Γ∕2 sec J X x →π∕2
9. ∕( x ) = sin 2x - s in x 10. ∕( x ) = sin x sin 2x 29. lim - (secx - tanx) 30.
2 2
lim ( V x + 4 - V x - 4
x → τr∕2 A→ +∞
11. ∕( x ) = ⅛ ⅛ 12. = g + 2x - 3
F in d the indefinite integrals o f the functions in Problems
31-34.
13. j∕( X ) = 2
, Λξ 7 i4 ∙ jΛv* ) =/ 4*ιt ---n 3
' ( x + l ) 22
(x -2 ) ' 1 2)
x (x + l)( x + 31. J (5x - 6) d x
Determ ine the m axim um and m inim um value o f each function

32. ∣ (/ - 1)(/ + 2) dt
on the interval given in Problems 15-16.
15. ∕( x ) = x 4 - 8x 2 + 12 on [ - 1, 2]
16. ∕( x ) = V x ( x - 5)1/3 on [0, 6] 33. ∣- dx
J x i
Evaluate the lim its in Problems 17-30.
17. lim ⅛ ⅛ 34. [ 5x2 - 2 x + l
18. lim 4 J d χ
x →+∞ 6x + 7 Vx
Λ∙→ +≈ X + 1 J
35. J OURNAL PROBLEM Mathematics Teacher* Which of

the graphs is the derivative and which is the function?
44. Oil from an offshore rig located 3 mi from the shore is to

be pumped to a location on the edge of the shore that is 8
mi east of the rig. The cost of constructing a pipe in the
ocean from the rig to the shore is 1.5 times more expen
sive than the cost of construction on land. How should the
pipe be laid to minimize cost?
36. Determine a, b, and c such that the graph of f(x ) =
ax 2 + bx 2 + c has an inflection point and slope 1 a t( - l,2 ) . 45. In order to obtain the maximum price, a shipment of fruit
should reach the market as early as possible after the fruit
37. ■ What Does This Say? Explain why the graph of a
has been picked. If a grower picks the fruit immediately
quadratic polynomial cannot have a point of inflection. for shipment, 100 cases can be shipped at a profit of $10
How many inflection points does the graph of a cubic per case. By waiting, the grower estimates that the crop
polynomial have? will yield an additional 25 cases per week, but because the
38. Find the points on the hyperbola x 2 —y 2 = 4 that are competitor’s yield will also increase, the grower’s profit
closest to the point (0, 1). will decrease by $1 per case per week. Use calculus to
39. A Norman window consists of a rectangle with a semi determine when the grower should ship the fruit in order
circle surmounted on the top. What are the dimensions of to maximize profit. What will be the maximum profit?
the Norman window of largest area with a fixed perimeter 46. The owner of a novelty store can obtain joy buzzers from
of ∕ , 0 meters? the manufacturer for 40Φ apiece. It is estimated that 60
40. Find numbers A, B, C, and D that guarantee that the buzzers will be sold when the price is $1.20 per buzzer
function and that 10 more buzzers will be sold for every 104
f(x ) = A x 3 + B x 2 + Cx + D decrease in price. What price should be charged in order
will have a relative maximum at (-1 , 1) and a relative to maximize profit?
minimum at (1, —1). 47. Westel Corporation manufactures telephones and has
developed a new cellular phone. Production analysis
41. Find the area of the rectangle of maximum area that can
be inscribed in the semicircle defined by the equation shows that its price must not be less than $50; if x units
are sold, then the price is given by the formula
y = Vα2 - x 2 p(x) = 150 —x. The total cost of producing x units is
for fixed a > 0. given by the formula C(x) = 2,500 + 30x. Find the maxi
42. An apartment complex has 200 units. When the monthly mum profit, and determine the price that should be
rent for each unit is $600, all units are occupied. Experi charged to maximize the profits.
ence indicates that for each $20-per-month increase in 48. The cost of producing a certain book is $25 per copy. The
rent, 5 units will become vacant. Each rented apartment number of books sold is given by the formula
costs the owners of the complex $80 per month to 50,000
maintain. What monthly rent should be charged to maxi n ~ x + 20 2,000
mize the owner’s profit? What is maximum profit? How
many units are rented when profit is a maximum? where x is the price of the book. The profit P(x) is the
43. A peach grower has determined that if 30 trees are planted number of books sold times the price less $29. That is,
per acre, each tree will average 200 lb of peaches per
season. However, for each tree grown in addition to the 30 P(x) = (x - 29)M
trees, the average yield for each of the trees in the grove In order to maximize the profit, what should be the selling
drops by 5 lb per tree. How many peach trees should be price of the book if the domain for x is [29, 150]?
planted on each acre in order to maximize the yield of
49. A manufacturer receives an order for 5,000 items. There
peaches per acre? What is the maximum yield?
are 12 machines available, each of which can produce 25
items per hour. The cost of setting up a machine for a
production run is $50. Once the machines are in opera
tion, the procedure is fully automated and can be super
*December 1990, p. 718. vised by a single worker earning $20 per hour. Find the
number o f machines that should be used in order to

minimize the total cost of filling the order. Describe how these inequalities qualitatively determine
50. The personnel manager of a department store estimates the shape o f the graph o f the functions over the given
that if N temporary salespersons are hired for the holiday interval.
season, the total net revenue derived (in hundreds of 58. f x ) = sin2x on [~ π , π]
dollars) from their efforts will be 59. f x ) = x 3 - x 2 - x + 1 on [ - ∣ , 2]
R(N) = - 3 A 4 + 5OA 3 - 261N 2 + 540A 60. f x ) = x 4 - 2x 2 on [ - 2 , 2]
for 0 ≤ N ≤ 9. How many salespersons should be hired 61. f x ) = x 3 - x + ^ + 1 on [ - 2 , 2]
in order to maximize total net revenue? Hint'. A critical
value for A(Λf ) is 2. 62. Consider a string 60 in. long that is formed into a
rectangle. Using a graphing calculator or a graphing
51. Show that the graph o f a polynomial of degree n, with
program, graph the area Alx) enclosed by the string as
n > 2, has at most n - 2 inflection points.
a function of the length x of a given side
52. Show that the graph of the function ∕( x ) = x" with n > 1 (0 < x < 30). From the graph, deduce that the
has either one or no inflection points, depending on maximum area is enclosed when the rectangle is a
whether n is odd or even. square. What is the minimum area enclosed? Com
53. Each tangent line to the circle x 2 + y 1 = 1 at a point in the pare this problem to the exact solution. What conclu
first quadrant will intersect the coordinate axes at points sions do you draw about the desirability of analytical
(x 1, 0) and (0, j 1). Determine the line for which x 1 + y 1 is solutions and the role o f the computer?
a minimum. 63. The shore o f a lake follows contours that are approxi
54. An accelerated particle moving at speed close to the speed mated by the curve 4x 2 + y 2 = 1, and a nearby road
of light emits power P in the direction θ given by lies along the curve y = 1/x for x > 0. Using your
graphing calculator or computer software, determine
the closest approach o f the road to the lake in the
north-south direction. Take the positive p-axis as
Find the value of cos θ for which P has the greatest value. pointing north.
55. Suppose that f is a continuous function defined on the 64. Using your graphing and differentiation programs,
closed interval [a, b] and that ∕'( x ) = c on the open locate and identify all the relative extrema for the
interval (a, b) for some constant c. Use M V T to show that function defined by
f ix ) = c(x - a) + f a ) f ix ) = ∣ x 5 - ¾x4 + 2x 2
for all x in the interval [a, b]. This may require a few judicious choices of the
56. Suppose that ∕" ( x ) exists and that f i x ) + c 2∕( x ) = 0, plotting interval.
with ∕'( 0 ) = 1. Show that for any xθ, there is a number w 65. Suppose you are a manager o f a fleet of delivery
between 0 and x for which ∕' ( x 0) + c 2∕( x 0 )x = 1. trucks. Each truck has a minimum speed of 15 mi/h
57. An electric charge Q o is distributed uniformly over a ring and a maximum speed of 55 mi/h. Suppose the truck
of radius R . The electric field intensity at any point x burns gas at the rate of
along the axis of the ring is given by
β nx
= (x 2 + A 2)3z2 gal/mi when traveling x mi/h. Using a graphing and
differentiation program (and/or root-solving pro
At what point on the axis is E(x) maximized? gram on your computer or calculator), answer the
following questions:
a. If gas costs S1.70∕gal, estimate the steady speed
that will minimize the cost o f fuel for a 500-mi
trip.
Using the graphing and differentiation programs o f your
computer software, graph f(x), f'(x ), f" ( x ) for each func b. Estimate the steady speed that will minimize the
tion in Problems 58-61. Print out a copy, i f possible. On cost if the driver is paid $28 per hour and the price
each graph, indicate the intervals where of gasoline remains constant at $ 1.70/gal.
a. f i x ) > 0 b. f i x ) < 0 c. f i x ) > 0 66. A mechanics problem The sketch shows a circular
wheel of radius 1 ft moving at an angular speed of 2
d. f i x ) < 0 e- f l f = 0 radians/s in a counterclockwise direction. A rod of
f . f i x ) does not exist g. f i x ) = 0 length L is attached to the circumference, and the
other end o f the rod can move only vertically, as the meter stick to a point of tangency. Recall that
shown in Figure 3.78. from the geometry involved, r > 25.
a. Make a plot of ∕(r ), thus obtaining a rough esti
mate of the desired “ zero” of f
b. Use the Newton-Raphson method to find the
radius of the sphere to 5-decimal-place accuracy.
68. P U T N A M E X A M IN A T IO N P RO BLEM Given the parabola

y2 = 2mx, what is the length of the shortest chord that is
normal to the curve at one end? Hint: If A B is normal to
the parabola, where A and B are the points (lm t 2, 2mt)
and (2ms2, 2ms), show that the slope of the tangent at A is
l∕2 t.
69. P U T N A M E X A M IN A T IO N P RO BLEM Prove that the poly
nomial
a. Argue that a point on the circumference can be (α - x) 6 — 3a(a - x) 5 + ⅜a2(a —x) 4 - yα4(α - x) 2
expressed as (x, y) = (cos2l, sin2t).
has only negative values for 0 < x < a. Hint: Show that
b. Derive the equation for q:
if x = 1g(l - y), the polynomial becomes a 6y 2g(y) where
q(t) = sin 2/ + V L 2 ⅛ cos22t
g( v) = y 4 - 3y 3 + ∣y 2 - i
c. Compute q'(t) without a calculator or computer.
d. From the physics of the problem, argue that and then prove that g(.v) < 0 for 0 < y < 1.
velocity q'(t) = 0 when the (x, y) point is at the top 70. P U T N A M E X A M IN A T IO N P RO BLEM Find the maximum
or the bottom o f the wheel. Then conjecture at value of f(x) = x 3 — 3x on the set of all real numbers x
what point(s) q' will achieve its maximum value. satisfying x 4 + 36 ≤ 13x2 .
71. P U T N A M E X A M IN A T IO N P RO BLEM Let T be an acute
From now on, use some computing help.
triangle with inscribed rectangles R and 5, as shown in
e. Compute q∖f) and compare it with your answer in
Figure 3.79.
part c.
f. For L = 2, plot q(t) and a'(t) and comment on
what you see. Any surprises?
g. Now, for L slightly larger than 1 (that is, 1.1. or
1.01) repeat part f.
h. Find for what values of t, q'(?) achieves its maxi
mum value. Specifically, find at least one t value
to two-decimal-place accuracy. Compare with
your answer in part d. Hint: You may want to look
for zeros o f q"(t).
i. Compare the maximum speed of point (0, q) with
that o f the other end of the rod, (x, y). Figure 3.79 Problem 71
j. Exploratory What do you think happens to q'm a
as L → 1.0? Let a, b, c and x , y, z be the lengths as indicated in Figure
67. Measuring a sphere In Chapter 1, you were intro
3.79. Let A(x) denote the area o f polygon X (rectangle or
duced to a method o f finding the radius of a sphere triangle). Find the maximum value of the ratio
by using a meter stick and a wire. This led to the zl(R) + /1(5)
equation for r: Λ(T)
for all possible choices of the acute triangle T and the

jf (vf )’ = cos— ----- j~ τ = 0
r r+ 1 inscribed rectangles R and 5. Hint: Show that the problem
where 25 represents the distance on the surface from can be restated as, “ Maximize ab + ac + be, subject to
a + b + c = h.”
"Wine Barrel Capacity”
This project is to be clone in groups o f three or four students. Each group will submit
A wine barrel has a hole in the middle of its side called a bung hole. To determine
the volume of wine in the barrel, a bung rod is inserted in the hole until it hits the
lower seam. Determine how to calibrate such a rod so that it will measure the
volume of the wine in the barrel.
You should make the following assumptions:
1. The barrel is cylindrical.
2. The distance from the bung hole to the corner is λ.
3. The ratio of the height to the diameter of the barrel is t. This ratio should be
chosen so that for a given λ value, the volume of the barrel is maximal.
Your paper is not limited to the following questions, but it should include these
concerns: You should show that the volume of the cylindrical barrel is
V = 2τrλ3t(4 + t2)~3,2, and you should find the approximate ideal value for t.
Johannes Kepler was the first person to show mathematically why coopers were
guided in their construction of wine barrels by one rule: make the staves (the
boards that make up the sides of the barrel) one and one-half times as long as the
diameter. (This is the approximate t-value.) You should provide dimensions for
the barrel as well as for the bung rod.
Johannes Kepler (1571-1630) is

usually remembered for his work in
astronomy, in particular for his
three laws of planetary motion.
However, Kepler was also a re
nowned mathematician and served
as court mathematician to the Aus
trian emperor Matthew I. While at
Matthew Γs court, Kepler observed
with admiration the ability of a
young vintner to declare quickly
and easily the capacities of a num
ber of different wine casks. He de
scribes how this can be done in his
book The New Stereometry o f Wine *The idea for this group research project comes from research done at Iowa State University as
Barrels, Mostly Austrian. part of a National Science Foundation grant. Our thanks to Elgin Johnston of Iowa State University.
4
Integration
■ CONTENTS
4.1 Area As the Limit of a Sum;
Summation Notation PREVIEW
Area as the limit of a sum; The
general approximation scheme; If an object moves along a straight line with velocity v(Z), what is its
Summation notation; Area
using summation formulas position s(f) at time Z? If∕( x ) ≥ 0 on the interval [a, b], what is the area
4.2 Riemann Sums and the Defi under the curve y = ∕( x ) over this interval? On Christmas Day, it starts
nite Integral snowing at a steady rate. A snowplow starts out at noon, going 2 miles
Riemann sums; The definite
integral; Area as an integral; during the first hour and 1 mile during the second. What time did it start
Distance as an integral; Proper snowing? At first glance, these problems may appear to have little in
ties of the definite integral common, but they are all typical of what we shall encounter in our study
4.3 The Fundamental Theorem of of integral calculus. In Section 1.1, we pointed out that there are two
Calculus; Integration by Substi
tution fundamental concepts in calculus, namely, the idea of a derivative and
The fundamental theorem of that of an integral. With this chapter we formally introduce you to this
calculus; Substitution with in
definite integration; Substitu second fundamental concept of calculus.
tion with definite integration
4.4 Introduction to Differential
Equations
Introduction and terminology; PERSPECTIVE
Separable differential equations;
Orthogonal trajectories; The
flow of fluid through an orifice;
The key concept in integral calculus is integration, a procedure that involves
Escape velocity of a projectile computing a special kind of limit of sums called the definite integral. We shall
4.5 The Mean Value Theorem for find that such limits can often be computed by reversing the process of
Integrals; Average Value differentiation; that is, given a function/ we find a function Fsuch that F' = f.
Mean value theorem for This is called indefinite integration, and the equation F' = f is an example of a
integrals; Average value of a
function; Variable limits of differential equation.
integration; Leibniz’s rule Finding integrals and solving differential equations are extremely important
4.6 Numerical Integration: The processes in calculus. We begin our study of these topics by defining definite
Trapezoidal Rule and and indefinite integration and showing how they are connected by a remarkable
Simpson’s Rule result called the fundamental theorem of calculus. Then we examine several
Approximation by rectangles;
Trapezoidal rule; Simpson’s techniques of integration and show how area, average value, and other
rule; Error estimation quantities can be set up and analyzed by integration. Our study of differential
4.7 Area Between Two Curves equations begins in this chapter and will continue in appropriate sections
Area between curves; Area by throughout this text. We also establish a mean value theorem for integrals and
vertical strips; Area by horizon develop numerical procedures for estimating the value of a definite integral.
tal strips; Two applications to
economics We close the chapter by showing how integration can be used to find the area
Chapter 4 Review between two curves and to compute certain useful quantities in economics.
Guest Essay, Ralph Boas
257
258 Ch. 4 Integration
4 .1 AREA AS THE LIMIT OF A SUM; SUMMATION NOTATION
Historical Note IN THIS SECTION Area as the limit of a sum, the general approximation
scheme, summation notation, area using summation formulas
In elementary school you were introduced to the concept of area, and you
calculated the areas of certain plane figures with straight-line boundaries. You
were even given a formula for the area of a circle, but, in general, you were not
taught to find the area of a plane region with curved boundaries. The process of
finding an area of a region with curved boundaries is called quadrature. The
area under a curve plays a central role in integral calculus. In this section, we
shall show that it is reasonable to define area as the limit of a sum. In the
process, we shall introduce ideas that play a key role in our general development
of integral calculus.
EARLY ASIAN CALCULUS ■ AREA AS THE LIMIT OF A SUM

The mathematician Seki Kδwa Computing area has been a problem of both theoretical and practical interest since
(1642-1708) measured the area
ancient times; except for a few special cases, the problem is not easy. For example,
of a circle using rectangles. This
you may know the formulas for computing the area of a rectangle, square, triangle,
sample was drawn by a student
of Kδwa. The Japanese called circle, or even a trapezoid. You have probably found the areas o f regions that were
this process for finding area more complicated but could be broken up into parts using these formulas.
yenri. However, how would you find the area between the parabola whose equation is
a. Problem: Compute the area b. The required area is

under the curve y = x~, above approximately the same
the .r-axis and between the lines as the total area bounded
x = 0 and x = 1. by the shaded rectangles.
Figure 4.1 Example of the area problem
EXAM PLE 1 Estimating the area under a parabola using rectangles and
right endpoints
Estimate the area under the parabola y = x 2 on the interval [0, 1].
Solution Observe that although we cannot compute the area by applying a simple
formula, we can certainly estimate the area by adding the area of approximat
ing rectangles constructed on subintervals of [0, 1], as shown in Figure 4. lb.
To simplify computations, we shall require all approximating rectangles to
have the same width and shall take the height o f each rectangle to be the y-
Sec. 4.1 Area As the Limit of a Sum; Summation Notation 259
coordinate of the parabola above the right endpoint of the subinterval on

which it is based.* For the first estimate, we divide the interval [0, 1] into 5
subintervals, as shown in Figure 4.2a. Because the approximating rectangles
all have the same width, the right endpoints are x 1 = 0.2, x 2 = 0.4, x , = 0.6,
x 4 = 0.8 and x 5 = 1. This subdivision is called partitioning the interval. The
width of each subdivision is denoted by ∆x and is found by taking the length
of the interval divided by the number of subintervals:
∆x = i j =-2 = ∣ = 0.2
If we let S n be the total area of n rectangles, we find for the case where n = 5
that
⅛ = ∕( X I) Δ ∙X + f( χ )^ χ
2
+ f ( χ 3)∆ χ + f( χ 4 )∆χ + ∕(∙^ 5)∆χ
= [∕(* 1) + Λ χ
2)
+ ∕(¾) + ∕(⅞) + ∕(⅞)1Δ *
= [/(0.2) + /(0.4) + /(0.6) + /(0.8) + ∕(l)](0.2)
= [0.22 + 0.4 2 + 0.6 2 + 0.8 2 + 12](0.2) = 0.44 _
Even though 1S,5 = 0.44 serves as a reasonable approximation of the area, we

see from Figure 4.2b in the margin that this area is too large. Let us rework
Example 1 using a general scheme rather than a specified number of rectangles.
Partition the interval [0, 1] into n equal parts, each with width
For k = 1, 2, 3, . . . , n, the ∕<th subinterval is

Γ∕c - 1 k ∖
L n , n∖
and on this subinterval, we then construct an approximating rectangle with width
∆x = ~ and height , because y = x 2 . The total area bounded by all n rectangles
is
=[( i R 2+∙∙∙ +f<lM
Δn t,∖n∕Y+∖n) ∖n) ∖∖n)
Consider different choices for n, as shown in Figure 4.3.
n= 4
Figure 4.3 The area estimate is improved by taking more rectangles.
*Actually, there is nothing special about right endpoints, and we could just as easily have used any
other well-defined point in the base subinterval— say, the left endpoint or the midpoint.
n If we increase the number of subdivisions n, the width ∆x = jj of each approximat

5 0.440 ing rectangle will decrease, and we would expect the area estimates S n to improve.
10 0.385 Thus, it is reasonable to define the area /1 under the parabola to be the limit ofS n as
50 0.343 ∆x → 0 or, equivalently, as n → +∞. Although we now have no way of formally
100 0.338 computing this limit, we can attempt to predict its value by seeing what happens
0.334
to the sum as n grows large without bound. It is both tedious and difficult to
1,000
evaluate such sums by hand, but fortunately we can use a calculator or computer
5,000 0.333
to obtain some (rounded) values for S n as shown in the margin. Notice that for
n = 5, the value 0.44 corresponds to the calculation in Example 1.
These computer-generated drawings show the area under the curve y = x 2 on

[1, 5] as approximated by rectangles using left endpoints (in contrast to right
endpoints used in Example 1). The sums in this output are called Riemann
sums, which we will define in the next section.
(Network points axe left endpoints.
NBR. RIEMANN SUM OVER

TERMS
[Network points are left endpoints.

WANT MORE?
NBR. RIEMANN SUM OVER
TERMS
30
38.375
39.84375
40.58594
■ THE GENERAL APPROXIMATION SCHEME
We now compute the area under any curve y = ∕(x ) on an interval [a, ft] where/is
a nonnegative continuous function. We first partition the interval [a, b] into n
equal subintervals, each of width
∆x = n
For k = 1, 2, 3, . . . , n the ∕cth subinterval is [a + (k - l)∆x, a + fc∆x], and the ∕dh
approximating rectangle is constructed with width ∆x and height f(a + ∕c∆x) equal
to the altitude of the curve y = fix ) above the right endpoint of the subinterval.
Adding the areas of these n rectangles, we obtain
Area of Area of Area of
first rectangle second rectangle ∙ ∙∙ nth rectangle
S n = / ( α + ∆x)∆x + ∕( α + 2∆x)∆x + ∙∙∙ + ∕( α + w∆x)∆x
as an estimate of the area under the curve. In advanced calculus, it is shown that
the continuity of f guarantees the existence of lim S n , and we use this limit to
define the required area, as shown in the following box.
Area As the Limit of a Sum Suppose/is continuous and ∕(x ) ≥ 0 throughout the interval [a, b]. Then the
area of the region under the curve y = ∕(x ) over this interval is
A = Jimθ [∕(α + ∆x) + / ( α + 2∆x) + ∙∙∙ + ∕( α + w∆x)]∆x where ∆x = ~ fl.
The definition of area as a limit of a sum is consistent with the area concept
introduced in plane geometry. For example, it would not be difficult to use this
formula to show that a rectangle has area A = fw or that a triangle has area
A = ∖bh. You will also note that we maintained everyday usage in saying that the
formula for the area of a rectangle is A = tw or the area of the region under the
curve y = fix ) is a limit, but it is actually more proper to say that the limit is the
definition of the area of the region.
The problem we now face is how to implement this definition of area as the
limit of a sum. The immediate answer (discussed in this section) is to use
summation formulas and technology. The long-range goal is to develop integral
calculus, which is discussed in the next section.
■ SUMMATION NOTATION
The expanded form of the sum for the definition of area makes it awkward to use.
Therefore, we shall digress to introduce a more compact notation for sums. Using
this summation notation, we express the sum a ∣ + a 7 + ∙∙∙ + an as follows:
n
a. 1+ «27 + ∙∙∙ + a∏ = Δ-Y
, a,
k =∖
The summation notation is sometimes called the sigma notation because the
uppercase Greek letter sigma (Σ) is used to denote the summation process. The
index k is called the index of summation (or running index). The terminology used
in connection with the summation notation is shown:
Upper limit of summation

General term
V ¾
k≡∖ ↑
j 1 Index of summation
Lower limit of summation
Note that as far as the summation process is concerned, the choice of

summation index is immaterial. For example, the following sums are all exactly
the same:
In general, an index (k,j, i, or λ) that represents a process in which it has no direct

effect on the result is called a dummy variable.
Several useful properties of sums and sum formulas are listed in Theorem 4.1.
We shall use the summation notation throughout the rest of this text, especially in
this chapter, Chapter 6, and Chapter 8.
T H E O R E M 4.1 Basic rules and formulas for sums
For any numbers c and d, and positive integers k, m, and n
n
1. Constant term rule £ c = c + c+ . . . + c = n c
k— 1
n terms
2. Sum rule Σ (¾ + bk ) = ∑ a k + ∑ bk
k=1 k-1 k— 1
n n / n ∖
3. Scalar multiple rule Σ M k = c∑ a = [ ∑ ak )c
k=1 k=1 k ×k=1 /
4. Linearity rule ∑ (cak + dbk ) = c∑ ak + d∑ bk

⅛=1 k= 1 k= 1
5. Subtotal rule If 1 < m < n, then
n m n
Σ a.κ = ∑ a,κ + ∑ a,κ
k=∖ ⅛=1 k=m+l
6. Dominance rule If ak ≤ bk for k = 1, 2, . . . , n, then
n n
Σ akk ≤ ∑ bkk
⅛=1 fc=1
Proof: These properties can all be established by applying well-known alge

braic rules (see Problem 47 of Problem Set 4.1). ■
Using summation notation, we can streamline the symbolism in the formula

for the area under the curve y = ∕(x ) on the interval [a, b∖. In particular, note that
the approximating sum S'n is
S n = lf ( a + ∆x) + f(a + 2∆x) + ∙∙∙ + f(a + π∆x)]∆x

n
= fc∑ f(a + A∆x) ∆x
where ΔΛ^ = a . Thus, the formula for the definition of area can be written as
Area under y = f(x ) above the x-axis n is the number of approximating rectangles
between x = a and x = b j, f Width of each rectangle
A = Λ→
lim S, , = ∆Λ'→0
+≈ 1 tl
lim ⅛÷,
Y ,f(a +, ∕<Δx)Xx
'
Height of the fcth rectangle
■ AREA USING SUMMATION FORMULAS
From algebra we recall certain summation formulas (proved using mathematical

induction) that we will need in order to find areas using the limit definition of
area.
Summation Formulas
n
∑ 1= n
⅛=ι
Λ , , n(n + 1)
Σ k — l + 2 + 3 + ∙∙∙ + w - ---- 5----
fc=ι 2
∑ k 2 = 12 + 2 2 + 32 + + n 2 = ⅛--t∙1¾ 2w + °
fc=ι o
A , ,2 ,,3 , n2(n + 1)2
∑ k = 1 + 2 3 + 33 + ∙∙∙ + n 3 = —½ — -
⅛=ι 4
EXAM PLE 2 Area using summation and the definition
Use the summation definition of area to find the area under the parabola
y = x 2 on the interval [0, 1]. You estimated this area in Example 1.
Solution Partition the interval [0, 1] into n subintervals with width ∆x = ---- -.
7 / L-∖ / 2
The right endpoint of the ∕cth subinterval is a + k∆x = and f∖ ^Λ = —. Thus,
from the definition of area we have
n
A = lim ∑ f(a + ZCΔΛ')ΔX = lim ∑
n ∕u 2 '
Γ∣
A ι→0 t = l n→+∞ k=∖ ∖ n.
1 n
= lim i ∑ kι2 Scalar multiple rule
n→+≈ n k=∖
,. ι [n(n + 1)(2« + 1)1 . , r
= lim — _--------- g--------- J Summation formula for squares
. n(n + 1)(2M + 1) ,. «(« + l)(2rc + 1)(⅛)

= 1hm —----- -Δ4------ - = lim ------------------------
6 " 3(⅛)
π →+≈ 6n- n→+∞
n( ZL±jΛ ( 2n + A f ι + 1V 2 + 1
1. n∖ n J∖ ) 1. ∖ n)∖ n
= lim ---------- τ---------- = lim --------- τ-------
n→+∞ 0 ∕t→+∞ O
PROBLEM SET 4 .1
θ Evaluate the sums in Problems 1-8 by using the summation
lim
formulas. n → +∞
6 250 15 10
4 ∙ Σ
1∙ Σ 1 2∙ Σ 1 3. ∑ k First sketch the region under the graph o f y = f(x) on the
k=∖ k≈∖ k=∖ k=∖
interval [a, b] in Problems 13-21. Then approximate the area
5 7 100 100
5. ∑ ∕<3 6. ∑ k ' 7. ∑ (2k - 3) 8. ∑ ( ⅛ - 1 ) 2 o f each region by using right endpoints and the formula
k=∖ k=∖ A-=l k=l
Use the properties o f sigma notation in Problems 9-12 to s n= Σ Ka + ^ΔΛ')ΔΛ∙

A-=l
evaluate the given limits.
n k
9. lim V -⅜ 10. lim for ∆x = and the indicated values o f n.
i
" →+∞⅛≡'ι n ’ k =ι n
13. ∕( x ) = 4x + 1on [0, 1] for 36. Show that the triangle with vertices (0, 0), (0, h), and (b, 0)
a. n = 4 b. n = 8 has area A = ⅛bh.
14. ∕( x ) = 3 - 2x on [0, 1]for 37. a. Compute the area under the parabola y = 2x 2 on the
a. n = 3 b. n - 6 interval [1, 2] as the limit of a sum.
15. ∕( x ) = x 2 on [1, 2] for b. Let ∕( x ) = 2 x 2 and note that g(x) = j x 3 defines a
a. n = 4 b. n = 6 function that satisfies g'(x) = ∕( x ) o n ⅛e interval
[1, 2]. Verify that the area computed in part a satisfies
16. ∕( x ) = cosx on [~ y , θ] for n = 4 A =g(2) - g ( l ) .
c. The function defined by
17. f(x ) = x + sin x on [θ, ⅞∣ for n = 3
∕z(x) = ∣ x 3 + C
18. f( x ) = -∑ on [1, 2] for n = 4 for any constant C also satisfies h '(x) = ∕( x ) . Is it true
that the area in part a satisfies A = A(2) - ∕z(l)7
19. f(x ) = j on [1, 2] for w = 4
20. f(x ) = V x on [1, 4] for n = 4

21. f( x ) = V l + x 2 on [0, 1] for n = 4 Computational Window
θ Find the exact area under the given curve on the interval pre
Use a tabular approach to compute the area under the
scribed in Problems 22-27 by using the summation formulas.
curve y = f(x) on each interval given in Problems 38-42 as
22. y = 4x 3 + 2x on [0, 2]
the limit o f a sum o f terms.
23. y = 4x 3 + 2x on [ 1, 2] 38. ∕( x ) = 4x on [0, 1]
24. y = 6x 2 + 2x + 4on [0, 3] 39. ∕( x ) = x 2 on [0, 4]
25. y = 6x 2 + 2x + 4on [1, 3]
40. ∕( x ) = cos x on [ - y , θ]∙ Compare with Problem 16.
26. y = 3x 2 + 2x + 1on [0, 1]
27. y = 4x 3 + 3x 2 on[0, 1] 41. ∕( x ) = x + sinxon ^0, j ] . Compare with Problem 17.
THINK TANK PROBLEMS Show that each statement in Prob 42. a. Use the tabular approach to compute the area
lems 28-33 about area is generally true or provide a counterex under the curve y = sinx + cosx on the interval
ample. It will probably help to sketch the indicated region for [θ, ■?] as the limit o f a sum.
each problem. b. Let ∕( x ) = sin x + cosx and note that g(x) =
28. If C > 0 is a constant, the region under the line y = C on - c o s x + sinx satisfies g'(x) = ∕( x ) on the in
the interval [a, ⅛] has area A = Cfb - a). terval [θ, f ] . Verify that the area computed in part
29. If C > 0 is a constant and b > a ≥ 0, the region under a satisfies A = g ( ∕) - g(0).
the line y = C x on the interval [a, b] has area
A = ∖c(b - a). c. The function ⅛(x) = - c o s x + sinx + C for con
stant C also satisfies ∕z'(x) = ∕( x ) . Is it true that
30. The region under the parabola y = x 2 on the interval the area in part a satisfies A = ⅛(⅛) - ∕z(0)7
[a, b] has area less than
∣ (⅛2 + a 2)(b — a)
31. The region under the curve y = V l —x 2 on the interval θ 43. Derive the formula
[ - 1, 1] has area A = %. ∑ y t= 1 + 2 + 3 + ... + w = ≤ ^ ± n
32. Let / b e a function that satisfies f( x ) ≥ 0 for x in the *=l 2
interval [a, b]. Then the area under the curve y = [∕(x ) ] 2 by completing these steps:
on the interval [a, b] must always be greater than the area
a. Use the properties of sum in Theorem 4.1 to show that
under y = f(x) on the same interval.
33. If / is an even function [that is, / satisfies ∕( x ) = ∕( - x ) ] ∑ ⅛ = ∣ ∑ [k2 - ( k - 1)2] + ∣ ∑ 1
and ∕( x ) ≥ 0 throughout the interval [ - α , a], then the k =l k=∖ k=∖
area under the curve y = ∕( x ) on this interval is twice the
area under y = ∕( x ) on [0, a], = ∣ Σ [ F - ( f c - l) 2] + ∣ n
34. Show that the region under the curve y = x 3 on the
interval [0, 1] has area ∣ square units.
b. Show that / [k2 —(k — 1)2] = n2
35. Use the definition of area to show that the area o f a ⅛=ι
rectangle equals the product of its length € and its width Hint'. Expand the sum by writing out a few terms. Note
w. the internal cancellation.
n n(n + 1) equal parts, the ∕<th subinterval is

c. Combine parts a and b to show that ∑ k =
⅛=ι 2 ' k - 1. k∖
n n)
44. First find constants a, b, c, and d such that
This formula uses approximating rectangles with heights
∕c3 = a[ki - (⅛ - 1)4] + bk2 + ck + d taken at the left endpoints.
and then modify the approach outlined in Problem 43 to 46. Develop a formula for area based on approximating
establish the formula rectangles with heights taken at the midpoints of subinter
vals.
f «(»+ l)(2w+ 1)
47. Use the properties of real numbers to establish the sum
mation formulas in Theorem 4.1. For example, to prove
the linearity rule, use the associative, commutative, and
45. The purpose of this problem is to show that the definition distributive properties of real numbers to note that
of area that uses the right endpoint of each subinterval is
not the only way to approximate the area under a curve. n
Specifically, we shall find the area A under the parabola ∑ (cak + dbk ) = (caj + dbl )
k=l
y = x 2 on the interval [0, 1] in several ways. + (ca2 + db2) + ∙ ■∙ + (can + dbn)
Verify that
n
= (cal + ca2 + ■■∙ + can) + (dbl + db2 + ■■■+ dbn)
lim ∑ 1 = c(α1 + a2 + ∙ ∙ ■αn) + √(⅛1 + b2 + ■■∙ ⅛h)
n→+∞ k=∖ 3
Compare this with the procedure outlined in Problem 43. = c ∑ a
k
+ d ∑ b
k
Note that when the interval [0, 1] is subdivided into n k= 1 k= 1
4.2 R IE M A N N SUM S A N D TH E D E FIN IT E INTEG RAL
Historical Note
IN THIS SECTION Riemann sum, definite integral, area as an integral,
distance as an integral, properties of the definite integral
We shall soon discover that not just area, but also such useful quantities as
distance, volume, mass, and work, can be first approximated by sums and then
obtained exactly by taking a limit involving the approximating sums. The
special kind of limit of a sum that appears in this context is called the definite
integral, and the process of finding integrals is called definite integration or
Riemann integration in honor of the German mathematician Bernhard Rie
mann (1826-1866), who pioneered this approach. In this section, we describe
integration as a generalization of the method for computing area introduced in
Section 4.1. You should also note the last paragraph of the Preview on page 273
in anticipation of one of the greatest labor-saving devices known to human
kind—namely, the fundamental theorem of calculus, which will be developed
in Section 4.3.
GEORG FRIEDRICH BERNHARD
RIEMANN (1826-1866)
We meet Riemann in this sec ■ R IE M A N N SUM S
tion in the context of what we
call Riemann sums. In his per Recall from Section 4.1 that to find the area under the continuous curve y = ∕(x )
sonal life he was frail, bashful, on the closed interval [a, b] where∕(x ) ≥ 0, we first partitioned the interval into n
and timid, but in his profes a
sional life he was one of the subintervals of equal width ∆x = . Then we evaluated/at the right endpoint
giants in mathematical history. a + ∕<∆x of the ∕cth subinterval for k = 1, 2, . . . , n and formed the approximating
He is known for his work in sum of the areas of the n rectangles, which we denoted by S n . Because we expect
geometry (for example, Rie the estimates S n to improve as ∆x decreases, we defined the area/l under the curve,
mannian geometry and Rie
above the x-axis, and bounded by∙z the lines x =n a and x = b to be the limit of S as
mann surfaces) and analysis. In
∆x → 0. Thus, we write
A = lim ∑ f(a + ∕<∆x)∆x
n→ +∞ k=∖
Sec. 4.2 Riemann Sums and the Definite Integral 26 7
his book, Space Through the if this limit exists. This means that A can be estimated to any desired degree of
Ages, Cornelius Lanczos said, accuracy by approximating the sum S n with ∆x sufficiently small (or, equivalently,
“Although Riemann’s collected n sufficiently large).
papers fill only one single vol This approach to the area problem contains the essentials of integration, but
ume of 538 pages, this volume there is no compelling reason for the partition points to be evenly spaced or to
weighs tons if measured intel insist on evaluating/at right endpoints. These conventions are for convenience of
lectually. Every one of his computation, but to accommodate applications other than area, it is necessary to
many discoveries was destined consider a more general type of approximating sum and to specify what is meant
to change the course of mathe by the limit of such sums. The approximating sums that occur in integration
matical science.” problems are called Riemann sums, and the following definition contains a step-
by-step description of how such sums are formed.
Riemann Sum Suppose a function/is given, along with a closed interval [a, ⅛] on which/is
defined. Then:
Step 1: Partition the interval [a, b∖ into n subintervals by choosing points
{x0, x l , . . . , x zι} arranged in such a way that
a = x0 < x 1 < x2 < - < x n . 1 < x n = b
Call this partition P. For k = 1, 2, . . . , n, the Ath subinterval width is
∆xjt = x k - x k _{ . The largest of these widths is called the norm of the
partition P and is denoted by ||P||; that is,
w = fc=1¾a x . n
Step 2: Choose a number arbitrarily from each subinterval. For k = 1, 2, . . . , n,
the number x k chosen from the Zrth subinterval is called the kth subinterval
representative of the partition P.
Step 3: Form the sum
R n = ∕(**)∆x 1 + ∕( x ∕∆ x 2 + ∙∙∙ + ∕(x*)∆x n = ∑ ∕(x*)∆x jt
k.—1
This is the Riemann sum associated with / the given partition P, and the
chosen subinterval representatives x* x 2, . . . , x*.
What This Says: This is a generalization of the area calculations of the last
section. Whenever we formulate the sum
∑ι ∕(x⅛)∆χfc
it is called a Riemann sum. Riemann sums are generally used to produce the
correct integral for a particular application. Note that the Riemann sum does
not require that the function / be nonnegative and also allows that x* is any
point in the ∕<th subinterval.
EXAMPLE 1 Formation of the Riemann sum for a given function
Suppose the interval [-2 , 1] is partitioned into 6 subintervals with subdivi

sion points a = x 0 = -2 , x 1 = -1.6, x 2 = -0.93, x 3 = -0.21, x 4 = 0.35,
x 5 = 0.82, and x 6 = 1 = b. Find the norm of this partition P and the Riemann
sum associated with the function ∕(x ) = 2x, the given partition, and the
subinterval representatives x* = -1.81, x* = -1.12, x* = -0.55, x* = -0.17,
x* = 0.43, and x* = 0.94.
Solution Before we can find the norm of the partition or the required Riemann
sum, we must compute the subinterval width ∆xfc and evaluate / at each
subinterval representative x ∕ These values are shown in Figure 4.4 and the
table of computations on page 268.
Given
Given
∣t ∕< A =
k x k " ⅞ -ι = ∆χ k x k 2(4)
1 -1 .6 - (-2 ) = 0.40 -1.81 -3.62

2 -0 .9 3 - (-1.6) = 0.67 -1 .12 -2 .2 4
3 -0.21 - (-0.93) = 0.72 -0.55 -1 .1 0
4 0.35 - (-0.21) = 0.56 -0 .17 -0 .3 4
5 0.82 - 0.35 = 0.47 0.43 0.86
6 1.00 - 0.82 = 0.18 0.94 1.88
Figure 4.4 Partition for a Riemann sum

From the table, we see that the largest subinterval width is ∆x 3 = 0.72, so the
partition has norm ∣∣P∣∣ = 0.72. Finally, by using the definition, we compute
the Riemann sum:
0 Notice from Example 1 that
the Riemann sum does not nec R 6 = (-3.62)(0.4) + (-2.24)(0.67) + (-1.10)(0.72) + (-0.34)(0.56)
essarily represent an area. The + (0.86)(0.47) + (1.88)(0.18) ≈ -3.1886 ≈ -3 .1 9 —
sum found is negative (and area
must be nonnegative). 0
■ THE DEFINITE INTEGRAL
By comparing the formula for Riemann sum with that of area in the last section,
we recognize that the sum S n used to approximate area is actually a special kind of
Riemann sum which has
∆x, k= ∆x = ~- and kx , = a + k∆x

n
for k = 1, 2, . . . , n. Because each subinterval in the partition P associated with S ll
has width ∆x, the norm of the partition is
∣∣P∣∣ = Λ x = ^
This kind of partition is called a regular partition. When we express the area under
the curve y = fix) as A = lim S . we are actually saying that A can be estimated to
any desired accuracy by finding a Riemann sum of the form S n with norm
H =⅛*
sufficiently small. We use this interpretation as a model for the following
definition.
Definite Integral I f/ is defined on the closed interval [a, b] we say/is integrable on [a, ⅛] if
n
I = lim ∑ fix*) ∆x,
∣∣P∣∣→0 k=l ' v k , k
exists. This limit is called the definite integral of / from a to b. The definite
integral is denoted by
rb rx==b
I ~ Ja
f ( x ) dχ 0r I = Jx=a
∕(∙χ ) dx
The latter is used if we wish to emphasize the variable of integration.

What this says: If the limit of a Riemann sum of f as ∣∣P∣∣ → 0 exists and is
finite, then we say that/is integrable. In particular, it means that the number I
can be approximated to any prescribed degree of accuracy by any Riemann
sum of/w ith norm sufficiently small. As long as we remember the conditions
of this theorem (that is ,/ is defined on a closed interval and that the limit of
the Riemann sum exists), it can be summarized by writing
^b n
J.⅛ ) Λ = t o Σ ⅛ > ,
Formally, I is the definite integral of/on [a, b] if for each number ε > 0, there
exists a number δ > 0 such that if
∑ ∕( 4 ) ∆ x fc
K—1
is any Riemann sum o f/ whose norm satisfies ∣∣P∣∣ < δ, then
I7 ^1 1M ) ΔXJ < Ε
In advanced calculus, it is shown that when this limit exists, it is unique.
Moreover, its value is independent of the particular way in which the partitions
of [a, b] and the subinterval representatives x*k are chosen.
Upper limit of integration Consider the notation used in the margin. The function / that is being
⅛ integrand integrated is called the integrand; the interval [a, b] is the interval of integration;
R 7U) Λ- and the endpoints a and b are called, respectively, the lower and upper limits of
a integration.
'k- Lower limit of integration In the special case where a = b, the interval of integration [a, b] is really just
a point, and the integral of any function on this “ interval” is defined to be 0; that
is, , fl
∕( x ) dx = 0
Ja
Also, at times, we shall consider integrals in which the lower limit of integration
is a larger number than the upper limit. To handle this case, we specify that
the integral from b to a is the opposite of the integral from a to b:
ca rb
∕( x ) dx = -
'b Ja
To summarize:
Definite Integral at a Point
Opposite of a Definite Integral
At first, the definition of the definite integral may seem rather imposing.
How are we to tell whether a given function/is integrable on an interval [a, b]1 If
/ is integrable, how are we supposed to compute the definite integral? Answering
these questions is not easy, but in advanced calculus, it is shown that /is integrable
on a closed interval [a, b] if it is continuous on the interval except at a finite num-
ber of points and if it is bounded on the interval (that is, there is a number A > 0
such that ∣∕ ( x ) ∣< A for all x in the interval). We will state a special case of this re
sult as a theorem.
THEOREM 4.2 Existence of Riemann sum

If f is continuous on an interval [a, b∖, th e n /is integrable on [a, b].
Proof: The proof is found in most advanced calculus textbooks. ■
Our next example illustrates how to use the definition to find a definite
integral.
EXAMPLE 2 Evaluating a definite integral using the definition
1
Evaluate f 4x dx.
J-2
Solution We note the variable of integration is x, so we interpret this as
4x dx
The integral exists because ∕(x ) = 4x is continuous on [—2, 1]. Because the
integral can be computed by any partition whose norm approaches 0 (that is,
the integral is independent of the sequence of partitions and the subinterval
representatives), we shall simplify matters by choosing a partition in which
the points are evenly spaced. Specifically, we divide the interval [-2 , 1] into n
subintervals, each of width
For each k, we choose the kth subinterval representative to be the right

endpoint of the Zrth subinterval; that is,
x*.k = - 2 + ∕<∆x = - 2 + kf-')
∖n)
Finally, we form the Riemann sum
r1 n
4x dx = lim ∑ 7f(x∏∆x
-2 ∣
∕>∣
∣
→o *=ι v k,
= lim ∑ 4 ( - 2 + ≡ ∏→ +∞ as ∣
∣∣
P∣→0
n →+∞ k=∖ ∖ ft J ∖ ft)
= lim -⅛ z
∑ ( - 2 H + 3∕<)
π →+≈ ∕γ fc=]
= lim -⅛f ∑ (-2w) + ∑ 3k}

n →+∞ ∏l ∖ k=∖ fc=l /
= lim ‰ + 3[ ^ T ∣ Formula for the
w→+∞ ∏z ∖ L 2 J∕ sum o f integers
= lim W - 4 » 2 + 3n 2 + 3n^
= lim - 6-^2 ↑- l ⅛ = - 6
n→+∞ n
Sec. 4.2 Riemann Sums and the Definite Integral 271
■ AREA AS AN INTEGRAL
Because we have used the development of area in Section 4.1 as the model for our
definition of the definite integral, it is no surprise to discover that the area under a
curve can be expressed as a definite integral. However, integrals can be positive,
zero, or negative (as in Example 2), and we certainly would not expect the area
under a curve to be a negative number! The actual relationship between integrals
and area under a curve is contained in the following observation, which follows
from the definition of area as the limit of a sum along with Theorem 4.2.
Area As an Integral Suppose/is continuous and f(x) ≥ 0 on the closed interval [a, ⅛]. Then the
area under the curve y = ∕(x ) on [a, ⅛] is given by the definite integral of f on
[a, b∖. That is,
0 We will find areas using a r⅛
Area = f{x) dx
definite integral, but not every Ja
definite integral can be inter
preted as an area. 0
Usually we find area by evaluating a definite integral, but sometimes area can
be used to help us evaluate the integral. At this stage of our study, it is not easy to
evaluate Riemann sums, so if you happen to recognize that the integral represents
the area of some common geometric figure, you can use the known formula
instead of the definite integral, as shown in Example 3.
EXAMPLE 3 Evaluating an integral using an area formula
f3 /---------
Evaluate V9 - x 2 dx.
J- 3
Solution Let ∕(x) = V9 - x 2 . The curve y = V9 - x 2 is a semicircle centered at
the origin of radius 3, as shown in Figure 4.5. From geometry, we know the
Figure 4.5 The curve area of the circle is A = πr 2 = ττ(3)2 = 9π. Thus, the area of the semicircle is
y = V9 - x 2 is a semicircle ¾r >so
of radius 3.
Γ √ 9 ^ U √ x = ⅛r
J-3 2
■ DISTANCE AS AN INTEGRAL
Many quantities other than area can be computed as the limit of a sum. For
example, suppose an object moving along a line is known to have continuous
velocity v(∕) for each time t between t = a and t = b, and we wish to compute the
total distance traveled by the object during this time period.
Suppose the interval [a, b] is partitioned into n equal subintervals, each
of length ∆t = ⅛ n a , as indicated in Figure 4.6. The ∕rth subinterval is
α + (⅛ -l)Δ r α + ⅛∆f
The distance is ∆s = via + (k - 1)∆rl(∆rt.
Figure 4.6 The distance traveled during the Uh time subinterval

[a + (k - 1)∆Z, a + k∆t], and if ∆t is small enough, the velocity v(Z) will not change
much over the subinterval; so it is reasonable to approximate v(Z) by the constant
value v[a + (k - 1)∆Z] throughout the entire subinterval.
The corresponding change in the object’s position will be approximated by the
product
v[a + ( k - 1) ∆Z] ∆t
This change will be positive if v[a + (k - 1) ∆Z] is positive and negative otherwise.
Both cases may be summarized by the formula
∣v[α + ( k - l)∆z]∣ ∆t
and the total distance traveled by the object as t varies from t = a to t = b is given
by the sum
S,, = ∑ ∖v[a + ( k - 1)∆Z]∣∆Z
We recognize this as a Riemann sum and can make it more and more precise by
taking more refined partitions (that is, shorter time intervals ∆Z). Therefore, it is
reasonable to define the exact distance S traveled as the limit of the sum S n as
∆t → 0, or, equivalently, as n → + ∞, so that
5 = lim Y ∣v[α + (k - 1)∆Z]∣∆Z = [ ∣v(Z)∣<Λ

n→+∞fc=l Ja
To summarize:
Distance The distance traveled by an object moving with continuous velocity v(Z) along a
straight line from time t = a to t = b f,is
b
Ja
∖v(f)∖dt
Sometimes we are more interested in where the object is at the end of a given
time interval than in the total distance it has traveled. The difference between the
object’s final position and its initial position is called its displacement and is given
by rb
[displacement] D = v(fi)dt
Ja
Thus, an object that moves forward 2 units and back 3 on a given time interval has
moved a total distance of 5 units, but its displacement is - 1 because it ends up 1
unit to the left of its initial position.
EXAM PLE 4 Distance moved by an object whose velocity is known
An object moves along a straight line with velocity v(Z) = t2 for t > 0. How far
does the object travel between times t = 1 and t = 2?
2 _ι ι
Solution We have a = 1, b = 2, and ∆Z = — — = therefore, the required
distance is
= lim [(M2 - 2n + 1) + k 1 + 2(« - 1)k]

= lim -3i (w2 - 2n + 1) ∑ 1 + ∑ k 2 + 2(n - 1) ∑ k ∖

n→∞n L ⅛=ι ⅛=ι ’ k=∖ J
ι∙ 1 Γ∕ 7 ι , n (n + l)(2w + 1) ~ , . Π (Λ + 1)Î
= lim μ ∕z 2 - 2M + l)xn + —------- ------------ + 2(z « - l ) - 2—^— ^J
= iim 14»3 - 9n 2 + n _ 14 _ 7
™ 6W3 6 3
Thus, we expect the object to travel ∖ units during the time interval [ 1, 2]. MM
PREVIEW
Students often become discouraged at this point, especially after looking at the
amount of “ work” necessary to complete Example 4, so we offer an encourag
ing word. Recall the definition of derivative in Chapter 1. It was difficult to find
derivatives using the definition, but we soon proved some theorems to make it
easier to find and evaluate derivatives. The same is true of integration. It is
difficult to apply the definition of a definite integral. However, we will soon
have some theorems to make your work easier when working with certain types
of functions. As a preview of what is to come, consider an object moving along
a straight line with displacement s(f) and velocity v(Z). We know from Chapter 2
that these are related by the equation v(f) = s'(f). If the object is always moving
forward, v(Z) > 0, for all t in the time interval [a, b∖, then the total distance
traveled by the object during this time interval is 5 = x(⅛) - x(α). However, by
comparing the definition of area as the limit of a sum with the formula for S:
5 = lim ∑ v[a + (k - 1)∆Z]∆Z
n→+∞ k=l
we see that the total distance S’ also equals the area under the graph of v(Z) on
the interval [a, /?]. That is:
If v(f) is a continuous function that satisfies v(t) > 0 on the interval [a, ⅛],
then the area under the curve y = v(t) on the interval [a, Λ∣ is equal to
s(Z>) - 5(α)
where s(t) satisfies s'(Z) = v(Z) throughout [a, b].
This observation anticipates the fundamental theorem o f calculus, which
provides a vital link between differential and integral calculus. We shall
formally introduce the fundamental theorem in Section 4.3 and use it in all
subsequent work.
Notice how our observation applies to Example 4, where v(Z) = t2
and the interval is [1, 2]. It is easy to verify that if g(Z) = ∣Z3, then g'(f) -
v(Z). Thus, the total distance traveled by the object during the time interval
[1, 2] is
∆ = g ( 2 ) - g ( l) = [∣(2 ) 3 - ∣ ( l ) 3] = j
which coincides with the result found numerically in Example 4.

Incidentally, note that h(f) = jZ3 + C also satisfies h'(f) = v(Z) for any
constant C . What happens if you try to compute the distance Δ using h(f)
instead of g (ff.
By considering the distance as an integral, we see that zero or negative values

of a definite integral can also be interpreted physically. When v(Z) > 0 on a time
interval [a, b], then the total distance S traveled by the object between times t = a
and t = b is the same as the area under the graph of v(∕) on [a, b∖.
When v(t) > 0, the object moves forward, but when v(∕) < 0, it reverses
direction and moves backward (to the left). In the general case where v(Z) changes
sign on the time interval [a, b], the integral
r∕>
v(∕) dt
Ja
a. Net displacement of 0
There is as much area under measures the net displacement of the object, taking into account both forward and
the positive part of the velocity backward motion.
curve as there is above the For example, for the velocity function v(∕) graphed in Figure 4.7a, the net
negative part.
distance traveled is 0 because the area above the Faxis is the same as the area
below, but in Figure 4.7b, there is more area below the Faxis, which means the net
distance is negative, and the object ends up “behind” (to the left of) its starting
position.
■ PROPERTIES OF THE DEFINITE INTEGRAL

b. Negative net displacement
In computations involving integrals, it is often helpful to use the three general
There is more area above
the negative part of the velocity
properties listed in the following theorem.
curve than there is under the
positive part.
Figure 4.7 Definite integral in THEOREM 4.3 General properties of the definite integral
terms of displacement
Linearity rule If f and g are integrable on [a, b], then so is rf + sg for
constants r, s.
rb rb
[rf(x) + sg(x)]dx = r f(x)d x + s g(x)dx
Ja Ja
Dominance rule Iff and g are integrable on [a, Z>] and f(x) ≤ g(x) throughout
y = fM this interval, then
rb rb
a b f(x )d x ≤ g(x~)dx
Ja Ja
a. The dominance rule
If g(x) ≥ f(x ), the area under Subdivision rule For any number c such that a < c < b
y - g(χ) can be no less than
the area under y =f(x). rb rc rb
f(x )d x = f(x)d x+ f(x)d x
la Ja Jc
assuming all three integrals exist.

Proof: Each of these rules can be established by using a familiar property of
sums with the definition of the definite integral. For example, to derive the
linearity rule, we note that any Riemann sum of the function rf + sg can be
expressed as
n Γ n ^1 Γ n
a c b Σ [r∕(4 ) + ^ ( ⅞ ) ] ¾ = 4I—∕C—
K.—1
∑1 M)Δ⅞] -1
+ 4l~∕t—∑1 M)Δ⅞
b. The subdivision rule
The area under y = f(x) on
and the linearity rule then follows by taking the limit on each side of this equation
[a, ft] equals the sum of the as the norm of the partition tends to 0. ■
areas [a, c] and [c, b].
Figure 4.8 Comparison of The dominance rule and the subdivision rule are interpreted geometrically for
dominance and subdivision rules nonnegative functions in Figure 4.8.
Sec. 4.2 Riemann Sums and the Definite Integral 275
Notice that if f(x) ≥ g(x) ≥ 0, the curve y = g(x) is always above (or
touching) the curve y = f(x), and the dominance rule expresses the fact that the
area under the upper curve y = g(x) cannot be less than the area under the area
under y = f(x). The subdivision rule says that the area under y = f(x) above [a, Z>]
is the sum of the area on [a, c] and the area on [c, b∖. We close with an example that
illustrates one way of using the subdivision rule.
EXAM PLE 5 Subdivision rule
I f f ∕(x )r∕x = 3 and f f∖ x)dx = -5 , what is f f(x )d x f

J-2 J-2 Jι
Solution According to the subdivision rule, we have
Γ7 f1 Γ7
f(x )d x = f(x )d x + f(x)d x
J-2 J-2 Jl
Therefore,
Γ7 f7 f1
f(x )d x = f(x )d x - f(x )d x = (-5 ) - 3 = - 8
Jl J-2 J-2
PROBLEM SET 4 .2
Q In Problems 1-8, estimate (using right endpoints) the given 14.
13. I x 2 dx (x 2 + x) dx
(^b 0 -1
integral f (x) dx by using a Riemann sum
Ja f 2
Γ2
15. I∣ (2x 2 - 3x) dx 16. I x 2 dx
S n = ∑ f(a + klsx) ∆x for n = 4. -1 0
∕∙1 rθ
2
2. ,f (4x + 2) dx 17 ∙ f0
1. , (2x + 1) dx 1 x dx 18. ∣ (3x 2 - 5x) dx
υ 0 -1 J-ι
f3 Γ2 Use the dominance property o f integrals to establish the given
3. x 1 dx 4 ∙ 1 x 3 dx
,ι
0 inequality in Problems 19-20.
r1 r3 19. f x 3 dx ≤ ∣ Hint: Note that x 3 ≤ x on [0, 1].
5. 1 (1 - 3x)dx 6∙ 1 (x 2 - x 3) dx Jo 2
0 I
rθ ', 7Γ
J
fττ∕4
7∙ 1 cos x dx 8. I (x + sinx) dx sin x d x ≤ 7τ Hint, sinx ≤ 1 for all x.
—π∕2 0 o
r4 r4
In Problems 9-12, v(t) is the velocity o f an object moving alont © 21. Given [5∕(x) + 2g(x)] dx = 7 and [3∕(x) + g(x)] dx = 4,
J-2 J-2
a straight line. Use the formula
r4 r4
find ∕( x ) dx and g(x) dx.
S„ = ∑ v[a + ( k - l)∆ ∕]∆ ∕ J-2 J-2
' k=∖
where ist = - — - to estimate the total distance traveled by the 22. Suppose
Γ2 f( x ) d x = 3, f 2 g(x) dx = - 1, and
n
f 2 Jo Jθ
object during the time interval [a, b]. Let n = 4 for Problems h(x)dx = 3.
9-12. Jθ r2
9. v(∕) = 3/ + 1 on [1, 4] 10. v(t) = 1 + 2t on [1, 2] a. Evaluate [2∕(x) + 5g(x) - 7A(x)] dx.
11. v(∕) = sin / on [0, τr] 12. v(f) = cost on [θ,
Jo "2
Evaluate each o f the integrals in Problems 13-18 by using the

following information together with the linearity and subdivi
23. Evaluate
f2
J0
[5∕(x) + sg(x) - 6∕z(x)] dx = 0
f(x ) dx given that

sion properties: J-ι
f2 f0 1 f2 3 f2 f f(x ) d x = 3', f f ( x ) d x = - 2 ,i f f(x )d x = 5
x 2 dx = 3; x 2 dx = i, ; x dx = 2 x dx = 2
J-ι J-ι J-ι Jo J -ι J2 Ji
Γ2 for - 1 ≤ x ≤ 1 30. Compute the Riemann sum associated with

24. L e t∕( x ) = <3 - x for 1 < x < 4 ∕( x ) = 4 - 5x, the partition P, and the subinterval repre
(,2x - 9 for 4 ≤ x ≤ 5 sentatives x* = -0 .5 , xζ = 0.8, x* = 1, X4 = 1.3, x* = 1.8.
31. Compute the Riemann sum associated w ith ∕(x) = x 3. the
Sketch the graph o ff on the interval [ -1 ,5 ] and show that
partition P, and the subinterval representatives x∣ = - 1,
f is continuous on this interval. Then use Theorem 4.3 to
x* = 0, x* = 1, x* = ⅛δ , and x* = ⅛p.
evaluate
32. If the numbers ak and bk satisfy ak ≤ bk for k = 1,2, . . . ,
n n
n, then ∑ ak ≤ ∑ bk . Use this dominance property of
A,∙=1 k= 1
5 for - 3 ≤ x ≤ -1 sums to establish the dominance property of integrals.
25. L e t∕( x ) = 4- x for -1 < x < 2
33. T H IN K T A N K P RO BLEM Either prove that the following
2 χ -2 for 2 ≤ x ≤ 5
result is generally true or find a counterexample: If∕( x ) is
Sketch the graph of/on the interval [—3, 5] and show that not identically zero on the interval [a, b], then
/is continuous on this interval. Then use Theorem 4.3 to f6
∕( x ) dx ≠ 0
evaluate Ja
Γ5 for some x-value on [a, b∖.
f( x ) d x
J-3 34. T H IN K T A N K P ROBLEM Either prove that the following
θ 26. Generalize the subdivision property by showing that for result is generally true or find a counterexample: If/and g
a ≤ c ≤ d ≤ b, are continuous on [a, ⅛], then
rb rc rd rb "∕> Γ rb "I Γ rb ∣
Ja
f( x ) d x =
Ja
f( x ) d x +
whenever all these integrals exist.

Jc
f( x ) d x +
Jd
f(x )d x
35.
J a
∕(x)g(x) dx =
'-Ja
∕( x ) dx
-^ ∙-J a
a. I f ∕( x ) < 0 on the interval [a, b], show that

g(x) dx
J
27. For any constants C and D, show that fA

∕( x ) dx = —A
f (Cx + D )d x = ( b - α)⅛ (6 + a) + Z>] Ja
where A is the area under the graph of y = —∕( x ) on

Hint. Sketch the region under the line y = C x + D , and [a, ⅛].
express the integral as an area. b. Combine the formula in part a with the formula of
28. For b > a > 0, show that Theorem 4.3 to show that
^b
Cb
J<7
x2
i
dx = ⅛(⅛3 - α 3)
J
a
f( x ) dx = P - N
where P is the sum o f the area of all positive regions

In Problems 29-31, use the partition
( ∕( x ) > 0) and N is the sum of the area of all negative
P = { - 1, —0.2, 0.9, 1.3,1.7, 2} on the interval [ - 1, 2].
regions (∕(x ) < 0).
29. Find the subinterval widths
c. Use the formula in part b to evaluate
δ ⅞ = Xk - Xk _ { Γ1
2x dx.
for k = 1, 2, . . . , 5. What is the norm o f P? J-2
4.3 THE FUNDAMENTAL THEOREM OF CALCULUS; INTEGRATION BY SUBSTITUTION
IN THIS SECTION The fundamental theorem of calculus, substitution with

indefinite integration, substitution with definite integration
We have found that the value of a definite integral can be approximated
numerically, and by using summation formulas we can compute some integrals
algebraically. These methods are messy, and it is doubtful that integration
would be the powerful mathematical tool it is if these were the only ways to
compute integrals. The fundamental theorem of calculus, presented in this
section, gives us a means o f evaluating many integrals. We shall study
techniques of integration in Chapter 7, but one such technique, the method o f
substitution, is so fundamental that it is useful to have it now.
Sec. 4.3 The Fundamental Theorem of Calculus; Integration by Substitution 27 7
■ THE FUNDAMENTAL THEOREM OF CALCULUS
In the last section we found that if v(Z) is the velocity of an object at time t as it
moves along a straight line, then
v(t) dt = s(b) - s(a)
where s(f) is the displacement of the object and satisfies s'(∕) = v(t). This result is
an application of a general theorem discovered by the English mathematician
Isaac Barrow (1630-1677), who was Newton’s mentor at Cambridge.
THEOREM 4.4 The fundamental theorem of calculus
If f is continuous on the interval [a, b] and F is any function that satisfies

F'(x) = ftx ) throughout this interval, then
f(x ) dx = F(b) - F(a)
Proof: Let P = {x0 , x l , x 2 , . . . , x f be a regular partition of the interval, with

subinterval widths ∆x = < Note that F satisfies the hypotheses of the mean
value theorem (Theorem 3.4, Section 3.2) on each of the closed subintervals
[xjt. 1, x i ,]. Thus, the M V T tells us that there is a point x'"k in each open subinterval
(-∖-∣, ⅞) for which
F ( x ,.) -F ( x . .)
— λv _ λγ - ■ — = F (⅞) θ∙' F(x k ) - F(x k _ 1) = F (χ)(χ k ~ x ^ 1)
* ⅛- ι
Because F'(x*k) = ∕(x⅛) and x k - x k _ l = ∆x = a , we can write F(x k) -
F(x k ~ l ) = ∕(⅞ )∆ x , SO that
F(x,) - F(x 0) = ∕(x * )∆ x
F(X 2) - F(x,) = ∕(x * ) ∆ x
F(x„) - F(x n . 1) = ∕(x * )∆ x

Thus, by adding both sides of all the equations, we obtain
∑1 f( x k) ∆x = ∕(x *) ∆x + ∕(x *) ∆x + - + ∕(x *) ∆x
= [F(x,) - F(x 0)] + [F(X 2) - F(x,)] + - + [F(xn) - F(x n .,)]
= F(x n) ~ F(x 0)
Because x 0 = a and x rι = b, we have
∑ ∕(x * ) ∆ x = F(∕>) - F(α)

k —1
Finally, we take the limit as ∣∣P∣∣ → 0, and because F(b) —F(a) is a constant, we
have
f ∕(x )√ x = F(b) - F(a) ■

Recall (from Section 3.10) that an antiderivative o f a function f is a function F

that satisfies F ' = f
rb
What this says: The definite integral f{ x ) d x can be computed by finding
Jα
an antiderivative F on the interval [a, ⅛] and evaluating it at the limits of
integration a and b. Also notice that this theorem does not say how to find the i
antiderivative, nor does it say that an antiderivative F exists.
The issue of existence of an antiderivative is important, and in Section 4.5 we

shall show that such an antiderivative always exists as long as f is continuous on
[a, b]. Indeed, we consider the issue of existence just as important as the formula
in the fundamental theorem. In fact, sometimes the two results are referred to as
the first and second fundamental theorems of calculus.
In order to give you some insight into why this theorem is important enough
to be named the fundamental theorem o f calculus, we repeat Example 2 from
Section 4.2 in the margin and then work the same example using the fundamental
theorem.
EXAM PLE 1 Evaluating an antiderivative
Γ1
Evaluate 4x dx using the definition of the definite integral and also using
the fundamental theorem.

Solution
Solution from Section 4.2 using Riemann sums: Solution using the fundamental theorem of calculus:
r1 n
4x dx = Wlim ∑ ∕( x t) ∆x 4x dx = F ( 1 ) - F (-2 )
1-2 →0⅛=I Λ K'
= lim ∑ 4 ( -2 + ¾ ( ∣') where F is an antiderivative of f If

∏
→+∞ ∣~ l ∖ n ) ∖n) F(x) = 2x 2 , then F'(x) = 4x, so F is an antideriv
ative of f Thus
= lim l ∑ (-2w + 37c)
κ →+≈ n k=∖
4x dx = 2(1)2 - 2 (-2 ) 2 = - 6
IQ / n n ∖
= lim 1 42 Y (-2n) + Y 3k
n →+≈ tt V⅛1 fc÷ι )
,. 12/, . f n ( n + 1)^∣∖
= hm — I (-2n)w + 3 ---- 5----
n→ +∞ n~∖ L z ∖)
= hm 1 2 (- 4 n 2 + 3 n 2 + 3n∖
n→ +∞ n ∖ Z /
= -6
Henceforth, when evaluating an integral by the fundamental theorem, we shall

denote the difference
F(⅛) - F(α) by F(x)∣*, which means F(x)∣∙^*
Sec. 4.3 The Fundamental Theorem of Calculus; Integration by Substitution 279
Sometimes we also write [∙F(x)]*, where F'(x) = ∕( x ) on [a, b∖. This notation is
illustrated in Example 2.
E X A M P L E 2 Finding the area under a curve using the fundamental
theorem of calculus
Find the area under the curve y = cosx on [-■J ’ f] .
Solution Because ∕'(x) = cosx is continuous on [—■ y], and because the deriva
tive of sinx is cosx, it follows (from the definition of antiderivative) that sinx
is an antiderivative of cosx. Also, because∕( x ) ≥ 0, the required area is given
by the integral
A = cosx dx = sinx∣2 πz2 = s in f - s in ( -f) = 1 - ( - 1 ) = 2
The region has area A = 2 square units.
I f you look at the graph o f∕( x ) = cos x on 0 It is important to remember that the definite integral I ∕( x ) dx is a
L ≡ , 5^∣, you can see that an area o f 2
L 2 2J f(x )d x is a family of functions. 0
seems reasonable.
The relationship between the indefinite and definite integral is given by
∕( x ) dx
EXAM PLE 3 Evaluating an integral using the fundamental theorem
f2
Evaluate a. c. ∣x∣ dx
J-2
Solution
a. Γ (-10) dx = -10x∣ " = -1 0 (5 + 3)= -8 0
1-1 I i
r 9Γ 1 1 r9 r ∕ ι∕2∖ 2^∣ 9
b. -fi=
L
+ x ∖dx = [x 1/2 + x l dx = ∣ - τ -r ∣ + ⅜- v2 4
J4 vx J J4 L∖ 2 / J
2 = (6 ÷ ⅜ )-(4 + ^ = 34∣
c. ∣x∣ dx Recall,
∙'^2 ∣x∣ = x if x ≥ 0
∣x∣ = - x if x < 0
So, for
χ ≥ 0, F 1(x) = y
X < 0, F 7(x) = -= y
We evaluate this integral by using the subdivision principle:

f2 f0 f2 f0 f2
∣x∣ dx = ∣x∣ dx + ∣x∣ dx = (-x ) dx + x dx
J-2 J-2 Jo J-2 Jo
The fundamental theorem of calculus reduces a problem of integration to a

problem of antidifferentiation. We stated the antidifferentiation formulas in
Theorem 3.11; they are repeated here for easy reference.
Integration Formulas
Constant rule: ∣ 0 du = C
Power rule: u n du - , + C; ∏Ψ 1
, 77+ 11
Trigonometric rules: sinw du = -cost/ + C
cos u du = sin u + C
∣ sec2 u du = tan u + C
cscu cot w du = - CSCM + C
1 CSC 2 M du = -cotw + C
secw tanw du = secι∕ + C
These formulas alone are not sufficient for many of the integrals we need to
evaluate. A most important procedure of integration is called substitution. We
now introduce this procedure with indefinite integration and then extend it to
definite integration.
■ SUBSTITUTION WITH INDEFINITE INTEGRATION
The method of substitution is the integration version of the chain rule. Recall that
according to the chain rule, the derivative of (x2 + 3x + 5)9 is
^ ( x 2 + 3x + 5)9 = 9(x2 + 3x + 5)8(2x + 3)
Thus, J 9(x2 + 3x + 5)8(2x + 3)dx - (x 2 + 3x + 5)9 + C
Note that the product is of the form g( u )~fc where, in this case, g(u) = 9z√8 and
u = x 2 + 3x + 5.
You can integrate many products of the form t>y applying the chain
rule in reverse, as indicated by the following theorem.
T H E O R E M 4.5 Integration by substitution
Let f g, and u be differentiable functions of x such that
Then
∕( x ) dx = [ g(ιι) ^ ς d x = [ g(u) du = G(u) + C
where G is an antiderivative of g.
Proof: If G is an antiderivative of g, then G'(u) = g(u) and, by the chain rule,
∕ ω = ⅛[G(<oι = G '(> )g = s ( u ) ⅛
Integrating both sides of this equation, we obtain

∫ f( x ) d x = ∣ [g(u)⅛ ] dx = ∫ [⅛G(W)]
dx = G(M) + C
as required.
EXAM PLE 4 Integration of an indefinite integral by substitution
Find J 9(x2 + 3x + 5)8(2x + 3) dx.
Solution Look at the problem and make the observations as shown in the boxes:
Let u = x 2 + 3x + 5. <— I f u = x 2 + 3x + 5, then du = (2x + 3) dx.
J 9(x2 + 3x + 5)8(2x + 3) dx = 9u 8 du
This is g(ιι). This is du.
Now complete the integration of g(u) and, when you are finished, back-
substitute to express u in terms of x:
f ∕,,9∖
9w8 du = 9(⅜-) + C = (x 2 + 3x + 5)9 + C
Sometimes the value for du is not quite as obvious as that shown in Example
4, so a more general procedure is shown next:
j 9(x 2 + 3x + 5)8(2x + 3) dx
Let u = x 2 + 3x + 5, so = 2x + 3, which implies dx -

Substitute these values:
du
u 2x + 3
9(x2* *4+ 3x + 5)8 (2x + 3) dx = 9w8(2x + 3) 2 ⅛
9u8 du
The goal here is to have an expression of

the form ∫ g(u) dιι— that is, all terms
involving x and dx should be eliminated.
EXAM PLE 5 Substitution with a radical function
Find J V 3 x + 7 dx.
Solution Let u = 3x + 7, so du = 3 dx or dx = η y substitute:
J V 3 x + 7 dx = j V u = j ∣ MIZ2 du
= V + C = ∣w 3z2 + C = ∣(3x + 7)3z2 + C

∖ 2 /
0 After you have made your substitution and simplified, there should be
no leftover x-values in the integrand. 0
EXAM PLE 6 Substitution with leftover x-values
Find J x(4x - 5)3 dx.
Solution Let u = 4x - 5, so du = 4 dx implies dx = substitute:
x(4x - 5)3 d x = [ x u 3 (⅛ ) = 4 f xw3 du

J *J ↑
There is a leftover x value.
We are not ready to integrate until the leftover x-term has been eliminated.
Because u = 4x - 5, we can solve for x:
_ u+ 5
xv ~ 4
so
4 f x u 3 du = 4 f ( ~4 ~ W du = -j⅛ f (μ* + 5 ι∕3) du
4 I 4 J ∖ 4 / 10 J v
= ⅛ (f+ 5 ⅝ ) + C = ⅛(4X - 5 ) 5 + ⅛(4X - 5 ) 4 + C _
If you are using a software program to integrate functions such as the one
shown in Example 6, you may obtain
x 5 - 60x 4 + 100x 3 - x2 +
j Z 04
This does not look anything like our answer to Example 6. However, by
expanding the answer shown in Example 6, you can see they are the same
(except for a constant).
EXAM PLE 7 Change of variable by substitution
x dx
Find
V x 2 + 1’
Solution Let u = x 2 + 1, so du = 2x dx and dx = substitute:
/du∖ . .
x dx f -⅛ 2 = 1 I u -m du = + C = (X 2 + i) ιz2 + C
J Vu 2 J 2 ∖ 1 ∕2 ∕
Vx2 + 1
EXAMPLE 8 Substitution with a trigonometric function
Find ( 4 - 2 cos θ)3 sin θ dθ.

Solution Let u = 4 - 2 cos θ, so ⅛ = -2 (v-s in θ) and dθ = ~ C⅛1- A', substitute:

aθ ’ 2 sin 0’
∣ (4 - 2 cos 0)3sin θ dθ = f w3sin θ ~ i

a = I ∣ ιι du
J J 2 sιnθ 2 J
= ∣ 0 θ + C = ∣(4 - 2 cos 0)4 + C _
As we have previously noted, the forms found by using software will

sometimes differ considerably from the forms obtained without the software.
For example, using software we obtain the following form for Example 8:
2 cos4 0 - 1 6 cos3 0 - 64 cos 0 - 4 8 sin2 0
To show the forms are equivalent, we can change this form to cosines:
2 cos4 0 - 1 6 cos 3 0 + 48 cos 2 0 - 64 cos 0 - 4 8
Now, we expand the answer shown in Example 8 to obtain
∣(4 - 2 cos 0)4 + C = 2 cos4 0 - 1 6 cos3 0 + 48 cos2 0 - 64 cos 0 + 32 + C
You can now see that for an appropriate choice of C , these forms are the
same. The process of reconciling a software answer and an answer from
human calculation is not very different from the process of proving that
equations in trigonometry are identities.
EXAM PLE 9 Using a trigonometric identity with substitution
Find j sin 2x dx.
Solution Let u = 2x, so du = 2 dx so that dx = substitute:
J sin 2x dx = J sin u = ∣ J sin u du = - j cos u + C = - ∣cos 2x + C
If you wish to check your work with integration problems, you can check by
using differentiation:
cos2x + C) = - ∣ ( - 2 sin2x) + 0 = sin2x
Here is a problem in which the rate of change of a quantity is known and we

use the method of substitution to find an expression for the quantity itself.
E X A M P L E 10 Find the volume when the rate of flow is known
Water is flowing into a tank at the rate of √3z + 1 ft 3∕m in. If the tank is empty
when t = 0, how much water does it contain 5 min later?
Solution Because the rate at which the volume V is changing is dV∕dt,

⅛ = √ 3zT T
dt
d V = √3z + 1 dt
∣ d V = ∫ V3z + 1 dt = ∫ √ i∕
u = 3t ÷ 1
du = 3 dt
V + C l = ∣ J u l/2 du
= 3 1/3/2 + We need different c o n s t a n t s f ° r the

different integrals.
V = ∣(3z + l) 3z2 + C where C = C 2 - C 1
F(0) = j(3 ∙ 0 + l) 3z2 + C = 0 so that C = - $ The initial

volume is 0.
F(z) = j(3z + l) 3z2 - j and K(5) = ∣(16) 3z2 - j = 14
The tank contains 14 ft 3 of water.
■ SUBSTITUTION WITH DEFINITE INTEGRATION

Example 10 could be considered as a definite integral:
√3z + 1 dt = j(3z + l) 3z2∣θ = j(16) 3z2 —∣ ( l ) 3z2 = 14
Notice that the definite integral eliminates the need for finding C. Furthermore,
the following theorem eliminates the need for returning to the original variable.
THEOREM 4.6 Substitution with the definite integral
Suppose/is continuous on the set of values taken on by g. If g' is continuous on

[a, b], and if/ h a s an antiderivative F on that interval, then
0 Note that when x = b, then fb rg(b)
u = g(b)∙, when x = a, then ∕[⅛(λ ')]^ , (λ ') dx = f(u) du where u = g(x), du = g , (x) dx
Ja ' Jg(a)
M = g(d). 0
provided these integrals exist.
Proof: Let u = g(x) and let F b e an antiderivative of/on the interval [a, ⅛], so
that F ' = / By the chain rule,
∕[ g ( Λ j] lg , (λ') dx = ∕( r ∕) du = F'(μ) du
Then the fundamental theorem of calculus tells us that
f /[.?(*)]<§■'(*) dx = F'(u) du
'a
I u=g(b)
Because u = g(x)
= F[g(b)] - F[(g(a)]
Because F is an antiderivative
o ff ■
Sec. 4.3 The Fundamental Theorem of Calculus; Integration by Substitution 2 85
EXAM PLE 11 Substitution with the definite integral
Evaluate (4Λ^ - 5)3 dx.
Solution Let u = 4 x - 5,
du = 4 dx
If x = 2, then u = 4(2) - 5 = 3.
I I
f2 f3 0 You cannot
(4x - 5)3 dx = ι
J∣ J-ι change variables
↑ ↑ and keep the
If v = 1, then ιι = 4( 1) - 5 = —1. original limits of
- 1 id ∣3 integration. 0
4 ' 4 ∣-ι
= ⅛ (8 1 -1 ) = 5
Notice that substitution with the definite integral does not require that you
return to the original variable.
PROBLEM SET 4 .3
Γ4 r°
In Problems 1-22, evaluate the definite integral. 19. a. 1 Vx dx b. V - x dx
rlθ r7 0 J-4
1. 7 dx 2. (—3) dx
J-10 J -5 20. a. I' ∖ x dx b. f dx
f5 f2 0 J—16
3. (fix + a) dx 4. (b - x) dx r3
J-3 1-2 f3 Γo
21. a. I ∣5x∣
<7x b. ∣
5x∣dx c. 5x dx
f2 f1 0 J-3 J-3
5. ax 3 dx 6. + (x 3 bx 2) dx r°
J-ι l-ι f2 f2
22. a. ∣
! ∣x ∣dx b. x dx C. v dx
f - ∣α 0 J-2 J-2
7. [ dx 8. — dx
J1 x 1-2 X
Γ1 b. ( (8x 7 + Vτr) dx Use substitution to evaluate the indefinite integrals in Problems

9. a. (5w7 + τr2) du
Jo JcI 23-39.
Γ4 y
10. V x (x + 1) dx 11. f >∕t(t — V~t) dt 23. 1I (2x + 3)4 dx 24. √3Z - 5 dt
Jo )o
4 2 4z5
12. f^ λ ^3 t 13. f x + χχ - 1 d χ 25. ∣ 26. I (11 - 2x)^ dx
1 dx I (x - 27)2z3 dx
Ji x ∣ι v
f3 Pπ∕4
14. (sin 2 x + cos2x) dx 15. (sec2 x —tan⅛) dx 27. I1 (x 2 - cos 3x) dx 28. csc2 5z dt
J-2 Jo
r4 b. [ 4 f i t + 4 )- 1/2 dt 2
16. a. (fit + 4) dt 29. ∣sin (4 - x) dx 30. 1 s"fis + 4 ds
Jo J()
rπ∕2 '7Γ∕2
f (6x - 9) dx
17. a. sin θ dθ b∙ I sin 20 dθ 31. ∣ V t(t i ' 2 + 5)3 dt 32.
Jo J() J (x 2 — 3x + 5)3
,y∕π
33. a. J x(3x 2 - 5)5 dx
18. a. ∣ cost dt b∙ 1 t cos t 2 dt b. x(3x - 5)5 dx
Jo J()
34. a. ∣x V l x 2 - 5 dx b. J xV'2x - 5 dx Use the results o f Problem 54 to evaluate the integrals given in
Problems 55-58.
rτr f π ∕2
b. J x V l x + 1 dx
35. a. 1 x V 2 x 2 + 1 dx 55. sin x√ x 56. cosxdx
J -π J - ιr ∕2
b. J x 3(x 2 + 9) 1/2 dx sπ x d*
X + J1
∣ x 2(x 3 + 9)1/2 dx
36. a. 1 57. Γ x V x 4 + 1 dx 58. f'
J—
3 J-l
b. J x 3(x 2 + 4) 1/2 dx 59. In each o f the following cases, determine whether the
37. a. I∣ x(x 2 + 4)1/2 dx given relationship is true or false.
r l7 5
a. (7x l0 ° l + 14x99) dx = 0
38. 1 x sin (3 + x 2) dx sin 3ZcosZ<Λ
39. J-175
', '7T CTT
Find the area o f the region under the curves given in Problems
40-47.
J 0
sin 2x dx =
Jo
cos⅛ dx
rττ∕2 rθ
40. y = x 2 + 1 on [ - 1, 1] 41. y = √Z on [0, 1] c. cosxd x = sin x√ x
J - ιτ ∕2 J -π
42. y = cos x on [ - y / ]
43. y = sin x + cosx on 0 ,τ 60. The slope at each point (x, y) on the graph o f y = F(x) is
L ' oj given by x (x 2 - l) ιz3, and the graph passes through the
point (3, 1). Use this information to find F. Sketch the
44. y = tV t 2 + 9 on [0, 4] 45. on ⅛ 1]
graph of F.
46. y = x (x - l ) ιz3 on [2, 9] 47. y = ∣x∣ on [2, 3] 61. A particle moves along the x-axis in such a way that at
time t, its velocity is v(Z) = Z2(Z3 - 8)ιz3.
θ 48. ■ What Does This Say? What is the relationship be- a. At what time does the particle turn around?
tween finding an area an< evaluating an integral?
b. If the particle starts at rest at x = 1, where does it turn
around?
49. ■ What Does This Say? Discuss x ιz2 dx.
J-4 62. A rectangular storage tank has a square base 10 ft on a
50. a. If F(x) = J - dx, find F so that F(l) = 0. side. Water is flowing into the tank at the rate of
Z(3Z2 + l)~,z2 ft 3∕s

b. Sketch the graphs of y = F(x), y = F(x) + 3, and
at time Z seconds. If the tank is empty at time Z = 0, how
y = ∙f,W - i∙
much water does it contain 4 s later? What is the depth of
c. Find a constant C such that the largest value of
the water (to the nearest quarter inch) at that time?
G(x) = F(x) + C i s 0.
63. J O U R N A L P R O B LE M * Evaluate
51. T H IN K T A N K P RO BLEM Find a function whose graph has
a relative minimum when x = 1 and a relative maximum [(x2 - l)(x + 1)] 2/3 dx
when x = 4.
52. Evaluate θ 64. T H IN K T A N K P RO BLEM The purpose of this problem is
to provide a counterexample showing that the integral of
∫0 ∕W Λ " h e r e Λ , v ) - { - J '; 4 J * ∕ ', a product (or quotient) is not necessarily equal to the
product (quotient) of the respective integrals,
a. Show that J x V x dx ≠ (∫ x dx)(∣ V x dx).
53. Evaluate
π f ∖∕ f ( V x dx
∕(r x∕ )∖ dx,
J ι ∕(rιx )∖ = fco sx i f θ ≤ x < τ τ ∕2
JQ
where
IΛ II al~~ X — ττ
b. Show that ——dx ≠ ----- -—.
J x ∫ x dx
65. Suppose/is a function with the property that f ∖x) = f(x)
54. a. Show that if f is continuous and odd [that is,
- for all x.
∕( - x ) = ∕(x )] on the interval [ - a , a], then
ra
f(x ) dx = 0 a. Show that ∕( x ) dx = f(b) - f(a).
J-α
b. Show that if / is continuous and even [/(—x) = ∕(x )] f b
[∕(x )] 2 dx = ∣ {[∕(⅛)]2 - [∕(f l )]2}.

b. Show that
on the interval [—a, a], then la
'0
f(x ) dx = 2 ∕( x ) dx = 2 ∕( x ) dx *From the College Mathematics Journal, Sept. 1989, p. 343.
Jo Problem by Murray Klamkin.
Sec. 4.4 Introduction to Differential Equations 287
4.4 INTRODUCTION TO DIFFERENTIAL EQUATIONS
IN THIS SECTION Introduction and terminology, separable differential

equations, orthogonal trajectories, the flow of a fluid through an orifice, escape
velocity of a projectile
The study of differential equations is as old as calculus itself. Today, it would be
virtually impossible to make a serious study of physics, astronomy, chemistry,
or engineering without encountering physical models based on differential
equations. In addition, differential equations are beginning to appear more
frequently in the biological and social sciences, especially in economics. In this
section we introduce you to the basic terminology of differential equations and
show how such equations can be used to model applications. Later in the text,
we shall investigate more and more sophisticated issues regarding differential
equations as we develop the necessary mathematical tools.
■ INTRODUCTION AND TERMINOLOGY
Any equation that contains a derivative or differential is called a differential

equation. For example, the equations
dy '<⅛Λ 2 .d v
= 3x 3 + 5 C = kF2 + 3 τx + ^ = χ y
⅛ + 2 ⅛ ' + 5t = sint
dx dt √X ∕ dt 2 dt
are all differential equations.
Many practical situations, especially those involving rates, can be described
mathematically by differential equations. For example, the assumption that
population P grows at a rate proportional to its size can be expressed by the
differential equation
C = kp
dt
where t is time and k is the constant of proportionality (we will examine this
equation in Chapter 6).
A solution of a given differential equation is a function that satisfies the
equation. A general solution is a characterization of all possible solutions of the
equation. We say that the equation is solved when we find a general solution.
For example, y = x 2 is a solution of the differential equation
dy
= lx
dx
because
Tx = T x ^ = T x ^ = 2 x
Moreover, because any solution of this equation must be an indefinite integral of

2x, it follows that
y = 2x dx = x 2 + C
is the general solution of the differential equation.
This section is required only for Section 6.4.

EXAM PLE 1 Solving a differential equation to find future revenue
An oil well that yields 300 barrels of crude oil a day will run dry in 3 years.
It is estimated that t days from now the price of the crude oil will be
p(f) = 30 + 0.3√Z dollars per barrel. If the oil is sold as soon as it is extracted
from the ground, what will be the total future revenue from the well?
Solution Let R(t) denote the total revenue up to time t. Then the rate of change of
revenue is the num ber of dollars received per barrel is p(t) = 30 + 0.3Vt,
and the num ber of barrels sold per day is 300. Thus, we have
Γ RATE OF CHANGE ^l _ ΓNUMBER OF DOLLARS! ΓNUMBER OF BARRELS!
LOF TOTAL REVENUE J ~ L PER BARREL J L SOLD PER DAY J
= (30 + 0.3√i)(300)
= 9,000 + 9 0 √ ∕
This is actually a statement of the chain rule . dR _ dR dB , where B denotes

’ dt dB d t
the num ber of barrels extracted and R denotes the revenue. It is often helpful
to use the chain rule in this way when setting up similar differential equations.
We solve this differential equation by integration:
R = f dt
J dt
R(f) = 9,000z + 60t 3z2 + C
Because Λ(0) = 0, it follows that C = 0. We also are given that the well will run
dry in 3 years or 1,095 days, so that the total revenue obtained during the life
of the well is
jR(l,095) = 9,000(1,095) + 60(l,095) 3z2

≈ 12,029,064.52
The total future revenue is $12,029,065.
■ SEPARABLE DIFFERENTIAL EQUATIONS
Many useful differential equations can be formally rewritten so that all the terms
containing the independent variable appear on one side of the equation and all
terms involving the dependent variable appear on the other side. Thus, such an
equation has the form
g(x) dx = ∕( p ) dy
or, equivalently (integrating both sides):
∣ g(x) dx = f f ( y ) dy
A differential equation that can be written in the form

d y . g(x)
dx f ( y )
is said to be separable. To solve a separable differential equation, separate the
variables and then separately integrate both sides o f the resulting equation. This
procedure is illustrated in Example 2.
EXAMPLE 2 Separable differential equation
Solve T = *
dx y
Solution dy = x
dx y
y dy = x dx
y dy = j"x dx
⅛y2 + c 1 = ¼ 2 + c 2
x1 - y 2 = C where C = 2(C, - C2)
Notice the treatment of constants in Example 2. Because all constants can be

combined into a single constant, it is customary not to write C = 2(Cj - C 2), but
Family of rather to simply replace all the arbitrary constants in the problem by a single
curves
arbitrary constant at the end of the problem.
The remainder of this section is devoted to selected applications involving
Orthogonal trajectory
separable differential equations.
The orthogonal trajectory inter
sects each curve in the given ■ ORTHOGONAL TRAJECTORIES
family at right angles.
A curve that intersects each member of a given family of curves at right angles is
called an orthogonal trajectory of that family. Orthogonal families arise in many
applications. For example, in thermodynamics, the heat flow across a planar
surface is orthogonal to the curves of constant temperature, called isotherms. In
the theory of fluid flow, the flow lines are orthogonal trajectories of velocity
potential curves. The basic procedure for finding orthogonal trajectories involves
differential equations and is demonstrated in Example 3.
EXAMPLE 3 Finding orthogonal trajectories
Find the orthogonal trajectories of the family of curves of the form

xy = C
Figure 4.9 Scattering experiment Solution We are seeking a family of curves. Each curve in that family intersects
can reveal the quantum mechani each curve in the family xy = C at right angles, as shown in Figure 4.10.
cal behavior of particles using or Assume that a typical point on a given curve in the family xy = C has
thogonal trajectories. Two guns coordinates (x, y) and that a typical point on the orthogonal trajectory curve
shoot particles toward one anoth has coordinates (X, T).
er. The particles that are deflected
by 90o are recorded in two detec
tors. The solid lines show one set
of possible trajectories for the de
flected particles. The dotted lines
show another set.
We use uppercase letters for one

kind of curve and lowercase for
the other to make it easier to tell
which curve is being mentioned Figure 4.10 The family of
at each stage of the following curves xy = C and their
discussion. orthogonal trajectories
Let P be a point where a particular curve of the form xy = C intersects the

orthogonal trajectory curve. At P, we have x = X and y = Y, and the slope
dY/dX of the orthogonal trajectory is the same as the negative reciprocal of the
slope dy/dx of the curve xy = C. Using implicit differentiation, we find
xy = C
X +y = 0 Product rule
^y = ∑y
dx x
dY
Thus, at the point of intersection P, the slope jγ - of the orthogonal trajec
tory is
dY = ____ L _ = ____ L = x = X
dX dy/dx —y/x y Y
According to this equation, the coordinates (X, Y ) of the orthogonal trajectory
curve satisfy the separable differential equation
dY = X
dX Y
discussed in Example 2. Using the result of that example, we see that the
orthogonal trajectories of the family xy = C are the curves in the family
X 2 - Y2 = K
where K is a constant. The given family of curves xy = C and the family of
Computer-generated family orthogonal trajectory curves X 2 — Y 2 = K are shown in Figure 4.10.
of orthogonal trajectories
■ THE FLOW OF A FLUID THROUGH AN ORIFICE
Consider a tank that is filled with a fluid being slowly drained through a small,
sharp-edged hole in its base, as shown in Figure 4.11.
By using a principle of physics known as Torricelli’s law,* we can show that
the rate of discharge dV/dt at time t is proportional to the square root of the depth
h at that time. Specifically, if all dimensions are given in terms of feet, the drain
hole has area ∠40, and the height above the hole is h at time t (seconds), then
f = -4 .S⅛ √ ⅛
Figure 4.11 The flow of a fluid is the rate of flow of water in cubic feet per second. This formula is used in
through an orifice
Example 4.
EXAM PLE 4 Fluid flow through an orifice
A cylindrical tank is filled with a liquid that is draining through a small

circular hole in its base. If the tank is 9 ft high with radius 4 ft, and the drain
hole has radius 1 in., how long does it take for the tank to empty?
Solution Because the drain hole is a circle of radius ⅛ ft (= I in.), its area is
πr2 = τr(⅛)2 , and the rate of flow (according to Torricelli’s law) is
dV= _ 4 o ∕S τ J∖ λ ∕τ
‘Torricelli’s law says that the stream of liquid through the orifice has velocity v = Y lg h , where
g = 32 ft∕s 2 is the acceleration due to gravity and h is the height of the liquid above the orifice. The
factor 4.8 that appears in the rate of flow equation is required to compensate for the effect of friction.
Because the tank is cylindrical, the amount of fluid in the tank at any
particular time will form a cylinder of radius 4 ft and height h. Also, because
we are using g = 32 ft∕s 2, it is implied that the time, t, is measured in seconds.
The volume of such a liquid cylinder is
V = τrr2h = ττ(4)2∕z = ∖ 6ττh
and by differentiating both sides of this equation with respect to t, we obtain
dV _ 1 6f
, dh
dt dt
- 44 v(y∑Λ∖v ∕∕,n - 16τr⅛J By substituting from Torricelli’s law

Λ 144∕ dt
dh _ -4.8V⅞ -0.0021V∕z
dt 144(16)
~7fdh = -0.0021 dt Separate the variables.

yh
∫ ∕z~ιz2 dh = f (-0.0021) dt Integrate both sides.
2h',2 + C 1 = -0.002 k + C2
2√Λ = -0 .0 0 2 k + C
To evaluate C, recall that the tank is full at time t = 0. Thus h = 9 when t = 0,
so that
2√9 = -0.0021(0) + C and C = 6
Therefore, the general formula is
2√A = -0 .0 0 2 k + 6
Now we can find the depth of the fluid at any given time or the time at
which a prescribed depth occurs. In particular, the tank is empty at the time te
when h = 0. By substituting h = 0 into this formula we find
2Vθ = - 0 .002k + 6 so that t ≈2,857 sec
Thus roughly 0.8 hr (47.62 min) are required to drain the tank. BBM
■ ESCAPE VELOCITY OF A PROJECTILE
Consider a projectile that is launched with initial velocity v0 from a planet’s

surface along a direct line through the center of the planet, as shown in Figure
4.12.
On April 12, 1981, the United

States launched the world’s first
space shuttle, Columbia. Figure 4.12 A projectile launched from the surface of a planet
We shall find a general formula for the velocity of the projectile and the minimal
value of v0 required to guarantee that the projectile will escape the planet’s
gravitational attraction.
We shall assume that the only force acting on the projectile is that due to
gravity, although in practice, factors such as air resistance must also be consid
ered. With this assumption, Newton’s law of gravitation can be used to show that
when the projectile is at a distance 5 from the center of the planet, its acceleration
is given by the formula
where R is the radius of the planet and g is the acceleration due to gravity at the
planet’s surface.*
Our first goal is to express the velocity v of the projectile in terms of the
distance s. Because the projectile travels along a straight line, we know that
a = Tt and v = ⅛
and by applying the chain rule, we see that

„ - d v - d v .d s _ d v ,
dt ds dt ds
Therefore, by substitution for a we have
-g R 2
7dsΓ v = cι2
v dv = - g R 2s~ 2 ds
J v dv = J - g R 2s~ 2 ds
∣v 2 + Cj = g R 2s~' + C 1
v2 = 2gR 2s~' + C
To evaluate the constant C , recall that the projectile was fired from the planet’s
surface with initial velocity v0 . Thus, v = v0 when s = R , and by substitution:
v0 2 = 2gR 2R~' + C which implies C = VQ 2 - 2gR
so
v2 = 2gR 2s~' + v0 2 - 2gR
Because the projectile is launched in a direction away from the center of the
planet, we would expect it to keep moving in that direction until it stops. In other
words, the projectile will keep moving away from the planet until it reaches a point
where v = 0. Because 2gR 2s~ λ > 0 for all s > 0, v2 will always be positive if
v0 2 - 2gR ≥ 0. On the other hand, if v0 2 - 2gR < 0, then sooner or later v will
become 0 and the projectile will eventually fall back to the surface of the planet.
Therefore, we conclude that the projectile will escape from the planet’s
*According to Newton’s law of gravitation, the force of gravity acting on a projectile of mass m
has magnitude F = mkts1 , where k is constant. If this is the only force acting on the projectile, then
F = ma, and we have ma = F = m k∕s 2. By canceling the m’s on each side of the equation, we obtain
a = kts^.
gravitational attraction if v0 2 ≥ 2gR— that is, if v0 ≥ V⅛R. For this reason, the
minimum speed for which this can occur, namely,
v0 = V⅛Λ
is called the escape velocity of the planet. In particular, for the earth, R = 3,956 mi
and g = 32 ft∕s 2 = 0.00606 mi∕s 2, and the escape velocity is
v0 = ∖i'2gR ≈ √2(0.006 06)(3,956) ≈ 6.924 357
The escape velocity for the earth is 6.92 mi∕s.
PROBLEM SET 4 .4
θ Verify in Problems 1-6 that i f y satisfies the prescribed relation 24. When a person is asked to recall a set of facts, the rate at
ship with x, then it will be a solution o f the given differential which the facts are recalled is proportional to the number
equation. of relevant facts in the person’s memory that have not yet
1. If x 2 + y 2 = 7, then ⅛∙ = — been recalled.
z ax y 25. The rate at which an epidemic spreads through a commu
2. If 5x 2 - 2y2 = 3, then ⅛ = nity of P susceptible people is proportional to the product
dy 2y of the number of people who have caught the disease and
3. If xy = C, then 4 - = — . the number who have not.
dx x 26. The rate at which people are implicated in a government
4. If x 2 — 3xy + y 2 = 5, then scandal is proportional to the product of the number of
(2x - 3y) dx + (2y — 3x) dy = 0. people already implicated and the number of people
involved who have not yet been implicated.
cp v
5. If y = sin (Ax + B), then / - + j ⅛2y ~ θ∙
θ Find the orthogonal trajectories o f the fam ily o f curves given in
6. If y = ⅛ - - + B, then x +2 = x 3. Problems 27-31. In each case, sketch several members o f the
20 x dx 1 dx given family o f curves and several members o f the fam ily o f
orthogonal trajectories on the same coordinate axes.
Find the general solution o f the separable differential equations
given in Problems 7-16. 27. the lines 2x —3y = C
7 d y = _x
28. the lines y = x + C
8,⅞ -x ≈ √ i
dx y 29. the cubic curves y = x 3 + C
9 - 7 Γ = 2β
30. the curves y = x 4 + C
dx ∖y 31. the curves xy 2 = C
11. c- γ = - V l - x 2 >2. ⅛> O'-0= 32. What do you think the orthogonal trajectories of the
ax y family of curves
13. xy dx + ∖ f^xy dy = 0 14. 4y3 <⅛ + 5z 2 dy = 0
x2 - y2 = C
15 ⅛ = s in ∙x
16. x 2 dy + secy dx = 0 will be? Hint. Take another look at Example 3.
dx cosy
33. The following is a list of six families of curves. Sketch
In Problems 17-20, find the general solution o f the given
several members of each family and then determine
differential equation by using either the product rule or the
which pairs are orthogonal trajectories of one other.
quotient ride.
a. the circles x 2 + y 2 = A
Π .x d y + y d x = 0 18. 4 = ⅛ -r⅛ . 0 b. the ellipses 2x 2 + y 2 = B 2
19. y dx = x dy, x > 0, y > 0 20. x 2y dy + xy 2 dx = 0 c. the ellipses x 2 + 2y2 = C
Write a differential equation describing the situation given in d. the lines y = Cx
Problems 21-26. Do not solve. e. the parabolas y 2 = Cx
21. The number of bacteria in a culture grows at a rate that is f. the parabolas y = Cx 2
proportional to the number present. 34. A cylindrical tank of radius 3 ft is filled with water to a
22. A sample of radium decays at a rate that is proportional to depth of 5 ft. Determine how long (to the nearest minute)
the amount of radium present in the sample. it takes to drain the tank through a sharp-edged circular
23. The rate at which the temperature of an object changes is hole in the bottom with radius 2 in.
proportional to the difference between its own tempera 35. Rework Problem 34 for a tank with a sharp-edged drain
ture and the temperature of the surrounding medium. hole that is square with length of side 1.5 in.
36. EXPERIMENT A rectangular tank has a square base 2 ft 44. After its brakes are applied, a certain sports car deceler
on a side that is filled with water to a depth of 4 ft. It is ates at a constant rate of 28 ft∕s 2 . Compute the stopping
being drained from the bottom of the tank through a distance if the car is going 60 mi/hr (88 ft∕s) when the
sharp-edged square hole that is 2 in. on a side. brakes are applied.
a. Show that at time t, the depth h satisfies the differen 45. SPY P ROBLEM The spy, having defused the bomb in
tial equation Problem 19, Section 3.7, is driving the sports car in
Problem 44 at a speed of 60 mi/hr on Highway 1 in the
4 ⅛ = -4.8A nV h remote republic of San Dimas. Suddenly he sees a camel
at 0
in the road 199 ft in front of him. After a reaction time of
00.7 seconds, he steps on the brakes. Will he stop before
where A g is the area of the drain hole.
hitting the camel?
b. How long will it take to empty the tank?
46. A scientist has discovered a radioactive substance that
c. Construct this tank and then drain it out of the 2 in.2 disintegrates in such a way that at time t, the rate of
hole. Is the time that it takes consistent with your disintegration is proportional to the square of the amount
answer to part b? present.
37. A toy rocket is launched from the surface of the earth with a. If a 100-g sample of the substance dwindles to only
initial velocity v0 = 150 ft∕s. (The radius of the earth is 80 g in 1 d, how much will be left after 6 d?
roughly 3,956 mi, and g = 32 ft∕s 2 .)
b. When will only 10 g be left?
a. Determine the velocity (to the nearest ft∕s) of the
47. In physics, it is shown that the amount of heat Q (calories)
rocket when it is first 200 feet above the ground.
that flows through an object by conduction will satisfy the
(Remember, this is not the same as 200 ft from the
differential equation
center of the earth.)
b. What is s when v = 0? Determine the maximum height d^ = - k A c^
dt fc 4 ds
above the ground that is attained by the rocket.
38. Determine the escape velocity of each of the following where t (seconds) is the time of flow, k is a physical
heavenly bodies: constant (the thermal conductivity of the object), A is the
a. moon (R = 1,080 mi; g = 5.5 ft∕s 2) surface area of the object measured at right angles to the
b. Mars (R = 2,050 mi; g = 12 ft∕s 2) direction of flow, and T is the temperature at a point 5
c. Venus (R = 3,800 mi; g = 28 ft∕s 2) centimeters within the object, measured in the direction
of the flow, as shown in Figure 4.13.
39. Population statistics indicate that x years after 1990, a
certain city was growing at a rate of approximately Figure 4.13 Heat conduction
1,5OOZ~lz2 people per year. In 1994 the population of the through an object
city was 39,000.
a. What was the population in 1990?
b. If this pattern continues, how many people will be
living in the city in the year 1999?
40. In a certain section of the country, the price of chicken is
currently $3 per pound. It is estimated that t weeks from Under certain conditions (equilibrium), the rate of heat
now the price will be increasing at a rate of V3Z + 1 cents flow dθldt will be constant. Assuming these conditions to
per week. How much will chicken cost 8 weeks from now? exist, find the number of calories that will flow each
41. It is estimated that t years from now a farmer’s crop will second across the face of a square pane of glass 2 cm thick
be increasing at a rate of 0.3Z2 + 0.6/ + 1 bushels per day. and 50 cm on a side if the temperature on one side of the
By how much will the value of the crop increase during pane is 50 and the temperature on the other side is 60°.
the next 20 days if the market price remains fixed at $3 The thermal conductivity of glass is approximately
per bushel? k = 0.0025.
42. The radius of planet X is one-fourth that of planet T, and 48. The shape of a tank is such that when it is filled to a depth
the acceleration due to gravity at the surface of X is eight of h feet, it contains
ninth that at the surface of T. If the escape velocity of
V = 9πΛ3 ft 3
planet X is 6 ft∕s, what is the escape velocity of planet T?
43. A survey indicates that the population of a certain town is of water. The tank is being drained through a sharp-edged
growing in such a way that the rate of growth at time t is circular hole of radius I in. If the tank is originally filled
proportional to the square root of the population P at the to a depth of 4 ft, how long does it take for the tank to
time. If the population was 4,000 ten years ago and is empty?
observed to be 9,000 now, how long will it take before 49. A rectangular tank has a square base 4 ft on a side and is
16,000 people live in the town? 10 ft high. Originally, it was filled with water to a depth of
Sec. 4.5 The Mean Value Theorem for Integrals; Average Value 295
6 feet, but now is being drained from the bottom of the b. On a certain planet, it is known that g = 25 ft∕s 2 . A
tank through a sharp-edged square hole 1 in. on a side, projectile is fired with an initial velocity of v0 = 2
a. Find an equation involving the rate dh∕dt. mi/sec and attains a maximum height of 450 mi. What
b. How long will it take to drain the tank? is the radius of the planet?
θ 50. A projectile is launched from the surface of a planet 51. A projectile is launched from the surface of a planet
whose radius is R and where the acceleration due to whose radius is R. The constant acceleration due to
gravity at the surface is g. gravity on the surface of the planet is g. According to
Newton’s law of gravitation, the force of gravity acting on
a. If the initial velocity v0 of the projectile satisfies
a projectile of mass m has magnitude
v0 < y ∕ rλgR, show that the projectile eventually re
turns to the surface of the planet. Also show that the
maximum height above the surface of the planet Fc∙ _ mk
- ^
reached by the projectile is
v0 -R where k is a constant. If this is the only force acting on the
h= n 2
2 projectile, then F = ma. Show that
⅛ R - v0
a = =J^
2 s
4.5 THE MEAN VALUE THEOREM FOR INTEGRALS; AVERAGE VALUE
IN THIS SECTION Mean value theorem for integrals, average value of a

function, variable limits of integration, Leibniz’s rule
The fundamental theorem o f calculus simplifies the procedure for evaluating
the definite integral providing we can find an antiderivative. However, this
theorem does not tell us whether every continuous function actually has an
antiderivative. In this section we prove the second fundamental theorem of
calculus, which says that a continuous function on an open interval has an
antiderivative on that interval. In order to prove this major result we first need
to establish a result called the mean value theorem for integrals.
■ MEAN VALUE THEOREM FOR INTEGRALS
In Section 3.2 we established a very useful theoretical tool called the mean value
theorem, which said that under reasonable conditions, there is at least one number
The red rectangle has the c on the interval [a, /?] such that the value of the derivative at x = c is exactly the
same area as the region under same as
Λ b) ~ fia )
b- a
The mean value theorem for integrals is similar, and in the special case where
∕' (x) ≥ 0 it has a geometric interpretation that makes the theorem easy to
understand. In particular, the theorem says that it is possible to find at least one
number c on the interval [a, b] such that the area of the rectangle with height ∕( c )
and width (b - a) has exactly the same area as the region under the curve y = fix')
on [a, b∖ . This is illustrated in Figure 4.14.
THEOREM 4.7 Mean value theorem for integrals
If f is continuous on the interval [a, b], there is at least one number c between a
Figure 4.14 Geometric and b such that
interpretation of the mean fd
f(x ) dx = fic)(b - a)
value theorem for integrals Ja
Proof: Suppose M and m are the largest and smallest values o f/ respectively,
on [a, b]. This means that
in ≤ f(x ) ≤ M when a ≤ x ≤ b
rb rb rb
m dx ≤ ∕( x ) dx ≤ Λ f dx Dominance property
Ja Ja Ja
rb
m(b - a) ≤ f(x ) dx ≤ M(b - a)
Ja
1 f' j
m ≤ ∕( x ) dx ≤ M
Ja
The intermediate value theorem (Section 1.7) says that because f is continuous on
the closed interval [a, b] and because the number
1 fb
J = I fix ) dx
lies between m and M , there exists a number c between a and b for which∕( c ) = I;
that is, b
-j^∑ra ∫a ∕W dx = ∕( c )
[ fix ) dx = ∕(c)(⅛ - a) ■
The mean value theorem for integrals does not specify how to determine c, it
simply guarantees the existence of at least one number c in the interval. However,
Example 1 shows how to find a value of c guaranteed by this theorem.
E X A M P L E 1 Finding c in the mean value theorem for integrals
Find a value of c guaranteed by the mean value theorem for integrals for
∕( x ) = sinx on [0, τr].
Solution sinx dx = -c o s x = - c o s τr + cos 0 = —( - 1) + 1 = 2

Jo ∣0
The region bounded by / and the x-axis on [0, π] is shaded in Figure 4.15.
The mean value theorem for integrals asserts the existence of a number c on
[0, π] such that f(c)(b - a) = 2. We can solve an equation to find this value:
∕(c)(∕> - a) = 2
(sin c)(τr - 0) = 2
sin c = —
77
c ≈ 0.690107, or 2.451486
Because c is between 0 and 77, we have found the value c guaranteed by the
Figure 4.15 Graph of∕(x ) = sinx mean value theorem for integrals. ∣=∑>
illustrating the mean value theo
rem on [0, ττ] ■ AVERAGE VALUE OF A FUNCTION
There are many practical situations in which one is interested in the average value
of a continuous function on an interval, such as the average level of air pollution
over a 24-hour period, the average speed of a truck during a 3-hour trip, or the
average productivity of a worker during a production run. Averages of this kind
can be computed by the formula in the following definition.
Average Value I f / is continuous on the interval [a, b], the average value off on this interval is
given by the integral
b
1
b⅜ Jfa dx
You probably know that the average value of n numbers x l , x 2, . . . , x is

x i + X 2 + ’‘’ + x n
n
but what if there are infinitely many numbers? Specifically, what is the average
value of ∕( x ) on the interval a ≤ x ≤ bc! To see how the definition of finite
average value can be used, imagine that the interval [a, £>] is divided into n equal
subintervals, each of width
∆x = ^
n
For k = 1, 2, . . . , n, let x"'k be a number chosen arbitrarily from the ∕<th
subinterval. Then the average value V o f/ o n [a, ⅛] is estimated by the weighted
sum
x
Q ∙∕z( x
v tl)z + f (∖ ?) +
J
+ f{∖ x n∖) 1ι ∖∕"
2 . / ________________ J
*∖ γ J f(γ "
- n ^ n *⅛ ' ' k'
Because ∆x = - — - , we know that - = — ∆x and

n ’ n b- a
s. - ⅛⅜ - [⅛ δ⅛ M) =⅛ I, M>iv
1
The sum on the right is a Riemann sum with norm ∣∣P∣∣ = and because
∣∣P∣∣ → 0 as n → +∞, we have
∣ n 1 rb
lim S = lim τ— - ∑ /(x?) ∆x = τ— - ∕( x ) dx
∣∣p∣∣→o n n→∞b - a kΞ∖j κ k , b - a ]a j x ,
Finally, because it is reasonable to expect the estimating average S n to approach
the “ true average value” as n → +∞, we are led to the formula given as the
definition of average value.
EXAM PLE 2 Average speed of traffic
Suppose a study suggests that between the hours of 1:00 P . M . and 4:00 P.M. on
a normal weekday, the speed of the traffic at a certain expressway exit is
modeled by the formula
S(t) = 2z3 - 2 1Z2 + 60Z + 20
kilometers per hour, where t is the number of hours past noon. Compute the
average speed of the traffic between the hours of 1:00 P.M. and 4:00 P.M.
Solution Our goal is to find the average value of 5'(∕) on the interval [1, 4]. The
average speed is
4⅛ γ ∫ (2t3 - 21/ + 60t + 20) dt = ⅜[∣Z4 - 7z3 + 30Z2 + 20z] ∣
= ∣ ( 2 4 0 - 43.5) = 65.5
Note that the mean value theorem for integrals can be interpreted as saying
that the average value of f(x) on [a, b] equals the value of f for some number c
between a and b. That is, a function must assume its average value at least once on
any closed, bounded interval [a, b∖ where it is continuous.
Actually, this is quite reasonable, because the intermediate value theorem of
continuous functions assures us that a continuous function /assumes every value
between its maximum M and minimum m, and we would expect the average value
lz to be between these two extremes. Example 3 illustrates one way of using these
ideas.
EXAM PLE 3 Average temperature
Suppose that x hours after midnight, the temperature (in degrees Celsius) in
Minneapolis one night can be modeled by the formula
T(x) = 2 - ⅜(x - 13)2
Find the average temperature between 2 A.M. and 2 P.M. and a time when the
average temperature actually occurs.
Solution We wish to find the average temperature of T on the interval [2, 14]
(because 2 P.M. is 14 hr after midnight). The average value is
T = T 4 ⅛ f 2 ' 4 I 2 “ ⅛x ~ 13 >2 ] d x = ⅛ [2 x - ÷ ∙ ⅛U - 1 3) 3 ∣ a"
=⅛⅛ z- ⅛ r ] ≈ -3.2857143
Thus the average temperature on the given time period is approximately 3.286
0C below zero. To determine when this temperature actually occurs, solve the
equation
AVERAGE TEMPERATURE = TEMPERATURE AT TIME X
-3.2857143 = 2 - ⅜(x - 13)2

37 = ( x - 13)2
x = 13 ± V 3 7 ≈ 19.082763 or 6.9172375
The first value is to be rejected because it is not in the interval [2, 14], so we
find that the average temperature occurs 6.917 hr after midnight, or at
approximately 6:55 A.M. ≡κ
■ VARIABLE LIMITS OF INTEGRATION
The fundamental theorem of calculus tells us that the definite integral can be
evaluated as a difference:
∕(x ) dx = F(b) - F(a)
where F is an antiderivative of f on [a, &]. But in general, what guarantee do we

have that such an antiderivative even exists? The answer is provided by the
following theorem, which is often referred to as the second fundamental theorem
of calculus.
THEOREM 4.8 Second fundamental theorem of calculus

Let /(/) be continuous on the interval [a, b] and define the function G by the
integral equation
G(x) = [ 7 ( 0 dt
Ja
for a ≤ x ≤ b. Then G is an antiderivative of f on [a, b]', that is,
G '^ - fx f(f)d t = f(x)
'a
on [a, b].
Proof: Apply the definition of derivative to G:
r ,.(x). = hm ---------
G
G(x + Δ x )-G (x )
---------—
V ,Δx→ 0 ∆X
= lim ⅛ [G(x + ∆x) - G(x)]
= lim -/- /( 0 dt - f(f) dt

∆A'→ 0 Δ JT Ja
1 Γ ΛX+ΔX
= Δx→
lim0 L
ΛΛ7- /( 0 dt +
Λ Ja
, Γ ΛX+ΔX
= lim -τ- jf(f)
y ’ dt Subdivision property
Δ,v→ 0 ∆x Jx
= lim ⅛ [ ∕ ( C ) Δ *1 This is true for c
ΔX→0 ∆1Λ
between x and x + Ax;
mean value theorem
for integrals.
- lim f(c) Since c is between
c→x x and x + ∆x and
Δx→ 0
= ∕w Since f is continuous at x
Notice that if F is any antiderivative of f on the interval [a, b], then the
antiderivative
G(x) = [ /(/) dt
Ja
found in Theorem 4.8 satisfies G(x) = F(x) + C for some constant C and all x on
the interval [a, b]. In particular, when x = a, we have
0 = [ 7 ( 0 dt = G(a) = F(a) + C
Ja
so that C = -F(a). Finally, by letting x = b, we find that
/(/) dt = G(Z>) = F(b) + C = F(⅛) + [-F(α)] = F(⅛) - F(α)
This provides an alternative proof of the (first) fundamental theorem.
EXAMPLE 4 Using the second fundamental theorem
Differentiate F(x) = [ ( 2 / - 3) dt.

J7
Solution From the second fundamental theorem, we can obtain F'(x) by simply
replacing t with x in the integrand ∕( t) = It - 3. Thus
F'(x) = ⅛ [ Γ (21 - 3) dt] = 2x - 3 -
The second fundamental theorem of calculus can also be applied to an integral

function with a variable lower limit of integration. For example, to differentiate
G(z) = fr≡ ■ du
Jz
reverse the order of integration and apply the second fundamental theorem of
calculus, as before:
σ'ω=⅛ ^ d u = τL ^ d u _ _ d_ f^ sin u i ∣u _ sinz
u dz l5 U d∑ J5 u z
■ LEIBNIZ’S RULE
Occasionally, it is necessary to differentiate an integral in which one (or both) of

the limits of integration is a function. More specifically, suppose ιι(x) is a
differentiable function of x and that F is the function defined by
r»W
^(w) = /(1) dt
Ja
for a continuous function/ Then F is a function of ιι, and the second fundamental
theorem of calculus tells us that
f =⅛[P<'>A ] =Λ">
However, 77 can also be regarded as a composite function of x, and by applying the
chain rule, we find that
dF _ dF du
dx du dx
so that
M f ∕< H = ⅛^λ ", ≡

More generally, an integral with differentiable functions in both limits of
integration can be differentiated by applying the following useful theorem.
THEOREM 4.9 Leibniz’s rule

If M(X) and v(x) are differentiable functions of x, then
⅛[C m λ ] =/(u) ⅛^/<v) ⅛

Proof: A formal proof of Leibniz’s rule requires the chain rule for a function
of two variables and will be given in Section 12.5. The plausibility of this rule can
be demonstrated by noting that for any number a between v and u
[ f(f) dt = f f(t) dt + f ∕(t) dt = - f f(t) dt + [ f(t) dt
Jv Jv Ja Ja Ja
and thus, by differentiating, we obtain

d
dx /(0 dt
EXAM PLE 5 Applying Leibniz’s rule
r x 2-3 x
D ifferentiate F(x) = tan t dt.
J∖⅛
Solution We shall apply L eibniz’s rule w ith ∕(Z ) = tanZ, u(x) = x 2 - 3x, and
v(x) = V x . Accordingly,
du _ _ √v _ 1 . 1
dx~ 29 x i1 and
dx 2 vGc
so that
= (2x - 3) tan (x 2 - 3x) - tan V x ∙ —-
= (2x - 3) tan (x 2 - 3x) -
PROBLEM SET 4 .5
θ In Problems 1-6 fin d c such that Determine the area o f the indicated region in Problems 13-18
and then draw a rectangle with base (b - a) and height f(c) for
∕ W dx = f(c)(b - a) some c on [a, b] so that the area o f the rectangle is equal to the
area o f the given region.
as guaranteed by the mean value theorem for integrals. I f you 13. y = ∣ x on [0, 10] 14. y = x 2 on [0, 3]
cannot find such a value, explain why the theorem does not
apply.
1. f(x ) = 4x 3 on [1, 2]
2. f(x ) = x 2 + 4x + 1 on [0, 2]
3. f(x ) - 15x^2 on [1, 5]
4. f(x ) = 12x^3 on (0, 3)
5. f(x ) = 2 csc2 x on [-j> -j]
6. ∕( x ) = cosx on [-J>
Use the second fundamental theorem o f calculus to find the

derivative o f F in Problems 7-12. Next, verify your answer by
integrating to fin d F and then differentiating the result. 15. y = x 2 + 2x + 3 on [0, 2] 16. y = p on [0.5, 2]
7. F(x) = ∣
| (4z + 9) dt 8. F(x) = I∣ (3t2 - 4t + 5) dt
0 0
z
9. F(x) = If (λ∕ Z- 4) dt P t dt
10. F(x) = Il √ i + 3t2
4 1
11. F(x) = IP dt 12. F(x) = ι sec t tan t dt

ι (1 -3 z ) 3 tτ∕3
J = -⅛ z ∣2 + t ∖t o + zo2l + ⅛v o(z ι + zo)

18. y = x + sinx on [0.5, 2]
36. What is the average velocity for the object described in

Problem 35 during the same time period?
37. Records indicate that t hours past midnight, the tempera
ture at the local airport was
/(/) = - 0 .3 ∕ 2 + 4 /+ 10
degrees Celsius. What was the average temperature at the

airport between 9:00 A.M. and noon?
38. Suppose a study indicates that t years from now, the level
Find the average value o f the function given in Problems 19-30 o f carbon dioxide in the air of a certain city will be
on the prescribed interval.
19. ∕( x ) = x 2 —x + 1 on [ - 1, 2] L (∕) = (/ + 1)2
20. ∕( x ) = x 3 - 3x 2 on [—2, 1] parts per million (ppm).
21. ∕( x ) = sinx on [θ, a. What is the average level of carbon dioxide in the first
3 years?
22. ∕( x ) = 2 sin x - cosx on [θ, f] b. At what time (or times) does the average level of
23. ∕( x ) = √ 4 - χ on [0, 4] 24. ∕( x ) = λ ∕ z l - x o n [ - 7 , 0] carbon dioxide actually occur? Answer to the nearest
25. ∕( x ) = (2x —3)3 o n[0,1] 26. ∕( x ) = x V 2 x 2 + 7 on [0,1] month.
39. Let /(/) be a function that is continuous and satisfies
27. ∕( x ) = x (x 2 + 1)3 on [ - 2 , 1] 28. ∕( x ) = v ^ 2x + ι on [0, 3] /(/) ≥ 0 on the interval [θ, f ] . Suppose it is known that
for any number x between 0 and y> the region under the
29. ∕( x ) = V 9 - x 2 on [ - 3 , 3] graph of / on [0, x] has area A(x) = tan x.
Hint. The integral can be evaluated as part o f a circle.
30. ∕( x ) = V 16 - x 2 on [0, 4] a. Explain why /(/) dt - tan x for 0 ≤ x ≤
Hint. The integral can be evaluated as part o f a circle. Jo
© Use Leibniz’s rule to differentiate the integral functions given b. Differentiate both sides o f the equation in part a and
deduce the identity o f f
in Problems 31-34.
40. T H IN K T A N K P ROBLEM Suppose that /(/) is continuous
31. F(x) = J cos/2 dt 32. G(x) = J
, J^^*dt
f for all / and that for any number x it is known that the
average value o f/ o n [ - 1 , x] is
r ∖-t ∙ rVv ι
33. H{t) = J 3 ≡ - dx 34. G(y) = ∫ ’ ⅛ 1 A(x) = sinx
35. If an object is propelled upward from ground level with an Use this information to deduce the identity of /
initial velocity vfl, then its height at time / is given by 41. Suppose that /(/) is continuous for all / and that for any
number x the average value o f / on [x, x 2] is
x = - ∖gt 1 + v0 ∕
where g is the constant acceleration due to gravity. Show A(x) = x

that between times t0 and ∕ 1, the average height o f the
object is Find/(2) given that/(4) = 2∕(2). Hint: Use Leibniz’s rule.
4.6 NUMERICAL INTEGRATION: THE TRAPEZOIDAL RULE AND SIMPSON'S RULE
IN THIS SECTION Approximation by rectangles, trapezoidal rule,

Simpson’s rule, error estimation
The fundamental theorem of calculus can be used to evaluate an integral
whenever an appropriate antiderivative is known. However, certain functions
have no simple antiderivative. To find a definite integral of such a function, it is
often necessary to use numerical approximation. In this section, we shall
examine approximation by rectangles, by trapezoids, and by parabolic arcs
(through Simpson’s rule). It will help if you have access to a calculator or
computer.
Sec. 4.6 Numerical Integration: The Trapezoidal Rule and Simpson's Rule 303
■ APPROXIMATION BY RECTANGLES
If f(x ) ≥ 0 on the interval [a, ⅛ the definite integral f( x ) dx is equal to area
under the graph of f on [a, ⅛]. As we saw in Section 4.1, one way to approximate
this area is to use n rectangles, as shown in Figure 4.16. In particular, divide the
interval [a, A] into n subintervals, each of width ∆x = ~ a , and let x k denote the
right endpoint of the ∕cth subinterval. The base of the Zdh rectangle is the Ath
subinterval, and its height is f(x*k ). Hence, the area of the Ath rectangle is∕(x ^ )∆ x
The sum of the areas of all n rectangles is an approximation for the area under the
curve and hence an approximation for the corresponding definite integral. Thus,
Figure 4.16 Approximation by r⅛
rectangles ∕( x ) dx ≈ ∕(x * ) ∆ x + ∕(x * ) ∆ x + ∙ ■∙ + ∕(x * ) ∆ x
This approximation improves as the number of rectangles increases, and we

can estimate the integral to any desired degree of accuracy by taking n large
enough. However, because fairly large values of n are usually required to achieve
reasonable accuracy, approximation by rectangles is rarely used in practice.
■ THE TRAPEZOIDAL RULE
The accuracy of the approximation generally improves significantly if trapezoids

are used instead of rectangles. Figure 4.17 shows the area from Figure 4.16
approximated by n trapezoids instead of rectangles. Even from these rough
illustrations you can see how much better the approximation is in this case.
Suppose the interval [a, b] is partitioned into n equal parts by the subdivision
points x 0 , x l , . . . , x n , where x 0 = a and x n = b. The ∕cth trapezoid is shown in
greater detail in Figure 4.18.
Figure 4.17 Approximation by

trapezoids
Recall the formula for
the area of a trapezoid:
A= ^ (b l + b 2 )h
Figure 4.18 The ∕cth trapezoid has area ⅜[∕-(¾ - l ) + ∕( x fc)]∆x.
If we let Tn denote the sum of the areas of n trapezoids, we see that

τ n = 2[∕(⅞ ) + ∕(*1 )] Δ * + ⅛[∕(⅞) + ∕(⅞ )]Δ ^ + " ∙ + I [/(*„- 1) + ∕(⅞ )]∆ X
= ∣[∕( ⅞ ) + W 1) + 2 ∕( X 2) + ∙∙∙ + 2 ∕(x n.,) + ∕( x n)]∆x
The sum Tn estimates the total area under the curve y = f(x) on the interval [a, ⅛]
and hence also estimates the integral
r⅛
∕( x ) dx
This approximation formula is known as the trapezoidal rule and applies as a

means of approximating the integral, even if the function/is not positive.
Trapezoidal Rule L e t/b e continuous on [a, ⅛]. The trapezoidal rule is

cb
f(χ ) r/x ≈ ∣ [ ∕( x 0) + 2 ∕(x,) + 2 ∕(x 2) + - + 2 ∕( x n . 1) + ∕( x n)]∆x
ja
where ∆x = - — - and, for the ∕<thκ subinterval, x l = a + ∕c∆x. Moreover, the

n
larger the value for n, the better the approximation.
Our first example uses the trapezoidal rule to estimate the value of an integral
that we can compute exactly by using the fundamental theorem.
E X A M P L E 1 Trapezoidal rule approximation
f2
Use the trapezoidal rule with n = 4 to estimate x 2 dx.
J-ι
Solution The interval is [a, b] = [ - 1, 2], so a = -1 and b = 2. Then,
2 -(-1 ) τ
Δ Λ- = ---- 1— '- = ¾ = 0.75. Thus,
⅞ = a = -1 ∕( x 0) = / (-1 ) = ( - l ) 2 = l
x ι = a + i ∙ ∆x = —i + 1 = —∣ 2 ∕(x,) = 2 ( 4 ) 2 = ∣ = 0.125
X 2 = a + 2 ∙∆ x = - l+ f = ∣ 2 ∕( Λ-2) = 2(⅛)2 = 2 = 0.5

2 ∕(x 3) = 2 ( ∣) 2 = f = 3.125
X 3 = a + 3 ∙ Δ Λ' = —1 + ∣ = ¾
Computational Window B I B
x
4
= a + 4 ∙ ∆x = b = 2 ∕(⅞ ) = 22 = 4
Software for numerical evalua
tion of integrals is common T4 = ⅛[l + 0.125 + 0.5 + 3.125 + 4](0.75) = 3.28125
(see the Technology Manual).
The output for Example 1 is The exact value of the integral in Example 1 is
shown:
f x 2 √x = ⅜-∣ = ⅞ — T^ = 3
Type of # of sub Estimate over J —1 3I * J J
estimate intervals [-1,2]
Therefore, the trapezoidal estimate T4 involves an error, which we denote by E 4 .
Trapezoid 4 3.28125
We find that
Trapezoid 10 3.045
Trapezoid 100 3.00045 ^4 = x 2 dx - T4 = 3 - 3.28125 = -0.28125
Trapezoid 1000 3.000005
The negative sign indicates that the trapezoidal formula overestimated the true
value of the integral in Example 1.
---------
■ SIMPSON’S RULE
Roughly speaking, the accuracy of a procedure for estimating the area under a
curve depends on how well the upper boundary of each approximating strip fits
the shape of the given curve. Trapezoidal strips often result in a better approxima-
tion than rectangles, and it is reasonable to expect even greater accuracy if the
approximating strips have curved upper boundaries, as shown in Figure 4.19. The
name given to the procedure in which the approximating strip has a parabolic arc
for its upper boundary is called Simpson’s rule.*
We shall derive Simpson’s rule as a means for approximating, as before, the
area under the curve y = f(x ) on the interval [a, b], where / is continuous and
satisfies ∕( x ) ≥ 0. First, we partition the given interval into a number of equal
subintervals, but this time, we require the number of subdivisions to be an even
number (because this requirement will simplify the formula associated with the
final result). If x 0 , x i , . . . , x n are the subdivision points in our partition (with
x 0 = a and x n = b), we pass a parabolic arc through the points, three at a time (the
points with x-coordinates x 0 , x 1, x 2 , then those with x 2, x y x 4, and so on). It can be
shown (see Problem 31) that the region under the parabolic curve y = f(x) on the
interval [x2*-2> x 2*l has area
∣ [ ∕( ⅞ t - 2) + 4∕( ⅛ - ι ) + ∕( ⅛ ) ] δ ∙v
where ∆x = ~ ~ ∙ This procedure is illustrated in Figure 4.20.
Figure 4.20 Approximation using parabolas
By adding the area of the approximating parabolic strips and combining terms, we
obtain the sum S n of n parabolic regions:
S n = ∣[ ∕( ⅞ ) + 4 ∕( x 1) + ∕( x 2)] ∆x + ⅜[∕(x2) + 4 ∕( X 3) + ∕( x 4)] ∆x
+ + 3 [∕( ⅞ - 2) + 4∕( ∙V ι) + ∕( x χ
n)]∆
= 3 [∕( χ 0) + 4∕( χ 1) + 2∕⅛ ) + 4∕(⅞ ) + ∙∙∙ + 4 ∕(x π. 1) + ∕(⅞ )]∆ x

These observations are summarized in the following box:
Simpson’s Rule L e t/b e continuous on [a, Z>]. Simpson’s rule is

f f(x ) dx ≈ ∣ [ ∕( x 0) + 4 ∕(x 1) + 2 ∕( X 2) + + 4 ∕(x n , 1) + ∕( x π)]∆x
Ja ∙j
where ∆x = x k = a + k∆x, k an integer and n an even integer. Moreover,

the larger the value for n, the better the approximation.
*This rule is named for Thomas Simpson (1710-1761), an English mathematician, who,
curiously, neither discovered nor made any special use of the formula that bears his name.
EXAMPLE 2 Approximation by Simpson’s rule
Use Simpson’s rule with n = 10 to approximate —.

Ji λ
Q Computational Window Q B
Solution We have ∆x = 1θ = 0.1, and x 0 = a = 1, x χ = 1.1, x 2 = 1.2, . . . ,
x g = 1.9, x ∣
0 = 2.
Computer output for Example
2 is shown: ft - d x ≈ S 10
Simpson’s rule to calculate estimate.
Type of # of sub Estimate over
_ 1( । 4 l 2 , 4 ∣ 2 . 4 . 2 l 4 ∣ 2 , 4 , jΛ∕∩ ∣ ∖
estimate intervals 11,2]
- 3 ∖1 1.1 1.2 + 1.3 1.4 + 1.5 + 1.6 + 1.7 + 1.8 1.9 + 2 ∕ υ ' υ
Simpson 10 0.6931502 0.6931502
Trapezoid 10 0.6937714
Trapezoid 100 0.6931534
■ ERROR ESTIMATION
Simpson 100 0.6931472
Simpson 1000 0.6931472 The difference between the value of an integral and its estimated value is called its
error. Because this error depends on n, we denote it by E n.
THEOREM 4.10 Error in the trapezoidal rule and Simpson’s rule

If/h as a continuous second derivative on [a, b∖, then the error E n in approximat
ing ∫" f(x )d x by the trapezoidal rule satisfies
(⅛ —α )3
Trapezoidal error: ∣
Ej ≤ 2 M, where M is the maximum value of ∣ o n [<≈, ⅛]
∕ " (x )∣
Moreover, if /h a s a continuous fourth derivative on [a, b∖, then the error E n

(n even) in approximating ∖ab f(x )d x by Simpson’s rule satisfies
Simpson’s error: ∣
Ej ≤ where K is the maximum value of | / (4)(x)| on [a, b↑
Historical Note Proof The proofs of these error estimates are beyond of scope of this book;
they can be found in many advanced calculus textbooks and in most numerical
analysis books. ■
EXAMPLE 3 Estimate of error when using Simpson’s rule
2 dx
Estimate the accuracy of the approximation of f — by Simpson’s rule with
n = 10 in Example 2. ∙^1
Solution If ∕(x ) = we find that / (4)(x) = 24x 5. The maximum value of this
function will occur at a critical value (there are none on [1, 2]) or at an
endpoint. Thus, the largest value of ∣∕ < 4,(x)∣on [1, 2] is | / <4)( 1)∣
= 24. Now,
apply the error formula with K = 24, a = 1, b = 2, and n = 10 to obtain
PIERRE SIMON LAPLACE K(b - α)5 = 24(2 - 1)5
(1749-1827) |£J ≤ ≈ 0.0000133
180n4 ~ 180(10)4
Newton and Leibniz have been
credited with the discovery of That is, the error in the approximation in Example 2 is guaranteed to be no
calculus, but much of its refιne- greater than O.OOOO133.
ment was due to the mathema

With the aid of the error estimates, we can decide in advance how many
ticians Laplace, Lagrange, and subintervals to use in order to achieve a desired degree of accuracy.
Gauss. Pierre Simon Laplace
was a French mathematician
EXAMPLE 4 Choosing number of subintervals to guarantee given accuracy
who did outstanding work in
differential equations, which
How many subintervals are required to guarantee that the error will be less
were introduced in this chapter. than 0.00005 in the approximation of
He was also known for his f 2 dx
work in celestial mechanics, ge Ji *
odesy, and probability. We will
meet Lagrange and Gauss in on [1, 2] using the trapezoidal rule?
other historical notes, but these
Solution Becausef(x ) = x ^ 1, we havef" (x ) = 2x~3. On [1, 2] the largest value of
three great mathematicians of
calculus were contrasted by W. ∣
∕" ( x ) ∣ ∕ ''( 1 ) ∣
is ∣ = 2, so M = 2, a = 1,6 = 2 and
W. Rouse Ball.
The great masters of modern
analysis are Lagrange, Laplace, The goal is to find the smallest positive integer n for which
and Gauss, who were contempo
raries. It is interesting to note ¼ < 0.00005
7
the marked contrast in their on-
styles. Lagrange is perfect both 10,000 < 3n2 Multiply by the positive number 60,000M2
in form and matter; he is care
ful to explain his procedure, and 10,000- 3n2 < 0
though his arguments are gener (100 - √3n)(100 + √3n) < 0
al they are easy to follow. La
place on the other hand explains n < -iθ fi or ∏ > 57.735027
nothing, is indifferent to style,
and, if satisfied that his results The smallest positive integer that satisfies this condition is n = 58; therefore,
are correct, is content to leave
them either with no proof or 58 subintervals are required to ensure the desired accuracy.
with a faulty one. Gauss is exact If/is the linear function f(x ) = Ax + B, then f" (x ) = 0 and we can take M = 0 as
and elegant as Lagrange, but the error estimate. In this case, the error in applying the trapezoidal rule satisfies
even more difficult to follow ∣E j ≤ 0. That is, the trapezoidal rule is exact for a linear function, which is what
than Laplace, for he removes ev we would expect, because the region under a line on an interval is a trapezoid.
ery trace of the analysis by In discussing the accuracy of the trapezoidal rule as a means of estimating the
which he reached his results,
and strives to give a proof which value of the definite integral
',Z>
while rigorous shall be as con
cise and synthetical as possible. Ja
∕(x ) dx
we have focused attention on the “error term,” but this only measures the error
A Short Account o f the History o f
Mathematics as quoted in Mathematical
that comes from estimating I by the trapezoidal or Simpson approximation sum.
Circles Adieu by Howard Eves (Boston: There are other kinds of error that must be considered in this or any other method
Prindle, Weber & Schmidt, Inc., 1977). of approximation. In particular, each time we cut off digits from a decimal, we
incur what is known as a round-off error. For example, a round-off error occurs
when we use 0.66666667 in place of ∣or 3.1415927 for the number π. Round-off
errors occur even in large computers and can accumulate to cause real problems.
Specialized methods for dealing with these and other errors are studied in
numerical analysis.
The examples of numerical integration examined in this section are intended
as illustrations and thus involve relatively simple computations, whereas in
practice, such computations often involve hundreds of terms and can be quite
tedious. Fortunately, these computations are extremely well-suited to automatic
computing, and the reader who is interested in pursuing computer methods in
numerical integration will find an introduction to this topic in the computer
supplement.
PROBLEM SET 4 .6
θ Approximate the integrals in Problems 1-2 using the trapezoi
dal rule and Simpson’s rule with the specified number of
subintervals and then compare your answers with the exact
value of the definite integral. 19. A quarter-circle of radius 1 has the equation y = Vl - x 2
for 0 ≤ x ≤ 1, which means that
x 2 dx with /7=4
V l - x 2 dx = 2∣
-4 r~
V x dx with n = 6
J
o
Approximate the integrals given in Problems 3-6 with the
Estimate τr correct to one decimal place by applying the
trapezoidal rule to this integral.
20. A quarter-circle of radius 1 has the equation y = V l - x 2
specified number of subintervals using: a. the trapezoidal rule for 0 ≤ x ≤ 1, which means that
b. Simpson's rule.
3. f' — with /7=4 V l - x 2 dx = 22

Jo 1 ÷
r 0 ,----------- =τ
Estimate τr correct to one decimal place by applying
4. V 1 + x 2 dx with n = 4 Simpson’s rule to this integral.
J-ι 21. Find the smallest value of n for which the trapezoidal rule
∙4 , estimates the value of the integral
J 2
"2
V l ÷ sinx dx with n = 4
x^ 1 dx
θ
J o
x cosx dx with n = 6
7. ■ What Does This Say? Describe the trapezoidal rule.

with six decimal-place accuracy.
22. The following graph (from the Wall Street Journal, Febru
ary 26, 1986) shows the 10-year treasury bond rate for the
8. ■ What Does This Say? Describe Simpson’s rule. years 1980-1985. Estimate the area in terms of the
rectangles shown on the grid under this curve using a
trapezoidal approximation where n = 6, and each year is
Estimate the value of the integrals in Problems 9-14 to within 1 unit.
the prescribed accuracy.
r, rm
9. -f÷-— with error less than 0.05
Jo x 2 + 1
r2 _____
10. Vl + x 2 dx with error less than 0.05
J-ι
*ι
J o
cos2 x dx accurate to three decimal places
12. x~ i dx accurate to three decimal places

Ji
23. The width of an irregularly shaped dam is measured at
^4 ______
5-m intervals, with the results indicated in Figure 4.21.
J
J
o
Λ TT
xV4 —x dx with error less than 0.1
Use Simpson’s rule to estimate the area of the face of the
dam.
θ cos2 θ dθ with error less than 0.01
o
In Problems 15-18, determine how many subintervals are
required to guarantee accuracy to within 0.00005 using
a. the trapezoidal ride b. Simpson's rule
f4
15. 16. (x3 + 2x2 + 1) dx
j-ι Figure 4.21 Area of the face of a dam
Sec. 4.6 Num erical Integration: The Trapezoidal Rule and Sim pson’s Rule 309
27. a. Repeat Problem 26, except now let ∕( x ) =

9x - x 3, on [0, π]. You may be surprised by the
result o f Simpson’s rule; but, after some thought,
24. Apply the trapezoidal rule to the following data explain this situation.
found on a spreadsheet (table):
b. Simpson struggles! In contrast to part a, here is an
example where Simpson’s rule does not live up to
expectations. For n = 10, 20, 40, and 80, use
Simpson’s rule and the rectangular rule (select
midpoints) for this integral and make a table of
errors. Then try to explain Simpson’s poor perfor
mance. Hint: Look at the formula for the error.
r2
v'4 - x 2 dx = π
Jo
θ 28. Show that if p(x) is any polynomial of degree less than or

equal to 3, then
£ p(x) dx = ^-=-^[p(α) + + P(⅛)]
25. Apply Simpson’s rule to the following data found on

a spreadsheet (table): This result is often called the prismoidal rule.
29. Use the prismoidal rule (Problem 28) to evaluate
f2
I (x 3 - 3x + 4) dx
J-1
30. Use the prismoidal rule (Problem 28) to evaluate
Γ3
(x 3 + 2x 2 - 7) dx
J-1
31. Let p(x) be a polynomial of degree at most 3.
a. Show there is a number c between 0 and 1 such that
r!
p(x) dx = p(c) + p (-c )
J —1
b. Show that there is a number c such that
r l/ 2
p(x) dx = ⅜[p(-c) + p(0) + p(c)]
J —1/2
26. The idea here is to study the “ order o f convergence” 32. Let p(x) = A x 3 + B x 2 + C x + D be a polynomial. Show
of three numerical integration methods; that is, how that Simpson’s rule gives the exact value for
the error goes to zero as n→ ∞. As summarized in f *i
Theorem 4.10, the error in the trapezoid rule behaves P(x) dx
Ja
like M∕n~ and the error in Simpson’s rule behaves
33. The object of this exercise is to prove Simpson’s rule for
like K !n ∖ hence for large n values Simpson’s rule is
usually much more accurate. It can also be shown the special case involving three points.
that the error in the rectangular rule (select mid a. Let P 1( - ∕r , ∕(-⅛ )), P 2(0, /(0)), P∕h, f(h)). Find the
points) is about ∣ that of the trapezoid rule. In this equation o f the form y = A x 2 + B x + C for the parab
problem, you are to demonstrate this notion. ola through the points P l , P 2, and P y
a. By using the fundamental theorem, show that b. If y = p(x) is the quadratic function found in part a,
show that
sinxcZx= 2 (exactly)
Jo Cl' ι
p(x) dx = ⅛[P(-h) + 4P(0) + P(⅛)]
b. Now, for several values of n (say, 10, 20, 40, and J-A i
80) approximate the above integral by the three
numerical methods. Build a table of the errors; c. Let β 1(x 1, ∕( x 1)), <2,(x,, ∕( x 2)), 0 3(x 3, ∕( x 3)) be points
thus, verify the above discussed order of conver with x 2 = x + h, and x 3 = x 1 + 2lι. Explain why
gence. In particular, from your table, find approx £ p(x) dx = ∣ [ P ( x 1) + 4P(X 2) + P(x 3)]
imate values o f M and K.
4 . 7 AREA BETWEEN TWO CURVES
IN THIS SECTION Area between curves, area by vertical strips, area by

horizontal strips, two applications to economics
We have seen how the area between the curves defined by y = ∕( x ) and the x-
rb
axis can be computed by the integral ∕( x ) dx on an interval a ≤ x ≤ b
Ja
where ∕( x ) ≥ 0. In this section, we shall find that integration can be used to
find the area of more general regions between curves. We shall also see how two
im portant economic applications can be considered as an area between two
curves.
■ AREA BETWEEN CURVES
In some practical problems, you may have to compute the area between two
curves. S u p p o se/an d g are functions such that ∕( x ) ≥ ,y(x) on the interval [a, b∖,
as shown in Figure 4.22. Note that we do not insist that both / and g are
nonnegative functions, but we begin by showing that case in Figure 4.22.
Area between y =f(x) and y = g(x) is equal to the area under y =f(x) minus the area under y = g(x)
Figure 4.22 Area between two curves
To find the area of the region R between the curves from x = a to x = b, we

subtract the area between the lower curve y = g(x) from the area between the
upper curve y = ∕( x ) and the x-axis; that is,
rb rb rb
a. Approximating rectangles
AREA OF R = ∕( x ) dx - g(x) dx = [∕(x ) - g(x)] dx
Ja Ja Ja
This formula seems obvious in the situation where both ∕( x ) ≥ 0 and tg(x) ≥ 0,
as shown in Figure 4.22. However, the following derivation requires only that f
and g be continuous and satisfy ∕( x ) ≥ <g(x) on the interval [a, b]. We wish to find
the area between the curves y = ∕( x ) and y = g(x) on this interval. Choose a
partition {x0 , x 1, x 2 , . . . , x } of the interval [a, b] and a representative num ber x*
from each subinterval [x, ,, x.]. Next, for each index k, with k = 1, 2, . . . , n,
construct a rectangle of width
∆x i , Xk Xk _ {
and height
∕( ⅞ ) - ∙sr(4)
b. A typical vertical strip
equal to the vertical distance between the two curves at x = x ∕ A typical
Figure 4.23 Using Riemann sums approximating rectangle is shown in Figure 4.23b. We refer to this approximating
to find the area between two curves rectangle as a vertical strip.
Sec. 4.7 Area Between Two Curves 311
The representative rectangle has area

[f( x k ) - M 4)1 Δ ⅜
and the total area between the curves y = f(x ) and y = g(x) can be estimated by
the Riemann sum
∑ L [∕(⅞ ) - M ) ] Δ X ⅛
It is reasonable to expect this estimate to improve if we increase the number of

subdivision points in the partition P in such a way that the norm ∣∣P∣∣ approaches 0.
Thus, the region between the two curves has area
A = lim ∑ l[f7(vx*kj), —gδ v( x↑k ), ]i ∆x,k

∣∣p∣∣→o k = ι ∙
which we recognize as the integral of the function ∕( x ) - g(x) on the interval [a, b∖.
These observations may be used to define the area between the curves.
Area Between Two Curves If/ a n d g are continuous and satisfy ∕( x ) ≥ g(x) on the closed interval [a, b},
then the area between the two curves y = ∕( x ) and y = g(x) is given by
A = f [∕(x ) - g(x)] dx
Ja
What this says: To find the area between two curves on a given closed
interval [a, b] use the formula
fZ>
A = [TOP CURVE - BOTTOM CURVE ] dx
Ja
It is no longer necessary to require either curve to be above the x-axis. In fact,

we will see later in this section that the curves might cross somewhere in the
domain so that one curve is on top for part of the interval and the other curve
is on top for the rest.
E X A M P L E 1 Area between two curves
Find the area of the region between the curves y = x 3 and y = x 2 - x on the
interval [0, 1].
Solution The region is shown in Figure 4.24. We need to know which curve is the
top curve on [0, 1]. Solve
x3 = x2 - x or x(x 2 - x + 1) = 0
The only real root is x = 0 (x2 - x + 1 = 0 has no real roots). Thus, the same
curve is on top throughout the interval [0, 1]. To see which curve is on top,
take some representative value, such as x = 0.5, and note that because
(O.5)3 > 0.5 2 - 0.5, the curve y = x 3 must be above y = x 2 - x. Thus, the
required area is given by
fl l i e
A = [x 3 - (x 2 - x)] dx = (⅛x4 - ⅛x3 + ∣ x 2) l = -⅛
Figure 4.24 The area between the J Q ‰-> k------- v ____ / ×m- j z∙ / IU 1z
curves y = x 3 and y = x 2 - x on Top Bottom
[0, 1] curve curve
In Section 4.1 we defined area for a continuous function f with the restriction
that ∕( x ) ≥ 0. Example 1 shows us that when considering the area between two
curves, we need to be concerned no longer with the nonnegative restriction but
only with whetherf(x ) ≥ g(x). We now use this idea to find the area for a function
that is negative.
E X A M P L E 2 Area with a negative function
Find the area of the region bounded by the curve y = x 2 - 4x and the x-axis.
Solution We find the points of intersection of the curve and the x-axis:
x 2 —4x = 0 The curve intersects the
x-axis where y = 0.
x(x - 4) = 0 so that x = 0, 4
The graph∕( x ) ≤ x 2 - 4x is shown in Figure 4.25. We see that on the interval
[0, 4 ],∕(x ) ≤ 0, but we can find the area of the given region by considering the
area between the curves defined by equations y = 0 and y = x 2 - 4x.
Figure 4.25 Area of region
A = I [0- (x 2 - 4x) ] dx = f - y + 2x2 )∣ 0 = y
J Top curve' ■ '
Bottom curve
■ AREA BY VERTICAL STRIPS
Although the only mathematically correct way to establish a formula involving

integrals is to form Riemann sums and take the limit according to the definition of
the definite integral, we can simulate this procedure with approximating strips.
This simplification is especially useful for finding the area of a complicated region
formed when two curves intersect one or more times, as shown in Figure 4.26.
Note that the vertical strip has height∕( x ) - g(x) if y = f(x) is above y = g(x) and
height g(x) - ∕( x ) if y = g(x) is above y = ∕( x ) . In either case, the height can be
represented by ∣∕(x) - g(x)∣, and the area of the vertical strip is
Δ4 = ∣∕(x) - <g(x)∣∆x = ∣∕(x) - g(x)∣ dx
Thus, we have a new integration formula for area; namely,
Figure 4.26 Area by vertical strips

EXAM PLE 3 Area by vertical strips
Find the area of the region bounded by the line y = 3x and the curve
y = x 3 + 2 x 2.
Solution The region between the curve and the line is the shaded portion of
Figure 4.27. Part of the process of graphing these curves is to find which is the
top curve and which is the bottom. To do this, we need to find where these
curves intersect:
x 3 + 2x 2 = 3x or x (x + 3)(x - 1) = 0
x = 0, - 3 , 1
The points of intersection are x = - 3 , 0, and 1. In the subinterval [ -3 , 0],

labeled A i in Figure 4.27, the curve y = x 3 + 2x 2 is on top (test a typical point
in the subinterval, such as x = -1 ) , and on [0, 1], the region labeled A 2 , curve
y = 3x is on top. The representative vertical strips are shown in Figure 4.27,
Figure 4.27 The area between the
and the area between the curve and the line is given by the sum
curve y = x 3 + 2x 2 and the line
y = 3x. rθ l∙∖
0 You cannot use the formula A = [(x3 + 2x 2) - (3x)] dx + [(3x) - (x 3 + 2x 2)] dx
A = ∫ [ f - g ]d x directly here J-3 Jθ
because the hypotheses f ≥ g is
not satisfied. In order to use
A = ∖∖f - g ∖dx over the entire
interval, you must remember
that ∖f - g∣ might b e / - g over = (or 1 1.83333333 by calculator)
part of the interval and g - f
over another part (see Figure
4.26). Make sure you check to
see which curve is on top. Be
cause the curves cross in Exam ■ AREA BY HORIZONTAL STRIPS
ple 3, the interval must be sub
divided accordingly. 0 For many regions, it is more appropriate to form horizontal strips rather than
vertical strips. The procedure for horizontal strips duplicates the procedure for
vertical strips. Suppose we want to find the area between two curves of the form
x = F(y) and x = G( 1') on the interval [c, d]. Such a region is shown in Figure 4.28,
together with a typical horizontal approximating rectangle of width ∆y, which we
refer to as a horizontal strip.
b. A typical horizontal strip
Figure 4.28 Area by horizontal strips

Note that regardless of which curve is “ ahead” or “ behind,” the horizontal strip
has length ∣F(y) - G(y)∣ and area
ΔA = ∣F(.y) - G(y)∣∆y
However, in practice, you must make sure to find the points of intersection of the
curves and divide the integrals so that in each region one curve is the leading curve
(“ right curve” ) and the other is the trailing curve (“ left curve” ). Suppose the curves
intersect where y = b for b on the interval [c, d}, as shown in Figure 4.28. Then
A = £ [G(y) - F(y)] dy + Γ [F(y) - G(y)] dy

Jb '--------.------- ‘
G ahead of F F ahead of G
EXAM PLE 4 Area by horizontal strips
Find the area of the region between the parabola x = 4y - y 2 and the line
x = 2j∙ - 3.
Solution Figure 4.29 shows the region between the parabola and the line,
together with a typical horizontal strip. To find where the line and the
parabola intersect, solve
4y - y 2 = 2y - 3 to obtain y = - 1 and y = 3
Throughout the interval [ -1 , 3], the parabola is to the right of the line (test a
typical point between - 1 and 3, such as y = 0). Thus, the horizontal strip has
m an area
ΔA = [ ( 4 y - ∕) - ( 2 y - 3 ) ] Δ y
κ____ v____ J s____ v____J
Right curve Left curve
and the area between the parabola and the line is given by
Figure 4.29 The area between the
curves x = 4y - y 2 and x = 2y - 3
using horizontal strips
A = f [(4y - y 2) - (2y - 3)] dy = ( (3 + 2y - y 2 ) dy
J-ι J-∣
= (3y + / - ⅜ ∕) ∣. 31 = (9 + 9 - 9) - ( -3 + 1 + ⅜) = 10j _
In Example 4, the area can also be found by using vertical strips, but the
procedure is more complicated. Note in Figure 4.30 that on the interval [ - 5 , 3], a
representative vertical strip would extend from the bottom half of the parabola
y 2 - 4y + x = 0 to the line y = ∣(x + 3), whereas on the interval [3, 4], a typical
vertical strip would extend from the bottom half of the parabola y 2 - 4y + x = 0
to the top half. Thus, the area is given by the sum of two integrals. It can be shown
that the computation of area by vertical strips gives the same result as that found
by horizontal strips in Example 4.
■ TWO APPLICATIONS TO ECONOMICS
Figure 4.30 The area between the Next, we shall see how definite integration can be used in economics to compute
curves x = 4y —y 1 and x = 2y —3 net earnings generated on time and a special quantity called consumer's surplus.
using vertical strips
Net Earnings. The net earnings generated by an industrial machine on a period
of time is the difference between the total revenue generated by the machine and
the total cost of operating and servicing the machine, as shown in Figure 4.31.
In the following example, the net earnings of a machine are calculated as a definite
integral and interpreted as the area between two curves.
EXAM PLE 5 Profitability of a piece of equipment
Suppose that a piece of equipment is purchased in 1990 and will generate

revenue of R(x) = 5,000 - 20x 2 dollars per year for x years after 1990. At the
same time the cost of maintaining and operating the equipment at that time is
C(x) = 2,000 + 10x2 dollars.
a. For how many years will the use of the equipment remain profitable?
Figure 4.31 Net earnings is the b. What are the net earnings generated by the machine during its period of
difference between total revenue profitability?
and total cost.
Solution The cost and revenue curves are sketched in Figure 4.32.
a. Note that the revenue curve is above the cost curve until they cross, so the
use of the equipment will be profitable as long as the revenue exceeds the
cost; that is, until
5,000 - 20x 2 = 2,000 + 10x2
30x2 = 3,000
x = ±10 (-1 0 is not in the domain)
The equipment will be profitable for 10 yr.
b. The difference F(x) = R(x) — C(x) represents the net earnings of the
machine after x years after 1990. To obtain the total net earnings, we
integrate over the period of profitability; that is, from t = 0 to t = 10.
H0
NET EARNINGS = [R(x) ~ C(x)] dx
JO
H0
= [(5,000 - 20x2) - (2,000 + 10x2)] dx
Jo
H0
= (3 ,0 0 0- 30x2)√x
Jo
, 110
4 j 8 12 ∣ 16 ' * M = (3,000x - 10x3)∣ 0 = 20,000
Figure 4.32 The net earnings from

a piece of equipment This piece of equipment should generate $20,000 of net earnings for the
years 1990-2000 (its 10-year period of profitability). M M
Consumer’s Surplus. In a competitive economy, the total amount that consum

ers actually spend on a commodity is usually less than the total amount they
would have been willing to spend. The difference between the two amounts can be
thought of as a savings realized by consumers and is known in economics as the
consumer’s surplus.
To get a better feel for the concept of consumer’s surplus, consider an example
of a couple who are willing to spend $500 for their first T V set, $300 for a second
set, and only $50 for a third set. If the market price is $300, then the couple would
buy only two sets and would spend a total of 2 × $300 = $600. This is less than the
$500 + $300 = $800 that the couple would have been willing to spend to get the
two sets. The savings of $800 - $600 = $200 is the couple’s consumer’s surplus.
Consumer’s surplus has a simple geometric interpretation, shown in Figure 4.33.
Figure 4.33 Geometric interpretation of consumer’s surplus

Note that p and q denote the market price and the corresponding demand,
respectively. Figure 4.33a shows the region under the demand curve p = D(q)
from q = 0 to q = q0, and the area of this region represents the total amount that
consumers are willing to spend to get q0 units of the commodity. The rectangle in
Figure 4.33b has an area of p 0 q0 and represents the actual consumer expenditure
for q0 units at p 0 dollars per unit. The difference between these two areas (Figure
4.33c) represents the consumer’s surplus. That is, consumer’s surplus is the area of
the region between the demand curve p = D(q) and the horizontal line p = p Q so
that
f'7o Γeô Γ⅜ f⅜
[D(q) - p 0] dq = D(q) dq - p 0 dq = D(q) dq - p q0
Jo Jo Jo Jo
Consumer’s Surplus If q0 units of a commodity are sold at a price of p 0 dollars per unit and if
p = D(r∕) is the consumer’s demand function for the commodity, then
’TOTAL AMOUNT CONSUM-' ACTUAL CONSUMER’
CONSUMER’S
ERS ARE WILLING TO EXPENDITURE FOR
SURPLUS
SPEND FOR qfj UNITS q0 UNITS
f?0
Z>(⅛) dq - p 0 q0
Io
EXAM PLE 6 Consumer’s surplus
Suppose the consumer’s demand function for a certain commodity is

D(q) = 4(25 - q2) dollars per unit. Find the consumer’s surplus if the com
modity is sold for $64 per unit.
Solution First find the number of units that will be bought by solving the
demand equation p = D(q) for q when p = $64:
64 = 4(25 - g 2) so that q= 3 Disregard-3 .
This says that q0 = 3 units will be bought when the price is p 0 = $64 per unit.
The corresponding consumer’s surplus is
r3
D(q) dq - 64(3) = j °
4(25 - q213
) dq - 192
Figure 4.34 Consumer’s demand = 4(25<7-⅛ 3)∣0 -1 9 2 = 72

curve p = 4(25 —q2) showing con
sumer’s surplus The consumer’s surplus is the shaded area in Figure 4.34.
PROBLEM SET 4 .7
Sketch a representative vertical or horizontal strip and fin d the 21. y = ∣4x - 1∣, y = x 2 - 5, x = 0, x = 4
area o f the given region in Problems 1-6. 22. x-axis, y = x 3 - 2x 2 - x + 2
23. y,-axis, x = y 3 - 3y 2 - 4y + 12
Find the consumer’s surplus for the given demand function
defined by D(q) at the point that corresponds to the sales level q
as given in Problems 24-27.
24. D(⅛) = 3.5 - 0.5<∕
a. ¾ = 1 b. ¾ = 1.5
25. D⅛) = 2.5 - 1.5√
a. ¾ = 1 b. q0 = 0
26. D(q) = 1 0 0 - 8q
a. ¾ = 4 b. q0 = 10
27. D(q) = 150 - 6q
a. √0 = 5 b. q0 = 12
© 28. ■ What Does This Say? When finding the area between
two curves, discuss reasons for deciding between vertical
and horizontal strips.
29. ■ What Does This Say? What is consumer’s surplus?
In Problems 30-33, fin d the consumer’s surplus at the point o f
market equilibrium [the level o f production where supply S(q)
equals demand D(q)].
Demand function Supply function
30. D⅛) = 1 4 - q2 S[q) = 2q 2 + l
31. D[q) = 2 5 - q 2 S(q) = 5q2 + 1
32. D(q) = 32 - 2q 2 S(√) = ∖q 2 + 2q + 5
33. D[q) = 2 1 - q 2 S(q) = ⅛ 2 + ⅛ + 5

34. Find the area of the region that contains the origin and is
bounded by the lines 2y = 11 - x and y = l x + 13 and
the curve y = x 2 - 5.
35. Show that the region defined by the inequalities
χ 2 + y 2 ≤ 8, x ≥ y, and y ≥ 0 has area τr.
36. Find the area of the region bounded by the curve
Sketch the area o f the region bounded by the given equations, V x + V y = 1 and the coordinate axes.
and fin d the area o f each region in Problems 7-23.
7. y = x 2 , y = x , x = - 1 , x = 1 37. Suppose an industrial machine that is x years old gene
rates revenue
8. y = x 3, y = x , x = - 1 , x = 1
9. y = x 2 , y = x 3 Λ(x) = 6,025 - 10x2
10. y = x 2, y = f i x dollars and costs
11. y = x 2 - 1, x = - 1 , x = 2, y = 0
12. y = 4x 2 - 9, x = 3, x = 0, y = 0 C(x) = 4,000 + 15x2
13. y = x 4 - 3x 2, y = 6x 2 dollars to operate and maintain.
14. x = 8 - y 2, x = y 2 a. For how many years is it profitable to use this ma
15. x = 2 - y 2 , x = y chine? [Recall that profit is P(x) = R(x) - C(x)].
16. y = x 2 + 3x - 5, y = - χ 2 + x + 7 b. What are the net earnings generated by the machine
17. y = 2x 3 + x 2 - x - l , y = x 3 + 2x 2 + 5x - 1 during its period of profitability? Interpret this
amount as the area between two curves.
18. y = sinx, y = cosx, x - 0, x = %
38. After t hours on the job, one factory worker is producing
19. y = sin x, y = sin 2x, x = 0, x = π
20. y - ∣x∣, y = x 2 - 6 β 1(t) = 60 - 2(t - 1)2
3 18 Ch. 4 Integration
units per hour, and a second is producing C(x) = 0.2x thousand dollars and that these x people will
bring in R(x) = V 3 x thousand dollars in additional reve
2,(Z) = 50 - 5t nue. How many new people would be hired to maximize
units per hour. If both arrive on the job at 8:00 A.M., how the profit? How much net revenue would the company
many more units (to the nearest unit) will the first worker gain by hiring these people?
have produced by noon than the second worker? Interpret 45. Suppose the demand function for a certain commodity is
your answer as the area between two curves. D(q) = 20 —4g 2, and that the marginal cost is C'{q~) =
39. Suppose that x years from now, one investment plan will 2q + 6, where q is the number of units produced. Find the
be generating profit at the rate of consumer’s surplus at the sales level q0 where profit is
maximized.
P'i (x) = 100 + x 2
dollars per year, whereas a second plan will be generating

profit at the rate of COMPUTATIONAL WINDOW
P'2(x) = 220 + 2x
46. Imagine a cylindrical fuel tank lying on its side (of
dollars per year. length L = 20 ft); the ends are circular with radius b.
a. For how many years will the second plan be more You will soon be asked to compute the amount o f fuel
profitable? in the tank for a given level.
a. Explain why the volume of the tank can be ex
b. How much excess profit will you earn if you invest in
pressed by
the second plan instead of the first for the period of
time in part a? Interpret the excess profit as the area 2
between two curves. V = 2L f f b - y 2 dy
40. Parts for a piece of heavy machinery are sold in units of J-b
1,000. The demand for the parts (in dollars) is given by b. Derive this formula for the volume of fuel at level
p(x) = 110 - x h, where - b ≤ h ≤ b ∖
ch ______
The total cost is given by 2 2
V(h) = 2L fb - j dy
C(x) = x3 —25x 2 + 2x + 30 (dollars) J-b
a. For what value o f x is the profit c. Finally, for b = 4, numerically compute F(A) for
h = —3, —2, . . . , 4. Note: F(0) and f'(4) will serve
P(x) = xp(x) - C(x)
as a check on your work.
maximized?
b. Find the consumer’s surplus with respect to the price
that corresponds to maximum profit. θ 47. The supply function represents the amount of a commodi
41. Repeat Problem 40 with p(x) = 124 - 2 x and C(x) = ty that would be supplied to the market at a given price. If
2x 3 - 59x2 + 4x + 76. the market price is s0 , then those producers who would be
42. Suppose when q units of a commodity are produced, the willing to supply the commodity for a price less than .s0
demand is p(q) = 45 - q 2 dollars per unit, and the mar realize a gain. For instance, if the market price is $ 12 and
ginal cost is the corresponding supply is only $10, then those produc
ers who are willing to supply the commodity at $10 gain
from the fact that the price is actually $ 12. The producer’s
surplus is defined to be the total gain realized by all
a. Find the total revenue and the marginal revenue. producers who are willing to supply the commodity for a
b. Find the value o f q (to the nearest unit) that maximizes price that is less than the market price.
profit.
Suppose the supply function for a certain commodity is
c. Find the consumer’s surplus (to the nearest unit) at the
s{q), where q is the number o f units o f the commodity that
value of q where profit is maximized. Hint'. Use the
will be supplied to the market when the price is 5 dollars
exact value (not the rounded value) of q.
per unit. Show that the producer’s surplus with respect to
43. Repeat Problem 42 with the fixed price ∙s,0 is given by
P(?) = ∣ ( 1 0 - ?) 2 and = ~q 2 + 5 ¾
[⅞ - ∙y ('7)] d Q
I
44. A company plans to hire additional personnel. Suppose it where q0 is the sales level that corresponds to the price s0 ;
is estimated that as x new people are hired, it will cost that is, .s,0 = s(g0).
Chapter 4 Review
IN T H IS R E V IE W Proficiency examination; supplementary problems

Problems 1-18 problems of the proficiency examination test your knowledge
of the concepts of this chapter, and Problems 19-30 test your understanding of
those concepts by asking you to work sample practice problems. The supple
mentary problems present questions relating to the material of this chapter in a
wide variety of settings.
Concept Problems 9. Describe in your own words the process of integration

1. What is the area as the limit of a sum? by substitution.
2. Complete these basic rules for sums: 10. What is a differential equation? How do you solve a
n
separable differential equation?
a ∙ ∑ C = ------------------------------ 11.What is an orthogonal trajectory?
k=∖
12. State the mean value theorem for integrals.
b∙ Σ (¾ + b k) = ----------------------------------- 13. What is the formula for the average value of a continu
k=i ous function?
14. State the second fundamental theorem of calculus.
c. ∑ cak = -----------------------------------
k=∖ 15. State Leibniz’s rule.
16. State the following approximation rules:
d. £ (cak + dbk) = ------------------------------------ a. rectangular approximation
k=∖
b. trapezoidal rule
n
e. Σ a k = _______________________ subtotal rule c. Simpson’s rule
⅛=1 17.How do you find the area between two curves?
for 1 < m < n 18. What is consumer’s surplus and how do you find it?
f. dominance rule:_______________________
n
Practice Problems
g∙ Σ 1 = ------------------------------------ 1 1
λ-=ι 19.Given that f x 4 dx = 5,I f x 2 dx = 3,I and f' dx = 1,
rl Jθ Jθ Jθ
h- ∑ k = ------------------------------ find [x2(2x 2 —3)] dx.
∕c = l Jo
n 20. Find 77 '(x) if Fix) = f t5 Vcos(2∕ + 1) dt.
i∙ ∑ ⅛2 = ------------------------------------- J3
k=∖
Evaluate the definite integral in Problems 21-24.
j∙ ∑ ⅛3 = -------------------------------------
⅛=1 21. f (Vx + x -3z2 ) dx 22. f (2x —6)(x2 - 6x + 2) dx
3. What is a Riemann sum? Ji Jo
4. Define a definite integral.
23. [ π
sl f x d x
λ 2 24. f (2x + l)V 2 x 2 + 2x + 5 dx
5. What conditions are necessary for an integral to repre Jo ' + COSX) J 2
sent an area?
6. How can distance be expressed as an integral? 25.a. Find the area under the curve f(x ) = 3x2 + 2 on
7. Complete these statements summarizing the general [ - l,3 ] .
properties of the definite integral. b. Find the area between the curves y = x 2 and y 3 = x.
ra
a. Definite integral at a point: f(x)dx 26. Find the average value of y = cos2x on the interval
_ Ja
ra [0. f l
b. Opposite of a definite integral: f(x)dx
_ _____________________ Jb 27.An object experiences linear acceleration given by
8. State the fundamental theorem of calculus. α(∕) = 2/ + 1 ft∕s 2
Find the velocity and position of the object, given that it 29. Solve the differential equation = y 2 sin 3x.
starts its motion (at t = 0) at 5 = 4 with initial velocity 2
ft∕s.
30. a. Find the necessary n to estimate the value of
28. When it is x years old, a certain industrial machine
generates revenue at the rate of R '(x) = 1,575 - 5x 2 <∙π∕2
dollars per year and results in costs that accumulate at
the rate o f C'(x) = 1,200 + 10x2 dollars per year. J o
cosx dx
a. For how many years is the use o f the machine to within 0.0005 o f its correct value using the trape
profitable? zoidal rule.
b. What are the total net earnings generated by the
machine during the period of profitability (see part a)? b. Find the necessary n if Simpson’s rule is used.
Γo f1 f io
1. Given that ∕( x ) dx = 3, ∕( x ) dx = - 1, and 21. [3 + 7x 73 - 100x1° l ] dx
J-ι Jo J-ιo
f g(x)dx = 7 find f [3g(x) + 2∕(x)] dx. r π ∕4
J-ι J-ι 22. [sin (4x) + 2 cos (4x)] dx
J--7 Γ∕4
2. Use the definition of definite integral to find fx3
∙ι 23. Find F'(x), where F(x) = Z2 cos4 ∕rft
J
Jχ 2
(3x 2 + 2x - l)dx
o 24. Find the area under ∕( x ) = x^ 2 on [1, 4].
25. Find the area under f(x ) = 2 + x - x 2 on [ - 1, 1].
3. Use the definition of definite integral to find
26. Find the area under∕( x ) = x 4 on [0, 2].
-ι
J o
(4x 3 + 6x 2 + 3) dx 27. Find the area under the curve y = x V x 2 + 5 on [—1, 2].
Find the area o f the regions bounded by the curves and lines in
Find the definite and indefinite integrals in Problems 4-22. I f Problems 28-31.
you do not have a technique for finding a closed (exact) answer, 28. 4y2 = x and 2y = x - 2 29. y = x 3, y = x 4
approximate the integral using numerical integration. 2z3 2
30. x = y , x = y 31. y = V'3 sinx, y = cosx,
r! ∕∙2
4. (5x 4 - 8x 3 + 1) dx 5. 3O(5x - 2)2 dx and the vertical lines
υ '-l x = 0 and x = J
2)2 Γ2 x2dx
6. ,f (x V x + dx 7. 32. a. Find f(f) if ∕" ( t ) = sin4t - cos2t and ∕( f ) =
0 l1 V x 3 + 1 1∙
Λ f) =
f2 r0 dx
8. (x + sinx) 3 dx 9. L V l - 2x b. Find f( x ) if ∕" '( x ) = 2x 3 + x 2 , given that ∕ "(1) = 2,
2 1
∕'( 1 ) = 1, and / ( 1 ) = 0 .
Γ2 dx Γo dx
10. ι 11. ιL 1vz l - 2x Solve the differential equations in Problems 33-40.
∣∣ V 3 x - 1 1 dy _ dy _ cos4x
33. (1 ~ J z)2 34.
dx dx y
1 V x (x 2 + V x + 1) dx
12∙ 1 13. 1∣ (x - 1)2 dx ⅛
dy = ( cosj∆ 2 + ι
35. 36. ∙s
dx ∖ s in x ∕
14. I⅛ ∕r 15. 1 (sin2 x + cos2 x) dx dy _ x - 1 + ,v_- < V - 1) dy x ly 2 + 2
χ- 37. 38. =
dx y y2 dχ y V λ -2 + 1
16. 1x(x + 4 )V x 3 + 6x 2 + 2 dx 17. I x(2x 2 + l) V x 4 + x 2 dy _ x sin x 2 dy ∕χ
39. 40. =
dx y 2 cos y 3 dx ∖y
18. 1 dx 19∙ 1 x V 1 - 5x 2 dx
V x (√ x + 1)2 41. Find the average value o f f ix ) = -s ι - ∕ on the interval
[θ, ⅛ ' cos x
20. 1V s in x - cosx(sinx + cos.x) dx

42. Find the average value o f∕( x ) = sinx
a. on [0, τr] b. on [0, 2ττ]
43. Use the trapezoidal rule with n = 6 to approximate be 0.4x dollars per unit. Suppose the manufacture’s profit
is $520 when the level of production is 16 units. What is
sinx dx. Compare your result with the exact value of the manufacturer’s profit when the level of production is
Jo
25 units?
this integral.
56. A tree has been transplanted and after t years is growing at
44. Estimate f V 1 + x 3 dx using the trapezoidal rule with a rate of
n = 6. J° 1+ (F ⅛
r1 d feet per year. After 2 years it has reached a height of 5 ft.
45. Estimate , x using the trapezoidal rule with
n = 8. J° V1 + x 3 How tall was the tree when it was transplanted?
57. A manufacturer estimates that the marginal revenue of a
3
46. Estimate f V 1 + x dx using Simpson’s rule with n = 6. certain commodity is
Jo
R '(x) = V x(x 3z2 + I)’ 1/2
dx
47. using Simpson’s rule with n = 8. dollars per unit when x units are produced. Assuming no
V 1 + x3
revenue is obtained when x = 0, how much revenue is
48. Use the trapezoidal rule to estimate to within 0.00005 the obtained from producing x = 4 units?
value of the integral 58. Find a function whose tangent has slope xV x 2 + 5 for
each value of x and whose graph passes through the point
x + - dx (2, 10).
59. A particle moves along the x-axis in such a way that after t
'≠ A w i t h a n seconds its acceleration is a(∕) = 12(2z + 1)~372. If it starts
49. Use the trapezoidal rule to approximate
error no greater than 0.005. at rest at x = 3, where will it be 4 seconds later?
50. Use Simpson’s rule to estimate to within 0.00005 the 60. An environmental study of a certain community suggests
value of the integral that t years from now, the level of carbon monoxide in the
air will be changing at the rate of 0 .It + 0.2 parts per
million per year. If the current level of carbon monoxide
x + - dx in the air is 3.4 parts per million, what will the level be 3
years from now?
51. Find the average value of the function defined by
61. A woman, driving on a straight, level road at the constant
speed vθ is forced to apply her brakes to avoid hitting a
f( χ ) = ≡ ⅛
1 2 cow, and the car comes to a stop 3 seconds later and s0 feet
from the point where the brakes were applied. Continuing
on [θ, y]. Hint'. Use the trapezoidal rule with n = 6. on her way, she increases her speed by 20 ft/s to make up
52. Find the area between the curves y - sin x, y = cos x, and time but is again forced to hit the brakes, and this time it
the lines x = 0 and x = 2ττ. takes her 5 seconds and s. feet to come to a full stop.
Assuming that her brakes supplied a constant decelera
53. The brakes of a certain automobile produce a constant tion d each time they were used, find d and determine v0,
deceleration of k m ∕s 2 . The car is traveling at 25 m/s when s0 , and s 1.
the driver is forced to hit the brakes. If it comes to rest at a
62. A company has purchased a machine at time t = 0 (in
point 50 m from the point where the brakes are applied,
years). It is estimated that the machine will earn
what is k?
E(t) = 180 - 0.25t 2 thousand dollars and that in the same
54. A particle moves along the x-axis in such a way that time, it will cost the company C(∕) = t 1 thousand dollars
α(∕) = - 4,v(Z), where s(f) and <r(∕) are its position and to maintain and repair the machine. Finally, it is known
acceleration, respectively, at time t. The particle starts that at time t, the salvage value of the machine is
from rest at 5 = 5.
a. Show that v2 + 4 ∕ 2 = 100.
Hint'. First use the chain rule to show that α(t) = v thousand dollars.
b. What is the velocity when the particle first reaches a. When should the machine be sold?
x = 3? Note: The sign of v is determined by the Hint: Let P(t^) = E(t) - C(f) and show that P(t) = 0
direction in which the particle is moving at the time in when t = 12. Explain why the machine should be sold
question. after T years, where T satisfies
55. A manufacturer estimates marginal revenue to be Γ'2
S(∕) = P(r) dt
100x-ιz 3 dollars per unit when the level of production is x jτ
units. The corresponding marginal cost has been found to Find T.
b. How much total net revenue does the machine gener 72. Solve the system of differential equations
ate in its lifetime?
2 <⅛ + 5 ⅛ = t
63. A company plans to hire additional advertising personnel. 2 dt dt l
Suppose it is estimated that if x new people are hired, they
⅛ + 3⅞- = 7 co st
will bring in additional revenue o f R(x) = V r2x thousand
dt dt
dollars and that the cost of adding these x people will be
C(x) = ∣ x thousand dollars. How many new people
Plint: Solve for and algebraically, and then integrate.
should be hired? How much total net revenue (that is,
revenue minus cost) is gained by hiring these people? 73. P U T N A M E X A M IN A T IO N P ROBLEM The time-varying
64. A study indicates that x months from now the population temperature of a certain body is given by a polynomial of
of a certain town will be increasing at the rate of the form ∕( t ) = ati + bt2 + ct + d, where t is time. Show
10 + 2 V x people per month. By how much will the that the average temperature of the body between 9 A.M.
population of the town increase over the next 9 months? and 3 P.M. can always be found by taking the average of
65. It is estimated that t days from now a farmer’s crop will be the temperatures at two fixed times, ti and / , which are
increasing at rate o f 0.3t2 + 0.6t + 1 bushels per day. By independent of a, b, c, and d. Find / and / to the nearest
how much will the value o f the crop increase during the minute.
next 6 days if the market price remains fixed at $2 per 74. P U T N A M E X A M IN A T IO N P RO BLEM Where on the parabo
bushel? la 4ay = x 2 (a > 0) should a chord be drawn so that it
66. Records indicate that t months after the beginning o f the will be normal to the parabola and cut off a parabolic
year, the price of turkey in local supermarkets was sector of minimum area? That is, find P so that the
shaded area in Figure 4.35 is as small as possible.
P(t) = 0.06Z2 —0.2t + 1.2
dollars per pound. What was the average price of turkey
during the first 6 months o f the year?
67. A retailer expects that x months from now, consumers
will be buying 50 cameras a month at a price of P(x) =
40 + 3 V x dollars per camera. Use the characterization of
the definite integral as the limit of a sum to find the total
revenue the retailer can expect from the sale of the
cameras over the next 9 months.
68. Economists predict that x months from now the demand
for beef will be Z>(x) pounds per months and the price will
be P(x) dollars per pound. Use the characterization o f the
definite integral as the limit of a sum to find an expression
for the total amount that consumers will spend on beef
Figure 4.35 Minimum area problem
this year.
69. If/ is continuous on [a, ⅛], show that
75. P U T N A M E X A M IN A T IO N P ROBLEM If α β . , α are
- ∣ ∕W ∣ ≤ ∕W ≤ I∕W ∣ θn [a, ⅛] real numbers that satisfy
70. If/ is continuous on [a, ⅛], show that
∕( x ) dx ∣∕(x)∣ dx
show that the equation a0 + a l x + a2x 2 + . . . + anx n = 0
has at least one real root. Hint: Use the mean value
71. Show that sinx dx ≤ τr. Hint: See Problem 70. theorem for integrals.
Chapter 4 Review 3 23
GUEST ESSAY
"Kinematics of Jogging” by Ralph Boas
Som e people think that calculus is dull, but it did not seem so three centuries ago,
when it was invented. Then, it produced unexpected results; and, now and then, it
still does. This essay is about such a result.
Y o u have learned about the intermediate value theorem (see Section 1.6), which
tells you, for instance, that if you jog at 8 min per mile, there must be some instant
when your speed is exactly ∣ mi per minute— assuming, as is only natural in a
course in calculus, that your elapsed time is a continuous function of the distance
covered. This principle is very intuitive and was recognized before calculus was
invented: Galileo was aware of it in 1638 and thought that it had been known to
Plato. On the other hand, there is a question with a much less intuitive answer that
was noticed only recently (and, as happens more often than mathematicians like
to admit, by a physicist). Suppose that you average 8 m in∕m i, must you cover
some one continuous mile (such as a “ measured mile” on a highway) in exactly 8
Ralph Boas (1912-1992), Pro
min? The answer is not intuitive at all: It depends on whether or not your total
fessor Emeritus of Mathematics
distance was an integral number of miles. More precisely, if you cover an integral
from the Northwestern Univer
number of miles, then you cover exactly one mile in some 8 min. However, if you
sity, wrote this guest essay. Pro
cover a nonintegral number of miles, there is not necessarily any one continuous
fessor Boas is well known for his
mile that you cover in 8 min.
papers and professional activi
ties. In addition to his work in T o prove this, let x be the distance (in miles) covered at any point during your
real and complex analysis, he trip, and suppose that when you stop you have covered an integral number of n
wrote many expository articles, miles. Let ∕( x ) be the time (in minutes) that it took to cover the first x miles; we
such as this guest essay, about will suppose that f is a continuous function. If you averaged 8 m in∕m i, then
teaching or using mathematics. f(x ) - 8.x = 0 when x = 0 and when x = n. Now suppose that you never did cover
any consecutive mile in 8 min; in mathematical terms,
R x + i ) - A x) ≠ 8
Because
A χ + i ) - A x) ~ 8
is continuous and never 0, it must either always be positive, or else always
negative; let us suppose the former. Write the corresponding facts for x = 0,
1, . . . , n:
Nature herself exhibits to us meas
urable and observable quantities in / (1 ) -/ ( 0 ) > 8
definite mathematical dependence', / (2 )-/ (1 ) > 8
the conception o f a function is sug
gested by all the processes o f nature
where we observe natural phenome
na varying according to distance or
/(«) - /(« - 1) > 8
to time. Nearly all the “ known” If we add these inequalities we obtain
functions have presented them
selves in the attempt to solve geo /(«) - /(0) > 8»
metrical, mechanical, or physical But we started with the assumption that∕(w) = 8« an d ∕(0 ) = 0, so assuming that
problems.
f ( x + 1 ) - f(x) is never 8 leads to a contradiction.
J. T. Mertz, A History o f European
Thought in the Nineteenth Century It is somewhat harder to show that only integral values of n will work. Suppose
(Edinburgh and London, 1903), p. 696. you jog so that your time to cover x miles is
fix ) = k sin 2 q+ + 8x
where n is not an integer and A:is a small number. This is a legitimate assumption,
because J is an increasing function (as a time has to be), if k is small enough. To be
sure of this, we calculate J'(x)— and here we actually have to use some calculus (or
have a calculator that will do it for us). We find
J'(vx )’ = n sin n + 8
If k is small enough {k < then J'(x) > 0. This shows not only that J
increases, but also that
J(x + i ) - J(x)
cannot be eight. Because J(x + 1) - Jr(x) is never negative, if you jog so that your
time is J(x), you will never even cover a whole mile in less than 8 min.
MATHEMATICAL ESSAYS
Use a library or references other than this textbook to research 9. Suppose that the time to cover x miles is given by
the information necessary to answer questions in Problems 1-6.
J(x) = sin 2 + 8x
1. One of the great ideas of calculus is that of derivative.
Write an essay of at least 500 words about some applica
Graph this function on [0, 8].
tion of the derivative that is not discussed in this text.
10. Use calculus to find how small k needs to be in the
2. The concept of the integral is one of the great ideas of expression
calculus. Write an essay of at least 500 words about the
relationship of integration and differentiation as it relates J(x) = k sin 2 + 8x
to the history of calculus.
3. At the beginning of Chapter 2, we wrote of the Fermat- so that J'(x) > 0.
Newton approach to finding the slope of a tangent to a 11. Find a number x, x ≠ 0, so that when it is divided by 2 the
curve at a point. Do some research in the history of display on your calculator shows 0. Write a paper describ
calculus and write a paper on the development of this ing your work as well as the processes or your calculator.
concept. At the conclusion of this paper, discuss why the 12. Calculate V2 using a calculator. Next, repeatedly subtract
concept of derivative is so important. the integer part of the displayed number and multiply the
4. Write a report on Georg Riemann. Include some recent result by 10. Describe the outcome, and then devise a
developments (1984) in the solution of the Riemann method for finding Vz2 using calculus. Write a paper
hypotheses. comparing these answers.
5. Write a report on Fermat’s last theorem. Include some Book Report Eli Maor, a native of Israel, has a long
recent developments (1993) in the solution of this theor standing interest in the relations between mathematics
em. and the arts. Read the fascinating book To Infinity and
6. Write a report on the Bernoulli family. Beyond, A Cultural History o f the Infinite (Boston:
7. As we saw in this chapter, the mathematician Seki Kδwa Birkhauser, 1987), and prepare a book report.
was doing a form of integration at about the same time 14. Make up a word problem involving an application of the
that Newton and Leibniz were inventing the calculus. derivative. Send your problem to:
Write a paper of the history of calculus from the eastern
viewpoint. Math Executive Editor
8. In the guest essay it was assumed that the time to cover x Bradley and Smith Calculus
miles is Prentice-Hall
Englewood Cliffs, NJ 07632
J(x) = k sin 2 — + 8x
The best one submitted will appear in the next edition
Suppose that n = 5. What choices for k seem reasonable? (along with credit to the problem-poser).
Exponential, Logarithmic,
and Inverse Trigonometric
Functions
■ CONTENTS
5.1 Exponential Functions; The
Number e
Exponential functions; Natural PREVIEW
base e; Continuous compound
ing of interest So far, all our work has been with polynomials, rational functions, root
5.2 Inverse functions; Logarithms functions, trigonometric functions, and combinations of these basic
Inverse functions; Criteria for functional types. However, many pure and applied mathematical topics
the existence of an inverse func
tion f^ l ∙, graph of f~ i ∙, The loga require functions of a different kind. We begin this chapter by defining
rithmic function; The natural the natural exponential function ex and then the natural logarithmic
logarithm
function, In x. We shall find that the derivatives of these functions
5.3 Derivatives Involving ex and
In x satisfy
Derivatives of natural logarith
mic functions; Derivatives of S < e ^'> = e ^ and ⅛ (ln x) = ⅜
exponential functions;
Logarithmic differentiation
5.4 Applications Involving Deriva Applications of these functions as they relate to both the integral and the
tives of ex and In x derivative are also considered. The fact that ex is equal to its own
Curve sketching; Optimization; derivative has important theoretical and practical uses. Indeed, it is this
Evaluating limits using loga
rithms and ΓHδpital, s rule feature of ex that causes mathematicians to regard it as “ natural.” We
5.5 Integrals Involving ex and In x also introduce and study the inverse trigonometric functions, which
Integration of ex and In x; have practical applications and also provide useful integration formulas.
Integration involving tanx and
cotx Finally, in Section 5.7 we provide an alternative development in which
5.6 The Inverse Trigonometric the natural logarithm is defined as an integral function with a variable
Functions upper limit.
Definition of inverse trigono
metric functions; Properties of
inverse trigonometric functions;
Derivatives and integrals involv
ing inverse trigonometric func PERSPECTIVE
tions
5.7 An Alternative Approach: The In 1964, an earthquake measuring 8.5 on the Richter scale devastated parts of
Logarithm As an Integral Alaska. The 1989 World Series was interrupted by a 6.5 quake, and newspaper
Definition of the natural loga accounts compared the two events by pointing out that the Alaska quake had
rithm as an integral; Geometric
interpretation of the natural been at least 100 times more powerful. But what kind of scale measures 8.5 as
logarithm; The natural exponen 100 times as large as 6.5? The Richter scale, which is often used to measure the
tial function magnitude of earthquakes, uses a logarithmic scale.
Chapter 5 Review
325
326 Ch. 5 Exponential, Logarithmic, and Inverse Trigonometric Functions
5 .1 EXPONENTIAL FUNCTIONS; THE NUMBER e
IN THIS SECTION Exponential functions, natural base e. continuous

compounding of interest
Linear, quadratic, polynomial, and rational functions are among those func
tions called algebraic functions. Functions that are not algebraic are called
transcendental functions. Trigonometric functions are examples of transcenden
tal functions, and in this section we introduce an additional important
transcendental function, the exponential function and the irrational number e.
We also discuss some graphical properties of exponential functions and an
interesting application dealing with continuous compounding of interest.
■ EXPONENTIAL FUNCTIONS
Exponential Function The function/is an exponential function if

∕( x ) = bx
where b is a positive constant other than 1 and x is any real number.
Recall that if n is a natural number, then

b'> = b - b - b - ■ ■ b
n factors
Furthermore, if b ≠ 0, then
b0 = 1, b " ~ ff, and h ιz" = v ∕
Also, if m and n are any integers, and m/n is a reduced fraction, then
j^m/n — Qj∖ ∕nyn
This means that we know what bx means for rational values of x. However, we
now wish to enlarge the domain of x to include all real numbers. To get a feeling
for what is involved in this problem, let us examine the special case where b = 2.
In Figure 5.1, we have plotted several points with coordinates (r, 2'), where r is a
rational number. We must now attach meaning to bx if x is not a rational number.
Figure 5.1 Graph of y = 2 x for se In order to enlarge the domain of x to include all real numbers, we need the help of
lected rational exponents the following theorem.
THEOREM 5.1 Bracketing theorem for exponents
Suppose b is a real number greater than 1. Then for any real number x, there is a
unique real number bx . Moreover, if p and q are any two rational numbers such
that p < x < q, then
bp < bx < bq
Proof: A formal proof of this theorem is beyond the scope of this course, but
we can outline the needed steps. The proof depends on a property of real numbers
called the completeness property of the real numbers. What the completeness
property says is that each real number can be approximated to any prescribed
Sec. 5.1 Exponential Functions: The Number e 327
degree of accuracy by a rational number. Using mathematical notation this is

equivalent to saying that if x is a real number, then for every positive integer n
there exists a rational number rn such that ∣x - rπ∖ < ~. In other words,
x = lim rn
n→+∞
Then, bx is defined by
bx = lim br"
n→+∞
The bracketing theorem follows from these definitions.
The bracketing theorem gives meaning to expressions such as 2v 3 . Because

1.732 < √3 < 1.733
the bracketing theorem says that
21.732 < 2 √3 < 2 1 ∙733
E X A M P L E 1 Graph of an exponential function
Graph ∕( x ) = 2x .
Solution Begin by plotting points using rational x-values. We see that the graph
has a fairly well defined shape (as shown in Figure 5.2a), but it is riddled with
“ holes” that correspond to those points whose x-coordinates are irrational
numbers.
We complete the graph by connecting the points with a smooth curve as shown
in Figure 5.2b. However, let us take a moment to relate the completed graph to
the definition of 2x for an irrational x-value. Consider the point (x, 2x ) for
x = V z3. Because V3 is irrational, the decimal representation is nontermina
ting and nonrepeating. In other words, V3 is a limit of a sequence of rational
numbers, specifically,
1, 1.7, 1.73, 1.732, 1.7320, 1.73205, 1.732050, 1.7320508, . . .
Then, 2v 3 is the limit of the sequence of numbers
2> 2 1∙7 2 1∙73 2 1∙732 2 1∙7320 2 1-732°5 2 1∙732050 2 1,7320508
Graphically, this means that as the rational numbers 1, 1.7, 1.73, . . . tend
toward V 3 , the points (1, 2), (1.7, 2 l ∙7), (1.73, 2 1∙73), . . . tend toward the
“ hole” in the graph of y = 2Λ that corresponds to x = V 3.
The shape of the graph of y = bx for any b > 1 is essentially the same as that
of y = 2x . The graph of y = bx for a typical base b > 1 is shown in Figure 5.3a,
and by reflecting this graph about the y-axis (see Problem 43) we obtain the graph
of y = b~x . The case where 0 1, and y = bx and
y = b~x are thus the same as y = c~x and y = cx , respectively.
We summarize the basic properties of exponential functions with the follow
ing theorem. Many of these properties were proved in previous courses for bases
a. b> 1 with rational exponents.
THEOREM 5.2 Properties of exponential functions
Let x, y be real numbers, and a and b positive real numbers.
Equality rule If b ≠ 1, then bx = by if and only if x = y.

Inequality rules If x > y and b > 1, then bx > by .
If x > y and 0 1.
b x y
Figure 5.3 Graph of y = bx
Quotient rule by = b° ~
Power rules (bx )y = bxy
a∖x _ ax
(ab)x = ax bx and .b∕ ~ bx
Limits involving exponential functions If b > 1,
lim bx =0 and lim bx = +∞
Λ,→+∞
χ→ -∞
lim b~x = +∞ and lim b~x = 0
x→ -∞ x→ +∞
Proof: The proof of this theorem is beyond the scope of this course. However,
some special cases and parts of the theorem are considered in the problem set. ■
Many parts of this theorem are used in the following example.

EXAM PLE 2 Exponential equations
Solve each of the following exponential equations.
a. 2χ2 ÷3 = 16 b. 2-V3Λ'+1 = 108 c. (√2)*2 =
Solution a. 2χ 2 + 3 =16
2x +3 = 24 Write 16 as 24 so that the equality rule can be used.
x 2 + 3 = 4, or x = ± 1
b. 2 -v3 .r+ι = 1 0 8
2Λ 3 V3 = 3 ∙ 36 Product rule
(2 ∙ 3)v = 36 Power rule; also divide both sides by 3.
6x
= 62
x =2 Identity rule
c. (√2)- 2 = ⅞
2 x ^'~ = 23X 2 Identity rule

y = 3x - 2
- - 6x + 4 = 0
6 ± √ 3 6 - 4(1)(4) ι zξ
A J— V J
You can use a graphing calculator to estimate solutions to equations like those
in Example 2. For example, in part a, you can graph both the left and right
sides of the equation and then look for the intersection point(s), if any.
The (TRACE) with the cursor on the curve y

= 2 x +3 gives the following values:
« = -1 .0 2 1 2 7 ? Y=lfi.4BH13 A root is between
X≡^.97B723H Y=15.539B37 ÷- these X-values.
and
M=.9?B723H Y=15.E39B37 <- A root is between
X=1.02127fifi V=lfi.H B H 13i <- these x-values.
V z il6
X n in = - 2 Yroin= - 5
Xnax=2 Vnax=2θ
XS G 1=1 YS G 1=1
The intermediate value theorem tells us the desired root is between the traced
values, so it looks like the roots are x = - 1 and x = 1. These values could be
checked by substitution.
For part b we have:
(TRACE) with the cursor on the curve y = 108; we
find X = 1.9574468 and X=2.0425532. From
the interm ediate value, we see the root is be
tween 1.9574468 and 2.0425532; you can check
to verify that x = 2 is the root.
For part c, we could be very clever in setting
scale to graph the given equations, but instead
this example illustrates that even though you
VIB2Λ X*3^<X+1)
V2≡10S
may have a graphing calculator there are many
X n in = - 4 l√r∣
problems that still require mathematical analy Xnax=4 i n = - 5Θ
Ynax=150
sis. Remember, a graphing calculator is a valu XΞ G 1=1 YS G 1=10
able tool, but it will not do your work for you.
■ NATURAL BASE e
In elementary algebra, exponential bases are usually 2 or 10, but in calculus, we

shall find it more useful to consider exponential functions with base e, where e is
the following num ber defined as a limit.
The Number e The limit
lim
∕7→+∞
n ( 1 + l ∕n ) ^ n
1 2 exists and is denoted by e. The number e is called the natural exponential base.
2 2 .2 5
3 2 . 37037037
4 2 .4 4 1 4 0 6 2 5 The letter e was chosen in honor of the great Swiss mathematician Leonhard
5 2 .4 8 8 3 2
6 2 .5 2 1 6 2 6 3 7 2
Euler (1707-1783), who investigated this limit and explored a number of
7 2 . 546499697 applications in which this limit plays a useful role. At first glance, there may seem
8 2 .5 6 5 7 8 4 5 1 4 to be very little about the number e that is “natural,” but later in this section and
9 2 .5 8 1 1 7 4 7 9 2
10 2 . 59374246
throughout the rest of this chapter, we shall point out features that make e an
extremely interesting and useful exponential base.
20 2 . 6 53 2 9 7 7 0 5 First, let us get a better idea of the nature of this number we have called e. It is
30 2 .6 7 4 3 1 8 7 7 6
40 2 .6 8 5 0 6 3 8 3 8
by no means obvious that this limit even exists (we shall show that it does in
50 2 .6 9 1 5 8 8 0 2 9 Chapter 8), but assuming that it does, it may seem to have the value 1. After all.
60 2 .6 9 5 9 7 0 1 3 9
70 2 .6 9 9 1 1 6 3 7 1
lim 1 + —) = 1 and 1” = 1.
80 2 .7 0 1 4 8 4 9 4 1 ∕ ∕ →+∞ n
90 2 .7 0 3 3 3 2 4 6 1
100 2 .7 0 4 8 1 3 8 2 9 However, the limit process does not work this way, and it turns out that
200 2 .7 1 1 5 1 7 1 2 3 e ≈ 2.718281828459045
300 2 .7 1 3 7 6 5 1 5 8
400 2 .7 1 4 8 9 1 7 4 4 The first four digits of this limit are substantiated by the table shown in the margin
500 2 .7 1 5 5 6 8 5 2 1 (which was done on a computer). It can be shown that the number e is irrational,
600 2 .7 1 6 0 2 0 0 4 9
700 2 . 716342739 so it does not have a terminating or repeating decimal representation.
800 2 .7 1 6 5 8 4 8 4 7 The function∕(x ) = ex is called the natural exponential function. This function
900 2 .7 1 6 7 7 3 2 0 8 obeys all the basic rules of Theorem 5.2 for exponential functions with base
1000 2 .7 1 6 9 2 3 9 3 2
2000 2 .7 1 7 6 0 2 5 6 9 b > 1. Because many of the applications we consider in calculus will have
3000 2 .7 1 7 8 2 8 9 2 complicated exponents, such as
4000 2 .7 1 7 9 4 2 1 2 1
5000 2 .7 1 8 0 1 0 0 5 ^3X2-2 sinx + 8
6000 2 .7 1 8 0 5 5 3 5
7 0 00 2 .7 1 8 0 8 7 6 9 4 we sometimes streamline this form by writing
8000 2 .7 1 8 1 1 1 9 5 5
9000 2 .7 1 8 1 3 0 8 2 5 exp(3x2 - 2 sinx + 8)
10000 2 .7 1 8 1 4 5 9 2 7
Exp Notation If∕(x ) is an expression in x, the notation

exp(∕(x))
means e ^ x ∖
■ CONTINUOUS COMPOUNDING OF INTEREST
One reason e is called the “natural” exponential base is that many natural growth
phenomena can be described in terms of ex . As an illustration of this fact, we close
this section by showing how ex can be used to describe the accounting procedure
called continuous compounding o f interest.
If a sum of money, called the present value or principal, is denoted by P and
invested at an annual rate of r for t years, then the future value is denoted by A and
is found by
A =P +I
where I denotes the amount of interest. Interest is an amount of money paid for
the use of another’s money. Simple interest is found by multiplication: I = Prt. For
Computational Window Q Q example, $1,000 invested for 3 years at a 15% simple annual interest rate
generates / = $1,000(0.15)(3) = $450, so the future value in 3 years is A =
You must be careful when $1,000 + $450 = $1,450.
using software to reach conclu Most businesses, however, pay interest on the interest as well as on the
sions in calculus. For example, principal, and when this is done, it is called compound interest. For example,
on most software or calcula $1,000 invested at 15% annual interest compounded annually for 3 years can be
tors, if you input
found as follows:
/ , ∖ ιo 2 °
(ι 1 + 102≡) F irst year: A = P + I
= P ■1 + P r I = Prt and t = 1
the output is 1 (which, of
course, is incorrect). We know = P ( l + r) For this example, A = $1,000(1 + 0.15)
(from calculus) that = $1,150
lim (l + ∣) = e > 1. S eco n d year: A = P ( l + r) + I The total amount from the first year
n
β→ 0 ' ' becomes the principal for the second year
A = P(l + r) ∙ 1 + P(l + r ) ∙r
= P(l + r)(l + r)
= P ( l + r)- For this example, A = $1,000(1 + 0.15)2
= $1,322.50
T hird year: A = P ( l + r) 2 + P ( l + r) 2 ∙ r
= P(1 + r)2(l + r)
= P ( l + r) 3 For this example, A = $1,000(1 + 0.15)3
≈ $1,520.88
Notice that with simple interest the amount in 3 years is $ 1,450, as compared with
$1,520.88 when interest is compounded annually.
Compound Interest Future If a principal of P dollars is invested at an interest rate of i per period for a total
Value Formula of N periods, then the future amount A is given by the formula
A =P(1 + ifi'
Compound interest is usually stated in terms of an annual interest rate r and a

given number of years t. The frequency of compounding (that is, the number of
compoundings per year) is denoted by n. Therefore, i = 22 and N = nt, in
the formula for A .
The first two graphs in Figure 5.4 show how an amount of money in an
account over a one-year period of time grows, first with quarterly compounding
and then with monthly compounding. Notice that these are “step” graphs, with
Figure 5.4 The growth of an account over a one-year period with different compounding
frequencies
With continuous compounding, we compound interest not quarterly, or

monthly, or daily, or even every second, but instantaneously, so that the future
amount of money A in the account grows continuously, as shown in Figure 5.4c.
To compute A , we take the limit as the number of compounding periods n
increases without bound. That is
nt
A = lim P( 1 + r ) To evaluate this limit, let
n→ +∞ ∖ A2∕ m = " so that ⅛= ∙‰and
nt = mrt. Also, as n → +∞
we see m → +∞.
Properties o f exponents
Scalar rule
= Pe r,
This is summarized in the following box.
Continuous Compounding of If P dollars are invested at an annual interest rate r compounded continuously,
Interest then the future value after t years is
A = Pe rt
EXAM PLE 3 Compounding of interest
$12,000 is invested for 5 years at 18%; find the future value at the end of 5
years if interest is compounded
a. monthly b. continuously
Solution P = $12,000; t = 5; and r = 18% = 0.18 are given.

12∙5
a. « = 12; A = P(1 + Z)ΛΓ = $12,000(1 + 2 ^ ) ≈ $29,318.64
b. A = $12,000e°18(5) ≈ $29,515.24
PROBLEM SET 5 .1
θ Sketch the graph o f the functions in Problems 1-4. 13. 5,θθθ(l + fl ⅛ 5 ) 12' 5> 14. 145,000(1 + a ⅛⅜i ) 365(5)
1. y = 3x 15. 2,589e0'45<6> 16. 850,000e-°04<'°)

3. y = - e ~ x Find all real numbers x that satisfy the equations in Problems
17-24.
Evaluate the expressions (with calculator accuracy) in Prob- 17. 3*2^X = 9 18. 4χ2+x = 16
lems 5-16. 19. 2*5X +2 == 25,000 20. 3λ 4x + ιz2 = 3,476
5. 322z5 + 9 3z2 6. (1 + 4 3 ∕2 ) - l∕2 21. C√z2)x+1 0 = 2χ2 22. (√5)X +2 = 5χ2
7. 2.5 - ιz 3 8. (1.5)- 2(1.5)3∙7 2 4 ∙⅛ = 1
23. 2*+3 = I
9. (e1∙3)2 10. √ F τ τ e
e2 θ 25. In the definition of the exponential function f( x ) = b x , we

11. e3e 2-3 12.
e + CL 3 require that b be a positive constant other than 1.
a. What happens if b = 1? Draw the graph of f( x ) = bx , years. What is the monthly payment? Hint. See Problem
where b = 1. Is this an algebraic or a transcendental 34 for the appropriate formula.
function? 36. A cool drink is removed from a refrigerator and is placed
b. What happens if b = 0? Draw the graph of ∕( x ) = bx , in a room where the temperature is 70 oF. According to a
where b = 0. Is this an algebraic or a transcendental result in physics known as Newton's law o f cooling, the
function? temperature of the drink in t minutes will be
26. T HINK T ANK P ROBLEM In the definition of the exponen
tial function f(x ) = bx , we require that b be a positive F(0 = 70 - Ae~kt
constant. What happens if b < 0— for example, if
b = -2 ? For what values of x is f defined? Describe the where A and k are positive constants. Suppose the temper
graph o f / i n this case. ature of the drink was 35 0F when it left the refrigerator,
27. Let E(x) = 2JC2^2∖ and 30 minutes later, it was 50 °F (that is, F(θ) = 35 and
a. Choose a number of values of x and plot the corre F(30) = 50).
sponding points (x, E(x)) on a coordinate plane. Pass a a. Find A and e^ 30*r.
smooth curve through the points to obtain a rough b. What will the temperature (to the nearest degree) of
sketch of the graph of E. the drink be after one hour?
b. Where does the graph cross the y-axis? What happens c. What would you expect to happen to the temperature
to the graph as x → +∞ and as x → -∞7 as t → +∞1
c. Based on your graph, what would you say is the 37. Newton’s law of cooling states that an object at tempera
smallest value of £ ? ture B surrounded by air temperature A will cool to a
28. If $3,600 is invested at 15% compounded daily, how temperature T after t minutes according to the formula
much money will there be in 7 years? (Use a 365-day year;
this is known as exact interest). T = A + ( B - A)e~kt
29. If $9,400 is invested at 14% compounded daily, how
much money will there be in 6 months? (Use a 360-day where k is a constant depending on the item being cooled.
year; this is known as ordinary interest). If you draw a tub of 120 oF water for a bath and let it stand
30. A certain bank pays 6% interest compounded continuous in a 75 0F room, what is the temperature of the water (to
ly. How much will a $1,000 investment be worth after 10 the nearest degree) after 30 min if k = 0.01?
years? 38. Biologists estimate that under ideal conditions, the popu
31. How much money must be invested at 5% interest com lation N(f) of a bacterial colony at time t (minutes) is given
pounded continuously in order to have $3,000 in 4 years? by
32. A bank pays 6% interest compounded monthly, and a N(f) = N 02 k,
savings institution pays 5.9% compounded continuously.
Over a one-year period, which account pays more inter where N o and k are positive constants. Suppose the colony
est? What about a five-year period? contains 2,000 bacteria at time t = 0, and 10 min later it
33. A manufacturer of car batteries estimates that p percent contains 5,000 bacteria. What is the population after 20
of the batteries will work for at least t months, where min? After one hour?
p(f) = 100e-°° 3' θ 39. Show that b"'b" = b m+ " for b > 0 and all positive integers
a. What percent of the batteries can be expected to last at m and n.
least 40 months? hm = b m n
40. Show that for b > 0 and all positive integers
b. What percent can be expected to fail before 50
months? m and n.
c. What percent can be expected to fail between the 40th 41. Show that (⅛m )'' = b mn for b > 0 and all positive integers
and 50th months? m and n.
34. If P dollars are borrowed for N months compounded 42. Show that (f∕b) m = f' rb"' for b > 0 and all positive inte
monthly at an annual interest rate of r, then the monthly gers m and n.
payment is found by the formula 43. If (x, y) is a point on the graph of y = f(x), note that
( - χ , yj is on the graph of y = ∕( - x ) . Use this fact to prove
that the graph of y = f ( - x ) can be obtained by reflecting
the graph of y = ∕(x ) about the y-axis.
z>Λ _ ι
Use this formula to determine the monthly car payment 44. Show that hm -—-,— = 1.
for a new car costing $12,487 with a down payment of A→O n
$2,487. The car is financed for 4 years at 12%. 45. Prove that lim ex = +∞ by showing that for any N > 0,
35. A home loan is made for $110,000 at 6% interest for 30 there is a number d such that ex > N whenever x > d.
5 .2 INVERSE FUNCTIONS; LOGARITHMS
IN THIS SECTION Inverse functions, criteria for the existence of an

inverse function f ~ l , graph of ∕ ~ 1, the logarithmic function, the natural
logarithm
Intuitively, an inverse function f ~ l “reverses” the effect of a given function f, so
/machine that both the composite functions fo f ~ l and f~ x o/have the effect of leaving x
unchanged; that is,
The inverse function ∕ ^ 1 reverses
the effect of f. This relationship ( f° f~ ') M = x and (f~ l o ∕)(x ) = x
can be illustrated by function
“machines.” This idea of inverse functions is exploited in this section to define a function
that is the inverse of the exponential function, namely, the logarithmic
function.
■ INVERSE FUNCTIONS
0 The symbol f 1 means the For a given function/ we write y0 = f(x 0) to indicate that /m aps the number X Q in
inverse of f and does not mean its domain into the corresponding number y0 in the range. If/ has an inverse∕ - 1 ,
1// 0 it is the function that reverses the effect of / i n the sense that
1(F
f 0) =⅞
For example, if
∕(x ) = 2x - 3, then /(0) = -3 , /(1) = -1 , /(2) = 1
and the inverse ∕ ^ l reverses / so that
/(0) = - 3 ∕ ^ 1( - 3 ) = 0 that i s ,∕- '[ ∕( 0 ) ] = 0
∕ - 1( - i ) = 1 that is,∕ - 1 [∕(l)] = 1
∕( i ) = - ι
/(2) = 1 ∕ - , (l) = 2 that i s , Γ 1[∕(2)] = 2
In the case where the inverse of a function is itself a function, we have the
following definition.
Inverse Function L e t/b e a function with domain D and range R. Then the function / 1 with
domain R and range D is the inverse of / if
/ , [∕W1 = * for all x in D
∕ [ ∕ ^ 1ω i = τ for all y in R
What this says: Suppose we consider a function defined by a set of ordered
pairs y = ∕(x). The image of x is y, as shown in Figure 5.5. If y is a member of
the domain of the function g = ∕ - 1 , then g(y) = x. This means that / matches
each element of x to exactly one y, and g matches those same elements of y
Figure 5.5 Inverse functions back to the original values of x. If you think of a function/as a set of ordered
/a n d # pairs (x, y), the inverse of / is the set of ordered pairs with the components
(y,χ)-
The language of this definition suggests that there is only one inverse function of/
Indeed, it can be shown (see Problem 61) that if / h a s an inverse, then the inverse
is unique.
Sec. 5.2 Inverse Functions; Logarithms 335
E X A M P L E 1 Inverse of a given function defined as a set of ordered pairs
L e t/ = {(0, 3), (1, 5), (3, 9), (5, 13)}; fin d / - 1 , if it exists.
Solution The inverse simply reverses the ordered pairs:
f - ' = {(3, 0),(5, 1),(9, 3),(13, 5)} —
EXAM PLE 2 Inverse of a given function defined by an equation
Let ∕( x ) = 2x - 3; find ∕ -1 , if it exists.

Solution To find ∕ -1 , let y = ∕( x ) and interchange the x and y variables; then
solve for y.
Given function: y = 2x - 3 Inverse: x = 2y - 3
2y = x + 3
y = ⅛(x + 3)
Thus, we represent the inverse function a s ∕ - 1 (x) = ∣(x + 3). To verify that
these functions are inverses of each other, we note that
∕ [ ∕ - 1 ω i = ∕[⅛ (* + 3 )] = 2⅛(x + 3 ) ] - 3 = x + 3 - 3 = x
and
∕ ^ 1[∕(∙v)] = ∕ - 1 (2x - 3) = ⅛[(2x - 3) + 3] = ∣(2x) = x
for all x. M M
■ CRITERIA FOR THE EXISTENCE OF AN INVERSE

FUNCTION f ~ l
The inverse of a function may not exist. For example,

∕= { ( 0 , 0),(1, 1 ),(-1 , 1), (2, 4 ) ,( - 2 , 4)}
and
g(x) = x 2
do not have inverses because if we attempt to find the inverses, we obtain relations
that are not functions. In the first case we find
Possible inverse o f/: {(0, 0), (1, 1), (1, -1 ) , (4, 2), (4, -2)}
This is not a function because not every member of the domain is associated with
a single member in the range: (1, 1) and (1, -2 ), for example.
In the second case, if we interchange the x and y in the equation for the
function g where y = x 2 and then solve for y we find:
x = y 2 or y = ± V x for x ≥ 0
But this is not a function of x, because for any positive value of x, there are two
corresponding values of y, namely, V x and - V x .
A function f will have an inverse/ -1 on the interval / when there is exactly one
number in the domain associated with each number in the range. That i s ,/ -1
b. A function that does not have exists if∕( x 1) a n d ∕(x ,) are equal only when x 1 = x 2 . A function with this property
an inverse is said to be one-to-one. This is equivalent to the graphical criterion shown in
Figure 5.6 Horizontal line test Figure 5.6.
Horizontal Line Test A function /h a s an inverse if and only if each horizontal line intersects the
graph of y = ∕(x ) in at most one point.
Remember that a function is strictly monotonic on some interval if it is either

strictly increasing or strictly decreasing on that interval (Section 3.3). Note that a
strictly monotonic function must be one-to-one and, therefore, must have an
inverse. For example, if / is strictly increasing on the interval I, we know that
x l > x 2 implies ∕( x 1) > ∕( x 2)
so there is no way to have ∕( x 1) = ∕( x 2) unless x l = x 2 . In fact, this idea is
formalized in the following theorem.
THEOREM 5.3 Existence of an inverse
Let / be a function that is continuous and strictly monotonic on an interval I.

Then∕ - ' exists and is monotonic on / (increasing if/ is increasing and decreasing
i f / is decreasing).
Proof: We have already commented on why ∕^^ l exists. To show that ∕ ^ 1 is
strictly increasing whenever / is increasing, let y 1 and y 2 be numbers in the range
o f/w ith y , > y 1. We shall show that∕ - l (w) > ∕ - l (y∣ )∙ Because y 1, y 2 are in the
range of / there exist numbers x 1, x 2 in the domain I such that y l = ∕( x 1) and
y 2 = ∕( x 2). Because y 2 > y l , it follows that∕( x 2) > ∕( x 1), and because/is strictly
increasing, we must have x 2 > x 1. Thus, ∕ - 1 (y2) > ∕ ^ 1(y 1), and ∕ ~ l is strictly
increasing. Similarly, if /is strictly decreasing, then so i s / - 1. (The details are left
for the reader.) ■
■ GRAPH OF f ~ 1
The graphs of /a n d its inverse ∕ - ' are closely related. In particular, if (a, b) is a
point on the graph of/ then b = ∕(α ) and a = f~∖ b), so ( / a) is on the graph of
∕ ^ 1. It can be shown that (a, b) and (b, a) are reflections of one another in the line
y = x. (See Figure 5.7, as well as Problem 60.) These observations yield the
following procedure for sketching the graph of an inverse function.
Procedure for Obtaining the If/ 1 exists, its graph may be obtained by reflecting the graph o f/in the line
Graph of ∕ - * y = x.
Figure 5.7 The graphs o f / a n d / 1 are reflections in the line y = x.

EXAM PLE 3 Graphing an inverse
Explain why the function f( x ) = x 3 + 2 has an inverse. Sketch the graph of

r 1∙
Solution The derivative of f is ∕'( x ) = 3x2 . Because f '( x ) > 0 for all x ≠ 0, it
follows that f is strictly increasing, and according to Theorem 5.3, ∕ ^ 1 exists.
The graph off is shown in Figure 5.8, and in the same figure, we obtain the
graph o f / " 1 by reflecting the graph of f in the line y = x.
In general, if the graph off has a nonhorizontal tangent line at a point P(a, b),
we expect the graph of the inverse function ∕ ^ 1 to have a nonvertical tangent line
at the reflected point Q(b, a). Thus, if/is differentiable at x 0 and ∕'( x 0) ≠ 0, then
∕ " 1 is differentiable aty 0 = ∕( x 0). This observation is formalized in Theorem 5.4.
THEOREM 5.4 Continuity and differentiability of inverse functions

Figure 5.8 The graphs of
f(x) = x 3 + 2 and its inverse Let / be a one-to-one function. (This means that it possesses an inverse.)
∕ ^ 1(*)
1. I f /is continuous on its domain, then ∕ ^ 1 is continuous on its domain.
2. I f /is differentiable at c and ∕'( c ) ≠ 0, then∕ ^ 1 is differentiable at ∕(c).
Historical Note
Pr oof: The reasonableness of this theorem follows from Figure 5.8. A proof is
given in Appendix B. ■
We can use Theorem 5.4 and the chain rule to provide a formula for the
derivative of∕ " 1(x) in terms of the derivative ∕'(x ).
THEOREM 5.5 Derivative of an inverse function
Suppose/is strictly monotonic and differentiable for all x in its domain. Then
σ ^' y w = ∕ ' [ ∕ - 1(χ)]
Proof: Let g(x) = ∕ ~ 1(x). From Theorem 5.4 we know that g is differentiable.
J OHN N APIER (1550-1617) By definition of inverse functions,
Napier was the Isaac Asimov of ∕[ f W ] = χ
his day, having envisioned the
tank, the machine gun, and the f [T(x )] = 1 Differentiate both sides with respect to x.
submarine. He also predicted
that the end of the world would ∕ ' [<?(*)] [<?(*)] = 1 Chain rule
occur between 1688 and 1700.
He is best known today as the 4 ^ [⅛,(x )] = f r ∖ Because f is monotonic, f ' ≠ 0, so
inventor of logarithms, which J LW(-λr)J ∕'[g(x)] ≠ 0, and we can divide both
until the advent of the calcula
sides by f ' [g(x)]. ■
tor were used extensively with
complicated calculations. Today
we use the logarithmic function EH What this says: Geometrically, the derivatives of two functions that are
to model many situations. Na inverses tell us that graphs of inverse functions have reciprocal slopes at the
pier considered mathematics points (a, b) and (b. a). Also notice that in the Leibniz notation, the formula in
only a diversion from his theo Theorem 5.5 can be written
logical work, on which he felt dx _ 1
his reputation would solely de dy cly
pend. dx
EXAMPLE 4 Derivative of an inverse function
Let ∕( x ) = ∖zx. Show that ∕^ '( x ) exists and use the derivative of an inverse
function theorem to evaluate the derivative o f∕^^1(x) at x = 4.
Solution Because ∕'( x ) = ⅜x-2z3 > 0 for all x ≠ 0, it follows that ∕~ 1 exists.
Because ./(64) = 4, we have ∕ ^ 1(4) = 64 and ∕'(6 4 ) = ∣(64)-2z3 = ⅛. We now
use the derivative of an inverse function theorem:
( Γ 1)'(4) i(4 )] /'(64) 1 = 48

n r
48
To check this result, note that because f(x ) = V x , we have f 1(x) = x 3. Thus,
( Γ 1)'(x) = 3x 2 and (∕->)'(4) = 3(4)2 = 48
as we found by applying the formula for the derivative of an inverse function

theorem. --τ
■ THE LOGARITHMIC FUNCTION
In Section 5.1, we observed that if b > 0 and b ≠ 1, the exponential function

∕( x ) = bx is continuous and monotonic (increasing if b > 1 and decreasing if
0 < b < 1). Therefore, according to Theorem 5.3 (the existence of an inverse),
∕( x ) = bx has an inverse that is also continuous and monotonic, and we shall refer
to this inverse function as the logarithm o f x to the base b. Here is a definition,
along with some notation.
Logarithmic Function If b > 0 and b ≠ 1, the logarithm of x to the base b is the function y = logi x
that satisfies by = x; that is
y = logi x means by = x
What this says: It is useful to think of a logarithm as an exponent. That is,
consider the following sequence of interpretations:
y = log**
y is the logarithm to the base b of x.
y is the exponent on a base b that yields x.
by = x
Notice that y —log6x is defined only for x > 0 because by > 0 for all y. We have
sketched the graph of y = log6x for b > 1 in Figure 5.9 by reflecting the graph of
y = bx in the line y = x. Notice that because y = bx is a continuous, increasing
function that satisfies bx > 0 for all x, logzx must also be continuous and
Figure 5.9 The graph of y = loghx increasing, and its graph must lie entirely to the right of the y-axis. Also, because
for b > 1 b0 = 1 and b l = b, we have logfi l = 0 and logi ⅛= 1
so (1, 0) and (b, 1) lie on the logarithmic curve. Theorem 5.6 lists several general
properties of logarithms.
T H E O R E M 5.6 Basic properties of logarithmic functions
If b > 0 and b ≠ 1, then
Equality rule If x = y , then logfix = logz,p

Inequality rules If x > y and b > 1, then log/;x > logz,y
If x > y and 0 0 Quotient rule l o ‰(y) = l o ≡⅛x ~ 1°g⅛T
and y > 0. 0
Power rule logz,xT = P ⅛ χ fθ r any real number p
Proof: Each part of this theorem can be derived by using the definition of
logarithm in conjunction with a suitable property of exponentials. ■
We will also need the following properties about limits involving logarithms.
Limits of Logarithms For b > 1,

J⅛ + 1°g ∕T = -∞
lim log,x = +∞
A'→+∞ D r η
0 Don’t forget this result. 0 lim log,x = log, lim x
x→c ° ° Lx→c J
What this says: A limit of a log is the log of a limit.
EXAM PLE 5 Evaluate a logarithmic expression
Evaluate log2(∣) + log9 128.
Solution You can evaluate this on your calculator, or you can use Theorem 5.6
log2 (∣) + log2 128 = log,(2)~3 + log,2 7
= log2 [2 ~3(27)] = log224 = 4 log22 = 4(1) = 4

What we have shown here is much longer than what you would do in your
work, which would probably be shortened as follows:
log2(∣) + log9 128 = log2 16 Because (∣)(128) = 16
= 4 Because 4 is the exponent on

2 that yields 16 ∙≡≡
EXAM PLE 6 Logarithmic equation
Solve the equation log3(2x + 1 ) - 2 log3(x - 3) = 2.

Solution Use Theorem 5.6, remembering that 2x + 1 > 0 and x - 3 > 0.
log3(2x + 1 ) - 2 log 3(x - 3) = 2

log3(2x + 1 ) - log3(x - 3)2 = 2
1θg ⅛ ⅛ =2
3
□2 — 2x + 1
(x - 3) 2
9(x - 3)2 = 2x + 1
9x 2 - 56x + 80 = 0
(x - 4)(9x - 20) = 0
x = 4 ,f
Notice that x - 3 < 0 if x = so the given logarithmic equation has only

x = 4 as a solution. Do not forget that we cannot take the logarithm of a
negative number. κβ
EXAM PLE 7 Exponential equation with base 10
Solve 105γ+3 = 195.

Solution 10 3γ+3 = 195
5x + 3 = log 10 l 95 Remember that 5x + 3 is the exponent
on a base 10 that yields 195.
5x = log 10 195 - 3
log 10 195 - 3
5
≈ —0.141 993 077 7 By calculator
■ THE NATURAL LOGARITHM
There are two frequently used bases for logarithms. If the base 10 is used, then the
logarithm is called a common logarithm; and if the base e is used, the logarithm is
called a natural logarithm. We use a special notation for each of these logarithms.
Common Logarithm The common logarithm, log 10x, is denoted by log x.

Natural Logarithm
(In x is pronounced “ell n x ” The natural logarithm, logex, is denoted by In x.
or “lawn x ”)
EXAM PLE 8 Solving an exponential equation with base e
Solve I = e 0 ∙000425 '.
Solution 1 — z , - o . 000425/
c
2
—0.000425/ = In 0.5 Remember that —0.000425/ is the exponent
on base e that yields ⅛ = 0.5.
, ° - 0 ln0 0 0 4 2 5 ~ 1 ∙6 3 0 934542
If a logarithmic equation has a base other than base 10 or e, you can solve it using
the equality rule of Theorem 5.6. The following theorem provides a useful formula
for converting from one base to another.
T H E O R E M 5.7 Change-of-base theorem
1°g⅛*
1
log
6α X = 1——
log6a
Remember, the definition o f logarithms
requires that b > 0, b ≠ 1.
Proof: Let y = logtfx
α' Definition o f logarithm
l o g χ = logftx Equality rule o f logarithms
y ∖ogba = Iogz,x Power rule o f logarithms
logzΛ
'1 log⅛α Divide both sides by logba.
LJ IVhat this says: To change from base a to another (possibly more familiar)
base b, you simply change the base on the given logarithm from a to b and then
divide by the logarithm to the base b of the old base a.
EXAM PLE 9 Solving an exponential equation using the change-of-base

theorem
Solve 63x+2 = 200.

Solution 6 3x+2 = 200
3x + 2 = log6200
3x = log6200 - 2
log,200 —2 ∣n 9∩∩
x = — ‰ ------- To evaluate log6200, use log6200 =
≈ 0.319 015 741 7
Because we most often use base e, it is worthwhile to state some useful

properties of natural logarithms.
T H E O R E M 5.8 Basic properties of the natural logarithm
a. In 1 = 0
b. In e = 1
c. ein x = x for all x > 0
d. In ey = y for all y
e. logzx = for any b > 0 (Z? ≠ 1)
f. bx = ex l n b for any b > 0 (b ≠ 1)
Proof: a. In 1 = log4,l = 0 Definition o f logarithm e'l = 1

b. In e = logt β = 1 Definition o f logarithm e' = e
c. eln γ = ey Let y = In x.
- x Definition o f logarithm-, i f y = In x, then ey = x.
y
d. In e = In x Let x = ey .
Definition o f logarithm: i f x = ey , then In x = y.
e ∙ 10¾ ∕ = ⅛3 Change-of-base theorem
f. bx = y Let y = b x .
= e'n y Property c o f this theorem ∣ j
= e-v ln i I f y = bx , then x = logz,y =
so that In y = x In b.
If you are using a software package such as Mathematica, Derive, or Maple,

be careful about the notation. Some versions do not distinguish between log x
(common logarithm) and In x (natural logarithm). All logarithms on these
versions are assumed to be natural logarithms, so evaluating a common
logarithm requires the change-of-base theorem.
We close this section with an application in which logarithms play a useful

role.
EXAMPLE 10 Exponential growth
A biological colony grows in such a way that at time t, the population is

P{t) = P,,ek '
where Po is the initial population and k is a positive constant. Suppose the
colony begins with 5,000 individuals and contains a population of 7,000 after
20 min. Find k and determine the population after 30 min.
Solution Because Po = 5,000, the population after t minutes will be
P(t) = 5,000efa
In particular, because the population is 7,000 after 20 min,
P(20) = 5,000CΛ <2°)

7,000 = 5,000e≡'
1 = e ,20A'
5 e
20k = ln(j)
k = ⅛ l n (i)
≈ 0.0168236
Finally, to determine the population after 30 min, substitute this value for k to
find:
P(30) = 5,000e3°fc
≈ 8,282.51 17
The expected population is approximately 8,000.
PROBLEM SET 5 .2
θ Determine which pairs o ffunctions defined by the equations in θ 35. Let ∕( x ) = sinx for 0 ≤ x ≤ J . Use Theorem 5.5 (the
Problems 1-6 are inverses o f each other. derivative of an inverse function theorem) to find the
derivative of the inverse function.
1. ∕( x ) = 5x + 3; g(x) =
2∙ ∕W = j x + 2; g(x) = ∣ x + 3 36. If logftl,296 = 4, what is θ⅛)3z27
3. ∕( x ) = ∣ x + 4; g(x) = ¾x + 3 37. If logv∕ 5 106 = 2, what is V⅛ - 25?

4. ∕( x ) = x ≠ 0; g(x) = 1 ’ x ≠ 0
Sketch the graph o f f in Problems 38-44 and then use the
5. ∕( x ) = x 2, x < 0; g(x) = V x , x > 0 horizontal line test to determine whether or not f has an
6. ∕( x ) = x 2 , x ≥ 0; g(x) = V x , x ≥ 0 inverse. I f f ~ ' exists, sketch its graph.
Find the inverse (if it exists) o f each function given in Problems 38. ∕( x ) = x 2, for all x 39. ∕( x ) = x 2 , x ≤ 0
7-12. 40. ∕( x ) = 1O , for all x
Λ
1 - f = { ( f i , 5), (6, 3), (7, 1), (2, 4)} 41. ∕( x ) = V l - x 2 on ( -1 , 1)

8. ∕( x ) = 2x + 3 42. ∕'(x) = x(x - l)(x - 2) on [1, 2]
9. g(x) = x 2 - 5, x ≥ 0 43. ∕( x ) = cosx, on [0, ττ]
10. g(x) = V x + 5 44. ∕( x ) = tanx, on (—■
j> ⅞)
" . *w - ∣⅛ ≈
45. According to the Bouguer-Lambert law, a beam of light
2a' + 1
that strikes the surface o f a body of water with intensity I o
12. A(x) = v will have intensity I at a depth of x meters, where
Evaluate the expressions given in Problems 13-22. I = I 0e kx k > 0

13. log,4 + log3 1 14. 2 ⅛ 3 ^ ⅛ 5 The constant k is called the absorption coefficient, and it
16. (log, ∣)(log 327) depends on such things as the wavelength o f the beam of
15. 5 log39 - 2 log2 16
light and the purity of the water. Suppose a given beam of
17. (3log71)(log5 0.04) 18. e 5 ln 2 light is only 5% as intense at a depth of 2 meters as it is at
19. log334 - In e>0∙5 20. ln(log 101') the surface. Find k and determine at what depth (to the
21. exp(ln 3 - In 10) 22. exp(logc 225) nearest meter) the intensity is 1% of the intensity at the
surface. (This explains why plant life exists only in the top
Solve the logarithmic and exponential equations to calculator 10 m of a lake or sea.)
accuracy in Problems 23-33. 46. A manufacturer of light bulbs estimates that the fraction
23. logτ 16 = 2 24. log x = 5.1 F(f) of bulbs that remain burning after t weeks is given by
25. e~ 3x = 0.5 26. In x 2 = 9
F(t) = e~k ,
27. 7^x = 15 28. e 2x = ln(4 + e)
where k is a positive constant. Suppose twice as many
29. ∣log3x = log,8 30. log,(x logA) = 4 bulbs are burning after 5 wk as after 9 wk.
31. log3x + log3(2x + 1) = 1 a. Find k and determine the fraction o f bulbs that are
burning after 7 wk.
32. l n ( √ ^ - ) = In x + lnf τ ⅛ -
∖1 - x) ∖1 + x ∕ b. What fraction of the bulbs burn out before 10 wk?
33. 2 31°g2* = 4 log39 c. What fraction of the bulbs can be expected to burn out
34. Let ∕( x ) = 3x + 5. Show that f ~ ∖x) exists and use the between the 4th and 5th weeks?
derivative of an inverse function theorem to evaluate the 47. A certain bank pays 6% interest compounded continuous
derivative o f ∕ - 1 (x) at x = 2. ly. How long will it take for $835 to double?
48. If an amount of money is compounded continuously at an a. Express E in terms of M.

annual rate r, how long will it take for that amount of b. How much more energy is released in an M = 8.5
money to double? earthquake (such as the devastating Alaska quake of
49. First National Bank pays 7% interest compounded 1964) than in an average quake of magnitude M = 6.5
monthly, and World Savings pays 6.95% interest com (such as the Los Angeles quake of 1994)?
pounded continuously. Which bank offers a better deal? θ 55. Let b be any positive number other than 1. Show that
50. A person invests $8,500 in a bank that compounds Xx = ⅛xlog ft x
interest continuously. If the investment doubles in 10
years, what is the (annual) rate of interest (correct to the 56. Let a, b be any positive numbers other than 1. Show that
nearest hundredth percent)?
51. In 1626, Peter Minuit traded trinkets worth $24 for land (logfl⅛)(logftα) = 1
on Manhattan Island. Assume that in 1990 the same land
was worth $25.2 billion. Find the annual rate of interest 57. Verify directly that if (w, v) is a point on the graph of
compounded continuously at which the $24 would have y = logfix, then (v, ιz) is on the graph of y = bx , and vice
had to be invested during this time to yield the same versa.
amount. 58. Suppose that / i s differentiable and that its inverse exists.
52. Biologists estimate that the population of a bacterial Let y = ∕(x ) and show that
colony is
P(∕) = P02kt d fd f~ ' 1
dx dy
at time t (in minutes). Suppose the population is found to
be 1,000 after 20 min and that it doubles every hour. Hint: Apply the chain rule to the equation ∕ - 1 (y) = x.
a. Find PQ and k. 59. Suppose th at/is differentiable and that its inverse exists.
b. When will the population be 5,000 (correct to the Let y = f(x) and show that
nearest minute)?
53. A decibel (named for Alexander Graham Bell) is the
smallest increase of the loudness of a sound that is
detectable by the human ear. In physics, it is shown that 1
¢7/°
when two sounds of intensity ∕ 1 and 7, (watts∕cm 3) occur, for all y in the domain of∕ ^ if > 0 for all x in the
the difference in loudness is D decibels, where domain of / Hint: Conclude that ∕ ^ 1 is increasing if / i s
D = 10 1og,o(τO increasing, and then show t h a t / -1 is decreasing if / i s
V
decreasing.
2∕
When sound is rated in relation to the threshold of human 60. Show that the line y = x is the perpendicular bisector of
hearing (IQ = 10^16) the level of normal conversation is 50 the line segment joining the points (a, b) and (b, a). Notice
decibels, whereas that of a rock concert is 110 decibels. that this means (a, b) and (b, a) are reflections of one
Show that a rock concert is 60 times as loud as normal another in the line y = x.
conversation but a million times as intense. 61. Show that i f / - 1 exists, it is unique. Hint: If g i and g7 both
54. The Richter scale measures the intensity of earthquakes. satisfy
Specifically, if E is the energy (watt∕cm 3) released by a
quake, then it is said to have magnitude M, where g l [∕(x)] = x = ∕[ g 1W]
g2[∕W J = χ = ∕[⅛ 2W]
= 1 0 g £~ 1L4
M
1.5 then show that y 1(x) = g,(x).
5 .3 DERIVATIVES INVOLVING e' AND In x
IN THIS SECTION Derivatives of natural logarithmic functions, derivatives

of exponential functions, logarithmic differentiation.
The goal of this section is to develop the differential calculus of logarithmic and
exponential functions. We shall begin by deriving differentiation formulas for
In x and ex , and we shall apply these formulas to a number of differentiation
problems and applications. We shall also develop a technique called logarithmic
differentiation, which can be used to simplify the differentiation of products
and quotients.
Sec. 5.3 Derivatives Involving ex and In x 345
■ DERIVATIVES OF NATURAL LOGARITHMIC FUNCTIONS
THEOREM 5.9 Derivative of In u
I f x > 0, then (In x) = ±
If u is a differentiable function of x, then ^(1∏ ») = ~ for u > 0.
Proof: To find the derivative of In x, we apply the definition of derivative.
-4- (In x) = lim ln(x + ∆x) - In x

dxv 7 ΔΛ-→ O ∆x
= lim -4- In ∕ x + ∆x∖
Δx→ 0 ∆x ∖ x )
lim -4- In
∆x
Δ,v→ 0
Let h = τ-> so ∆x = ÷.
∆x h
As Δ x→ 0, h→ ∞
= lim Inf 1 + Γ)
h→ +∞ x ∖ nJ
= - lim h Inf 1 + 4-)

X h→ ÷∞ ∖ fl/
1 / 1∖ a
= - hm In 1 + 4-) Power rule for logarithms
X h→ +∞ ∖ fl/
1 / 1∖λ
= - In hm 1 + 4-) Lim it o f a log property
X h→ +∞ ∖ rl/
= In e Definition o f e
= 1
In e = 1, because e' = e.
x
For the second part of this theorem, let u be a differentiable function of x and
apply the chain rule to In u to find
≠ ( l n w) = Γ ≠ (ln =
dxv ’ Ldu v 7J dx u dx
EXAM PLE 1 Derivative of a natural logarithm
Differentiate y = ln(5x2 + 2x + 1).

Solution Use the chain rule where ιι = 5x 2 + 2x + 1, so = 10x + 2 and
⅞ = —X τ (lθ * + 2) 1Ox + 2
dx 5x 2 + 2x + Γ ,
5x 2 + 2x + 1
V = 5x 2 + 2x + 1 shows In Example 1, we did not specifically note that 5x2 + 2x + 1 ≥ 0 but it is
5x 2 + 2x + 1 > 0 . worth noting because In u is defined only for u > 0. See Figure 5.10.
T H E O R E M 5.10 Derivative of In ∣
w∣
I f ∕( x ) = In ∣x∣, x ≠ 0, then ∕' ( x ) = γ
Also, if u is a differentiable function of x , then In ∖u∖= ~
Proof Using the definition of absolute value,

ct ∖ _ ∏n x i f x > 0
2 A [ln (-χ) if x < 0
so
∣ if % > 0
∕ ,ω = 1 1
j - ( - l ) = -l ifx < 0
L~x ’ x
T h u s,∕'(x ) = - for all x ≠ 0. The second part (for w, a differentiable function of x)
follows from the chain rule. ■
EXAM PLE 2 Derivative of a function involving a natural logarithm
Differentiate ∕( x ) = sin x + V in x.
Solution Write ∕( x ) = sinx + (In x ) ιz2.
f (x) = cosx + ∖ (In x) ιz2 Q = cosx +
EXAM PLE 3 Derivative of a function involving a natural logarithm
Differentiate ∕( x ) = (x2 + ln V x) 3.
Solution Write ∕( x ) = (x 2 + ⅜In x) 3. Use the power rule of logarithms.
∕'( x ) = 3(x2 + A In x) 2 (2x + J-') = ⅛ (x 2 + lnV x) 2(4x2 + 1)

∖ J Λ,X ∖ /
■ DERIVATIVES OF EXPONENTIAL FUNCTIONS
The next theorem is easy to prove and remember and also is one of the most
important results in all of differential calculus.
T H E O R E M 5.11 Derivative of eu
If ιι is a differentiable function of x, then -4- (e lt) = e u 4r.

’ dxκ ’ ax
Proof: Let v = ex so that
x = In v
⅛ W ≡ ⅛ ( > " ''> Implicit differentiation

Sec. 5.3 Derivatives Involving ex and In x "5^1
∣ _ 1 r/v
v dx
v = ⅛-
dx
ex =⅛Γ⅛Sex} βecause v ~ e'
For the second part of this theorem, let u be a differentiable function of x , and
apply the chain rule to e “ to find
d (&u\ — d ( tι∖ du — u du
d x ( e ’ ~ d u { e > d x ~ e dx ■
EXAM PLE 4 Derivative of an exponential function

7
Differentiate ∕( x ) = e x + *.
Solution e χ l + x (2x + 1)
EXAM PLE 5 Derivative of a product involving an exponential function

L__ / . U~ ~ . Z ∙----------------------------------------------------------------------------------------------------------------------------------------------------------------------
Differentiate e" 3λ'sinx.

Solution Use the product rule:
= e 3x(cosx) + [e 3x (-3)] sinx

= e - 3 x (cosx - 3 sinx)
0 It is important to remember that a power function x" has a constant

exponent and a variable base, whereas an exponential function bx has a
constant base and a variable exponent. You should not expect a function like
2x to behave like x 2 . 0 The following theorem illustrates a major
behavioral difference between the monomial x n and the exponential bx .
T H E O R E M 5.12 Derivative of an exponential function with base b
For base b {b > 0, b ≠ 1),
≡ i^ = (in *)*-
and if u is a differentiable function of x, then
- ^ b u = (In b)bu for all u.
Proof: Because b x = e x ln b, (Theorem 5.8f), we can apply the chain rule

formula (e") with u = x In b to obtain
~ J x = d x ( e X ln = exλnb d x 'n = e X 'n *θn b) ~
The second part of this theorem follows from the chain rule.
EXAM PLE 6 Derivative of an exponential function with base b ≠ e
Differentiate f(x ) = x(2 1 - x ).

Solution Apply the product rule.
∕' W = ⅛ C v 2 > -) = x ⅛ ( 2 ' - n + 2 '- ⅛ ( ,v )
= x(ln 2)(21- χ ) ( - l) + 2 , - χ (l) = 2 1- χ (l - x l n 2) .
T H E O R E M 5.13 Derivative of a natural logarithm with base b
For base b (b > 0, b ≠ 1)
■4- (log,x) = τi—U — for all x > 0

d x v 66 ' (In b) x
and if u is a differentiable function of x , then
-y-(lθg,w) = 71—U — ~ r for all u > 0

d x v 6ft ’ (In o) u dx
Proof: Because ∣ogzx = and because (In x) = for x > 0, we have
d_ ∩ A d_ Γ ln xl = _ L _ . 1 = ___ 1___
1 &b- > = √χLln⅛J
c∕ λ Λ In b x (In b)x
Once again, the second part follows from the chain rule.
HI What this says: The derivatives of bx and logftx are the same as the
derivatives of ex and In x, respectively, except for a factor of In b that appears
as a multiplier in the formula
⅛ ( fr ') - 0" i )fr,
and as a divisor in the formula
⅛ ( 10¾χ ) = ( 1 ⅛
EXAM PLE 7 Derivative of a logarithmic function with a base b ≠ e
Differentiate g(x) = log3 1 ∣x2 - x∣.

Solution Let u = x 2 - x and b = 3.1.
■?'« - i ⅛ ( ⅛ ,k 2 -*1) = (ln 3.i)∣^-,τ)⅛ ^- *)

2 γ∙ _ ι
= 71— „ ζ ,. ------r
j For all x such that u = x 2 - x ≠ 0;
(In 3 .1 ) ( χ - - x ) t h a t i s , χ ψ (y ι
■ LOGARITHMIC DIFFERENTIATION
Logarithmic differentiation is a procedure in which logarithms are used to trade

the task of differentiating products and quotients for that of differentiating sums
and differences. It is especially valuable as a means for handling complicated
Sec. 5.3 Derivatives Involving e x and In x 349
product or quotient functions and power functions where variables appear in both
the base and the exponent.
EXAM PLE 8 Logarithmic differentiation
κ . , the
., , ∙ .∙ r e 2x (2x - 1)6
Find derivative ot y = — z—.
(x 3 + 5)2(4 - 7x)
Solution The procedure called logarithmic differentiation requires that we first
take the logarithm of both sides and then apply properties of logarithms
before attem pting to take the derivative.
e 2x (2x - 1)6
y ~ (χ 3 + 5)2(4 - 7x)
e 2x (2x - 1)6
In y = In
(x 3 + 5)2 (4 - 7x)
= In e 2x + ln(2x - 1)6 - ln(x 3 + 5)2 - ln(4 - 7x)
= 2x + 6 ln(2x - 1 ) - 2 ln(x 3 + 5 ) - ln(4 - 7x)
Next, differentiate both sides with respect to x and then solve for
j I =2 + ⅛ ⅛ > ] - 2[⅛ < 3*2>] - [ 4 ⅛ - 7)]

⅛ = v Γ2 + - ^ ______6 x ^ 7 . .1
+
dx ∙' Lz 2x - 1 χ 3 + 5 4 - 7xJ
This is the derivative in terms of x and y. If we want the derivative in terms of

x alone, we can substitute the value of y:
dy = e 2x (2x - 1)6 Γ 12 _ 6x 2 + 7 1
dx (χ 3 + 5)2(4 - 7x) L 2x - 1 χ 3 + 5 4 - 7xJ
EXAM PLE 9 Derivative of a function with a variable in both the base term
and the exponent
Find where y = (x + l) 2x .
Solution y = ( x + l ) 2x
In y = ln[(x + 1)2Λ'] = 2x ln(x + 1)
Differentiate both sides of this equation:
~ 2 tl n (x + 1 >1} + [ ⅛ ( 2 x ) ] l n (x + 1> p r °d u c l r u l e
y Tx = ⅛
= 2 x [ ⅛ ( l ) ] + 2 ln(x + 1 ) = ⅛ + 2 ln(x + 1)
Finally, multiply both sides by y = (x + 1)2+

l ^ [ ⅛ + 21 * + ⅜ + 1 +
Logarithmic differentiation is also a valuable theoretical tool. For example, we

can use it to prove that the power rule holds even when the exponent is an
irrational number.
EXAMPLE 10 Proof of power rule for irrational exponents
Prove (xr ) = rx'' 1 holds for all real numbers r.
Solution If y = x r, then y = e r i n x , so that

In y = r In x
L⅛ J, L∖
y dx ∖ x)
dy _ ( A
dx ~ ^∖ x)
Substitute y - x r.
= rx r~i
PROBLEM SET 5 .3
θ Differentiate the functions given in Problems 1-26. 36. y = V ( x l ° + l) 3(x 7 - 3)8 47 .. 1∕
1. y = e3∙v 2. y = In 5x
J V X - 9 ( X + 3)2
3. y = ln(3x 4 + 5x) 4. y = e 2 x + ' 4e — e ~ ' _______ 40 P
c
3X2
r υ
J (X - 2 3)2√4 - 5x
5. y = x 2 + 2x 6. y = 3λ' + x 3 (x 3 + l) 2(4x - 7)^2
1. y = π e + e π + ex + x 8. y = ττx + x π 40. y = xx 41. y = x lnVx
9. y = eχ2 10. y = ex + x e 42. y = (x 2 + 3)2x 43. y = (sinx) v ∙*
4x
11. f(x ) = e^ + e ~ 3 2x 12. ∕( x ) = e 3x - e^ 5x 44. y = χ∙v'v 45. y = (χ∙v)x
2
13. /( /) = exp(∕ + / + 5) 14. f(x ) = exp(x2 - 3x^ l ) 46. y = x s ,n ∙v 47. y = (sinx)-v
15. ∕( u ) = ∣(e 2κ - e^2") 16. /( /) = e 3,+ 1 + e l - 3 ' Find an equation for the tangent line to the curves given in
1 e~x
Problems 48-62 at the prescribed point.
17. ∕( x ) = x + e~ 18. ∕( x ) = (x 2 + 3)e1+x
48. y = (In x)-v where x = e 49. y = x ln x where x = 1
19. g(∕) = + π2 20. g(t) = e'sin(2∕ + 1) 50. y = (sinx) co sx where x = ? 51. y = x x 'λ where x = 1
21. g(∕) = (ev l^ - ∕)2 22. ∕( x ) = (ex + In 2x)2 52. y = xe~ x where x = 0 53. y = — where x = 1
x
23. ∕ ( M) = ln(ln u) 24. ∕( x ) = exp[exp(exp x)] vsinx
54. y = e where x = 0 55. x 2 + 2y = ey at (1, 0)
25. y = ln(sinx + cosx) 26. y = ln(secx + tanx)
56. x log2y = x - y at (1, 1) 57. y = x2^ 3-v where x = 1
Use implicit differentiation to fin d in Problems 27-32. 58. y = x In x 2 where x = 1 59. _ I n V x w jl e r e χ = 4
27. xe^ x = y 28. xe y + ye x = 2x 60. ln(xy) = 2(x - y) at (1, 1)
29. exy + 1 = x 2 30. ln(xy) = e 2x (2x2 - 3)(x + 2)2 ,
61. y = , ------ <7s- where x = 1
(z x 3 + 03x - 5)5
31. exy + In y 2 = x 32. l o g . ( - ) = x
∖^y j 62. y = ( 1 - 3x)2∙γ where x = - 1
θ 33. ■ What Does This Say? Compare and contrast the 63. Find the equation o f the line that is normal to the graph of
derivatives of the following functions. the equation
a. y = x 2 b. y = 2λ^ c. y = e x d. y = x e 2x + y = 1 + ey
34. ■ What Does This Say? Compare and contrast the at the point (1,0)
derivatives o f the following functions: 64. Find an equation for the normal to the curve y = V x In x
a. y = log x b. y = In x at the point where x = e~ 2.
35. ■ What Does This Say? Discuss logarithmic differenti θ 65. a. If F(x) = ln∣cosx∣, show that F'(x) = -ta n x .
ation. b. If F(x) = ln∣secx + tanx∣, show that F'(x) = secx.
Use logarithmic differentiation in Problems 36 -4 7 fo r finding 66. Evaluate lim — -,—-.
dv Λ→o h
You may express your answer in terms o f both x and y, and
67. I f∕( x ) = ex , show th a t∕'(x ) = ex by using the definition of
you do not need to simplify the resulting rational expressions. derivative. Hint: You will need the limit from Problem 66.
Sec. 5.4 Applications Involving Derivatives of ex and In x 351
5 . 4 APPLICATIO NS INVOLVING DERIVATIVES OF e x A N D In x
Computational Window ^ π IN THIS SECTION Curve sketching, optimization, evaluating limits using
Remember, even though a logarithms and ΓH6pitaΓs rule
graphing calculator can be very We now study a few applications involving the derivatives of exponential and
helpful, you cannot assume logarithmic functions. We begin with curve sketching, then consider examples
that the calculator will do all of optimization and rectilinear motion, and, finally, use ΓH6pitaΓs rule to
your work for you. Indeed, the evaluate limits involving exponential and logarithmic functions.
ideas of calculus are often nec
essary to be sure you have the
correct graph. ■ CURVE SKETCHING
EXAMPLE 1 Graphing a function involving a logarithmic function
Determine where the function ∕(x ) = is increasing and where it is

decreasing. Sketch the graph of f.
Computational Window ~ ∏
Solution First of all, note th a t/is defined only for x > 0 and that
⅛) - 1 In x 1 - In x
The graph of f(x) = is (x) =
shown:
Because x 2 > 0 for all x > 0, it follows that ∕'( x ) = 0 only when In x = 1;
thus, e is the only critical number off. We use the first-derivative test and look
at the sign of ∕'( x ) to the left of e (at x = 1, for instance) and to the right (at
x = 3, for example). We see that
f'(x ) > 0 for x < e and
∕'( x ) < 0 for x > e
Y ιB ln X×X T hus,/is increasing to the left of e and decreasing to the right, as shown in
X n in = -IS Y m in = -10 Figure 5.1 la.
X n a x= lθ Ynax=10 We now have most of the information we need to sketch the graph of / In
X s c l= l Y s c l= l particular, we know that the graph lies entirely to the right of the y-axis
(because/ is defined only for x > 0) and that it rises to a relative maximum at
χ = e, after which it falls indefinitely. Moreover, because we obtain
= 0 only for x = 1
Xmax=10 Ym ax=.4
X s c l= l Y s c l= . 1
(-
o
a. Tendency diagram for b. The graph rises to a maximum c. Further analysis shows
f ω =⅛* at e and then “flattens out” as that the graph approaches
it falls. the x-axis as x →+ ∞ .
Figure 5.11 The graph of ∕(x ) =

it follows that the graph crosses the x-axis on the way up but not on the way
down, and this causes us to suspect that the graph must “flatten out” in some
way as x → +∞ (see Figure 5.1 lb).
Note that ∕(x ) = becomes large and negative as x approaches 0
through positive values; that is,
lim = -∞
x→o+ x
To see what happens to ∕(x ) as x increases without bound, we use ΓHδpitaΓs
rule (Section 3.9) and find that
1
lim
x→+∞ X = χ→+∞ 1 = 0
lim f
The graph of f approaches the y-axis asymptotically as x tends toward 0,

and approaches the x-axis as x increases without bound. The completed graph
is shown in Figure 5.lie. Notice that as the graph comes down from the
relative maximum at e, its concavity appears to change from down to up.
Indeed,
x 2( 4 ) - (1 - In x)(2x) _ _ 3 + 2 In χ
∕" (x ) =
(x2)2 x3
Solving the equation ∕" (x ) = 0 for x, we find:
- 3 + 2 In x _ Q
v3
2 In x = 3
In x = j, or x = e 3/2
Because e 3'2 ≈4.48, we find f''(4) < 0 and ∕" (5 ) > 0, so that / has an
inflection point at x = e 3l2. w
Functions such as or xe 5λ', which have the general form 4 r or x nekx ,

X X
appear frequently in applications, and it is convenient to have the limit formulas
given in the following theorem.
THEOREM 5.14 Limits involving natural logarithms and exponentials

If k and n are positive integers, then
lim ⅛ = - ∞ lim ⅛ n = 0
Λ→ 0÷X’' x →+∞
x
,lim
∙ — ekx
→ +∞ xτ-
n = +∞ lim x ne~kx = 0
x
Proof: These can all be verified directly or by applying ΓHδpitaΓs rule. For
example,
lim 4 r = lim —- lim —¼n = θ
.-→+∞ x x→+∞ nx l∙→+∞ nx
The other parts are left for you to verify.
EXAMPLE 2 Graphing a function involving an exponential function

Determine where the function
∕( X ) = e - χ 2 ∕2
V 2ττ
Sec. 5.4 Applications Involving Derivatives of e x and In x 35 3
is increasing, decreasing, concave up, and concave down. Find the relative
extrema and inflection points and sketch the graph. This function plays an
important role in statistics, where it is called the standard normal density
function.
Solution The first derivative is
Because e χ l ' 2 is always positive, ∕'( x ) = 0 if and only if x = 0. Hence, the

corresponding point
0. ξ ⅛ ) - (0. 0.4)
is the only critical point. Checking the sign off ' on each side of 0, we find that
f is increasing for x < 0 and decreasing for x > 0, so there is a relative
maximum at x = 0, as indicated in Figure 5.12a.
a. Intervals of increase and

decrease for∕(%)
1 2,2
Figure 5.12 Preliminary sketches for the graph of ∕( x ) = ~ y = e^x
χ 2ττ
We find that the second derivative o f/ is
- X2∕2 - L_e :-/2
which is zero when x = ±1. We find that/(1) = / (-1 ) ≈ 0.24, and that the
concavity of the graph o f/ is as indicated in Figure 5.12b.
Finally, we draw a smooth curve through the known points, as shown in
Figure 5.13. The graph o f / rises to the high point at (0, 0.4) and then falls
indefinitely, approaching the x-axis asymptotically because e~x 12 approaches
0 as ∣x∣ increases without bound. Note that the graph has no x-intercepts,
because e~x z2 is always positive. c= ∣
■ OPTIMIZATION
Optimization problems involving exponential and logarithmic functions occur in

many applied areas and can be analyzed by the methods of Chapter 3. Here are
Figure 5.13 Graph o f the standard
normal density function: two examples.
∕W = ~ ⅛ = e ^ λ^2z2
∖ lπ EXAM PLE 3 Maximum concentration of a drug
Let C(∕) denote the concentration in the blood at time t of a drug injected into
the body intramuscularly. In a now classic paper by E. Heinz, it was observed
that the concentration is given by
Ctf) = ~ e~b') t≥ 0
where a, b (with b > a) and k are positive constants that depend on the drug.
At what time does the largest concentration occur? What happens to the
concentration as t → +∞7
Solution To locate the extrema, we solve C'(∕) = 0.
= ⅛ [ ( - α ) ^ β' - {-b)e~ b∙] = τ ^ - a {be~bt - ae~<")
We see that C'(ζ) = 0 when

be~btc = ae~alc
e atc -btc = a
a tc ~ b tc = ln f Definition of logarithm
t = —½ ^ l n T
r a- b b
Using the second-derivative test, it can be shown that the largest value of
C(Z) occurs at tc (see Problem 57).
To see what happens to the concentration as t → +∞, we compute the limit
Maximum
lim at bl
∕→ +∞ C(t) = ∕→ +∞ D Cl [e~ - e~ ]
lim
_ k Him 4 / - lim -⅛ = T ~ ~ [θ ~ θ]
b - a Lz→+ =0 e at t→+∞ e bt∖ b ~ a 1 j
This tells us that the longer the drug is in the blood, the closer the
concentration is to 0. The graph of C is shown in Figure 5.14.
Intuitively, we would expect the Heinz concentration function to begin at
0, increase to a maximum, and then gradually drop off to 0. Figure 5.14
indicates that C(Z) does not have these characteristics, because it does not
Figure 5.14 Graph of quite get back to 0 in finite time. This suggests that the Heinz model may
C(0 = r ⅛ ^ a' - e~bt) apply most reliably to the period of time right after the drug has been
injected.
EXAMPLE 4 Greatest consumer expenditure

Sales Expenditure Consumer demand for a certain commodity is
0 0 D(p) = 5,000e-00 2p
10 40,936.54
units per month when the market price is p dollars per unit. Determine the
20 67,032.00
market price that results in the greatest consumer expenditure.
30 82,321.75
40 89,865.79 Solution The consumer expenditure E(p) for the commodity is the price per unit
→50 91,969.86<-Maximum on list p times the number of units sold; that is,
60 90,358.26
70 86,308.94 E(p) = pD(p) = 5,OOOpe~o °2p
80 80.758.61
90 74,384.50 Because only nonnegative values of p are meaningful in this context, the goal is
100 67,667.64 to find the absolute maximum of E(p) on the interval [0, ∞). The derivative of
150 37,340.30 E(p) is
Sec. 5.4 Applications Involving Derivatives of ex and In x 355
E '(p ) = 5,000(-0.02pe- 0 ∙02p + e~002 P)

= 5,000(1 - 0.02∕t)e-° 0⅛
Because e~0 -02p is never zero, E'(p) = 0 if and only if
1 - 0.02p = 0, or p = 50
The easiest way to verify that p = 50 actually gives the absolute maximum of
E(p) on the interval p > 0 is to check the sign of E '(p) for 0 50:
sumer expenditure function
£(p) = 5,OO(⅛<ro o ⅛ If 0 0.
If p > 50, then E ∖p) = 5,000t>-°°⅛(l - 0.02p) < 0.
Hence, E(p) is increasing on (0, 50) and decreasing on (50, +∞), and there is
an absolute maximum when p = 50 as shown in Figure 5.15. Thus, consumer
expenditure will be greatest when the market price is $50 per unit.
■ EVALUATING LIM ITS USING LO G ARITHM S A N D L’H OPITAL’S

RULE
Certain limits of the form lim [∕(,x)] u w will result in forms of the type 0°, ∞0 , or
1” , which are indeterminate forms (recall Section 3.9). We shall find that
logarithms can often be used to reduce such forms to forms of the type θ or ∣ ,
which can then be handled by ΓH6pitaΓs rule. The usual approach is to let L equal
the limit we wish to evaluate and then take the natural logarithm of both sides,
which should then allow us to use ΓH6pitaΓs rule. This procedure is illustrated in
the following example.
Computational Window J3E∏

You can sometimes use a EXAM PLE 5 Limit of the form ∞ 0
graphing calculator to verify
limits like the one in Example Find lim x xlx .
χ-→+∞
5: Solution This is a limit of the indeterminate form ∞0 .
If L = lim x xlx then
x → +∞
In L = In χlim x Vx
→ ÷∞
= lim In x x'x The limit of a log is the log of a limit
X → +∞
= lim - In x
χ→+∞X
VιBX^X-ι = lim
.v→ +∞ X
X h in = 0 Vnin=Θ
Xnax=10Θ0 Ymax=2 1
X s c l= 1 0 0 Y s c l= .5 = lim 7 ΓH∂pitaTs rule
χ→+∞ 1
Notice that the graph looks like
the line y = 1, which confirms = 0
that L = 1.
Thus, we have In L = 0; therefore, L = e 0 = 1.
EXAM PLE 6 Limit of the form 1 “

To use a graphing calculator on Find lim f l +
a limit like the one in Example
6, you can choose a value for k Solution This is a limit of the indeterminate form 1“
(such as k = 2) and graph the
resulting function. If L = lim ( 1 + ⅜) , then
A'→÷∞ ∖ X∕
In L = In lim ( 1 + ⅜)
X→+∞ ∖ /
/ AΛX
= lim In 1 + -
x →+∞ ∖ x)
= lim [x Inf 1 + m ]
x→+∞ L ∖ X∕J
To convert this ∞ ∙ 0 indeterminate form into a θ form, we write x as 1/( 1/x):

Xnin=Θ V n in = 0
X∣ηax=1000 Y∩ax=10
X ≥ cl= 1 0 0 V s c l= l In L = lim Form ∞ ■0
It looks like the limit is some

where between 7 and 8. This is F77orm 0Q
consistent with the answer in
Example 6 (for k = 2):
L = e 1 ≈ 7.389056099 ΓH∂pitaΓs rule
Thus, L = e k .
PROBLEM SET 5 .4
0 In Problems 1-4, the graph o f the given function f(x ) for x > 0
in Figure 5.16. In each case,
3. ∕( x ) = — 4. ∕( χ ) = e x cosx
In Problems 5-20, answer the following questions about the
given function, and then sketch its graph.
a. What is the domain o f the function?
b. What are the critical numbers, and where is the graph o f the
function rising and falling?
c. What are the coordinates o f all high and low points? Where
does the graph cross the coordinate axes?
Sec. 5.4 Applications Involving Derivatives o f e x and In x 357
d. Where is the graph concave up and concave down? What are where m is a real number and σ is a positive constant.
the second-order critical numbers? a. Find D'(t) and determine all relative extrema o f D(t).
5. ∕( x ) = 3x ^ 1 6. ∕( x ) = 2 3 " x
b. What happens to D (∕) as t → +∞ and as t → —∞7
7. ∕( x ) = x - In x 8. ∕( x ) = x In x
c. Sketch the graph of £>(/).
9. ∕( x ) = xe~ x 10. ∕(w ) = u 2e~ 2u
41. A Gompertz curve is the graph o f a function of the form
l l .∕ ( x ) = ^ 12. ∕( x ) = log32x
G'(x) = A exp (-B e~ kx ) for x > 0
9w
13. ∕( x ) = e χ2 14- / ( -) = ⅛ where A , B, and k are positive constants. Such graphs
appear in certain population studies. Find G , (x) and
/ /2 ∖ G'"(x), and sketch the graph of G.
15. /(/) = ln(^τ ⅛ 7 J 16. g(∕) = ln(4 - t 2)
42. In an experiment, a biologist introduces a toxin into a
17. ∕(x) = ex + e x bacterial colony and then measures the effect on the
18. j∕ xW = ~ xx 7 g -rx
’ e + e population of the colony. Suppose that at time t (in
19. ∕( x ) = (In x) 2 20. ∕( x ) = In x 2 minutes) the population is
21. Let S(x) = ∣(e v - e x) and C(x) = ∣(e x + e x ). These P(f) = 5 + e-°∙04'(t + 1)
functions are known as the hyperbolic sine and cosine,
thousand. At what time will the population be the largest?
respectively, and are examined in Section 7.6.
What happens to the population in the long run (as
a. Show that 5"(x) = C(x) and C'(x) = 5,(x). t → +∞)7 Find where the graph of P has an inflection
b. Sketch the graphs of S and C. point, and interpret this point in terms of the population.
22. Find all points of intersection of the curves Sketch the graph of P.
43. Some psychologists believe that a child’s ability to memo
y = 2x ^ 2x and y = 4λ'
rize is measured by a function o f the form
Sketch the two curves on the same set of coordinate axes.
.. f t In / + 1 if 0 < t ≤ 4
23. Let b be a positive number other than 1. Sketch the graph ^ ) = (1 i f ∕= 0
of y = logi x for each case.
a. b > 1 b. b < 1 where t is time, measured in years. Determine when the
largest and smallest values of g occur.
Find the maximum and minimum values o f the functions given 44. Suppose a manufacturer estimates that when the market
in Problems 24-27 on the prescribed interval. price of a certain product is p, that
24. ∕( x ) = xe~ 2x on [ - 1 , 2] 25. ∕( x ) = ξ on ⅛ 2]
x = ~6 ln(⅛)
26. /(I) = on ⅛, 9] 27. /(/) = te- t l on [ - 1 . 1]
units will be sold. It is also estimated that the cost of
Use ΓH∂pitaΓs rule to evaluate the limit in Problems 28-37. producing these x units will be
28. lim 29. lim, x In x
t → 0* r →0
30. lim x s'n x 31. lim (sinx) In x C(x) = 4xe~ x ,6 + 30
r →0 '
ln(ln x) ι-iΓ
32. lim 33. lim a. Find the average cost, C (x)∕x, the marginal cost, C'(x),
and the marginal revenue for this production process.
34. lim (In x ) I/x 35. lim (ex + x ) 14t
Λ→ 0 b. What level o f production x corresponds to maximum
lim+ χ ιz<χ - 1> 37. lim (ei ^- 1 —x) v
36. profit?
ι →0
38. Sketch the graph of ∕( x ) = e^⅛inx. Note that the graph
crosses the x-axis only where x = nτr for any integer n. In Problems 45-47 you are asked for the percentage rate. We
Where does the graph have high points and low points? define the percentage rate o f a function f at the value x to be
What happens to the graph as x → +∞7
100∕'(x)
θ 39. Show that for positive constants m , k,
.m
∕(-v)
if k > m
lim 45. It is projected that t years from now, the population of a
if k < m
certain country will be P(t) = 50e0 °2' million.
What is the limit when k = m l
a. At what rate will the population be changing with
40. A function that plays a prominent role in statistics is respect to time 10 years from now?
b. At what percentage rate will the population be chang
ing with respect to time t years from now?
46. A certain industrial machine depreciates in such a way 52. An epidemic spreads through a community in such a way
that its value after t years is that t weeks after its outbreak, the number of residents
who have been infected is given by a function of the form
<2(0 = 20,000e-°∙4'
dollars.
a. At what rate is the value of the machine changing with ∕( z) ^ 1 + c e -kt
respect to time after 5 years? where A is the total number of susceptible residents. Show
b. At what percentage rate is the value of the machine that the epidemic is spreading most rapidly when half the
changing with respect to time after I years? susceptible residents have been infected.
47. A quantity decreases exponentially (that is, f(t) = Ae~k '). 53. A particle moves along the x-axis in such a way that at
a. Show that its rate of decrease is proportional to its size, time t it is at x(zj = V~te~, .
b. Show that the percentage rate of change is constant. a. Find the velocity and acceleration of the particle at
48. A cool drink is removed from a refrigerator on a hot time t.
summer day and placed in a room whose temperature is b. Show that the particle initially moves in a positive
30 0C. According to Newton’s law of cooling, the tempera direction and then at a certain time T it turns and
ture of the drink I minutes later is given by a function of moves back toward the origin. What is T? Where is the
the form / / ) = 30 - Ae~k , . Show that the rate of change particle when it turns?
of the temperature of the drink with respect to time is c. What is the total distance traveled by the particle
proportional to the difference between the temperature of between times t = 0 and t = 1?
the room and that of the drink. 54. A particle moves along the x-axis in such a way that at
49. The mathematics editor at a major publishing house time t, its position is given by
estimates that if x thousand complimentary copies are
x(t) = A e k, + Be~k,
distributed to professors, the first-year sales of a new text
will be Show that at any given time, the acceleration of the
f(x ) = 2 0 - 1 5e~0-2x particle is proportional to its position.
thousand copies. Currently, the editor is planning to θ 55. Suppose the percentage rate of change of the square of a
distribute 10,000 complimentary copies. function/is proportional to cos x and that/(0) = 1. What
a. Use marginal analysis to estimate the increase in first- is / ( 77)?
year sales that will result if 1,000 additional compli a. Show that its rate of decrease is proportional to its size.
mentary copies are distributed. b. Show that its percentage rate of increase is constant.
b. Calculate the actual increase in first-year sales that will 56. Show that the graphs of y = e~k x and y = ekx are reflec
result from the distribution of an additional 1,000 tions of one another in the y-axis.
complimentary copies.
57. In this section we showed that the Heinz concentration
50. A manufacturer can produce shoes at a cost of $50 a pair function,
and estimates that if they are sold for x dollars a pair,
consumers will buy approximately e
C(t) = - ~bt>
s(x) = l,000e~° lx
pairs of shoes per week. At what price should the manu for b > a has exactly one critical number, namely,
facturer sell the shoes to maximize profit?
51. It is estimated that t years from now the population of a b- a a)
certain country will be
Find C"(t) and use the second derivative test to show that
a relative maximum for the concentration (It) occurs at
time t =tc .
million. When will the population grow most rapidly?
5 .5 INTEGRALS INVOLVING e x AND In x
IN THIS SECTION Integration of exponential and logarithmic functions,

integration involving tanx and cotx
We reverse the differentiation of exponentials and logarithms to find integra
tion formulas and then apply these results to the motion of a particle.
Sec. 5.5 Integrals Involving ex and In x 359
■ INTEGRATION OF e x AND in x
THEOREM 5.15 Integration of ex
J ex dx = ex + C
Proof: This follows directly from the differentiation formula
⅛(<--) - <"
EXAM PLE 1 Integration involving an exponential function
Show that for any nonzero constant k

∣ e kx dx = ⅛e kx + C
Solution Let u = kx, so du = k dx.

[ e kx dx = f e'' Φ = τ [ e" du = 1 e u + C = j- e kx + C
k k k k
EXAM PLE 2 Substitution involving an exponential function
r2 Px' t
Evaluate —7- dt.
Ji Vi
Solution Let u = V t, so that du = ~ r d t and ⅛ = 2 du. Also, if t = 1, then u = 1
2Vt Vt
and if t = 2, u = V 2 , and we have
/-2 vγ ∕-V2 lx/?
dt = e ll[2 du] = 2C' 11, = 2 ( d 2 - c l ) ≈ 2.789 937 1
Ji V t Jι ■■■
THEOREM 5.16 Integration of 1/x
± d x = ln∣x∣ + C
j
Proof: This proof follows directly from Theorem 5.10. Because
^(ln∣x∣) = ⅛ for all x ≠ 0
we have I dx = ln∣x∣ + C _
In Chapter 4, we found ∖x n dx for n ≠ - 1 , and now we know what happens

when n = - 1 , so we restate the complete power rule for integration.
Power Rule y -n+1

, 1 + C n+
{n+ 1
ln∣x∣ + C n = -1
EXAM PLE 3 Integration of a rational function
F ta d ∫ ⅛ ⅛
Solution Let w = l x - 5, so that du = 2 dx and dx = ∣ du.

J 2χd-5 = J ^∖ ~ι~ ~ 1 / = 5 1 π ∣m∣ + C = 5 l∏2χ - 5) + C
If you use a computer software package for integration of a function, such as

the one shown in Example 3, you may find small differences between the
computer output and the result shown in the example. We have already
mentioned that the software does not show the constant of integration. It also
may not show the absolute value symbols. For Example 3, you may see the
computer result
7LN(2X - 5)
2
Compare this with the result ? In ∣2x - 5∣ + C , as obtained in Example 3.
EXAM PLE 4 Integration of a trigonometric function
Find sinx dx
3 + 2 cosx
Solution Let u = 3 + 2 cosx, SO that du = - 2 sinx dx.
sinx dx 5 = - 7 ∙ n M + C = - ∣ ln∣3 + 2 cosx∣ + C

3 + 2 cosx
= —∣ ln(3 + 2 cosx) + C Because 3 + 2 cosx ≥ 0 cz⊂∑∣
In general, it is a good idea to begin work on an integration problem by using

algebra to simplify the integrand as much as possible. This principle is illustrated
in the next example.
EXAM PLE 5 Integration of an exponential to a base other than e
Find J 35 x dx.
Solution We have a formula for integrating exponentials to a base e, so we write
35 χ = e (5 Λ-)in 3 Theorem 5.8f, page 341
35~-v dx = [ c ,5 ~+ln 3 dx = ( = ÷ ⅛ eu + C
J J ∖ln 3/ In 3
Let u = (5 - x) In 3;
du = (-In 3) dx, so dx = J~J -
-1
In 3 In 3
Sec. 5.5 Integrals Involving e x and In x 3 61
■ INTEGRATION INVOLVING tan x AND cot x
We derived integration formulas for sin x and cosx in Chapter 4, and now that we
have the natural logarithm at our disposal, we can also integrate certain expres
sions involving tanx and cotx. The key to these integration problems is the
following theorem.
THEOREM 5.17 Integral of tangent and cotangent
J tanx dx = -ln∣cosx∣ + C j^ cotx dx = ln∣sinx∣ + C
Proof: tanx dx — dx Let u = cosx; then du = -sinx dx.

J J -,v
= [ = -ln∣w∣ + C = -ln∣cosx∣ + C
I u
The formula for the integral of the cotangent may be derived in a similar fashion
(see Problem 62). ■
EXAM PLE 6 Substitution with a cotangent function
Find J cot(l - 2x) dx.
Solution I cot (1 —2x) dx = I cot u ∖- ^ d ι ι ∖ Let u = 1 - lx, then

•1 ' ' " ' du = -2 dx.
= I cotw du = - ∣ ln∣sinw∣ + C = ln∣sin(l - 2x)∣ + C
We complete this section with an applied example.
EXAM PLE 7 The motion of a particle
A particle moves along a coordinate axis in such a way that its acceleration at
time t is α(f) = ( 5 - Z)- 2 . If the velocity of the particle is 2 when t = 4, how far
does the particle move between times t = 0 and t = 4?
Solution Because α(Z) = v'(t), it follows that
v(Z) = ∫ α(Z) dt = ∫ 4 ψ = ∫ z ⅛Γ = + C = (5 - Z)-' + C
f l
Let u = 5 - t;
du = -dt.
Because v(4) = 2, we have 2 = (5 - 4)^^1 + C , which means C = 1.

Next, if s(f) is the position of the particle at time t, then
S , (I) = v(Z) = (5 - Z)~1 + 1
It follows that the total distance traveled by the particle between times t = 0
and t = 4 is given by
f4 f4 ∣4
5,(Z) = lv (0l dt = [(5 - t) 1 + ∖]dt = (-ln∣5 - t∖+ /) u
Jo Jo
Note: v(Z) > 0 on [0, 4]
= —ln∣ 5 - 4∣ + 4 + ln∣ 5 - 0 ∣ - 0 = 4 + l n 5 ≈ 5.609 437 9 _

PROBLEM SET 5 .5
θ Find the given integral in Problems 1-32.
35. I
pin 2 ,
∕( < V - e~x )dx
f 2
36. I (ex - e~x )2 dx

—In 2 2
Q —2x Γ dx 0
1. e 5x dx 2. dx 3. 1
2x + 1 f2
Γ2 e 1''λ dx 38. I 5 ~x dx
37. jI x2
[ 3 dx ς x dx
1
6. I∣ In ex dx
4. 1 2x - 5 J In e rττ∕6 rιr∕3
39. 1 tan 2x dx 40. I cot x esc x dx
0 π∕4
7. (αw2 - M 1 + e u) du, for constant a
f 5 0∙58 dx
r ,2 5,ooo
41. 42. l l + 1 0 X z5
∣0 1 + e~0-2x 0
8. (au4 + bit 1 + ce u) du, for constants a, b, and c
Hint for Problems 41 and 42: Multiply numerator and denom
inator by ex .
9. x exp( 1 - x 2) dx
Find the area o f each region in Problems 43-46.
43. The region under the curve y = ∣(e x - e~x ) above the x-
10. ι (x + 1) exp(x2 + 2x + 3) dx axis such that x ≤ In 2.
44. The region under the curve y = on the interval [ 1, e].
Γ / r2 ∣ 7 4x + 6 ,, ,λ'2^ dx
11. , , + xe x dx 12. /—3---- — λβ
, ∖1 - x / ∣ ∖ V x 2 + 3x 45. The region bounded by the curves y = 2 v and y = 2 1 ~ x
and the y-axis.
f x dx 2
13. ∣
∣ 2x 2 + 3 14. ∣1 xx3 +dx1 46. The region under the curve y = cotx on the interval
47. Find the average value of the function ∕( x ) = e3x on the

f½ ⅛ r fx ∣ 4x dx interval [ - 1 , 4].
15. 1, 16. 1I 2x + 1
x - 5
48. Find the average value of the function f (x) = xe~ x ^ on the
interval [1, 3].
xλ x f X dx
17∙ 1∣ V x e ' dx 18. 1
∣ x 2/3 49. Find the average value of f(x ) = tan x on the interval
[°4 ]∙
19. I∣ 2 3 + " du 20. 1∣ 102z dt © 50. Find f ⅛½ dx. 51. Find f

J x J x In x
52. Find I . . Hint: Rewrite the integrand as τ -γ—τz.
21∙ f ^ d x f In (x + 1) , J ex + 1 b 1+ e x
1 22. I , ι ax
X x + 1 53. A particle moves along the x-axis in such a way that at
time t its velocity (in ft∕s) is
23. Ii dx 24. 1Γ dx
V x (√ x + 7) x 2z3(T x + 1) v(t) = ∣ + t
Γ e l dt Γ e^t dt How far does it move (to the nearest foot) between times
25. ∣ 26. ∣
1 v 7 ( X z + 1) t = 1 and t = e 2 1
et + 1
54. A manufacturer estimates that the marginal cost o f a
2 certain production process is given by
27. 1∣ cosx dx 28. IΓ sec 0 dθ
5 + 2 sin x 1 + tan θ C'(x) = e 0∙0 lx + 3 V x
when x units are produced. If the cost o f producing 4 units
29. 1Γ ∞ t V ⅛ 30. IΓ tan x dx
Vx secx + 1 is $1,000, what does it cost to produce 9 units?
55. The operators of a new computer dating service estimate
31. 1∣ sinx - cosx 32. 1Γ secx tanx dx that the fraction of people who will retain their
sinx + cosx 5 - 3 secx memberships in the service for at least t months is given
by the function
Evaluate the definite integrals given in Problems 33-42. Ap
proximate the answers to Problems 41 and 42 to two significant / / ) = e-t' i0
figures.
There are 8,000 charter members, and the operators
,, f l 5X 2 dx ,4 f 4 e~'⅛ dx expect to attract 200 new members per month. How many
J o 2x 3 + 1 J1 Vx members will the service have 10 months from now?
Sec. 5.6 The Inverse Trigonometric Functions 36 3
56. A national consumer’s association has compiled statistics b. The total number of people who have lived from A.D . 0
suggesting that the fraction of its members who are still to A.D. 2000 is given by the integral
active t months after joining is given by the function r2000
P(t) dt
/ ( 0 = e-°∙2' Jo
Evaluate this integral and determine what percentage of
A new local chapter has 200 charter members and expects
all people who have lived since A.D. 0 will be alive in A.D .
to attract new members at the rate of 10 per month. How
2000.
many members can the chapter expect to have at the end
of 8 months? 61. Show that for each positive integer n
9
57. Find the enclosed area between the curve y = - and the
line x + y = 3. x
χ2
58. Find the area between the curve y = xe and the x-axis by completing steps a-d.
on [-3 , 3]. f 1+ ιz" √∕ fi +ι⅛
2 a. Show that —≤ dt.
Ji , Ji
59. Find the area between the curve ι’ = —; and the x-axis on
[1,5]. ’ X
b. Show that Inf 1 + —) ≥ —γ-r.
∖ n) n + 1
@ 60. Under favorable ecological conditions (that is, plenty of
1 1 i
food and growth space), the population of a biological Hint. 7t ≥ 1 , ι for I ≤ t ≤ I + -
+⅛ «
colony is given by
c. Show that
P = P0 e k '
where Po is the initial population and k is a positive
constant. Suppose the human population of the earth has d. Complete the proof.
followed this pattern since A.D. 0.
62. Derive the integration formula
a. Assuming that the world population was 1 billion in
1825 and 5 billion in 1986, find the population Po at cotx dx = ln∣
sinx∣+ C
A.D. 0 and the constant k.
5 .6 THE INVERSE TRIGONOMETRIC FUNCTIONS
IN THIS SECTION Definition of inverse trigonometric functions, properties

of inverse trigonometric functions, derivatives and integrals involving inverse
trigonometric functions
You were introduced to the inverse trigonometric functions in trigonometry,
but because of their importance in calculus we define and review them in this
section. We also introduce some important derivative and integral formulas
involving these inverse functions.
■ DEFINITION OF INVERSE TRIGONOMETRIC FUNCTIONS
Recall from Section 5.2 that a function f has an inverse only when it is one-to-one
on its domain—that is, when each number y in the range of f corresponds to
exactly one number x in the domain. This condition is satisfied by the natural
logarithm because In x is monotonic on its domain x > 0. The trigonometric
functions are not one-to-one, so their inverses do not exist. However, if we restrict
the domains of the trigonometric functions, then the inverses exist. Thus, we can
define the inverse trigonometric functions if we restrict their domains. In
trigonometry you probably distinguished between the sine curve with unrestricted
domain and the sine curve with restricted domain by writing y = sinx and
y = Sin x, respectively. In calculus, it is not customary to make such a distinction
by using a capital letter.
Let us consider the sine function first. We know that the sine function is
strictly increasing on the closed interval [ - f , f], and if we restrict sinx to this
interval, it does have an inverse, as shown in Figure 5.17.
a. The graph of sin 1x is

obtained by reflecting the
part of the sine on Γ-5, 5
about y = x. L z z.
Figure 5.17 Inverse sine function
Inverse Sine Function y = s in~ l x if and only if x = sinp and -j ≤ y ≤

.................................................................................................................................................................................................................................. . ∣
0 The function sin^'x is NOT Inverses of the other five trigonometric functions may be constructed in a
the reciprocal of sinx. To de similar manner. For example, by restricting tanx to the open interval (-■J, y)
note the reciprocal, write where it is one-to-one, we can define the inverse tangent function as follows.
(sinx)- 1 . 0
Inverse Tangent Function y = tan lx if and only if x = tan y and —j < y
The graph of y = tan l x is shown in Figure 5.18.
Figure 5.18 Graph of the inverse tangent

Sec. 5.6 The Inverse Trigonometric Functions 365
0 It is easier to remember the TABLE 5.1 Definition of Inverse Trigonometric Functions

restrictions on the domain and
the range if you do so in terms Value of x can be
of quadrants, as shown in Table Inverse Function Domain Range Pos Neg Zero
5.1. Note the last column, which
gives the range (or value of the Quadrant
y = sin^ 1x -1 ≤ x ≤ 1 -f ≤ y ≤ f I IV 0
angle y). For example, if ∖_____ v____ 2>
y = sin - 1 x and if x is positive, Quadrants I and IV
then y term inates in Q uadrant I; y = cos^ 1x -1 ≤ x ≤ 1 0 ≤ y ≤ π I II π
2
in other words, 0 ≤ y ≤ f ≈ Quadrants I and II
1.57; on the other hand, if x is y = tan - 1 x — ∞ < x < +∞ -f < y < f I IV 0

negative, then the terminal side Quadrants I and IV
of angle y is in Quadrant IV,
y = sec^ lx x ≥ 1 or x ≤ -1 0 ≤ y ≤ π , y≠ j I II undefined
with -1 .5 7 ≈ --? ≤ y ≤ 0. Fi
Quadrants I and II
nally, if x is 0, then y = 0. 0
y = csc^ l x x ≥ 1 or x ≤ -1 - f ≤ y ≤ f, y ≠ 0 I IV undefined
Quadrants I and IV
y = cot~ lx —∞ < χ < 4-∞ 0 < y < π I II 7T
2
Quadrants I and II
y = C O S Î Λ^ y = cot - 1 x y = sec^l x y = esc l x

-1 < x < 1 All .v .∏ > 1 ∣x ∣≥ 1
0≤y ≤ π 0<y≤π 0 < y ≤ π ,y ≠ ½
Figure 5.19 Graphs of four inverse trigonometric functions*
EXAM PLE 1 Evaluating inverse trigonometric functions
1(— b. sin 1 0.21 c. cos 10 d. tan 1⅛ )

Evaluate a. sin
a. sin 1 = -f, Think: x = is negative, so y is in

Solution
Quadrant TV, the reference angle is the
angle whose sine is it is %, so in
Quadrant IV the angle is
b. sin~ 1 0.21 =≈ 0.21 1 575 0 By calculator, be sure to use radian mode
and inverse sine {not reciprocal).
1
C. COS 0 = y Memorized exact value
d. tan~ 1 (τV) = ? Think: x = is positive, so y is in Quadrant I; the
∖ V 3∕ 6
reference angle is the same as the value o f the
inverse tangent in Quadrant I.
0 The inversion formulas for ’‘There are no standard domain restrictions for inverse secant and inverse cosecant. Other
sin - 1 and ta n - 1 are valid only calculus textbooks may use different restrictions from the ones we have chosen here. However, the ones
on the specified domains. 0 we have used are the ones that are most convenient to use with standard calculators.
■ PROPERTIES OF INVERSE TRIGONOMETRIC FUNCTIONS
The definition of inverse functions yields four formulas, which we call the
inversion formulas for sine and tangent.
Inversion Formulas sin (sin 1x) = x for -1 ≤ x ≤ 1

sin -1 (siny) = y for - f ≤ y ≤ f
tan(tan - 1 x) = x for all x
tan^^l (tany) = y for - f < y < f
EXAMPLE 2 Inversion formula for x inside and outside domain
Evaluate the given functions.

a. sin(sin^ 1 0.5) b. sin(sin~1 2) c. sin -1 (sin 0.5) d. sin - 1 (sin2)
Solution a. sin (sin-1 0.5) = 0.5, because -1 ≤ 0.5 ≤ 1.
b. sin (sin- 1 2) does not exist, because 2 is not between -1 and 1.
c. sin -1 (sin0.5) = 0.5, because ≤ 0.5 ≤ f.
d. -1
sin (sin 2) = 1.1415927, by calculator. For exact values, notice that
(reduction principle) sin 2 = sin(τr - 2) (so that —f ≤ τr - 2 ≤ y)
and we have sin -1 (sin2) = sin-1 [sin(ττ —2)] = ττ - 2.
Some trigonometric identities correspond to inverse trigonometric identities.
For example,
sin -1 (-χ ) = —sin-1 (x) is true
but
0 cos- 1 (-x ) = cos- 1 x IS NOT TRUE 0 ;
try x = 1, for example
EXAMPLE 3 Proving an inverse identity
Show that for - 1 ≤ x ≤ 1, sin- 1 (-x ) = -s in - l x.

Solution Let y = sin - 1 ( - x)
siny = —x Definition of inverse sine
-sin y = x
sin(-y) = x Opposite angle identity
-y = sin - 1 x Definition of inverse sine
y = -s in -1 x
Thus, sin- 1 ( - x) = -s in - l x.
A useful device for remembering inverse trigonometric relationships is the
reference triangle. For example, suppose α = sin - 1 x, so that sinα = x, where
0 < a < We construct a right triangle with an acute angle equal to a, which
Figure 5.20 A reference triangle means we can assume the side opposite has length x and the hypotenuse has length
Sec. 5.6 The Inverse Trigonometric Functions 36 7
1, as shown in Figure 5.20. Then, if β is the angle complementary to a in this

triangle, we have
O_ ADJACENT SIDE = X =
p HYPOTENUSE 1
which means that β = cos 1x. Similarly, because the side adjacent to angle a has
length V l - χ 2 (by the Pythagorean theorem), we have
tan a = „ , y so a = tan 1
VI - x 2 .V1 - x 2.
Note that the relationships displayed in the reference triangle in Figure 5.20
indicate that
cos 1x + sin 1x = y for 0 ≤ x ≤ 1
because β = cos 1x, a = sin 1x and a + β = y in the reference triangle. This

complementary angle relationship can be rewritten as
cos l x = y - sin 1x
which is often used as the definition of the inverse cosine function.
EXAMPLE 4 Using a reference triangle
Find cos (sin 1x), where 0 < sin 1x ≤ y.

Solution Let y = sin 1x, so we want to find cosy. Form a right triangle that has an
acute angle y such that y = sin- l x. Then
γ ∕∣ — γ-2
cos (sin 1x) = cosy = ----
cos (sin 1x) = V 1 - x 2
■ DERIVATIVES AND INTEGRALS INVOLVING INVERSE

TRIGONOMETRIC FUNCTIONS
Differentiation formulas for the six fundamental inverse trigonometric functions

are listed in Theorem 5.18. Notice that these derivatives are not inverse trigono
metric functions or even trigonometric functions, but are instead rational func
tions or roots of rational functions.
THEOREM 5.18 Differentiation formulas for six inverse trigonometric

functions
If u is a differentiable function of x, then
^y- (COS 1U) = / 1 —Φ
⅛ s,n" u*'√ Γ ⅛ ⅛ αxv , v i - u 2 ax
-y-(tan~ 1w) = , 1 Φ ⅛ ∞ t - '" ) = ⅛ ⅛
τ
αxv ’ 1 + u 27 ax
⅛ SEC = ∣
M∣
V U2 - 1 ⅛ ⅛ c s c "'") ^ ∣,,∣v ,, 1' - 1 ⅛
Proof: We shall prove the first formula and leave the others as problems. Let
y = sin - l x, so x = siny. Because the sine is 1-1 and differentiable on [~2> j ] , the
inverse sine is also differentiable. To find its derivative, we proceed implicitly:
siny = x
⅛ sin y ) = ⅛ ω
cos^
dy 1.
⅛ r
Because - f ≤ y ≤ f , cosy ≥ 0, so
dy = ι _ 1 = 1
dx cosy V1 - s in 2 y √1 - x 2
If M is a differentiable function of x , then the chain rule gives
E X A M P L E 5 Derivatives involving inverse trigonometric functions
a. f(x ) = tan^ l V x b. g(t) = sin -1 (1 - 0 c∙ K χ ) ~ sec-1 e 2x
Solution a. Let u = V x in the formula for ⅛ ( s i n ^ ' ^
∕ ,ω = ⅛ ta n -1 vχ)
= _____ 1_____ -d-(Vχ} = - - 1 (i - Q = _______ i ______

1 + (V x) 2 dx ' 1 + x ∖2 V x ) 2 V χ (l + χ)
b. Let ιι = (1 - Z);
^ '( Z ) = ⅛ s in - 1 (l -Z)]
= _______ 1_______ d_< ι _ A = ______ → ______ = -1

V l —(1 — Z)2 dp , V l —(1 — Z)2 V 2 t - z2
c. Let u = e2x ;
Λ'(x) - ⅛ [see" = V i because cl' > ° —

E X A M P L E 6 Graph of an inverse trigonometric function
Sketch the graph of f(x ) = tan^ 1x 2 .

Solution We find that
Λ χ >- T ⅛
,,, , ( I + x 4)(2) - 2Λ (4X ≈) 2 - 6x 4
fr w = — — ^
T h u s ,∕ , (x) = 0 only when x = 0, and because ∕" ( 0 ) > 0, the graph is concave
up, so there is a relative minimum at x = 0. Because/(0) = tan -1 (0) = 0, the
minimum occurs at (0, 0).
Next, note that ∕" ( x ) = 0 when 6x 4 = 2— that is, when x = ± y∣⅛
≈ ±0.7598357. Because ∕" ( x ) < 0 for ∣x∣ > and ∕" ( x ) > 0 for
- < x < the concavity of the graph changes at X = ± and points
of inflection must occur there. The graph o f/ is shown in Figure 5.21.
EXAM PLE 7 Optimization application

∕( x ) = tan - 1 x 2 A painting is hung on a wall in such a way that its upper and lower edges are 10
ft and 7 ft above the floor, respectively. An observer whose eyes are 5 ft above
the floor stands x feet from the wall, as shown in Figure 5.22. How far away
from the wall should the observer stand to maximize the angle subtended by
the painting?
Solution In Figure 5.22, θ is the angle whose vertex occurs at the observer’s eyes
at a point O located x feet from the wall. Note that a is the angle between the
line OA drawn directly from the observer’s eyes to the wall and the line OB
from the eyes to the bottom edge of the painting. In Δ O A B the angle at O is a,
with cot α = j . In Δ O A C , the angle at O is (α + θ), and cot(α + θ) = f . By
using the definition of inverse cotangent we have
θ = (α + θ) —a = cot 1
Figure 5.22 The angle θ is sub

tended by observer’s eye. To maximize θ, we first compute the derivative
~5 + - 2 _2
25 + x 2 4+ x
dθ
Solving the equation = 0 yields
-5(4 + x 2) + 2(25 + x 2) = 0
- 3 x 2 + 30 = 0
x = ±√10
Because distance must be nonnegative, we reject the negative value. We apply
the first-derivative test to show that the positive critical number VTθ
corresponds to a relative maximum (verify). Thus, the angle θ is maximized
when the observer stands V 10 ft away from the wall.
Each of the six derivative formulas for the inverse trigonometric functions has
an integral counterpart, but as shown in the following theorem, we can pare down
the list of integral formulas from six to just three.
THEOREM 5.19 Integrals involving the inverse trigonometric functions
f du sin - 1 ι∕ + C for ∣w∣ < 1

J V l - u2
f du tan - 1 u + C for all u
J 1 + ιι-
dιι
sec- 1 w + C for ∣M∣ > 1
∣w∣Vw2 - 1
Ch. 5 Exponential, Logarithmic, and Inverse Trigonometric Functions
Proof: The proofs are straightforward reversals of the derivatives of the

inverse trigonometric formulas. We might show, however, how two of the
derivative formulas can be pared down to one formula. We know, for instance,
that for - 1 ≤ u ≤ 1
⅛ ≡ i" " " ∙ = √ r ⅛ and ⅛ ° r t >= √ f ⅛
and the corresponding integration formulas are

, du~ - = sin 1w + C and du -c o s l ιι +C
V l - M2 Vl - M2
Clearly, only one of these integration formulas is necessary, and we choose the one
involving sin -1 u. We can even use a trigonometric identity to show these formulas
are identical:
—cos 1M + C = (sin 1M - y) + C = sin l u + C 1
where C 1 = - J + C. A similar reason applies to our choice of the integral formulas

involving ta n - 1 w and sec~ 1M. ■
EXAM PLE 8 An integral involving an inverse trigonometric function
dx
Find
V4 - x r
Solution Because the integral involves V α 2 - x 2 , we think of using an inverse
sine, but first we must do some algebra. We write:
Let u = ¾
du = ⅛ dx.
j_ f 2 d u iu 15
2 J vK V ? sin + C = sin + C
The method used in Example 8 can be applied to obtain the following

integration formulas for the inverse trigonometric functions.
Generalized Integration For a constant a > 0:

Formulas for Inverse
Trigonometric Functions -,~Y L .-.
2
= sin~ 1 -
Vα^ - M «
- ≠ ¼ = lta n -^
α 2 + M2 a a
EXAM PLE 9 Integrating with a generalized inverse trigonometric formula
dx
Find
V9 - 4 V
Solution Compare this with Example 8.

f clx _ f dx , t _ _
_ f | du
J √ 3 2 - ιι2
= 1 s ∙i n” 1 ∕1z√
β∖I + C From the formula with a = 3
= sin” 1 + C
E X A M P L E 10 Integrating an inverse trigonometric form
Find
Solution
du _ 1 _ -ι u
z∕√<r - 4 2 s e c 2
Let u — ex ;
du = ex dχ∙,
, du du
dx = —λ = — .
e u
2 sec ' |2 I + C
E X A M P L E 11 Inverse trigonometric form by completing the square
Find dx
x 2 - 4x + 10'
Solution First, we complete the square in the denominator:
x 2 - 4x + 10 = x 2 - 4x + 22 - 4 + 10
= (x - 2)2 + 6
dx dx _
x λ - 4x + 10 (x - 2)2 + 6
Let u = x - 2;
dx = du.
E X A M P L E 12 Area under a curve
Find the area under the curve y = ------- 1, on the interval [ -1 , 11.
1+ x
Solution The region under the curve is shaded in Figure 5.23. The area is given
by the integral
= [tan^1x ] ∣J 1 = tan^ l l - tan” 1 ( - 1)
_ 77 _ ( _7T )
Figure 5.23 The area under 4 ∖ 4/

y = -f⅛ on [~ 1 > 11 — 2Γ
2
r5T 2 Ch. 5 Exponential, Logarithmic, and Inverse Trigonometric Functions
PROBLEM SET 5 .6
θ 1. ■ What Does This Say? Discuss the restrictions on the Evaluate the definite integrals in Problems 43-48.
domain and range in the definition of the inverse trigono ,-. f ' dx f~''4 sin x√ x f ln 2 v 5 ex dx
4J. λλ 45.
metric functions. 1∕V3 λ'V4χ- 1 j '0 2
1 + cos x ∣o e2→ 4
2. ■ What Does This Say? Discuss the use of reference 1
triangles in relation to the inverse trigonometric func 46. if ______ ⅛ ______ 47 ΓΛ^ x d x 48.
f l x 3 dx
4
∣1 x[4 + (In x) 2] ’ J !0 √ i-x l0 1 + x 8
tions.
Give the exact values fo r functions in Problems 3-14. θ Sim plify each expression in Problems 49-54.
11
49. sin(2 tan - 1 x) 50. tan(2 tan - 1 x)
3. a. cos b. sin - 1
51. tan(cos x) 52. cos(2 sin - 1 x)
- l
4. a. sin - l (- b. cos (-⅛) 53. sin(sin - 1 x + cos - 1 x)
5. a. tan -1 ( - 1) b. cot - l ( -V 3 ) 54. cos(sin - 1 x + cos - 1 x)
l
55. Use reference triangles, if necessary, to justify each o f the
6. a. sec^1 (-V 2 ) b. cot - (-⅛ ) following identities.
7. a. sin -1 (~⅛r) b. sec - ‘ ( - I ) a. cot - 1 x = - tan - l x for all x
8. a. sec" 1 ( ⅛ b. cot - 1 ( - l ) b. sec- 1 x = cos -1 for all ∣x∣ ≥ 1
9. cos (sin - 1 2) -1
10. sin (cos ⅛ c. csc - 1 x = sin -1 - J for all ∣x∣ ≥ 1
-1
11. cot(tan - l j) 12. tan (sin 1) 56. a. Show th a t∕'( x ) = 0 where
8 13. cos (sin- 1 5 + 2 cos - , ⅜)

Hint: Use the addition law for cos(α + β).
14. sin (sin- 1 ∣ + cos-1
∕( x ) = sin lx + cos 1x
What does this prove about /(x)?
b. Show that tan -1 x + cot - 1 x is a constant by examining
Hint: Use the addition law for sin (α + β). its derivative. What is the value of this constant?
57. Find the area under the curve
Find the derivative ÷ in Problems 15-30.
dx y
15. y = sin -1 (2x + 1) 16. y = cos 1 (4x + 3) xVx2 - 1
17. y = tan 1 λ⅛ 2 + 1 18. y = cot - 1 x 2 on the interval (V 2 , 2).
-1
19. y = (sin 2x) 3
20. y = (tan- 1 x 2)4 58. Find the point of inflection of the curve
-1
21. y = V ta n (2x) 22. y = ln∣sin- 1 x∣ y = (x + l)tan -1 x
23. y = ln∣cos x∣- 1 24. y = sin -1 ex
59. Find the critical values for the function and identify each
25. y = tan - l (l) 26. y = cos 1 (sinx) as a relative maximum, relative minimum, or neither.
27. y = sin -1 ∣cosx∣ 28. y = ln[sin - ' (ev)] f ix ) = tan -1 ( - ) - tan -1 (⅜), a > b
∖a l ∖b
29. x sin ly + y tan 1x = x
30. sin - 1 v + y = 2xy Sketch the graph o f each function in Problems 60-63.
60. ∕( x ) = x + tan - 1 x 61. ∕( x ) = x - tan -1 (2x)
Find the integral indicated in Problems 31-42. -1
62. ∕( x ) = cos x + sin x -1 63. ∕( x ) = 2x - sin - 1 x
f dx f dx 64. A worker standing 4 m from a hoist sees the hoist at eye
31. 32. , x V x 2 - 9
' x 2 + 16 level. If the hoist is raised at the rate of 2 m ∕s, how fast is
Γ dx Γ dx the worker’s angle of sight changing when the hoist is 1.5
33. I 34. I
V^5 - 2x 2 V 4 -7 x 2 m above eye level?
f 2λ' + 5 dx
35. I1 7 f ⅜ ~ √ dx
36. I1
x 2 + 4x + 5 x2 + 1
Γ x dx Γ 2 x + 1 d,x
37. ∣
' x2 + x + 1 38. I √1 - x 2
f (2 + x) dx Γ (8x — 1) dx
39. I 2 40. I
I √ 4 -2 x -x V 1 - 4x 2
( dx 1
42. ∣ dx
41. I
' x 2 + 2x + 2 V -x 2 + 3x - 2
Sec. 5.7 An Alternative Approach: The Logarithm As an Integral 37 3
θ 66. Prove f d lt ; = sin 1(-) + C

65. The bottom of an 8-ft-high mural painted on a vertical
wall is 13 ft above the ground. The lens of a camera fixed J V a - u2
2 ∖a)
to a tripod is 4 ft above the ground. How far from the wall 67. Prove f 1 d u 2 = - ta n " 1 6 A + C
should the camera be placed in order to photograph the I a2 + u2 a ∖a j
mural with the largest possible angle?
68. Prove I — , df 7 = - sec
1 u
+ C
J ι∣V u 2 - a 2 a a
69. Use the identity cot 1x = •? - tan 'x for all x to find
4~ (cot^lx).
axv ’
70. Use the identity sec- 1 x = cos^ 1 ( - ) for ∣x∣ ≥ 1 to show
that W
4χ (sec 1x) = , , A2— = for ∣x∣ > 1

d ∣x∣V x - 1
5 .7 AN ALTERNATIVE APPROACH: THE LOGARITHM AS AN INTEGRAL*
IN THIS SECTION Definition of the natural logarithm as an integral,

geometric interpretation of the natural logarithm, the natural exponential
function
“ When I use a word,” Humpty Charles Dodgson, whose pen name was Lewis Carroll, was a mathematician
Dumpty said, “ it means just what I who knew that a standard procedure in mathematics is to define a function by
choose it to mean— nothing more “ fitting” it to a list of desired characteristics. In this optional section, we
or less.” explore an alternative treatment of the natural logarithm and exponential
Lewis Carroll (1832-1898)
functions in which In x is defined as an integral and its properties are derived by
referring to properties of integrals. The natural exponential function ex can
then be defined as the inverse of In x.
■ DEFINITION OF THE NATURAL LOGARITHM AS AN

INTEGRAL
You may have noticed in Section 5.1 that we did not prove the properties of
exponential functions for all real number exponents. In order to treat exponentials
and logarithms rigorously, we use the alternative approach provided in this
section. Specifically, we use a definite integral to introduce the natural logarithmic
function and then use this function to define the natural exponential function. In
this way we can develop precise meanings for bx and logfrx.
Natural Logarithm The natural logarithm is the function defined by
At first glance, it appears there is nothing “ natural” about this definition, but
if this integral function has the properties of a logarithm, why should we not call it
a logarithm? We begin with a theorem that shows that In x does indeed have the
properties we would expect of a logarithm.
The material of this section is not required for subsequent sections.

x dt
Historical Note THEOREM 5.20 Properties of a logarithm as defined by In x = . t
Let x > 0 and y > 0 be positive numbers. Then

a. In 1 = 0
b. In xy = In x + In y
c. In - = In x - In y
y
d. In x p = p In x for all rational numbers p
Proof: a. Let x = 1. Then y = 0.
b. For fixed positive numbers x and y, we use the additive property of
integrals as follows:
ln(xy) = Γ ⅞ = Γ ⅞ +
LEONHARD EULER (1707-1783) Ji r JI F Let u = -., so t = uχ∙,
Euler’s name is attached to al r dt = x du.
most every branch of mathe = In x + x du If t = x, then u = 1;
matics. He was the most prolif
, ιιx if t = xy, then u = y.
ic writer on the subject of = In x + In y
mathematics, and his mathe
matical textbooks were master c. and d. The proofs are outlined in the problem set.
fully written. His writing was
not at all slowed down by his
total blindness for the last 17 ■ GEOMETRIC INTERPRETATION OF THE NATURAL
years of his life. He possessed a LOGARITHM
phenomenal memory, had al
most total recall, and could An advantage of defining the logarithm by the integral formula is that calculus can
mentally calculate long and be used to study the properties of In x from the beginning. For example, note that
complicated problems. if x > 1, the integral
When Euler was 13 he regis
tered at the University of Basel In x = (
and was introduced to another Ji l
famous mathematician, John
Bernoulli, who was an instruc may be interpreted geometrically as the area under the graph of y = γ from t = 1
tor there at the time. If Ber to t = x, as shown in Figure 5.24. Because the area shown in Figure 5.24 is
noulli thought that a student positive, we have In x > 0 if x > 1. If 0 < x < 1, then
was promising, he would pro
In x = [ ⅛ = ~ [ v < 0
vide, sometimes gratis, private
instruction. Here is Euler’s own Jl Jx
account of this first encounter
with Bernoulli:
‘7 soon found an opportunity to
gain introduction to the famous
professor John Bernoulli, whose
good pleasure it was to advance
me further in the mathematical
sciences. True, because o f his
business he flatly refused me
private lessons, but he gave me
much wiser advice, namely, to
get some more difficult mathe
matical books and work through
them with all industry, and Figure 5.24 If x > 1, In x = ∣y is the area under y = ∣on [1, x].
Ji t t
Sec. 5.7 An Alternative Approach: The Logarithm As an Integral 375
so that
wherever I should find some
check or difficulties, he gave me In x > 0 if x > 1
free access to him every Satur In 1 = 0
day afternoon and was so kind
as to elucidate all difficulties, In x < 0 if 0 < x < 1
which happened with such great
ly desired advantage that when The definition In x = f y makes it easy to differentiate In x. Recall from
ever he had obviated one check It L
for me, because o f that ten oth Section 4.3 that according to the second fundamental theorem of calculus, if /is
ers disappeared right away, continuous on [a, b∖, then .x
which is certainly the way to
make a happy advance in the
F(x) = / ( / dt
ja
mathematics sciences.” HP
is a differentiable function of x with derivative y : = ∕(x ) on any interval [a, x].
From Elements o f Algebra by x
Leonhard Euler Therefore, because γ is continuous for all t > 0, it follows that In x = ' ⅞ is
Ji L
I
differentiable for all x > 0 with derivative x) = - , and by applying the chain
rule, we find that
dxv ’ u ax
for any differentiable function u of x with ιι > 0.
To analyze the graph o f∕(x ) = In x, we use the curve-sketching methods of
Chapter 3:
a. In x is continuous for all x > 0 (because it is differentiable), so its graph is
“unbroken.”
b. The graph of In x is always rising, because the derivative
4~ v(in x) = -
dx ’ x
is positive for x > 0. (Recall that the natural logarithm is defined only for
x > 0.)
c. The graph of In x is concave down, because the second derivative
⅛ (2vl n x ) = y - M
t∕x > dx ∖ x
= Ξ42
x
is negative for all x > 0.
2
d. Note that In 2 > 0 (because r dyt > 0) and because In 2p = p In 2, it
follows that lim In 2p = +∞. But the graph of∕(x ) = In x is always rising,
p →+∞
and thus lim In x = +∞. Similarly, it can be shown that lim In x = -∞.
e. I f b is any positive number, there is exactly one number a such that In a = b
(because the graph of In x is always rising for x > 0). In particular, we
define x = e as the unique number that satisfies In x = 1.
Figure 5.25 The graph of the nat
ural logarithm function, In x These features are shown in Figure 5.25.
■ THE NATURAL EXPONENTIAL FUNCTION
Originally, we introduced the natural exponential function ex and then defined the
natural logarithm In x as the inverse of ex . In this alternative approach, we note
that because the natural logarithm is an increasing function, it must be one-to-one.
Therefore it has an inverse function, which we denote by E(x).
Because In x and E(x) are inverses of one another, we have

E(x) = y if and only if In y = x
From the definition of inverse formulas we have
E,(ln x) = x and ln[E,(x)] = x
We call these formulas the inversion formulas. Therefore,
E(0) = jE(ln 1) Because In 1 = 0
= 1 Because £(in x) = x
and
£'( 1) = E,(ln e) Because In e = 1 (see e above)
=e Because E(ln <?) = e
To obtain the graph of E(x), we re∩ect the graph of y = In x in the line y = x.
The graph is shown in Figure 5.26. Notice that In x = 1 has the unique solution
x = e.
Figure 5.26 The graph of E(x) is the reflection of the graph of In x in the line y = x.
The following algebraic properties of the natural exponential function can be
obtained by using the properties of the natural logarithm given in Theorem 5.20
along with the inversion formulas.
THEOREM 5.21 Properties of the exponential as defined by E(x)

For any numbers x and y,
a. E(x + y) = E(x)E(y)
κ er ∖ _ ∙fe (χ )
b ∙ E (x E(y)
c. [E(x)]p = E(px)
Proof: a. We use the fact that In AB = In A + In B to show that
ln[∕f(x)E(y)] = In E(x) + In E(y) Property of logarithms
= X+ y Inversion formula
= In E(x + y) Inversion formula
Because In x is a one-to-one function, we conclude that
£(x)£(j) = E(x + y)
b. and c. The proofs are similar and are left as exercises. ■
Sec. 5.7 An Alternative Approach: The Logarithm As an Integral 377
Next, we use implicit differentiation along with the differentiation formula

4 - γ(In x) = i
dx ’ x
to obtain a differentiation formula for E(x).
THEOREM 5.22 Derivative of E(x)

The function defined by F(x) is differentiable and
f =£ W
Proof: We know from Theorem 5.4 that E is differentiable because it is the
inverse of the differentiable function In x. If y = F(x), we know that
In y = ln[E(x)]
In y = x Inversion formula
1⅛ = 1 Differentiate implicitly.
y dx
dy
dx ~ -v
d dy dE and
,
ydrx = £ (vx )’ Because y r = yr^ E(x) = y
Finally, we observe that if r is a rational number, then

ln(er) = r In e = r
so that
E(r) = e r
This means that E(x) = ex when x is a rational number, and we define E(x) to be ex
for irrational x as well. In particular, we now have an alternative definition for e:
E(l) = e' = e.
PROBLEM SET 5 .7
θ 1. Use the integral definition vals should be used in Simpson’s rule to estimate In 3 with
an error not greater than 0.00005?
4. Prove the quotient rule for logarithms,
In = In x - In y
to show that In x → —∞ as x → 0 + . Hint: What happens to
ln(2~n ) as N grows large without bound?
for all x > 0, y > 0. Hint: Note that the product rule for
2. Use Simpson’s rule with n = 8 subintervals to estimate logarithms implies
In I - y = In( - ) + In y
L ∖ yf J U7
Compare your estimate with the value of In 3 obtained 5. Show that !n(x, ') = p In x for x > 0 and all rational
from your calculator. exponents p by completing the following steps.
3. Use the error estimate for Simpson’s rule to determine the a. Let F(x) = ln(xp ) and G(x) = p In x. Show that F'(x) =
accuracy of the estimate in Problem 2. How many subinter G'(x) for x > 0. Conclude that F(x) = G(x) + C.
b. Let x = 1 and conclude that d. If∕'( x ) is defined for each x ≠ 0, show that
F(x) = G(Λ) ∕'( 1 )
That is, that ln(x,') = p In x.
Then show that
6. Use Rolle’s theorem to show that
∕W = ∕ ' ( l ) ∫ ' f fo rx > 0
In M = In N
if and only if M = N. Hint. Show that if M ≠ N, Rolle’s
e. Conclude that any solution of
theorem implies that
f(χy) = f(,χ) + ∕( τ )
⅛ < h v )-0
that is not identically 0 and has a derivative for all x ≠ 0
must be a multiple of
for some number c between M and N. Why is this impossi
ble? F(x) = f l T for x ≠ 0
7. The product rule for logarithms states Ji ‘
logft(Λ∕7V) = log,, 37 + log/;7V 8. Show that for each number A, there is only one number x
for which In x = A. Hint: If not, then In x = In y = A for
For this reason, we want the natural logarithm function x ≠ y. Why is this impossible?
In x to satisfy the functional equation 9. By comparing areas under the curve
f(xy) = f(x) + f(y)
Suppose ∕( x ) is a function that satisfies this equation
throughout its domain D. between t = 2 and t = 3, show that
a. Show that i f ∕( l ) is defined, then /(1) = 0.
b. Show that i f /(-1 ) is defined, then /(-1 ) = 0. In 2 < 1 < In 3
c. Show that ∕(-.γ) = f(x) for all x in D. Why does this imply that 2 < e < 3?
Chapter 5 Review

Concept Problems 7. What is the formula for the derivative of an inverse

1. What is an exponential function? function?
2. What is e? 8. a. What is a logarithm?
3. What is the future value formula for interest b. What is a common logarithm?
a. compounded n times a year? c. What is a natural logarithm?
b. compounded continuously? 9. State the change of base theorem.
4. Define an inverse function. 10. State the derivative formulas for the given functions.
5. What is the horizontal line test? a. In u b. log4w c. e" d. bu
6. What is the procedure for graphing the inverse of a given 11. Outline the procedure for logarithmic differentiation.
function?
12. State the integration formulas for the given functions, endpoint behavior (for example, as x → +∞ and
a. ex b. x n c. tanx d. cotx x → -∞).
13. Define the six inverse trigonometric functions; include 21. Calculate the indicated integral.
the domain, range, and graph of each. -In 2
14. State the inversion formulas for the inverse trigonomet

ric functions.
b. ∣
J o
(e2x - e~ 2x ) dx
ln x d x c. J t a n 2 x √ x
15. State the differentiation formulas for the six inverse
trigonometric functions.
22. Evaluate each of the given limits.
16. State integration formulas involving the inverse trigono
metric functions. a. lim+ (1 + x) 4/x
x →0
Practice Problems b. lim (-r -J— ) c. lim x tan x

j →∣ ∖1 - x ) λ-→o÷
17. Find 23. Find the area of the region above the x-axis and under
ax
y = e2x - ex on [0, 1].
a. y = x 2e~'^x b. y = !n 7x
In 3x 24. How quickly will $2,000 grow to $5,000 when invested at
an annual interest rate of 8% if interest is compounded
18. Find a. quarterly? b. monthly? c. continuously?
dx
a. y = sin^^l (3x + 2) b. y = tan -1 2x 25. A manufacturer can produce cameras at a cost o f $40
19. Use logarithmic differentiation to find y ' if each and estimates that if they are sold for p dollars
each, the consumers will buy approximately D(p) =
_ ln(x 2 - 1)
y 800e"0∙t,lp cameras per week. At what price (to the
^ x ( l - 3x)3 '
nearest dollar) should the manufacturer sell the cameras
20. Graph y = (x 2 - 3)e"jc. Be sure to show high and low to maximize profit?
points, points of inflection, intercepts, asymptotes, and
1. Evaluate each of the given numbers. 12. Find f 1(x) i f f( x ) = *1∣∙ What is the domain o f f ‘?
a. In 4.5 b. e2∙8 c. tan -1 2
2. Evaluate each of the given numbers. dy
Find the derivative ÷ in Problems 13-27.
dx
a. e ln 7 ∙4 b. In e 3∙7 c. sin " 1 (cos-j¾) 13. y = exp(2x2 + 5x - 3) 14. y = ln(x 2 - 1)
3. Evaluate each of the given numbers. 15. y = x3 2 " x 16. y = log3(x 2 - 1)
a. sin (2 ta n " 1 3) b. sin icos j + sin " 1 -⅛) 17. exy + 2 = In - 18. y = V x sin " 1 (3x + 2)
X
Solve each equation in Problems 4 -7 for x. 19. y = ^ s in x
20. y = 2 x log,x
4. a. 4x ^ 1 = 8 b. 2χ2+4x = ⅛ 21. y = e~x log5 3x 22. x2∙r + y 2 x = 3
5. a. ln(x - 1 ) + ln(x + 1 ) = 2 lnV12 sin " l x
23. ln(x + y 2) = x 2 + 2y 24. y =
x + ta n " 1x
b. V x = cos" 1 0.317 + sin - 1 0.317
-x tan " 1x
6. a. log,2 χ2 = 4 b. log4V x (x - 15) = 1 25. j 26. y = e" x V ln 2x
sin l x x
7. a. log2x + log,(x — 15) = 4 b. 32Λ'^1 = 6 v3 1"
27. y = (sinx)(sin 1x) + x cot 1x
In Problems 8-9 , fin d f 1 i f it exists.
8. a. f(x ) = 2x 3 - 7 b. ∕( x ) = ~ flx + 1, x ≥ Find the standard form equation for the tangent line to the
given curve at the prescribed point in Problems 28-32.
9. a. f(x ) = ∖ , ex - 1, x ≥ 0 b. f( x ) = + ∣'x ≠ 7
28. y = x In ex where x = 1
10. Show that for any a Ψ 1, the function ∕( x ) = * * is its 29. y = x e 2x ~' where x = ∣
own inverse. 30. ex y = x - y at (1, 0)
11. Letj f(x)
v ’
= a x t j . Findj ∕ ^ 1 in terms of a, b, c, and d. 31. y = (1 - X )Λ where x = 0
ex + d
Under what conditions does ∕ " 1 exist? 32. y = 2Λ — log, x where x = 1
In Problems 6 9 - 72, fin d the area o f the region under each curve
Use logarithmic differentiation to fin d in Problems 33-38. and above the x-a xis on the prescribed interval.
52 f i x 2 - x ) 3 69. y = 2 ^ ∕[ 0 , 1] 70. y = il⅛ [1. e 2 ]
33.
(2x 3 - 3x) 4
(x 2 + 3) ιz3V x 4 + x + 1 71. y = ta n x , Γθ, f l 72. v = ----- [0, 1]

34. l 1 ÷ e~ x
"Vzx 3 + 5
73. L e t∕' ( x ) = (x 2 + bx + b)e~ x for constant b. Find all criti
35. y = (x 2 + l) 2v 36. y = (In x) λ'
cal numbers o f f Determine the relative maxima and
37. y = (3x 2 + 2)2-x 38. y = x s i n x minima o f / in terms o f b.
Sketch the graph o f each function given in Problems 3 9-42. Be 74. Find the average value o f the function ∕( x ) = on the
sure to show high and low points, points o f inflection, inter interval [1, e].
cepts, asymptotes, and endpoint behavior (as x → +∞ and 75. How much should you invest now at an annual interest
X →, -oo). rate o f 6.25% so that your balance 10 years from now will
39. y = x 2 ln∖ zx 40. y = sin - 1 x - cos - 1 x be $2,000 if interest is compounded
41. ∕( x ) = l ∏ f e f ) 42. ∕( x ) = x e i 'x a. quarterly? b. continuously?
76. A lighthouse is 4,000 ft from a straight shore. Watching
the beam on the shore from the point P on the shore that
Calculate the indicated integral in Problems 43-64. is closest to the lighthouse, an observer notes that the light
, λ' ÷ 1 dx χ2 is moving at the rate o f 3 ft/s when it is 1000 ft from P.
43. 44. xe dx
How fast is the light revolving at this instant?
f dx 77. A light is 4 mi from a straight shore line. The light
45. 1∣ 7 +d xx 2 46. I V 2 5 - 9 X 2
revolves at the rate o f 2 rev/min. Find the speed o f the
fπ∕2 ∙ 7 r l∏2 spot o f light along the shore when the light spot is 2 mi
sm x d x
47. ι 48. ∣ (e 2 x - e~ 2 x ) d x past the point on the shore closest to the source o f light.
∣ 2 + co sx 0 JQ -∣- y ∖
49. ι
rln 2
(e 2x + 5)ex dx 50. ,
r, 7r∕6
sec 2x d x
(
7----- — for
i - χ y∕
'o 'o x y ≠ 1 whenever —f < tan^ l ( -√ + l' ∣ < f . Thenestab-
f e'''x dx
51. 52. ι1 sin 2x d2x lish the following equations.
∖∕x ' 1 + sin x
a. tan - 1 (⅜) + tan - 1 (∣) = f
f e 2x d x f (x 2 + x + 1) d x
53. 1 x 54. 1 b. 2 t a n - 1 ( ∣ ) + t a n - 1 ( ∣ ) = f
∣ e + 1 x2 + 1
Γ co sx - sin x Γ sec 2x d x
55. ∣ 56. 1' 1 + tan x c. 4 t a n - 1 (∣) - t a n - 1 (5 ⅛) = f
' co sx + sin x
2 Γ (2 + x) d x Note: The identity in part c will be used in Chapter 8 to
57. IΓ 1 + V l - χ2 ,
58. 1 estimate the value o f π.
Vl - x 1+ x2
79. Use the identity co t - l x = f - tan^ 1x for all x to find
59. 1 co sx dx
2
60. I x dx ⅛ ( c o t ^1 ^ ∙
1 - sin x x4 + 1
2
61. 1Γ x d x 62. 1' ⅛ ⅜ ∙ W < 1 80. Use the identity sec^ 1x = cos^ 1 - for ∣x∣ > 1 to verify
√1 + x 6
dx r d x , ∣Y∣ 1
dχ ∣x ∣V x 2 - 1
Evaluate each o f the lim its in Problems 65-68.
65. a. lim (1 + x) 4z* b. lim — ) 81. Spy Problem The spy has finally escaped from the angry
1 → o* Λ∙→ I ∖ 1 -X∕ members o f the Cam el Lovers Club o f San Dim as (Prob
lem 45, Section 4.4) only to discover that his best friend,
66. a. lim , x ta n x b. lim ------- -
.v→0 .v→0 x Sigmund (“ Siggy” ) Leiter, has been found murdered in a
6 7 . a , lin , !⅛ i± l> b , lin , fιY freezer where the air temperature is 10 oF. The spy
x→0 X χ → +∞ ∖X J remembers that t hours after death, a body has tempera
e r c o s ∙γ 1
ture
68. a. lim (4 - ⅛ )v b. lim ~
λ χ→+∞ × ∙×
τ = T a + (T d - τ a )(0.97)'
where Ta is the air temperature and Td is the temperature

o f the body at the time of death. He is told that Siggy’s Computational Window
body temperature was 40 0F when it was discovered at
1 P.M. on Thursday, and he knows that the dark deed was
Use a computer or the Newton-Raphson method to esti
done by either Boldfinger or Andre Scelerat. If Scelerat
mate a solution correct to the nearest hundredth for the
was in jail from Monday until noon on Wednesday and
equations in Problems 89-90.
Boldfinger was at the annual Villain’s Conference from
noon Wednesday until Friday, who “ iced” Siggy and 89. 2jc = -
x
when? (You may assume Siggy’s body temperature at the
90. sin - l (x - 1) = tan - 1 x
time of death was a normal 98.6 oF.)
91. Use a computer and the trapezoidal rule with n = 8
82. A manufacturer of light bulbs estimates that the fraction
to estimate
F(t) of bulbs that remain burning after l weeks is given by
e^χ2 dx
F(f) = e~kt
92. The prime number theorem states that the number of
where k is a positive constant. Suppose twice as many
primes between a and b (for 0 < a < ⅛) is approxi
bulbs are burning after 5 wk as after 9 wk.
mately equal to
a. Find k and determine the fraction of bulbs still burn
ing after 7 wk.
b. What fraction of the bulbs burn out before 10 wk?
c. What fraction of the bulbs can be expected to burn out Use a computer and the trapezoidal rule with n = 20
between the 4th and 5th weeks? to estimate the number of primes between
83. First National Bank pays 7% interest compounded a = 50,000 and b = 60,000.
monthly and Fells Cargo Bank pays 6.95% interest com
pounded continuously. Which bank offers the better deal?
84. Let k be a fixed positive number, and suppose A(l) is the
93. P U T N A M E X A M IN A T IO N P RO BLEM
A heavy object is at
area under the curve y = e~ k x and above the x-axis on the
interval [0, t]. Find lim H(t). tached to end A of a light rod A B of length a. The rod is
z→+∞ hinged at B so that it can turn freely in a vertical plane. If
85. Show that for any positive number N
the rod is balanced in the vertical position above the
lim x (N l ,x - 1) = In N hinge and then slightly disturbed, show that the time
required for the rod to pass from the horizontal position
86. a. Use the definition of the definite integral to show that to the lowest position is
lim (e i "' + e2jn + . . . + e"", )(i ) = e - 1
n→+∞ W∕ T =
b. Evaluate lim (-j—}---- F----- F ∙∙∙ 4-----2 — ). 94. P U T N A M E X A M IN A T IO N P ROBLEM Which is greater
,,→+ ∞ ∖1 + n 2 + « n + nJ
, - / 1 , 2 , , n (v π )V∏+∣ 0r (V7Γ-ΓT)v "
c. Evaluate
π →+∞ ∖n~ + 1 n2 + 4 n~ + n 2
87. Show that the one-sided limits
where n > 8? Hint: Use the function ∕( x ) = and
Jim1+ 12 ι⅛- and J⅞- ι + 21⅛ show that x y > y x when e ≤ x < y.
95. P U T N A M E X A M IN A T IO N P RO BLEM Show that
both exist but are not equal.
88. T H IN K T A N K P ROBLEM Each of the following equations 1
ln( 1 + -
may be either true or false. In each case, either show that 1+ x
the equation is generally true or find a counterexample.
The graph of the
a. tan~1x = s 'n x
1x b. e l , x = e~ l ,x for x > 1 equation x y = y v for x > 0, y > 0 consists of a straight
cos
line and a curve. Find the coordinates of the point where
c. tan - 1 x = —-— d. cot - l x = ? - tan - 1 x the line and the curve intersect.
tan x 2
e. cos(sin~'x) = V I - x2 f. sec- 1 = cos- l x
GROUP RESEARCH PROJECT
"Quality Control”
This project is to be done in groups o f three or four students. Each group will submit
I own a plant that manufactures disk drives. My problem is to decide how many
hours of labor to put into the drive. The minimum labor time required to get a
drive is 5 hours, but a 5-hour drive will fail at some point during the warranty
period with a probability of 10%. Any extra labor time will reduce this failure
probability at an exponential rate: Specifically, each additional hour of labor cuts
the failure probability by 40%. A failed disk drive is returned to us for repair,
requiring an average of 2 additional hours of labor; in addition it incurs an
estimated cost of $200 for handling and damage to the reputation of the product.
Suppose that the labor cost is $7 per hour.
Your paper is not limited to the following questions but should include these
concerns: How many hours of labor should be put into each machine to minimize
The calculus is the greatest aid we my cost per machine? Suppose that hourly labor cost is w dollars instead of $7 per
have to the appreciation o f physical hour; answer the same question. The labor cost w is a parameter, and the optimum
truth in the broadest sense o f the labor time depends on w. Exactly what is the nature of this dependence? For what
word. w should we go with the basic 5-hour product, and for what w should this be
W. F. Osgood
Bulletin American Mathematical
increased, and by how much?
Society.
It is certainly true that all physical

phenomena are subject to strictly
mathematical conditions, and
mathematical processes are unas
sailable in themselves. The trouble
arises from the data employed.
Most phenomena are so highly
complex that one can never be
quite sure that he is dealing with
all the factors until the experiment
proves it. So that experiment is
rather the criterion o f mathemati
cal conclusions and must lead the
way.
X E. Dolbear
Matter, Ether, Motion (Boston,
1894), p. 89.
*This problem is from a set of course notes for an introductory calculus course taught by Peter D.
Taylor of Queen’s University, Kingston, Ontario.
6
Additional Applications
of the Integral
■ CONTENTS PREVIEW
6.1 Volume: Disks, Washers, and
Shells In Chapter 4, we found that area and average value can be expressed in
Method of cross sections;
Volumes of revolution: disks terms of the definite integral. The goal of this chapter is to consider
and washers; The method of cy various other applications of integration such as computing volume, arc
lindrical shells
length, surface area, work, hydrostatic force, and centroids of planar
6.2 Arc Length and Surface Area regions. We shall also see how separable differential equations can be
The arc length of a graph; The
area of a surface of revolution used to study growth and decay phenomena such as population models
6.3 Physical Applications: Work, and carbon dating. Finally, we shall introduce first-order linear differen
Liquid Force, and Centroids tial equations and show how such equations can be used to study
Work; Fluid pressure and force;
Centroid and moment of a dilution problems and current in an electrical circuit.
plane region; Volume theorem
of Pappus
6.4 Growth, Decay, and First-Order
Linear Differential Equations
Radioactive decay: carbon PERSPECTIVE
dating; Exponential and logistic
growth; First-order linear differ What is the volume of the material that remains when a hole of radius of 0.5 J? is
ential equations; Applications drilled into a sphere of radius 7?? The “center” of a rectangle is the point where
of first-order linear equations its diagonals intersect, but what is the center of a quarter circle? If a reservoir is
Chapter 6 Review filled to the top of a dam shaped like a parabola, what is the total force of the
water on the face of the dam? If a worker is carrying a bag of sand up a ladder
and the bag is leaking sand through a hole in such a way that all the sand is gone
when the worker reaches the top, how much work is done by the worker? If a
closed electrical circuit has a resistor of 8 ohms, an inductor with 2 henries
inductance, and an EMF of 5 V, what is the current in the circuit at time t?
These questions are typical of the kind we shall answer in this chapter.
6.1 VOLUME: DISKS, WASHERS, AND SHELLS
IN THIS SECTION Method of cross sections, volumes of revolution: disks

and washers, method of cylindrical shells, summary
We have seen how to compute area by integration, and our first goal in this
chapter is to examine several methods for computing volume.
383
384 Ch. 6 Additional Applications of the Integral
■ METHOD OF CROSS SECTIONS
A number describing the three-dimensional extent of a set is called volume and is

measured in cubic units. We say that a cube with side 1 has unit volume. Table 6.1
reviews some of the common solids whose volume formulas you may remember
from precalculus.
TABLE 6.1 Volume Formulas

Name of Solid Characteristics Volume Formula Picture
Cube Side 5 V = s3
Sphere Radius r V = fτr r 3

r'-
Right circular Height h
Z
cone Circular base of Y = j π r 2h
radius r
Regular A pyramid formed

tetrahedron by 4 equilateral V = ⅛ √2α3 a
triangles, each of
side a
Cylinder Height h
A
Area of base B V = Bh
Λ
Circular cylinder Height h

Circular base of V = ττr2h
radius r
Λ
A right cylinder can be regarded as a number of congruent disks stacked in a

b. Top of a single vertical pile, as shown in Figure 6.la, and its volume can be computed by taking
coin (a cross the product of the common cross-sectional area B and the height h of the stack.
section) has
area B. The For example, a right circular cylinder of height h and radius r has circular cross
volume of the sections of area B = ττr2, and its volume is V = Bh = πr 2h.
stack of coins is Bh. A similar approach can be used to find the volume of other solids whose cross
a. Stack of coins sections are known. However, when the cross-sectional areas are not constant, it
with height h may be necessary to use calculus.
Figure 6.1 The volume of a cylin Let S be a solid and suppose that for a ≤ x ≤ b, the cross section of 5 that is
der perpendicular to the x-axis at x has area A(x). Think of cutting the solid with a
Sec. 6.1 Volume: Disks, Washers, and Shells 385
knife and removing a very thin slab whose face has area A(x) and whose thickness
is ∆x, as shown in Figure 6.2.
To find the volume of S, we first take a partition
P = {x0 , x 1, . . . , x π} of the interval [a, b∖
and choose a representative number x* in each subinterval [xfc-1 , X A.]∙
Next, we consider a cylindrical slab with width
a. Horizontal cross-section ∖ χ k = χ
k ~ χ
k - i
and constant cross-sectional area √4(x⅛). This slab has volume

ΔK fc = ^ Δ x fe
and by adding up the volumes of all such slabs, we obtain an approximation to the
volume of the solid 5: «
Δ Γ ≈ j Σ Λ(X I)Δ¾
The approximation improves as the number of partition points increases, and it is
b. Vertical cross-section
reasonable to define the volume V of the solid 5 as the limit of Δ Kas the norm of
the partition ∣∣P∣∣ tends to 0. That is,
Figure 6.2 Volumes of solids with
known cross sections V = lim ∑ A(x*.') ∆x,k
∣∣p∣∣→o ⅛≡i v v
which we recognize as the definite integral A(x) dx. To summarize:

JCl
Volume of a Solid with Known A solid S with cross-sectional area A(x) at each point perpendicular to the
Cross-Sectional Area x-axis on the interval [a, /?] has volume
rd
V= A(x) dx
EXAM PLE 1 Volume of a solid using square cross sections
The base of a solid is the region in the xy-plane bounded by the j-axis and the
lines y = 1 - x, y = 2x + 5, and x = 3. Each cross section perpendicular to the
x-axis is a square. Find the volume of the solid.
Solution The solid resembles a tapered brick, and it may be constructed by
“ gluing” together a number of thin slabs with square cross sections, like the
one shown in Figure 6.3. We begin by drawing the base in two dimensions and
then find the volume of the Zrth slice.
∆x
c. Cross-sectional representative element:

Δ V = L 2 ΔX
a. Two-dimensional graph of the

given base
Figure 6.3 A solid with a square cross section
To model this construction mathematically, we subdivide the interval

[0, 3], form a vertical approximating rectangle on each resulting subinterval,
and then construct a slab with square cross section on each approximating
rectangle. If we choose the width of a typical slab to be ΔΛ' and the height to be
L , the slab will have volume Δ V, where

Δ V = L 2 Δx
= L 2(x) ∆x
= [(2x + 5) - (1 - x ) ] 2 ∆x
= (3x + 4)2 ∆x
The volume of the entire solid is obtained by integrating to “ add up” all the
volumes Δ F, and we find that the volume of the solid is
r3
V= (3x + 4)2 dx
Jo
= [ (9x 2 + 24x + 16) dx = (3x3 + 12x2 + 16x)∣∩ = 237
Jo '0 ≡≡≡
E X A M P L E 2 Volume of a regular pyramid with a square base
A regular pyramid has a square base of side L and has its apex located H units
above the center of its base. Derive the formula V = ¼HL2 .
Solution The pyramid is shown in Figure 6.4. This pyramid can be constructed
by stacking a number of thin square slabs. Suppose that a representative slab
has side £ and thickness ∆∕z and that it is located h units above the base of the
pyramid as shown in Figure 6.4. By creating a proportion from corresponding
parts of similar triangles, we see that
£ _ H -h so that £ = L ( 1- ⅛
L H ∖ 'Λ
Therefore, the volume of the representative slab is
Δ K = £2 A/z = L 2 ( l -⅛ Y ∆ ∕v
∖ Γ7∕
To compute the volume V of the entire pyramid, we integrate with respect to h
from the base of the pyramid (A = 0) to the apex (A = H ). Thus,
V = [ " L 2 ( 1 - ⅛ Y dh = L 2 ∏ 1 - ⅜A + ⅛ A 21 dh
Jo ∖ h ) JO L -H H- J
Figure 6.4 The volume of a ∖ H + 3H 2) ∖o H 3∕√ 2∕ «

pyramid
Other volume formulas (including those in Table 6.1) may be found in a
similar fashion (see Problems 21-28).
■ VOLUMES OF REVOLUTION: DISKS AND WASHERS
A solid of revolution is a solid figure 5 obtained by revolving a region R in the xy

plane about a line £ (called the axis of revolution) that lies outside R . Note that
such a solid S may be thought of as having circular cross sections in the direction
perpendicular to L .
Suppose the function f is continuous and satisfies ∕( x ) ≥ 0 on the interval
[a, b], and suppose we wish to find the volume of the solid 5’ generated when the
region R under the curve y = ∕( x ) on [a, b] is revolved about the x-axis. That is,
the axis o f revolution is horizontal and it is a boundary o f the region R . Our strategy
shall be to form vertical strips and revolve them about the x-axis, generating what
are called disks (that is, thin right circular cylinders) that approximate a portion of
the solid of revolution S, as shown in Figure 6.5.
Now we can compute the total volume of 5 by using integration to sum the
volumes of all the approximating disks. Recall that the formula for the volume of
a cylinder of height h and cross-sectional area A isd∕z. Figure 6.5a shows a typical
vertical strip with height f( x l) and width ∆xjt. You might notice that the width of
the strip is the same as the thickness of the disk. The solid of revolution can be
thought of as having cross sections perpendicular to the x-axis that are circular
disks of volume ΔK(x) = 7Γ[ ∕( X *)]2
a. A representative vertical strip
has height f( x k ) and width ∆ x fc. Area o f circular cross section
The total volume may be found by integration:

Area o f base Thickness
V= A(x) dx = ττ[∕(x)]2 dx
Ja Ja
This procedure may be summarized as follows:
The Disk Method

The disk method is used to find a volume generated when a region R is
b. The representative disk is formed
revolved about an axis L that is perpendicular to a typical approximating strip
by revolving the representative in R. In particular, if the region R bounded by the curve y = ∕(x), the x-axis,
strip about the x-axis. This disk and the vertical lines x = a and x = b is revolved about the x-axis, it generates
has radius∕(% t ) and thickness ΔΛJ.
a solid with volume r⅛ r⅛
Figure 6.5 The disk method K = ττy2 dx = π [∕(x )] 2 dx
Ja Ja
What This Says: The following diagram may help you remember the key
ideas behind the disk method. I
From precalculus Use formula to find Integrate on the interval
I KNOWN FORM ULA ∣ I REPRESENTATIVE ELEMENT ∣ → I INTEGRATION FORM ULA]
rb
Volume of disk: Δ V = rry2 ∆x V = 77 y 2 dx
V = Bh = τrr2h Ja
≡≡≡≡M≡≡≡≡≡≡≡≡
EXAMPLE 3 Volume of a solid of revolution: disk method
Find the volume of the solid S formed by revolving the region under the curve
y = x 2 + 1 on the interval [0, 2] about the x-axis.
Solution The region is shown in Figure 6.6.
Figure 6.6 Volume of a solid of revolution: Disk method

r2
V = 77 (χ 2 + 1)2 dx = 7τ (x 4 + 2x 2 + 1) dx
JO Jo
= ^ 7 7 ≈ 4 3 .1 4 4 53 9 1 1
With a small modification of the disk method, we can find the volume of a
solid figure generated by revolving about the x-axis the region between two curves
y = f(x ) and y = g(x) where ∕( x ) ≥ ^(x) for a ≤ x ≤ b. When a typical vertical
strip is revolved about the x-axis, a “ washer” is formed that has cross-sectional
The volume of the solid of revolution is found by the formula given in the
following box.
The Washer Method

Suppose f and g are continuous functions with ∕( x ) ≥ g(x) on the interval
0 Remember [ ∕ 2 - g 2 ] ≠ [a, ⅛], and let R be the region bounded above by y = ∕( x ) , below by y = g(x),
( / - g)2 0 and on the sides by x = a and x = b. The washer method says that the volume
generated by revolving R about the x-axis is
v = [ π ([∕(x )] 2 - [g(x)]2) dx
Outer radius Inner radius

rb / ∖
= 7τ ([R(x)]2 - [r(x)]2 l dx
s
Ja ×---*--- j x j
- '
Outer radius Inner radius
rb rb
= ττ [R(x)]2 dx - 7τ [r(∙x)]2 dx
The disk method and the washer method also apply when the axis of
revolution is a line other than the x-axis. In Example 4, we consider what happens
when a particular region R is revolved not only about the x-axis, but also about
other axes.
EXAM PLE 4 Volume of a solid of revolution: washer method
Let R be the solid region bounded by the parabola y = x 2 and the line y = x.
Find the volume of the solid generated when R is revolved about the
a. x-axis b. y-axis c. line y = 2

Solution First, find the points of intersection of the parabola and the line by
solving the following system of equations:
This is equivalent to solving x = x 2 , which has solution x = 0, 1.

Draw the region R in the xy plane as shown in Figure 6.8.
a. We form the solid by rotating the region
R about the X-axis. Note that the line
y = x is always above the parabola on the
interval [0, 1], so when we form a washer
(∩. 0)
to approximate the volume of revolution,
the outer radius is y = R(x) = x and the
Figure 6.8 The region bounded by inner radius is y = r(x) = x 2 . Thus, the
y = x 2 and y = x required volume is
V = ττ f [x 2 - (x 2)2] dx =
Jo
τr [ (x 2 ~ x 4) dx = τr(Xc3 - ∣ x 5 ) = ⅛
Jθ ∖D ∙-× / 10 1D
b. Because we are revolving R about the y-
axis, we use horizontal strips to approxi
mate the solid of revolution, as shown at
the right. Note that the parabola x = V y
is to the right of the line x = y on the
interval [0, 1], so the approximating
washer has outer radius R = V y and in
ner radius r = y; thus,
A = ττ [ [(Vy) 2 - (τ)2] dy
Jo
c. The outer radius is R = 2 - x 2 and the

inner radius is r = 2 - x , as shown at the
right. The volume is
A = ττ[ [(2 - x 2)2 - ( 2 - x) 2] dx

Jo
Γ1
= 77 (x 4 - 5x 2 + 4x) dx
Jo
ΓxJ _ 5x 3 , 4x 2^∣ 1 _ 8ττ
['5 3 2 J 0 15
0 For the disk method or the washer method, make sure that the
approximating strips are perpendicular to the axis of revolution. 0
■ THE METHOD OF CYLINDRICAL SHELLS
Sometimes it is easier (or even necessary) to compute a volume by taking the

approximating strip parallel to the axis of rotation instead of perpendicular to
the axis as in the disk or washer methods. Figure 6.9a shows a region R under the
curve y = ∕( x ) on the interval [a, b], with a representative vertical strip. When this
strip is revolved about the y-axis, it forms a solid called a cylindrical shell.
When the approximating strip is rotated about the y-axis, it generates a
cylindrical shell of height ∕( x ) and thickness ∆x, as shown in Figure 6.9b. Because
the strip is x units from the axis of rotation and is assumed to be very thin, the
cross section of the shell (perpendicular to the y-axis) will be a circle of radius x
and circumference 2τrx. If we imagine the shell to be cut and flattened out, it is
seen to be a rectangular slab of volume
Thickness
AJ ' = V
2πxf(x)
_ J■Ax
Area of the rectangular cross section
(See Figure 6.9c.) Thus, the total volume of the

solid is given by the integral
a. A vertical strip in the region cb

R under the curve y =f(x)
V= 2πxf(x) dx
on the interval [a. b] Ja
The approximating shells are shown in Figure Figure 6.10 Approximating a

6.10 and the formula is repeated in the following solid o f revolution by cylin
box. drical shells
Method of Cylindrical Shells

The shell method is used to find a volume
generated when a region R is revolved about an
axis L parallel to a typical approximating strip
in R. In particular, if R is the region bounded
by the curve y = f(x), the x-axis, and the verti
cal lines x = a and x = b where 0 ≤ a < b,
b. When the strip is revolved then the solid generated by revolving R about
about the y-axis, a shell is
generated. the v-axis has volume
fA
Jz = 2τrxf(x) dx
Ja
EXAMPLE 5 Volume of a solid of revolution: shell method
Find the volume of the solid formed by revolving the region bounded by the
1πx graphs of y = x 3 + x 2 + 1, x = 1, and x = 3 about the y-axis.
c. The unwrapped Solution This example illustrates why we need the shell method. A typical
"flattened" shell has
volume ΔV = 2πx∕(x)∆x. vertical strip (shown in Figure 6.11) has height f(x) = x 3 + x 2 + 1 so the
Figure 6.9 Method of cylindrical volume, by the method of shells, is
shells
ri
V(x) = 2 77 x(x 3 + x 2 + 1) dx
= 2 77 (x4 + x 3 + x) dx
= 2ττ∖⅞-
5
+⅜
4
+⅜
2 ∕∣ ι
= 144.877
Figure 6.11 A volume of revolution

by the shell method ■■■
We conclude this section with Table 6.2, which compares and contrasts the
disk, washer, and shel, methods for finding volume.
TABLE 6.2 Volumes of Revolution for Which the Axis of Revolution Is Either the x-axis or the j-axis
SUMMARY
a. Disk method: A representative rectangle is perpen Horizontal axis of revolution Vertical axis of revolution
dicular to the axis of revolution. The axis of revolu
tion is a boundary of the region.
Width of rectangle is ∆x. Width of rectangle is ∆y.

[∕(x )] 2^fc
⅛(T)]2 dy
Length (Height) of rectangle Length of rectangle
b. Washer method: A representative rectangle is per
pendicular to the axis of revolution. The axis of
revolution is not part of the boundary.
Length of rectangle Length of rectangle
c. Shell method: A representative rectangle is parallel

to the axis of revolution.
Length (Height)
Distance to axis
Distance to axis
PROBLEM SET 6 .1
In Problems 1-6, sketch the given region and then find the 15. y = V4 - x 2, 0 ≤ x ≤ 2
volume of the solid whose base is the given region and which a. about the x-axis
has the property that each cross section perpendicular to the x- b. about the y-axis
axis is a square.
c. about the line x = -1
1. the triangular region bounded by the coordinates axes and
the line y = 3 - x
2. the region bounded by the x-axis and the semicircle
y = V 16 - x 2
3. the region bounded by the line y = x + 1 and the curve
y = x 2 - 2x + 3 16. y = V4 - x 2, ≤ χ ≤ 2
4. the region bounded above by y = Vsinx and below by the a. about the x-axis
x-axis on the interval [05 τr] b. about the y-axis
5. the region bounded above by y = Vcosx and below by the c. about the line y = -1
x-axis on the interval [—J>
6. the triangular region with vertices (1, 1), (3, 5), and
(3, -2)
In Problems 7 -12, sketch the region and then find the volume
of the solid whose base is the given region and which has the
property that each cross section perpendicular to the x-axis is In Problems 17-20, draw a representative strip and find the
an equilateral triangle. integral for the volume o f the solid formed by revolving the
7. the region bounded by the circle x 2 + y 2 = 9 given region
8. the region bounded by the curves y = x 3 and y = x 2 a. about the x-axis
9. the region bounded by the y-axis, the parabola y = x 2, and b. about the y-axis
the line 2x + y —3 = 0
10. the region bounded above by y = cosx, below by Set up the integral only; DO NOT EVALUATE.
y = sinx, and on the left by the y-axis
17. the region bounded by
11. the region bounded above by the curve y = tanx and
x = y 2 and y = x 2
below by the x-axis, on the interval [θ,
12. the region bounded by the x-axis and the curve y = ex
between x = 1 and x = 3
Name the most appropriate method (disks, washers, or shells)

for finding the solid of revolution in Problems 13-16. Use the
representative rectangle that is shown to set up the integral, but
DO NOT EVALUATE.
13. y = 4 - x, 0 ≤ x ≤ 4
a. about the x-axis
b. about the y-axis
c. about the line y = —1
14. y = 4 - x 2, 0 ≤ x ≤ 2 19. the region bounded by

a. about the x-axis the lines y = 1, x = 1,
b. about the y-axis and the curve
c. about the line y = —1 y = x 3 + 2x + 1
20. the region bounded by the curves y = x 2 and y = 28. semicircles

-4 x
In Problems 29-38, find the volume o f the solid formed when

θ In Problems 21-24, find the volume o f the solid whose base is the region described is revolved about the x-axis.
bounded by the circle x 2 + y 2 = 9 with the indicated cross 29. the region under the curve y = V x on the interval [0, 1]
sections taken perpendicular to the x-axis. 30. the region under the curve y = V x on the interval [0, 8]
21. squares 22. equilateral triangles
31. the region bounded by the lines y = x, y = 2x, and x = 1
32. the region bounded by the lines x = 0, x = 1, y = x + 1,
and y = x + 2
33. the region under the curve y = x 2 + x 3 on the interval
[0, 77]; approximate your answer to the nearest unit
34. the region bounded by the curves y = 2 - x 2 and y = x 2
35. the region bounded by the curves y = x 2 and y = x 3
36. the region under the curve y = V2 sin.v on the interval
[0, 77]
37. the region under the curve v = V sinx on the interval
[0. 77]
23. isosceles right triangles 24. semicircles 38. the region bounded by the curves y = ex and y = e~x
between x = 0 and x = 2; approximate your answer to the
nearest unit
In Problems 39-46, find the volume o f the solid formed by

revolving the given regions about the y-axis.
39. the region under the line y = 2x on the interval [0, 1]
40. the region under the curve y = V x on the interval [0, 1]
41. the region bounded by the curves y = x 2 and y = x 3
42. the region bounded by the lines y = x, y = 2x, and x = 1
43. the region bounded by the parabola y = 1 - x 2 , the y-axis,
and the positive x-axis
In Problems 25-28, find the volume o f the solid whose base is
bounded by the graphs o f y = x + 1 and y = x 2 — 1 with the 44. the region bounded by the curve y = 8 —x 3, the y-axis,
indicated cross sections taken perpendicular to the x-axis. and the positive x-axis
25. squares 26. equilateral triangles 45. the region bounded by the parabolas y = 1 —x 2 , y = x 2,
and the j>-axis
46. the region under the curve y = e~x ~ on the interval [0, 1]
In Problems 47-52, find the volume o f the solid generated by

revolving each region about the prescribed axis.
47. the region bounded by y = x 2 and y = x 3, about the line
y = -1
48. the region bounded by x = V4 - y, x = 0, and y = - 1,
about the y-axis
49. the region bounded by y = x 3, y = 12 - x 2 , and x = 1,
about the line x = - 1
50. the region bounded by y = ∣( 12x - x 3), y = 2, and x = 0,

about the y-axis
51. the triangular region with vertices (1, 1),(2, 5), and (4, 1),
about the x-axis 59. Repeat Example 1 using a spreadsheet to find the
52. the triangular region with vertices (1, 1), (3, 4), and (5, 1), volume.
about the y-axis 60. Repeat Example 3 using a spreadsheet to find the
In Problems 53-55, fin d the volume V o f the solid using the volume.
given information about its cross section. 61. Cross-sectional areas are measured at one-foot inter
53. The base of the solid is a circle with radius 1, and the cross vals along the length o f an irregularly shaped object,
sections perpendicular to a fixed diameter of the base are with the results listed in the following table (both x
squares. and A are measured in feet):
X 0 1 2 3 4 5
A 1.12 1.09 1.05 1.03 0.99 1.01
X 6 7 8 9 10
A 0.98 0.99 0.96 0.93 0.91
54. The base of the solid is an equilateral triangle each side of

which has length 4. The cross sections perpendicular to a Estimate the volume (correct to the nearest hun
given altitude of the triangle are squares. dredth) by using the trapezoidal rule.
62. Cross-sectional areas are measured at 1-ft intervals
along the length of an irregularly shaped object, with
the results listed in the following table (both x and A
are measured in feet):
X 0 1 2 3 4 5
A 1.12 1.09 1.05 1.03 0.99 1.01
55. The base o f the solid is an isosceles right triangle whose

legs a and b are each 4 units long. Each cross section X 6 7 8 9 10
perpendicular to a is a semicircle.
A 0.98 0.99 0.96 0.93 0.91
Estimate the volume (correct to the nearest hun

dredth) by using Simpson’s rule.
θ 63. A hemisphere o f radius r may be regarded as a solid whose

base is the region bounded by the circle x 2 + y 2 = r2 and
56. Find the volume of the solid generated when the region with the property that each cross-section perpendicular to
the x-axis is a semicircle with a diameter in the base. Use
y = on the interval [1, 4] is revolved about this characterization and the method o f cross-sections to
show that a sphere of radius r has volume V = ∣π r 3 . Note:
a. the x-axis b. the y-axis c. the line y = —2
This is a formula given in Table 6.1.
57. The great pyramid o f Cheops is approximately 480 ft tall
64. Use the method of cross-sections to show that the volume
and 750 ft square at the base. Find the volume of this
of a regular tetrahedron of side a is ∙⅛V2α3. Note-. This is
pyramid by using the cross-section method.
a formula given in Table 6.1.
58. When viewed from above, a swimming pool has the shape
of the ellipse 65. Find the volume of the football-shaped solid (called an
ellipsoid} formed by revolving the ellipse
* i J±= 1
900 + 400 ^ + ^=1
a2 b2
The cross sections of the pool perpendicular to the ground
and parallel to the y-axis are squares. If the units are in about the x-axis. What is the volume if the ellipse is
feet, what is the volume of the pool? revolved about the y-axis?
66. Find the volume of the doughnut-shaped solid (called a

torus) formed by rotating the circle
(x - a)2 + y 2 = b 2
about the r-axis. You may assume that a > b.

67. A “ cap” is formed by truncating a sphere of radius A at a
point h units from the center, as shown in Figure 6.12.
Find the volume of the cap.
Figure 6.13 Volume of the frustum o f a cone
69. A spherical sector (or a “ gem” ) is formed from a section

of a sphere as indicated in Figure 6.14. Assume that the
spherical base has radius R and that the parallel truncat
ing planes are A and h2 units from the center of the
sphere, where
0 < A, < Λ2 < R

Find the volume o f the gem.
Figure 6.12 Volume of the cap of a sphere
68. The frustum o f a cone is the solid region bounded by a

cone and two parallel planes, as shown in Figure 6.13. If
the planes are h units apart and intersect the cone in plane
regions of area∠41 and A , use integration to show that the
frustum has volume
F=∣(.1, + ∖ < ξ ξ + ,l 2)
Figure 6.14 Volume o f a section of a sphere
6.2 ARC LENGTH AND SURFACE AREA
IN THIS SECTION The arc length of a graph, the area of a surface of

revolution
If you were asked to measure the length of a given curve by hand, you might
proceed by first fitting a piece of string to the curve and then measuring its
length with a ruler. In this section, we see how such measurements can be
carried out mathematically using integration. We also develop a general
procedure for computing the surface area of certain solid figures.
Physical method for measuring the
length o f a curve
■ THE ARC LENGTH OF A GRAPH
If a function f has a derivative that is continuous on some interval, then / is said to

be continuously differentiable on the interval. The portion of the graph of a
continuously differentiable function/that lies between x = a and x = b is called
the arc of the graph on the interval [a, b]. To find the length of this arc, let P be a
partition of the interval [a, /?], with subdivision points x 0, x 1 , . . . , x where
x 0 = a and x n = b. Let P k denote the point (xk , yk ) on the graph, where yk = f(xf).
By joining the points P Q, P i , . . . , P n, we obtain a polygonal path whose length
approximates that of the arc. Figure 6.15 demonstrates the labeling for the case
where n = 6.
The length of the polygonal path connecting the points Pθ, P i , . . . , P on the
graph o f/ is the sum
n
Σ sx
k=∖
Sec. 6.2 Arc Length and Surface Area 397
Figure 6.15 A polygonal path approximating an arc of a curve
where sk is the length of the segment joining Pk ~ l to P k . By applying the distance

formula and rearranging terms with ΔX A = x k - x k _ { and ∆y k = y k - y k _ v we find
that _______________________________
⅜ = √ ( ⅞ - ⅞ - I)2 + ( ^ - Λ - I ) 2
= √(Δ⅞)2 + ( ∆ γ j 2
U ⅞ ) 2 + (∆yfcj 2
(Δ⅞)2
It is reasonable to expect a connection between the ratio ∆y k ∕∆x k and the

derivative dy!dx = ∕'( x ) . Indeed, it can be shown that
⅞ = √ι + [ ∕'( 4 ) ] 2 δ ⅞
for some number x k between x k _ 1 and x k . Therefore, the arc length of the graph off
on the interval [a, b] may be estimated by the Riemann sum
∑ ∖ i + [ ∕'( 4 ) ] 2 δ ⅞
fc=ι
This estimate may be improved by increasing the number of subdivision

points in the partition P of the interval [a, b∖ in such a way that the subinterval
lengths tend to zero. Thus, it is reasonable to define the actual arc length to be the
limit
lim T ∖ 1 + u[∕'(⅞ 2 ∆x,

H →o⅛ι v k)]
'i k
Notice that iff ' is continuous on the interval [a, b], then so is V 1 + [ ∕'( ∙ x )]2 ∙ Thus,
this limit exists and is the integral o f V l + [∕'( x ) ] 2 with respect to x on the interval
[a, ⅛]. These observations lead us to the following definition.*
*We have used integration to obtain a meaningful definition of arc length for a function f{x ) with
a continuous derivative f'{x). It is possible to define arc length for certain curves y = fix ), where f(x )
does not have a continuous derivative, but we will not pursue this more general topic. It may surprise
you to know there are curves in the plane whose length cannot be defined. Curves that have length are
said to be rectifiable, and those that do not are nonrectifiable. You might wish to look up references to
fractals or the snowflake curve.
Arc Length Let / b e a function whose derivative f ' is continuous on the interval [a, b].
Then the arc length, s, of the graph of y = f(x) between x = a to x = b is given
by the integral
^b
a J
V l + [∕'( x ) ] 2 dx
Similarly, for the graph of x = g(y), where g' is continuous on the interval
[c, d], the arc length from y = c to y = d is
<∙d
c J
V l + [ / ( / ] 2 dy
E X A M P L E 1 Arc length of a curve
Find the arc length of the curve y = x 3'2 on the interval [0, 4].
Solution Let / ( / = x 3Z2; therefore,/'/) = ∣ x ιz2.
= ⅛(10) 3z2 - (I) 372] ≈ 9.073415289
⊂3 Computational Window ▲
∣ Once again, we remind you of alternative forms common on computer

∣ software. For Example 1, an alternative form is
80∕ l ° ~ ψ j ≈ 9.073415289
E X A M P L E 2 Arc length of function of y
Find the arc length of the curve x = ∣ ∕ + ⅛y 1 from y = 1 to y = 3.
4y 4 - 1
Solution Because g(y) = ∣ ∕ + ⅛y 1, we have g'(y) = y 2 - ⅛y 2
, which
4/
is continuous throughout the interval from y = 1 to y = 3. Therefore, the arc
length is _________________
f3 ,_____________2 ri / M √ - 1∖2
V l + [ ∕ ( ∕ ] dy = J i + ' 7 dy
Ji Ji v ∖ 4j- /
16/ - 8 / + 1
dy
16/
16/ + 16/ - 8 / + 1
A graphing calculator will not dy
16/
give you the arc length, but you
can graph the curve and imag 3 ∣(4y4 + 1)2
ine it to be a “piece of string.”
If you then measure the string, V ( 4 ∕) 2
is the calculus answer reason
able? + ⅛ ≈ 8.833333333
3 ∕ 4 '^ T
--------------------- - ⅛at≡m⅛⅜≡
Some advanced scientific calculators will give the arc length of a function
, correct to a specified tolerance. For Example 2, if the tolerance is 0.01 with
j δ = 0.001, we find (on a TI-85, for example),
arc(expression, variable, left endpoint, right endpoint)
or
arc(X^3∕3 + .25X~ 1, X , 1, 3)
∣ to obtain 8.83333345015.
EXAM PLE 3 Estimating arc length using numerical integration
Find the length of the curve defined by y = sinx on [0, 2π].

Solution Because y ' = cosx, we have, from the arc length formula,
5= ∣ V 1 + cos2 x dx
Jo
We do not have techniques to allow us to evaluate this integral, so we turn to
numerical integration. The region is shown in Figure 6.16. If we consider four
rectangles we find, by various methods:
Method (n = 4) Approximation
Rectangles
Left endpoints 7.584476
Right endpoints 7.584476
Midpoints 7.695299
Trapezoids 7.584476
Simpson’s rule 7.150712
Figure 6.16 Right-endpoint

approximation for n = 4
In practice, you would choose just one of the methods whose approximate
values are given. You might also note that, for this example, Simpson’s rule
performs worse than the trapezoidal, midpoint, or even the left- and right
endpoint methods. For more accurate results, you might wish to use a
computer program. A simple output is shown in the computer window at left.
■ THE AREA OF A SURFACE OF REVOLUTION
When the arc of a graph is revolved about a line L , it generates a surface that we
shall refer to as a surface of revolution. Figure 6.17 shows an arc, together with a
typical approximating line segment. Notice that when the segment is revolved
about L , it generates the surface of the frustum of a right circular cone.
Red shaded region is the
∆s frustrum of a cone.
Rather than argue formally with Riemann sums, we will simply proceed
intuitively. It can be shown that a frustum with slant height € and base radii r1 and
r2 has lateral surface area ττ(r1 + r,)€ (see Problem 32). Figure 6.17 shows an
approximating line segment we shall assume has length Δ Δ, and is located h units
above the axis of revolution L . Then, by revolving this segment about L , we
generate a frustum whose slant height is ∆.s' and which is so thin (because ∆s is very
small) that the radii of both its circular bases are essentially the same as h.
To find the area of the surface of revolution let the radii of the bases be r1 and
r7. Then the approximating frustum has surface area ττ(r1 + r7)Δ∆, . Because both r 1
and r2 are essentially the same as h, we can estimate the surface area of the frustum
by 1
Lateral length
Δ5 = ττ(h + h)∆s = 2ττh∆s = 2τr∕(x)Vl + [∕'( x ) ] 2 ∆x
Distance to axis of revolution is h = f(x).
The surface area .S’ can be obtained by integrating this expression on the interval
An increment of surface area A .S’ [a, b]. These considerations lead us to the following definition.
Surface Area Suppose f ' is continuous on the interval [a, Z>]. Then the surface generated by
revolving about the x-axis the arc of the curve y = ∕( x ) on [a, b] has surface
area
5 = 2 77 f ( x ) V l + [∕'( x ) ] 2 dx
Ja
You may find it instructive to derive this formula by the more rigorous
approach used to establish the arc length formula (see Problem 34). We close with
two examples that illustrate the use of the surface area formula.
EXAM PLE 4 Area of a surface of revolution
Find the area of the surface generated by revolving about the x-axis the arc of
the curve y = x 3 on [0, 1].
Solution Because f(x ) = x 3, we have ∕'( x ) = 3x 2, which is certainly continuous
on the interval [0, 1].
"1 ____________ r1 _________ r 10

J0
x 3V 1 + (3x2)2 dx = 2π x 3V 1 + 9x 4 dx = 2π
Jo Jl
w172∣
Let it = 1 + 9x 4 ; du = 36 x i dx.
If x = 1, then ιι = 10.
If x = 0, then u = 1.
= ⅞ 0 " 3z2)Γι° = ∑7[1 0 v τ ° - 1 1≈3∙56
EXAM PLE 5 Derive the formula for the surface area of a sphere
Find a formula for the surface area of a sphere of radius r.

Solution We can generate the surface of the sphere by revolving the semicircle
y = >∕r2 - x 2 about the x-axis. We find that
Because the semicircle intersects the x-axis at x = r and x = -r , the interval of

integration will be [-r, r]. Finally, by applying the formula for the surface
area, we find
= 2ττ r dx = 2πr(2r) = 4ττr2
To generalize the formula for surface area to apply to any vertical or

horizontal axis of revolution, suppose that an axis of revolution is R(x) units from
a typical element of arc on the graph of y = fix), as shown in Figure 6.18.
Figure 6.18 Surface of revolution about a general vertical or horizontal axis
Then 2ττR(x) is the circumference of a circle of radius R(x). and it can be shown
that an element of surface area is
Δ5 = 2πΛ(x)Δs = 2πΛ(x)Vl + [∕'(x )] *2 ∆x
Thus, the surface of revolution on the interval [a, 6] has area
rb
S = 2ττ ∖ Λ(x)Vl + [.∕'(x)] 2 dx
Ja
In particular, if the graph of y = ∕(x ) is revolved about the y-axis, an element of
the arc is R(x) = x units from the y-axis, and the resulting surface has area
rb
S = 2π x √ 1 + [∕'(x )] 2 dx
Ja
PROBLEM SET 6 .2
θ Find the length o f the arc o f the curve y = f(x) on the intervals
'^ 1. fix ) = 3x + 2 on [ - 1 , 2] 5. f ix ) = ∣ x 3z2 + 1 on [0, 4]
2. ∕( x ) = 5 - 4x on [—2, 0] 6. fix') = ⅜(2 + x 2) 3z2 on [0, 3]
3. fix ) = 1 - 2x on [1, 3] 7. f ix ) = ⅛x 5 + ∣x ^ 3 on [1, 2]

4. ffx ) = x 312 on [0, 4] 8. f ix ) = ∣ x 3 + ∣x ^ l on [1, 4]
9. f ix ) = ∣ x 4 + ∣x ^ 2 on [1, 2]
Find the surface area generated when the graph o f each

function given in Problems 21-24 on the prescribed interval is
revolved about the y-axis.
21. ∕( x ) = ∣(12 - x) on [0, 3]
22. ∕( x ) = j x 3z2 on [0, 3]
23. ∕( x ) = ⅜Vx(3 - x) on [1, 3]

12. f(x ) = y∕e2x - 1 - sec l (<T') on [0, In 2]
13. Find the length o f the curve defined by 9x2 = 4 j∕3 between
the points (0, 0) and (2Vz3, 3).
14. Find the length of the curve defined by (y + 1)2 = 4x 3
between the points (0, —1) and (1, 1).
15. Find the area of the surface generated when the arc of the
curve
y = ⅜x3 + (4x)^l
between x = 1 and x = 3 is revolved about 24. ∕( x ) = 2Vz4 - x on [1, 4]

a. the x-axis b. the y-axis
16. Find the area of the surface generated when the arc of the
curve
„ _ 3,,5/3 _ 3.,,1/3
x - 5.r 4>
between y = 0 and y = 1 is revolved about

a. the y-axis b. the x-axis c. the line y = - 1
θ Find the surface area generated when the graph o f each
function given in Problems 17-20 on the prescribed interval is
revolved about the x-axis. The graph o f the equation
17. ∕( x ) = 2x + 1 on [0, 2]
X 2' 3 ÷ y 2'3 = 1
18. ∕( x ) = V x on [2, 6]
19. ∕( x ) = ∣ x 3 + ⅛x^l on [1, 2] is an astroid.
Find the length o f this particular astroid by finding the

length of the first quadrant portion, y = (1 —x 2z3)3z2 on
[0, 1]. By symmetry, the entire curve is four times the
length o f the arc in the first quadrant.
26. Find the area o f the surface generated by revolving the
portion o f the astroid with y ≥ 0 (see Problem 25)
x 2/3 + y 2'3 = 1 on [ - 1, 1]
about the x-axis.

27. Find the length of the curve
χ2∕3 -∣- j 12∕3 =
q
Give your answer correct to the nearest hundredth. between x = 1 and x = 27.
28. Find the area o f the surface generated when the arc o f the c. Compare the results of parts a and b to show that the
curve frustum has surface area
X 2'3 2∕3
+ y = 9 A = ττ(rl + ∕' 2)f
between x = 1 and x = 27 is revolved about the x-axis. 33. If∕'( x ) exists throughout the interval [x⅛- 1 , x j , show that
29. Use Simpson’s rule with n = 20 to estimate the area of the there exists a number x^ in this interval for which
surface obtained by revolving y = tanx about the x-axis
V (∆ x fc)2 + (Δ¾)2 = V l + [∕'(x *)] 2 ∆ x ,
on the interval [0, 1].
30. Show that when the arc of the graph of∕( x ) between x = a where ∆x f,. = x k - x k _ } and ∆y k = f ( x k ) - ∕( x i . 1). Hint:
and x = b is revolved about the y-axis, the surface gener Use the mean value theorem.
ated has area
cb 34. Verify the surface area formula by completing the follow
S = 2π x V 1 + [ ∕'( x ) ] 2 dx ing steps (see Figure 6.20).
Ja
31. Show that a cone of radius r and height Λ has surface area a. Let P = {x 0 , x 1, . . . , x j∙ be a partition of the interval
[a, b], and for k = 0, 1, . . . , n, let P k denote the point
S = π rV r 2 + h 2 (x k , y k ) where y k = J ∖x k)∙ Show that when the line
segment between P,.-1 and P k is revolved about the x-
Hint. Revolve part o f the line hy = rx about the x-axis. axis, it generates a frustum o f a cone with surface area
32. Figure 6.19 shows the frustum of a cone with base radii r1
and r, and slant height f. π ( ⅛ . 1 + y k )<k
where £k is the distance between P k _ 1 and P k .
Figure 6.20 Surface area formula

Figure 6.19 Frustum of a cone
b. Notice that ∣(y fc. + T⅛) is a number between y k and

1 l
Suppose that a regular polygon of n sides is inscribed in y k . Use the intermediate value theorem (Section 1.7) to
each circular base. Each side of the polygon in the upper show that
base has length .s, 1 and each side of the polygon in the y k . t + y k = 2f(cl)
lower base has length s2 . Notice that a trapezoid is formed
for some number ck between x j,.-1 and x k .
when the ends of corresponding sides in these polygons
are joined by line segments. c. Explain why the surface generated when the arc of the
graph of/between x = a and x = b is revolved about
a. The surface area o f the frustum can be estimated by
the x-axis has area
adding the areas o f the approximating trapezoids.
Show that this estimate is
2 √ (⅛ )V 1 + [ ∕'( x ∕) ] 2 ∆x k
ll⅛ Σ
⅛ + s 2)hn
where h is the distance between corresponding sides d. Notice that as ∣∣P∣∣ → 0, the numbers ck and x*k are
in the polygons inscribed in the circular bases. “ squeezed” together. Use this observation to show that
the surface area is given by the integral
b. Show that the polygon inscribed in the larger base has
perimeter fb /---------
5, = 2τr ∕( x ) V l + [∕'( x ) ] 2 dx
lim ns. = 2πr Ja
n→+∞ 1
Use similar reasoning to establish the following limit 35. The center of a circle of radius r is located at (R. 0), where
statements: R > r. When the circle is revolved about the j-axis, it
generates a torus (a doughnut-shaped solid). What is the
lim ns, = 2ττr,2 and lim h = t
n→+∞ 2 n→+≈ surface area of the torus?
6 .3 PHYSICAL APPLICATIONS: WORK, LIQUID FORCE, AND CENTROIDS
IN THIS SECTION Work, fluid pressure and force, centroid and moment of
a plane region, the volume theorem of Pappus
Integration plays an important role in many different areas of physics.
Mechanics is the branch of physics that deals with the effects of forces on
objects. In this section, we show how integration can be used to compute work
and the force exerted by a liquid. We also use integration to compute centroids
of plane objects and to apply a result called the volume theorem o f Pappus. This
classic result provides a simple procedure for computing the volume of certain
solids of revolution.
■ WORK
In physics, considerable time is devoted to the study of forces. Intuitively, a force is

that which pushes, pulls, compresses, distends, distorts, or in any way changes the
state of rest or state of motion of a body. There are several types of forces, such as
centrifugal force, centripetal force, and electromotive force. In order to measure
forces, we define a unit of force to be the force that will give unit acceleration to a
unit mass. When a constant force of magnitude F is applied to an object through a
distance d, the work performed is defined to be the product of force and distance.
Work Done by a Constant Force If a body moves a distance d in the direction of an applied constant force F,
the work W done by the force is
W = Fd
COMMON UNITS OF In the U.S. system of measurements, work is typically expressed in foot-pounds,
WORK AND FORCE inch-pounds, or foot-tons. In the International System (SI), work is expressed in
Mass Distance Force Unit WORK
newton-meters (called joules'), and in the centimeter-gram-second (CGS) system,
the basic unit of work is the dyne-centimeter (called an erg).
kg m newton (N) joule For example, the work done in lifting a 90 lb bag of concrete 3 ft is W =
g cm dyne (dyn) erg Fd = (90 lb)(3 ft) = 270 ft-lb. Notice that this definition does not conform to
slug ft pound ft-lb everyday use of the word work. If you labor at lifting the concrete all day but are
not able to move the sack of concrete, then no work has been done.
It is not necessary for F and d to be constants, but if they are not, then calculus
is used to find the work done. Suppose F is a continuous function and F{x) is a
variable force that acts on an object moving along the x-axis from x = a to x = b.
We shall define the work done by this force. Partition the interval [a, b] into
subintervals, and let ΔXA, be the length of the kth subinterval f,. If ∆xk is
sufficiently small, we can expect the force to be essentially constant on Ik , equal,
for instance, to F(x*k ), where x£ is a point chosen arbitrarily from the interval I k . It
is reasonable to expect that the work ∆uj,. required to move the object on the
interval Ik is approximately ∆wi = F(x*) ∆xjfc, and the total work is estimated by the
sum
n
Σ, F(x*)∆x jt
as indicated in Figure 6.21.
Sec. 6.3 Physical Applications: Work, Liquid Force, and Centroids 405
c v *.)
F{x* λ The work ∆w,k done when
k → ,■ e
an object moves from
-4---χ*1-------- 1
x ,,
---- 1
x,
—I—
b
χ
i - ι to χ
force
w ith the
F isk approximately
∆wt = F(x*) ∆xj .
Δ ¾
Figure 6.21 Work performed by a variable force
By taking the limit as the norm of the partition approaches 0, we find the total
work is given by
W = lim ∑ F(x↑)∆x, = [ F(x) dx
∣∣P ∣∣→O FC⅛ V ⅛' A Ja v ’
Work Done by a Variable Force The work done by the variable force F(x) in moving an object along the x-axis
from x = a to x = b is given by
r⅛
W= F (x )d x
Ja
EXAM PLE 1 Work with a variable force
An object located x ft from a fixed starting position is moved along a straight

road by a force of F(x) = 3x2 + 5 lb. What work is done by the force to move
the object:
a. through the first 4 ft? b. from 1 ft to 4 ft?
Solution a. W = f (3x2 + 5) dx = (x 3 + 5x) ∣θ = 84 ft-lb
b. W= (3x2 + 5) dx = (x 3 + 5x) 11 = 78 ft-lb
Hooke’s law, named for the English physicist, Robert Hooke (1635-1703),
states that when a spring is pulled x units past its equilibrium (rest) position, there is
a restoring force F(x) = kx that pulls the spring back toward equilibrium* (See
Figure 6.22.) The constant k in this formula is called the spring constant.
H≡≡k Hooke's law: A

force F(χy) = kx acts
to restore the spring
to its equilibrium
)------------ F position.
o o
A spring at rest A spring stretched x units
(equilibrium position) past equilibrium
Figure 6.22 Hooke’s law
‘ Actually, Hooke’s law applies only in ideal circumstances, and a spring for which the law applies
is sometimes called an ideal (or linear) spring.
EXAM PLE 2 Hooke’s law as an application of the work formula
The natural length of a certain spring is 10 cm. If it takes 2 dyn ■cm of work to
stretch the spring to a total length of 18 cm, how much work will be performed
in stretching the spring to a total length of 20 cm?*
Solution Assume that the point of equilibrium is at 0 on a num ber line, and let x
be the amount the free end of the spring is extended past equilibrium. Because
the stretching force of the spring is F(x) = kx, the work done in stretching the
spring b cm beyond equilibrium is
r⅛ rb
W = F(x) dx = k x dx = ⅛kb2
Jo Jo
We are given that W = 2 when b = 8, so
2 = ⅛k(8)2 implies k = γg
Thus, the work done in stretching the spring b cm beyond equilibrium is
IF = ⅛(⅛) ~ ^hP1 d Yn ' c m

In particular, when the total length of the spring is 20 cm, it is extended 6 = 10
cm, and the required work is
IF = ⅛ (10) 2 = ⅛ = 3.125 d y n -c m r__ 1
EXAM PLE 3 Work performed in pumping out a tank of water
A tank in the shape of a right circular cone of height 12 ft and radius 3 ft is

inserted into the ground with its vertex pointing down and its top at ground
level, as shown in Figure 6.23a. If the tank is half-filled with water (density
p = 62.4 lb ∕ft 3), how much work is performed in pumping all the water in the
tank to ground level?
Solution Set up a coordinate system with the origin at the vertex of the cone and
the j-axis as the axis of symmetry (Figure 6.23b). Partition the interval
0 ≤ y ≤ 6 (remember, the tank is only half full). Choose a representative
point y k in the kth subinterval, and construct a thin disk-like slab of water
a. A conical water tank units above the vertex of the cone. Our plan is to think of the water in the tank
as a collection of these slabs of water piled on top of each other. We shall find
the work done to raise a typical water disk and then compute the total work by
using integration to add the contributions of all such slabs.
Note that the force required to lift the slab is equal to the weight of the slab,
which equals its volume multiplied by the weight per cubic foot of water. Let
x k be the radius of the ∕<th slab. By similar triangles (Figure 6.23b), we have
* *
χ 3 * yk
k or
7 = T2 x k = 4
b. A disk of water is 12 - units Hence, the volume of the slab is

jy*.
k
from the top.
j , Thickness
Figure 6.23 Work in pumping ∕y , ∖ 2
water ΔE= ⅛ Δ¾ = ^ ΔΛ
Radius
*A unit of measurement for work is the erg and is defined as 1 erg = I dyn ■cm. It is used in the
CGS system, where length is in centimeters, mass is in grams, and time is in seconds. The unit 2
dyn ∙ cm used in this problem would be more commonly expressed as 2 ergs.
so that the weight o f the slab is

∕y * ∖ 2
62.477(-4 ) ∆v jt lb
From Figure 6.23b, we see that the slab o f water must be raised 12 - y* feet,
which means that the work Δ1F required to raise this slab is
∕y * ∖ 2
Δ W = 6 2 . 4 τ r ( ( 1 2 - y*) ∆ y k
Finally, to compute the total work, we add the work required to lift each thin
slab and take the limit o f the sum as the norm o f the partition approaches 0,
(that is, as n → ∞)ι
PF =B⅛ 62∙47Γ⅛ (J P 1 2 - Λ)Δ¾

= 62.4τr ∫q θ f p 12 - y) dy = ¾ 4 ττ ∣ θ (12y2 - y≡) dy
a. Pascal's principle: In an = 3.9τ7^4y3 - ∣ y 4 Ψ = 2,106ττ≈ 6,616 ft-lb m

equilibrium state, the fluid
pressure is the same in all
directions.
■ FLUID PRESSURE AND FORCE
Anyone who has dived into water has probably noticed that the pressure (that is,
the force per unit area) due to the water’s weight increases with depth. Careful
observations show that the water pressure at any given point is directly propor
tional to the depth at that point. The same principle applies to other fluids.
In physics, Pascal’s principle (named for Blaise Pascal (1623-1662), a French
mathematician and scientist) states that fluid pressure is the same in all directions.
b. The pressure is constant on a
horizontal surface but varies This means that the pressure must be the same at all points on a surface
with depth on a surface sub submerged horizontally, but if the surface is submerged vertically, or at an angle,
merged vertically or at an angle. the pressure varies with the depth at each point on the surface, as shown in Figure
Figure 6.24 Pressure on a 6.24.
submerged object We can now state the following formula for fluid force.
Fluid Force I f a surface o f area A is submerged horizontally at a depth h in a fluid, the

weight of the fluid exerts a force of
F = (PRESSURE)(AREA) = phA
on the surface, where p is the fluid density (weight per unit volume). This force
is called the fluid pressure or hydrostatic pressure.
Weight-density, p (lb∕ft 3) It turns out that this fluid force does not depend on the shape o f the container
or its size. For instance, each of the containers in Figure 6.25 has the same
Water 62.4 pressure on its base because the fluid depth h and the base area A is the same in
Seawater 64.0 each case.
Gasoline 42.0
Kerosene 51.2
Milk 64.5
Mercury 849.0
Figure 6.25 The fluid force F = phA does not depend on the shape or size of the container.
To see how integration can be used to compute a fluid force, let us consider a
plate that is submerged vertically in a fluid of density p as shown in Figure 6.26.
First, we set up a coordinate system with the horizontal axis on the surface of
the fluid and the positive vertical axis (the “//-axis”) pointing down, so greater
depths correspond to larger values of h (see Figure 6.26b). For simplicity, we
assume the plate is oriented so its top and bottom are located a units and b units
below the surface, respectively.
Our strategy will be to think of the plate as a pile of subplates or slabs, each so
thin that we may regard it as being at a constant depth below the surface. A typical
slab (a horizontal strip) is shown in Figure 6.26b. Note that it has thickness ∆∕v and
a. A plate submerged vertically length L = L(h), so that its area is ΔA = L∆,h. Because the slab is at a constant
in a fluid o f density p depth h below the surface, the fluid force on its surface is
Fluid____________
Δ F = (PRESSURE)(AREA) = ph LA = ρhL(h) Lh
surface
and by integrating as h varies from a to b, we obtain the total force on the plate.
h
width
’ ___ Λ,' Fluid (Hydrostatic) Force
h length L(A)
Suppose a flat surface (a plate) is submerged vertically in a fluid of weight
density p (lb∕ft3) and that the submerged portion of the plate extends from
b. A typical slab has area x = α to x = ⅛ona vertical axis. Then the total force, F, exerted by the fluid is
A = L ( ∕∣) Δ Λ . given by
Figure 6.26 Hydrostatic pressure r,Z>
F= ph(x) L(x) dx
Ja
where h(x) is the depth at x and L(x) is the corresponding length of a typical
horizontal approximating strip.
EXAMPLE 4 Fluid force on a vertical surface
The cross sections of a certain trough are inverted isosceles triangles with
height 6 ft and base 4 ft, as shown in Figure 6.27. If the trough contains water
to a depth of 3 ft, find the total fluid force on one end.
Ψ∕1
a. A trough with half-filled b. Side view o f trough c. Use similar triangles

triangular cross sections to find ~ - 2— 4
4 6
Figure 6.27 Fluid force
Solution First, we set up a coordinate system in which the horizontal axis lies at
the fluid surface and the positive vertical axis (the //-axis) points down (see
Figure 6.27b). Next, we find expressions for the length and depth of a typical
thin slab (horizontal strip) in terms of the variables. We assume the slab has
thickness ∆∕z, and if its length is L , then by similar triangles (see Figure 6.27c)
we have
⅜= so that L = ∣(3 - Λ)
We see that the approximating slab has area
NA = L ∆ h = ⅜ (3- h)∆h
Finally, multiply the product of the fluid’s weight-density at a given depth (ph)
and the area of the approximating slab (ΔA) and integrate on the interval of
depths occupied by the vertical plate.
F = f jp∕z(3 - ∕z) dh
Jo χ___ .___ /
ΔF = force on a typical slab
= ⅞(62.4) (3h - h~) dh p = 62.4 lb ∕ft 3 for water

Jo
= 41.6(∣∕z2 - ∣∕z 3)∣θ = 187.2 lb
The fluid force formula applies equally well if the plate is submerged in a fluid
whose density varies with the depth h.
EXAM PLE 5 Total force on one face of a dam
A reservoir is filled with water to the top of a dam. If the dam is in the shape of
a parabola 40 ft high and 20 ft wide at the top, as shown in Figure 6.28, what is
the total fluid force on the face of the dam?
Solution Instead of putting the x-axis on the surface of the water (at the top of the
dam), we place the origin at the vertex of the parabola and the positive y-axis
along the parabola’s axis of symmetry (Figure 6.28b). The advantage of this
choice of axes is that the parabola can be represented by an equation of the
a. Cross section of a dam
form v = ex2 . Because we know that y = 40 when x = 10, it follows that
40 = c(10)2 so that c = $ and thus y = ∣ x 2
The typical horizontal strip on the face of the dam is located y feet above the
x-axis, which means it is h = 40 —y feet below the surface of the water. The
strip is ∆y feet wide and L = 2x feet long. Write L as a function of y:
y = ∣ x 2 so that x = J⅜y
and
b. A typical horizontal strip is

L = 2x = = V 10ι,
40 - y ft below the water
surface. Therefore
Figure 6.28 Force on one face of a NA = L ∆ y = 2 y∣ ^y ∆y = V 10r ∆y
dam
Finally,
Depth below water: h = 40 - y
/-40 1 ∕∙40
F = ph U h ) clh = (62.4)(√Tθ )(40 - y )V y dy
Jo Λ '- ^— ‘ Jo
Area of the strip
Density of water = 62.4
,___ f4 0
= 62.4√10 (40,V IZ2 - y 3z2) dy
Jo
= 62.4VTθ (ψ y 3z2 - j v 5z2)∣40°
= 62.4V10 (y 322) ( ^ - ∣ v ) ∣ 400
= 532,480 lb
■ CENTROID AND MOMENT OF A PLANE REGION
In mechanics, it is often important to determine the point where an irregularly

shaped plate will balance. The moment of a force measures its tendency to produce
rotation in an object and depends on the magnitude of the force and the point on
the object where it is applied. Since the time of Archimedes (287-212 B.c.), it has
been known that the balance point of an object occurs where all its moments
cancel out (so there is no rotation).
The mass of an object is a measure of its inertia; that is, its propensity to
maintain a state of rest or uniform motion. A thin plate whose material is
distributed uniformly, so that its density p (mass per unit area) is constant, is
called a homogeneous lamina. The balance point of such a lamina is called its
centroid and may be thought of as its geometrical center. We shall see how
centroids can be computed by integration in this section, and shall examine the
Theoretically, the lamina should balance topic even further in Section 13.6.
on a point placed at its centroid.
Consider a homogeneous lamina that covers a region R bounded by the curves
y = f(x ) and y = g(x) on the interval [a, ⅛], and consider a thin, vertical
approximating strip within R , as shown in Figure 6.29. The mass of the strip is
given by
∆m = p ∙ [∕( x ) - g(x)]∆Λ-
Area of strip
Density
and the total mass of the lamina may be found by integration:

r∕)
m = P [∕(Λ-) - g(Λ-)] dx
Ja
The geometrical center of the approximating strip is (x, y), where x = x and
y = ⅛ [∕W + g(x)]∙ The moment of the approximating strip about the j-axis is
defined to be the product
Δ Λ f, = x ∙ ∆m = x{p[∕(x) - g(xj] ∆x}
Mass of strip
Distance of the
strip from y-axis
Figure 6.29 Homogeneous lamina
This product provides a measure of the tendency of the strip to rotate about the
with a vertical approximating strip
y-axis. Similarly, the moment of the strip about the x-axis is defined to be the
product
⅛m x = y ∆m = ∣[ ∕( x ) + g(x)] {p[f(x) - g(x)] ∆x}
Average distance o f the
strip from x-axis
= ⅛{[∕(λ')]2 - [g(x)]2}∆ x
This product provides a measure of the tendency of the strip to rotate about the
x-axis.
Integrating on [a, b], we find the moments of the entire lamina R about the
y-axis and x-axis are given by
M = P ( Λ'[∕(x) - g(x)] dx and M χ = ∣p f {[∕(x )] 2 - [g(x)]2} dx
y
Ja Ja
If the entire mass m of the lamina R were located at the point (x, y), then its
moment about the x-axis and y-axis would be my and m x, respectively. Thus, we
have nix - M and my = M χ, so that
rb
m P x [f(x ) - £(*)] dx
m ⅛∫ {[∕W]2- l⅛(∙x)]2} dx
x = —- = — -------------------------- and
m rb m ~ rb
P [ ∕W “ T(-x )] dx P [f(x) - #(*)] dx
We can now summarize these results.
Let f and g be continuous and satisfy f(x ) ≥ g(x) on the interval [a, ⅛ and
consider a thin plate (lamina) of uniform density p that covers the region R
between the graphs of y = ∕( x ) and y = g(x) on the interval [a, b]. Then
fb
Mass The mass of R is: m = p [∕(x) - g(x)] dx.
Ja
Centroid The centroid of R is the point (x, y) such that
px[∕(x) - g(x)] dx 2Jaf P{ [ ∕W ] 2 - LgrC0J2} dx

-------------------------- a n d y=~⅛ rb
p [∕(*) - g(x)∖ dx P∖M ~ ■
?(*)] dx
Ja
EXAM PLE 6 Centroid on a thin plate
A homogeneous lamina R has constant density p = 1 and is bounded by the

parabola y = x 2 and the line y = x. Find the mass and the centroid of R.
Solution We see that the line and the parabola intersect at the origin and at the
point (1, 1), as shown in Figure 6.30. Because p - 1 , the mass is the same as
the area of the region R . That is,
rl ∕ r 2 r 3 ∖ ∣l ι
m = A = (x - x 2) dx = ⅞- - ⅜- I = -7
Jo ∖2 3 / ∣o 6
and we find that the region has moments M γ and M χ about the y-axis and
x-axis, respectively, where
Figure 6.30 The centroid of a M y= ∕ 0 x (x ^^x2) dx = ( τ ^^ ⅞ ) l o = ⅛
planar region
and
Thus, the centroid of the region R has coordinates

ι , - < t ⅛ 2
12 _ 1
and =W =
I 2
6
■ VOLUME THEOREM OF PAPPUS
In certain geometric applications, it is convenient to speak of the center of mass

(χ, y) without referring to a plate covering R . In this context, the center of mass is
the centroid of the region R . We shall use this interpretation in the following
theorem, which is usually attributed to the 4th century Greek geometer Pappus.
T H E O R E M 6.1 Volume theorem of Pappus
The solid generated by revolving a region R about a line outside its boundary (but
in the same plane) has volume V = As, where A is the area of R and 5 is the
distance traveled by the centroid of R.
Proof: First, choose a coordinate system in which the y-axis coincides with the
axis of revolution. Figure 6.31 shows a typical region R together with a vertical
approximating rectangle. We shall assume that this rectangle has area ∆ ∕l and that
it is located x units from the y-axis.
Notice that when the rectangle is revolved about the y-axis, it generates a shell
of volume Δ K = 2πxΔA. Thus, by partitioning the region R into a number of
rectangles and taking the limit of the sum of the volumes of all the related
approximating shells as the norm of the partition approaches 0, we find that
n
V = lim ∑ 2 π xk.Δ A ,k = f 2ττx dA = 2τr f x dA
∣∣∕>∣∣→0 k≡∖
_ _ f x dA
— 2 τrx f dA Because x = j-
Figure 6.31 The volume theorem = 2πx A Because f dA = A
of Pappus
= As Where s = 2πx is the
distance traveled by the
centroid (the circumference
of a circle of radius x) ■
We close this section with an example that illustrates the use of the volume
theorem of Pappus.
EXAM PLE 7 Volume of a torus using the volume theorem of Pappus
When a circle of radius r is revolved about a line in the plane of the circle
located R units from its center (R > r), the solid figure so generated is called a
torus, as shown in Figure 6.32. Show that the torus has volume V = 2π 2r 2 R.
Axis of revolution
The centroid travels 2πR units.

Figure 6.32 The volume of a torus
Solution The circle has area A = πr2 and its center (which is its centroid) travels
2πR units. In this example s = R. Therefore, according to the theorem of
Pappus, the torus has volume
V = (2ττR)A = 2ττR(πr2) = 2τr2r2R h m
Incidentally, the volume found in Example 7 using the theorem of Pappus was
computed in Problem 66 of Section 6.1 by using shells.
PROBLEM SET 6 .3
θ 1. ■ What Does This Say? Discuss work and how to find force against the end o f the tank. The weight densities are given
it. on page 407.
2. ■ What Does This Say? Discuss fluid pressure and how 12. water 13. seawater
to find it.
3. ■ What Does This Say? What is meant by a centroid?
Outline a procedure for finding both mass and centroid.
4. What is the work done in lifting a 50-lb bag of salt 5 ft?
5. What is the work done in lifting an 850-lb billiard table 15
ft?
6. How much work is done in lifting a 50-lb weight to a
height of 6 ft?
7. A 5-lb force will stretch a spring 9 in. beyond its natural
length. How much work is required to stretch it 1 ft
beyond its natural length? 15. kerosene
8. Suppose it takes 4 dyn ■cm of work to stretch a spring 10
Radius of semicircle
cm beyond its natural length. How much work is needed
to stretch it 4 cm further?
9. A bucket weighing 75 lb when filled and 10 lb when empty
is pulled up the side of a 100-ft building. How much more
work is done in pulling up the full bucket than the empty
bucket?
10. A 30-ft rope weighing 0.4 lb∕ft hangs over the edge of a
building 100 ft high. How much work is done in pulling
16. water 17. mercury
the rope to the top of the building?
11. A 30-lb ball hangs at the bottom of a cable that is 50 ft
long and weighs 20 lb. The entire length of cable hangs
over a cliff. Find the work done to raise the cable and get
the ball to the top of the cliff.
In Problems 12-17, the given figure is the vertical cross section
o f a tank containing the indicated substance. Find the liquid
Find the centroid o f each planar region in Problems 18-23. 33. If water leaks from the bottom of the tank in Problem 32,
18. the region in the first quadrant bounded by the curves how much work is done?
y = x i and y = Λ√ZX 34. A cylindrical tank with base radius 2 ft and height 12 ft is
19. the region bounded by the parabola y = x 2 - 9 and the sunk into the ground so that the top of the tank is at
x-axis surface level. Find the work done to pump the liquid to
the top of the tank if the liquid density is 80 lb ∕ft 3 and the
20. the region bounded by the parabola y = 4 —x 2 and the
tank is filled to the top.
line y = x + 2
35. A hemispherical bowl with radius 10 ft is filled with water
21. the region bounded by the curve y = x ~ ', the x-axis, and
to a level o f 6 ft. Find the work done to the nearest 100 ft-
the lines x = 1 and x = 2
lb to pump all the water to the top of the bowl.
22. the region bounded by the curve y = x 1 and the line
2x + 2y = 5 36. A holding tank has the shape of a rectangular parallelepi
ped 20 ft by 30 ft by 10 ft. How much work is done in
23. the region bounded by y — ~ = and the x-axis on [1, 4] pumping all the water to the top of the tank? How much
work is done in pumping all the water out of the tank to a
Use the theorem o f Pappus to compute the volume o f the solids height of 2 ft above the top o f the tank?
generated by revolving the regions given in Problems 24-27
about the prescribed axis.
24. the region bounded by the parabola y = V x , the x-axis,
and the line x = 4 about the line x = - 1
25. the triangular region with vertices ( -3 , 0), (0, 5), and
(2, 0), about the line y = —1
26. the semicircular region y = V 1 —x 2 , about the line
y = -1
27. the semicircular region x = V 4 - y 2 , about the line
x = -2
28. The centroid of any triangle lies at the intersection of
20 ft
the medians, two-thirds the distance from each vertex to
the midpoint of the opposite side. Locate the centroid of
the triangle with vertices (0, 0), (7, 3), and (7, -2 ) using 37. A cylindrical tank of radius 3 ft and
this method, and then check your result by calculus. height 10 ft is filled to a depth of 2 ft
with a liquid of density p = 40 lb ∕ft 3.
29. The centroid of any triangle lies at the intersection of the
Find the work done in pumping all
medians, two-thirds the distance from each vertex to the
the liquid to a height o f 2 ft above the
midpoint o f the opposite side. Locate the centroid for
top of the tank.
the triangle with vertices ( -2 , 0), (3, 5), and (3, -2 ) using
this method, and then check your result by calculus. 38. An oil can in the shape o f a rectangular parallelepiped is
filled with oil of density 0.87 g ∕cm 3 . What is the total
30. An object moving along the x-axis is acted upon by a force
force on a side of the can if it is 20 cm high and has a
F(x) = x 4 + 2x 2 . Find the total work done by the force in
square base 15 cm on a side?
moving the object from x = 1 cm to x = 2 cm. Note:
Distance is in cm, so force is in dynes. 39. A swimming pool 20 ft by 15 ft by 10 ft deep is filled with
water. A cube I ft on a side lies on the bottom o f the pool.
31. An object moving along the x-axis is acted upon by a force
Find the total liquid force on a lateral face of the cube.
F(x) = ∣sinx∣. Find the total work done by the force in
moving the object from x = 0 cm to x = 2ιr cm. Note: 40. In Problem 39, suppose a log of radius I ft lies at the
Distance is in cm, so force is in dynes. bottom of the pool. What is the total liquid force on one
end of the log?
32. A tank in the shape of an inverted right circular cone of
height 6 ft and radius 3 ft is full of water. How much work 41. In Problem 39, suppose a brick with sides o f lengths 2 in.,
is performed in pumping all the water in the tank over the 3 in., and 6 in. stands upright on the bottom of the pool.
top edge? What is the total liquid force on the other five sides?
42. An object weighing 800 lb on the surface of the earth is
propelled to a height o f 200 mi above the earth. How
much work is done against gravity? Hint: Assume the
radius of the earth is 4,000 mi, and use Newton’s law
F = -k ∕ x 2 , for the force on an object x miles from the
center o f the earth.
43. A dam is in the shape of an isosceles trapezoid 200 ft at
the top, 100 ft at the bottom, and 75 ft high. When water
is up to the top of the dam, what is the force on its face?
44. According to Coulomb’s law in physics, two similarly

charged particles repel each other with a force inversely
c. Derive this formula for the volume o f the fuel at
proportional to the square of the distance between them.
level h, where - b ≤ h ≤ b:
Suppose the force is 12 dynes when they are 5 cm apart,
rh
a. How much work is done in moving one particle from a
V(h) = 2La
distance o f 10 cm to a distance of 8 cm from the other? J -b
b. How much work is done in moving one particle from
Hint. Recall that the volume of a cylinder, o f any
an “ infinite” distance to a distance of 8 cm from the
shape, is the “ area × length.”
other? Hint. Find the work done in moving it from
distance .v to distance 8 and then let .s, → +∞. d. Finally, for a = 5 and b = 4, numerically compute
V(h) for h = - 3 , —2, . . . ,3, 4. Note: F(0) and
A region R with uniform density is said to have moments o f Iz (4) will serve as a check on your work.
inertia I χ about the x-axis and I about the y -a x is, where 49. We think of the last part of Problem 48 as the
“ direct” problem: Given a value o f h, compute F(Λ)
I χ = ∫ y 2 dA and I y = ∫ x 2 dA
by a formula. Now we turn to the “ inverse” problem:
The dA represents the area o f an approximating strip. In For a set value of V, say F o, find h such that
Problems 45-46, fin d Iχ and I y for the given region. V(h~) = Ffl. Typically, the inverse problem is more
45. the region bounded by the parabola y = V x , the x-axis, difficult than the direct problem and requires some
and the line x = 4. sort o f iterative method to approximate the solution
(for example, Newton-Raphson’s method). For ex
46. the region bounded by the curve y = ΛVZ 1 - x 3 and the
ample, if F o = 500 ft 3, we could define
coordinate axes
47. If a region has area A and moment of inertia Z about the f(h) = F(A) - 500
j-axis, then it is said to have a radius of gyration
and seek the relevant “ zero” off . Set up the necessary
py ~ 'h computer program to do this and find the h values, to
two-decimal place accuracy corresponding to these
values o f V:
with respect to the jy-axis. Compute the radius of gyration a. 500 b. 800
for the triangular region bounded by the line y = 4 - x
Hint: A key to this problem is thinking about V'(fι),
and the coordinate axes. where V(h) is expressed in 48c. Also, for starting
values h 0 , refer to your work in Problem 48d.
Computational Window θ 50. Show that the centroid of a rectangle is the point of
intersection o f its diagonals.
In these computational problems we revisit a problem 51. A right circular cone is generated when the region bound
similar to the one we considered at the end o f Chapter 4, ed by the line y = x and the vertical lines x = 0 and x = r
but add a couple o f complications. Imagine a cylindrical is revolved about the x-axis. Show that the volume o f this
fuel tank lying on its side [o f length L = 20 ft); the ends are cone is
elliptically shaped and are defined by the equation
V = ∣π r 3
52. Find the volume of the solid figure generated when a

JVe are concerned with the amount offuel in the tank for a square of side L is revolved about a line that is outside the
square and parallel to two of its sides, and located 5 units
given level.
from the closer side.
48. a. Explain why the area of an end o f the tank can be
expressed by 53. Find the volume of the solid figure generated by revolving
an equilateral triangle o f side L about one of its sides.
54. Let R be a region in the plane, and suppose that R can be
partitioned into two subregions R 1 and R y Let (x 1, y 1) and
(x2, y 2) be the centroids of R l and R 2 respectively, and let
b. By making the right substitution in the integral in
∠41 and A 2 be the areas o f the subregions. Show that (x, y)
part a, one can get an integral representing half the
is the centroid o f R , where
area inside the circle (with unit radius). Do this
and obtain the area inside the above ellipse to be ^41x 1 + A 2X 2 √ V I + Λ 2T2
πab. and
A l + T, Λ∣ + A 2
6 . 4 GROWTH, DECAY, AND FIRST-ORDER LINEAR DIFFERENTIAL EQUATIONS
IN THIS SECTION Radioactive decay: carbon dating, exponential and

logistic growth, first-order linear differential equations, applications of first-
order linear equations
We investigate applications involving separable differential equations of the
general forms
dy =
— + b, d, _dy = (m j + b)y
,.
m y an
We also see how to solve equations of the form
+ P(x)y = Q(x)
We see how this type of differential equation can be used to model dilution
problems and problems involving electrical circuits.
■ RADIOACTIVE DECAY: CARBON DATING
The decomposition of a radioactive substance is a very complicated process, but

experiments indicate that such substances disintegrate at a rate proportional to
the amount present at each instant. Thus, if Q(t) is the amount of a given
radioactive substance present at time Z, we have
where k is a positive constant (the minus sign indicates that Q decreases with
time). This is called the decay equation for the substance. Let QQ be the initial
amount of the substance present.
To solve the decay equation, separate the variables and integrate both sides.
d Q ,k nQ
dt ~
fd Q =
f , ,
H∑ I ~k c t
ln∣(2∣ = —kt + C 1
e -kt+c l — Q Definition o f natural logarithm
e ~kt e C∖ = Q
Thus, Q = Ce~ kt , where C = e c ι. Finally, because Q = Q o when Z = 0, we see that

O 0 = C c 0 or C=Q 0
Decay Equation The decay equation of a radioactive substance is

2(Z) = <20e~k '
where Q(t) is the amount of the substance present at time Z, Q o is the initial
amount of the substance, and k is a positive constant that depends on the ∣
substance.
We shall refer to the graph of the decay equation as the decay curve. Such a
curve is shown in Figure 6.33. In Figure 6.33 we have also indicated the time tlι
required for half of a given substance to disintegrate. The time th is called the half
life of the substance, and it provides a measure of the substance’s rate of
disintegration.
Sec. 6.4 Growth, Decay, and First-Order Linear Differential Equations 417
There is a simple relationship between k and the half-life th . In particular,

because a sample Q o grams will contain only Ô0 grams after th years, we have
<2⅛) = Q 0 e~kt'∙ = ⅛β0

so that
ρ-kt ∣j —
By taking logarithms on both sides of this equation, we find that
(half life) —kth = In I = —In 2
ln 2
Figure 6.33 The decay curve for a t ∣' = k
radioactive substance
Half-Life Formula A radioactive substance that decays according to the decay equation Q = Q 0 e~k ,
has half-life
. In 2
t ∣>~ k I⅛
⅛⅛
⅛≡
⅛⅛
∙⅛W<τ≡⅛ ' . . ⅛⅛≡
This formula can be used to determine th if we know k or k if we know th . Note that

the half-life th is independent of the initial amount Q o .
E X A M P L E 1 Amount of a radioactive substance present, given half-life
A particular radioactive substance has a half-life of 600 yr. Find k for this
substance and determine how much of a 50-gm sample will remain after
125 yr.
Solution According to the half-life formula, we have
k = ⅛ 2 = ⅛ ≈ 0.001 155 245 3

th o(Jl)
Next, to see how much of a 50-gm sample will remain after 125 yr, substitute
Q o = 50, k = 0.001 155 245 3, and t = 125 into the decay equation:
2(125) = 50e-° 0° 1 155 245 3(125) ≈ 43.27682805
There will be about 43 gm present. ∣≡∑≡
One of the more interesting applications of radioactive decay is a technique

known as carbon dating, which is used by geologists, anthropologists, and
archeologists to estimate the age of fossils and other objects.** The technique is
based on the fact that all animal and vegetable systems (whether living or dead)
contain both stable carbon l2 C and a radioactive isotope 14C. Scientists assume
that the ratio of 14C to l2 C in the air has remained approximately constant
throughout history. Living systems absorb carbon dioxide from the air, and so the
ratio of 14C to 12C in a living system is the same as that in the air itself. When a
living system dies, the absorption of carbon dioxide ceases. The 12C already in the
system remains while the 14C decays, and the ratio of 14C to 12C decreases
exponentially. The half-life of 14C is approximately 5,730 years. The ratio of l4 C to
*Carbon dating is used primarily for estimating the age of relatively “recent” specimens. For
example, it was used (along with other methods) to determine that the Dead Sea Scrolls were written
and deposited in the Caves of Qumran approximately 2,000 years ago. For dating older specimens, it is
better to use techniques based on radioactive substances with longer half-lives. In particular, potassium
40 is often used as a “clock” for events that occurred between 5 and 15 million years ago.
Paleoanthropologists find this substance especially valuable because it often occurs in volcanic
deposits and can be used to date fossils trapped in such deposits.
12C in a fossil t years after it was alive is approximately

R = R 0e~kl
where k = g γgθ and R Q is the ratio of l4 C to 12C in the atmosphere. By comparing

R(t) with R o , the age of the object can be estimated.
EXAM PLE 2 Carbon dating
An archaeologist has found a fossil in which the ratio of 14C to l2 C is ∣ the ratio
found in the atmosphere. Approximately how old is the fossil?
Solution The age of the fossil is the value of t for which R(t) = J R Q :
15^/, x?0 = R 0eL ~kt
1 = p-kt
e
5
- k t = In |
t= In 0.2 k= 5 730
≈ 13,304 64798
The fossil is approximately 13,305 yr old. M ≡
■ EXPONENTIAL AND LOGISTIC GROWTH
When the population P(Z) of a colony of living organisms (humans, bacteria, etc.)
is small, it is reasonable to expect the relative rate of change of the population to
be constant. In other words,
dP
j⅛- = k or ^P = kP
where A: is a constant (the unrestricted growth rate). This is called exponential

growth. As long as the colony has plenty of food and living space, its population
will obey this unrestricted growth rate formula and P(f) will grow exponentially.
However, in practice, there always comes a time when environmental factors
begin to restrict the further expansion of the colony, and at this point, the growth
ceases to be purely exponential in nature. To construct a population model that
takes into account the effect of diminishing resources and cramping, we assume
that the relative rate of change in population is decreased by a braking factor that
is proportional to the population itself. Thus, we have
dP
⅛L = k - L P or ^ = ( k - LP)P
where L is a positive constant (the braking rate). This is called a logistic equation,
and it arises not only in connection with population models, but in a variety of
other situations. The following example illustrates one way such an equation can
arise.
EXAM PLE 3 Logistic equation for inhibited growth
The rate at which an epidemic spreads through a community is proportional

to the product of the number of residents who have been infected and the
number of susceptible residents who have not. Express the number of
residents who have been infected as a function of time.
Solution Let Q(f) denote the number of residents who have been infected by time
t and B the total number of susceptible residents. Then the number of
susceptible residents who have not been infected is B - Q, and the differential
equation describing the spread of the epidemic is
W = ∕<β(B - β) k is the constant o f

proportionality.
dQ , j
Q(B - Q) k cit
f dQ _ f
J Q(B ~ Q) ~ J k dt
The integral on the right causes no difficulty, but the integral on the left is not
easy to integrate in its present form. Consider the following algebraic
manipulation, which allows us to rewrite the troublesome integral in a useful
form for our purposes in this application:
1 v i pg - <2 + Q I = 1
B -Q ) B YQ(B-Q) ∖ Q(B - Q)
We now substitute the form on the left into the equation and complete the
integration:
1 f dQ ι [ dQ = [
B J Q + B J B ~ Q J kdt
p ln∣β∣ - ln∣5 — Q ∖= kt + C Integrate each.
± ln Q
B m B - Q —kt + C Division property o f logs
Inf p = Bkt + B C Multiply both sides by B.

∖B - β ∕ {Note that Q > 0, B > Q.)
Q — o Bkt+BC
Definition o f In
B -Q e
Q = ( B - Q)e Bkt e B C Because eBC is a constant,

we can let A = e BC .
Q = Λ ,B e flkt —A l Qe βkt
Q + A χQe ε k t = A l Be βkt
A l Be ε k t
2 - 1 + A l e βkt
To simplify the equation, we make another substitution; let A = J - , so that

(after several simplification steps) we have 1
Be akt
A B
, 1 + Ae~ βkt
i1 + A
The graph of β(Z) is shown in Figure 6.34. Note that the curve has an
inflection point where
Q(0 = f
(The details are left for the reader.) This corresponds to the fact that the
epidemic is spreading most rapidly when half the susceptible residents have
been infected.
TABLE 6.3 Summary of Growth Models
Inhibited growth ⅛ = ^ - β ) Long-term population

or logistic growth growth (with a limiting
Rate is jointly propor Solution: value); spread of a disease
tional to the amount in a population; sale of fad
present and to the 0 = ___ B___ items (for example, singing
1 + Ae~β kl
difference between flowers); growth of a
the amount present business
and a fixed amount
(k > 0).
⅛ = k(β - 0
Limited growth Learning curve; diffusion of
Rate is proportional information by mass media;
to the difference Solution: intravenous infusion of a
between the amount medication; Newton’s law of
present and a fixed Q = B - Ae~kt cooling; sales of new products;
limit (k > 0). growth of a business
■ FIRST-ORDER LINEAR DIFFERENTIAL EQUATIONS
Equations that can be expressed in the form

^^∙ + P(x)y = <2(x)
are called first-order linear differential equations. For example,

χ2 ^
c ~ ^ + 2 ^y = χ 5
is such an equation, with

P(x) = - f v -y>2) = - 1 - 2x^2 and Q(x) = p = x 3
The following theorem shows how the general solution of a first-order linear
differential equation y ' + P(x)y = Q(x) may be obtained.
THEOREM 6.2 General solution of a first-order linear differential equation

The general solution of the first-order linear differential equation
+ P(x)j = <2(x)
is given by
y = 7 ⅛ [/ 2( χ )z (λ∙) dx + C
where Z(x) = e ∣p <χ )d-r .

Proof: First, note that
= e ∖p(x)dx∖jL f P(Λ ) D Χ ∖= e {f(x)dx p( χ ) = ∕( λ-)jp(χ)
Ct LCl∙λ, I J
because = ^v'c also will need to note that
⅛ [I(x)y ] = z W + T ⅛ I(x) = I(*) f f + y z (λ ') p (x ) Product rule
We now begin the proof by starting with the given differential equation.
+ P(x)y = β(x) Given

z (*)[⅛j + F(x)τ] = Z(χ)β(χ) Multiply both sides by I(x).
∕U ) + Kχ)P{×)y = Jr(χ)Q(χ)
^[7(χ)y] = f(χ)<2(χ) Substitute.
^ [∕(x )y ] dx = J ∕(x)Q(x) dx Integrate both sides.

∕(x)y = J ∕(x)β(x) dx + C
y = 7⅛) / 7 (x )β( x ) dx+ c
Because multiplying both sides of the differential equation

Historical Note
+ P(x)y = Q(x)
by ∕(x) makes the left side an exact derivative, the function ∕(x) is called an
integrating factor of the differential equation.
A first-order linear differential equation together with the condition y = y 0
when x = x0 is called an initial value problem. Here is an example of such a
problem.
EXAMPLE 4 First-order linear differential equation
Solve = e~x - 2 y,x ≥ 0, subject to the initial condition y = 2 when x = 0.
Solution The differential equation can be expressed in the proper form by

adding 2y to both sides:
GUSTAV DIRICHLET (1805-1859)
Gustav Lejune Dirichlet was a
⅛ + 2 ∙ y = e ~x for x ≥ 0
professor of mathematics at the
University of Berlin and is We have F(x) = 2 and β(x) = e~x . Both ∕ j (x) and Q(x) are continuous in the
known because he was instru domain x ≥ 0. An integrating factor is given by
mental in formulating a rigor
ous foundation for calculus. He ∕(x) = e^F(x> dx = e^-dx = e2x for x ≥ 0
was not known as a good
We now use the first order linear differential equation theorem, where
teacher. His nephew wrote that
the mathematics instruction he ∕(x) = e 2x and <9(x) = e~x , to find y.
received from Dirichlet was the y = j⅛*[J e2xe~x dx + c]
most dreadful experience of his
life. Howard Eves tells of the
time Dirichlet was to deliver a = -¼ [ex + C I = e~x + e~2x C for x ≥ 0
lecture on definite integrals, but exL J
because of illness he posted the
following note: To find C, note that because y = 2 when x = 0, 1 = C. So the solution is
y = e~x + e~2x , x ≥ 0.
Because of illness I cannot
lecture today APPLICATIONS OF FIRST-ORDER LINEAR EQUATIONS
Dirichlet
The students then doctored the EXAMPLE 5 A dilution problem

note to read:
A tank contains 20 lb of salt dissolved in 50 gal of water. Suppose 3 gal of
brine containing 2 lb of dissolved salt per gallon runs into the tank every
'M ichaelmas
minute and that the mixture (kept uniform by stirring) runs out of the tank at
JEaster
(Because of illness 1
cannot lecture today) ¢/(1 Frdor)

Dirichlet
the rate of 2 gal/min. Find the amount of salt in the tank at any time t. How
much salt is in the tank at the end of one hour?
Solution Let S(t) denote the amount of salt in the tank at the end of t minutes.
Michaelmas and Easter were Because 3 gal of brine flows into the tank each minute and each gallon
school holidays, and 1 Frdor contains 2 lb of salt, it follows that (3)(2) = 6 lb of salt flow into the tank each
(Fredrichsd’or) was the custom minute (see Figure 6.35). This is the inflow rate.
ary honorarium for a semester’s
For the outflow, note that at time t, there are S(f) lb of salt and 50 +
worth of lectures.
(3 - 2)t gallons of solution (because solution flows in at 3 gal/min and out at
2 gal/min). Thus, the concentration of salt in the solution at time t is
5 0 + l l b ⅛a l
Sec. 6.4 Growth, Decay, and First-Order Linear Differential Equations 42 3
Inflow rate of salt: and the outflow rate of salt is

6 Ib/min
θ-
,
■SV) lb∕gal] ΓO 1/ ■ 1 -*S ( 0 11 /
[2 gal∕mιn] = y lb∕mιn
50 ÷ t f
Combining these observations, we see that the net rate of change of salt dS/dt
is given by the first-order linear differential equation
dS _ , 2S
dt →- 50 + Z
↑
τ rι Outflow
Inflow
Outflow rate of salt: This differential equation can be written as

*L lb ∕ min dS , I S _ ,
50 + r dt 50 + l 0
Figure 6.35 Rate o f flow equals
inflow rate minus outflow rate. which we recognize as a first-order linear differential equation with
p <'∣ = 5 < ⅛
and Q(t) = 6. The integrating factor is

/(/) = e ∖p(t)dt = J( 2 Λ ) ∕( 5 O + ∕) = e 21n∣50+Z∣ = (5 Q + ; )2
and the general solution is

■SV) = ∙(5Q+7)2 ∫ 6(50 + 0 2 dt + C
- p o ⅛ P ( 5θ÷')≈ + q
= 2 (5 0 + '> + ( 5 δ ⅛
To evaluate C , we recall that there are 20 lb of salt initially in the solution.

This means 5(0) = 20, so that
^(°) = 2(S0 ÷ 0) + - ξ- ⅛
2 0= 100 +
—80(502) = C
Thus, the solution to the given differential equation, subject to the initial
condition 5(0) = 20, is
Specifically, at the end of 1 hour (60 min), the tank contains
5(60) = 2(50 + 60) - z580°^5θL ≈ 203.471 0744

The tank contains about 200 lb of salt. M M
A second application of first-order linear differential equations involves the

current in an R L electric circuit. An R L circuit is one with a constant resistance,
R , and a constant inductance, L . Figure 6.36 shows an electric circuit with an

electromotive force (EMF), a resistor, and an inductor connected in series. The
EM F Current
EM F source, which is usually a battery or generator, supplies voltage that causes a
£(r) /(0 current to flow in the circuit. According to Kirchhoffs second law, if the circuit is
closed at time t = 0, then the applied electromotive force is equal to the sum of the
voltage drops in the rest of the circuit. It can be shown that this implies that the
current* ∕(t) that flows in the circuit at time t must satisfy the first-order linear
Figure 6.36 An RL circuit dia differential equation
gram
L cM-t + R I = E
at
where E is the EM F and L (the inductance) and R (the resistance) are positive
constants. Write
77
to see that P(x) = (a constant) and Q(x) = £ (a constant). Then,
Z(t) = e ~∣
m t
= e -(R ∕L )t
= e ~(R∕L)t E .L
L R c
= ^ + Ce~^R 'Et
R
It is reasonable to assume that no current flows when t = 0. That is, I = 0
77
when t = 0, so we have 0 = ⅛ + C e 0 and C = -E ∕ R . The solution of the differen-
∕∖
tial equation is
_ E _ ( E ∖ - RUL
R ∖R R
Notice that because e ^r v l ^ → 0 as t → +∞, we have
E ∖ p - R t∕L - E
lim I(t) = 4 - lim
∕-→ +∞ JX I —> ÷ ∞ R) R
This means that, in the long run, the current / must approach E∕R. The solution of
the differential equation consists of two parts, which are given special names:
U,
is the steady-state current.
E e r i 'l is the transient current.
R
Figure 6.37 shows how the current ∕(Z) varies with time t.
We have seen four common growth and decay models, which are summarized
in Table 6.3. This presentation only touches on a topic that could fill an entire
Figure 6.37 The current in an RL course.
circuit with constant EMF
*It is common practice in physics and applied mathematics to use I as the symbol for current.
Naturally, this has nothing to do with the concept of integrating factor introduced earlier in this
section.
PROBLEM SET 6 .4
© Find the general solution o f each separable differential equa 27. In 1980 the gross domestic product (GDP) of the United
tion given in Problems 1-10. States in constant 1972 dollars was $1,481 billion. If the
growth rate from 1980 to 1984 was 2.5% per year, predict
1. ÷ = 3y
x 2. = y2 3. -τ- = ey the GDP in 1990. Consult an almanac to see if this
dx dx dx
dy
prediction is correct.
r+v dy y dy
4. dx e 5. =
6. Tx = x y 28. According to the Department of Health and Human
dx X
Services, the divorce rate in 1984 in the United States was
dy _ cosx dy dy ιn
7. 8.
= ex 9. -j— y = 10 4.9% and there were 1,155,000 divorces that year. How
dx cosy dx dx
many divorces will there be in 1996 if the divorce rate is
⅛ + y - SO constant?
10.
dx ■
29. According to the Department of Health and Human
11. Each function on the left satisfies exactly one differential Services, the marriage rate in 1984 in the United States
equation on the right. Match each function with the was 10.5% and there were 2,487,000 marriages that year.
proper equation. How many marriages will there be in 1996 if the marriage
a. y = e~2x i. y'" + 3y" - 4y' = rate is constant?
b. y = ex — <?- 4 ii. y " - 4y = 0 30. A tank contains 10 lb of salt dissolved in 30 gal of water.
c. y = e - v sinx Hi. y " + 2y' - 3y = 0 Suppose 2 gal of brine containing 1 lb of dissolved salt per
gallon runs into the tank every minute and that the
d. y = xe 3x iv. y " — 6y' + 9y = 0
mixture (kept uniform by stirring) runs out at the same
e. y = ex + e~3x v. y " + 2y' + 2y = 0 rate.
a. Find the amount of salt in the tank at time t.
Solve the differential equations in Problems 12-17. b. How long does it take (to the nearest second) for the
12 ⅛ + ⅛ = χ 13⅛ + ⅛ - √ J + l tank to contain 15 lb of salt?
ax x dx x 31. In Problem 30, suppose the tank has a capacity of 100 gal
14. x 4 ⅞- + 2x 3y = 5 15. x 2 ⅞- + xy = 2 and that the mixture flows out at the rate of 1 gal/min
dx dx ■ (instead of 2 gal/min).
16. x ⅛ → 2y = xe χ3 17. J + ~ 1 )y = e~2x a. How long will it take for the tank to fill?
dx - ax ∖ x F b. How much salt will be in the tank when it is full?
In Problems 18-19, solve the initial value problems, 32. The rate at which a drug is absorbed into the blood system
dv is given by
18. χ -j— 2y = 2x 3, x > 0 with v = 2 when x = 1
dx ■
19. y ' + = 0, x > 0 with y = 2 when x = 2
where M∕) is the concentration of the drug in the blood
stream at time t. What does b(f) approach in the long run
O In Problems 20-23, fin d the orthogonal trajectories o f the given (that is, as t → +∞)7 At what time is ⅛(f) equal to half this
fam ily o f curves. Recall from Section 4.4 that a curve is an limiting value? Assume that ⅛(0) = 0.
orthogonal trajectory o f a given fam ily i f it intersects each 33. An ΛL-circuit has a resistor of R ohms, inductance of L
member o f that family at right angles. henries, and emf of E volts where R, L, and E are
constant. Suppose no current flows in the circuit at time
20. the family of parabolas y 2 = 4/oc
t = 0. If L is doubled and E and R are held constant, what
21. the family of hyperbolas xy - c effect does this have on the “long-run” current in the
22. the family of circles x 2 + y 2 = r 2 circuit (that is, the current as t → +∞)7
23. the family of exponential curves y = Ce~x 34. If the 1990 census recorded 16,098,000 Hispanics and if
24. The Dead Sea Scrolls were written on parchment at about growth is proportional to the population, how many
100 B.C. What percentage of 14C originally contained in Hispanics will be recorded in 2000 if the 1984 population
the parchment remained when the scrolls were discovered was 15,575,000?
in 1947? 35. A population of animals on Catalina Island is limited by
25. Tests of an artifact discovered at the Debert site in Nova the amount of food available. If there were 1,800 animals
Scotia show that 28% of the original l4 C is still present. present in 1980 and 2,000 in 1986 and it is estimated that
What is the probable age of the artifact? the population cannot exceed 5,000, estimate the popula
26. In 1990 the gross domestic product (GDP) of the United tion in the year 2000.
States was $5,465 billion. If the growth rate from 1989 to 36. In 1986 the Chernobyl nuclear disaster in the Soviet
1990 was 5.08%, predict the GDP in 2000. Union contaminated the atmosphere. The buildup of
radioactive material in the atmosphere satisfies the differ

ential equation a. Replace the derivative 0'(7) in the differential
equation in Problem 39 by the difference quotient
dM= r (k_ M \ M = Q when r = 0 0(7 + ∆7) - 0(7)
α7 ∖r / ∆7
where M = mass o f radioactive material in the atmos Now, for simplicity, take ∆7 = 1, thus obtaining
phere after time t (in years); k is the rate at which the
radioactive material is introduced into the atmosphere; r 0 (7 + l) = 0 ( 7 ) [ l + k ( 5 - 0 ( ∕) ) ]
is the annual decay rate of the radioactive material. Find b. Starting with 0(0), we can compute 0(1), then
the solution, Λ∕(Z), o f this differential equation in terms of 0(2), and so on. Write a simple program to
k and r. generate a sequence of 0(n). For example, the
37. S P Y P RO BLEM Sweet revenge! The spy has finally following Mathematica code sets the parameters
tracked down Boldfinger and is ready to pay him back for and 0(0), and generates 0(1), 0(2), . . . ,0(16).
killing his friend, Siggy Leiter (See Problem 81 in the Your own code should be similar.
Supplementary problems from Chapter 5). He is in an
airplane flying at 10,000 ft, when he sees the chateau of k = 0.05; B = 6.0; 0[O] = 5.0; Table [Q[n + 1] =
his nemesis. There is no place to land, so he decides to 0[∏](1.O + k(B - 0[n])), {n, 0, 15}
parachute into Boldfinger’s lair. In order to have some (Your sequence o f 0(w) should move toward the
control over the spot where he lands, he plans to free-fall solution 0 = £.)
for 40 seconds before opening his parachute, but for c. In the above difference equation you may encoun
safety’s sake, he doesn’t want his speed, v, to be greater ter a difficulty not seen in the corresponding
than 20 ft/s when he lands. If the spy weighs 192 lb and differential equation; it can become unstable. In
the air resistance is 0.75v(Z) until the parachute opens and fact, if the product kB exceeds 2, the problem
10v(7) afterward, does he land safely? Hint. See Problem becomes unstable. This can have serious implica
43. tions for such populations. To see this, pick k and
38. An RL-circuit has a resistance of R = 10 ohms and an B so that kB > 2 (but not too big!) and run the
inductance o f L = 5 henries. Find the current /(7) in the program. Comment on your results; including
circuit at time t if 7(0) = 0 and the electromotive force your interpretation of the parameters k and B,
(emf) is: which will lead to this instability.
a. E = 15 volts b. £ = 5e- 2 ' sin 7 d. A challenge problem. To see why kB > 2 leads to
instability, you must do the following. We want to
Computational Window study what happens as solution 0(w) get close to
the key value B. So let y(n) = 0(n) - B, and look
at small y(n). Replace 0(n) and y(n) + B in the
In the following two problems we pursue the logistic
difference equation and discard the y(n)2 term
growth equation from a couple o f new perspectives. First,
(since it is assumed to be very small). Thus, you
by simply studying the differential equation Q'(t) =
obtain the new simple equation:
kQ (B - Q), we can make some important conclusions
about the solutions, without actually finding the solutions. y(n + 1) = y(π)(I - kB)
Second, many populations do not reproduce throughout
the year. In fact, most o f animal, insect, and planet life Comment on what will happen to the sequence
breed and reproduce periodically (for example, once a y(ri) in the unfortunate event that kB > 2. In
year). In this case the correct model is no longer a particular, see if you can explain why we say that
differntial equation, but rather a “difference equation,” the solution y(n) = 0 is unstable (thus, the corre
which you will see in Problem 40. This latter model has a sponding solution Q(ri) = B is unstable).
rather surprising complication.
39. By simply examining the differential equation itself,
make these arguments: θ 41. A Bernoulli equation is a differential equation of the form
a. Argue that Q(t) = B is a solution to Q'(f) = y ' + P(x)y = O(x)y"
k β (β - Q).
b. If 0 < 0(0) < B, then 0(z) will increase for all 7; where n is a real number, n ≠ 0, n ≠ 1.
but, it can never exceed B. a. Show that the change of variable u = y 1 - , ' transforms
c. Similarly, what happens if 0(0) > B I Note'. Since such an equation into one of the form
all solutions are funneled toward the solution M '(Λ ) + (1 - M) P(X )U(X ) = (1 - ri) 0(x)
Q = B, we say that this solution is highly stable.
40. Now, suppose the population 0(7) only produces This transformed equation is a first-order linear differ
offspring once each ∆7 period o f time. This could be ential equation in u and can be solved by the methods
1 year or 3 weeks, for instance. of this section, yielding a solution to the given Ber
noulli equation.
b. Use the change of variable suggested to solve the Solve the equation and characterize the family of all
Bernoulli equation curves that satisfy the given geometric condition.
43. The Motion of a Body Falling Through a Resisting Medi
y ' + Λ, = i y 2 um To discuss the motion of a body falling through the
42. Consider a curve with the earth’s atmosphere, we assume that the only forces acting
property that when hori on the body are gravitational attraction and air resistance
zontal and vertical lines that is proportional to the velocity of the body.
are drawn from each a. Suppose the body has weight W so that
point P(x, y) on the curve Weight o f body
to the coordinate axes, , '∣'
the area A l under the m ¾ = W — kv
curve is twice the area /1, at <v--
'—∙—' Air resistance
above the curve, as Force on body
shown in Figure 6.38. Remember W = mg where m is the object’s mass and
g = 32 ft∕s 2 is the acceleration due to gravity. Also,
v ~ v (θ)∙ Solve this differential equation to express
Show that x and y satisfy the differential equation 0
velocity v in terms of time t.
dy _ ∣- b. Use integration to find the displacement s of the body
dx l x at time t. (Assume that s = .s,0 when t = 0.)
CHAPTER 6 REVIEW

In Chapter 1 we introduced the idea of mathematical modeling, and in Chapter
4 we introduced one of the most important and revolutionary ideas in the
history of mathematics, namely, that of a Riemann sum. In this chapter we put
these two ideas together to model a great many real-life situations. Table 6.4
summarizes this relationship between modeling and the Riemann sum.
TABLE 6.4 Modeling Applications Involving Integration

Section Situation Riemann Sum Model Outcome
4.2 Introduction Model something in terms 1. Partition interval. The approximations improve as the
o f one or more continuous functions 2. Choose a number x'k arbitrarily norm o f the partitions goes to 0. The
on a closed interval [a, ⅛]. from each subinterval of width ∆x z,. Riemann sums approach a limiting
3. Form a sum of the type integral. Use this integral to define
n and calculate the quantity we origi
R „ = ∑ ι ∕( ¾ ) ∆ x t nally wanted to measure.
4.2 Area
Find the area under a curve defined by Find the sum o f n rectangles.
y = f(x) where f( x ) ≥ 0, above the x- rb
axis and bounded by the lines x = a Σ ∕( ⅞ ) ∆Λ-λ Area = f(x ) dx

∕<= 1 v y Ja
and x = b.
Area of ∕√th rectangle
4.2 Displacement
An object is moving along a line having Find the sum of n distances.
continuous velocity v(Z) for each time t rb
between t = a and t = b. Σ v(∕i*)∆ ∕
fc= 1 , j
Displacement = ∣ v(Z) dt
Ja
Distance traveled over
∕<th time interval
TABLE 6.4 (continued) Modeling Applications Involving Integration

Section Situation Riemann Sum Model Outcome
4.6 Average value

1 f' ,
Find the average of a continuous function on an Find the average of a Average = ∕W dx
interval. weighted sum o f n values. Ja
y ,Λ ⅞ >
4.7 Area between curves

Σ [∕⅛ *) - s⅛*)] ¾ rb
Find the area between the curves y = ∕( x ) and Area = [∕(x ) - g(x)] dx
K.—1 ∖________ j
y = £(*) Ja
Top curve minus
bottom curve
6.1 Volume by cross section rb

Find the volume of a solid with a cross section Σ Λ ⅞ ) ∆x λ. Volume = .∙4(x) dx
k= 1 Ja
perpendicular to the x-axis with face area .∙l(x).
6.1 Volumes of revolution

Find the volume of a region revolved about the
specified axis. rb
Disks: x-axis ∑ π [ ∕( x t*)]2 ∆ x , Volume = τr [∕(x )] 2 dx
k=↑ Ja
∑ √ [ ∕( ⅞ ) ] 2 -⅛ ⅛ * j] 2) ∆ χ t V = r r f ([∕(x )] 2 -⅛ ( x ) ] 2)rfx
Washers: x-axis
⅛=1
2
cb
Shells: v-axis Σ ¾ ∕( x * ) ∆ x i Volume = 2τr 1 x∕(x) dx
fc=l Ja
6.2 Arc length n rb

Find the length o f a continuously differentiable Σ V ι + [ ∕'⅛ * ) l 2 ∆x i Length = V l + [ ∕'( x ) ] 2 dx
k= 1 Ja
curve y = ∕( x ) on [a, b].
6.2 Area of a surface of revolution n rb

Find the area o f a surface o f revolution (about the ∑ 2√⅛*)Δ⅞ Area = 2τr 1 ∕( x ) V l + [ f l (x)]2 dx
k= 1 Ja
x-axis).
6.3 M ass
cb
Find the mass o f a thin plate with constant ∑ p [f(x k ) ^ #(4)] ¼ ∙ m = p [∕(x ) - ⅛∙(x)] dx
density p. k=l Ja
6.3 Moment
M y = p f x [∕( x ) - g(x)]dx
Find the moment of a thin plate with constant Moment about the v-axis
density p (about the v-axis). Ja
Σ pχ*k [Λχ*k ) - ^⅛*)]∆x jt
Centroid k= 1 p Cb
Find the centroid of a homogeneous lamina. M
X = 2 {[/(*)] 2 - [g(x)]-}dx
Moment about the x-axis Ja
∕M M ∖
,∑ 2<[∕(⅞ )] 2 “ ⅛(⅞)]2}Δ⅞ (x, v) = —
∖m m )
6.3 Work rb
Find the work done by a variable force F(x) fc∑ F(x t*)∆ x t Work = 1 F(x) dx
directed along the x-axis from a to b. Ja
6.3 Hydrostatic force rb

Find the total force of a fluid against one side of ∑ p(^*)L⅛*)Δ⅛ F = 1 phL(h) dh
a vertical plate. k= 1 Ju
C o n c e p t P r o b le m s 16. Which expressions represent volumes of a solid with

1. Describe the process of finding volumes of solids with a parallel cross sections that are perpendicular to an axis?
known cross-sectional area. 17. Which expressions represent volumes of solids of revo
2. Compare and contrast the methods of finding volumes lution revolved about the x-axis?
by disks, washers, and shells. 18. Which expressions represent volumes of solids of revo
3. What is the formula for arc length of a graph? lution revolved about the y-axis?
4. What is the formula for area of a surface of revolution? 19. Let R be a region bounded by the graphs of the equa
5. How do you find the work done by a variable force? tions y = (x - 2)2 and y = 4.
6. How do you find the force exerted by a liquid on an a. Sketch the graph of R and find its area.
object submerged vertically? b. Find the volume of a solid of revolution of R about
7. What are the formulas for the mass and centroid of a the x-axis.
homogeneous lamina? c. Find the volume of a solid of revolution of R about
8. State the volume theorem of Pappus. the y-axis.
9. State Hooke’s law. 20. The base of a solid in the xy-plane is the circular region
10. What is Pascal’s principle? given by x 2 + j 2 = 4. Every cross section perpendicular
11. What is the formula for solving the first-order linear to the x-axis is a rectangle whose height is twice the
differential equation y ' + P(x)y = Q(x)l length of the side that lies in the xy-plane. Find the
12. What is an RL circuit? volume of this solid.
21. Find the arc length of the curve y = 2 —x 3/2 from (0, 2)
P r a c tic e P r o b le m s to (1, 1).
JOURNAL PROBLEMS [Mathematics Teacher, December 22. Find the surface area obtained by revolving the para
1990, ρ. 695.) Use the integrals A-G to answer the questions bolic arc y 2 = x from (0, 0) to (1, 1) about the x-axis.
in Problems 13-18. Give answer to the nearest hundredth unit.
<∙b rd rb
23. Solve the first-order linear differential equation
J
D. f A(y) dy
a
[∕(x)] 2 r/x B. 7τ [f[y)∖ 1 dy
Jc
C.
Ja
E. τr f [f(x) - g(x)] dx
A(x) dx
dx x + 1
Jc Ja subject to the condition y = 1 when x = 0.
rb rd 24. A tank contains 200 gal of saturated brine with 2 lb of
F. 77 {[∕(x)] 2 - [g(x)]2}dx G. 77 {[∕∙(X)]2 - te(y)]2} dy salt per gallon. A salt solution containing 1.3 lb of salt
Ja Jc per gallon flows in at the rate of 5 gal/min, and the
13. Which formula does not seem to belong to this list? uniform mixture flows out at 3 gal/min. Find the
14. Which expressions represent volumes of solids of revo maximum concentration of salt in the tank and the time
lution? required to reach 2/3 the maximum concentration.
15. Which expressions represent volumes of solids contain 25. Find the centroid of the region between y = x 2 - x and
ing holes? y = x - x 3 on [0, 1].
Find the volume of a solid of revolution obtained by revolving 6. The base of a solid is the region R in the xy plane bounded
the region R about the indicated axis in Problems 1-4. by the curve y 2 = x and the line y = x. Find the volume of
the solid if each cross section perpendicular to the x-axis
1. R is bounded by y = Vx, x = 9, y = 0; about the x-axis.
is a semicircle with a diameter in the xy-plane.
2. R is bounded by y = x 2, y = 9 —x 2; x = 0 about the 7. Find the centroid of a homogeneous lamina covering the
y-axis. region R that is bounded by the parabola y = 2 - x 2 and
3. R is bounded by y = x 2 and y = x 4; about the x-axis. the line y = x.
8. A force of 100 lb is required to compress a spring from its
4. R is bounded by y = Vcosx, x = f , x = f, and y = 0; natural length of 10 in. to a length of 8 in. Find the work
about the x-axis. required to compress it to a length of 7 in.
5. The base of a solid is the region in the xy plane that is 9. A force of 1000 lb is required to compress a spring from
bounded by the curve y = 4 - x 2 and the x-axis. Every its natural length of 10 in. to a length of 8 in. Find the
cross section of the solid perpendicular to the v-axis is a work required to stretch it to a length of 11 in.
rectangle whose height is twice the length of the side that 10. A tank has the shape of the solid obtained by revolving
lies in the xy plane. Find the volume of the solid. the parabolic arc y = 2x2 from x = 0 to x = 2 about the
y-axis (units are in feet). The tank is filled with water. work required to pump all the fluid to a height of 2 ft
Find the work done in pumping the water to the top of the above the top of the tank.
tank. 23. Each end of a container filled with water has the shape of
11. A tank has the shape of the solid obtained by revolving an isosceles triangle, as shown in Figure 6.40. Find the
the parabolic arc y = 2x 2 from x = 0 to x = 2 about the y- force exerted by the water on an end of the container.
axis (units are in feet). The tank is filled with water. Find
the work done in pumping the water to a point 5 ft above
the top of the tank.
12. Find the volume of a sphere of radius a by revolving the
region bounded by the semicircle y = Vα 2 —x 2 about the
x-axis.
13. Find the volume of a sphere of radius a by revolving the
Figure 6.40 Water pressure
region bounded by the semicircle x = Vα 2 —y 2 about the
y-axis.
14. Two solids 5j and S 2 have as their base the region in the 24. A container has the shape of the solid formed by rotating
the region bounded by the curve y = x 3 ft and the lines
xy-plane bounded by the parabola y = x 2 and the line
y = 1. For solid 5 1, each cross section perpendicular to x = 0 ft and y = 1 ft about the y-axis. If the container is
the x-axis is a square, and for S 2 , each cross section per filled with water, how much work is required to pump all
pendicular to the y-axis is a square. Find the volumes of the water out of the container?
*Sj and S 2 . 25. A gate of a water main has the shape of a circle with
15. Use calculus to find the lateral surface area of a right diameter 10 ft. If water stands 2 ft high in the main, what
circular cylinder with radius r and height h. is the force exerted by the water on the gate?
16. Use calculus to find the surface area of a right circular Suppose a particle o f mass m moves along a number line under
cone with (top) radius r and height h. the influence o f a variable force F, which always acts in a
17. A 20-lb bucket in a well is attached to the end of a 60-ft direction parallel to the line o f motion. Then, at a given time
chain that weighs 16 lb. If the bucket is filled with 50 lb of the particle is said to have kinetic energy
dirt, how much work is done in lifting the bucket and K = ∖ mv2
chain to the top of the well?
where v is the particle’s velocity, and it has potential energy
18. A conical tank whose vertex points down is 10 ft high with
(top) diameter 4 ft. The tank is filled with a fluid of P = - Γ Fdx
density p = 22 lb∕ft 3. Find the work required to pump all Js0
the fluid to the top of the tank. relative to the point s on the line o f motion. Use these
19. A swimming pool is 3 ft deep at one end and 8 ft deep at definitions in Problems 26-27.
the other, as shown in Figure 6.39. The pool is 30 ft long 26. A spring at rest has potential energy P = 0. If the spring
and 25 ft wide with vertical sides. Find the fluid force constant is 30 lb∕ft, how far must the spring be stretched
against one of the 30-ft sides. so that its potential energy is 20 ft-lb?
27. A particle with mass m falls freely near the earth’s
surface. At time t = 0, the particle is .s0 units above the
ground and is falling with velocity vθ.
a. Find the potential energy P of the particle in terms of s
if the potential energy is 0 when s = sθ.
b. Find the potential energy P and the kinetic energy K of
Figure 6.39 Diagram of a swimming pool the particle in terms of time t.
28. An object weighs w = l,000r^ 2 grams, where r is the
20. The ends of a container filled with water have the shape of object’s distance from the center of the earth. What work
the region R bounded by the curve y = x 4 and the line is required to lift the object from the surface of the earth
y = 16. Find the fluid force exerted by the water on one (r = 6,400 km) to a point 3,600 km above the surface of
end of the container. the earth?
21. An observation porthole on a ship is circular with diame 29. A certain species has population P(t) at time t. The growth
ter 2 ft and is located so that its center is 3 ft below the rate of the population is affected by seasonal variations in
water line. Find the fluid force on the porthole, assuming the food supply and may be modeled by the differential
that the boat is in seawater (density 64 lb∕ft 3). equation
22. A conical tank whose vertex points down is 12 ft high with ^ = k ( P - A ' ) cost
(top) diameter 6 ft. The tank is filled with a fluid of
density p = 22 lb∕ft 3 . If the tank is only half full, find the
where k and A are physical constants. Solve the differen 35. A container of a certain fluid is raised from the ground on
tial equation and express P(t) in terms of k, A , and P o , the a pulley. While resting on the ground, the container and
initial population. fluid together weigh 200 lb. However, as the container is
30. A company plans to hire additional advertising personnel. raised, fluid leaks out in such a way that at height x feet
Suppose it is estimated that if x new people are hired, they above the ground, the total weight of the container and
will bring in addition revenue o f ∕∕( x ) = V∑x thousand fluid is 200 —0.5x lb. Find the work done in raising the
dollars and that the cost o f adding these x people will be container 100 ft.
C(x) = ∣ x thousand dollars. How many new people 36. A bag of sand is being lifted vertically upward at the rate
should be hired? How much total net revenue (that is, of 31 ft/min. If the bag originally weighed 40 lb and leaks
revenue minus cost) is gained by hiring these people? at the rate of 0.2 lb∕s, how much work is done in raising it
31. Find the center of mass o f a thin plate o f constant density until it is empty?
if the plate occupies the region bounded by the lines 37. The vertical end o f a trough is an equilateral triangle, 3 ft
2j<, = x, x + 3y = 5. and the x-axis. on a side. Assuming the trough is filled with a fluid of
32. The portion o f the ellipse density p = 40 lb ∕ft 3, find the force on the end o f the
trough.
38. The radius of a circular water main is 2 ft. Assuming the
main is half full of water, find the total force exerted by
the water on a gate that crosses the main at one end.
that lies in the first quad
rant between x = r and
x = a, for 0 < r < a, is
revolved about the x-axis
to form a solid 5, ,. Find
the volume o f 5 e.
39. The force acting on an object moving along a straight line

is known to be a linear function of its position. Find the
force function F if it is known that 13 ft-lb of work is
33. The portion of the hyper required to move the object over 6 ft and 44 ft-lb o f work
bola is required to move it 12 ft.
40. A tank has the shape of a rectangular parallelepiped with
length 10 ft, width 6 ft, and height 8 ft. The tank contains
a liquid o f density p = 20 lb ∕ft 3. Find the level of the
liquid in the tank if it will require 2,400 ft-lb of work to
move all the liquid in the tank to the top.
41. A tank is constructed by placing a right
circular cylinder of height 10 ft and
radius 2 ft on top of a hemisphere with
the same radius. If the tank is filled with
a fluid of density p = 24 lb ∕ft 3, how
much work is done in bringing all the
fluid to the top o f the tank?
34. Let R be the part o f the ellipse
42. A piston in a cylinder causes a gas either to expand or to

contract. Assume the pressure P and the volume of gas V
that lies in the first quadrant. Use the theorem of Pappus in the cylinder satisfy the adiabatic law (no exchange of
to find the coordinates of the centroid of R . You may use heat)
the facts that R has area /1 = ∣π ab, and the semiellipsoid
P U 1∙4 = C
it generates when it is revolved about the x-axis has
volume V = ⅜πab2 . where C is a constant.
a. If the gas goes from volume V∖ to volume V2, show that 47. A child is building a sand structure at the beach. Find the
the work done is given by work done if the density o f the sand is p = 140 lb ∕ft 3 and
the structure has the shape o f a cone o f height 1 ft and
W = Γ 2 CV~'A d V diameter 2 ft.
J ', 48. A child is building a sand structure at the beach. Find the
b. How much work is done if 0.6 m 3 of steam at a work done if the density o f the sand is p = 140 lb ∕ft 3 and
pressure o f 2,500 newtons∕m 2 expands 50%? the structure is a tower (cylinder) of height 1 ft and radius
43. A thin sheet of tin is in the shape o f a square, 16 cm on a 4 in.
side. A rectangular corner o f area 156 cm 2 is cut from the 49. Let 5 be the solid generated by revolving about the x-axis
square, and the centroid of the portion that remains is the region bounded by the parabola y = x 2 and the line
found to be 4.88 cm from one side o f the original square, y = x. Find the number A so that the plane perpendicular
as shown in Figure 6.41. How far is it from the other three to the x-axis at x = A divides S in half. That is, the
sides? volume on one side o f A equals the volume on the other
side.
50. A plate in the shape o f an isosceles triangle is submerged
vertically in water, as shown in Figure 6.43. Find the force
156
cm^
on one side of the plate if the two equal sides have length
3, the third side has length 5, and the top vertex of the
16 cm
triangle is 4 units below the surface.
16 cm
Figure 6.41 Centroid of a square
44. The front and back faces o f a container are parabolic
surfaces with vertical axes and the vertex of the parabola
at the bottom (assume the coordinate axes are located so
the parabola has the form y = k x 2). Each parabolic face is
10 ft across the top and 15 ft deep. If the container is filled Figure 6.43 Hydrostatic pressure
with a liquid of density p = 40 lb ∕ft 3 , find the force on
either parabolic face.
51. Repeat Problem 50 for equal sides o f length A , the third
45. Solve the differential equation side with length B, and the top vertex D units below the
dy _ y surface, as shown in Figure 6.43.
dx χ + cyy 2 52. Find the arc length o f the curve
by reversing the roles o f x and y (in other words, treat y as
the independent variable). yz = A x 3 + -
x
46. Let R be the region bounded by the semicircle
y = ∖z 4 - x 2 and the x-axis, and let .S' be the solid with on the interval [a, b], where A and B are positive constants
base R and trapezoidal cross sections perpendicular to the with A B = y 2 .
x-axis. Assume that the two slant sides and the shorter 53. T H IN K T A N K P ROBLEM When the line segment
parallel sides are all half as long as the side that lies in the y = x —2 between x = 1 and x = 3 is revolved about the
base of each cross-sectional trapezoid are congruent as x-axis, it generates part o f a cone. Explain why the surface
shown in Figure 6.42. Set up but do not evaluate an area of this cone is not given by the integral
integral for the volume o f ,S'.
f3
2π(x - 2)V2 dx
Ji
What is the actual surface area?
54. An object below the surface of the earth is attracted by a

gravitational force that is directly proportional to the
distance from the object to the center of the earth. Find
the work done in lifting an object weighing W lb from a
depth of .v feet (below the surface) to the surface. Assume
Figure 6.42 Volume of a region 5, the earth’s radius is 4,000 miles.
In Problems 55-56, consider a continuous curve y = f(x) that 67. Find the volume of the solid whose base is the region R
passes through ( - 1, 4), (0, 3), (1, 4), (2, 6), (3, 5),'(4, 6), (5, 7), bounded by the curve y = ex and the lines y = 0, x = 0,
(6, 6), (7, 4), αw√(8, 3). x = 1, if cross sections perpendicular to the x-axis are
equilateral triangles.
55. Use Simpson’s rule to estimate the volume o f the solid
formed by revolving this curve about the x-axis for 68. Find the length of the curve y = In secx on the interval
[ - l ,9 ] . [o, ⅛
56. Use the trapezoidal rule to estimate the volume o f the 69. Find the length of the curve
solid formed by revolving this curve about the line
y = -1 on [ - 1, 9].
b2 8α l n ⅛
57. Find the centroid of a right triangle whose legs have
lengths a and b. Hint: Use the theorem of Pappus. between y = b and y = 1b.
58. The largest of the pyramids at Giza in Egypt is roughly
70. Snowplow Problem of R. P. Agnew (This problem was
480 ft high and has a square base, 750 ft on a side. If the
mentioned in the preview for Chapter 4.) One day it starts
rock in the pyramid averages 160 lb ∕ft 3 , how much work
snowing at a steady rate sometime before noon. At noon,
was required to lift the stones into place to construct the
a snowplow starts to clear a straight, level section of road.
pyramid?
If the plow clears 1 mi of road during the first hour but
59. The half-life of the radioactive isotope cobalt-60 is 5.25 requires 2 hr to clear the second mile, at what time did it
yr∙ start snowing? Answer this question by completing the
a. What percentage o f a given sample of cobalt 60 re following steps:
mains after 5 yr? a. Let t be the time (in hours) from noon. Let h be the
b. How long will it take for 90% of a given sample to depth of the snow at time t, and let 5 be the distance
disintegrate? moved by the plow. If the plow has width w and clears
60. The rate at which salt dissolves in water is directly snow at the constant rate p, explain why
proportional to the amount that remains undissolved. If 8
lb of salt is placed in a container o f water and 2 lb wΛ⅛ = p
dissolves in 30 min, how long will it take for 1 lb to
dissolve?
b. Suppose it started snowing tβ hours before noon. Let r
61. Scientists are observing a species o f insect in a certain denote the (constant) rate of snowfall. Explain why
swamp region. The insect population is estimated to be 10
million and is expected to grow at the rate o f 2% per year. Λ(∕) = r(t + ∕0)
Assuming that the growth is exponential and stays that
way for a period of years, what will the insect population By combining this equation with the differential equa
be in 10 yr? How long will it take to double? tion in part a, note that
62. Find the length of the curve y = ex on the interval [0, 1]. ∕7 c
63. Find the volume of the solid generated when the region wr(ι + t0) f t = P
under the curve
Solve this differential equation (with appropriate con
= 1 ditions) to obtain t0 and answer the question posed in
y
the problem.
on the interval [0, 2] is revolved about the y-axis. 71. P U T N A M E X A M IN A T IO N P ROBLEM
Find the length of the
curve y 2 = x 3 from the origin to the point where the
64. Find the volume o f the solid formed by revolving about
tangent makes an angle of 45° with the x-axis.
the x-axis the region bounded by the curve
A solid is bounded by
. " ⅛ two bases in the horizontal planes z = h/2 and z = - A ∕2 ,
and by a surface with the property that every horizontal
cross section has area given by an expression of the form
the x-axis, and the lines x = 1 and x = 2.
65. Find the area between the curves y = tan 2 x and A = a0z 3 + a i z 2 + a2z + α3
y = sec2 x on the interval [θ, j ] .
a. Show that the solid has volume V = ∣Λ (5 1 + 5 , + 4Λ f)
66. Find the surface area of the solid generated by revolving where 5 1 and 5 2 are the area of the bases and M is the
the region bounded by the x-axis and the curve area o f the middle horizontal cross section.
y = ex + ⅛e~x b. Show that the formulas for the volume of a cone and a
sphere can be obtained from the formula in part a for
on [0, 1] about the x-axis. certain special choices o f α0, α 1, α,, and a γ
"Houdini’s Escape”
This project is to be done in groups of three or four students. Each group will submit
Harry Houdini was a famous escape artist. In this project we relive a trick of his
that challenged his mathematical prowess, as well as his skill and bravery. It may
challenge these qualities in you as well.
Houdini had his feet shackled to the top of a concrete block which was placed on
the bottom of a giant laboratory flask. The cross sectional radius of the flask,
measured in feet, was given as a function of height z from the ground by the
formula r(z) = 10z~ιz2, with the bottom of the flask at z = 1 ft. The flask was then
filled with water at a steady rate of 22ττ ft3∕min. Houdini’s job was to escape the
shackles before he was drowned by the rising water in the flask.
H ARRY H OUDINI (born Ehrich
Weiss) (1874-1926) Now Houdini knew it would take him exactly 10 minutes to escape the shackles.
For dramatic impact, he wanted to time his escape so it was completed precisely at
the moment the water level reached the top of his head. Houdini was exactly 6 ft
Mathematics is the gate and key o f tall. In the design of the apparatus, he was allowed to specify only one thing: the
the sciences. . . . Neglect o f mathe height of the concrete block he stood on.
matics works injury to all knowl Your paper is not limited to the following questions but should include these
edge, since he who is ignorant o f it concerns: How high should the block be? (You can neglect Houdini’s volume and
cannot know the other sciences or
the volume of the block.) How fast is the water level changing when the flask first
the things o f this world. And what
is worse, those who are thus igno starts to fill? How fast is the water level changing at the instant when the water
rant are unable to perceive their reaches the top of his head? You might also help Houdini with any size flask by
own ignorance and so do not seek generalizing the derivation: Consider a flask with cross-sectional radius r(z) an
a remedy. arbitrary function of z with a constant inflow rate of dV/dt = A. Can you find
Roger Bacon dh!dt as a function of h(f)c!
Opus Majus, Part 4, Distinctia Pri-
ma, cap. 1
It seems to be expected o f every

pilgrim up the slopes o f the mathe
matical Parnassus, that he will at
some point or other o f his journey
sit down and invent a definite inte
gral or two towards the increase o f
the common stock.
J. J. Sylvester
Notes to the Meditation on
Poncelet’s Theorem; Mathematical Pa *MAA Notes 17 (1991), “Priming the Calculus Pump: Innovations and Resources,” by Marcus S.
pers, Vol. 2, p. 214 Cohen, Edward D. Gaughan, R. Arthur Knoebel, Douglas S. Kurtz, and David J. Pengelley.
CUMULATIVE REVIEW FOR CHAPTERS 1 - 6
1. ■ What Does This Say? Define limit. Explain what 33. Find the area bounded by the curve y = x V x 2 + 8 and
this definition is saying using your own words. the x-axis on [—2, 1].
2. ■ What Does This Say? Define derivative. Explain 34. Find the area between the curves y = sinx and
what this definition is saying using your own words. y = cosx and the lines x = 0 and x = 1.
3. ■ What Does This Say? Define a definite integral.
35. Find the slope of the curve y = at the point x = J .
Explain what this definition is saying using your own
words.
4. ■ What Does This Say? Define a differential equa Solve the differential equations in Problems 36-37.
tion, and in your own words describe the procedure for
36. = x 2 y2V 4 - x 3 37. y dx - x dy = x In x dx
solving a first-order linear differential equation.
Evaluate the limits in Problems 5-13.

Find the particular solution o f the differential equations in
3 2 5χ
5. lim * 2 ~ π - 2
2
6. lim 3x 2 + 7x + 2 Problems 38-39.
Λ→2 3X - 7x + 2 x → + ≈ 5x —3x + 3
38. = 2(5 - y); y = 3 when x = 0
7. lim ( V x 2 + x - x) 8. lim ≤≡⅜'
Λ'→ +∞ A'→ 77/2 C OSX dy
39. = ey sin x with y = 5 when x = 0
9. lim "∙ ∙v s π 10. li m s ½
.χ→o x + sιn*x Λ∙→O x 40. Find the volume of the solid formed by revolving about
,∙ ⅛(x2 + 50) the x-axis the region bounded by the curve
11. lim (1 + X ) 27A' 12. lim π
x→o 2x _ 2
•V V 3 x - 2
13. lim+ x sιnjc
x→0
Find the derivatives in Problems 14-22.
41. Find the arc length of the curve
14. y = 6x 3 - 4x + 2 15. y = (x 2 + l) 3(3x - 4)2
_ x2 - 4 y - V e 2x - 1 — sec^ , ex
16. y 3x+ 1 17. = — -—
’
y X + c osx on the interval [0, 1].
18. x 2 + 3>xy + y 2 = 0 19. y = csc2 3x 42. A mason lifts a 25-lb bucket of mortar from ground level
20. y = e 5x ~4 21. y = ln(5x 2 + 3x - 2) to the top of a 50-ft-high building using a rope that
22. y = cos 1 (x 2 - 3) weighs 0.25 lb∕ft. Find the work done in lifting the
bucket.
Find the integrals in Problems 23-28. 43. A rocket is launched vertically from a point on the
f9 f1 ground 5,000 ft from a T V camera. If the rocket is rising
23. I dθ 24. 1 50(2x - 5)3 dx vertically at the rate o f 850 ft/s at the instant the rocket
4 -1
is 4,000 ft above the ground, how fast must the T V
Γ, x dx camera change the angle of elevation in order to keep
25. 1 26. 1
| esc 30 cot 30 dθ the rocket in the picture?
∣0 √9 + x 2
44. A lighthouse is 4,000 ft from a straight shore. Measur
Γ ex dx
27. ∣
' ex + 2
3 f
28. I x + 2x - 5 d χ ing the beam on the shore from the point P on the shore
X that is closest to the lighthouse, it is observed that the
f4 dx rotating light is moving at the rate o f 5 ft/s when it is
29. Evaluate using Simpson’s rule with n = 6. 1000 ft from P. How fast is the light revolving at this
J o V tl + x 3
instant?
30. Sketch the graph o f y = x 3 - 5x 2 + 2x + 8. 45. A tank contains 50 gal o f pure water. Brine containing 3
4— lb o f dissolved salt per gallon runs into the tank at the
31. Sketch the gra1 ph o f y = -i- , — j.
° ∙ 4 + χ- rate of 2 gal/min, and the well-stirred mixture runs out
32. Find the maximum value of at the rate of 1 lb per gallon. How much salt does the
∕( x ) = ∣ x 3 - 2x 2 + 2>x — 10 on [0, 6]. tank contain after 15 min?
Methods of Integration
■ CONTENTS
7.1 Review of Substitution and
Integration by Table
Review of substitution; Using
tables of integrals;
Trigonometric substitutions
7.2 Integration by Parts PREVIEW
Integration by parts formula;
Repeated use of integration by
parts; Definite integration by In this chapter we increase the number of techniques and procedures for
parts; Solving differential equa integrating a function. One of the most important techniques, substitu
tions by parts tion, is reviewed in the first section and is then expanded in several
7.3 The Method of Partial different contexts. Other important integration procedures include
Fractions
Decomposition; Integrating ra using tables, integration by parts, and partial fractions. In addition,
tional functions by partial improper integrals, hyperbolic functions, and inverse hyperbolic func
fraction decomposition;
Substitution with partial tions are discussed.
fractions; Rational trigonomet
ric integrals
7.4 Summary of Integration
Techniques PERSPECTIVE
7.5 Improper Integrals It is possible to differentiate most functions that arise in practice by applying a
Improper integrals with infinite
limits of integration; Improper fairly short list of rules and formulas, but integration is a more complicated
integrals with unbounded inte process. The purpose of this chapter is to increase your ability to integrate a
grands variety of different functions. Learning to integrate is like learning to play a
7.6 The Hyperbolic and Inverse musical instrument: At first, it may seem impossibly complicated, but if you
Hyperbolic Functions persevere, after a while music starts to happen. It should be noted that as
Hyperbolic functions;
Derivatives and integrals involv technology becomes more available, techniques become less important and
ing hyperbolic functions; ideas become more important. If you have such technology available, you might
Inverse hyperbolic functions wish to consult the Technology Manual.
Chapter 7 Review
436
Sec. 7.1 Review of Substitution and Integration by Table 437
7 .1 REVIEW OF SUBSTITUTION AND INTEGRATION BY TABLE
IN THIS SECTION Review of substitution, integral tables, trigonometric

substitutions
In Chapter 4 we derived a number of integration formulas and examined
various algebraic procedures for reducing a given integral to a form that can be
handled by these formulas. In this section we review those formulas as well as
the important method of integration by substitution. Today’s technology has
greatly enhanced the ability of professional mathematicians and scientists in
techniques of integration, thereby minimizing the need for spending hours
learning rarely used integration techniques. Along with this new technology, the
use of integral tables has increased in importance. Most mathematical hand
books contain extensive integral tables, and we have included a complete
integration table in the accompanying book, Student Mathematics Handbook.
For your convenience, we have also included a short table of integrals in
Appendix D .
For convenient reference, the most important integration rules and formulas are
listed inside the back cover. Notice that the first four integration formulas are the
procedural rules, which allow us to break up integrals into simpler forms, whereas
those on the remainder of the page form the building blocks of integration.
■ REVIEW OF SUBSTITUTION
Remember when doing substitution that you must choose u, calculate du, and then
substitute to make the form you are integrating look exactly like one of the known
integration formulas. We will review substitution by looking at different situations
and special substitutions that prove useful.
EXAM PLE 1 Integration by substitution
Find x 2 dx
(x 3 - 2)5'
Solution Let u be the value in parentheses; that is, let u = x 3 - 2. Then
du = 3x 2 dx, so by substitution:
, r du r
x dχ = A = 1 u~ 5du = - - ---- 1- eC = — - ( Λx 3 - 2l'^'4 + CC
(χ 3 _ 7)5 j l l 5 3 J u u u 3 -4 12'' ’
EXAM PLE 2 Fitting the form of a known integration formula by substitution
.-1
∙ ,d f t dt
J √ι - ∕4'
Solution We notice the similarity between this and the formula for inverse sine if
we let a = 1 and u = t2 . If u = t2, then du = 2t dt and
du
t dt 2 If u = t2 , then du = 2/ dt, t dt - ⅛±, and
Vl - z4 Vl - M2 V l - F = V l - (F) 2 = V l - u-.
_ 1 ‘ ,l = 2 sin 1u + C = ∣ sin 1 t2 + C
- 2 2
438 Ch. 7 Methods of Integration
The art of substitution (Section 4.3) is very important, because many of the
techniques developed in this chapter will be used in conjunction with substitution.
Examples 3 and 4 illustrate additional ways substitution can be used in integration
problems.
EXAM PLE 3 Substitution to derive an integration formula
Find J secx dx.

Solution Multiply the integrand secx by secx + tanx and divide by the same
Computational Window quantity:
A computer form for the inte- secx(secx + tanx) secx tanx + sec2 x ,
gral in Example 3 is secx dx = ---------------- r^ t------------- dx
secx + ta n x
secx + tanx
In I tan f—
— ) . Note that the
The advantage of this rearrangement is that the numerator is now the
constant C is assumed to be 0 derivative of the denominator. That is, using the substitution
in many software programs.
u = secx + tan x
du = (secx ta n x + sec2 x) dx
we find
secx tanx + sec~x dx = f —■= ln∣w∣ + C = ln∣secx + tanx∣ + C
secx + tanx
You may wonder why anyone would think to multiply and divide the
integrand secx in Example 3 by secx + tanx. To say that we do it “ because it
works” is probably not a very satisfying answer. However, techniques like this are
passed on from generation to generation, and it should be noted that multiplica
tion by 1 is a common method in mathematics for changing the form of an
expression.
EXAM PLE 4 Substitution after an algebraic manipulation
Find i - ~ λ -
J 1+ t
Solution The straightforward substitution u = 1 + ex does not work:
dx C du r J
I £1 = I This is not an appropriate form

1 + ex
J u J because x has not been eliminated.
w= 1 + du = ex dx
Instead of the straightforward substitution, rewrite the integrand as follows:

1 _ e~x ( 1 j _ e~x
1 + ex e~x ∖ 1 + ex ) e~x + 1
u = e~x + 1; du = -e ~ x dx
I τ ⅛ = f ≠ ⅛ ⅞ = ∖ ^ f Γ = -ln∣M∣ + C = -ln(e-'- + 1) + C
Remember, e~x + 1 > 0 for all x, so ln∣e^x + 1∣ = ln(e~* + 1). B M
Algebraic procedures can also be used to handle integrals of the form

dx
Ax~ ÷ B x ÷ C
If B 2 —4AC ≥ 0, the quadratic A x 2 + B x + C can be factored using only real

numbers, and the integral can be evaluated by a method ofpartial fractions, which
is developed in Section 7.3. If B 2 - 4A C < 0, the denominator cannot be factored
using only real numbers, and we proceed by first completing the square, as shown
in the following example.
EXAM PLE 5 Substitution by completing the square
dx
2 x 2 + 4x + 5’
Solution You might begin by trying the “ obvious” substitution u = 2x 2 + 4x + 5,
but this substitution does not work (try it). Notice, however, that the
denominator is a quadratic form with A = 2, B = 4, and C = 5. Because
B 2 - 4A C = 4 2 - 4(2)(5) < 0
we complete the square in the denominator of the integrand:
2x 2 + 4x + 5 = 2[(x 2 + 2x) + j] = 2[(x 2 + 2x + 1) - 1 + £| = 2∣((x + 1)2 + 5]
Thus,
dx _ f _______ dx_______ _ f du
2x 2 + 4x + 5 J 2 [(x + 1)2 + 1] J 2 [u 2 + ∣
]
u = x + 1; du = dx
Compare this with the rule for the inverse tangent (with a 2 = ⅛j.
When the integrand involves terms with fractional exponents, it is usually a

good idea to choose the substitution x = u n , where n is the smallest integer that is
divisible by all the denominators of the exponents. For example, if the integrand
involves terms such as x 1z4, x 2z3, and x 1z6, then the substitution x = M' 2 is
suggested, because 12 is the smallest integer divisible by the exponential denomi
nators 4, 3, and 6. The advantage of this policy is that it guarantees that each
fractional power of x becomes an integral power of u. Thus,
x 1z6 = (w12) ιz6 = u 2 , x 1z4 = (w12) ιz4 = u i , x 2z3 = (M12)2Z3 = w8
EXAM PLE 6 Substitution with fractional exponents
Find ∣ 1z3^ x 1z2.

J x 1z3 + x 1z2
Solution Because 6 is the smallest integer divisible by the denominators 2 and 3,
we set u = x 1z6, so that w6* = x and 6M5 du = dx. We now use substitution
f dx = f 6M5 du τ = 6
Jx 1z3 + x 1z2 J (M6) IZ 3 + (w6) ιz2
_ ! 6u 5 du _ l^ 6u 5 du _ f 6u 3* du
J u 2 + u 3 J u 2(u + 1) J 1 + u
Substitution does not guarantee an integrable form. When the degree of the
numerator is larger than the degree of the denominator, division is often
helpful. By long division
,16+y u = 6u2 —6M + 6 + 11 + u∙

f = [(βu 2 - 6ιι + 6 + , du
J 1 + u )∖ 1 + u)
= 2w3 - 3M2 + 6ιι —6 ln∣
1 + w∣
+C
l + x ιz6∣
= 2(x ιz6)3 —3(x ιz6)2 + 6x ιz6 —6 ln∣ +C Because u = x ιz6
l + x ιz6∣
= 2x ιz2 - 3x ιz3 + 6x ιz6 - 6 ln∣ +C ■
■ USING TABLES OF INTEGRALS
Example 6 (especially the part involving long division) seems particularly lengthy.
When faced with the necessity of integrating a function such as
f 6u3 du
J I 3- u
most would turn to a computer, calculator, or a table of integrals. If you look, for
example, at the short integration table (Appendix D), you will find (formula 37)
f u 3 du _ («» + b)3 3b(au + b)2 3b2(au + b) b3
J au + b ~ 3a4 2a4 «4 a4 n (a U ,
If we let a = 1 and b = 1, we find

f 6u3 du _ J ( « + 1)3 3 ( u + l) 2 + 3(M +1) I
J τ + i - 6π ------------ 2— — τ------- 1 Φ + 1∣
J
= 2(M + 1)3 - 9(M + l) 2 + 18(M + I) - 6 ln∣
w + 1∣
We note that when using most tables, it is necessary to add the constant C to the
form given by the table. We also note that the algebraic form does not always
match the form we obtain by direct integration. You might wish to show
algebraically that the form we have just found and the form
2u 3 - 3u2 + 6u - 6 ln∣
l + w∣+ C
from Example 6 are the same.
If we use some computer software (such as Derive, Maple, or Mathematica),

for [ &■' ⅛u w e 0 bta in
J 1+ u
—6 LN(M + 1) + 2u 3 - 3M2 + 6u
On the other hand, if we are using computer software, we would not need the
simplified form for Example 6, and we find
ιz6 ιz3 ιz6
γ 1/2, = - 6 LN' (x + 1 ,) + 2Vx - 3x + 6x
If we compare this answer with the one we found in Example 6, we note that
most software programs do not necessarily give the terms in the usual order,
they do not show the absolute value symbols, and they also do not provide the
constant C.
Sec. 7.1 Review of Substitution and Integration by Table 44 1
(cM∏''t∣ T° u s e a n integral table, first classify the integral by form. The forms listed in the
..- J table of integrals included in the Student Mathematics Handbook are as follows:
Elementary forms (formulas 1-29; these were developed in the text)

Linear and quadratic forms (formulas 30-134)
Forms involving au + b', u 2 + a 2; u 2 - a 2 ; a 2 - w2; an + b
and pu + q∖ au 2 + bιι + c
Radical forms (formulas 135-270)
Forms involving Vαw + &; Vαw + b and pu + q∖V , au + b and
Vpz∕ + q; Vw 2 + a 2; λ∕u2 - a 2 ; V a 2 —w2; Vaw 2 + bιι + c
Higher degree binomials (formulas 271-310)
Forms involving u 3 + a 3; w4 + a 4 and u 4 - a 4 ;
u" + a" and u', - a n
Trigonometric forms (formulas 311-444)
Forms involving cosaw; sin aw; sin aw and cosaw; tan aw;
cotaw; sec aw; esc aw
Inverse trigonometric forms (formulas 445-482)
Exponential and logarithmic forms (formulas 483-513)
Forms involving eau ^, In ∣w∣
Hyperbolic forms (formulas 514-619; we study these in Section 7.6)
Forms involving cosh aw; sinh aw; sinh aw and cosh aw;
tanh aw; coth aw; sech aw; csch aw
Inverse hyperbolic forms (formulas 620-650; Section 7.6)
A condensed version of this table is provided in Appendix D.

There is a common misconception that integration will be easy if a table is
provided, but even with a table available there is a considerable amount of work.
After deciding which form applies, match the individual type with the problem at
hand by making appropriate choices for the arbitrary constants. More than one
form may apply, but the results derived by using different formulas will be the
same (except for the constants) even though they may look quite different. We
shall not include the constants in the table listing, but you should remember to
include it with your answer when using the table for integration.
Take a few moments to look at the integration table in the Student Mathemat
SM H
ics Handbook (or Appendix D). Notice that the table has two basic forms of
integration formulas. The first gives a formula that is the antiderivative, whereas
the second, called a reduction formula simply rewrites the integral in another
form.
Table of Integrals
INTEGRALS INVOLVING UU + b
(au + b)n + '

30. (αw + b)ndu =
(n + l)α
(au + ⅛)"+2 b(au + ⅛)"+ l
u(au + b~)ndu =
(n + 2)α2 (n + l)α 2
,. ,. „ , (au + ⅛)"+3 2b(au + ⅛)"+2 b 2(au + b)"+ '

u 2(au
v
+ b)ndu = ——— ,
' (n + 3)α3 (n + 2)α3 (n + l)α 3
Historical Note INTEGRALS INVOLVING a u + b
u m + ∖ au + b) nb u", (au + b)n 'dιt

m +n+ 1 m +n+ 1
u m (au + b)n+ ' mb
33. u m (au + bydu = ιιm '(au + bydu
(m + n + l)α (m + n + l)α
~U ' (au + b)"+ ' +n +2
m um (au + b)n + 'du
(/7 + 1)⅛ (H + 1)b
— ~7-7 = - ln[αw + b∖
au + b a 1
EXAM PLE 7 Integration using a table of integrals— antiderivative given

AMALIE (“EMMY”) N OETHER
(1882-1935)
Find J x 2(3 - x) 5 dx.
Emmy Noether has been called
the greatest woman mathemati Solution This is an integral of the form au + tr, we find that this is formula 32,
cian in the history of mathe where u = x , a = -1,⅛ = 3, and n = 5. Remember, that the integral table does
matics. Once again we see a not show the added constant, so do not forget it when writing your answers
woman overcoming great obsta using the table.
cles in achieving her education.
At the time she received her (3 ~ x ,)5+3 2(3)(3- x ) 3÷2 3 2( 3 ~ X ) ^ '
Ph.D. at the University of (5 + 3 ) ( - l) 3 (5 + 2 ) ( - l) 3 (5 + 1 )(-1)3
Erlangen (in 1898), the Aca
demic Senate declared that the = - ∣( 3 - x) 8 + ¾3 - x) 7 - ∣(3 - x ) 6 + C
admission of women students
would “overthrow all academic This form is acceptable as an answer. However, you may wish to complete the
order.” Emmy Noether is fa algebraic simplification:
mous for her work in physics
and mathematics. In mathemat x 2(3 - x y d x = -⅛ (3 - x) 6[7(3 - x) 2 - 48(3 - x) + 84] + C
ics, she opened up entire areas
for study in the theory of ide = -⅛ (3 - x) 6(7x2 + 6x + 3) + C
als, and in 1921 she published
a paper on rings, which have One of the difficult considerations when using tables or computer software to
since been called Noetherian carry out integration is recognizing the variety of different forms for accepta
rings. (Rings and ideals are ob ble answers. Note that it is not easy to show that these forms are algebraically
jects of study in modern alge
equivalent. If we use computer software for this evaluation, we find yet
bra.) She was not only a re
nowned mathematician, but another algebraically equivalent form:
also an excellent teacher. It has ∫ x 2(3 - x ) 5dx = + i y - - 15x6 + 54x5 - + 81x 3
been reported that she often
lectured by “working it out as
she went.” Nevertheless, her Once again, remember that you must add the + C to the computer software
students flocked around her “as form. M M
if following the Pied Piper.”
EXAM PLE 8 Integration using a table of integrals— reduction formula
Find J (In x) 4 dx.

Solution The integrand is in logarithmic form; from the table of integrals we see
that formula 501 applies, where u = x and n = 4. Note that tables of integrals
do not usually show the absolute value symbols. You need to remember that
In x is defined only for positive x.
SM H
Sec. 7.1 Review of Substitution and Integration by Table 4 43
Take a close look at formulas 499 to 502, and note that formula 501 gives
another integral as part of the result. This is called a reduction formula because
it enables us to compute the given integral in terms of an integral of a similar
type, only with a lower power in the integral.
(In x)4 dx = x(ln x)4 - 4 j (In x)4 ^ 1 dx
= x(ln x)4 —4^x(ln x)3 —3 J (In x) 3-1 dx^ Formula 501 again
= x(ln x)4 - 4x(ln x)3 + 12 (In x) 2 dx
= x(ln x)4 - 4x(ln x) 3 + 12[x(ln x) 2 - 2x In x + 2x] + C

This is formula 500.
= x(ln x)4 - 4x(ln x) 3 + 12x(ln x)2 - 24x In x + 24x + C s≡≡
Note from the previous example that we follow the convention of adding the
constant C only after eliminating the last integral sign (rather than being
technically correct and writing C1, C2, . . . for each integral). The reason we can do
this is because C1 + C2 + ∙∙∙ = C, for arbitrary constants.
It is often necessary to make substitutions before using one of the integration
formulas, as shown in the following example.
EXAMPLE 9 Using an integral table after substitution
x dx
√8 - 5x2 '
Solution This is an integral of the form V<z2 - M2, but it does not exactly match
any of the formulas.
Note, however, that except for the coefficient of 5, it is like formula 225. Let
u = V5 x (so id = 5x2); then du = Vz5 dx∖
u du u
f X dx _ C √5 ' √ 5 Let u = V5 x so that x =
J V8 - 5x2 J VΓ ς T2 DU
du = V5 dx, so that dx = ~~p.
_Xf u dιι v5
5J
Now apply formula 225 where a 2 = 8:
1 f K2‰ = l ( - √ 8 ^ 5 ) + C = - ∣ √ 8 - 5x2 + C
J V o ~ U~
As you can see from Example 9, using an integral table is not a trivial task. In
fact, other methods of integration may be preferable. For Example 9, you can let
u = 8 - 5x 2 and integrate by substitution:
Let u = 8 - 5x 2;
du = -1 Ox dx.
x dx u 1/2 du
√8 - 5x 2
= -⅛(2w ιz2) + C = - ∣ V 8 - 5x 2 + C
O f course, this answer is the same as the one we obtained in Example 9. The point
of this calculation is to emphasize that you should try simple methods of
integration before turning to the table of integrals.
Sometimes, by choosing different formulas you may obtain two different
looking (but equivalent) forms.
E X A M P L E 10 Integration by table (multiple forms with substitution)
Find 5x 2V 3 x 2 + 1 dx.
Solution This is similar to formula 170, but you must take care of the 5 (constant
multiple) and the 3 (by making a substitution).
Let u = V 3 x; then du = V'3 dx.
5x 2V 3 x 2 + 1 dx = 5 [ + 1⅛
↑ ∖
u 1 = 3x 2 , so that x 2 = y
Computational Window - π
M2V M2 + 1 du Use formula 170 where a = 1.
The form that we obtain when
using computer software can
be shown:
5√3 LN (√3x 2 + 1 + √3x)
72 = [2√3x(3x 2 + l) 3z2 - √ 3 x√ 3 x 2 + 1 - ln∣√3 x + V 3 x 2 + 1∣ + C
5xV3x 2 + l(6x 2 + 1)
+ 24 = ^ ^ 2 x (3 x 2 + l) 3z2 - x V 3 x 2 + 1 - —τ= ln∣λ∕3 x + V 3 x 2 + l∣j
We now show how one of the formulas in a table of integrals might be derived
by using substitution.
E X A M P L E 11 Deriving an integral formula
Derive a formula for j tan 3 ax dx (formula 405).

Solution We will derive this form by using substitution and a trigonometric
identity.
j~ tan 3 ax dx = j~ tan ax(tan 2 ax) dx = J tanαx (sec2 ax - 1) dx
= tan ax sec2 ax dx - tan ax dx

Let ιι = tan ax,
so du = a sec2 ax dx.
Formula 403
=taιΓax
2a
+ - ln∣cosαx∣
a ' 1
+ C
This corresponds to formula 405 in the Student Mathematics Handbook.
■ TRIGONOMETRIC SUBSTITUTIONS
Another useful type of substitution is known as a trigonometric substitution. In

order to see why trigonometric substitutions can be useful, suppose an integrand
contains the term V α 2 - M2, where a > 0. Then by setting u = a sin θ and
using the identity cos2 θ = 1 - sin 2 θ, we obtain
V a 2 - u 2 = V α 2 - α2sin 2 θ = a V l - sin 2 θ = a cos θ
Thus, the substitution u = a sin θ, du = a cos θ dθ eliminates the square root and
converts the given integral into one involving only sine and cosine. This
substitution can best be remembered by setting up a reference triangle. This
process is illustrated by the following example.
E X A M P L E 12 Trigonometric substitution with form y∣

a2 - u2
Find J V 4 - x 2 dx.
Solution First, using a table of integration we find (formula 231; a = 2)

∫ vT z T 2 dx = λ 'v 4 χ l + 2 sin" 1 θ θ + C
2~
Our goal with this example is to show how we might obtain this formula with a
trigonometric substitution. Refer to the triangle shown in Figure 7.1. Notice
that sin 0 = so that x = 2 sin θ. Then
cos θ = 2 X s o t *ι a t v *4 - x 2 = 2 cos θ
J V4 - x 2 dx = j^ (2 cos 0)(2 cos 0 dθ)
Let x = 2 sin θ so Vz4 —x 2 = 2 cos 6;

dx = 2 cos θ dθ.
sin θ = ∣
= 4 J cos2 0 dθ
Figure 7.1 Reference triangle with
form √ 4 - x
2 = 4 J 1 + cos 20 c∣
g Half-angle identity
= 2 J (1 + cos 20) dθ
=2 dθ + 2 J cos 20 dθ
Let u = 26;
du = 2 dθ.
= 20 + 2 C
= 20 + sin20 + C
The final step is to convert this answer back in terms of x. Because x = 2 sin 0,
we know
j = sin 0
0 = sin -1 (jθ
To simplify sin 20 we need a double-angle identity:

sin20 = 2 sin 0 cos 0 = x V 1 - sin2 0 = x -^ 1 - (j) = ^V4 - x 2
Thus,
J V4 - x 2 dx = 20 + sin 20 + C = 2 sin -1 + ^V4 - x 2 + C
Similar methods can be used to convert integrals that involve terms of the
form Vα 2 + u 2 or V M2 - a 2 into trigonometric integrals.
PROBLEM SET 7 .1
Find each integral in Problems 1-12. 4Λ-3 - 4x 10 [ ------ χ 3 ~ x ___ dx
f 2x - 1 d f 2x + 5 x 4 - 2x 2 + 3 J (x 4 - 2x 2 + 3)2
1∙ 1 2. dx
' (4x 2 - 4x)2 a ∣V x 2 + 5x
2x + 4
3. 1∖ ⅛ ^dx fln(x + 1) x 2 + 4x + 3
1 4. ∣ dx
X x + 1
5. 1I cosx e sιn x dx 6 ∙ 1Γ x dx Integrate the expressions in Problems 13-24 using the short
'4 + x 4 table o f integrals given in Appendix D.
-ι ∣t 1 dt (1 + cotx) 4 csc2 x dx dx dx
J 9 + t6 x 2V x 2 - a 2 x 2∖∕a 2 - x 2
15. x In x dx 16. I∣ In x dx 17. ∣

x e ax dx 48. J sin 3 x cos4 x dx Hint: Let u = cosx.
Γ dx ∣ x 2 dx
19∙ 1 Γ dx 49. sin 2 x cos2 x dx Hint: Use the identities shown in
18. 1
' a + be 2x 20. 1
∣ x 2V x 2 + 16
Vx2 + 1 Problems 45 and 46.
f x dx
22. I∣ dx 50. ■ What Does This Say? Using Problems 47-49, formu
21. ∣
∣ √ 4x2 + 1 23. 1e - 4 ∙v sin 5x dx
'x √ i - 9x 2 late a procedure for integrals of the form
24. I x sin 1x dx sin'"x cos"x dx
Evaluate the integrals in Problems 25-34. I f you use an 51. ■ What Does This Say? Explain the process of using
integral table, state the number o f the formula used, and i f you a trigonometric substitution on integrals involving
use substitution, show each step. I f you use an alternative table V α 2 - M2 .
o f integrals, then cite the source as well as the formula number. 52. ■ What Does This Say? Explain the process of using
25. f( l + bx)~ l dx 26. f , x ^x ≡ 27. f x (l + x) 3 dx a trigonometric substitution on integrals involving
J J v a — χ- I V α2 + if.
x V 1 + x dx x e 4x dx x In 2x dx 53. ■ What Does This Say? Explain the process of using
a trigonometric substitution on integrals involving
V M2 - of.
ln 3x dx
Find each integral in Problems 54-59.
x 3 dx ex dx f dx
√ 4x4 + 1 54. IΓ
1 1 + eλ'z2 55. 1| χ l∕2 _|_ χ l∕4
Work Problems 35-38 by two methods. Γ18 tan 2 t sec2 / f 4√x

56. I1 (2 + tan 3 t)2 57. ∣
∣ x ιz3 + 2 x ιz2
35. [ ∕ ‰ 2
J √9 - x
36. [ - *±1= dx
J √4 + x 2 f dx e x
58. I (x + ⅜)√4x2 + 4x 59. ∣2x ~ ~2 z√
37. J( 9 + x 2) ιz2 dx 38. I e + e^ * + 2
60. Find the area of the region bounded by the graphs of
Use the Student Mathematics Handbook or other available
integration table to integrate the integrals given in Problems y = , ^ -, y = 0 from x = 0 to x = 4.
V x2 + 9
39-44.
39. ∫
v4 ζ + 1∙ dx 40. sec3 dx y = cos 2x, y = 0, x = 0, and x = J .
62. Find the volume o f the solid generated when the curve
41. J s in 6 xcZx 42. dx y = x (l - x 2) ιz4 from x = 0 to x = 1 is revolved about the
9x 2 + 6x + 1 x-axis.
43. f (9 - x 2) 3z2 dx 44. 5≡ 2 ^ 63. Find the volume of the solid generated when the curve
∕x
cosx
45. Derive the sine squared formula shown on the inside back
cover (formula 348):
sin 2 x dx = ∣ x - ∣ sin 2x
Hinf. Use the identity sin2 x = ----- CQ⅞-2x.

46. Derive the cosine squared formula shown on the inside
back cover (formula 317):
cos2 x dx = ⅛x + ∣ sin 2x
Hinf. Use the identity cos2 x = 1 + 5,os--.
Problems 47-49 use substitution to integrate certain powers o f

sine and cosine.
x=< z4≡P
47. sin4 x cosx dx Hint'. Let u = sinx.
between y = 1 and y = 2 is revolved about the y-axis.
65. Find the arc length o f the curve y = x 2 on the interval Derive the integral formulas in Problems 71-73. Assume that
[0, 1]∙ m and n are integers.
66. Find the arc length of the curve y = ln(cosx) on the π rπ
interval [0, π∕4].

67. Find the surface area o f the surface generated when the
J
f
o
x ∕(sin x) dx = ?
∙ ∙
2
;
Jo
f0
/(sin x) dx
if m Ψ n
curve y = x 2 on the interval [0, 1] is revolved about the 72. sin nx sin m x dx = i ∙
x-axis. 17τ itr m = n
68. Find the surface area o f the surface generated when the f" . 0 ιh ≠ m
curve y = x 2 on the interval [0, 1] is revolved about the 73. cos nx cos m x dx = i ∙
J —π 1 7τ iff n = m
j--axis.
θ 69. Show that 74. Find the surface area of the torus generated when the
curve
cscx dx = -ln ∣cscx + cotx∣ + C
x 2 + (y - b)2 = 1,
Hint: Multiply the integrand by
is revolved about the x-axis.
cscx + cotx
cscx + cotx’
70. Find J 2 sin x cosx dx by using the indicated substitution,
a. Let u = cosx. b. Let u = sinx.

c. Write 2 sinx cosx = sin2x and carry out the integra
tion.
d. Show that the answers you obtained for parts a-c are
the same.
7.2 INTEGRATION BY PARTS
IN THIS SECTION Integration by parts formula, repeated use of

integration by parts, definite integration by parts, solving a differential
equation by parts
Integration by parts is a procedure based on reversing the product rule for
differentiation. We present this section not as a technique of integration but as
a procedure that is useful in a variety of different applications.
■ INTEGRATION BY PARTS FORMULA
Recall the formula for differentiation of a product. If u and v are differentiable

functions, then
d(uv) = u dv + v du
Integrate both sides of this equation to find the formula for integration by parts:
J d(uv) = J ιι dv + J v du
ιιv = J u dv + J v du
If we rewrite this last equation, we obtain the formula summarized in the
following box.
Formula for Integration by

Parts u dv = uv - v du
Sec. 7.2 Integration by Parts 449
EXAM PLE 1 Integration by parts
Find j xe *x dx.
Solution To use integration by parts, we must choose u and dv so that the new
integral is easier to integrate than the original.
J x e^-dxy
u dv
Let u = x , dv = ex dx
du = dx an v = ∫ ex dx = ex
xe x dx = x ex c' dx = xe x - ex + C
u v v du
You can check our work with integration by parts by differentiating, using
software, or an integration table. This is found in Appendix D or the Student
Mathematics Handbook (formula 484, with a = 1). t—= >
SM H
COMMENT: You may wonder why an arbitrary constant (call it K) was not
included when performing the integration associated with ∫ dv in integration
by parts. The reason is that when applying integration by parts, we need just
one function v whose derivative is dv, so we take the simplest one—the one
with K = 0. You may find it instructive to see that taking v = ex + K gives the
same result.
Integration by parts is often difficult the first time you try to do it, because
there is no absolute choice for u and dv. In Example 1, you might have chosen
u = ex and dv = x dx
γ 2
du = ex dx v = ∣x dx = zy
Then J xe x dx = ex - J y ∙ ex dx
u v v du
= J x e - J J x 2ex dx
2 x
Note, however, that this choice of u and dv leads to a more complicated form than
the original. In general, when you are integrating by parts, if you make a choice for
u and dv that leads to a more complicated form than when you started, consider
going back and making another choice for u and dv.
EXAM PLE 2 When the differentiable part is the entire integrand
Find J In x dx for x > 0.
Solution Let u = In x
and
, dv = dx
1
du = - dx v=x
x
In x dx = (In x)x - x( - d x ) = x In x - dx = x ∖n x - x + C
I I ∖√ι ) I
Check with formula 499 (Appendix D) where a = 1, which we worked as

Problem 16 of Problem Set 7.1. E ∏ΞΞ 3
■ REPEATED USE OF INTEGRATION BY PARTS
Sometimes integration by parts must be applied several times in order to evaluate

a given integral.
EXAM PLE 3 Repeated integration by parts
Find x2 e x dx.
Solution Let u - x 2 dv = e x dx
and
du = 2x dx v = -e~ x
J x2 e x dx = x 2( - e x) - J ( - e λ')(2x dx)
= - x 2e~x + 2 j^χ e ~x dx
Let u = x , dv = e~x dx
djιι - dιx a∏θ v = —e x
= - x 2e~x + 2∣∖r(-e - ∙γ) - [ ( - e ~ x ) dx
= - x 2e~x - 1xe~ x + 2 ( ∖ jj + C
= - e ~ x (x 2 + 2x + 2) + C
Check with formula 485 where a = - 1 .
In the following example it is necessary to apply integration by parts more

than once, but as you will see, when we do so a second time we return to the
original integral. Note carefully how this situation can be handled algebraically.
EXAM PLE 4 Substitution with an algebraic manipulation
Find j e 2x sm x dx.
Solution For this problem you will see that it will be useful to call the original
integral I. That is, we let
I = J e 2x sinx dx
Let u = e2x , dv = sin x dx

, . , v , τ and
du = 2e~ dx v = -c o s x
I = J e2x sinx dx = e2x(-co sx ) - ∣
^(-co sx)(2 e 2x dx)
= - e 2λ'cosx + 2 j e 2x co sx dx
Let u = e2x , dv = cos x dx

du = 2e2x dx an v = sin x
= -β 2xcosx + 2ê2λ(sinx) - j" sinx(2e 2x dx)^
= - e 2xcosx + 2e~γsinx - 4 e2γsinx dx

Notice that this last integral is I. Thus, we have

I = -<j2γcosx + 2<?~Ysinx - 4/ Solve this equation for I.
51 = - e 2'γcosx + 2e2-γsinx
I = ∣ c 2γ(2 sinx - cosx)
Thus, J e2γsinx dx = ∣ e2-γ(2 sinx - cosx) + C

Check with the integration table (formula 492, where a = 2 and b = 1). ( = J
■ DEFINITE INTEGRATION BY PARTS
In order to get a clear picture of the integration by parts formula for definite
integrals, it is instructive to interpret the integration by parts formula in terms of
area. Consider the area of the rectangle with sides w2 and v2, as shown in Figure
7.2.
AREA OF RECTANGLE = BLUE AREA + PINK AREA + PURPLE AREA
N__________________ v __________________ / v √ I J V √
Entire colored rectangle .u - v,

u2y 2 = u t vl + v du + u dv
Jul Jvl
This equation can be written as
Figure 7.2 Integration by parts ^w dv = u2 v2 - u l v.

using areas
This relationship is summarized in the following box, where we use the usual
limits of integration, namely, x = a to x = b.
Integration by Parts for rb ∣

⅛ rb
Definite Integrals u dv = uv∖α - I
v du
la Ja
You should recognize that this formula for definite integrals is the same as the
formula for indefinite integrals, where the first term after the equal sign has been
evaluated at the appropriate limits of integration. This is illustrated by the
following example.
EXAM PLE 5 Integration by parts with a definite integral
x e l x dx.
Solution Let u = x . dv = e2xdx

and
Ξ Computational Window f~ ∣ 1 du = dx v = ⅛e2f
You can use computer software
or many calculators to evaluate
definite integrals. For example,
with CONVERGE software and It is usually easier to simplify the algebra
1,000 subintervals, we evaluate and then do one evaluation here at the end
the integral in Example 5 as of the integration part of the problem.
2.093571. Compare with the
decimal approximation of = (⅛xe2* - ⅛ 2γ)∣' = ⅛ 2 + ⅛
∣e 2 + { ≈ 2.097264025.
Check in Appendix D (formula 484, with a = 2).
EXAM PLE 6 Integration by parts followed by substitution
rτr∕4
Evaluate tan - l xα⅛.
Jo
Solution Let u = tan 1 x dv = dx
, dx and
Computational Window du = ζ------ 25 v= x
1+ x
For n = 1,000, we find I π∕4
Γτr∕4 tan~ l x√x = (tan^ 1x) x n
tan ^ l x dx ≈ 0.282396 -'0 ' u ' civ ' u ' v
Jo
as compared to the answer in π∕4
Example 6:
0
¾ an^1 - 0.5 Inf 1 + ⅛ Use substitution where w = 1 + x 2 and
4 ∖4∕ ∖ 16/ dw = 2x dx.
J- In w = ln( 1 + x 2)
≈ 0.2826574895
You must think about the an
swers you obtain when using
computer software. There are
entire courses that discuss the = ⅛ ( t a n ^l f ) - ⅛ ln (1 + ⅛ ] - [o - ⅛In 1]
accuracy of various computer
approximations. ta n
=f ^ 'f “ 2 ln ( 1 + ⅛ )
Check with an integration table (formula 457, for example, with a = 1). ≈ M
■ SOLVING DIFFERENTIAL EQUATIONS BY PARTS*
Integration by parts can be used in any of the applications we have considered.

The following example involves a differential equation.
EXAM PLE 7 First-order linear differential equation
Solve the differential equation + f = e2x.
Solution Recall from Section 6.4 that the general solution of the equation
+ P(x)y = <2(x) is given by y = J ^ J [∫(2(X )7(X ) dx + C]
where ∕(x) = e⅛ x >dx is the integrating factor. For the differential equation in
this example, we have P(x) = x~ 1, O(x) = e2x, and
∕(x) = efχ ~1dx = eln -v = x
Thus, the general solution is
y = [χe 2x dx + c j Integrate fxe2x dx by parts.
= ±[⅛xe2x —∣
e2∙γ + C] See Example 5 for details.
= ⅛ 2 .v _
2
β ≤ + C
-e 4x + x _
*This material is not required for subsequent material.
PROBLEM SET 7 .2
O Find each integral in Problems 1-14. 33. Find the volume of the solid generated when the region
under the curve y = e~x on the interval [0, 2] is revolved
2x
ι . 1xe dx 2. x sin x dx 3. x In x dx about the j-,-axis.
34. Find the volume of the solid generated when the region
∣x ta n - l x √ x
4. 1 5. 1sin - 1 xrZx 6. 1x 2 sinx dx under the curve y = sinx + cosx on the interval [θ, is
revolved about the v,-axis.
f⅛ ‰ 8. 1
7. I ∣ e l x sin 3x dx 9. 1x cos2 x dx 35. Find the volume of the solid generated when the region
Vx J under the curve y = In x on the interval [1, e] is revolved
10. IΓ x 3 dx 1 about the indicated axis.
, 2 Vx + l
11
π ∙J x 2 ln x r ∕x 12. I (x + sinx) 2 dx
a. x-axis b. y-axis
13. ∣ 36. Find the centroid (with coordinates rounded to the near
I e 2λ√ 1 - ex dx 14. 1x sin x cosx dx
est hundredth) of the region bounded by the curves
y = sin x and y = cos x and the y-axis on 0 ≤ y ≤ 1.
Evaluate the definite integrals in Problems 15-20. 37. Find the centroid (with coordinates rounded to the near
f4 Γ ce 3 est hundredth) of the region bounded by the curves
15. V x In x dx 16. x In x dx y = e x , y = e~x , and the line x = 1.
Ji Ji
Problems 38-42 require the material dealing with differential
17. Γ(ln x) 2 dx 2
18. f 3(In 3x) dx equations.
Jι J1∕3 38. Solve ⅞- + 7^√∙ = x(l + x).
dx 1+ x v ’
19. c 2 'cos 2X dx 20. [ x(sinx + cosx) dx
Jo Jo dy 2xy = s , n x
39. Solve + subject to the condition y = 1
θ 21. ■ What Does This Say? Describe the process known as
where x = 0.
integration by parts.
40. Find a function ∕( x ) whose graph passes through (1, 1)
In Problems 22-25 first use an appropriate substitution and and has the property that at each point (x, y) on the graph,
then integrate by parts to evaluate the integral. Remember to the slope of the tangent line is x 2 + 2y.
give your answers in terms o f x. 41. Suppose it is known that /(0) = 3 and
J f,ττ
fln x s in ( ln x ) [ .
22. -------- - -------- dx 23. [sin 2x In(cosx)]αx [ ∕( x ) + ∕"( x ) ]s in x dx = 0
o
24. J e 2-v sin ex dx 25. J[ s in x ln(2 + cosx)]Vx What is/ ( TT)?
42. Because a rocket burns fuel in flight, its mass decreases
f v-3 with time and this in turn affects its velocity. It can be
26. a. Evaluate — - dx by parts. shown that the velocity v(∕) of the rocket at time t in its
flight is given by
b. Evaluate the integral in part a by first dividing the
v(z) = - r l n ^ r ^ - gt
integrand.
Evaluate the integrals in Problems 27-28 by parts. where w is the initial weight o f the rocket (including its
fuel) and r and k are, respectively, the expulsion speed and
27. J sin 2 x √ x 28. J cos2 x dx the rate o f consumption of the fuel, which are assumed to
be constant. As usual, g = 32 ft∕s 2 is the constant accelera
29. Find x" In x dx, where n is any positive real number. tion due to gravity. Suppose w = 30,000 lb, r —8,000 ft∕s,
and k = 200 lb∕s. What is the height of rocket after 2 min
30. After t seconds, an object is moving at a speed o f te~υ l
(120 sec)?
meters per second. Express the distance the object travels 43. In a medium for which air resistance is proportional to
as a function o f time. the velocity of a falling body, the velocity of a body with
mass m satisfies the differential equation
31. After t hours on the job, a factory worker can produce
∖OOte~o ∙5t units per hour. How many units does the m = mg - kv
worker produce during the first 3 hours?
32. After t weeks, contributions in response to a local fund- where k is a positive constant and g is the constant of
raising campaign were coming in at the rate o f 2,OOO∕e- o -2z acceleration due to gravity. Find v(t) if the initial velocity
dollars per week. How much money was raised during the of the body is v0 in feet per second.
first 5 weeks? 44. A lake has a volume o f 6 billion ft 3, and its initial
pollutant content is 0.22%. A river whose waters contain 47. If n moles of an ideal gas expand at constant temperature
only 0.06% pollutants flows into the lake at the rate of 350 T, then its pressure p and volume V satisfy the equation
million ft 3∕day, and another river flows out of the lake p V = ∏RT, for constant R. It can be shown that the work
also carrying 350 million ft 3∕day. Assume that the water done by the gas in expanding from volume lz 1 to V is
in the two rivers and the lake is always well mixed. How
long does it take for the pollutant content to be reduced to W = nRT In y
0.15%?
45. The displacement from equilibrium of a mass oscillating What is the average work done as V increases from K1 to
at the end of a spring hanging from the ceiling is given by V1 = 10F,?
48. In physics, it is known that loudness L of a sound is
y = 2.3e - 0 -25'cos 5/
related to its intensity I by the equation
feet. What is the average displacement (rounded to the
nearest hundredth) of the mass between times t = 0 and L = 10 lo g y
-, o
t = τr∕5 seconds?
decibels where I o = 12-1 2 w att∕m 2 is the threshold of
audibility (the lowest intensity that can be heard). What is
the average value of L as the intensity of a TV show ranges
between IQ and Zl = 3 ■10^5 watt/m 2?
θ 49. Derive the formula
( x πe x dx = x"e x — n jx " ~ l e x dx
(This is formula 486 with a = 1.)

46. A photographer is taking a picture of a clever sign on the 50. Derive the formula
back of a truck. The sign is 5 ft high and its lower edge is J(ln x)"dx = x(ln x) n — n j(ln x)"^ 1 dx
1 ft above the lens of the camera. At first the truck is 4 ft
away from the photographer, but then it begins to move (This is formula 501.)
away. What is the average value of the angle θ subtended
51. Wallis’s formula If n > 0 is an even integer, use reduction
by the camera lens as the truck moves from 4 ft to 20 ft
away from the photographer? formulas to show
Γ z2∙ √ Γ z2 J Γl ∙ 3 ∙ 5 — (« - 1)^]
Jo sm∕7√Λ- = jθ c o s n d x = [ 2 ∙ 4 ∙ 6 - ∕ 7 'J 7r2
52. State and prove a result similar to the one in Problem 51

for the case for odd exponents.
7 .3 THE METHOD OF PARTIAL FRACTIONS
IN THIS SECTION Decomposition, integrating rational functions by

partial-fraction decomposition, substitution with partial fractions, rational
trigonometric integrals
Partial-fraction decomposition has great value as a tool for integration. This
process may be thought of as the “reverse” of adding fractional algebraic
expressions, and it allows us to break up rational expressions into simpler
terms.
■ DECOMPOSITION
Partial-fraction decomposition is an algebraic procedure for expressing a reduced

rational function as a sum of fractional parts. For example, consider a typical
Sec. 7.3 The Method of Partial Fractions 4 55
algebraic process of adding rational expressions (and the necessity of finding a

common denominator):
2 , -3 = 2(X + 2) + (-3)(x + 1) = -x+ ι
χ + l x+2 (x + l)(x + 2) x 2 + 3x + 2
In partial-fraction decomposition, we do just the opposite: We start with the
reduced fraction -x + j
2
x + 3x + 2
Computational Window E K ∣
Software programs will decom and write it as the sum of fractions

pose rational expressions quite
handily using the direction +1 x+2
“expand.”
The rational expression
Λ x)
f(x ) = D(x)
can be decomposed into partial fractions if P and D have no common factors and if
the degree of P is less than the degree of D. If the degree of P is greater than or
equal to the degree of D, then use division to obtain a polynomial plus a proper
fraction. For example,
x 4 + 2x 3 - 4x 2 + x - 3 = x 2 + 3x + 1 + — -— — This was found by
xλ - x - 2 5 x long division.
polynomial
Proper fraction:
This is the part
that is decomposed
into partial fractions.
Partial-fraction decomposition has great value as a tool for integration. In
Example 5, we shall find
x 4 + 2x 3 - 4x2 + x - 3
x2 - x - 2
which is difficult (if not impossible) without partial-fraction decomposition or
appropriate computer software. The process consists of dividing to find a
polynomial (which is easy to integrate) plus a proper fraction, which is simplified
using partial-fraction decomposition.
In algebra, it is shown that a proper fraction can be written as the sum F + F,
+ ■∙ ■+ w h≡re each F. is of the form
A Ax + B
(x ~ r)n (χ 2 + sx + t)m
for constants r, s, and t. Also assume that x 2 + sx + t is an irreducible (that is, can
not be factored into linear factors) quadratic term. We begin by focusing on the
first form.
Partial-Fraction Decomposition: A. A^
-------- f 7---- ~ 2-, + ∙∙∙ +
A Single Linear Power X- r (χ - r) (x - r)n
EXAM PLE 1 Partial-fraction decomposition with a single linear power
Decompose λ into a sum of partial fractions.

(x —2)
Solution 2- 6Λ^+ 3 = A \ , A 2 , A i
X
(x - 2)3 x - 2 (χ - 2)2 (x - 2)3
Multiply both sides by (x - 2)3 to obtain:
x 2 —6x + 3 = √4l (x —2)2 ÷ Λ f x —2) + A 3
Let x = 2-. 2 2 -6(2) + 3 = Λ 1( 0 ) + Λ ,( 0 ) + Λ 3
-5 = Λ3
Substitute √43 = —5 and expand the right side to obtain
x 2 —6x + 3 = √41x 2 ÷ ( —4.41 + A 1)x + (4,41 —2√4, —5)
This implies (by equating the coefficients of the similar terms) that
1 = .4 l x 2 terms
—6 = —4^ 1 + A 2 x terms
3 = 4.41 —2∕12 —5 Constants
Because √41 = 1, we find A 2 = - 2 , so the decomposition is
x 2 - 6x + 3 _ 1 + -2 + -5
(x - 2)3 x - 2 (x - 2)2 (x - 2)3 —
If there are two or more linear factors in the denominator, the fraction should
be decomposed into distinct linear powers. For example,
_____________ P{χ)_____________
( x - r l ) ( x - r2) ∙∙∙ ( x - rn)
is decomposed into separate terms
This process is illustrated with the following example.
EXAM PLE 2 Partial-fraction decomposition with distinct linear factors
Q γ _ 1
Decompose 2 2 γ _ 7∙
s ° lu t i° n J - x - 2 = ( x - 2 ) ( χ 1+ 1) First factor the denominator,
if possible, and make sure
there are no common factors.
= , A 2 Break up the fraction in parts,
x - 2 x + 1 each with a linear denominator.
The task is to find A l and A 2.
0 Note that the degree of the A f x + 1 ) + A 2(x - 2) Obtain a common denominator
denominator is the same num on the right.
(x - 2)(x + 1)
ber as the number of arbitrary
constants, A 1 and ∠42 . This pro Now, multiply both sides o f this equation by the least common denominator,
vides a quick intermediate which is (x - 2)(x + 1) for this example.
check on the correct procedure.
8x - 1 = Λ 1( x + 1) + A 2(X - 2 )
Make this check a standard step
in your developmental task. 0 Substitute, one at a time, the values that cause each of the factors in the least
Sec. 7.3 The Method of Partial Fractions 457
common denominator to be zero.

Let x = -1 : 8.x - 1 = A 1(x + 1 ) + √42(x - 2)
8(-1) - 1 = Λ 1(-1 + 1) + Λ 2(-1 -2 )
- 9 = -3 A 2
3 = A2
Let x = 2; 8x - 1 = ∠41(x + 1 ) + A 2(x - 2)
8(2) - 1 = Λ ,( 2 + 1) + A 2( 2 - 2 )
15 = 3A l
5= At
If there is a mixture of distinct and repeated linear factors, we combine the

procedure illustrated in the preceding examples. For example,
5x 2 + 21x + 4 ■ ddpepc∩oτmn n ∩ς p d a s A t +
A i +
A i
(x + l) 2(x - 3) P°sed as χ + 1 + + 1)2 + γ _ 3∙
Note that the degree of the denominator is three and we use three arbitrary
constants √4p A 2, and A y
5x 2 + 2 lx + 4 ∙ d p p ∩r n r ι ∩s p d a ς
A \ A i A i A 4
(x + 1)3(x - 3) decomposed χ + 1 ++ (λ - + 1)2 ++ (x + 1)≡ ++ x - 3’
The degree of the denominator is four, so we use four constants.

If some of the factors of the denominator are irreducible quadratic factors,
then the numerators have the form A x + B , as indicated in the following box.
Partial-Fraction Decomposition: A ,x + B. A 2x + B-, A x + B

A Single Irreducible Quadratic x 2 + sx + t (x 2 + sx + t)1 " ' (x 2 + sx + t)m
Factor Because the degree of the denominator is 2m, we have 2m arbitrary constants,
namely,
-z , A ,,
1’ √4,,
2’ . . . ’, A m ,, B 1’
,, B 2’
., . . . 5, B m .
EXAM PLE 3 Partial-fraction decomposition with a single quadratic power
—3x*3 — V
Decompose 2 2.
Solution The decomposition gives

_ a v 3_ r A .x + B . A 1x + B ,,
(x 2 + 1)2 ^ (x 2 + 1)2 x2 + 1
Multiply both sides of this equation by (x2 + 1)2, and simplify algebraically.
- 3 x 3 - x = (A l x + B ∣
) + (A 2X + B 2 )(x2 + 1)
= ∕ l 2x 3 + β 7x 2 + (√41 + A-,)x + (B l + B 2)
Next, equate the coefficients on each side of this equation and solve the
resulting system of equations to find A } = 2 ,A 2 = - 3 , 5 l - 0, and B 2 —0. This

means that
—3x 3 - x _ 2x , -3 x
(x 2 + l) 2 (x 2 + l) 2 x 2 + l —
Many of the examples we encounter will offer a mixture of linear and
quadratic factors. For example,
2 ∣ 4 _7 3 A .x + B A2 Ai
- δr27---- 77rτ— ⅛⅛2 is decomposed as — 27 , . -I----- 7-∑∙ + -— — rp?.
(x + 4)(x + 3) x + 4 x + 3 (x + 3)2
The degree of the denominator is four, and there are four constants.
In algebra, the theory of equations tells us that any polynomial P with real
coefficients can be expressed as a product of linear and irreducible quadratic
powers, some of which may be repeated. This fact can be used to justify the
following general procedure for obtaining the partial-fraction decomposition of a
rational function.
Partial-Fraction Decomposition P(χ)

L e t ∕( x ) = where P(x) and D(x) have no common factors and Z)(x) ≠ 0.
of the Rational Function
P(x)∕D(x) Step 1. If the degree of P is greater than or equal to the degree of D, use long
P(x)
(or synthetic) division to express jA -τ as the sum of a polynomial and a
R(χ} n '
fraction i which the degree of the remainder polynomial R(x) is less
than the degree of the denominator polynomial D(x).
Step 2. Factor the denominator Z)(x) into the product of linear and irreducible
quadratic powers.
P(x)
Step 3. Express as a cascading sum of partial fractions of the form
A. a ∣x + B k
(x - r)n a n dj (χ 2 + ,sx + t)m
Verify that the number of constants used is identical to the degree of the
denominator.
■ INTEGRATING RATIONAL FUNCTIONS BY PARTIAL

FRACTION DECOMPOSITION
We will now apply the procedure of partial-fraction decomposition to integration.
EXAM PLE 4 Integrating a rational function with a repeated factor
Find f 'v — 6λ Λ3 3 dx.

J (x - 2)
Solution From Example 1, we have
[ ∙γ ~6-ξ+3 3 √A' = f[— 7 + ~2 + ~3 ^l dx

J (x~ 2 ) J Lx - 2 (x - 2)2 (x - 2)3J
= J (x - 2)^1√x - 2 J (x - 2)~2dx - 5 ∣(x - 2)~3dx
= l∏∣x - 2∣ + ξ ⅛i + 2 & ~ 2 )2 + C
EXAM PLE 5 Integrating a rational fraction with distinct linear factors
Find f -Γ 4 + 2A-; - 4x 2 + A - 3 d χ
J A2 - A - 2
Solution We have P(x) = A 4 + 2A 3 - 4A 2 + A - 3 and D(x) = A 2 - A - 2; because

the degree of P is higher than the degree of D, we carry out the long division
and write
A4 + 2A 3 - 4A 2 + A - 3 dx = I ( A 2 + 3A + 1 4— — L j cjx
A2 - A - 2 J ∖ A2 - A - 2/
The polynomial part is easy to integrate. The rational expression was
decomposed into partial fractions in Example 2.
A4 + 2A^ - 4A 2 + A - 3 _ ΓG + 3 χ + 1 + 8A - 1 μ χ
A2 - A -2 J ∖ A2 - A -2 ∕
= jI [yA 2 ÷ 3A + 1 4----
JV
^f 2,y JCι )∣dx
iy
= j~x dx + 3 ∣A dx + J√ A + 5 J ( A - 2)- 1 √A + 3 J ( A + l) - 1 tZx
2
= + ¾,- + A ÷ 5 ln∣A —2∣ + 3 ln∣A ÷ 1∣ ÷ C
EXAM PLE 6 Repeated quadratic factors
Find I ⅞⅛¾f λ
Solution From Example 3,
[ — τττ dx = f . ," v
dx + f , 3-γ c}x
J (x 2 + 1)2 J (A 2 + 1)2 J A2 + 1
2
Let u = x + 1;
du = 2x dx.
= I u 2 du - 21 u 'du
= -u ~ x - j In ∣w∣ + C
ln
= τ ⅛ 7 - i (χ 2 + i ) + c
- Λ , ∣ JL
EXAM PLE 7 Repeated linear factors
Find 5A 2 + 2 1A + 4
(A + 1)2(A - 3)
Solution The partial-fraction decomposition of the integrand is

5A 2 + 21A + 4 _ A + A2 A3
(A + 1 ) 2(A - 3 ) (A + 1 ) 2 A +1 A - 3
Multiply both sides by (A + 1)2(A - 3):
5A 2 + 21A + 4 = Λ 1(A - 3 ) + √42(A + 1)(A - 3 ) + A 3(A + 1)2
As in Example 1, we substitute x = -1 and x = 3 into both sides of the

equation to obtain A l = 3 and A 3 = 7, but Λ 2 cannot be obtained in this
fashion. To find A 2 , multiply out the polynomial on the right:
5x 2 + 21x + 4 = (∠4, + A 3)x 2 + (A l - 2A 2 + 2√43)x + (-3^4, - 3√42 + ∠43)
Equate the coefficients of x 2, x , and 1 (the constant term) on each side of the
equation:
5 = A2 + A3 X 2 terms
21 = A χ - 2,4, + 2√43 x terms

4 = -3√41 - 3∠4-, + ∕1 3 Constants
Because we already know that ∠41 = 3 and ∕1 3 = 7, we use the equation
5 = A 2 + Λ 3 to obtain A 2 = - 2 . We now turn to the integration:
f 5x 2 + 21x + 4 , f 3 dx , f - 2 dx + f 7 dx
J ( x + l ) 2( x - 3 ) J(x + 1 )2 J ∙ x + 1 J x ~ 3
= -3 ( x + l)^^l - 2 ln∣x + 1∣ + 7 ln∣x - 3∣ +
EXAM PLE 8 Distinct linear and quadratic factors
Find [ x 22 j^ ⅛
x ~ 23 dx.
J (x + 4)(x + 3)
Solution The partial-fraction decomposition of the integrand has the form

x 2 + 4x - 23 _ A Λ' + G ∣ , A
(x 2 + 4)(x + 3) x2 + 4 x + 3
Multiply both sides by (x 2 + 4)(x + 3) and then combine the terms on the
right:
x 2 + 4x - 23 = (A,x + 5 1)(x + 3 )+ A 2(x 2 + 4)
= (∠41 + A 7)X 2 + (3∕1∣ + B l )x + (35 l + 4√42)
Equate the coefficients to set up the following system of equations:
A 1 + √42 = 1 x 2 terms
{
3/11 + B i = 4
3B i + 4/1, = —23
x terms
Constants
Solve this system (the details are not shown) to find A i = 3, B i = - 5 , and
A 2 = - 2 . We now turn to the integration:
C x 2 + 4x - 23 J _ f 3x - 5 J + Γ - 2 dx
J (x 2 + 4)(x + 3 ) ’ J x2 + 4 J x + 3
_ , f x dx _ s [ dx _ 7 f dx
J x2 + 4 J x2 + 4 Jx + 3
↑
Let ιι = x 2 + 4;
du = 2x dx.
= 3 [∣ ln(x2 + 4)] - 5 ⅛tan^ 1 ⅛ ) - 2 ln∣x + 3∣ + C

■ SUBSTITUTION WITH PARTIAL FRACTIONS
The method of partial fractions can also be used in conjunction with substitution
and other methods of integration.
EXAM PLE 9 Substitution followed by partial-fraction decomposition
Fi n d [ 2/3 l v-5∕3∙
J A A
Solution In order to guarantee an integral with integer powers, we use the

substitution x = u 3 and dx = 3w2 dir.
f dx _ f 3u 2 du _ [ 3u 2 du _ [ _______ - 3 du_______
J x 2' 3 - x 5' 3 J u 2 - u 5 J w2(l - u 3) J (u - l)(u 2 + u + 1)
We use partial-fraction decomposition:
_3 √4∣ Z42M + B,
(μ — 1)(M2 + u + 1) u - 1 + u 2 + ιι + 1
—3 = A 1(w2 ÷ u + I) + (A∙,u + B l )(u — 1)

—3 = (/11 ÷ √4,)w2 + (∕1∣ - A 2 + B 1)u + (√41 —B 1)
By equating the coefficients, we obtain the system of equations
A χ + A, = 0 u2 terms
< A 1 —A 2 + B 1 = 0 u terms
k√41 —B t = —3 Constants
with the solution A i = - 1 , A 1 = I , and B l = 2.
We now continue with the integration:
_______ - 3 du_______
(w - 1)(M2 + u + 1)
u +I ψ ∣ By Pa r tial fractions
1^ +
----
u- 1 1
J M2- a+. M a+. ⅛ - ∖⅛, +t1 du
Let v = ιι +
dv = dιι.
-d u + f v dv , 3 [ dv Let w = v2 +
u - 1 J v2 + I 2 J v2 + I dw = 2v dv.
f - d u , 1 f dw , 3 f dv
J u- 1 2J w 2 J v 2 + ∕√3∖2
= -ln∣w - 1 ∣ + ∣ ln ∣ w ∣ + ∣ ∕ ' X λ∖tan 1 f⅛ ∖ + C

∖^2^∕ ∖ ~T∕
= - ln∣M - 1∣ + I ln∣v2 + ¾∣ + V3 tan 1 (—i j + C
= -ln ∣ι∕ — 1∣ + J ln∣u 2 + u + 1∣ + V3 tan -1 +
= - ln ∣ x ιz3 - 1∣ + ⅜ln∣x 2z3 + x 1'3 + 1∣ + √ 3 tan " 1 ( 2 * ¾ h~ ) + C
■ RATIONAL TRIGONOMETRIC INTEGRALS
The German mathematician Karl Weierstrass (1815-1897; see the historical note
on page 59) noticed that the substitution
u = tan ⅞ -ττ < x < π
will convert any rational function of sinx and cosx into an ordinary rational
function.
In order to show why this is a worthwhile substitution we use the double-angle
identities for sine and cosine (see Figure 7.3):
sinx = 2 siny cos^ and cosx = cos2 ^ - sin2 5
Figure 7.3 Weierstrass substitu
tion = 2( u
^∖V1 +
(
M2∕∖ V 1
r-
1
+ u 2∕
=f∖V1r- + u V
1
)
- fW rl + u )V
2
ll
2
_ 1u _ 1 - ιι2
1 + u2 1 + u2
Finally, because u = tan^> we have x = 2 tan -1 u, so dx = + du.
Weierstrass Substitution For -ττ < x < IT, let u = ta∏y so that
sinx = . 2» cosx = -∣— and dx = , ,d u

1 + u2 1 + u2 1 + u2
This is called the Weierstrass substitution.
The following example illustrates how this substitution can be used along with
partial fractions to integrate a rational trigonometric function.
E X A M P L E 10 Integrating a rational trigonometric function
________ dx________
3 cosx - 4 s in x ’
Solution Use the Weierstrass substitution, that is, let u = tan^.

Remember, dx = ∖^ 5; cosx = * ~ u ∙,∙, and sinx = , ~--5.
1 + u2 ’ 1 + u2 1 + u2
2 du
f ________ dx________ __ f __________ 1 + u 2_________ ,. ,./■
J 3 c o s x - 4 sinx J ∕∣ _ u 2∖ ( 2u ∖ c ιm P lfy∙
3 ∖1 + w2∕ ^^ ∖1 + w2∕
_ f - 2 du _ f - 2 du
J 3u2 + 8u - 3 J (3u - 1)(M + 3)
Sec. 7.3 The Method of Partial Fractions 46 3
This integral can be handled by the method of partial fractions.

_______ - 2 _ A i A 2
(3M - l)(w + 3) ^^ 3w - 1 + ιι + 3
Solve this to find A γ = - j and √t2 = ∣. We continue with the integration.
f ________ dx________ __ f - 2 du
J 3 cosx - 4 sinx J (3u — 1)(M + 3)
_ C - ∣ du + [ ∣ du
J 3ιι - 1 J u + 3
= - j∙⅛ l n ∣3 u ^^ 1 ∣ + 5 ∙ l n lw + 3∣ + c
= - ∣ ln∣3 tan J - 1∣ + ∣ ln∣tan^ + 3∣ + C
Once again, observe that when carrying out integration, you may obtain very
different forms for the result. For Example 10, you might use an integration table
(formula 393 in the Student Mathematics Handbook) to find
_________ du_________ _ 1 ∣
p sin aw + q cos an a∖∕p- + q 1 *
Let p = - 4 , q = 3, a = 1, so that
________ dx________ = l l n
3 cosx - 4 sinx 5
To illustrate further the point of alternative forms for the same integral, we
add still another form for the indefinite integral of Example 10. Computer
software programs often give alternative forms. For Example 10, we obtained
LN(3cos(x) + sin(x) + 3) LN(cos(x) - 3 sin(x) + 1)
5 5
Problem 64 asks you to derive the formula

secx dx = ln∣secx + tanx∣ + C
from scratch. You might recall that we derived this formula using an unusual
algebraic step in Example 3 of Section 7.1. You can now derive this by using a
Weierstrass substitution.
PROBLEM SET 7 .3
θ Write each rational function given in Problems 1-12 as a sum
2x 2 + x
x2 - x + 3
x 2(x - 1)
o f partial fractions.
ι 1 3x - 1 7 4x 3 + 4x 2 + x - 1 o x 2 —5x - 4
7
x(x - 3 ) x2 - 1 x 2(x + 1)2 • (x 2 + l ) ( x - 3 )
, 3x 2 + 2x - 1 2x 2 + 5x - 1 □ x 3 + 3x 2 + 3x - 4 1
4
χ (x + l) ’ χ (χ 2 - l ) • x 2(x + 3)2 x3 - 1
12.
x 2(x - 1)
Compute the integrals given in Problems 13-18. Notice that in 55. Find the area of the region bounded by the curve
each case, the integrand is a rational function decomposed into
partial fractions in Problems 1-6. 1
' 6 - 5x + x 2
13. IΓ dx [ ⅛ ~ 1⅛
14. 1, x 2 - 1
' x(x - 3) and the lines x = ⅜, x = 5, and y = 0.
2 2
15. IΓ 3x + 2x - 1 , 16. If 2x + 2 5x - 1 J 56. Find the volume of the solid generated when the curve
x ( x + l) x (x - l )
2 - x + 3
17. IΓ 4 dx 18. If, xx 2(x - 1)
y
' 2x 2 + x
is revolved about the y-axis.
Find the indicated integrals in Problems 19-34. 57. Find the volume (rounded to four decimal places) of the
3 4 2
19. 1Γ 2x 2 + 29x - 1 J 20. 1Γ x 2- x + 2 d solid generated when the curve
x (x - 1) x ( x -l) ∙
Γ *2 ÷ -L -^ f dx
21. 11 22. ' x 3 - 8
x2 + x - 2
is revolved about the x-axis.
23. i1 24. 1
x - 1 Xj - 1 58. Find the volume of the solid generated when the region
under the curve
25. I∣Γ ( xx+dx1)2 26. Γ 2x dx 2 1
∣ (x -2 )
⅛ ⅛ A V x 2 + 4x + 3
dx 28. ,
27. 1
x(x + 1)(x - 2) on the interval [0, 3] is revolved about the x-axis.
x dx [ x + ' ⅛ 59. Find the volume of the solid generated when the region
29. , 30. 1 z 22
(x + l)(x + 2)2 1 x(x + 2) under the curve
[ 5∖ + 7 .d x 4 —x ∖
31. 2 32. ∣ 5x dx ∕ = x2 4 + x)
∣ x 2 - 6x + 9
f 3x 2 - 2x + 4 , f 33 ∖ + 24 -v + ∖ √x on the interval [0, 4] is revolved about the x-axis.
33. ∣ x 3 - x 2 + 4x - 4 34. 1
∣ x 3 + 2x 2 + x - 2 60. Find the volume of the solid generated when the region
© 35. ■ What Does This Say? Describe the process of partial- under the curve
fraction decomposition. y = - ------ i -------
sιnx + cosx
Find the indicated integral in Problems 36-53. on the interval [θ, is revolved about the x-axis.
x
36. 1 cosx dx 37. I 2 x e dxx 61. S P Y P ROBLEM The spy manages to parachute safely into
' sin2 x - sinx - 2 2e - 5 e -2> the small village next to Boldfinger’s castle (see Problems
x
39. II sιnx<7x 37 in Section 6.4). He lands at night and is observed by
38. IΓ e2x dx
' e - 1 (1 + cosx) 2 two villagers, the Redselig twins, who promise to keep
2 silent. The spy cleverly disguises himself as an old duck
40. I1 tan2 x dx 41. IΓ sec x dx plucker, but despite his best efforts, word begins to
sec x + 4 tan x + 4
circulate among the 60 citizens of the village about the
42. IΓ dx 43. I, dx spy’s true identity. His keen mind determines that the
, x 1/4 - X χ l∕4 _ χ 5∕4
rate at which the rumor concerning his identity will
dx Γ
45. ∣ dx spread is jointly proportional to the product of the
44. 1
' sinx - cosx 3 cosx + 4 sinx number N of those who know already and the number
60 —N who do not, so that
Γ dx
46. I' 5 sin x + 4 47. IΓ sinx - cosx dχ
sinx + cosx
Γ dx = ≡ (6 0 - N )
48. IΓ dx 49. 1
4 cosx + 5 secx - tanx
for a constant k. He needs a week to get the information
50. 1f dx 51. ∣ dx
he requires, but he figures that as soon as 20 or more
' 3 sin x + 4 cosx + 5 4 sinx - 3 cosx —5
villagers know his identity, Boldfinger is sure to find out
f dx dx
52. I∣ 2 cscx - co tx + 2 53. I and he will be a dead duck plucker. If three people know
x(3 - In x)( 1 - In x)
his secret after one day, will he be able to complete his
54. Find the area under the curve
mission?
@ 62. Use partial fractions to derive the integration formula
x 2 + 5x + 4
~ A ~ . = ≠- ln ∣ ω ∣
+ C
between x = 0 and x = 3. a 2 - x 2 2a la - xl
Sec. 7.4 Summary of Integration Techniques 465
63. Use partial fractions to derive the integration formula 65. Derive the formula
dx _ 1 , ∣x ∣ J cscx dx = —ln∣
cscx + cotx∣+ C
x(ax + ⅛) b ∖ax + b ∖
using a Weierstrass substitution.
64. Derive the formula 66. Recall from Section 6.4 that a population P(∕) follows a
secx + tan x ∣+ C
J secλ' dx = ln∣ logistic growth pattern if the rate of population change is
proportional to the product P(A - P), where A is a
using a Weierstrass substitution. positive constant. Set up and solve an appropriate differ
ential equation forP. Assume that the initial population is
Λ)∙
7 .4 SUMMARY OF INTEGRATION TECHNIQUES
IN THIS SECTION Integration strategy

It is worthwhile to step back from our discussion of integration techniques and
to take some time summarizing what we have done. In this section, we present a
mixed bag of integration problems with no hints about how to carry out the
integration.
We conclude our study of integration techniques with a table summarizing

integration techniques.
TABLE 7.1 Integration Strategy
Step 1: Simplify. Simplify the integrand, if possible, and use one of the procedural rules (see the
inside back cover or the table of integrals).
Step 2: Use basic formulas. Use the basic integration formulas (1-29 in the integration table). These are the
fundamental building blocks for integration. Almost every integration will involve
some basic formulas somewhere in the process, which means that you should
memorize these forms.
Step 3: Substitute. Make any substitution that will transform the integral into one of the basic forms.
Step 4: Classify. Classify the integrand according to form in order to use a table of integrals. You
may need to use substitution to transform the integrand into a form contained in
the integration table. Some special types of substitution are contained in the
following list:
I. Integration by parts
A. Forms ∫ x ne ax dx, ∫ x"sinαx dx, ∫ x"cosαx dx
Let u = x n.
B. Forms ∫ x'1 In x dx, ∫ x n sin^ l ax dx, ∫ x n tan^ l ax dx
Let dv = x n dx.
C. Forms ∫ eax sin bx dx, ∫ e ax cos bx dx
Let dv = e ax dx.
II. Trigonometric functions
A. f sinm x cos"x dx
m odd: Let u = cosx.
n odd: Let u = sinx.
m and n both even: Use half-angle identities.
B. ∫ tan m x sec"x dx
m odd: Let u = secx.
n even: Let u = tanx.
m even, n odd: Write using powers of secant and use integration tables
(Formula 428).
C. For a rational trigonometric integral, try the Weierstrass substitution.

Let u - tan^’ so that smx - , 2 w √ cosx - J j ιz 2 , a n d d u ^^ 1 1 ι d x -
2 1 + u2 1 + u1 1+ M
III. Radical function; try a trigonometric substitution.
A. Form Vα 2 - u2: Let u = a sin θ.
B. Form Vα 2 + » 2: Let u = a tan θ.
C. Form V M2 - a2: Let u = a sec θ.
IV. Rational function; try partial-fraction decomposition.
Step 5: Try again. Still stuck?
1. Manipulate the integrand.
Multiply by 1 (clever choice of numerator or denominator).
Rationalize the numerator.
Rationalize the denominator.
2. Relate the problem to a previously worked problem.
3. Look at another table of integrals or consult computer software that does
integration.
4. Some integrals do not have simple antiderivatives, so all these methods
may fail. We will look at some of these forms later in the text.
5. If dealing with a definite integral, an approximation may suffice. It may be
appropriate to use a calculator, computer, or the techniques introduced in
Section 4.6.
EXAMPLE 1 Deciding on an integration procedure
Indicate a procedure to set up each integral. It is not necessary to carry out the
integration.
a. f s3n^Y^ dx b. f (1 + tan 2 θ) dθ c. f sin3 x cos2 x dx

J NX J J
2
d.. Jf 4xx2 cos3->x√Jx e. J f v ψx rdx
72 f∙e Jf x dx
g, f ∕x h f e 3xs ι n 2 χ ^ i. [ ∞ s4x d x
J x√x2 - 9 J J 1 - sιn2 x
Solution Keep in mind that there may be several correct approaches, and the way
you proceed from problem to solution is often a matter of personal preference.
However, in order to give you some practice, we will show several hints and
suggestions in the context of this example.
a. f dx The integrand is simplified and is not a fundamental type. Substitution will
J Vx work well for this problem if you let u = Vx. After you make this substitution
you can integrate using a basic formula.
b. J (1 + tan 2 θ) dθ The integrand can be simplified using a trigonometric identity (1 + tan 2 0 =
sec2 θ), so that it can now be integrated using a basic formula.
c. J sin3 x cos2 x dx The integrand is simplified, it is not a basic formula, and there does not seem
to be an easy substitution. Next, try to classify the integrand. It involves
powers of trigonometric functions of the type ∫ sin'"x cos"x dx, where m = 3
(odd) and n = 2 (even), so we make the substitution u = cosx.
Sec. 7.4 Summary of Integration Techniques 467
d. ∣4x 2 cos 3x dx The integrand is simplified (you can bring the 4 out in front of the integral, if
you wish), it is not a basic formula, and there is no easy substitution. It does
not seem to have an easy classification, so we check the table of integrals and
find it to be formula 313.
f x dx
J V9 - x 2 The integrand is simplified and is not a basic formula. However, it looks like
we might try the substitution u = 9 - x 2. It looks like this will work because
the degree of the numerator is one less (namely x). You might also find this in
a table of integrals (formula 225).
f x 2 dx
The integrand is simplified, it is not a basic formula, and it does not look as
J V 9 -x2 though a substitution will work because of the degree of the numerator. It also
does not seem to have an easy classification, so we turn to formula 226 of the
table of integrals.
[ dx
8' J x V X ~ τ) The integrand is simplified, and it is a basic formula for inverse secant
(formula 26).
h. J ei x sin 2x dx The integrand is simplified, is not a basic formula, and there does not seem to
be an easy formula. If we try to classify the integrand, we see that integration
by parts will work with u = sin2x and dv = e3x dx. We also find that this form
is formula 492 in the integration table.
f cos4 x dx The integration can be simplified by writing 1 - sin2 x = cos2 x. After doing
J 1 - sin 2 x
this substitution you will obtain
cos4 x dx f co s4 * ^ = f cos2 x dx
1 - sin 2 x J cos2 x J
This form can be integrated by using a half-angle identity or formula 317 in
the table of integrals. a aa
PROBLEM SET 7 .4
θ Find each integral in Problems 1-50. 26. I∣ xc'sin x dx 27∙ 1
∣ e - λ cos x dx
1. f 2x
~ 21 3 dx 2. [ j x + 3 dx
J (x - x 2)3 J V x 2 + 3x 28. 1 e2x sin 3x dx 29. I cos- 1 ( -x ) √ x
3. J (x sec 2x 2) dx 4. J (x 2 csc2 2x 3) dx
30. 1 sin^ 1 2xrfx 31. ι sin 3 x √ x
5. [ (ex co te x ) d x 6. ∫ ta n ^ ⅛
32. 1 cos5 x √ x 33. I∣ sin 3 x cos2 x dx

7. f ta n θn x '> 8. V c o tx csc2 x dx
J x
9. f 3 + 2.sιn t dt 1θ 2 + cosx dχ
34. ∣sin 3 x cos3 x dx 35. I sin 2 x cos4 x dx
J cost J sιnx
11. f X X 1 L 12. [ s in 2 x ^ v 13 f x2 + x + ⅛ 36. 1 sin 2 x cos5 x dx 37. 1 sin 5 x cos4 x dx
J 1+ eî l J 1 + sm 4 x J x2 + 9
14. [ 3x + 2
J √4 - x 2
dx 15. f P L≤ dx
J 1 - ex
16. [
J
e '~v x~ d x
Vx
38. 1 sin4 x cos2 x dx 39. 1 tan 5 x sec4 x dx
,-, f 2t2 dt ,q ∣
^ t3 dt 1 ft f dx 40. I tan 4 x sec4 x dx 41. 1∣
J V l-t 6 J 2 8 + t8 J l+ e 2x X
42. ∣
-*⅝
∏ (' dx ∣ dx '>'> ∣ dx dx Γ 2x + 3 ,
43. I1 ∕ - 52- dx
J 4- e x J x 2 + 2x + 2 Jx 2 + x + 4 ∣ V x 2 - 16 V2x - 1
23. ∣ ,∙4 ⅛ ----- 24. f , c lx — - 25. fta n ~ 1 x √ x 44.
f dx
Jx 2 + x+ l J x2 - x + 1 J I x√ x2 - 1
J tan m x sec"x dx
45. f
I X√ X 2 + 1 J
2
46. I∣ x V x + 1 dx
Γ (2x + 1) dx What substitution would you use in the case where m

a.
2
47. l √ 4 χ - χ 2 - 2 48∙ J ∣ V 3 + 4x - 4x dx and n are both odd?
b. What substitution would you use if n is even?
∖ cos x dx -n ∣ sec2 x<r∕x
4v. c. What would you do if n is odd and m is even?
I V I + sιn-χ J' Vsec 2 x - 2
Evaluate the definite integrals in Problems 51-64. under the curve y = cosx between x = 0 and x = J is
f2 fl Jv
revolved about the x-axis.
2 UΛ
51. ,1 √ 4 - χ √x 52. 1 2 83. Find the average value of secx on [θ, j ] .
o J ∣0 √9 - x
84. Find the average value (to the nearest hundredth) o f cscx
f1 dt on [1, 2].
JJ. ∣
l0 (x 2 + 2)3/2 54∙ j lθ 4z2 ÷ 4Z ÷ 5
Γ2 dx
55. I, x 4V x 2 + 3
-s r2 x3
j ∣ (3 + x 2, .),3z2
, , dx y =
'
, Vc -
, y = 0, x = 0, and x = 4.
VX2 + 9 j
0
86. What is the volume o f the solid obtained when the region
r2√3 bounded by y = V x e^ χ2, y = 0, x = 0, and x = 2 is re
57. 1 x 3V x 2 + 4 dx 58. 1I x 2V 5 - x2 dx
'-2 J0 volved about the x-axis?
rln 2 87. What is the volume of the solid obtained when the region
f, y √ + 9 ι f c
59. e zV 1 + e 2t dt 60. I bounded by y = x V 9 - x 2 and y = 0 is revolved about
0 J1 X the x-axis?
Γ7τ∕4 r∙7τ∕4 88. Find the length of the curve y = ln(secx) from x = 0 to
61. 1 sin 5 x<√x 62. 1 cos 6 x t ∕x
,o J0 x = f.
89. Find the centroid (to the nearest hundredth) of the region
Γ7 Γ ∕4 Γ7T∕3
63. I tan 4 x d x 64. ∣| sec 4 x<√x bounded by the curve y = x 2e~x and the x-axis, between
o J0 x = 0 and x = 1 revolved about the x-axis.
θ 90. Use integration by parts to verify that
Find each integral in Problems 65-77.
x
1Γ e dx 2x fi, ∣Γ dx . , , (a sin⅛x - b cos b x ) e a x
03. Vl + e ’ J
x √ 4 x 2 + 4x + 2 ∫e αχsιn
b x d x = ------------ 25---- --1-------------1- C
a + b
2
67. Ifx + 4x + 3 d χ 68. ∣f⅛ ⅛ ⅜ Λ (This is formula 492.)
i
' X + χ- + X J x 3 + 2x 2 91. Use integration by parts to verify that
2
69. Ir 5x + 18x + 34 dx 70. ∣f - 3 x 2 + 9x + 21 ,
(x - 7)(x + 2)2 J (x + 2 ) 2 (2X + 1) f aλ , , (a cos bx + b sin bx')ea x „
e cosbx dx = -------------25---- - 2,------— + C
J a + b
f ■ 3 V 5 -,- dx 72. 1f x 2 + 4x + 3 ^x
71. I 22
x + 2x + 1 J Xi + X1 + X (This is formula 493.)
[ 5 λ ' 2 + 3 -γ
~ - dx 74. ∣f ,6 x + 5
dx 92. Let f " be continuous on the closed interval [a, ⅛]. Use
73. ∣
x 3 + 2x 2 J 4x 2 + 4x + 1 integration by parts to show that
5X 2 - 4 X + 9 [ V '' W dx = b f(b ) - a f ∖a) - f(b) + ∕(α )
75. 1Γ __________x dx_________ IΓ
(x + l)(x + 2)(x + 3)
7 fi
Jx 3 - x 2 + 4x - 4
DΧ
93. Derive the integration formula

77. IΓ _________dx_________
γw + 1 ∩n χ∖n C m
θ
5 cosx - 12 sinx
78. Find the area (to the nearest hundredth) under the curve ∫
x'', (ln x)"dx
' ,
= -------'
m + 1
, -
--------- d
m + 1J
x v(ln x) w~l dx
’
y = sinx + cos2x on [0, 4]. where m and n are positive integers. (This is formula 510.)
79. Find the average value (to the nearest hundredth) of the Use the formula to find
function ∕( x ) = x sin 3 x 2 between x = 0 and x = 1. j x 2(ln x) 3 dx
80. An object moves along the x-axis in such a way that its
velocity at time t is v(t) = sin? + sin 2 Z cos3 t. Find the 94. Derive the integration formula
distance moved by the object between times t = 0 and
t = f∙ I sin M x dx = → n '- M x c o s Λ x + n^± f s i n ,1-2 λ χ dχ
J An n J
81. ■ What Does This Say? Integrals o f the general form
(This is formula 352.) Use this formula to evaluate
cot'"x csc"x dx J sin 3 4x dx
are handled in much the same way as those of the form
Sec. 7.5 Improper Integrals 469
7 .5 IMPROPER INTEGRALS
IN THIS SECTION Improper integrals with infinite limits of integration,

improper integrals with unbounded integrands
rb
We have defined the definite integral f(x ) dx on a closed, finite interval [a, b]
Ja
where the integrand ∕( x ) is bounded. In this section, we extend the concept of
integral to the case where the interval of integration is infinite and also to the
case where f is unbounded at a finite num ber of points on the interval of
integration. Collectively, these are called improper integrals.
■ IMPROPER INTEGRALS WITH INFINITE LIMITS OF

INTEGRATION
In physics, economics, probability and statistics, and other applied areas, it is

useful to have a concept of integral that is defined on the entire real line or on half
lines of the form x ≥ a or x ≤ a. If ∕( x ) ≥ 0, the integral of f on the interval
x ≥ a can be thought of as the area under the curve y = ∕( x ) on this unbounded
interval, as shown in Figure 7.4. A reasonable strategy for finding this area is first
to use a definite integral to compute the area from x = a to some finite number
Figure 7.4 Finding the area under x = N and then to let N approach infinity in the resulting expression. Here is a
a curve on an unbounded region definition.
rN
Improper Integrals (First Type) Let a be a fixed num ber and assume ∕( x ) dx exists for all N ≥ a. Then if
lim ∕( x ) dx exists, we define the improper integral

Λi→ +≈ Ja
∣
∙+≈ pιV
f(x) dx = lim fix ) dx
Ja ' N→+∞ Ja
^+∞
Ja
num ber and to diverge otherwise.
f( x ) dx is said to converge if this limit is a finite
EXAM PLE 1 Converging improper integral
dx
v2∙
Solution We draw the graph of y = l ∕x 2, as

shown in Figure 7.5. This region is un
bounded, so it might seem reasonable to
conclude that the area is also infinite.
Beginbycom putingtheintegralfrom 1tot.
t
Consider ' dx _ 1 z = -1t
X ∣I
Let us consider the situation more carefully, as shown in Figure 7.6. For t = 2,
3, 10, or 100, it appears that the area is not infinite, but rather is finite.
L e tr= 100, A =99/100.
Figure 7.6 Area enclosed by y = - γ the x-axis, y = 1, and y = t
Instead of picking a particular t, suppose we find the limit as t → ∞, as shown

in Figure 7.7.
Figure 7.7 The area bounded

by the curve y = l∕x 2, the line
x = 1, and the x-axis is finite.
Thus, the improper integral converges and has the value 1.
EXAM PLE 2 Diverging improper integral
Evaluate
Solution The graph of y = l∕x , shown in

Figure 7.8, looks very much like the graph
of y = l ∕x 2 in Figure 7.7. We compute the
integral from 1 to t and then let t go to
infinity.
Figure 7.8 The area bounded by

the curve y = l∕x, the line x = 1,
and the x-axis is infinite.
The limit is not a finite number, which means the im proper integral
diverges.
We have shown that the improper integral

converges and that the improper integral diverges.
In geometric terms, this says that the area to the right of x = 1 under the curve
y = l ∕x 2 is finite, whereas the corresponding area under the curve y = l/x is
infinite. The reason for the difference is that as x increases, l∕x 2 approaches zero
more quickly than does l/x, as shown in Figure 7.9.
EXAMPLE 3 Determining convergence of an improper integral
+
Figure 7.9 An unbounded region Show that the improper integral If °°dx converges only for p > 1.
may have either a finite or an infi
nite area.
What this says: The improper integral
+*dx ' 1 if p > 1
_
xp •p- 1
diverges if p ≤ 1
Solution We already know the integral diverges for p = 1 (Example 2), so we can
assume p ≠ 1.
f + ∞z√v clpγ Γ y~P+ > 1 l 1
≡ = lim lim x j ,
p = t→ = lim τ j - [ ZL l ^ - 11j
∣1 χP t →+∞ Jj x +∞L-p + 1J । z→+ ≈ l -p
If p > 1, then 1 - p < 0, so t l ~p → 0 as Z→ +∞, and
f "$ = ⅛ ( , " ' - 1] = τ ⅛ - ‘1=j⅛ f

If p < 1, then 1 - p > 0, so t'~p → +∞ as t → +∞, and
f + °5⅜p = lim τ~ 1~- [Z 1~', - 1] = +∞

∣1 x →
z + ≈ 1 - p v j
Thus, the improper integral diverges for p ≤ 1 and converges for p > 1. ≡ ≡
EXAMPLE 4 Improper integral for which technology may give an incorrect

result
Evaluate the given integrals.
√5 + 1
X c ,
χ 3.24 “
r - v Vz5 _i_ ∣ 1 ∣
Solution We note that v5 ≈ 2.236067 978; consider λ --,,ft = -I— ~-~z.
a. Because 3.236 - Vz5 < 1, the integral diverges.
b. Because 3.24 - V5 > 1, the integral converges:
(3.24 - U i - ■ ≈ 254322
c. Because 1 + Vz5 - V5 = 1, the integral diverges.

EXAM PLE 5 Improper integral using ΓH6pitaΓs rule
Evaluate xe 2x dx.
'o
1
4
E X A M P L E 6 T H IN K TA NK example: Gabriel’s horn, a solid with a finite

volume but infinite surface area
Gabriel’s horn is the name given to the solid formed by revolving about the x-
axis the unbounded region under the curve y = -~for x ≥ I . Show that this
solid has finite volume but infinite surface area.
Solution We will find the volume by using disks, as shown in Figure 7.10.
’ _ f, / ι ∖
(x ') 2 dx = ττ lim x 2 dx = τr lim - 7 + 1 = π
t→+∞ j ∣ t→+∞∖ ι /
A = 2 77 ∕( x ) V l + [ ∕ , (x)]2 dx = 2
Ji
= 2τ7 Γ ≡ ⅛
Let u = x2
Jι xi du - 2x dx
Figure 7.10 Gabriel’s horn.
You can fill Gabriel’s horn with a
finite amount of paint, but it takes
an infinite amount of paint to col
= 2 τr [
Jι 2ιd
t *dιι = 77 ι→+∞
lim f —
J∣ id
m
2
+ * du
or the inside of the horn!
- 77 lim
t→+∞ \_ u
+ 1- +
I
ιn L + √M2IJ+ 11∣^∣
= +∞ (The details are left for the reader.)
We can also define an improper integral on an interval that is unbounded to

the left or on the entire real number line.
Improper Integrals (First Type,

Extended) Let b be a fixed number and assume ∕(x ) dx exists for all t < b. Then if
rb Jt
liιn^ f(x ) dx exists, we define the improper integral
rb rb
f(x) dx = lim f(x ) dx
J-∞ t→-∞Jf
rb
The improper integral ∕(x ) dx is said to converge if this limit is a finite
number and to diverge otherwise. If both
-+ ∞ ra
Ja
f(x ) dx and
J-∞
∕(x ) dx
converge for some number a, the improper integral of∕(x ) on the entire x-axis
is defined by
r+∞ ra r+ ∞
f(x ) dx = ∕(x ) dx + f(x) dx
EXAM PLE 7 Improper integral to negative infinity
0
Find f ex dx.
J —x
f0 rθ
Solution ex dx = llimc ex dx = lim (1 - e t ) = 1
J -o o →~° Jt t→ -∞
■ IMPROPER INTEGRALS WITH UNBOUNDED INTEGRANDS
A function f is unbounded at c if it has arbitrarily large values near c. Geometrical

ly, this occurs when the line x = c is a vertical asymptote to the graph off at c, as
If f is unbounded at c and a < c < b, the Riemann integral £ f(x) dx is not even
defined (only bounded functions are Riemann integrable). However, it may still be
possible to define £ f(x) dx as an improper integral in certain cases.
Let us examine a specific problem. Suppose
∕ω = ⅛ for 0 < x ≤ 1
T hen /is unbounded at x = 0 and ∕ 0' ∕( x ) d x is not defined. However,
∕< τ) - ⅛
is continuous on every interval [t, 1] for t > 0, as shown in Figure 7.12. For any
such interval [t, 1] we have
f 1 dx _ Γ x -ιz 2 β(x = 2Vx∣' = 2 - 2 √ Z
'z √ x
I ∣,
If we let t → 0 through positive values, we see that
lim f ⅛ = lim (2 - 2√t) = 2
Figure 7.12 Graph of y = ⅛ '→θt Jι ∖ x <→0 *
This is called a convergent improper integral with value 2, and it seems reasonable
to say
ι f'-⅛ = 2
+ Jι V⅛
In this example, f is unbounded at the left endpoint of the interval of
integration, but similar reasoning would apply if it were unbounded at the right
endpoint or at an interior point. Here is a definition of this kind of improper
integral.
rb
Improper Integral If/ is unbounded at a and ∕( x ) dx exists for all t such that a < t ≤ b, then
(Second Type) Jt
rb rb
fix) dx = lim ∕( x ) dx
Ja t ~*a+ J∕
If the limit exists (as a finite number) we say that the improper integral
converges; otherwise, the improper integral diverges. Similarly, if/is unbound
ed at b and ∕( x ) dx exists for all t such that a ≤ t < b, then
J<7
fb Γ
f(x) dx = lim ∕( x ) dx
Ja /
<, C
and I ∕( x ) dx both converge, then

J
a
f(x) dx
f f(x) dx = f fix ) dx + f ∕( x ) dx
Ja Ja /
We say that the integral on the left diverges if either of the integrals on the right
diverges.
What this says: If a continuous function is unbounded at one of its
endpoints, then replace that endpoint by t, evaluate the integral, and then take
the limit as t approaches that endpoint. On the other hand, i f / is continuous
on the interval [a, b], except for some c in (a, b) at w hich/has an infinite
discontinuity, then rewrite the integral as
rZ? rc rb
fix ) dx = ∕( x ) dx + ∕( x ) dx
Ja Ja Jc
You will then need limits to evaluate each of the integrals on the right.
EXAM PLE 8 Improper integral at a right endpoint
l
Find Jfo ( ^ dx
- l ) 2z3'
Solution Let f(x ) = (x - I) -273 . This function is unbounded at the right endpoint
of the interval of integration and is continuous on [0, Z] for any t with
0 ≤ t < 1. We find that
fo ( x ~ Γ ) 2' 3 ∕⅛1- f 0 (x - 1 ) 273 1)1/3] |o
= 3 h m [ ( i - l ) 173- ( - l ) ] = 3
That is, the improper integral converges and has the value 3.
EXAM PLE 9 Improper integral at a left endpoint
Find secx dx.

Jττ∕2
Solution Because secx is unbounded at the left endpoint f of the interval of

integration and is continuous on [Z, τr] for any t with J < t ≤ ττ, we find that
77
secx dx = lim secx dx = lim ln∣secx + t tanx∣1

'ττ∕2 r→ (τr∕2) + f→ (τ∕2) t
- lim [ln ∣-l + 0∣ - ln∣secZ + tanZ∣] = -∞

f→ (π∕2) +
Thus, the integral diverges.
E X A M P L E 10 Improper integral at an interior point
Find f (x - 2) 1dx.
Jo
Solution The integral is improper because the integrand is unbounded at x = 2.
If the improper integral converges, we have
0 A common mistake is to fail
f3 /-2 <∙3
to notice the discontinuity at ( x - 2 ) ^ 1√ x = (X - 2 ) ^ 1√X + (X - 2 ) - 1 √X
x = 2. It is W R O N G to write ∣o Jo J2
f3 3
(x - 2)- l dx = lim f (x - 2) l dx + lim f (x —2) l dx
2
z→ ^ J o t →2 +
Jo 't
= ln∣x —2∣∣0 = In 1 —In 2 If either of these limits fails to exist, then the original integral diverges.
Because
= -In 2 r, _ μ
Notice that this mistake leads to lim I (x —2) γdx = lim ln∣x - 2∣∣θ
the conclusion that the integral
= lim [ln(2 - f) - In 2]
converges, which, as you can
see, is incorrect. 0 = —∞
we find that the original integral diverges.

PROBLEM SET 7 .5
0 1. I What Does This Say? What is an improper integral? c+∞
13. I 5e - 2 x dx 14. ∣ e 1 x dx
2. ■ What Does This Say? Discuss the different types of Jo '1
improper integrals. -> ⅛
In Problems 3-42, either show that the improper integral 15. ∣^+°° x 2 dx 16. ι--- F
J 1 (x 3 + 2)2 -— +
converges and find its value, or show that it diverges.
Γ+∞ J f + ” x 2 dx r+≈ 2
Γ 'd x
3. ,1, dx
X 2 4. ∣ , V~x
17.
J1 V x 3 +2
18. ι xe x dx
0
Γ+∞ 1 r + x e - ^ dx f, +∞
5. ι∣ dx 6∙
f +≈dx 19. 20. I xe~ x dx
0.99 1 Ji Vχ
χ
l1 V χ 0
r+∞ +5
7∙
∖+*, dx, 2/3
8. ιf x~ dx 21. I 5xe' 0 ~x dx 22. 1Γ ° In x dx
1, x 1∙' x
1 Jo 1
+ ∞ dx + r+
9. I r+”
dx 10. 1Γ 23. I,[ °°x dx 24. I “ dx
∣
3 2x - 1
L ⅛2x - 1 In x
2 x∖∕ln x
+ ∕, +∞ r+≈
n . Ifl ”( 2 xdx- l ) 2 12. ι e~x dx 25. 1 x 2e~x dx 26. 1 x 3e~ x2 dx
3 'o Jo to
f 0 2x dx Γ° x dx
27. 28. 48. Find all values of p for which X ≡ F converges, and
L χ2 + i I, (1 + x 2)2
find the value of the integral when it exists.
f0 dx f4 dx
29. . 30. ∣
L √2 - x L ( 5 - x) 2 r 1√jt-
49. Find all values of p for which — converges, and find
Jo x
∣ xe~W dx
31. 1 fx dx
32. .
L χ2 + ι the value of the integral when it exists.
f ‘ /2 c∕ x
f 1 dx f 4 dx 50. Find all values of p for which —∩----- converges, and
33. ι l∕5 34. I
0 -<√x
'o λv
find the value o f the integral when it exists.
f1 dx C+ ”3x dx 51. Discuss the calculation
35. Ilo ( l - x ) ,z2 36. I
L (3x2 + 2)3
r! .⅛ -v l' l - -I1 -(-oι- - 2
37. I∣ In x dx 38. I∣ In x dx
0 1 Is the calculation correct? Explain.
+ l 52. JOURNAL P R O B LE M .* Peter Lindstrom of North Lake
39. IΓ ∞ dx 40. IΓ x dx
'e x(ln x) 2 o 1 - x2 College in Irving, Texas, had a student who handled an
improper form as follows:
Γ+ " dx
41. I 42. ∣
l I (x - l)e x dx = f x — dx
0 ex + e~x
0 e
ι Ji
© 43. Find the area of the unbounded region between the x-axis
and the curve ΓHδpitaΓs rule
2 for x ≥ 6
τ (x - 4)3 1
e
44. Find the area of the unbounded region between the x-axis
and the curve What is wrong, if anything, with this student’s solution?
for x ≤ 2 f2
(x - 4)3 53. f( x ) dx, where
'o
45. The total amount of radioactive material present in the
' 1
for 0 ≤ x ≤ 1
atmosphere at time T is given by
^r ∕w = ∙ 1
o
r
Pe~ dtl
J .< ,(2 - x) 3
for 1 < x < 2
where P is a constant and t is the number of years. A © 54. T H IN K T A N K P ROBLEM Give an example of a region of
recent United Nations publication indicates that at the finite surface area that has an infinite volume— that is, it
present time r = 0.002 and P = 200 millirads. Estimate can be painted with a finite amount of paint, but holds an
the total future buildup of radioactive material in the infinite amount o f paint.
atmosphere if these values remain constant.
The Laplace transform o f the function f is defined by the
46. Suppose that an oil well produces P(t) thousand barrels of improper integral
crude oil per month according to the formula
F(s) = %{f(f)} = Γ e s t f(t) dt
0 2r
P(∕) = 100e^ ° - 100e^°∙ u
Jo
where t is the number of months the well has been in where s is a constant. This notation is used in Problems 55-60.
55. Show that for constant a (with 5 - a > 0):
production. What is the total amount of oil produced by
the oil well? a. T {e a , } = y ~ b. T{α} = “
for x ≥ 1
47. L e t∕( x ) = 'J c. ££{sin at} = —τ
for —1 < x < 1 s + αi
1 e λ+1 for x ≤ -1
Sketch the graph of f and evaluate d. T{cos at} = - r - ~ i
∕W dx
“College Mathematics Journal, 24, No. 4 (September 1993): 343.
56. Show that ⅛{af + bg} = a∙E{f} + b'E∖g} Evaluate this improper integral to find S{z} = 4j. From
57. Let T!~l {F(5)} denote the inverse Laplace transform of
F(s). That is, S Γ 1{.F(s)} = ∕( t) if and only if ££{/(/)} = F(s). this, and Problem 58, we know !E{te2l} =
(A
J J) 2.
Find (use Problems 55 and 56): a. -£{Fe 2z} b. H{e 3 z cos2z}
a. S!~l b. ⅛~' c. g ~ l ( , 5 2J
l(5 - 1) J
d. 2 ~ '[ -2— ----
ls + 45 + 51
4
c. ;2 —35 + Hint: Complete the square.
s 2(s - 1) 60. If F(s) = ^{∕(Z)}, show that
Hint. Use partial-fraction decomposition.
3
+ 2s2 - 35 - 17 %{tf(t)} = -F '(5)
d.
(52 + 4)(52 - 1) You may assume that
58. If F(s) = H{f(l)}, show that ⅛{ealf(t)} = F(s - a).
⅛ ∫ θ e - z∕( Z ) z ∕z = ∫ θ ⅛ - z∕(Z)zZz]
59. Use the result in Problem 58 to find the following Laplace
transforms and inverse transforms. For example,
SE{t} = e~st t dι (because ∕(Z) = z)
Jo
7 .6 THE HYPERBOLIC AND INVERSE HYPERBOLIC FUNCTIONS
IN THIS SECTION Hyperbolic functions, derivatives and integrals

involving hyperbolic functions, inverse hyperbolic functions
Hyperbolic and inverse hyperbolic functions are introduced and studied;
integration formulas to be used throughout the rest of the text are derived.
■ HYPERBOLIC FUNCTIONS
In physics, it is shown that a heavy, flexible cable (for example, a power line) that
is suspended between two points at the same height assumes the shape of a curve
called a catenary (see Figure 7.13), with an equation of the form
y = ∣(e'7α + e~x,a )
This is one of several important applications that involve combinations of

exponential functions. The goal of this section is to study such combinations and
their inverses.
Figure 7.13 The hanging cable In certain ways, the functions we shall study are analogous to the trigonomet
problem ric functions, and they have essentially the same relationship to the hyperbola that
the trigonometric functions have to the circle. For this reason, these functions are
called hyperbolic functions. Three basic hyperbolic functions are the hyperbolic
sine (denoted “sinhx” and pronounced “cinch”), the hyperbolic cosine (coshx;
pronounced “kosh”), and the hyperbolic tangent (tanh; pronounced “tansh”).
They are defined as follows.
Hyperbolic Functions
sinhx = — — for all x
cosh x = — ^l7 e for all x
1 ∙x = C 7 g -v
tanhx = s ^n ⅜ for all x
Sec. 7.6 The Hyperbolic and Inverse Hyperbolic Functions 479
Graphs of the three basic hyperbolic functions are shown in Figure 7.14.
a. The hyperbolic sine b. The hyperbolic cosine c. The hyperbolic tangent

y = sinh .r y = cosh x y = tanh x
Figure 7.14 G raphs o f the three basic hyperbolic functions
The list of properties in the following theorem shows how the basic hyperbolic
functions are similar to the trigonometric functions.
THEOREM 7.1 Properties of the basic hyperbolic functions

0 A major difference between cosh 2 x - sinh 2 x = 1
the trigonometric and hyperbol
s in h (-x ) = - s in h x (sinhx is odd)
ic functions is that the trigono
metric functions are periodic, c o sh (-x ) = coshx (coshx iseven)
but the hyperbolic functions are
ta n h ( - x ) = - ta n h x (tanhx is odd)
not. 0
sinh(x + y) = sinhx coshy + coshx sinhy
cosh(x + y) = coshx coshy + sinhx sinhy
Proof: We will verify the first identity and leave the others for the reader.
cosh 2 x - sinh 2 x = ---- ) - (----- ------J
_ e 2x + 2 + e~2x _ e 2x —2 + e~2x _
4 4
There are three additional hyperbolic functions, the hyperbolic cotangent

(cothx), the hyperbolic secant (sechxj, and the hyperbolic cosecant (cschx). These
functions are defined as follows:
coth x = -—⅛— = e x + e - x
tan h x ex - e x
sech x = — V— = γ ,~ -77
cosh x ex + e
csch x = -τ-⅛-— = x,. x

sm hx e - e
Two identities involving these functions are
sech 2 x = 1 - tanh 2 x and csch 2 x = coth 2 x - 1
You will be asked to verify these identities in Problem 58.
■ DERIVATIVES AND INTEGRALS INVOLVING HYPERBOLIC

FUNCTIONS
Rules for differentiating the hyperbolic functions are listed in Theorem 7.2.
THEOREM 7.2 Rules for differentiating hyperbolic functions
Let u be a differentiable function of x. Then:
^ (sin h u) = cosh u ^ (c o s h M) = sinh M ^
⅛ (tan h w ) = sech2 u g ⅛ c o th l 0 = -csch 2 i ∕ g
^(sech ι∕) = -sech ι∕ tanhw ^(csch ιι) = -csch u coth u
Proof: Each of these rules can be obtained by differentiating the exponential

functions that make up the appropriate hyperbolic function. For example, to
differentiate sinhx, we use the definition of sinhx:
⅛ s ιn h x ) = cl ex - e x
dx 2
= ⅛ex - ⅛e λ'(- 1) = + e λ') = coshx
The proofs for the other derivatives can be handled similarly.
EXAM PLE 1 Derivatives involving hyperbolic functions
Find for each of the following functions:
a. y = cosh Ax (A is constant)
b. y = tanh(x 2 + 1)
c. y = ln(sinhx)
Solution a. We have
z7 /7
^∙-(coshAx) = smh (Ax) ^(A x) = A sinh Ax
b. We find that
^ [ta n h (x 2 + 1)] = sech2 (x2 + 1) ~y^C*2 + 1) = 2x sech2 (x2 + 1)
c. Using the chain rule, with u = sinhx, we obtain
⅛ tln (s in M l = sT⅛7- ⅛ s in h χ ) =≡ < c o sh χ ) = c o th x .
Each basic differentiating formula for hyperbolic functions corresponds to a

basic integration formula. These formulas are listed in the following theorem.
Sec. 7.6 The Hyperbolic and Inverse Hyperbolic Functions 48 1
T H E O R E M 7.3 Integration formulas involving the hyperbolic functions
sinhx dx = coshx + C J coshx dx = sinhx + C
sech2 x dx = tanhx + C ∣ csch2 x dx = -c o th x + C
sechx tanhx dx = -s e c h x + C ∣~cschx cothx dx = -c s c h x + C
Proof: The proof of each of these formulas follows directly from the corre
sponding derivative formula. ■
EXAM PLE 2 Integration involving hyperbolic forms
Find each of the following integrals:

a. J cosh3 x sinhx dx b. J x sech2 (x 2) dx c. J tanhx √x
f 7√4 1
∫ cosh x sinhx dx =
3 w3 du = + C = ∣ cosh4 x + C
Let u = coshx;
dιι = sinh x dx.
'1∙ | x sech2 (x 2) dx = sech2 (x 2) (x dx)
Let u = x 2;
du = 2x dx.
= | sech2 w(∣ dll)
= ∣ tanh u + C
= I tanhx 2 + C
c. [ t a n h x r ∕x = [ s in h
^ x
J J coshx
Let u = coshx;
du = sinh x dx.
= J = ln∣M∣ + C = ln(coshx) + C
The graph o f y = sinh - , x is obtained
by reflecting y = sinh x in the line
y = x.
■ INVERSE HYPERBOLIC FUNCTIONS
Inverse hyperbolic functions are also of interest, mainly because they enable us to
express certain integrals in simple terms.
Because sinhx is continuous and strictly monotonic (increasing), it is one-to-
one and has an inverse, which is defined by
y = sinh - 1 x if and only if x = sinhy
for all x and y. This is called the inverse hyperbolic sine function, and its graph is
obtained by reflecting the graph o fy = sinhx in the line y = x, as shown in Figure
Figure 7.15 The graph of 7.15. Other inverse hyperbolic functions are defined in a similar manner (see
y = sinh-1 x Problem 60).
Because the hyperbolic functions are defined in terms of exponential func

tions, we may expect to be able to express the inverse hyperbolic functions in
terms of logarithmic functions. We summarize this relationship in the following
theorem.
T H E O R E M 7.4 Logarithmic formulas for inverse hyperbolic functions
sinh- 1 x = ln(χ + V x 2 + 1), all x csch^ 1x = lnθ^ -l----- x ) , x ≠ 0
cosh~1 x = ln(x + V x 2 - 1), x ≥ 1 sech^1x = ln ^ + Yj ~χ2), 0 < x ≤ 1
tanh - 1 x = I In j +x > ∣x∣ < 1 coth~1x = ∣ In + ∣∙> ∣x∣ > 1
Proof: We will prove the first part and leave the next two parts for youto
verify (see Problem 61). The other parts are proved similarly. Let y = sinh^^1xj
then its inverse is
x = sinhy
= ∣(e v - e~y )
2x = ey - j - y
e 2y - 2xey - 1 = 0
_ 2x ± ^√4x2 + 4 Quadratic formula with
ey 2 a = 1, b = -2 x, c = -1
= x ± Vx2 + 1
Because ey > 0 for all y , the only solution is ey = x + V x 2 + 1, and from the
definition of logarithms,
y - ln(x + V x 2 + 1) ■
Differentiation and integration formulas involving inverse hyperbolic func

tions are listed in Theorem 7.5.
T H E O R E M 7.5 Differentiation and integration formulas involving the

inverse hyperbolic functions
[ . /------- 7 - sinh 1u + C
-τ~(sinh^1 ») = 1 . Φ
axy ’ Vl + M2 dx J V l + u2
1 it
X<cosh ' u > v f → d i∙ l f ∣" ∣ > 1 [ . r -^— - - cosh + C
J ∖ u1 - 1
1 M) 1u
⅛ ta n h = 1 , 1 M2 ⅛ ι f ∣ w∣ < 1 f , 1 = tanh + C , if lw∣ < 1
J 1 - u2 > ∣∣
csch 1 1"∣+ c
X<csch , "> ^ H √ f ⅛ Λ ∫ 11√ τ⅛
sech , 0 < a < 1 [ . γ .------ 5 - -sech 1u + C

⅛ " " ∙≈ a √ Γ ⅛ Λ
J u∖∕ 1 - u~
X ( c °th 1M) = 1 -1 M 2 ⅛ L F W∣ ∣> 1 [ = coth- , u + C, if IM∣ > 1
1- u1 > ∣∣
Sec. 7.6 The Hyperbolic and Inverse Hyperbolic Functions 483
Historical Note Proof: The derivative of each inverse hyperbolic function can be found either
by differentiating the appropriate logarithmic function or by using the definition
in terms of hyperbolic functions. We will show how this is done for y = sinh~1 x,
and then will leave it for you to apply the chain rule for u, a differentiable function
of x. By definition of inverse, we see that
x = sinh y and ^ = c°sh T

Thus,
⅛ n h - 'x ) - ⅛ - ^ = 1
1 _ 1
d x 's ' d x d x coshy V 1 + sinh2 y V l + x 2
dy
because cosh2 y = sinh2 y + 1 and sinhy = x.
The integration formula that corresponds to this differentiation formula is
dx = sinh 1x + C
PIERRE DE F ERMAT
(1601-1665) V l + x2
A lawyer by profession, Fermat The other differentiation and integration formulas follow similarly.
liked to do mathematics in his
spare time. He wrote well over
3,000 mathematical papers and EXAM PLE 3 Derivatives involving inverse hyperbolic functions
notes. In Chapter 2 we saw how
his work served to inspire Isaac Find for a. y = sinh , (ax + b), and b. y = cosh l (secx), 0 ≤ x ≤ f .
Newton. In addition to his in
novative work with tangents, he a x + b>
obtained the first procedure for Solution a. ⅛ [s in h - (ax + ⅛)] = y ∕ l + ^ + b y ⅛
differentiating polynomials.
However, Fermat’s real love a
was in the area of number
_
V l + (ax + ⅛)2
theory. His most famous prob
lem has come to be known as b. ⅛ cosh-> (secx)] = 1
vsec2 χ _ 1 ⅛(secx)
Fermat’s last theorem. He
wrote in the margin of a book:
“To divide a cube into two = s y x ta n x
cubes, a fourth power, or in gen V tan 2 x
eral, any power whatever above
the second, into powers of the Note that the result of Example 3b implies
same denominations, is impos
secx dx = cosh 1 (secx) + C
sible, and I have assuredly
found an admirable proof of = ln(secx + V sec 2 x - 1) + C
this, but the margin is too nar
row to contain it.” = ln(secx + tanx) + C
In 1993, this problem was re
ported solved, and the proof Notice that we do not need absolute values because 0 ≤ x ≤ J . We will use this
looks promising, but to date formula in the next chapter.
has not been verified.
EXAM PLE 4 Integral involving an inverse hyperbolic function
f 1 dx
Evaluate
∣0 V l + x 2
Γ1 dx
Solution = [sinh 1 x]∣ 0
Jo V l + x 2
= ln(x + V x 2 + l)∣θ
= ln(l + V2) - ln(l)

= ln(l + V2)
PROBLEM SET 7 .6
θ Evaluate (correct to four decimal places) the indicated hyper 50. Sketch the graph of y = cothx. Be sure to show key
bolic or inverse hyperbolic function in Problems 1-12. features such as intercepts, high and low points, and
1. sinh 2 2. cosh 3 3. tanh ( - 1) points of inflection.
4. sinh -1 0 5. coth 1.2 6. tanh -1 0 51. First show that coshx + sinhx = ex , and then use this
7. cosh-1 1.5 8. sinh(ln2) 9. cosh (In 3) result to prove that
-1
10. sech 0.2 11. sech 1 12. coth -1 (-3 ) (sinhx + coshx)" = cosh ux + sinhwx
Find ~τ~ in Problems 13-29. for positive integers n.
dx
13. y = sinh 3x 14. y = cosh (1 - 2x 2) 52. If x = a cosh t and y = b sinh t where a, b are positive
2
15. y = cosh (2x + 3x) 16. y = sinh V x constants and t is any number, show that
17. y = sin h - 18. y = cosh -1 (x 2) x f_ f = 1

a 2 b2
19. y = sinh -1 (x 3) 20. y = x tanh -1 (3x)
53. a. Verify that y = a cosh ex + b sinh ex satisfies the dif
21. y = sinh -1 (tanx) 22. y = cosh -1 (secx)
ferential equation
23. y = tanh -1 (sinx) 24. y = sech ( )
y " - c 2y = G
∖1 + χ∕
_ sinh 1 x b. Use part a to find a solution of the differential equa
25. 26. y = sinh -1 x - V l + χ∙ tion
cosh - 1 x
27. y = x cosh - 1 x - V x 2 - 1 y" -4 y =0
28. x coshy = y sinhx + 5
subject to the initial conditions y(0) = 1, y '(0) = 2
29. e vsinh - 1 x + e - -vcosh - 1 ι- = 1
54. Find the area of the region bounded by the curves
Compute the integrals in Problems 30-45. y = sinhx and y = coshx, the v-axis, and the line x = 1.
30. ∣
x cosh(l - x 2)dx 31. I sin h ^ Λ ' 55. Find the length of the catenary
J J x2
, , fsech 2 (In x) dx ,, f y = a co sh -
32. — -— 33. cothx⅛ c a
between x = - a and x = a.
-, Λ I _____dx_____ Γ__ dt
J √ 4 x 2 + 16 J V 9r 2 - 16 56. Find the volume of the solid formed by revolving the
region bounded by the curve y = tanhx on the interval
Λ6 f 17 f c o s x dχ [0, 1] about the x-axis.
J 36 - 16 ∕ 2 J V 1 + sin 2 x
57. Find the surface area of the solid generated by revolving
38. [ r d x . 39. [ χ 2 d x , the curve y = coshx on the interval [ - 1 , 1] about the
J V l + x4 J 1- x 6 X-axis.
© 58. Prove the formulas
r , ' 2 dx r 3 dx
40. 41. — a. sech2 x + tanh 2 x = 1 b. coth 2 x - csch2 x = 1
Jo 1 - x J2 1“ x
59. Derive the differentiation formulas for cosh u, tanh u, and
Γ1 t 5 dt Γ2 ex dx sech u, where w is a differentiable function of x.
42. ,∣ √ l + z>2 43. l V e 2 x - 1
0 l 60. First give definitions for cosh-1 x , tanh -1 x , and sech- 1 x,
pin 2 r1 and then derive the differentiation formulas for these
44. ∣ sinh 3x dx 45. , x sech2 x 2 dx functions by using these definitions.
0 0
61. Prove
θ 46. Show that tanh (x + y) = ⅛ + tanhy
v j '
a. cosh -1 x = ln(x + V x 2 - 1)
1 + tanh x tan v
47. Show that b. tanh - 1x = ∣ In } + x
λ 1- x
a. sinh2x = 2 sinhx coshx 62. Show that
b. cosh2x = cosh2 x + sinh2 x flf-x' = s inh - 1 -
48. Show that V x2 + V a
a. -1 < tanhx < 1 for all x for constant a > 0. State and prove a similar formula for
b. lim tanhx = 1 and lim tanhx = -1
dx
49. Determine where the graph of y = tanhx is rising and
Xa2 - x2
falling and where it is concave up and concave down.
Sketch the graph and compare with Figure 7.14c. with α > 0 and ∣x∣ < a.
Chapter 7 Review
IN TH IS R E V IE W Proficiency examination; supplementary problems

Concept Problems f dx
17. ∣
∣ sinh (1 - 2x) dx 18. j
V⅞ —x 2
1. Discuss the method o f integration by substitution for
Γ x 2 dx 3
both indefinite and definite integrals. 19∙ 20. j, x2 dx
2. What is the formula for integration by parts?
1(x 2 + 1)(x - 1) x - 1
3. What is a reduction integration formula? Evaluate the definite integrals in Problems 21-24.
4. When should you consider a trigonometric substitution? f2
21. x In x 3 dx 'y,y Γ _______ dx_______
5. When should you consider the method of partial frac J 2 (x - l) 2(x + 2)
Ji
tions?
*-zr∕4
6. What are the substitutions to use when integrating
rational trigonometric integrals?
f 45 dx
∣
3 2x —x
1
Jo
sec3 x tan x dx
7. Outline a strategy for integration.

Determine whether each improper integral in Problems
8. What is an improper integral?
25-28 converges, and i f it does, fin d its value.
9. Define the hyperbolic sine, hyperbolic cosine, and hy
perbolic tangent. 25. Γ ° x e ~ 2xd x 26. [ * c o s x dx
Jo Jo 1 “ cosx
10. State the rules for differentiating the hyperbolic func
tions. 1
f 1y -L 7 f +∞
11. State the rules for integrating the hyperbolic functions. 27. — ------ — dx 28. e -'sin x dx
Jo x 2 (x - 2) Jo
12. What are the logarithmic formulas for the inverse hyper
29. Find ⅜ for y = V t a n h 1 2x.
bolic formulas? z
dx
13. State the differentiation and integration formulas for
the inverse hyperbolic functions.
under the curve
Practice Problems
14. Evaluate each of the given numbers. y = J
’ V9 - X2
a. tanh - l (0.5) b. sinh (In 3) c. coth - l (2)
on the interval [0, 2] is revolved about the y-axis.
Find the integrals in Problem 15-20.
7x + 1 I"
∫
i 21
Vx + 1
dx 16. x sin 2x dx
J
Find the derivative in Problems 1-5. Find the integrals in Problems 6-35.
dx 2
1. y = tanh - 1 ^ 2. y = x cosh - 1 (3x + 1) cos 1 x dx 7. Ilf x dx 2
√ 4 -x
3. y = s ln h x 4. y = sinhx coshx 8.
∣ x dx 9. jf 3 x - 2 dx
z ex 1x 2 - 2x + 5 x 3 —2x 2
5. y = x sinh % + sinh(e ' + e~x )
10. 11. 44. Compute x ∕" ( 3 x ) dx, given that ∕'( 0 ) is defined,
(x 2 + x + l) 3z2 V x ( l + 1vzx) Jo
/(0) = 1 ,/ ( 3 ) = 4 , a n d ∕'(3 ) = - 2 .
1V 9 x 2 - 1 dx
12∙ 1 13. 1x 2 tan - 1 x dx
(, π∕2
X
45. What is [∕( x ) + ∕"(x )]c o s x dx if /(0) is defined,
Jo
14∙ I∖ dx 15. 11e vV 4 - e 2x dx
∣ sinx + tanx ∕( f ) = 5 a n d ∕' ( O ) = -1 7
17∙ 46. S P Y P ROBLEM After plucking the spy from the village
16. 1∣ c o s j sin z∣ dx 1/5 + ]∕
(Problem 61, Section 7.3), Boldfinger has him thrown
into the Closet of Doom in the confines of his castle
18. 1sin x 5dx 19∙ 1
1V 1 + sin x dx dungeon. He is tied to a chain inside the Closet and
∣ cos x
poison gas is slowly pumped in at the rate of 0.2 ft 3∕m in.
He quickly determines that the Closet has volume 400 ft3
20. 1cosx ln(sinx) dx 21. 1sin (In x) dx
and that a mixture o f gas and fresh air is continuously
Γ sinhx dx escaping from the Closet at the same rate the gas is being
22. e2x sech (e2λ) dx 23. 1 pumped in. He knows that the gas becomes deadly when
∣ 2 + coshx
the air in the Closet contains at least 0.8 ft 3 of poison gas.
1 x dx
24. ιtanh^
∣ 1- x 2
25. x 2cot^ 1 x dx If he takes 3 min to get untied and another 45 sec to kick
open the door, will our hero survive to reach Chapter 8?
f ∕ 2±2 - d x 47. A prime number is a positive number n with exactly two
26. Ix (l + x ) ιz3 dx 27. I
∣ x 3 + 6x + 1 divisors. In number theory, it is shown that the total
Γ sin x - cosx c∣χ number o f primes less than or equal to x is approximated
28. 29∙ II cos (V x + 2) dx by the function
I (sinx + cosx) 1/4
f χ 3 + 2λ' dx
30. ι1√5 + 2 sin 2 x sin 2x dx 31. 1x 4 + 4x 2 + 3
f x dx V 5 - x 2 dx Use the trapezoidal rule to estimate the number of primes

32. 1√5 - x 2 33. 1
, X between x = 2 and x = 1,000.
2 48. Find the centroid o f a thin plate of constant density δ
34. 1V x + x dx 35. ∣
x 3(x 2 + 4)-ιz 2 dx that occupies the region bounded by the graph of
y = sin x + cosx, the coordinate axes, and the line
x = π∕4.
Determine whether the improper integral in Problems 36-39 49. Repeat Problem 48 for a plate that occupies the region
converges, and i f it does, fin d its value. bounded by y = sec2 x, the coordinate axes, and the line
πz2 x = τr∕3.
36. ∣ cosx dx τf -j f + °° dx
' L 4 + x2 50. Find the volume obtained by revolving about the y-axis
Jo V s in x
the region bounded by the curve y = sinhx, the x-axis,
and the line x = 1.
38.
x 2 + 4x + 6 51. Find the volume obtained by revolving about the y-axis
the region bounded by the curve y = coshx, the v-axis,
40. If n is a positive integer and a is a positive number, find
and the line y = 2.
52. The region bounded by the graph of y = sinx and the x-
axis between x = 0 and x = ττ is revolved about the x-axis
to generate a solid. Find the volume of the solid and its
41. Find s i⅞
surface area.
J 1 + COS2 X
53. Find the length of the arc of the curve y = ∣ x 5z4 that lies
-l ∕∙l
o J
x'"(l —x) ndx =
Jo
x' 1(l —x) md x for positive
54.
between x = 0 and x = 1.
Find the volume of the solid obtained by revolving about
integers m, n. the x-axis, the region bounded by the curve
43. T H IN K T A N K P RO BLEM Each of the following equations y = (9 - x 2) ιz4 and the x-axis.
may be either true or false. In each case, either show that 55. Find the volume of the solid formed by revolving about
the equation is generally true or find a number x for the y-axis the region bounded by the curve
which it fails.
a. tanh (I In x) = x ?x > 0
- ' x + 1
b. sinh 1 (tanx) = tanh 1 (sinx) for - J < x < J between x = 0 and x = 4.
56. Find the volume o f the solid formed by revolving about

the x-axis the region bounded by the curve gested here. I f your computer can compute Γ(7.3)
explicitly, compare with your approximate result.
_ 2
y √ 3 χ -2 64. Stirling’s formula This problem requires Problem
62 and a rapid evaluation o f Γ(x), see Problem 63.
You have probably noticed that n∖ grows incredibly
57. Find the surface area o f the solid generated by revolving fast (that is, faster than exponentials, even e" ). This
the region bounded by the curve
problem investigates this fact. It has been shown that
y = e x + {e^ jc Γ(x + 1) ≈ C x x + l l 2 e~ x
on [0, 1] about the x-axis. for large x; so the major growth factor is x λ'. You are
58. Find the volume o f the solid whose base is the region R to find the constant C .
bounded by the curve y = e x and the lines y = 0, x = 0, a. By plotting, for large x , the ratio o f Γ(x + 1) and
x = 1, i f cross sections perpendicular to the x-axis are x + ',2 e^ x , p j
x ιn a nu r n e r j c a ι estimate o f C .
equilateral triangles. b. Given that C is the square root o f a familiar
59. Show that the area under y = 4 on the interval [1, a] number, find C . W hen you are finished you will
equals the area under the same curve on [k, ka] for any have Stirling’s formula.
number k > 0. c. Stirling’s formula is called an asymptotic result;
60. Find the length o f the curve y = ln(secx) on the interval that is, it becomes more accurate as x → +∞. To
[θ j]∙ see this (provided your computer can handle very
large numbers), approximate n∖ for n = 10 and
61. Find the length o f the curve
n = 100, and calculate the relative error in each
x case. Hint-. The ratio o f the two values works well
b2 8a n b
for this calculation o f relative error.
between y = b and y = 2b.
62. If 5 > 0, the integral function defined by 65. Derive the integration formula
f+∞
f , 2, , x(α 2 - x 2)" 2a 2 n f , ,, , ,
Jo
e~ l ts ~ 1 dt
is called the gam m a function. It was introduced by

J
(α - x 2)''r∕x = — 5— —
’ 2n + 1
1- 5— — v(α2 - x 2) n 'd x
2M + 1 J
Use this formula to find J ( 9 — x 2) 5z2tfx
’
Leonhard Euler in 1729 and has some useful properties,

a. Show that Γ(.sj converges for all s > 0. 66. Derive the integration formula
b. Show that Γ(s + 1 ) = .sΓ(s,). f sin''x d x _ 1 sin''^ 1 x _ n - 1 f sin"^ 2 x d x
J co s m x m - 1 cos" 1~1 x m - 1 J cos m ^2 x
c. Show that Γ(n + 1) = n! for any positive integer n.
67. Derive an integration formula for J °θj n mχ x ∙
68. Derive an integration formula for l^x"(x2 + a 2)~ m dx.
63. This problem is a continuation o f Problem 62. 69. Derive the integration formula
a. Show that for a positive integer n:
f dx _ - N , q χ + b _ (2« - 3)a (^ dx
Γ(x + n) = (x + n - l)∙∙∙(x + l)xΓ(x) J x " V a x + b (n - l)b x n ~ , (2n - 2)b J χ n ~ , V a x + b
b. Suppose we need to approximate Γ(x) by
numerically integrating over an inteval (0, T ), for f + °T 2,4 Y 3 1 1
70. The improper integral ---- ------—γ d x converges
T quite large. O f course, you want T to be no J∣ Lx4 + 1 x + 1J
larger than necessary. For 0 < x ≤ 1, find T so for exactly one value o f the constant A . Find A and then
that *∞
compute the value o f the integral.
J τ
e^ , tx ~' dt < 0.005
Thus, your approximation to Γ(x) by integrating

71. Evaluate
∣0 (x + l)(x 2 + x + 1)
if it exists.
over (0, T ) is accurate to within 0.005 units. Note. 72. P UTNAM EXAM INATION P ROBLEM Show that
in general, t*^1 ≤ 1. 22 _ f 1x 4(l ^^ x )4 d x
ι-t
π I ^
c. We will now use part a. Suppose you need to 7 Jo 1+ x2
approximate Γ(7.3). Explain how to use part a to 73. P UTNAM EXAM INATION P ROBLEM Evaluate
numerically compute Γ(0.3), and thus Γ(7.3), and dx
be sure the resulting relative error is less than 1 + (tan x) 3/2
0.005. 74. P UTNAM EXAM INATION P ROBLEM Evaluate
d. Explain why you would rather computer Γ(0.3) 1/2
thanΓ(7.3). t e x p [ - 1985(/ + ∣)] dt
Jo ' '
e. Numerically compute Γ(7.3) by the method sug Γx 2 /—
You may assume that e~ x dx = v τ τ .
"Buoy Design”
You have been hired as a special consultant by the U.S. Coast Guard to evaluate
some proposed new designs for navigational aids (buoys).
The buoys are floating cans that need to be visible from some distance away,
without rising too far out of the water. Each buoy has a circular cross section and
will be fitted with a superstructure that carries equipment such as lights and
batteries.
To be acceptable, a fully equipped buoy must float with not less than 1.5 ft nor
more than 3 ft of freeboard. (Freeboard is the distance from the water level to the
top of the device.)
The book o f nature is written in
the language o f mathematics.
Galileo
a. A floating object will displace a volume of water whose weight equals the
weight of the floating object. (This is Archimedes’ principle.)
b. Your devices will be floating in salt water, which weighs 65.5500 lb∕ft3.
c. The buoys will be constructed of ⅛in. thick sheet metal that weighs 490
The pseudomath is a person who lb∕ft 3.
handles mathematics as a monkey d. You can estimate an additional 20% in weight attributable to welds, bolts,
handles the razor. The creature and the like.
tried to shave himself as he had e. Each buoy will be fitted with a superstructure and equipment weighing a
seen his master do; but, not having
total of 2,000 lb.
any notion o f the angle at which
the razor was to be held, he cut his f. Each design you are given to evaluate will be presented to you in the form
own throat. He never tried it a sec of a curve x = f(y), to be revolved about the y-axis.
ond time, poor animal! But the
pseudomath keeps on in his work,
proclaims himself clean shaved, concerns: You are to design your evaluation procedure using a top-down struc
and all the rest o f the world hairy. tured approach. The design should take the form of an outline of the procedure
A. De Morgan used, with headings and subheadings as appropriate.
Budget o f Paradoxes (London,
1872), p. 473. *This project is adapted from a computer project used at the U.S. Coast Guard Academy.
Infinite Series
8.1 Sequences and Their Limits
Sequences; The limit of a Is it possible for the sum of infinitely many numbers to be finite? This
sequence; Bounded, monotonic
sequences concept, which may seem paradoxical at first, plays a central role in
8.2 Introduction to Infinite Series: mathematics and has a variety of important applications. The goal of
Geometric Series this chapter is to examine the theory and applications of infinite sums,
Definition of infinite series; which shall be referred to as infinite series. Geometric series, introduced
General properties of infinite
series; Geometric series; in Section 8.2, are among the simplest infinite series we shall encounter
Applications of geometric series and, in some ways, the most important. In Sections 8.3-8.6, we shall
8.3 The Integral Test: P-series develop convergence tests, which provide ways of determining quickly
Divergence test; Series of non
negative numbers; The integral whether or not certain infinite series have a finite sum. Next, we shall
test; p-series turn our attention to series in which the individual terms are functions
8.4 Comparison Tests instead of numbers. We shall be especially interested in the properties of
Direct comparison test; Limit power series, which may be thought of as polynomials of infinite degree,
comparison tests
although some of their properties are quite different from those of
8.5 The Ratio Test and the Root
Test polynomials. We shall find that many common functions, such as ex ,
Ratio test; Root test; Summary ln(x + 1), sin A, cos A, and tan -1 A can be represented as power series,
of convergence tests for positive
term and nonnegative-term and we shall discuss some important theoretical and computational
series aspects of this kind of representation.
8.6 Alternating Series; Absolute
and Conditional Convergence
Leibniz’s alternating series test;
Error estimates for alternating
series; Absolute and conditional PERSPECTIVE
convergence; Rearrangement of
the terms in an absolutely Series, or sums, arise in many different ways. For example, suppose it is known
convergent series that a certain pollutant is released into the atmosphere at weekly intervals and
8.7 Power Series is dissipated at the rate of 2% per week. If m grams of pollutant are released
Convergence of a power series; each week, then at the beginning of the first week, there will be Sj = m grams in
Power series in (x - c); Term-
by-term differentiation and inte the atmosphere, and at the beginning of the second, there will be 0.98m grams
gration of power series of “old” pollutant left plus m grams of “new” pollutant, to yield a total of 5'2 =
8.8 Taylor and Maclaurin Series m + 0.98m grams. Continuing, at the beginning of the «th week there will be
Taylor and Maclaurin
polynomials; Taylor’s theorem; S n = m + 0.98m + (0.98)2m + ∙∙∙ + (0.98)"- l m
Taylor and Maclaurin series;
Operations with Maclaurin and grams. It is natural to wonder how much pollutant will accumulate in the “long
Taylor series run” (as n → +∞), and the answer is given by the infinite sum
Chapter 8 Review .S'+ .j, = m + 0.98m + (0.98)2m + ∙∙∙
But just exactly what do we mean by such a sum, and if the total is a finite
number, how can we compute its value?
We seek answers to these questions in this chapter.
489
490 Ch. 8 Infinite Series
8 .1 SEQUENCES AND THEIR LIMITS
IN THIS SECTION Sequences; the limit of a sequence; bounded, monotonic

sequences
Most phenomena we have considered occur continuously, but in practically
every field of inquiry, there are situations that can be described by cataloging
individual items in a numerical listing. In this section we define a mathematical
tool, called a sequence, in order to do this cataloging, and then we define what
we mean by the limit of a sequence. In the next section we shall use our
knowledge of sequences to define and study infinite sums.
The making of a motion picture is a complex process, and editing the film into a
movie requires that all the frames of the action be labeled in chronological order.
For example, R 2 1-435 might signify the 435th frame of the 21st reel. A
mathematician might refer to the movie editor’s labeling procedure by saying the
frames are arranged in a sequence.
■ SEQUENCES
A sequence is a succession of numbers that are listed according to a given

prescription or rule. Specifically, if n is a positive integer, the sequence whose wth
term is the number an can be written as
α 1, α2, . . . , a , . . .
or, more simply,
{αn}
The number a n is called the general term of the sequence. We will deal only with
infinite sequences, so each term an has a successor an + l and for n > 1, a
predecessor α 1. For example, by associating each positive integer n with its
reciprocal we obtain the sequence denoted by
This drawing represents consecu
tive frames from a 16-mm movie jjj, which represents the succession of numbers 1, ⅛, ∣ , . . . , jt , . . .
film showing a golfer from the
moment he is at the top o f his
backswing until he hits the ball. If The general term is denoted by an = ±.
this film were projected at a rate
of several frames per second, the The following examples illustrate the notation and terminology used in
viewer would have the illusion of connection with sequences.
seeing the golfer in action as he
makes his downswing. EXAM PLE 1 Given the general term, find particular terms of a sequence
Find the 1st, 2nd, and 15th terms of the sequence {a }, where the general term
is
Solution If n = 1, then a χ = (j) 1 1 = 1. Similarly,
^ = ⅛ rι = r= 2 - i4
Sec. 8.1 Sequences and Their Limits 49 1
The reverse question, that of finding a general term given certain terms of a
sequence, is a more difficult task, and even if we find a general term, we have no
assurance that the general term is unique. For example, consider the sequence
2, 4, 6, 8, . . .
This seems to have a general term an = 2n. However, the general term
an = (« - 1)(∏ - 2)(« - 3)(« - 4 ) + 2n
has the same first four terms, but a 5 = 34 (not 10, as we would expect from the
sequence 2, 4, 6, 8).
It is sometimes useful to start a sequence with a0 instead of α 1j that is, to have
a sequence of the form
βθ, 6Z। ∙ ∙ ∙
So far, we have been discussing the concept of a sequence informally, without

a definition. We have observed that a sequence {an} associates the number an with
the positive integer n. Hence, a sequence is really a special kind of function, one
whose domain is the set of all positive (or possibly, nonnegative) integers.
Sequence A sequence { α j is a function whose domain is a set of nonnegative integers

and whose range is a subset of the real numbers. The functional values a l , a 1, ∣
a3 , . . . are called the terms of the sequence, and an is called the nth term,
or general term, of the sequence.
0 Even though we write a , In this text, we shall only consider real-valued sequences— that is, sequences in
remember this is a function, which each an is a real number. Unless otherwise specified, “ sequence” will always
mean “ real-valued sequence.” Although a sequence is a function, we usually
a(ri) ~ n + ∣where the domain
represent sequences as {<2n} rather than with functional notation.
is the set of nonnegative
integers. 0
■ THE LIMIT OF A SEQUENCE

a. Graphing a sequence in one
dimension It is often desirable to examine the behavior of a given sequence {an} as n gets
arbitrarily large. For example, consider the sequence
« = — r-r
'1 77 + 1
1 7 7
Because a l = α, = α3 = ∣> . . . , we can plot the terms of this sequence on a
number line, as shown in Figure 8 .la, or the sequence can be plotted in two
dimensions, as shown in Figure 8. lb. By looking at either graph in Figure 8.1, we
see that it appears the terms of the sequence are approaching 1. In general, if the
terms of the sequence approach the number L as n increases without bound, we
say that the sequence converges to the limit L and write
L = lim a l
n→ ∞ '
b. Graphing a sequence in two
dimensions
(Note that for simplicity, the limiting behavior is denoted by n → ∞ instead of the
usual n → +∞). For instance, in our example, we would expect
Figure 8.1 Graphing the sequence
lim n
n→ ∞ n+ 1
This limiting behavior is analogous to the continuous case (discussed in Section

3.5), and may be defined formally as follows.
Convergent Sequence The sequence {an} converges to the number L , and we write
L = Iim a n
α5 α 4 a3 a2 αl
n→∞
* * * ------- γ p * -------- > - if for every ε > 0, there is an integer N such that
L -ε 'L
an L+ε
∖a - L ∖ < ε whenever n > N
a. one dimension
Otherwise, the sequence diverges.
What this says: The notation L = lim an means that the terms of the
sequence {a } can be made as close to L as may be desired by taking n
sufficiently large.
A geometric interpretation of this definition is shown in Figure 8.2.

Note that if
L = lim an
n→∞
the numbers a may be practically anywhere at first (that is, for “ small” n), but
eventually, the an must “ cluster” near the limiting value L .
The theorem on limits of functions carries over to sequences. We have the
Figure 8.2 Geometric interpreta following useful result.
tion of a converging sequence
T H E O R E M 8.1 Limit theorem for sequences
If lim an = L and limn b = M , then the following limits exist, and:

n→∞ n→∞
Linearity rule lim (raπ + sbn) = rL + sM

Product rule li m (αA) =lm
n—>∞ 'l 'l
α τ
Quotient rule lim √≈ = ~ provided M ≠ 0
n →∞bn M
Root rule lim v a = V L provided ∖f f is defined for all n and
AZ-→∞ ,1
m∕-
∖ L exists
Proof: The proof of these rules follows from the limit rules that were stated in
Section 3.5. ■
E X A M P L E 2 Convergent sequences
Find the limit of each of these convergent sequences:
∫ 1001 . [ 2n 2 + 5» - 7) ∫ 3n4 + n - 1 ∣
I n J , I n3 J C ∙ 15M4 + 2M2 + 1J
Solution a . As n grows arbitrarily large, 100//7 gets smaller and smaller. Thus,
lim ≡ = 0
Λ∕→∞ ∕Z
Figure 8.3 Graphical representa
tion of an = 100/n A graphical representation is shown in Figure 8.3.
Sec. 8.1 Sequences and Their Limits 493
b. We cannot use the quotient rule of Theorem 8.1 because neither the limit
in the numerator nor the one in the denominator exists. However,
2n 2 + 5 n - l = 2 + J . _ J7.
ni n ∏2 ∏i
and by using the linearity rule, we find that
lim ~n + — - = lim ⅜ + 5 lim -⅛ - 7 lim -k
∕7→∞ ∕7 j n→ ∞ rl n→ ∞ ∏ n→ ∞
=0+0+0
=0
Figure 8.4 Graph of A graph is shown in Figure 8.4.
c. Divide the numerator and denominator by n4 to obtain
A graph of this sequence is shown in Figure 8.5.
EXAMPLE 3 Divergent sequences

Show that the following sequences diverge:
a. {(-l)''} . f w5 + n 3 + 2 1
I?/?4 + n 2 + 3-1
Figure 8.5 Graph of
_ 3κ4 + n - 1 Solution a. The sequence defined by { (-1)'1} is -1 , 1, -1 , 1, . . . . This sequence
', 5n 4 + 2Λ2 + 1 diverges by oscillation because the nth term is always either 1 or -1 , so a
cannot approach one specific number L as n grows large. The graph is shown
in Figure 8.6a.
b . lim + »3 + 2 = * + γ +
t ,→-x Jn4 + M2 + 3 2 + J-i + ^-,

n n n
The numerator tends toward 1 as n → ∞, and the denominator approaches
0. Hence the quotient increases without bound, and the sequence must
diverge. The graph is shown in Figure 8.6b.
Y = (X 5 + X 3 + 2 ) ∕( 7 X 4 +X2 + 3 )
X - f ∏ ’ y ∙
•
X Y
■2.5 e *
•
6 .8774009 •
1.0 1 7 3 7 9 ■2.0 ∙
8 1 .1 5 8 0 7 8 •
9 1.2 9 9 2 5 5 •
10 1.4 4 0 7 6 6 - 1 .5
11 1 .58252 •
12 1.7 2 4 4 5 8 •
13 1.8 6 6 5 3 6 - 1 .0 ∙
14 2 .0 0 8 7 2 5 •
15 2 .1 5 1 0 0 3 •
16 2.2 9 3 3 5 2
17 2.4 3 5 7 6 2
18 2.5 7 8 2 2 1
19 2.7 2 0 7 2 1
20 2.8 6 3 2 5 7
4 8 12 16 x
n 5 + n3 + 2
b. a
7n4 + n2 + 3
Figure 8.6 Graph and values for two divergent sequences

If nlim
—⅛□ca does not exist because the numbers a become arbitrarily large as n → ∞,
we write lim αn = ∞. We summarize this more precisely in the following box.
n→∞
Limit Notation lim a = ∞ means that for any real number A, we have an ' > A for all
n→∞ nrι
sufficiently large n.
means that for any real number B, we have bn <; B for all
,l,i→
m≈ b∏= -∞
sufficiently large n.
Rewriting the answer to Example 3b in this notation, we find that

5 3
lim » 4 + π 2 + 2
n→∞ 7w + n + 3
Also notice that lim ( - 5») = -∞ (that is, decreases without bound) whereas
∕7 → ∞
lim (—1)" does not exist. Thus, the answer to Example 3a is neither +∞ or -∞.
n→∞
EXAMPLE 4 Determining the convergence or divergence of a sequence
Determine the convergence or divergence of the sequence {Vn2 + 3» - n}.

Solution Consider
lim (V M2 + 3n - ri)
n→∞
It would not be correct to apply the linearity property for sequences (because
neither lim Vw2 + 3M nor lim n exists). It is also not correct to use this as a
reason to say that the limit does not exist. You might even try some values of n
(shown in the Figure 8.7) to guess that there is some limit. In order to find the
limit, however, we shall rewrite the general term algebraically as follows:
Vtt2 + 3« - n = (Vn 2 + 3« - ri)∖ n 2 + 3n + n

, 2
Vri + 3n + n
Figure 8.7 Graph of
a = V « 2 + 3« —n: convergent ∕ι 2 + 3« - n 2 n _ 3 _3
or divergent?
Vw2 + 3n + n ∣ √ H2 + 3n + 1 l∏2 + 3n , 1
Hence, lim (Vw2 + 3n - ri) = lim . ----- = ⅞

"→x '≈→≈ ∕ 1 + 3 + 1 z
Note the graph of the sequence in Example 4. The graph of a sequence consists
of a succession of isolated points. This can be compared with the graph of
y = Vx 2 + 3x - x ,x ≥ 1, which is a continuous curve (shown in Figure 8.8). The
only difference between ∕7lim
→χ an= L and n→lim∞f(x) = L is that n is required to be an
integer. This is stated in the hypothesis of the following theorem.
THEOREM 8.2 Limit of a sequence

Figure 8.8 Graph of Suppose / i s a function such that an = f(ri) for n = 1, 2, . . . . If lim f(x) exists and
f(x ) = V x 2 + 3x - x, x ≥ 1 lim f(x) = L, the sequence { a ∖ converges and H→ lim∞ a = L.
χ→∞ ,1 tl
Proof: Let ε > 0 be given. Because lim ∕'(x) = L, there exists a number
N > 0 such that
∖f ( x ) - L ∖ < ε whenever x > N

In particular, if n > N, it follows that ∖f(ri) - L∣ = ∖a n - L ∖ < ε.
Be sure you read this theorem correctly. In particular, note that it does not say
that if lim a n = L , then lim ∕( x ) = L (see Figure 8.9b).
does NOT necessarily equal L. (^)
Figure 8.9 Graphical comparison of lim a and lim ∕( x ) , where f(ri) = a for n = 1, 2, . . .
11
M —»20 r → ≡c
EXAM PLE 5 Convergence using ΓH6pitaΓs rule
Given that the sequence ] 1 n r converges, evaluate lim

11 C J n→ ∞ 1 C
Solution Let L = lim ∕( x ) , where ∕( x ) = ∣ gV. Because /(77) = aιl for n = 1, 2,

„2
. . . , Theorem 8.2 tells us that lim -∣---- is the same as lim ∕( x ) .
∕7→∞ 1 C n→ ∞
' γ,- ∙ 2v
lim η------x? = lim ----x? ΓH∂pitaΓs
r ride
Λ-→=C 1 — e x →≈ - e
= lim —2-7. l ,H∂pitaΓs rule again

χ→ ∞ C'
= 0
Thus, by Theorem 8.2, lim 7-3—-1 = lim ∕( x ) = L = 0.
∏→∞ 1 β χ→ ∞
From the graph, it looks like The squeeze theorem (Theorem 1.9) can be reformulated in terms of sequences.
∣>n3∕(α,,) ≤ 0.
THEOREM 8.3 Squeeze theorem for sequences
If an ≤ b« ≤ cn for all n >’ N , and n a - lim

lim n c =’ L , then
,,→x lim
n b = L
n →x
Proof: The proof follows from Theorem 1.9 and Theorem 8.2.
EXAM PLE 6 Using ΓH6pitaΓs rule and the squeeze theorem
Show that the following sequences converge, and find their limits.
a. lim ∕7ιz" b. lim -⅛

n→ ∞ n→ 0 /7
1
Solution a. Let L = lim n l/n; then In L = lim = lim τ = 0 (ΓHopitaΓs rule)
n→ ∞ n→ ∞ /7 n→ ∞ 1
Thus, L = e0 = 1.
b. We cannot use ΓH6pital, s rule because x! is not defined as an elementary

function when x is not an integer. Instead, we use the squeeze theorem for
sequences. Accordingly, note that
s o t ^ι a t ^, ι - ⅛ - 2 ∙ 2’ - 3 ∙ 3 ∙ 3’ ‘ , ’
_ n ∙ ( n - 1) ■■
■3 ∙2 ∙ 1 _ Γn(n - 1) ∙∙∙ 3 ∙ 2^∣/ A
b n~ n ∙n ∙n ∙∙∙n L n ∙ n ∙ n ∙ ∙ ∙ n J∖ n ∕
n factors n — 1 factors
Thus, if an = 0, and cn = ⅛, for all n, we have
⅛ bn f
0 ≤ - ⅛n ≤ - ∙⅛ ≤ - because n ∙ n∖ < πn
n n πn n
Because lim an = 0 and lim c = 0, it follows from the squeeze theorem

∕7→x n→∞ n
that lim ⅛ = 0.
n→≈ ιl
■ BOUNDED, MONOTONIC SEQUENCES
A sequence {αn} is said to be
Increasing Increasing if α1 < a2 < a 3 < < αjt-1 < a k < ∙ ∙ ∙

Nondecreasing Nondecreasing if a χ ≤ a 2 ≤ ∙∙∙ ≤ αfc-1 ≤ ak ≤ ∙∙∙
Decreasing Decreasing if α1 > α9 > α3 > ∙∙∙ > a jt-1 > ak > ∙∙∙
Nonincreasing Nonincreasing if α1 ≥ a 2 ≥ ≥ ak _ { ≥ ak ≥
Monotonic Monotonic if it is nondecreasing or nonincreasing

Strictly Monotonic Strictly monotonic if it is increasing or decreasing
I
Bounded Above Bounded above by M if an ≤ M for n = 1, 2, 3, . . .
Bounded Below Bounded below by m if m ≤ a for n = 1, 2, 3, . . .
Bounded Bounded if it is bounded both above and below
In general, it is difficult to tell whether or not a given sequence converges or

diverges, but thanks to the following theorem, it is easy to make this determina
tion if we know the sequence is monotonic.
T H E O R E M 8.4 B M CT : The bounded, monotonic convergence theorem
A monotonic sequence {an} converges if it is bounded and diverges otherwise.

Proof: A formal proof of the BM CT is outlined in Problem 54. For the
following informal argument we will assume that {an} is a nondecreasing sequence.
(You might wish to see whether you can give a similar informal argument for the
increasing case.) Because the terms of the sequence satisfy a l ≤ a2 ≤ ai ≤ ∙∙∙,w e
know that the sequence is bounded from below by α1 and that the graph of the
corresponding points (rc, a n) will be rising in the plane. Two cases can occur, as
shown in Figure 8.10. Suppose the sequence {a } is also bounded from above by a
number M , so that a l ≤ an ≤ M for n = 1, 2, . . . . Then the graph of the points
(«, αn) must continually rise (because the sequence is monotonic) and yet it must
stay below the line y = M . The only way this can happen is for the graph to
approach a “ barrier” line y = L (where L ≤ M ) , and we have lim a = L , as
shown in Figure 8 .10a. However, if the sequence is not bounded from above, the
graph will rise indefinitely (Figure 8.1 Ob), and the terms in the sequence { α j
cannot approach any finite number L . ■
a. If∏n < Λ ∕forn = 1,2, ..., the graph
of the points (n, an ) will approach a EXAM PLE 7 Convergence using the BMCT
horizontal “barrier” line y = L.
11 ∙ 3 ∙ 5 ∙∙∙ (2/7 - 1)1
Show that the sequence t 2 ∙ 4 ∙ 6 ∙∙∙ (2//) J converges.
Solution The first few terms of this sequence are
aα =1 a =1∆3 = 3 a = 1 ■3 ∙ 5 _5_
ι 2 α 2 2 -4 8 α 3 2 ∙4 ∙6 16
Because ⅛ > ∣ > ^⅛, it appears that the sequence is decreasing (that is, it is
monotonic). We can prove this by showing that a x < a , or equivalently,
< 1. (Note that an ≠ 0 for all n.)
1 ∙ 3 ■5 ∙∙∙ [2(∕∕ + 1) - 1]
b. If (a } is not bounded from above,
the graph rises indefinitely. an + l 2 ∙ 4 ∙ 6 ∙ ∙ ∙ [2(//+1)]
an 1∙ 3 ∙ 5 (2//- 1)
Figure 8.10 Graphical possibilities
for the BMCT theorem 2 ∙4 ■6 ∙∙∙ (2/7)
1 ■3 ∙ 5 (2//+ 1) 2 ∙4 ∙6 (2//) 2/7 + 1
“ 2∙4∙6 (2/7 + 2) ‘ 1 ∙ 3 ∙ 5 ∙∙∙ (2/7 - 1) ~ 2/7 + 2
for any n > 0. Hence, an+χ < a for all n, and {αzι} is a decreasing sequence.
Because a > 0 for all n, it follows that {aιj} is bounded below by 0. Applying
the B M C T , we see that {an} converges, but the B M C T tells us nothing about
the limit. b m
COMMENT: Technically, the sequence {τzzι} is monotonic only when its terms
are either always nonincreasing or always nondecreasing, but the B M CT also
applies to sequences whose terms are eventually monotonic. That is, it can be
shown that the sequence {α } converges if it is bounded and there exists an
integer N such that {<2κ} is monotonic for all n > N. This modified form of the
B M CT is illustrated in the following example.
EXAM PLE 8 Convergence of a sequence that is eventually monotonic
Show that the sequence [ converges.
Solution We will apply the B M CT. Some initial values are shown in Figure 8.11.
The succession of numbers suggests that the sequence increases at first and
then gradually begins to decrease. To verify this behavior, we let
/« =⅛
V-A
2 V x - V x In x
Figure 8.11 a ιι = and find that ∕ , (x)
2x 2
Find the critical values: .∖ √ - _ θ

2χ-
2 V x = V x In x
In x = 2
e2 = x
Thus, x = e 2 is the only critical value, and you can show that ∕'( x ) > 0 for
χ < e2 and ∕'( x ) < 0 for x > e2 . This means that / is a decreasing function
for x > e 2 . Thus, the sequence ⅛ must be decreasing for n > 8 (because
e2 is between 7 and 8). We see that the sequence is bounded from below
because
0 1
V/7
Therefore, the given sequence is bounded from below and is eventually
decreasing, so it must converge. ■■■■
The B M C T is an extremely valuable theoretical tool. For example, in Chapter

5, we defined the number e by the limit
lim (1 +7ι)n =e
but to do so we assumed that this limit exists. We can now show this assumption is
warranted, because it turns out that the sequence {(1 + £)"; is increasing and
bounded from above by 3 (the details are outlined in Problem 53). Thus, the
B M CT assures us that the sequence converges, and this in turn guarantees the
existence of the limit. We end this section with a result that will be useful in our
subsequent work. The proof makes use of the formal definition of limit.
THEOREM 8.5 Convergence of a power sequence
If r is a number such that ∣r∣ < 1, then lim rn = 0.

∕z→∞
Proof: The case where r = 0 is trivial. We shall prove the theorem for the case
0 < r < 1 and leave the case - 1 < r < 0 as an exercise. Because 0 < r < l,w e
can write r as
for some h > 0, so

= -----i----n
' (l + ∕7)
Because (for all fixed ri)
n(π — 1)
(1 + /7)" = 1 + nh H-----J]— " 2 + ” ■+ Binomial theorem
> nh Because h > 0
we see
r" = (∏ ⅛ < ⅛ lfa > b ' t h e n ka < k b-
We now use the definition of limit. Given ε > 0, we see that if n > -±r> then
1 εn 1
< ε because both n and ε are greater than zero. We can then choose N = -^∙>
and if n > N, then r" < ε and by the definition of limit we conclude
lim r', = 0. ■
PROBLEM SET 8 .1
@ 1. ■ W hat D oes T h is Say? W h at do we m ean by the lim it E x p la in why each sequence in Problem s 4 2 -4 5 diverges.
o f a sequence? 42. { l + ( - l ) " } 43. {cos nπ}
2. ■ W hat D oes T h is Say? W h at does the bounded, m o n o 44 b 3 - 7 ∕t + 5 1
45. {V n }
tonic, convergence theorem say? ■ L lP P n 2 + 2 19J
Write out the fir s t fiv e term s (beginning with n = 1) o f the 46. Suppose a particle o f mass m moves back and forth along
sequences given in Problem s 3 -1 1 . a line segment o f length ∣α∣. In classical m echanics, the
particle can m ove at any speed, and thus, its energy can be
3. {1 + ( - l ) " } any positive num ber. H ow ever, quantum m echanics re
places this continuous m odel o f the p article’s behavior
w ith one in w hich the p article’s energy level can have only
certain discrete values, say, E l , E 2 , . . . . Sp ecifically, it
can be shown that the nth term in this q uantum sequence
has the value
9. {α π} where α 1 = 256 and a = V α 1 for n ≥ 2
n 2h 2
n = 0, 1, 2, . . .
10. {a n} where a l = - 1 and a = n + an [ for n ≥ 2 8m a2
11. {a n} where α l = 1 and a n = ( α ,,. 1) 2 + a „ _ , + 1 for n ≥ 2
where h is a physical constant know n as P la n ck ’s constant
C o m p u te the lim it o f the convergent sequences in P roblem s (h ≈ 6.63 × 10^^27 erg∕s). List the first five values o f E for
1 2 -3 5 . a particle with mass 8 m g m oving along a segment o f
length 100 cm .
5n + 81 ( 5/7 1
12. 13. 'In + 7 J 47. A drug is adm inistered into the body. A t the end o f each
n J
hour, the am ount o f drug present is h a lf what it was at the
14
2/7 + 11 14 - 7nl end o f the previous hour. W hat percent o f the drug is
. 15. L 8 + /7 J
3/7 - 4 J present at the end o f 4 hr? A t the end o f n hours?
16. { 8n 2 + 800/7 + 5,0001 (100n + 7 ,pool 48. T he Fibonacci rabbit problem T he general term in a
17. < 2
2 n 2 - 1,000/7 + 2 J ( n - n —1 J sequence can be defined in a num ber o f ways. In this
18. { 8∕72 + 6/7 + 4,0001 ( ∕73 - 6∕7 2 + 85 problem , we consider a sequence whose nth term is
19. defined by a recursion form ula', that is, a form ula in which
∕73 + 1 J l-2∕73 — 5/7 + 17P
the ∕7th term is given in terms o f previous terms in the
20. 1 In 1 ( 8/7 — 5 P θ V n l
21. sequence. T h e problem was originally exam ined by L e o
/7 + ^1∖f n ) (2n + 8 O θ V nJ nardo Pisano (also called Fibonacci) in the 13th century.
22. ] 3√n 1 (in «1 T h e so-called F ib o n a cci sequence has applications in
2 3 .. m any areas, including biology and botany.
.5 V ∕7 + 5√zn J I n2 J
Suppose rabbits breed in such a way that each pair o f
24. {:2 5/n} 25. {∕7 3 Z " }
adult rabbits produces a pair o f baby rabbits each m onth.
26. ∣( ' + i )"l
27. {(n + 4 ) ιz "} T h e first m onth after birth , the rabbits are adolescents
A nJ J and produce no offspring. H ow ever, beginning with the
28. { W '∕( M+2)} 29. {(In ∕7) lz "}
Number Number
Pairs of Rabbits (the pink rabbits are ready to
30. 1 ∣ e~ n x d x I 31. {V ∕7 2 + n - nj
of of
reproduce)
Months Pairs
Jo J
32. { √∕7 + 5 V Z∕7 - V n j 33. {ΛvΛm
∕~) Start 1
34. {(an + b ) ιz "}, a and b positive constants

. 1
35. {ln n — ln(∕7 + 1)}
Q S h o w that each sequence given in Problem s 3 6 -4 1 converges

2 2 ⅛1 .. 1
either by show ing it is increasing with an upper bound or
decreasing with a lower bound.
3 3 ⅛x ; V. x
36∙ {⅜} 37∙{ln(⅛1)} 4 5 ⅛: ∖ "...; 1 - V. x .
; ∖ ; ∖ ;
V A V y Λ τ ¼ v
38. ( 3 w ~ 2 ∣ 39. ( 4 z7 + 5 ∣ 5 8 .,,.....- ■ ∖∖ ⅛ ∕∕
I n J I n J
40. { ⅛ 2 } 41. {X^} V Ψ ψ V
Same pair
(rabbits never die)
second month, the rabbits are adults and each pair 52. If lim e " = 0, find N so that -⅛ < 0.001 if n > N.
n→× e"
produces a pair of offspring every month. Let an denote
the number of adult pairs of rabbits in her “colony” at the 53. Convergence of the sequence {(1 +
end of n months. a. Suppose n is a positive integer. Use the binomial
a. Explain why α 1 = 1, a2 = 1, a3 = 2, α4 = 3, and in theorem to show that
general
a,l+l = an_, + an for n = 2, 3, 4, . . .
b. The growth rate of the colony during the (n + l)th
month is
Compute rn for n = 1, 2, 3, . . . , 10. b. Use part a to show that 1 + - ) < 3 for all n.
n∕
c. Assume that the growth rate sequence {rι} defined in
part b converges, and let
c. Use part a to show that is an increasing
L = lim r sequence. Then use the BMCT to show that this
n →∞ "
Use the recursion formula in part a to show that sequence converges.
54. Proof of the BMCT Suppose there exists a number M
and a positive integer N so that a ≤ a l ≤ M for
all n > N.
and conclude that L must satisfy the equation
a. It is a fundamental property of numbers that a se
quence with an upper bound must have a least upper
bound; that is, there must be a number A with the
Use this information to compute L. property that an ≤ A for all n but if c is a number such
that c < A, then am > c for at least one m. Use this
In Problems 49-52, use thefact that lim an = L means ∖arl —L∖ property to show that if ε > 0 is given, there exists a
positive integer N such that
is arbitrarily small when n is sufficiently large.
< A - ε < aN < A
49. If →
lim ——r = 1, find N so
In that
+ 1 —τ~r
I ~ 1 θ∙θl if
n ∞n + 1
n > N. b. If N is the integer that satisfies the condition in part a,
show that
50. If lim 2 n ,+ ⅛1- = 2, find N so that 2/7 ,+ , 1 - 2 < 0.01 if
n→ ∞ 77 + 3 In + 3 I A - ε < an < A
n > N.
for all n > N.
51. If lim z⅛ 1 = 0, find N so that zl⅛ i < 0.001 if
n→= n 3 I ni I c. Show that ∖a —A∖ < ε for all n > N and conclude
n > N. that ,,-.-
lim a'> = A.
r
8.2 INTRODUCTION TO IN FIN ITE SERIES: GEOMETRIC SE R IE S
IN TH IS SECTION Definition of infinite series, general properties of

infinite series, geometric series, applications of geometric series
We define an infinite series (sum) as the limit of a special type of sequence. We
then study a few basic properties of infinite series and examine geometric series,
a special kind of series with numerous applications.
■ D EFIN ITIO N OF IN FIN ITE SE R IE S
One way to sum a list of numbers is to form subtotals until the end of the list is
reached. Similarly, to give meaning to the infinite sum
Sec. 8.2 Introduction to Infinite Series: Geometric Series 501
S —<2∣+ Cl-l + <7. + fl4 +
it is natural to examine the “ partial sums”
ai , α 1 ÷ α,, a 1 + α ,+ a v a i + a7 ÷ ct-i ÷ α4, . . .

If it is possible to attach a numerical value to the infinite sum, we would expect the
partial sums S n = a λ + a 1 + + an to approach that value as n increases without
bound. These ideas lead us to the following definition.
Infinite Series An infinite series is an expression of the form

Ω1 + d!9 + + ∙ ∙∙ = ∑ Ct∣f,
Historical Note
The mathematician and the nth partial sum of the series is
Brahmagupta was the most n
$n ~ a ∖ + a 2 + "' + a n~ a
j,∑ ι k
prominent of the Hindu mathe
maticians of the 7th century. The series is said to converge with sum 5 if the sequence of partial sums {S }
He did some early work on converges to S. In this case, we write
arithmetic series and the sum X
of the squares of natural num ∑ a . = lim
κ n
= 5,
k=∖ n→ ∞
bers. He wrote an astronomical
work that contained several If the sequence {5J∙ does not converge, the series
x
chapters on mathematics. He ,Σ a kκ
was an exceptionally clear ex k=l
positor, and before each exam diverges.
ple he worked, he stated a rule
that was appropriate to the What this says: An infinite series converges if its sequence of partial sums
problem. converges and diverges otherwise. If it converges, its sum is defined to be the
Source: Chandra Sharma, Univer
limit of the sequence of partial sums.
sity of Manchester Institute of Science
and Technology.
R E M A R K : We shall use the symbol ∑ a. to denote the series a. + a , +

⅛=ι
α3 + ∙∙∙ regardless of whether this series converges or diverges. If the sequence
of partial sums {S } converges, then
X
∑ a.κ = lim
k= 1 n→ ∞
and the symbol ∑ ak is used to represent both the series and its sum.
Also, we shall consider certain series in which the starting point is not 1; for
example, the series
1 1 1 ι ι
⅜+ -⅛+ ⅛ + ∙∙∙ can be denoted by Σ τ or ∑ -, ,
3 4 5 ∙ t⅛3
⅛ k≡2 k + 1
EXAM PLE 1 Convergent series
≈ ∣
Show that the senes ∑ converges.
A-= ι 4
Solution This series has the following partial sums.

S, = ⅛
S' = l ⅛ l = i
*-, 2 2 4 4
S' =
d - 1 + 1+ 1= 2
3 2 4 8 8
$n ~ 2 + 4 + ■" + 2"
The sequence of partial sums is ∣> 5. . . ,

∣⅜>-J-J ⅞ and in general,
c = ι1
_ X w = 2n -π 1.
*j π 2 2
2n — ι / 1∖
Because lιm — = lim 1 - y j = 1, we conclude that the series converges
n→∞ Λ fj→∞∖ JL /
and its sum is 1. M M
EXAM PLE 2 Divergent series
Show that the series ∑ ( - 1 ∕' diverges.

* = ι
Solution The series can be expanded (written out) as

X
∑ ( - l) * = -1 + 1 - 1 + 1 - 1 + 1 - ∙ ∙ ∙
⅛ =1
and we see that the nth partial sum is

_ C - 1 if n is odd
X - ( 0 if n is even
Because the sequence {S } has no limit, the given series must diverge, ∣= ι
A series is called a telescoping series (or collapsing series) if there is internal
cancellation in the partial sums, as illustrated by the following example.
EXAM PLE 3 A telescoping series
X 1
Show that the series ∑ -∩---- τ converges and find its sum.
k=∖ k~ + k
Solution Using partial fractions we find that

i = i = i , -i
k^ + k k(k + 1) k k + 1
Thus, the nth partial sum of the given series can be represented as follows:
n+ 1
The limit of the sequence of partial sums is

lim .S' = lim [ 1 ----- 3—r ∣ = 1
n→∞ n N →,X L n + 1J
Thus, the series converges, with sum 5 = 1 .
■ GENERAL PROPERTIES OF INFINITE SERIES
Next, we shall examine two general properties of infinite series. Here and
elsewhere, when the starting point o f a series is not important, we may denote the
series by writing
~Sta k instead of ∑ ak
THEOREM 8.6 Linearity of infinite series
If 1a k and 'Zbk are convergent series, then so is ∑(c¾ + dbk} for constants c, d, and
∑(cα, + db,) = c ⅛a, + d ⅛b,
Proof: Compare this with the linearity property in Theorem 4.1. In Chapter 4
the limit properties are for finite sums, and this theorem is for infinite sums. The
proof of this theorem is similar to the proof of Theorem 4.1, but in this case it
follows from the linearity rule for sequences (Theorem 8.1). The details are left as
a problem. ■
EXAM PLE 4 Linearity used to establish convergence
∞ r Λ 1
Show that the series ∑ 77---- 7 - 37 converges, and find its sum.
⅛ =ι L /c I- κ 2 -1
” 1 x 1
Solution Because we know that ∑ 7^---- r and £ - k both converge, the linearity
k= 1 k P h k= 1 - '
property allows us to write the given series as

00 1 “ 1
4 ∑ ι 1 -u ∕ “ 6 Σ ⅛
k=∖ ki + k k = ∖ 2k
We can now use Examples 1 and 3 to conclude that the series converges and
that the sum is
4 Σ 7τ⅛ 7 - 6 ∑ ⅜ = 4(1) - 6(1) = - 2

*= ι k- + k k=ι 2 lc M B
The linearity property also provides useful information about a series of the
form ~⅛(cak + dbk) when either ⅛ak or ⅛bk diverges and the other converges.
THEOREM 8.7 Divergence of the sum of a convergent and a divergent

series
If either ∑αt or ⅛bk diverges and the other converges, then the series ∑(αt + bk)
must diverge.
Proof: Suppose ⅛ak diverges and ∑⅛λ. converges. Then if the series ∑(αj,. + bk)
also converges, the linearity property tells us that the series
∑[(¾ + bk) - bk] = ⅛ak
must converge, contrary to hypotheses. It follows that the series S(¾ + bk)
diverges. ■
This theorem tells us, for example, that Y , 7 ∖ , + ( - l) t must diverge because
L/C + ∕C J
∞ 1 ∞
, 1---- τ^ converges (Example 3), ∑ ( - l ) t diverges (Example 2).
even though ∑ ■
⅛=∣k~ + k ∕<=ι
■ GEOMETRIC SERIES
We are still very limited in the available techniques for determining whether a
given series is convergent or divergent. In fact, the purpose of a great part of this
chapter is to develop efficient techniques for making this determination. We begin
this quest by considering an important special kind of series.
Geometric Series A geometric series is an infinite series in which the ratio of successive terms is
constant. If this constant ratio is r, then the series has the form
∞
∑ ark = a + ar + ar2 + ar 3 + ∙ ∙ ■+ arn + ∙ ■∙ a≠ 0
k=0
For example, 3 + ∣ + ¾+ ∣ + . . . i s a geometric series because each term is one-

half the preceding term. The ratio of a geometric series may be positive or
negative. For example,
is a geometric series with r = - ∣ . The following theorem tells us how to determine

whether a given geometric series converges or diverges and, if it does converge,
what its sum must be.
T H E O R E M 8.8 Geometric series theorem
The geometric series ∑ ark with a ≠ 0 diverges if ∣r∣ ≥ 1 and converges if ∣r∣ < 1
with sum k=0
∑ ark = a
k≈0 1- r
Proof: Note that the nth partial sum of the geometric series is
S n = a + ar + ar 2 + + arn ~ l
rS n = ar + ar2 + ar3 + ∙ ∙ ∙ + arn Multiply both sides by r.
rS n - S n = (ar + ar2 + ar3 + + ar") - (a + ar + ar2 + ∙∙∙ + ar"~')
(r - 1)S 1 = ar" - a
a(rn - 1)
( r -l)
If ∣r∣ > 1, the sequence of partial sums {Sιι} has no limit, so the geometric series
must diverge. However, if ∣r∣ < 1, Theorem 8.5 tells us that rn → 0 as n → ∞, and
we have
, fl( 0 -l) - a
∑ ark = lim S n = lim a r' - 1
k=0 n→∞ n→∞ r - 1 r- 1 1- r
To complete the proof, it must be shown that the geometric series diverges when
∣r∣ = 1, and we leave this final step as a problem. ■
EXAM PLE 5 Convergence or divergence for a geometric series
Determine whether each of the following geometric series converges or

diverges. If the series converges, find its sum.
Solution a. Because r = j satisfies ∣r∣ ≥ 1, the series diverges.

b. We have r = - ∣ so ∣r∣ < 1, and the geometric series converges. We
see that the first term (for k = 2) is a = 3 (-∣) 2 = so
y √ - B5 t - a = 25 - X
ι⅛ ' 1- r 1 _ (_1) 10
■ APPLICATIONS OF GEOMETRIC SERIES
Geometric series can be used in many different ways. Our next three examples
illustrate several applications involving geometric series.
Recall that a rational number r is one that can be written as r = p/q for integer
p and nonzero integer q. It can be shown that any such number has a decimal
representation in which a pattern of numbers repeats. For instance,
= 0.5 = 0.50 = 0.454545 . . . = 0.45
where the bar indicates that the pattern repeats indefinitely. The next example
shows how geometric series can be used to reverse this process by writing a given
repeating decimal as a rational number.
EXAM PLE 6 Repeating decimals
— P
Write 15.423 as a rational number —.
q
Solution The bar over the 23 indicates that this block of digits is to be repeated;
that is 15.423232323 ∙ ∙ ■. The repeating part of the decimal can be written as
a geometric series as follows.
15.423 = 15 + ⅛ + ⅛ + ⅛ + ⅛ + ∙ ∙ ∙
- 15 + ⅛ + ⅜ [1 + ⅛ + T⅛ + '■']
Sum o f a geometric
series with a = 1 and
r = 1/100
15,269
990
A tax rebate that returns a certain amount of money to taxpayers can result in
spending that is many times this amount. This phenomenon is known in
economics as the multiplier effect. It occurs because the portion of the rebate that
is spent by one individual becomes income for one or more others who, in turn,
spend some of it again, creating income for yet other individuals to spend. If the
fraction of income that is saved remains constant as this process continues
indefinitely, the total amount spent as a result of the rebate is the sum of a
geometric series.
EXAM PLE 7 Multiplier effect in economics
Suppose that nationwide, approximately 90% of all income is spent and 10%
is saved. How much additional spending will be generated by a $40 billion tax
rebate if savings habits do not change?
Solution The amount (in billions) spent by original recipients of the rebate is
0.9(40). This becomes new income, of which 90%, or 0.92(40), is spent. This,
in turn, generates additional spending of 0.9 3(40), and so on. The total
amount spent if this process continues indefinitely is
0.9(40) + 0.9 2(40) + 0.9 3(40) + ∙∙∙ = 0.9(40)[l + 0.9 + 0.9 2 + ∙ ∙ ∙]
∞
= 36 ∑ 0.9A-' Geometric series
with a = 1, r = 0.9
- 1
j0 ∖l - 0.9/
= 360 (billion) c= zz3
EXAM PLE 8 Accumulation of medication in a body
A patient is given an injection of 10 units of a certain drug every 24 hr. The

drug is eliminated exponentially so that the fraction that remains in the
patient’s body after t days is ∕( t ) = e~t' i . If the treatment is continued
indefinitely, approximately how many units of the drug will eventually be in
the patient’s body just prior to an injection?
Solution O f the original dose of 10 units, only 10e^ιz5 are left in the patient’s
body after the first day (just prior to the second injection). That is
5, = 10<T1/5
The medication in the patient’s body after 2 days consists of what remains
from the first two doses. O f the original dose, only 10c^2z5 units are left
(because 2 days have elapsed), and of the second dose, 10e,- ιz 5 units remain:
,S'2 = 10<J - IZ5 + 10e"2z5
Similarly, for n days,
S n = 10e-ιz 5 + 10e^2z5 + ∙∙∙ + 10e~n'5
The amount 5 of medication in the patient’s body in the long run is the limit
of S ll as n → ∞. That is,
A = lim An
n →∞
= f 10e-4z5 Geometric series with

*=1 a = 10e- ι z , and r = e~',5
10e~ιz5
1 - e^ ιz5
≈ 45.166 556
We see that about 45 units remain in the patient’s body.
0 A sa final note, remember that a sequence is a mere succession of terms,

while series is a sum of such terms. Do not confuse the two concepts. For
example, a sequence of terms may converge, but the series of the same terms
may diverge:
11 35917
1 + y J∙ is the sequence y , 4 ’ 8’ T6’ ■∙ ∙ >which converges to 1.
/ 1∖ 3 5 9
∑ I 1 + —A is the series f + 4 + g + ∙ ∙ ∙ > which diverges. 0
PROBLEM SET 8 .2
θ 1. ■ What Does This Say? Explain the difference between In Problems 23-30, each series telescopes. In each case, express
sequences and series. the nth partial sum S n in terms o f n and determine whether the
2. ■ What Does This Say? What is a geometric series? series converges or diverges by examining lim S n .
n →∞
Include a derivation o f the formula for the sum of a
geometric series. a ∑⅛^(Γ⅛J u ⅛ -≡ ]
Determine whether the geometric series given in Problems 00 00
1 1
3-22 converges or diverges, and fin d the sum o f each conver 25' ∑0 (∕c ÷ l)(Ac + 2) 26∙ ∑0 (∕< + 2)(∕< + 3)
gent series.
v ( i ∖k v (-^ ∖ k 27. Y ln( 1+4 -) 28. y +77------. 1 , , —I—+
3. A I 5 / 4. Σ I 5 /
fc÷ι ∖ k) fc⅛1 ( 2 Z c -1)(2⅛+1)
j
k=0 ∖ ∕ j
k=0 ∖ ∕
∞ 7 ∞ 7
5. Σ ⅜ 6. ς (-3 fi 29. Σ 2/7 + t∖2 Hint∖ Use partial-fraction decomposition,
⅛=0 i k=0 ( V ⅛=ι k (k + 1)-
∑ (ff ” V⅛ + 1 - Vfe Hint'. Note that
7. 8. Σ ⅜
k=∖ ∖z ∕ k=l 2 ∑ 1 V'∕c 2 + k f i k 2 + k = ∖f i y/k + 1.
j
9. y - - t 10. ∑ 5(0.9)^ Express each decimal given in Problems 31-34 as a common
⅛¾ ( -4 ) ∕<=1
(reduced) fraction.
∞ -ιk
11. ∑ e -0.2fc 12. 31. 0.01 32. 2.231
k=∖ 4A ÷2
33.1.405 405 34. 41.201 OθΓθ
$ (~2)λ-
13. 14. Σ M ) ⅛ S θ 35. a. Find numbers A and B such that
⅛¾ 3A ÷' k≈2
— -I
+ c1π - + k - 1 _ A k _ Bk + 1
15. Icq 1 1eq ∣<-⅛ q>^ ⅜ '
2A+1
2 ∕<+ι 2*
—+ t + Λ
(, I
16. ≈ + π3 + ∙
— "-- —7
b. Evaluate Y ∖ ,. √.
+ -- 1 — + ( ι ) 7 ÷ ( τ ∖1° k=∖ 27 κ+ 1
17. let -- -4
' ∙
+
'- r
,4 y
cq - q — + ∕2 ∖ 5 _ / t7 ∞ 9∕√ - 1
∣m 1 --- ∣r 'Σ 36. Evaluate y — a,+ 1 .
18. - c -s
- ∖3) ' ,3>1 + ⅛=ι 3
+ 1
Hint: Follow the procedure in Problem 35.
19. 2 + √2 + 1 ÷ ⅛ + ∙∙∙ / 1 ∖
2 ∞ In —- ------
20. 3 + √3 + 1 + _ L + 1 + ... 37. Evaluate ∑ ∖(¾ + 1)”/ .
√3 3 "= l n(n + 1)
21. (1 + √2) + 1 + (-1 + √2) + ∙∙∙ Hint: Use the properties of logarithms to show that the
22. (√2 - 1) + 1 + (√2 + 1) + series telescopes.
, ” n Hint: Express as a telescoping 52. Winnie Winner wins $100 in a pie-baking contest run by
38. Evaluate > -— —1~τ. . the Hi-Do Pie Co. The company gives Winnie the $100.
,,÷ι (w + 1)! series.
However, the tax collector wants 20% of the $ 100. Winnie
39. Find ∑ (2al . + 2^fc) given that ∑ a k = 3.57. pays the tax. But then she realizes that she didn’t really
⅛=0 ⅛=0 win a $100 prize and tells her story to the Hi-Do Co. The
x A —3 - *-' ∞
40. Find ∑ — ? ---- given that ∑ bk = 0.54. friendly Hi-Do Co. gives Winnie the $20 she paid in taxes.
k=o - k=0 Unfortunately, the tax collector now wants 20% of the
41. Evaluate ∑ (⅛ + A ) ∙ $20. She pays the tax again and then goes back to the
Hi-Do Co. with her story. Assume that this can go on
indefinitely. How much money does the Hi-Do Co. have
42. Evaluate ∑ Γ(= Y + (j Y Y .
to give Winnie so that she will really win $100? How
A÷ O L∖3 ∕ W J much does she pay in taxes?
43. Show that ∑ -.—+1 , <7— 7 - , . , τ = - if a > 0. 53. A patient is given an injection of 20 units of a certain drug
A⅛O (A ^>^∙A + * + 1) a every 24 hr. The drug is eliminated exponentially, so the
fraction that remains in the patient’s body after t days is
44. Evaluate ∑ —---7-r—
2 ∑----- r ∙ j ∖t) = e~ υ 2 . If the treatment is continued indefinitely,
⅛÷o (2 + k) + 2 + k
approximately how many units of the drug will eventually
45. If ∑ al 2 = ∑ b , 2 = 4 and Y a,b. = 3, what is be in the patient’s body just prior to an injection?
k=0 k=0 k=Q
54. A certain drug is injected into a body. It is known that the
Σ (⅞ - bk )22 amount of drug in the body at the end of a given hour is ∣
k=0
the amount that was present at the end of the previous
46. A pendulum is swung through an arc o f length 20 cm from
hour. One gram is injected initially, and to keep the
a vertical position and allowed to swing free until it
concentration up, an additional gram is injected at the
eventually comes to rest. Each subsequent swing of the
end of each subsequent hour. What is the highest possible
bob o f the pendulum is 90% as far as the preceding swing.
amount of the drug that can ever be present in the body?
How far will the bob travel before coming to rest?
55. Each January 1, the administration of a certain private
47. A rotating flywheel is allowed to slow to a stop from a
college adds 6 new members to its board of trustees. If the
speed of 500 rpm. That is, it revolves 500 revolutions in
fraction o f trustees who remain active for at least t years is
the first minute, and after the first minute it rotates two-
f( x ) = e^0∙2z, approximately how many active trustees will
thirds as many times as in the preceding minute. How
the college have on December 31, in the long run? Assume
many revolutions will the flywheel make before coming to
that there were 6 members on the board on January 1 of
rest?
the present year.
48. A ball is dropped from a height of 10 ft. Each time the ball
56. How much should you invest today at an annual interest
bounces, it rises 0.6 the distance it had previously fallen.
rate of 15% compounded continuously so that, starting
What is the total distance traveled by the ball?
next year, you can make annual withdrawals of $2,000
forever?
57. Racecourse paradox The Greek philosopher Zeno of
Elea (495-435 B.C.) presented a number of paradoxes that
caused a great deal of concern to the mathematicians of
his period (see Section 1.1, for example). Perhaps the
most famous is the racecourse paradox, which can be
stated as follows:
49. Suppose the ball in Problem 48 is dropped from a height
of /7 feet and each time rises 75% of the distance it had
A runner can never reach the end of a race: As he runs the
previously fallen. If it travels a total distance of 21 ft, what
track, he must first run ⅛the length of the track, then ⅛the
is A?
50. Suppose that nationwide, approximately 92% of all remaining distance, so that at this point he has run 7 + 5
income is spent and 8% is saved. How much additional = ∣ of the length of the track, and ∣ remains to be run.
spending will be generated by a $50 billion dollar tax cut After running half this distance, he finds he still has ∣ of
if savings habits do not change? the track to run, and so on, indefinitely.
51. Suppose that a piece of machinery costing $10,000 depre
ciates 20% of its present value each year. That is, the first
year $10,000(0.20) = $2,000 is depreciated. The second
year’s depreciation is $8,000(0.20) = $1,600, because the
value for the second year is $10,000 - $2,000 = $8,000.
-I------------1------ 1-4-
If the depreciation is calculated this way indefinitely, 1 3 7
what is the total depreciation? 2 4 8 ■"
Sec. 8.3 The Integral Test: p-Series 509
Suppose the runner runs at a constant pace and that 63. Show that ∑ ark diverges if ∣r∣ ≥ 1.
it takes him T minutes to run the first half o f the track. Set k=0
up an infinite series that gives the total time required to n
run the track, and verify that the total time is 2T, as 64. a. Show that ∑ (¾ - ¾ + 2 ) = (α ∣ + a J ~ (α ,,+ ∣ + α ,,+ )∙
2
∕<= 1
intuition would lead us to expect.
b. Use part a to show that if lim ak = A , then
58. S P Y PROBLEM The spy just barely survived the gas in the ∞ - *x
Closet of Doom (Problem 46, Chapter 7 review), but in ∑ (a k - a k+ ι~) = a l + a 2 - 2 A
the process was rendered unconscious and taken prisoner. K—1
He is questioned every 3 hours and if he doesn’t talk, a c. Use part b to evaluate
new dose o f a drug is administered. His sadistic captors ∑ ∖k ' lk - (k + 2)'ifk ^
use the same small dose of drug each time, to prolong the ⅛=ι L
spy’s agony and their fun. The spy knows from previous
experience with this drug that each time a dose is given it d. Evaluate Y —r'—r.
results in a concentration of 0.04% in his bloodstream and k=2 k 2 - 1
that t hours after a dose is given, the concentration in his
bloodstream from that particular dose is given by 65. Square A B C D has sides of length 1. Square E F G H is
0.04e- 0 ∙322z. The spy has a high tolerance for the drug but formed by connecting the midpoints of the sides of the
estimates he will crack when the cumulative concentra first square, as shown in Figure 8.12. Assume that the
tion of drug in his bloodstream is 80% of the ultimate pattern of shaded regions in the square is continued
concentration (as n → ∞). How much time does he have to indefinitely. What is the total area of the shaded regions?
escape?
59. a. Let ∑ ak = S n be the nth partial sum of an infinite
k=∖ "
series. Show that S — S n ~ l = a .
b. Use part a to find the nth term of the infinite series
whose nth partial sum is
c = — -—
» 2n + 3
60. Prove that if Xαz, = A and X⅛t = 5, then

X(αjt + bk } = A + B.
61. Show that if Xα t converges and X⅛z, diverges, then
X(<2A. - bk ) diverges. Hint. Note that bk = ak - (μk - bk ).
62. Find two divergent series 1a k and X⅛t such that: Figure 8.12 Find the area of the shaded regions.
a. ∑(αt + bk ) is also divergent.
b. X(<¾ + bk ) is convergent. 66. Repeat Problem 65 for a square whose sides have length a.
8.3 THE INTEGRAL TEST: p-SERIES
IN THIS SECTION Divergence test; series of nonnegative numbers,

integral test; p-series
In Sections 8.3-8.6, we shall develop “tests” for determining the convergence
of a given series. These tests have the advantage that they do not require
knowledge of an explicit formula (such as the one we found for a geometric
series) for the nth partial sum S n and the disadvantage that they usually cannot
be used to find the actual sum of a convergent series.
The convergence or divergence of an infinite series is determined by the behavior

of its nth partial sum S n . In Section 8.2, we saw how algebraic methods can
sometimes be used to find formulas for the nth partial sum of series. Unfortunate
ly, it is often difficult or even impossible to find a compact formula for the nth
partial sum S n of a series, and other techniques must be used to determine

convergence or divergence.
■ DIVERGENCE TEST
When investigating the infinite series 1a k , it is easy to confuse the sequence of

general terms {αfc} with the sequence of partial sums {5,fl}, where
5 "=
Because the sequence {αjt} is usually more accessible than {Sn}, it would be
convenient if the convergence of {5n} could be settled by examining
lim a,k
k →∞
Even though we do not have one simple, definitive test for convergence, we do
have a first test that tells us that if certain conditions are met, then the series
diverges.
THEOREM 8.9 The divergence test
If the series ⅛ak converges, then the sequence {αn} converges, and then lim ak = 0.
Thus, if lim a k ≠ 0, then the series ∑αfc must diverge.
0 Theorem 8.9 gives a neces
sary but not a sufficient condi Proof: Suppose the sequence of partial sums {Sn} converges with sum L, so
tion for convergence, in the that lim S n = L. Then we also have
n→∞
sense that clouds are necessary lim S *,= L
for rain, but clouds are not suffi n—⅛∞
cient to guarantee rain. 0 Because S k - <S,jt-1 = ak , it follows that

lim ak = lim (S. - S ._ .) = lim S k - lim S._ = L - L = 0
EXAM PLE 1 Divergence test to show divergence
Show that the series

∣ 0⅛+τ⅛diver≡es∙
Solution Taking the limit of the Zcth term as k → ∞, we find
lim k - 300 : 1
k→∞ 4/c + 750 4
Because this limit is not 0, the divergence test tells us that the series must
diverge.
0 The divergence test is often m isinterpreted as saying that 1a k converges

if lim ak = 0, but this is not necessarily true. For example, we will show in
Example 2 that the series diverges even though lim £ = 0. 0
What this says: The divergence test cannot be used to show convergence. It
can only tell us the series diverges if lim a =/=0.
n→∞
■ SERIES OF NONNEGATIVE NUMBERS; THE INTEGRAL TEST
Series whose terms are all nonnegative numbers play an important role in the
general theory of infinite series and in applications. Our next goal is to develop
convergence tests for nonnegative term series, and we begin by establishing the
following general principle.
THEOREM 8.10 Convergence criterion for series with nonnegative terms
A series 1a k with ak ≥ 0 for all k converges if its sequence of partial sums is

bounded from above and diverges otherwise.
Proof: Suppose ⅛ak is a series of nonnegative terms, and let S n denote its «th
partial sum; that is
n
S n = ∑ a,k = a.1 + a,2 + + a
⅛=ι ''
Because .8’n+1 = .S’n + an+,1. and because aκ. ≥ 0 for all k,, it follows that
s 4, l ≥ s
for all n, so that {5 } is a nondecreasing sequence. According to the BMCT, the
nondecreasing sequence {5, } converges if it is bounded from above and diverges
otherwise. Hence, because the series ~Sl ak represents the limit of the sequence {S },
the series ⅛ak converges if the sequence {∑j} is bounded from above and diverges if
it is unbounded. ■
This convergence criterion is often difficult to apply because it is not easy to

determine whether or not the sequence of partial sums {.S’ } is bounded from
above. Our next goal in this section and in Sections 8.4-8.6 is to examine various
tests for convergence. These are procedures that allow us to determine indirectly
whether a given series converges or diverges without actually having to compute
the limit of the partial sums. We begin with a convergence test that relates the
convergence of a series to that of an improper integral.
THEOREM 8.11 The integral test
If ak = f(k) for k = 1, 2, . . . , w h ere/is a positive, continuous, and decreasing

function of x for x ≥ 1, then
∑ a, and
⅛=ι κ
either both converge or both diverge.
Proof: We shall use a geometric argument to show that the sequence of partial
sums {S,π} of the series Y ak is bounded from above if the improper integral
A:=0
∕(x ) dx converges and that it is unbounded if the integral diverges. Accordingly,
Jι
Figures 8.13a and 8.13b show the graph of a continuous decreasing function f that
satisfies/(«) = a for n = 1, 2, 3, . . . . Rectangles have been constructed at unit
intervals in both figures, but in Figure 8.13a the ∕<th rectangle has height
j ∖k + 1) = a k + χ , whereas in Figure 8.13b, the comparable rectangle has height
f(k ) = ak . Notice that in Figure %Λîa, the rectangles are inscribed so that
AREA OF THE FIRST n - 1 RECTANGLES < AREA UNDER y = f(x) OVER [1, n]
cn
a 2 + a 3 + ∙∙∙ + an < f(x) dx
, Ji
rn
a i + a 1 + a 3 + ∙∙∙ + an < α1 + f(x ) dx Add al to both sides.
Ji
Similarly, in Figure 8.13b, the rectangles are circumscribed so that
AREA UNDER y = f(x) OVER [1, n + 1] < AREA OF THE FIRST n RECTANGLES
a. The Ath rectangle has height
∕ α + n = ¾ + lΓ
f(x ) dx a ∣ + n, ÷ a 3 + + an
Let S = α l + a 1 + a3 + + a be the nth partial sum of the series so that
Now, suppose the improper integral ∕( x ) dx converges and has the value /;
that is, ∕( x ) dx = I. Then
b. The A'th rectangle has height

f<F> = ak . It follows that the sequence of partial sums is bounded from above (by a i + /),
Figure 8.13 Integral test
and the convergence criterion for nonnegative term series tells us that the series
7 a must converge,
⅛=ι k
On the other hand, if the improper integral f( x ) dx diverges, then
Ji
rn + 1
Iim f(x) dx = ∞
H→∞Jl
rn + 1
It follows that lim S n = ∞ also because n f(x) dx < 5 ; this means that the
n →∞ Ji
series must diverge. Thus, the series and the improper integral either both
converge or both diverge, as claimed. ■
0 When applying the integral test, the function f need not be decreasing
for all x ≥ 1, only for all x ≥ N for some number N. That is, f must be
decreasing in “ the long run.” 0
x 1
An important senes is y τ>

κ
which is called the harmonic series. In our next
k=ι
example, we use the integral test to show that this series diverges.
EXAM PLE 2 Harmonic series diverges

X 1
Test the series Y -r for convergence.

k≈∖ κ
Solution Because∕( x ) = γ. is positive, continuous, and decreasing for x > 1, the

conditions of the integral test are satisfied.
'b 1
- dx = lim - d x = lim [In b —In 1] = ∞
X b→≈
The integral diverges, so the series diverges.
EXAM PLE 3 Integral test for convergence
∞ Λ-
Test the series ∑ for convergence.
Solution The function ∕( x ) = — = xe λ','5 is positive and continuous for all

x > 0. We find that (
α Computational Window C U B ∕'( x ) = e - ∙v 5 ^ + e~x,5 = (1 -
If you have a graphing calcula

The critical number is found when f ( x ) = 0, so we solve
tor, you can easily see that the
conditions for the integral test
are met. (1^ f)^ 5=0
! - ∣ =° Set each factor equal to 0, but note that
e ^x!5 ψ 0
x = 5
We see that ∕'( x ) < 0 for x > 5, so it follows that / is decreasing fo rx > 5.
Thus, the conditions for the integral test have been established, and the given
series and the improper integral either both converge or both diverge.
Computing the improper integral, we find
X nin=0 Vιηin=0 r∞ rb
Xnax=10 Yroax=3 xe~ x,5 dx = lim xe~ x ' 5 dx
X s c l= l V s c l= l b ⅛→∞j5
Letu=x dv ~ e~x'idx
Although this graph suggests du = dx v = -5e~x'i
that the conditions for the inte
gral test are met, you still must
Γ , ∣⅛ Cb
use calculus to evaluate the in = lim -5 x e x/5L5 - (-5 e x ,5 ) dx
tegral. b→∞L 1 √5 ■
= lim [-S x e ~ xj5 - 25e“ -v/5]l
b→∞ j ∣5
= lim ∖-5 b e~ b li - 15e~ b li + 25<T1 + 25<T1]

b→∞
= - 5 lim '"' ^ t ∕ + lim 50e^,
⅛→∞ e b ,i b→∞
0 Note that the series converg
es because the integral converg
es, but this does not mean that
= ^5 l⅛⅛)^ + 50e^1 ΓH∂pitaΓs rule lim =0
= 50e^1
the series Σ [⅛ eQ
uals
Thus, the improper integral converges, which in turn assures the convergence
50e>~1. 0 of the given series. ≡≡≡
■ p-SERIES
The harmonic series Σ⅛ is a special case of a more general series form called a
p-series.
p-Series A series of the form

y _L = X -L + -L ...
*⅛ k p IP + V p +3
where p is a positive constant is called a p-series. If p = 1, then the p-series is
the harmonic series.
The convergence of the p-series is dependent on the value of p. Remember, we

have already seen (in Example 2) that the series diverges if p = 1.
THEOREM 8.12 The p-series test
∞ 1
The p-series y converges if p > 1 and diverges if p ≤ 1.
Proof: We leave it for the reader to verify that ∕(x ) = is continuous,

positive, and decreasing for x ≥ 1 and p > 1. We know the harmonic series
(p = 1) diverges, and for p > 0, p ≠ 1, we have
dx _ lim C x p dx = lim “=' 1 if p > 1

χP b→∞ Ji ⅛→≈ i P f∞ if 0 1 and diverges if 0 < p < 1. For
p = 0, the series becomes
y ±0 = l l l ...
fc⅛ ∕< 1+ 1+ 1+
and if p < 0, we have lim -X = ∞, so the series diverges by the divergence test
∕c→∞ K-l
(Theorem 8.9). Thus, a p-series converges only when p > 1. ■
HI What this says: The statement that the series converges if p > 1 and
diverges if p ≤ 1 is clear enough. A more subtle question, however, is why?
The curves y = ∖ !xp (p > 0) all decrease as x increases. This also seems clear
enough. The answer we seek, namely why, lies in looking at the rate of decrease
for the curves y = l∕x p . If p > 1, the curve decreases fast enough that the
shaded area in Figure 8.14a is bounded by a fixed number, no matter how large
the value b. On the other hand, if the curve decreases slowly enough, then the
shaded area in Figure 8.14b increases without bound as b → ∞.
Figure 8.14 Graphs for the p-series test

Sec. 8.3 The Integral Test: p-Series 5 15
E X A M P L E 4 p-series test
Test each of the following series for convergence.

a' J“ 1 b'∑≈ ⅛
( ι
^ ⅛ι )∖
1 VF
Solution a. Here V Σ 3 = Zc3z2, so p = ∣ > 1 and the series converges.

, co 1 1 ι ∖
b. Y -7k - ~7= ; We note that
t⅛∖∖e ∖'ld
i ∑ι converges, because it is geometric with ∣r∣ < ∣Λ < 1,

∞ j
and ∑ —τ= diverges, because it is a p-series with p = ⅛< 1.
k—1 v ∕C
Because one series in the difference converges and the other diverges, the
given series must diverge (Theorem 8.7). B M
PROBLEM SET 8 .3
θ 1. ■ What Does This Say? What is a p-series? k 2
29. ∑ 30. ∑ k + 1 31. ∑
k2 + 1
2. ■ What Does This Say? What is the integral test? k + 1 k + 5 k3
For the p-series given in Problems 3 -6 , specify the value o f p 2k i + 3 1 -k 5+ ∕c 2 + 1

32. ∑ 33. ∑ 34. ∑
and tell whether the series converges or diverges. ⅛=ι k5 ⅛=ι V k=l ki + 2
3∙ ” ι y lθ θ
∞ 1
∑
A-=l
⅛ κ
4
J Vk 6 f 35. ∑
n⅛ + 1
„ 2 .7 3 2 ! 36. Σ n„2 .2 3 6
+ 1
Test the series in Problems 7-2 8 fo r convergence. ⅛- I

∞ / o∖At 37. ∑ 38. Σ
V l∏-⅛ k + 2
7. JVi l∏∑
k2 8. Λ k 9. Σ 2 + kτ
k=∖ κ ⅛÷ι ∖ )
J _ , 2fc+ 3
∞ tr 39. ∑ .2 k 40. Σ
1 + ∞ 1 3k + 4.
10. Σ t 11. ⅛ ⅛ 12. J , v r ⅛=ι k=l
*=ι ∖ λ- ∕
Show that the function f determined by the nth term o f each

13. ∑ k ~w 14. Σ ⅛-4z3 15. V k series given in Problems 4 1 -4 6 satisfies the hypotheses o f the
k= 1 k=l fc÷ι k 2 + 1 integral test. Then use the integral test to determine whether
k2 ∞< ∕vL.2 the series converges or diverges.
16. ≠∙ 17. 18. ∑ ke~ k2
⅛÷ι (Zc3 + 2) 2 ⅛1 V k 3 + 2 k=∖ 4i y — i----- 43. f
’ ⅛ι (2 + 3A)2 42. ∑ (2 + k)~ 3'2
∞ 1 ⅛=ι k=2
19. y ι 20. J i 4 fc 21. y ‘
⅛i (0.25)* ⅛÷2 fc(l∏ k) 2 44 y — i— 2 ∞ (tan^ 1 ⅛)2
⅛¾ ∕c(ln k) Ji ι+ ⅛ 2 46. ∑ ke~ k2
∞ 1 k2
⅛=1
24. y
k
22. 23. y k k
⅛⅛2 AV in k k=∖ pe k=l e © F in d all values o f p fo r which the series given in Problems
t a n ' 1 A: ∞ 1 4 7 -5 2 converges.
25. f 26. ∑ cot - 1 k 27. y k 1
t÷ι 1 + k 2 k=∖ ⅛⅛1 e + e^ k y - k- y - V-
∖ ⅛ ( fc 2 - l ∑ 448
8∖ ¾ ( ^ + 4)∕'
28. 7. 3e~ 2k
49 ⅛y ^—l n-—^
4∖ y ' nk p
so ' ∑
50
Use the divergence test and the p-series test to determine which
o f the series in Problems 2 9 -4 0 converge. Write "inconclusive” °° 1 °° 1
51' i∑3 k In fc[ln(ln k)]p 52∙ ∑2 W∑T∑
i f there is not yet enough information to settle convergence.
exactly if p is an even integer. On the other hand, this

Computational Window sum is unknown for p = 3.
a. Use the work of Problem 53a to find N sufficient
In these problems we explore some o f the numerical to give an approximation to the sum with p = 3
implications o f the integral test for series o f the form with an error no greater than 0.0005.
b. Since we do not have an exact answer to compare
with, the following idea is often used. Compute S N
53. a. Argue that the “ tail” of the series (the error after for your N found in part a; then compute S 2 N .
taking N terms) is bound by Compare the results and comment on the confi
dence you have in your work. Was the calculation
V ± < Γ dx
Σ k∣ , ~ , of S , v necessary to give you this confidence?
k=N+l κ - J N Λχi
b. Suppose you want to approximate the sum with
p = 2 by the partial sum S N , SO that the error is no
bigger than 0.01. Use part a to decide how many θ 55. T H IN K T A N K P ROBLEM Find two sequences {at } and {⅛fc}
terms are needed; that is, find N. for which lim ak = 0, lim bk = 0. 'Zak bk converges, and
c. It can be shown that X¾ converges but ∑bk diverges.
x 1 ττ2 56. T H IN K T A N K P ROBLEM Find two sequences { α j and {bk}
I⅛ " ^ 6 ^
for which lim ak = 0, lim bk = 0, t a k bk converges, and
Using the N found in part b, compute ∑ v and 1a k and ⅛bk both diverge.
compare with this answer. How good was your
estimate of N in part b? 57. Show that if the series ∑ ak converges, then
“ ∣ ⅛=ι
rχ x r∞
54. It is a bit surprising that sums ∑ Jf> c a n t,e computed
∕(.v ) dx ≤ ∑ ak ≤ a. + f(x) dx
Ji *=ι Jo
8 . 4 COMPARISON TESTS
IN THIS SECTION Direct comparison test, limit comparison tests

Often a given series will have the same general appearance as another series
whose convergence properties are known. In such cases, it is convenient to use
the properties of the known series to determine those of the given series. The
goal of this section is to examine three comparison tests for making such
determinations.
■ DIRECT COMPARISON TEST

THEOREM 8.13 Direct comparison test
Let 0 ≤ ak ≤ ck for all k ≥ N for some N.
X X
If ∑ ck converges, then ∑ ak also converges.
k= 1 k= 1
Let 0 ≤ d,κ ≤ a,κ for all k.

x x
If ∑ dk diverges, then ∑ ak also diverges.
k= 1 k= 1
Proof: Suppose Σ¾ is a given nonnegative term series that is dominated by a

convergent series 1ck in the sense that for sufficiently large k, 0 ≤ ak ≤ ck for all
k. Then, because the series ⅛ck converges, its sequence of partial sums is bounded
from above (say, by M), and we have
n n
Σ α , ≤ kΣ= I cl < M
k= 1 i κ κ
for all n
Thus, the sequence of partial sums of the dominated series ⅛ak is also bounded
from above by M, and it, too, must converge.
Sec. 8.4 Comparison Tests 517
On the other hand, suppose the given series 3iak dominates a divergent series
∑dk so that 0 ≤ clk ≤ ak . Then because the sequence o f partial sums o f "∑l d k is
unbounded, the same must be true for the partial sums o f the dominating series
X¾ , and Xa k must also diverge. ■
B What this says: Let Xα z, and l d k be series with positive terms. The series
1 a k converges if it is “ smaller” than (dominated by) a known convergent series
1c k and diverges if it is “ larger” than (dominates) a known divergent series X√z,.
That is, “ smaller than small is small,” and “ bigger than big is big.”
EXAM PLE 1 Convergence using the direct comparison test
Test the series ∑ ^k + ι for convergence.

k=∖
Solution For k ≥ 0, we have 3 k + 1 > 3k > 0, and 0 < τ-U—- < X . Thus, the
3λ + 1 3z'
given series is dominated by the convergent geometric series Σ β⅛(convergent
because r = j). Direct comparison test tells us the given series must converge.
EXAM PLE 2 Divergence using the direct comparison test

fc=l
Solution For k > 25, we have ⅛ > 0, so the given series dominates
the divergent p-series ∑ τ∣p (divergent because p = ∣ < 1). The direct
*=ι κ ~
comparison test tells us the given series must diverge.
QB
It is sometimes possible to program your calculator or series convergence or divergence. However, you must
computer to give the partial sums o f the given series. keep in mind that such conclusions are fallible. Con
I f the series does not converge or diverge too slowly, sider the partial sums for Examples 1 and 2.
you may be able to come to a conclusion about the
τz______ ι. £ 1 ∞ 1
L.AC11 UΛdllipiC Δ.
" κ *v i ∖ ⅛ 3 k + 1 *⅛ V k - 5
N F(l) + . . . + F(N)
1 .25 N F(l) + . . . + F(N)
2 .35 1 -.25
3 .3857143 2 -.5288789
4 .3979094 3 -.8348812 Also, notice how you can
5 .4020078 4
6 .4033776 5 —1 Η .5 ,∩
Ro 0 0
ΓΠ O
±o
be misled by/
a computer.
r
7 .4038347 6 -ι.922096 The series in Example 2

8 .4039871 7 -2.34686 does not exist when
9 .4040379 8 -2.807356 /< = 25, so we cannot sum
10 .4040548 9 -3.307356 beyond k = 25.
11 .4040605 10 -3.851508 j
12 .4040623 100 20.10886
13 .404063 200 34.38251 *-
14 .4040632 1018 80.66898
IS .4040632 10018 229.6481
The partial sums se em to converge in Example 1. The part al sums do not seem to converge in Example 2.
EXAM PLE 3 Direct comparison test
∞ 1
Test the series ∑ p- for convergence,
⅛=ι κ-∙
Solution We shall compare the given series with a geometric series. Specifically,
note that if k ≥ 2, then l∕k∖ > 0 and
k∖= k(k - ∖)(k - 2) ∙∙∙ 3 ∙ 2 ■1 ≥ 2 ∙ 2 ∙ 2 - 2 ∙2 ∙ 1 = 2fc" 1
k — 1 terms
Therefore, we have 0 < ⅛ ≤ τ⅛[∙, and because the given series is dominated
by the convergent geometric series τppγ (with r = ∣), it must also
converge. A=l
■ LIMIT COMPARISON TEST
It is not always easy or even possible to make a suitable direct comparison between
two similar series. For example, we would expect the series ∑ 1 ∕(2 * - 5) to
converge because it is so much like the convergent geometric series ∑ l∕2 fc. To
compare the series, we must first note that is negative for k = 1 and k = 2
and positive for k ≥ 3. Thus, if k ≥ 3,
1 1
1k 2i - 5
That is, S 1∕(2A∙' - 5) dominates the convergent series ∑ l∕2 t , which means the direct
comparison test cannot be used to determine the convergence of X1∕(2 A^- 5) by
comparing with X l∕2 ∖ In such situations the following test is often useful.
T H E O R E M 8.14 Limit comparison test
Suppose ak > 0 and bk > 0 for all sufficiently large k and that
α.
Inn -3 = L
k→∞bk
where L is finite and positive (0 < L < ∞). Then
∑αλ, and Σ⅛λ either both converge or both diverge.
Proof: The proof of this theorem is given in Appendix B.
H What this says: We have the following procedure for testing the conver
gence of X<2fc.
Step 1: Find a series ⅛bk whose convergence properties are known and
whose general term bk is “ essentially the same” as ak .
a k
Step 2: Verify that lim τ- exists and is positive.
k →∞ Dk
Step 3: Determine whether ΣZ∖ converges or diverges. Then the limit
comparison test tells us that X<2t does the same.
EXAM PLE 4 Convergence using the limit comparison test
∞ ι
Test the series y —⅛ —
-
- for convergence.
k= I 2 5
Solution Because the given series has the same general appearance as the
convergent geometric series ∑ l∕2 fc, compute the limit
1
2 k - 5 ^)k
lim = hm A/ = 1
∕<→∞ 1 FC→∞ 2 - 5
2A∙
This limit is finite and positive, so the limit comparison test tells us that the
given series will have the same convergence properties as ∑ l∕2 ∖ Thus, the
given series converges. ■■
EXAM PLE 5 Divergence using the limit comparison test
3k + 2
Test the series y for convergence.
k=∖ V k ( 3 k - 5 )
Solution For large values of k, the general term of the given series
3k + 2
ak
V k (3 k -5 )
seems to be similar to
To apply the limit comparison test, we compute the limit

3k + 2
wk V k (3 k -5 ) ,.
hm τ- = hm --------------- = hm 3k + 2 .
∕<→∞ k→∞ 1 k→∞ 3 k - 5
Vk
Because the limit is finite and positive, it follows that the given series behaves
like the series Σ l ∕V k , which is a divergent p-series [p = £). We conclude that
the given series diverges. ■■■
EXAM PLE 6 Limit comparison test
Test the series y

+ 1θθfor convergence.
⅛÷ι - 70
Solution For large k,
ak= ⅞5+-1⅞ seems to behave like bk =
and indeed, we find that
7/c + 100
k li e k,5 (7k + 100)
lim e - 70
k→∞ k k(e k,5 - 70)
e k' 5
In Example 3, Section 8.3, we showed that the series 1kle kli converges, and
therefore the limit comparison test tells us that the given series also
converges.
Occasionally, two series Σ¾ and ∑⅛jfc appear to have similar convergence

properties, but it turns out that
is either 0 or ∞, and the limit comparison test therefore does not apply. In such
cases, it is often useful to have the following generalization of the limit compari
son test.
THEOREM 8.15 The zero-infinity limit comparison test
Suppose ak > 0 and bk > 0 for all sufficiently large k. Then,

a, „
If lim 77 = 0 and ∑o,κ converges, the series za κk converges.
⅛→≈ Ok
d.
If lim 77 = ∞ and ΣZ>Λ diverges, the series za k diverges.
A-→≈ ok
Proof: The first part of the proof is outlined in Problem 62; the second part
follows similarly. ■
EXAM PLE 7 Convergence of the log-power quotient series
∞ l L-
Show that the series ∑ nfcp converges if q > 1 and diverges if q ≤ 1. We call
∕<=ι '
this the log-power quotient series.
Solution We will carry out this proof in three parts:
Case I: (q > 1)
Let p be a number that satisfies q > p > 1, and let
and ⅛ = ⅛
Then
In ck
Cl, k∣
lim 77 = lim ——
⅛→∞ bk k→∞ 1
kp
= lim fe⅛ This is the indeterminate form
k→∞k'l p ∞∕∞ because q - p > 0.
1
= lim ____ k ΓH∂pitaΓs rule
k ~*x (q - p)k q~p ~i
1→∞(∩ —n∖k cl~P ) Because q - p > 0
Because Σl∕Zd, converges (∕ι-series with p > 1), the series converges by the
zero-infinity comparison test.
Case II: (⅛ < 1)

Now, let p satisfy q 0
∞ ∞ |
Because ∑ b k = ∑ 77> diverges (we know p < 1), it follows from the zero-
^=1 co In ∕c
infinity limit comparison test that ∑ t ^⅛ diverges when q < 1.
Case III: (q = 1)
=c ∣n fc
Here, the series is ∑ —κ7—’ which diverges by the integral test:
*=ι
⅛ ⅛ = lim ⅛ ⅛ l ∣ i = 00
X b → ∞ L 2 J∣ι
Let u = In .r;
du = 7 dx.
PROBLEM SET 8 .4
θ The most common series used fo r comparison are given in ' in v, 1
19. ∑
Problems 1 and 2. Tell when each converges and when it k=∖ V k3 + 2 ' a⅛1 V k 2 + i
diverges.
X 21. ∑ 2k2 77 V ∕< ÷ 1
1. geometric series: ∑ rk A=l k4 - 4 ⅛ ∕<2 + 1
k=0
∞ 1 (k + 2)(k + 3)
2. p-senes: ∑∣ 23. ∑ 24 1∑
k=∖ fc7/2 ⅛1 V '2
Each series in Problems 3 -1 2 can be compared to the geomet X

2k + 3 26 y 3 F + 2
72 iς V
ric series, or p-series given in Problems 1-2. State which, and *∙ 2. k 2 + 3k + 2 ∖÷1 k2 + 3k + 2
k=∖
then determine whether it converges or diverges.
∞ X
CO S? 277 V k 28 V 3
3∙ ∑ 4. ∑ 0∙5 fc 7∙ λ
A=l 0
k=0 k=∖ (k + 2)2 A ’ Λ÷ I 4 A + 3
X
5. Σ 1.5A 6. ∑ 2 w2 7Q
2 ^∙ λV 1 30 V 1
A=0 k=0 k=∖

k (k + 2) ⅛1 + 2)(k + 3)
∞
1 X
1 ≈ 1,000
7. Σ 8. l . ⅛ 31. ∑ 32∙ s
A=l k
A= 1 V k 2k ⅞*
∞
1 ⅛ √ i ∣
si∏(k!)∣ “ cos⅛ 3
9. Σ 10. 33. ∑
A=i k 3' 2 k2 z⅛ι V k
k=∖
X X
X
11. Σ 1 12. ∑ e k 2k i + k + 1 336
A=0 k=∖ J73ς. V
2- 6' y V - —k k2 -— 34
a⅛
k=∖ k2 + k2 + 1
Test the series in Problems 13-44 fo r convergence.
k 3-,c' ∙ v In k
∞
1 x 1 317/. V
2, 6∙, V 2 t + 3
13. λ 14. y 2 1 k=∖ 4ki - 5
2
k=∖ k + k k Ξ∖ k + 3k + 2
X
1 Zc~ + 1 40. y sin ∣
15. Σ 16. j ⅛ 39. ∑
k=∖ (k 2 + 2)k 2 kV k
A-=l Vk
X
1 y 1________
Λ1 ∙ 6 k 2 + 2k + 1 y 6k 2 + 2k + 1
17. vλ 18. ⅛⅛ V∕<(k + 1) 4 1 λV 42
0 ∙9(4k 2 + k + 4)
k=∖ √2∕< + 3 k=∖ k l l (4k 2 + k + 4) ÷
4 1 k
52 2 Ch. 8 Infinite Series
43. ∑ 56. For what values of p and q does the series

⅛=ι y ∕F Z 2 ⅜
y (In k f
44. f Vk i⅛ k»
⅛=1 ^∖ 'k'' + 1 '√ P converge?
θ 57. a. Let {α } be a positive sequence such that
lim k p a,
θ Test the series given in Problems 45-52 for convergence.
k →∞ '
exists. Use the limit comparison test to show that X¾
45. 1t -÷ 4 46. Λ In k converges if p > 1.
a⅛1 - 1
b. Show that Xe- t '^ converges. Hint: Show that
” ln(fc + 1) lim k- e~k exists and use part a.
47. 48. ∕÷≠∙, Ink 2k
A ( ⅛ + i) 3 58. Let Xuz be a series of positive terms and let {⅛n} be a
sequence of positive numbers that converge to a positive
y 1
49. y 1
1
50. o9
number. Show that Xαt converges if and only if "Zak bk
z÷ 2 (⅛ + 3)(ln<-) -' λ⅛ (k + 3)(ln A') ∙ converges. Hint: Use the limit comparison test.
- k '>'k £ (1 + k)!k
59. Suppose 0 ≤ cιk ≤ A and bk ≥ k 2 for some number A.
51. ∑ fc<' 52. y
a.
k= ∖ ∕<= 1 Show that Xyi converges.
53. Show that the series
60. Suppose al. > 0 for all k and that X¾ converges. Show
y ⅛+ 3 _ 4 , 5 , 6 + that X— diverges.
fc÷ι (∕c + 3)! 4! 5! 6! ' ak
61. Suppose 0 < ak < 1 for every k, and suppose that Xαfc
converges by using the limit comparison test.
converges. Use a comparison test to show that 1a k 2
54. Show that the series
converges.
i + _L ι ___ 1___ ... 62. Show that if ak > 0, and bk > 0, and
1 ∙ 3 + 1 ■3 ∙ 5 + 1 ■3 ∙ 5 ■7+ + ___ 1__
(2∕< + 1)!
converges. Hint: Compare with the convergent series
Xl∕k!.
then Xαt converges whenever 1b k converges. Hint: Show
55. Use a comparison test to show that that for any ε > 0 there is an integer N for which
l ⅛ - θ l = ⅞ < ε if∕< ≥ Λr
⅛=2 (In ∕c)ln k ∣
⅛ 1 b k
converges. Hint: Use the fact that In k > e2 for sufficient Then show that ak < εbk and complete the proof using
ly large k. the limit comparison test.
8 .5 THE RATIO TEST AND THE ROOT TEST
IN THIS SECTION Ratio test, root test, summary of convergence tests for
positive-term and nonnegative-term series
Comparison tests can be applied to a given series ∑ajt only if ∑αfc strongly
resembles a “ known” series ∑⅛fc, Otherwise it is usually necessary to apply a
convergence test that uses only the properties of Xα4 itself. We have already seen
how the divergence test and the integral test can be used for this purpose, and in
this section we shall develop two additional tests, the ratio test and the root test.
■ RATIO TEST
Intuitively, a series of positive terms Σ ¾ converges if and only if the sequence {ak }
decreases rapidly toward 0. One way to measure the rate at which the sequence {αfc}
is decreasing is to examine the ratio a k + J a k as k grows large. This approach leads
to the following result.
Sec. 8.5 The Ratio Test and the Root Test 523
THEOREM 8.16 The ratio test

Given the series ∑ΛA, with ak > 0, suppose that
The ratio test states the following:
If L < 1, then Xαλ, converges.

If L > 1 or if L is infinite, then Xαfc diverges.
If L = 1, the test is inconclusive.
Proof: In a sense, the ratio test is a limit comparison test in which Xαjt is
compared to itself. We shall use the direct comparison test to show that Xαfr
converges if L < 1. Choose R such that 0 < L < R < I. By the definition of the
limit of a sequence, there exists some N > 0 such that
⅞÷ι L <R for all k > N
⅞
Therefore, for k ≥ N,
a N+∖ <:~ a N^∙
a N+2 < a N+l^ < a N^^
a N+3 < a N+JR^ < a N+i^ < a N^
We now form two series, the geometric series

∞
∑i aN R k = CLN R + a γ R 2 + ∙∙∙ + ak R k + ∙∙∙
which converges because R < 1, and the series
∞
fc?i a Njrk ~ a N+1+ a N+2 ÷ - ’ ÷ a N+k + "'
which also converges by the direct comparison test (because each term is less than
the corresponding term of aN R k ). Thus, 1a k converges, because we can discard a
finite number of terms (∕< ≤ N - 1).
The proof of the second part is similar, except that we choose R such that
a, ,
lim - ^j i = L > R > 1
k→∞ CLfc
and show that there exists some M > 0 so that α vz+fc > aMR k .
To prove that the ratio test is inconclusive if L = 1, it is enough to note that
the harmonic series
1
∑ γ∙ diverges
b with lim = lim — = 1
k=ι k k→∞ a, k→∞ 1
k
and the p-series
“ 1 ■, ,. a k+ ι ,. (k + 1)2 1
∑ 75 converges
b with lim — —= h m --------- = 1
fc=ι k- k→∞ ak k→∞ j_
k2 ■
You will find the ratio test most useful with series involving factorials or
exponentials.
EXAM PLE 1 Convergence using the ratio test
∞ ∩k
Test the series ∑ TT fθ r convergence,
⅛=ι κ ∙
2k
Solution Let a k = and note that
2 fc+1
a ... . (k + 1)! ,. ∕<∣7<∙+ι .. 2
L = 1 m - ^ 1 = hm - - -k - - = hm zΛ z n π ⅛ = hm 7 ÷ r = 0
k→∞ ak k→∞ 2 ⅛→≈ ( k + l ) ! Z k→∞k + 1
If
Thus L < 1, and the ratio test tells us that the given series converges, M B
EXAM PLE 2 Divergence using the ratio test
∞ L-∕c
kk
Solution Let a k = and note that
(k + l) fc+1
,. a k + l (⅛ + l)! ,. k ! ( k + l )7λ + l ,. ( k + l ) k ,. ∕ ιY
Lr = hm -------= hm —— , , k 7 = hm ∣∣ i ,, 1 ,, = hm-— ττ—k = hm 1 +ι f = e
k→∞ a k k →∞ k k →∞ k'∖k + 1)! ⅛ →∞ k A-→ ≈ ∖ k)
k∖
Because L > 1, the given series diverges. M B
Computational Window “1
It is instructive to compare and contrast two different A second procedure is to use the ratio test. In this
ways you can use technology to help you with prob ( 1Y c
case, we are interested in hm 1 + τ . We can look at
lems like Example 2. First, as you have seen, you can k →∞ ∖ rC∕
use a calculator or a computer to calculate the partial ( 1V
terms of the sequence ¾ = I 1 + j-1 to see if this limit is
sums directly. This table is shown here on the left.
greater than, less than, or equal to 1. This is shown below.
y kk
*⅜ι k ∖
N F (l) + . . . + F(W)
1 1 Y = ( l + l ∕X ) x
8 3 y
X = 1000000
3 7.5
4 1 8.16667 2 .5 - e
X Y •
•
5 4 4 .2 0 8 3 3 1 2 2 .0 - - ∙
6 1 09.0083 2 2 .2 5
3 2 .3 7 0 3 7
7 2 7 2 .4 0 9 7 4 2 .4 4 1 4 0 6 1 .5 -
8 6 8 8 .5 1 1 3 5 2 .4 8 8 3 2
6 2 .5 2 1 6 2 6
9 1 7 5 6 .1 3 8 7 2 .5 4 6 5 1 .0 -
8 2 .5 6 5 7 8 5
10 4 5 1 1 .8 7 9 2 .5 8 1 1 7 5
10 2 . 593742 0 .5 -
11 11659.53 100 2 .7 0 4 8 1 4
12 3 0 2 7 3 .4 6 1000 2 .7 1 6 9 2 4 1 11 11 11 11 11 11 11 11
10000 2 .7 1 8 1 4 6 1
13 789 1 2 .3 100000 2 .7 1 8 2 6 8 1 2 3 4 5 6 7 8 9 X
14 2 0 6 3 7 5 .3 1000000 2 .7 1 8 2 8 1
15 5 4 1 2 3 9 .9
We see that the limit o f the ratio o f consecutive terms
The partials sums seem to diverge. o f (ak + 1)/'ak is greater than 1.
EXAM PLE 3 Ratio test fails
Test the series Y” , 1_ for convergence.

⅛=2 α jq
Solution Let ak = L 3and find

1
,∙ 2(⅛ + l)-3
Lτ = lι m ---------------= lim y r— γ
As→∞ [ k→∞ 2k - 1
2 ∕c - 3
The ratio test fails. We can use the integral test or the limit comparison test to
test convergence. Note that
J” . 1_ , is similar to the known divergent series ” y1t

Y
⅛=2 7 -, k=2
so we suspect the given series is divergent.

EXAM PLE 4 Convergence of a series of power functions
Find all x > 0 for which the series ∑ k 3x k converges.

k= l
Solution Applying the ratio test,
(fc ÷ l) 3x fc+1 (⅛ + l) 3
k ix k k3 X
(k + 1)3
The series converges when lim -—,k'
-, x < 1, which is true whenever x < 1,
k ,. (fc+l)≈ ,
because khm
→∞
— k i — = 1. BHM
COMMENT: There are more-refined versions of the ratio test that can be used
to handle certain series where L = 1. A consideration of such tests is delayed
until advanced calculus.
■ ROOT TEST
Of all the tests we have developed, the divergence test is perhaps the easiest to
apply, because it involves simply computing lim a. and observing whether or not
that limit is zero. Unfortunately, most “interesting” series have lim a, = 0, so the
divergence test cannot be used to determine whether they converge or diverge.
However, the following result shows that it may be possible to say more about the
convergence of X¾ by examining what happens to w ζ as k → ∞. This test will
prove particularly useful with a series involving a ∕<th power.
THEOREM 8.17 The root test

∕ζ __
Given the series 1a k with ak ≥ 0, suppose that lim ∖ ak = L. The root test states
the following:
If L < 1, then ⅛ak converges.
If L > 1 or if L is infinite, then ⅛ak diverges.
If L = 1, the root test is inconclusive.
Proof: The proof of this theorem is similar to that of the ratio test.
EXAM PLE 5 Convergence with the root test
∞ 1
Test the series 7j—ττ⅛ for convergence.
Solution Let a,k = - —τ-τ and note that

(In kp
L = lim ⅜=J⅛V(ln kyk = ⅛ h⅛= °

Because L < 1, the root test tells us that the given series converges.
EXAM PLE 6 Failure of the root test
∞ 1
Test the p-series ∑ -∏ p for convergence,
⅛=ι κ -
Solution We know that the p-series converges for p > 1 and diverges for p ≤ 1.
The point of this example is to show that the root test gives £ = 1 for all
jp-series.
This confirms the statement that the root test is inconclusive for
£ = 1. Let ak = l∕k p so that
L = lim
To find this limit, use logarithms:

plk p,k
- p In k
In £ = In lim k = lim In k = lim k
This is of the form ∞∕∞, and we apply ΓHδpitaΓs rule.
In £ = ∕lim = k→
lim∞ 0
<→χ K ∣ = 0 so that £ = e = 1
Thus, £ = 1 for any /^-series, yet those with p > 1 converge, whereas those
with p ≤ 1 diverge. t=>
■ SUMMARY OF CONVERGENCE TESTS FOR POSITIVE-TERM

AND NONNEGATIVE-TERM SERIES
This concludes our study of basic convergence tests for series of positive terms.
There are no firm rules for deciding how to test the convergence of a given series
~Zak , but we can offer a few observations.
In situations where the ratio test and the root test both apply, it is often easier
to use the ratio test. However, the root test turns out to be more discriminating in
the sense that any series whose convergence can be determined by the ratio test
can also be handled by the root test, but the root test can be used to determine the
convergence of certain series for which the ratio test is inconclusive (see Problem
56). For this reason we suggest that you try the ratio test first, and if that fails then
try the root test. All tests (except for the ratio test) can also be applied for
nonnegative terms, as well as for positive terms. These procedures are summarized
in Table 8.1.
TABLE 8.1
Guidelines for Determining Convergence of the Series Xu, for ak > 0
Series with known convergence properties

∞
Geometric series, y a r k diverges if ∣r∣ ≥ 1 and converges if ∣r∣ < 1,
k= 1
with sum 5 =
1 —r
^series, ∑ ⅛ converges if p > 1 and diverges if p ≤ 1
(special case p = 1)
∞ 1
Harmonic series, ∑ y diverges (special case, p = 1, of a p-series)
⅛=ι K
Telescoping series, ∑ (bk - ⅛λ + 1 ) converges if lim b.κ ++ . = L with sum S = b.1

k=∖ k→ ∞
- L
1 L-
Log-power quotient series, ∑ -yy- converges if q > 1 and diverges if q ≤ 1
k= ι K ,
Review of Convergence Tests, Presented in a Suggested Order for Checking
1. Divergence test Compute lim ak . If this limit is not 0, the series diverges.
2. Limit comparison test Check to see whether ∑ak is similar in appearance to a
series ∑(r,. whose convergence properties are known,
and apply the limit comparison test. If lim ak ∕bk = L
where L is finite and positive, then ∑ak and X⅛A either
both converge or both diverge.
, 1
3. Ratio test If ak involves fc!, k p , or a , try the ratio test, lim ----- = L.
k→ ∞ a k
Converges if L < 1, diverges if L > 1, fails if L = 1.
4. Root test If it is easy to find lim Va k = L , try the root test.
Converges if L < 1, diverges if L > 1, fails if L = 1.
5. Integral test If/ is continuous, positive, and decreasing, and ak = ∕(fc)
for all k, then ∑ a.κ and the improper integral
k=∖
Γx
f(x) x d both converge or diverge. Think of using
Jl
this test if/ is easy to integrate or if ak involves a
logarithm, a trigonometric function, or an inverse
trigonometric function.
6. Direct comparison test If 0 ≤ ak ≤ ck and ∑ ck converges, then ∑ a k
converges. ^
If 0 ≤ dk ≤ ak and ∑ dk diverges, then ∑ ak
diverges.
^k
7. Zero-infinity limit If lim = 0 and X⅛A, converges, then the series Xαj,
comparison test
converges.
a k
If lim y = ∞ and X ∑ diverges, then the series Xαi, diverges.
If a test is inconclusive, do not quit; try another test.

EXAM PLE 7 Testing for convergence

Computational Window Π H
π, . the
., ∙ v' V3∕∙c — 1 In k
The divergence test, of course,
Test series Y — , , , , 9λ —
is the first test. You can often *÷1 k(k + 2)
check this using a graphing cal
culator. Solution We would expect the given series to behave like the series
y V⅛ In k _ y In k
⅛≡,ι k *2 k=∖ k il2
We apply the limit comparison test by considering the following limit:
y∕îk — 1 In Zc
u> k(k + 2) ∕C3Z2V 3 F Γ T
lim -r- = lim TnT lim
k → ∞ D∣ζ k→ ∞ k(k + 2)
Vι Ξ<-Γ<3X-1) In X) £3/2
z<XCX+2))
Xnin=Θ V n in = " .5 1
Xnax=25 Vmax=.5 V ⅛ √ 3 F 7 T A: _ lim
lim
X s c l= 5 V ≥ c l= . 1 k+ 2 1 ∕√→χ
k
It appears that lim a k = 0, so
this test fails, k →x Thus, by the limit comparison test we see that both series either converge or
diverge (because ∖z 3 is finite and positive). Because the comparison series
∑(ln k)lk 3*'2 is a log-power quotient series with q = 3/2, it must converge
(because q > 1), and thus the given series also converges. ■■
EXAM PLE 8 Testing for convergence
∞ ∕c ι II 2, 31
Test the series JΓ + 1 ∙ 4 ∙ 7 ∙ 10 +
1 . 4 . 7 ...’ ( 3 ^ + jy = 1 + γ∑4 + 1 .4 .7
for convergence.
Solution Let ak - 1 .4 .7 .^ .' tfk ~+ iy B e c a u s e ak involves kl, we try the ratio

test. ' , ( 1
(Ar + 1)!
¾÷, _ 1 ∙ 4 ∙ 7 [3(A + 1) + 1] _ ( ⅛ + l ) ! ∙ l ∙ 4 ∙ 7 ∙ ∙ ∙ ( 3 ⅛ + l ) = k + i
a, k∖ k∖∙ 1 ∙4 ∙ 7 (3∕c + 4) 3A + 4
1 ∙4 ∙ 7 (3∕c+ 1)
Thus,
< 7 .,,
L = lim ~j l l 1 = lim k + i _- 1
⅛→ ∞ d , n → ∞ 3k + 4 3
and since L < 1, the given series converges.
PROBLEM SET 8 .5
θ 1 . 1 What Does This Say? Compare and contrast the ratio 6 ∙ f∞ ⅛
τk y k 8'
∞ ')k
test and root test.

7.
4⅛ 2 t' Σ⅛
2. ■ What Does This Say? Outline a procedure for deter Σ e -i
” /a∖t
9. 10. 11.
mining the convergence of 'S,ak for ak > 0. k=∖ k=∖ ∖j ∕
Use either the ratio test or the root test to test for convergence o f 12.
∞
Σ k / n ∖ A'
4)
13.
Σ { 1k )
14.
V ⅛102A'
the series given in Problems 3-26. k=1 ∙h ∕ k=∖ S ^ J
x nk
4. Σ ⅛ y J< L k
3 ∙ Σ ∏ 5 y — 15. 16. 17.
t÷, k'. A '= l
' ⅛, 2« t⅛U +
Sec. 8.6 Alternating Series; Absolute and Conditional Convergence 529
y 3⅛ + 1
19. y k∖ 5
20. y k + 100
y (x + 0.5)*' (3 x + 0 .4) fc
18. ¾
i 2k ⅛÷1 ( * + 2) 4 4i k∖
47.
*=ι kV χ
48. y k2
k=∖
y (∑ ∑ y (^ )2 ∞ Yk x 2k
21. ¾ (2⅛)!
fc
22. ∑ k 2 2~ k
k=l
23.
i ÷ 1 [(2∕<)!]2
49. so. y k
k=∖
y (^ 2)! y ( k ∖k X X
24. y k 4 3~k
k=∖
25.
⅛÷ι [(2k) 2 ]!
26. ⅛ Λ 2 k + ι y 51. y (a x )k for a > 0 52. y k x 2k
k=∖ k=∖
θ 53. Use the root test to show that ⅛kp e~k converges for any
Test the series in Problems 27-44for convergence. Justify your
fixed positive real number p. What does this imply about
answers (that is, state explicitly which test you are using).
∞ 1,000 the improper integral
y 5,000 5∕< + 2
27. Σ ■
k=l k
28. ⅛¾ k V k 29. Σ
k=l k2 k x pe x dx ,>
y (⅛!∑ y v⅛∣ 3k + 5 k
30.
t⅛ kk
31. 2-,
k=l
Ok
z
32. Σ k2>k
54. Consider the series y 2 - t + <~1l . What can you conclude
k=∖ ⅛=ι
y 2⅛ y 2⅜ fk + 1 about the convergence of this series if you use each of the
33. ⅛¾ k k 34. ⅛ kk 35. Σ
k=∖
fck+0.5 following?
a. the ratio test b. the root test
1 y k∖ 21,OOOA-
36. V - k 37. 38. λv k,2
55. Let {ak} be a sequence of positive numbers and suppose
kΞ∖ 'k ÷
i 1 (⅛ + l)! k=∖ k
that lim = L, where 0 < L < 1.
X ( + r ) -⅛2
1 y (k + i ∖^k l t →≈ ak
cos k∖
39. Σ 1
k=∖
40. ^- ι ∖ k /
k=∖ X κ '
41. Σ
k=∖
2k ∖ a. Show that lim ak = 0. Hint. Note that ∑αz. converges.
A,∙→≈
∞ sin k∖ y ΛnkY b. Show that lim 77 = 0 for any
3 x > 0.
1 ⅛→≈ k∖
42. λ
k=∖
1 k 2>
43. A l
k=2 X
lr
K /
/ 44. Σ
k=2 (In k) k
56. Consider the series
θ Find all x > 0 for which each series in Problems 45-52 11 ÷ - ÷ f + f ÷ - + ∙∙,
+ -2 +
+ - f +÷ - +
2 4+ 4+ 8+ 8+ 8+
converges.
45. f k 2x k 46. ∑ kxk a. Show that the ratio test fails.
λ-=ι k=l b. What is the result of applying the root test?
8.6 ALTERNATING SERIES; ABSOLUTE AND CONDITIONAL CONVERGENCE
IN THIS SECTION Leibniz’s alternating series test, error estimates for

alternating series, absolute and conditional convergence, rearrangement of
terms in an absolutely convergent series
We now turn our attention from nonnegative term-series to series with both
positive and negative terms. We begin by examining series whose terms
alternate in sign. We shall examine conditions that guarantee the convergence
of such alternating series and then discuss the speed of convergence. Finally, we
make a few general observations about series whose terms are neither all
nonnegative nor strictly alternating in sign. This leads to a consideration of
absolute and conditional convergence.
There are two classes of series for which the successive terms alternate in sign, and
each of these series is appropriately called an alternating series:
odd terms are negative: y ( - l ) ⅛ , = - α 1 + a 7 - a 3 + ∙∙∙

k=i
even terms are negative: y ( - l) z-'+ 1 α, = a. - tz7 + α, -

⅛=ι
where in both cases ak > 0.
■ LEIBNIZ’S ALTERNATING SERIES TEST
In general, just knowing that lim ak = 0 tells us very little about the convergence
properties of the series X¾, but it turns out that an alternating series must
converge if the absolute value of its terms decreases monotonically toward zero.
This fact was first proved in the 17th century by Leibniz and may be stated as in
the following theorem.
THEOREM 8.18 Alternating series test
If ak > 0, then an alternating series

∑1( - ! ) X or ∑ ι ( - i ) fc+1¾
converges if both of the following two conditions are satisfied:
1. lim a.κ = 0
k→∞
2. {ak} is a decreasing sequence; that is, ¾ +1 < ak for all k.
Proof: We will show that when an alternating series of the form
∑ , ( - ι r ,¾
satisfies the two required properties, it converges. The steps for the other type of
alternating series are similar. For this proof we are given that
lim a.k = 0 and α,..
k →∞ k+1
< a.k for all k
We need to prove that the sequence of partial sums {5n} converges, where
S n = ∑ι (-V ) k+ l ak = a l - a2 + ai - a4 + ■■■ + ( - l) " + l a n
Our strategy will be to show first that the sequence of even partial sums {S2 } is
increasing and converges to a certain limit L. Then we shall show that the
sequence of odd partial sums {<S,2n. 1} also converges to L.
To understand why we would think to break up S n to look at the even and odd
partial sums, consider Figure 8.15. Start at the origin; and because a l > 0, S 1 will
be somewhere to the right of 0 as shown in Figure 8.15a. We know that a2 < a i
(decreasing sequence) so S, is to the left of Sj but to the right of 0 as shown in
Figure 8.15b. As you continue this process you can see that the partial sums
oscillate back and forth. Because a → 0, the successive steps are getting smaller
and smaller, as shown in Figure 8.15c.
Note that the even partial sums of the alternating series satisfy
, a ∣ a2
S4 = (<21 - α2) + (α3 - α4)
^2 n - *⅞) L '" L (¾∏-1 ^^ ¾∏)

Because {ak} is a decreasing sequence, each of the quantities in parentheses
(β,7- 1 - α2√) ' s positive, and it follows that {*S,2zι} is an increasing sequence:
Figure 8.15 Alternating series test s2 < .s,4 < s6 < -
Moreover, because
(^ 2
$2n fl l (¾ T>) (f l4 ^5) ” ' (¾ n -2 ¾ Λ -1 ) ^2n <^
the sequence of even partial sums {5,n} is bounded above by α l (see Figure 8.15c),
and because it is also an increasing sequence, the B M CT tells us that the sequence
must converge, say, to L; that is,
lim S , n = L
n →∞ -
Next, consider {S2n~l }, the sequence of odd partial sums. Because S 1 -
¾ π-ι = a a r ,d li m a = θ> it follows that S,2JJ- 1 = S 2n - an and
n n
lim 5 , , = lim S∖ - lim a = L - 0 = L

n→∞ 2n~ *i n →∞ 2n n→-<- n
Since the sequence of even partial sums and the sequence of odd partial sums both
converge to L , it follows that {5,H} converges to L , and the alternating series must
converge. ■
R E M A R K : When you are testing the alternating series Σ ( - l) k + 'a k for conver
gence, it is wise to begin by computing lim ak . If this limit is 0, you can show
that the alternating series converges by verifying that {ak} is a decreasing
sequence. However, if lim a.κ ≠ 0 you know immediately that the series
k→∞
diverges and no further computation is necessary.
E X A M P L E 1 Convergence using the alternating series test
≈ ( - l ) fc
Test the series ∑ —-,κ— for convergence. This series is called the alternating
k=∖
harmonic series.
Solution The series can be expressed as ∑(~l)⅛ k , where ak = γr - We have

lim 7- = 0 and because
k→∞k
1 . 1
(k + 1) k
for all k > 0, the sequence {ak} is decreasing. Thus, the alternating series test
tells us that the given series must converge. ≡β
E X A M P L E 2 Using l, H6pitaΓs rule with the alternating series test
_ , . ≈ ( - l) i+ 1 In k c
Test the series ∑ ------- -------- for convergence.
Solution Express the series in the form ∑ (- l) k + 'a k , where ak = We begin

with the divergence test. (Notice that the graph in Figure 8.16 shows that the
limit of the ∕<th term as k → ∞ seems to be 0.) For ak > 0 where k > 1 and by
applying l , H6pitaΓs rule, we find that
1
In lc
Figure 8.16 Graph of y = -
It remains to show that the sequence { α j is decreasing. We can do this by

setting∕ ( Λ^) = and noting that the derivative
rω =⅛ b j Ji
satisfies f'( x ) < 0 for all x > e. Thus, the sequence {ak} is decreasing for all
k > 3 > e. The conditions of the alternating series test are satisfied, and the
given alternating series must converge. BOB
EXAM PLE 3 Divergence using the alternating series test
≈ ∕- ] U + l
Test the series ∑ -----_ . , for convergence.
⅛=∣ tan k
Solution The series can be expressed as Σ ( - l ) k+1 ¾, where ak = t n~-Γ^' The

1 is shown in Figure 8 .17. It seems that the limit as x → ∞ is
graph of j-, =
tan 1 k
not 0, so we suspect the series diverges. We can show this as follows:
lim a, = lim 1 l= 2 ≠ 0
∕<→≈ * k→∞ .tan -1 k, 77 77
2
■' tan 1 k Thus, the series diverges by the divergence test.
∞ (^~lT+ l
Alternating p-Series The series — is called the alternating p-series.
The following example shows that the alternating jp-series converges for all p > 0.
EXAM PLE 4 Alternating /^-series
X / — ] V<+1
Prove that the alternating /^-series ∑ — — converges for p > 0.
Solution Let ak = p-, and note that lim ak = 0 for p > 0. To show that the
sequence {ak} is decreasing, we find that
1
⅞ +ι = + iχ = ∣
CP
ak X (λ-+iχ
kp
so ak > ak + l . Thus, the alternating /^-series converges for all p. ≡≡≡
■ ERROR ESTIMATES FOR ALTERNATING SERIES
For any convergent series with sum L , the «th partial sum is an approximation to
L that we expect to improve as n increases. In general, it is difficult to know how
large an n to pick in order to ensure that the approximation will have a desired
degree of accuracy. However, if the series in question satisfies the conditions of

the alternating series test, the absolute value of the error incurred by using the sum
for the first n terms to estimate the entire sum turns out to be no greater than the
(« + l)th excluded term. More formally, we have the following theorem.
T H E O R E M 8.19 The error estimate for an alternating series
Suppose an alternating series

∑ (-1 )S or Σ ( - 1 ) A'÷1¾
satisfies the conditions of the alternating series test; namely,
lim ak = 0 and {ak} is a decreasing sequence (αfc+1 < ak )
I f the series has sum 5, then
∣ ^ - ∑ 1∣ < Σ,÷ 1
where S rι is the wth partial sum of the series.
Proof: We will prove the result for the second form of the alternating series
and leave the first form for the reader. That is, let
S = ∑ (-1 Γ ⅛ and S ι, = ∑ ( - 1 ∕'÷ ⅛
Begin with A —S n :
S -S ιι= Σ ( - l Γ ' ¾ - Σ ( - l ) fc÷'a k = Σ ( - l Γ 1¾

κ—1 K—1 K.—∕<^∣^1
= ( - ι r χ +1 + ( - ιr χ , +2 + ( - ι r χ +3 + ∙∙∙
= (-ιr(-i)⅛ n+1 + (-i)"(-i)X +2 + (-i)"(-i)‰ + -
= ( -i) ‰ 1 -‰ + ‰ -‰ + ••■]
= (-l)" [α n+1 - (α,,+2 - ‰ ) - K ÷ 4 - ‰ ) - - ]
Because the sequence {απ} is decreasing, we have ak - αt+1 ≥ 0 for all k, and it
follows that
∣s - ∑,I = ∣αn+1 - (an+2 - α i+3 ) - (⅛+4 - an + 5 ) - ∙∙∙ ∣≤ an+χ
because every quantity in parentheses is positive. ■
M What this says: If an alternating series satisfies the conditions of the

alternating series test, you can approximate the sum of the series by using the
«th partial sum (5 ), and your error will have an absolute value no greater than
the first term left off (namely, αn + ∣).
EXAM PLE 5 Error estimate for an alternating series
Consider the convergent series ∑ — — .

⅛=ι k
a. Estimate the sum of the series by taking the sum of the first four terms.
How accurate is this estimate?
b. Estimate the sum of the series with three-decimal-place accuracy.
Solution
a. Let a, = ⅛ and let .8' denote the actual sum o f the series. The error estimate
κ k
tells us that
∣S - S 4 ∣ ≤ a 5
where ∙S'4 is the sum o f the first four terms o f the series. Using a calculator,
we find
¾ = p -⅜ + ⅜ -⅜ -0 M 5 9 3 9 4
and α 5 = y 4 = 0.001 6. Thus, if we estimate X by X'4 ≈ 0.945 9, we incur an

error o f about 0.001 6, which means that
0 Estimating the remainder is
∣S, - S 4 ∣ ≤ 0.0016
pointless if you are not careful
in rounding. 0 0.945 9 3 9 4 -0 .0 0 1 6 ≤ S ≤ 0.945 939 4 + 0.001 6
0.9443394 ≤ 5 ≤ 0.947 5394
So, we have S', correct to two decimal places, as 0.94.
b. Because we want to approximate 5 by a partial sum S n with three-decimal-
place accuracy, we can allow an error o f no more than 0.0005. The error is
measured by
a 'l ' = (n + 1)4
Thus, we wish to find n so that
0.000 5 ≤ (" + 1 )4
V z2,000 ≤ n + 1
6.687403 - 1 ≤ n
This says that n must be an integer greater than 5.687 403; that is,
n ≥ 6. ⊂~~)
■ ABSOLUTE AND CONDITIONAL CONVERGENCE
The convergence tests we have developed cannot be applied to a series that has
mixed terms or does not strictly alternate. In such cases, it is often useful to apply
the following result.
THEOREM 8.20 The absolute convergence test
A series of real numbers Xα fc must converge if the related absolute value series X∣<¾J
converges.
Proof: Assume ∑∣αj converges and let bk = a k + ∣αfc∣ for all k. Note that
, _ f2∣¾∣ if ak > 0
k [0 if ak ≤ 0
Thus, we have 0 ≤ bk < 2 ∣αi ∣ for all k. Because the series X∣αJ converges and both
1,bk and X∣αJ are series of nonnegative terms, the direct comparison test tells us
that the dominated series X¾ also converges. Finally, because
a k = b k ~ ∖a κ∖
and both ⅛bk and X∣αfc∣ converge, it follows that ∑<y, must also converge. ■
EXAM PLE 6 Convergence using the absolute convergence test
Test the series l+ ⅜ + ⅛ - ⅛ - ⅛ + ⅛ + ⅛ + ⅛ - ⅛ - -⅛θ ” ■for conver

gence. This is the series in which the absolute value of the general term is
and the pattern of the signs is + + H-------+ + H-------- ∙∙∙.
Solution This is not a series of nonnegative terms, nor is it strictly alternating.
However, we find that the corresponding series of absolute values is the
convergent jp-series
∣1 -+∣-i-ι-i-∣__
4 + 9 + 161__+∣-... = ⅛
y —
L2
a 1 yt
Because the absolute value series converges, the absolute convergence test
assures us that the given series also converges.
EXAM PLE 7 Convergence of a trigonometric series
• cc si n k r*
Test the series ∑ -τy∣~ for convergence.
k= i 2
Solution Because sink takes on both positive and negative values, the series
cannot be analyzed by methods that apply only to series of nonnegative terms.
Moreover, the series is not strictly alternating.
The corresponding series of absolute values is
sin k
Σ
k=∖ 2k
which is dominated by the convergent geometric series X l ∕2 fc because
∣sin k∖ ≤ I for all k; that is,
sin k ≤ for all k

2k 2k
Therefore, the absolute convergence test assures us that the given series
converges. The graph of the related continuous function f(x ) = (sinx)∕2 x is
Figure 8.18 Graph of y = shown in Figure 8.18. ⊂ Ξ□
If Xαi , converges, then X∣tzJ may either converge or diverge. The two cases that
can occur are given the following special names.
Absolutely Convergent Series The series 1a k is absolutely convergent if the related series X∣¾∣ converges.
Conditionally Convergent Series The series X¾ is conditionally convergent if it converges but X∣αJ diverges.
For example, the series in Examples 6 and 7 are both absolutely convergent, but
the alternating harmonic series
” ( - l ) zc+1
t⅛ι k
is conditionally convergent because it converges (see Example 4), whereas the

related series of absolute value terms X∣. (the harmonic series) diverges.
The ratio test and the root test for nonnegative-term series can be generalized
to apply to arbitrary series. The following is a statement and proof of the
generalized ratio test. The proof of the generalized root test is left to the problems
(see Problem 60).
THEOREM 8.21 The generalized ratio test
For the series ^Stak , suppose ak ≠ 0 for k ≥ 1 and that
where L is a real number or ∞. Then:
If L < 1, the series ∑ak converges absolutely and hence converges.

If L > 1 or if L is infinite, the series Xαfc diverges.
If L = 1, the test fails.
Proof: Assume L < 1; then the positive series X∣αJ converges by the ratio
test, and hence Xαjl converges absolutely.
Assume L > 1; then the sequence {∣αj} is eventually increasing, which means
that {<2,.} cannot converge to 0 as k → ∞. Thus Xa i diverges by the divergence
test.
Assume L = 1. See Problem 61 for this part of the proof.
EXAM PLE 8 Convergence and divergence using the generalized ratio test
Find all values of x for which the series 1 k x k converges and all x for which it
diverges.
Solution Let ak = kx k .
a k+∖ (k + l)x 7c+1 k+l
lim = lim = lim f- '-t 12x
j = 1⅛(1 + 0 x ∣= w
k→ ∞ <2κ
, ∕<→ ∞ kx k k→ ∞ ∖ K / x k
Thus, by the generalized ratio test we see for ∣x∣ = L that the series converges
for ∣x∣ < 1 and diverges if ∣x∣ > 1. If x = 1, the series is X⅛(l)fc, which clearly
diverges by the divergence test. Similarly, if x = - 1 , XZc(- l) k diverges by the
divergence test. Thus, the series converges for ∣.v∣ < 1 and diverges for
∣X∣ ≥ 1. ≡B
Additional Guidelines 1. Test for absolute convergence Take the absolute value of each term and
for Determining Convergence proceed to test for convergence of positive series, as summarized in Table
of Series 8.1.
1. i f S K I converges, then the alternating series will converge.
2. If X∣αfc∣ diverges, then the series is not absolutely convergent, but the
alternating series might still converge conditionally.
2. Alternating series test For¾ > 0, the series converges if ∣αfc+1∣ < ∣αj for
all k and ⅛→limχ a, = 0. k
Sec. 8.6 Alternating Series; Absolute and Conditional Convergence 53 7
■ REARRANGEMENT OF TERMS IN AN ABSOLUTELY

CONVERGENT SERIES
You may be surprised to learn that if the terms in a conditionally convergent series
are rearranged (that is, the order of the summands is changed), the new series may
not converge or it may converge to a different sum from that of the original series!
For example, we know that the alternating harmonic series
∞ ( - l) * ÷ 1
*≡i k
converges conditionally, and it can be shown (see Problem 52)
1 _ 12 +' 3 1 _ ±4 +' 5 1 _ 16
1 ++ 1 _ 18 ++ 91 ____ -- imn ?2
7
If we rearrange this series by placing two of the subtracted terms after each added
term, we obtain
1 2 4 + 3 6 l + l- ... = ( l- l) - l÷ ( l- * ) - l+ ...
- 2 4 + 6 8
=⅛ 4 + M + ∙∙∙)
= ∣ In 2
In general, it can be shown that if ~∑,ak is conditionally convergent, there is a

rearrangement of the terms of t a k so that the sum is equal to any given finite
number. In Problem 53, for example, you are asked to rearrange the terms of the
series
∞ ( - l) * ÷ 1
k=∖ k
so that the sum is jin 2.

This information may be somewhat unsettling, because it is reasonable to
expect a sum to be unaffected by the order in which the summands are taken.
Absolutely convergent series behave more in the way we would expect. In fact, i f
the series t a k converges absolutely with sum S, then any rearrangement o f the terms
also converges absolutely to S. A detailed discussion of rearrangement of series
requires the techniques of advanced calculus.
PROBLEM SET 8 .6
θ 1. ■ What Does This Say? Discuss absolute versus condi ≈ ( - ι Γ '⅛ “ ( - l ) fc÷1k 2
tional convergence.
6 ' ∑ 2 ∕< + l 7.
⅛ k2 + 1
2. ■ What Does This Say? Discuss the convergence of the “ (-[U + l
alternating p-series. ” ( - l) * +1 ⅛
8 ∙ J, 9.
3. ■ What Does This Say? Look at Problems 4-31 and ⅛ι
notice that most starting values are k = 1, but some have ” (—i) fc
10. f (-l)⅛÷>g 11.
starting values k = 2. Why do you think we did not use k=l e ⅛≡) F∑
k = 1 for all these problems?
” ( - l) fc÷1∕c!
Determine i f each series in Problems 4-31 converges abso 12. ∑ ( 13. Λ kk
lutely, converges conditionally, or diverges. k=∖ κ
“ ( - l ) t'+1 ⅜ (-l)⅛ +1 ⅛2 ,4 ∙ ∑ < -1 )⅛ ⅛ 15.

⅛ F + l 5∙ Σ ∕c3 + 1
∞ ')k ∞
(~2) fc 46. T HINK T ANK P ROBLEM Find an upper bound for the
16. Σ ( - ι ∕ +1⅛ ι
17. Σ
k=l k= ∖ k∖ ≈ (—1V.
error if the alternating series y η—r~ is approximated by
∞ ( - 1 )L-+1 L=2 LN Κ
18. Σ ( - 1)a+1 ⅛ - 19. Σ
In k the partial sum
∕<=1 κ-∙ k=2
∞, (-i)*∙∙+∣
^ ∞ ( - 1 )L∙÷1 ς,
7
_ _J___ L + ..!____L- + J ____L
20. + 1 )(fc
21. Σ In 2 In 3 In 4 In 5 In 6 In 7
⅛ + 2) k=2 (In k f
47. T HINK T ANK P ROBLEM Find an upper bound for the
≈ ( - 1)⅛+ l ∞ ( - 1 )⅛+1 ∞ t - { ∖k + i k
22. 23. Σ
z∑2 ln(ln k) k=2 k In k
error if the alternating series y ---- ~ k— is approximated
k=ι -
∞ (—l) fc+1ln k (-l)* c+1⅛
∞
by the partial sum
24. Σ k 25. Σ
k= ∖ κ
k=2 In k
ς = 1 - 2 . 3 _ _ _4 + _5__ A
X( - l ) t+ 1 ⅛ A 2 22 23 24 25 26
26. , ÷ ∕ , k2 27. Σ (Λ + 2)2
k= ∖
48. For what numbers p does the alternating series
∞ / iλ∕c+I
⅛ < ^ 1H ⅛ Σ ) , nj ⅛+ι
l n (l n
y converge? For what numbers p does it con
28. 29. Σ l Zrln⅛
k=2
verge absolutely?
∞ / <∖ ∖∕k
k i 53k+2 ∞ ’ -ΛZ2
30. Σ ( - i) i+ 1 ⅛ 31. Σ (-n⅛+ι
<
■ U i 49. Test the series y s ι l ), for convergence.
k -∖ ∖κ - ∕ λ'-1 2
k= l κ^
k
θ Given the series in Problems 32-37, 50. Use series methods to evaluate lim -∏- where x is a real
χ
, k→∞ kl
a. Estimate the sum o f the series by taking the sum o f the first number.
four terms. How accurate is this estimate? θ 51. Show that
b. Estimate the sum o f the series with three-decimal-place
accuracy. lim (4 + 4 ÷ 4 + ■■■ ÷ —- In n]
∞ (-n⅛+ι ≈ (-l)*+> ≈ ( - n* n →∞ M 2 3 n !
32. y ∖ √ , 33. y l -∏— 34. y l 7 √2 - converges.
⅛⅛ 2 2^ 2 ⅛ 1 k∖ k½i k
” ( - l ) k+1
52. Show that >, -- r— = In 2 by completing the following
k= 1 K
steps:
m / _ ∣∖k + 1 w 1
Use the generalized ratio test in Problems 38-43 to find all a. LetSm, = ∑ -— -,---and let H m= y τ . Show that S, 2m
⅛⅛ k ⅛=∣k
numbers x for which the given series converges.
= H , - H .
∞ γk 2m in
38. 39. y ⅜ b. It can be shown (see Problem 51) that
⅛÷ι V%
lim (1 + 4 + ∙ ∙ ∙ + —— In m) = γ
2 kx k 41 » (⅛ + V x k. m →χ × Σ /77 '
40. Σ
⅛=ι k ∖ i≡i k ∖ k + 3)
where γ is a number called Euler’s constant (γ = 0.57722).
42. f (-l)H f Use this fact, along with the relationship in part a to show
y k p (-Γ ) kx k for p > 0
43.
k= ∖ ⅛=ι that lim S n = In 2. Hint. Note that
A→∞
44. T HINK T ANK P ROBLEM Find an upper bound for the S 1 = H , - H
2m 2m m
» (-l)* + '
error if the alternating series y — r — is approximated = [Hlm - ln(2m)] - [7⅛rm - In m] + ln(2m) - In m
⅛=ι "-
by the partial sum 53. Consider the following rearrangement of the conditional-
=o (_ n t+ ι
1+ 1 - I + 1 ly convergent harmonic series y ---- -,---- :
k= 1 ^-
5 2 3 4 5
45. T HINK T ANK P ROBLEM Find an upper bound for the l+ 1 -1 1 1 -1
3 2+ 5+ 7 4 + 1 J _ _l ...
9 ++ 11 6 +
≈ (_ n t+ l
error if the alternating series y — rj— is approximated a. Let S and H n denote the wth partial sum of the given
k=ι k ∞ ,
by the partial sum rearranged series and the harmonic series ∑ 4∙>
⅛=ι k
respectively. Show that if n is a multiple of n = 3m,
± + ± - ± + ±
22 32 42 52 then S,5m = H.4m - 2
⅛ 2m - ⅛H
2 nι .
Sec. 8.7 Power Series 53 9
b. Show that lim 5n = ∣x In 2. Hint. You may find it 58. T H IN K T A N K P ROBLEM Give an example of a sequence
n→∞ {ak} with the property that Xα 2, converges and Xαj,. also
helpful to use
converges.
lim ( l + ∣ + ∙ ∙ ∙ + ⅛ - l n n ) = γ 59. Show that if ∑ a, converges then ∑ α,2 must also
converge. fc=1 λ =0
given in part b of Problem 52. 60. Prove the generalized root test, which may be stated as
54. T H IN K T A N K P ROBLEM
What (if anything) is wrong with follows: Suppose ak ≠ 0 for k ≥ 1 and that
the following computation? lim Λ√Z[1<'ZkJ = L
k →∞ '
Then:
If L < 1, the series '∑ak converges absolutely.
If L > 1, the series '2.ak diverges.
61. Show that if
55. Test the given series for convergence. L = lim

⅛→≈
a. 2 - 2 1/2 + 2 1/3 - 2 1/4 + ∙∙∙
b. (1 - 2 ιz ι ) - ( 1 - 2 lz2) + ( 1 - 2 lz3) ---- the series Xα t can either converge or diverge. Hint. Find a
convergent series with L = 1 and a divergent series that
56. Suppose {ak} is a sequence with the property that satisfies the same condition.
∖a ∣ < A" for some positive number A and all positive 62. Let Σ ( - l ) t + 1 αi , be an alternating series such that {αjt} is
integers n. Show that the series ∑ak x k converges absolute decreasing and
ly for ∣x∣ ≤ ∖!A.
lim a,k = 0
k →∞
57. T H IN K T A N K P ROBLEM Give an example of a sequence Show that the sequence of odd partial sums <S,2n-ι *s
{ak} with the property that Xα 2, converges but ⅛ak diverges. decreasing.
8 .7 POWER SERIES
IN THIS SECTION Convergence of a power series, power series in x — c,

term-by-term differentiation, and integration of power series
A power series is a series whose terms are power functions o f the form
ak(x — c)k . We shall examine the properties of such series and shall find that
under reasonable conditions, they can be differentiated and integrated term by
term.
An infinite series of the form

^∑o a f x - c) k = aQ + a l (x - c) + a f x - c) 2 + ∙∙∙
is called a power series in (x — c). The numbers a0, a i ,a 2, . . . are the coefficients of
the power series, and we will be concerned only with the case where these
coefficients, as well as x and c, are real numbers. If c = 0, the series has the form
∞
k χ 2 + a x 3 + ∙∙∙
i o akx = a0 + a l x + a2
∑ y
■ CONVERGENCE OF A POWER SERIES
How can we determine the set of all numbers x for which a given power series
converges? This question is answered by the following fundamental result on
convergence. We begin by considering the case where c = 0.
T H E O R E M 8.22 Convergence of a power series
For a power series ∑ ak x k , exactly one of the following is true:
1. The series converges for all x.

2. The series converges only for x = 0.
3. The series converges absolutely for all x in an open interval ( - R , R) and
diverges for ∣x∣ > R.
Note∖The series should be checked separately at the endpoints, because it
could converge absolutely, or converge conditionally, or diverge at x = R
and x = - R .
Proof: The proof is found in most advanced calculus textbooks.
The following three examples show each of these possibilities.
E X A M P L E 1 When the convergence set is the entire x-axis
∞ γk
Show that the power series ∑ ⅛∙ converges for all x.
⅛=ι κ ∙
Solution If x = 0, then the series is trivial and converges. For x ≠ 0, we use the
generalized ratio test to find
jςk+ 1
L = lim = lim ∕T⅞T u~⅛

fc
= 1*m /-'+ 1= θ
k→∞ xk >--∙≈ {k + l)!x k→∞k + I
k∖
Because L = 0 satisfies L < I, the series converges for all x.
E X A M P L E 2 Convergence only at the point x = 0
X
Show that the power series ∑ ^∙χ k converges only when x = 0.
*= ι
Solution We use the generalized ratio test to find

,. (k + l) ! x * +1 ,. , ,.,,
Lτ = hm ----- —-7k----- = hm (vz k + l ) x 1
k→∞ k∖x k→∞ "
For x = 0, the limit is 0, but for x ≠ 0, the limit is ∞. Hence, the power series
converges only for x = 0. M M
EXAM PLE 3 When the convergence set is bounded
Find the convergence set for the geometric series

∞
∑ xk = I + x + x2 + x3 +
k=0
Solution Because this is a geometric series, we know it converges for ∣x∣ < 1.
This is the same as saying it converges on the open interval ( -1 , 1) and
diverges elsewhere. ≡i3
The generalized ratio test can be used to show that the convergence set of a
power series has one of the three forms described by the following theorem.
Sec. 8.7 Power Series 541
T H E O R E M 8.23 The convergence set of a power series

Let t a k u k be a power series, and let
L = lim
∕<→∞
Then:
If L = ∞, the power series converges only at ιι = 0.
If L = 0, the power series converges for all real u.
If 0 < L < ∞, let R = ∖∕L. Then the power series converges absolutely for
|»| < R (or - R R . It may either converge
or diverge at the endpoints u = - R and u = R.
Proof: We shall use the generalized ratio test. First, in the case where
k+
L = lim exists, consider the limit
k→∞ a
a k+ f k + ' a k+ l
M = lim = lim ∣u∣ = L ∖u∖
k→∞ ak uk k→∞ a k
The generalized ratio test tells us that ⅛ak u k converges absolutely if M < 1 and
diverges if M > 1.
If L is infinite, the series ⅛ak ιιk converges only at u = 0 (because this is the
only value of u such that L ∖u∖ < 1). If L = 0, then L ∖ιι∖ < 1 is always satisfied, and
the series converges absolutely for all u. Suppose L is finite and L ≠ 0. Then the
series converges absolutely for all u such that
∣M∣ < 4- = γ = R Because R = j~
R
Thus,
—R < it < R Because ∣u∣ < R
We see the series diverges for ∖u∖ > R , and we can have either convergence or
divergence at u = R and u = - R . ■
The number R that appears in Theorem 8.23 is called the radius of conver
gence of the power series '½al u k , and the interval
- R < u < R is the interval of absolute convergence.
Thus, the interval of convergence of a power series consists of the interval of
absolute convergence - R < u < R , and possibly one or both endpoints u = - R
and u = R , as shown in Figure 8.19. If the convergence set is all real numbers, we
write R = ∞.
Diverges Converges Diverges
-R 0 R *
Figure 8.19 The interval of convergence of a power series
In Example 1, we found that 2 x k∕k! converges for all x , so R is infinite and the
interval of convergence is the entire real line. In Example 2, we found that the
power series ⅛k∖x k has radius of convergence 0. In Example 3, we found that X x k
converges on ( - 1, 1), so R = 1.
We now turn our attention to finding R for an example that is not a geometric
series.
EXAM PLE 4 Finding an interval of absolute convergence
∞ 2 kx k
Find the interval of absolute convergence for ∑ —κ≠ - .
*=ι
2k
Solution We note that ak = jτ and u = x.
2 fc+1
L = lim ^—7 2k = 2
⅛→∞ 22 k + 1
k
Thus, the interval of absolute convergence [R = £) is -⅛ < x < ∣.
EXAM PLE 5 Finding an interval of convergence
Find the radius of convergence and the interval of convergence for the power
series
Solution We find that ak = j~_ and u = x.
so that R = £ = 1. The series converges absolutely for ∣x∣ < 1. The interval of
absolute convergence is ( -1 , 1), and the radius of convergence is 1.
0 Do not forget to check the Next, check the endpoints -1 and 1 for convergence. When x = - 1 , the
endpoints. 0 power series
Σ
fc=ι
is the convergent alternating harmonic series. If x = 1, the power series
becomes
y ι
A⅜1 k
which is the divergent harmonic series. This means that the endpoint
corresponding to x = 1 is not included in the interval of convergence. Thus,
the interval of convergence is [ -1 , 1). ∣= 1
Theorem 8.23 uses the ratio test for finding the radius of convergence.
However, sometimes it is more convenient to use the root test. The next theorem
is a companion to Theorem 8.23, which simply replaces the ratio test with the root
test.
T H E O R E M 8.24 Root test for the radius of convergence

Let ⅛ak ιιk be a power series, and let
L = lim *l{∣∖a, I
k→∞ vl *l
Then:
If L = ∞, the power series converges only at u = 0.

If 0 < L < ∞, let R = ML. Then the power series converges absolutely for
∣M∣ < R and diverges for ∣M∣ > R . It may either converge or diverge at the
endpoints u = - R and u = R.
Proof: The proof is similar to the proof of Theorem 8.23 and is omitted. ■
EXAM PLE 6 Radius of convergence using the root test

≈ / >L i 1 U - 2 3
Find the radius of convergence of the power series ∑ —τ - x k .
k=∖ ∖ K- /
/O 1∖P *
1
Solution Setting ak = I—+ — I and u = x , we have
,,— r ∕τ , a- 1V 2 W / ι∖fe
L = lim ‰,∣
vι j = lim — = lim 1 + I = e
k→∞ ' ⅛→∞L∖ κ / J ⅛→≈∖ Zc∕
Thus, the radius of convergence of the power series is R = £ = ± and the series
converges absolutely for ∣x∣ < e~ l and diverges for ∣x∣ > e~ l .
■ POWER SERIES IN (x - c)
In some applications, we will encounter power series of the form

∞
∑ ak(x - c)k = a0 + α 1(x - c) + a f x - c)2 + ∙ ∙ ∙
k=Q
in which each term is a constant times a power of x - c. The intervals of
convergence are intervals of the form —R < x —c < R , including possibly one or
both of the endpoints x = c - R and x = c + R . You can easily show this by letting
u = x - c in Theorems 8.23 and 8.24.
EXAM PLE 7 Power series in (x - c)
Find the interval of absolute convergence of the power series

” (x + l) fc
t⅜ι 3k
Solution We find that ak = j k and c = - 1:
1
3;<+1 = lim
L = lim
1 k→∞
3k
Thus, R = 3, so the interval of absolute convergence is
-3 < x + 1 < 3
—4 < x <2
that is, the interval of absolute convergence is ( - 4 , 2).
■ TERM-BY-TERM DIFFERENTIATION AND INTEGRATION

OF POWER SERIES
Consider the power series of the form

∞
∑ a kx k = a0 + a i x + α2x 2 + a i x 3 + ∙∙∙
where x is a variable and all a n are constants (often called the coefficients of the
series). For each fixed x, the power series may converge or diverge. In other words,
le t/b e a function defined b y ∕( x ) = a 0 + α 1x + α2x 2 + α 3x 3 + whose domain is
1 + x ÷ x + ∙ ∙■ the set of all values of x for which the series converges. The function / looks like a
1 - x)l + Ox + 0x 2 + Ox3 + polynomial function, except it has infinitely many terms. For example, consider
1- x the function defined by
x + 0x 2
∕( χ ) = ι£ = 1 + x + x 2 + x 3 + ∙∙∙ By long division (see margin)
x 2 + 0x 3 We recognize this as a geometric series, which has an interval of convergence of

x2 - x3 ( - 1, 1), as shown in Example 3. In Figure 8.20 we show the graphs o f/ as well as
x3 + some of the graphs of n terms of the corresponding geometric series, as follows:
∕( x ) = ι 2x ~ 1 + x interval
+x +x +2 3 + x" + ∙∙∙
of convergence ( - 1, 1)
« = 1 ~ ≈ 1
n = 2: τ ⅛ - ≈ 1 + x
1 -V
n = 3: -p4— ≈ 1 + x + x 2
1 X
n = 4: i . 2X„≈ 1 + x + x 2 + x3
a. Polynomial approximations b. Detail showing interval of convergence
Figure 8.20 The graphs of y = _ and the polynomials from the first terms
of the associated power series
Notice from Figure 8.20 that the larger the n, the closer the series approximation
is to the graph of /
If we regard a power series as an “ infinite polynomial,” we would expect to be
able to differentiate and integrate it term by term. The following theorem shows
that this procedure is legitimate on the interval of absolute convergence.
T H E O R E M 8.25 Term-by-term differentiation and integration of a power

series
A power series ∑ ak u k with radius of convergence R > 0 can be differentiated or

⅛=o
integrated term by term on its interval of absolute convergence. More specifically,
∞
if f(u) = ∑ ak u k for ∣w∣ < R , then for jι∕∣ < R we have
∕'(w ) = Σ ka k u k 1
= a 1 + 2Ω2M + 3α3u2 + 4U4M3 + ∙∙∙
k k
. x fc+1 + C
∕ ( M) du = I f ak u k du = Y ( f a ,u d ιt ] = Y 1
J ∖⅛=o t÷o∖J k
J k=o k + 1
i
Proof The proof of this result is outside the scope of this text, but details can
be found in practically any advanced calculus text. ≡≡≡
I What this says: In many ways, a function defined by a power series behaves
like a polynomial. It is continuous in its interval of absolute convergence, and
its derivative and integral can be determined by differentiating and integrating
term by term, respectively.
E X A M P L E 8 Term-by-term differentiation of a power series
Let f be a function defined by the power series

∕W = ∑θ ⅛ for all x
Show that f'( x ) = ∕( x ) for all x , and deduce that f(x ) = e*x .
Solution The given power series converges for all x (by the generalized ratio
test— see Example 1), and Theorem 8.25 tells us that it is differentiable for all
x. Differentiating term by term, we find
= ≡ [1 + x + 2 ! + ¾^ + ⅜^+ ■"]
= 0 + l+ ⅛ + ¾ → ⅛ + -
γ ∙2 -v-3
= 1 ÷ X ÷ ⅛τ ÷ ⅛τ ÷ = fix
v )
2! 3!
In Chapter 6, we found that the differential equation f'( x ) = f(x ) has the
general solution f(x ) = Ce x . Substituting x = 0 into the power series for ∕( x ) ,
we obtain
/(0) = 1 + 0 + ^ + ^ + - = 1
and by solving the equation 1 = Ce 0 for C , we find C = 1; therefore,

∕( x ) = ex . —
A power series can be differentiated term by term— not just once, but
infinitely often in its interval of absolute convergence. The key to this fact lies in
showing that if f satisfies
∕ ( Λ ) = ∑θ akx k
and R is the radius of convergence of the power series on the right, then the
derivative series
∞
∕'( x ) = Σ ka k x k ~x
also has the radius of convergence R . (You are asked to show this in Problem 44).
Therefore, Theorem 8.25 can be applied to the derivative series to obtain the
second derivative
f" ( x ) = ∑2 k{k - ∖)akx k ~2
for ∣x∣ < R . Continuing in this fashion, we can apply Theorem 8.25 to / "'(x ),/ (4),
and all other higher derivatives of f.
For example, we know that the geometric series
∞ ,
∑ x k converges absolutely to f(x ) = ∣ _ - for ∣x∣ 1
Thus, the term-by-term derivative
converges to f'( x ) = iT ⅛ for w 1, and the term-by-term second derivative
d2 ∑ xk = _d. y κ∣i χx k- = Σ k(k - 1)x k 2

dx 2 *=o dx ⅛⅛
2 and so on. These ideas are illustrated in our next

converges to f" ( x ) =
(1 - x ) 3
example.
EXAM PLE 9 Second derivative of a power series
Let f be the function defined by the power series

∞ (-l)k χ 2k
∕ W = 1∑ L W - 0 Do not forget, 0! = 1. 0
for all x. Show that ∕" ( x ) = - ∕( x ) for all x.

Solution First, we use the ratio test to verify that the given power series converges
absolutely for all x.
χ 2(fc+l)
[2(∕<+ 1)]! χ 2Zt+2 (2Λ)! 1. γ2

L = lim = lim υ
[2(∕c + 1)]! χ 2⅛ (2⅛ + 2)(2⅛+1)
A,'→ ∞ v 2⅛ A'→ χ
(W
Because L < 1, the series converges for all x. Next, differentiate the series,
term by term:
^ ) =≡ [ι-⅛ +⅞-⅛ + - 2x , 4x 3 _ 6x 5
2! 4! 6!
- _i_ ×L _ 2C5
1! 3! 5!
Finally, by differentiating term by term again, we obtain
f ( γ∖- d _ \ + x 3 x 5 - .. 1 _ ι . 3x 2 5x 4 ,
l! + 3 ! - 5 ! +4 J ~ ~1 + 3 Γ - 3 Γ +
Γ v 2 v4 ^l
= ~L - 2!^ + 4f + "J =
EXAMPLE 10 Term-by-term integration of a power series
By integrating an appropriate geometric series term by term, show that

∞ Λ-+ 1
J 0 F T T = ^ l n <1 ^ x ) fo r - 1 < x < l
φ X I
Solution Integrating the geometric series £ uk - 1 _ term by term in the
interval-1 < ιι < 1, we obtain k=0
∑ uk du = [1 + u + u 2 + u 3 + ■■•] du
k=0 ∣o
∞ γk+ 1
= Σ ⅛ fo r - 1 < x < 1
k=0 κ X 1
We also know that ( , c∣u - ln ( l - x).

Jo 1 - w
Thus, - ln ( l - x) = [ 1 d u = ∑ Λ , . fo r -1 < x < 1
v 7 Jo 1 - w ⅛≡b k + 1
PROBLEM SET 8 .7
© Find the radius o f convergence and interval o f convergence for 27. ∑ 3 Z-(5.X)4Λ' 28. ∑ k∖ {x + 1)3*
the power series given in Problems 1-28. k=1 ∕<= 1
kxk k 2x k © Find the interval o f convergence in Problems 29-34.

1. v 2. v
÷
fc 1 * +l k=∖ k + 1 X 11 χ k
29. ∑ k ∖ x + 1)M + I 30. Σ - v rk
∞ k fk + 1)x k k=∖ k~ι k
3. h k + 2 4. ∑ ∑ k - 1 x k
k=∖
31. ∑ 2'Λ I×- 1 / 32. y ιfi k x k
“ k ∖ x - 2)k k=∖ k=∖
5. f ZC2 3%Y - 3)fc 6. Λ ~ιk X
∕c= l k=∖ J
33. ∑ k(ax) k for constant a 34. y (a 2x ) k for constant a
£ 3 k [x + 3 / 4 λ (x + l) t k=∖ k=∖
7. 8.
⅛÷1 4* k=∖ j-χk In Problems 35-38, fin d the derivative f'( x ) by differentiating
“ ∕c !(x - 1 / ≈ (X - 1 5 / as a power series.
9. Z ς3l k 10. 35. ∕( x ) = ∑ ( f Y =° Yk
k=[ ⅛⅛1 hι(^ + 1) 36. f i x ) = ∑ j ~
⅛≈o ∖z ∕ k=∖ K
k
» 2 k lx - 3)k
11. Σ γ k ( * - v 12. ⅛⅜1 k l k + 1)
∕<= 1 2
37. f i x ) = ∑ (k + 2)x k 38. f i x ) = ∑ k x k
t=o ⅛=o
≈ k f3 x -4 )k ≈ (2Λ- + 3)k
13. 14.
⅛ ( ∕< + 0 2 ∕÷1 4k In Problems 39-42, fin d f(u)du by integrating a power
series. ∙'6
v ∑∑ “ I2k)!x k
15. i⅛ 16. 39. f i x ) = f ⅛ γ 40. f i x ) = ∑ ⅛
7'< r⅛ (3⅛)!
k=o k ∙, k=l κ
£ (∕C!)2XA l ~ l ) kk x k
17. Z k 18. 41. f i x ) = ∑ Ik + 2)x k 42. f i x ) = ∑ k x k
k=∖ kκ ' ⅛1 + 2)
k=0 k=0
l~ ∖ )k χ k V (3 x ∕ 43. T HINK T ANK P ROBLEM Show that the series
19. 20. ⅛ 2 fc÷1
⅛≡⅞ ∕c(ln k) 2 a
y sin(fc!x)
≈ I2x) 2k » lx + 2) 2k fc¾ k2
21. 22. Z "Ik
⅛ « k=∖ j
converges for all x. Differentiate term by term to obtain
S (3 x )κ ∞ I2>x)i k the series
23. Σ 24. Σ -k
k=i z k=∖ xy y kl cos (∕c! x)
≈ r)k ∕÷ ∣ F
25. 26. ∑ 2 k l 3 x f k
k=∖ Show that this series diverges for all x.
θ 44. Suppose {al } is a sequence for which 48. For what values of x does the series
≈ ∣
∏m⅜j = ⅛ Σ — k7 converge?
k=∖ k x
Show that the power series
Note: This is not a power series.
Σ k aκ,x k ~'
∕<-ι 49. Find the radius of absolute convergence for
has radius o f convergence R.
∞ y (k + 3)!√
45. Show that if ∕"(Λ') = y ak x k has radius of convergence ⅛¾ k∖{k + 4)!
⅛=ι
R > 0, then the series 50. Find the radius of absolute convergence for
∞
Σ akx kP ≠, 1 ∙ 2 ∙ 3 ∙ ■■k ( - χ ) 2k ~'
z,-=ι
where p is a positive integer, has radius of convergence ⅛÷ι l ∙ 3 ∙ 5 ∙ ∙ ∙ ( 2 f c - l )
R ifp .
51. Let f be the function defined by the power series
46. For what values of x does the series
f( x ∖ = y l - . l1 2 —
X ⅛ c°nver≡e7 ⅛ ( 2 fc + l) !
Note: This is not a power series. for all x. Show that ∕''( x ) = ~ ∕(x) for all x .
47. For what values o f x does the series 52. Let f be the function defined by the power series
≈ , , ” χ 2k
Σ. -s κk converge? /z( X ) “ J
■ 1(¾
k= 1 X
Note: This is not a power series. for all x . Show that f" ( x ) = ∕( x ) .
8 .8 TAYLOR AND MACLAURIN SERIES
IN THIS SECTION Taylor and Maclaurin polynomials, Taylor’s theorem,

Taylor and Maclaurin series, operations with Maclaurin and Taylor series
In this section, we seek to represent a function as an infinite series in such a way
that the value of the given function at x = c and the value of the series differ by
no more than some specified tolerance. To do this, we introduce an entire new
class of power series. We will then conclude by considering operations with
power series.
■ TAYLOR AND MACLAURIN POLYNOMIALS
Consider a function f that can be differentiated n times on some interval I. Our

goal is to find a polynomial function that approximates f at a number c in its
domain. For simplicity, we begin by considering an im portant special case where
c = 0. For example, consider the function ∕( x ) = e x at the point x = 0, as shown in
Figure 8.21. In order to approximate / by a polynomial function M(x), we begin
by making sure that both the polynomial function and f pass through the point
x = 0. That is, M(0) = ∕(0 ). We say that the polynomial is expanded about c = 0 or
that it is centered at 0.
There are many polynomial functions that we could choose to approximate /
at x = 0, and we proceed by making sure that b o th /a n d Λ√ have the same slope at
Figure 8.21 Graph o f∕(.r ) = ex x = 0. That is,
and approximating polynomials
Λ ∕ l (.λ') and M ,(x) M '(0) = ∕'( 0 )
Sec. 8.8 Taylor and Maclaurin Series 549
Historical Note The graph in Figure 8.21 shows that we have ∕( x ) = e x , so f'( x ) = e x and
/(0) = / (0 ) = 1∙ To find M l , we let
Λ f l (x) = a0 + a i x Λ ∕( x ) = α l
and require Λ√^l (0) = 1 and Λ ∕(0 ) = 1.

Because M'i (0) = a l = 1 and Λ f 1(0) = a0 = 1, we find that
Λ f l (x) = 1 + x
We see from Figure 8.21 that close to x = c the approximation off by is good,
but as we move away from (0, 1), M χ no longer serves as a good approximation. To
improve the approximation, we impose the requirement that the values of the
second derivatives of M and f agree at x = 0. Using this criterion, we find
Λ ∕(x ) = a0 + a l x + a2χ-
so that
BROOK T AYLOR (1685-1731)
Λ ∕( x ) = a l + 1a2x and M'2 (x) = l a 2
The two mathematicians most
closely associated with power Because ∕" ( x ) = ex and ∕''(0 ) = 1 we want
series are Brook Taylor and
Colin Maclaurin (1698-1746). 2a2 = 1 or α2 = ∣
The result stated in this sec
tion, called Taylor’s theorem, Thus,
was published by Taylor in Λf 2(x) = 1 + x + ∣ x 2
1712, but the result itself was
discovered by the Scottish To improve the approximation even further, we can require that the values of the
mathematician James Gregory approximating polynomials M y M y . . . , M n at x = 0 have derivatives that match
(1638-1675) in 1670 and was those o f/ a t x = 0. We find that
published by Gregory in 1688.
Taylor wrote extensively on a Λ ∕( x ) = 1 + x ÷ ∣ x 2 + ∣ x 3
variety of subjects in addition
to mathematics, ranging from Λ ∕4 (X ) = 1 + x + ∣ x 2 + ∣ x 3 + 74X4
physics to music and philoso
phy. He became secretary of Λ ∕( x ) = 1 + x + j x 2 + ∣ x 3 + j4X 4 + ⅛ x 5
the Royal Society, but his writ
ing style prevented him from
receiving full credit for all his ΛT,,(χ) = ι + f τ + f + f [ + - + ⅞τ
achievements, and his personal
life was marked with great
This 7?th degree polynomial approximation of f at x = 0 iscalled a
tragedy.
Maclaurin polynomial. If we repeat the steps for x = c ratherthan for
x = 0,we
find
,r, , (x - c) (x - c)2 (x - c)3 (x - c)n
= ec + ' 1 , ’ + ½ r r + - √ + - + √
which is called the nth degree Taylor polynomial of the function ∕( x ) = ex at x = c.
■ TAYLOR’S TH EO REM
Instead of stopping at the wth term, we will now approximate a function/by an

infinite series. A function/is said to be represented by the power series
∞
∑ akA x - c)k
k=0
on an interval I if
∕( x ) = ∑θ ak(x - c)k
for all x in I. Power series representation of functions is extremely useful, but

before we can deal effectively with such representations, we must answer two
questions.
1. Existence: Under what conditions does a given function have a power series
representation?
2. Uniqueness: When f can be represented by a power series, is there only one
such series, and if so, what is it?
The uniqueness issue is addressed in the following theorem.
THEOREM 8.26 The uniqueness theorem for power series representation
Suppose an infinitely differentiable function/is known to have the power series

representation
∕(x ) = ∑θ ak (x - c)k
for —R < x - c < R. Then there is exactly one such representation, and the
coefficients ak must satisfy
- f w W
n
a k~ k∖
fork = 0, 1, 2, . . . .
Proof: The uniqueness theorem may be established by differentiating the
given power series term by term and evaluating successive derivatives at c. We
start with
∕ ( Λ') = Σ Q a fx - c)k
and substitute x = c to obtain
∕(c ) = a0 + a i (c - c) + a2(c - c)2 + ∙∙∙ = a0
Next, differentiate the original series, term by term,
f'(x ) = a l + 2a2(x - c) + 3a3(x - c)2 +
and thus,
∕'( c ) = a l + 2Ω2(0) + 3α3(0)2 + ∙∙∙ = rz1
Differentiating once again and substituting x = c we find
f"(c}
j'( c ) = 2a2 so that a2 = —
In general, the ∕<th derivative o f /a t x = c is given by
β k ∖ x) = k∖ak so that ak = -
The question of existence—that is, under what conditions a given function

has a power series representation—is answered by considering what is called the
Taylor remainder function:
*,,(x) = ∕(x ) - Tn(x)

We see that/exists as a series if and only if

lim R,(x) = 0 for all x in I
n→∞ n
This relationship am ong/ its Taylor polynomial T n(x), and the Taylor remainder
function R n(x), is summarized in the following theorem.
THEOREM 8.27 Taylor’s theorem
If/an d all its derivatives exist in an open interval I containing c, then for each x
in I
∕W = ∕(< , ) + γ↑ ∖ x c) + ^ ^ ∖ x c)2 + ∙∙∙ + ^ ^ ∖ x c)n + R,,(x)
where the remainder function R n(x) is given by

∕ ( ,,+ l )(z )
W = ( B + 1 ) !^ C )"
for some z n that depends on x and lies between c and x.

Proof: The proof of this theorem is given in Appendix B. ■
The formula for R n(x) is called the Lagrange form of the Taylor remainder
function, after the French/Italian mathematician Joseph Lagrange (1736-1813).
When applying Taylor’s theorem, we do not expect to be able to find the exact
value of z . Indeed, if we could find that value then an approximation would not
have been necessary. However, we can often determine an upper bound for ∣R (x)∣
in an open interval. We will illustrate this situation later in this section.
■ TAYLOR AND MACLAURIN SERIES
The uniqueness theorem tells us that if/ has a power series representation at c, it
must be the series
//) + γ↑ ∖ x c) + ∙^2 j ∖ x c)2 + ? 3( ∖x c)3 + ∙ ∙ ∙
Suppose there is an open interval I containing c throughout which the function

/ and all its derivatives exist. Then the power series
f(c ) + ^ ^ ∖ x c) + j r 2 J ∖ x c)2 + j " 3 P ( x C) 3 + ∙∙∙ I
Taylor Series
is called the Taylor series of / at c. The special case where c = 0 is called the
Maclaurin series of /:
∕'( 0 ) χ ∕" ( θ ) , . ∕" '( θ ) .
Maclaurin Series ∕(zvn
θ )x ^h I] ∣ 2! x ∣ 3j x +
0 The uniqueness theorem may be summarized by saying that the Taylor

series of / at c is the only power series of the form X α ∕x - c)k that can
possibly represent/ on I. But you must be careful. All we really know is that i f
/h as a power series representation, then its representation must have the
Taylor form. 0
E X A M P L E 1 Maclaurin series
Find the Maclaurin series for f(x ) = cosx.

Solution First, note that/is infinitely differentiable at x = 0. We find
∕( x ) = cosx /(0) = 1
f'( x ) = - s in x / (0 ) = 0
∕" ( x ) = -c o s x ∕''(0 ) = - ι
∕" '( x ) = sinx ∕'" ( 0 ) = 0
/ (4)(x) = cosx / (4)(0) = 1
Thus, by using the definition of a Maclaurin series, we have
1
x2 x4 x6 / (- l) zbc2fc
cosx = ~ ⅛T + ⅞ι - 6∣ + - ∑0 (2fc)!
You might ask about the relationship between a Maclaurin series and a
Maclaurin polynomial. Figure 8.22 shows the function∕( x ) = cosx from Example
1 along with some successive Maclaurin polynomials. We use the notation Λfn(x)
to represent the first n + 1 terms of the corresponding Maclaurin polynomial.
Note that the polynomials are quite close to the function near x = 0, but that they
become very different as x moves away from the origin.
E X A M P L E 2 Maximum error when approximating a function with a
Maclaurin polynomial
Find the Maclaurin polynomial Λ ∕( x ) for the function ∕( x ) = e* and use this
polynomial to approximate e. Use Taylor’s theorem to determine the accuracy
of this approximation.
Solution First find the Maclaurin series for ∕( x ) = e x .
∕( x ) = e x /(0) = 1
∕'( x ) = ex ∕'( 0 ) = 1
The Maclaurin series for e x is

- 1+ 1 + 2∣X 2 + +
1 i^ SV ■
” fcl0 ⅛
20!
Figure 8.22 Comparison of a Then find T/(x):
function and its Maclaurin
polynomials M 5(x) = l + x + f + ⅝ +g + ⅛ ≈ ^
A comparison of/and Λ ∕( x ) is shown in Figure 8.23a. Notice from the graphs

of these functions that the error seems to increase as x moves away from the
origin. Suppose we draw the graphs on the interval [0, 1], as in Figure 8.23b.
To determine the accuracy we use Taylor’s theorem
e ≈ l + l + ⅛+ ⅜+ ⅛ + i⅛ + Λ 5(l)
= 2.716+ Λ 5(1)
0.996 1.000 1.004
a. Comparison of/a n d M. b. Comparison on [0, 1] c. Eσor atx = 1
Figure 8.23 Comparison of ∕(x ) = ex and M i (x) = 1 + ∙x + ^^^t^^^ + 24 + p/j
0 Be careful with rounding. 0 -r , ∙ , ∙ ∙ , , . ∕ , "+ l , (⅞) ,

b The remainder term is given by Rn n(x) = + ∩∣ x
c
for some number z n
between 0 and x. In particular,
* 5ω = ( 5 ⅜ ! W 5+1
for some z i between 0 and 1. Because 0 < z5 < 1, we have e" 5 < e and
Λ 5(l) < < ⅛ ≈ 0.0041667 because e< 3
Thus, e is between 2.716 666 7 + 0.004166 7 and 2.7166667 - 0.004166 7;

that is,
2.712500 < e < 2.720834
You might wonder about using a bound on e (e < 3) to find another bound on
e. This is not a new idea (remember the Newton-Raphson method). If we had
picked 2.73 instead of 3 as a first choice, we would obtain an even better
bound for e. Figure 8.23c shows a representation of the actual error at
x = 1. MH
EXAM PLE 3 Taylor series
Find the Taylor series for f(x ) = In x at c = 1.

Solution Note that/is infinitely differentiable at x = 1. We find
∕( x ) = ln x / (1 ) = 0
∕'( x ) = ⅛ ∕' ( i ) = ι
∕''( χ ) = ⅞r ∕" ( 0 = - ι
∕" ' ( * ) = ⅛ ∕" '( 1 ) = 2

A
∕ ,4 V ) = ⅛ ∕ ( 4>(1) = - 6
A
z n t ,.<-+÷y-i)i ∕<*>(1) = ( - 1 )⅛+i(fc- 1 )∣

Then, use the definition of a Taylor series to write

In x = 0 + ⅛ x - 1) - ⅛(x - 1)2 + ⅜(x - 1)3 - ⅜(x - 1)4 + -
= (x - 1) - ∣(x - 1)2 + ⅜(x - 1)3 - ±(x - 1)4 + -
_ ≈ ( - l ) fc+1( χ - 1)*
α
The computer format of the Taylor series expansion for a function may vary
considerably from that shown in the text. For example, the Taylor series
expansion for 4 terms of In x expanded about c = 1 may be
- ⅞ + ⅜^ “ 3x 2 + 4x ~ γ ∣
You notice this is equivalent to the expansion shown in Example 3.
Suppose/is a function that is infinitely differentiable at c. Now we have two

mathematical quantities, / and its Taylor series. There are several possibilities:
1. The Taylor series of f may converge to f on the interval of absolute

convergence, - R < x - c < R (or ∣x - c∣ < R).
2. The Taylor series may only converge at x = c, in which case it certainly
does not represent f on any interval containing c.
3. The Taylor series of f may have a positive radius of convergence (even
R = ∞), but it may converge to a function g that does not equal f on
the interval ∣x - c∣ < R.
E X A M P L E 4 A function that is defined at points where its Taylor series
does not converge
Show that the function In x is defined at points for which its Taylor series at
c = 1 does not converge.
Solution We know that In x is defined for all x > 0. From the previous example,
we know the Taylor series for In x at c = 1 is
( - l ) fe^l ( x - 1)A'
We find the interval of convergence for this series centered at c = 1 by

computing
(→ ) fc
Zc ÷ 1
lim
(-l)*~ 1
k
This means that the interval of absolute convergence is
∣x - 1∣ < 1
-1 + 1 < x < 1 + 1
Sec. 8.8 Taylor and Maclaurin Series 5 55
Test the endpoints:

( -l) * -( 0 -l) ∖ ≈ - i
x = 0:
k k=∖ k
which is known to diverge (harmonic series), and
≈ (-l)^ ^ 1(2 - l) fc
⅛⅜ι k
which converges (alternating harmonic series). Thus, the series converges only
on (0, 2], but the function itself is defined for all x > 0. The function is
compared with the sixth-degree Taylor approximation in Figure 8.24.
Figure 8.24 Graphs of τ∕ r1 -z γ - n - ( - < - i ) 2 , U ^ -1 ) 3 ( x -l) 4 . ( x -l) 5 ( x -l) 6

f(x) = In x and T6 (λj ⅞W (Λ 2 3 4 5 6
The next example illustrates a function whose Maclaurin series does not
represent that function on any interval.
EXAM PLE 5 A Maclaurin series that represents a function only at a single

point
Let j∕' be the function defined by

j i10
„ . „
ifc x = 0
Show that f has a Maclaurin series that represents it only at x = 0.
Solution It can be shown that f is infinitely differentiable at 0 and t h a t ∕w (O) = 0
for all k, so that f has the Maclaurin series
0 ÷ -βx ÷ y ↑∕ ^ ÷ ^f = 0
This series converges to the function ∕( x ) = 0 for all x and thus represents
∕( x ) only at x = 0. The graph of f and the Maclaurin series is shown in
Figure 8.25. ≡κ
Figure 8.25 Graph of f and its

Maclaurin series y = 0 ■ OPERATIONS WITH MACLAURIN AND TAYLOR SERIES
According to the uniqueness theorem, the Taylor series for/at c is the only power
series in x - c that can satisfy
∕( x ) = ∑θ ak (x - c)k
for all x in some interval containing c. The coefficients in this series can always be
found by substituting into the formula
za7 - ∕⅛
k~ ∕d
but occasionally they can also be found by algebraic manipulation.
EXAM PLE 6 Maclaurin series using substitution
Find the Maclaurin series for -—-—χ

1+ x z
∞
Solution We want a series of the form ∑ a.u k that represents , , on an
⅛=o 1+ x 1
interval containing 0. We could proceed directly using the definition of a

Maclaurin series. We could also carry out long division of 1 + x 2 into 1, but
instead we will modify a known series. We know that if ∣w∣ < 1 we can write
-i- -— = 1 + M+ w2 +
1- u
so by substitution (ιι = - x 2) we have
1___ 1
k v2
Historical Note provided ∣ - x 2∣ < 1; that i s ,-1 1. Hence, by the uniqueness theorem,
the desired representation is
, 1 7 = ∑ ( - l ) fex 2fc fo r - 1 < x < 1

1 + x 2 fc=o
EXAM PLE 7 Maclaurin series using substitution and subtraction
1 | x*∖
Find the Maclaurin series for f(x ) = In
(1_ M and use this series to compute
In 2 correct to five decimal places.
Solution For this function, we first use a property of logarithms:
∕( x ) = In
C OLIN M ACLAURIN
(1698-1746)
From Example 3, we know
Maclaurin was a mathematical
prodigy. He entered the Univer In x = (x - 1 ) - ∣(x - 1)2 + ∣(x - I) 3 - ∙
sity o f Glasgow at the age of
11. At 15 he took his master’s so that
degree and gave a remarkable
public defense of his thesis on ln(l + x) = [1 + x - 1] - ∣[1 + x - 1]2 + ⅜[1 + x - I] 2 - -
gravitational attraction. At 19
he was elected to the chair of , X 3 _ χ4
mathematics at the Marischal 2 +^ 3 4
College in Aberdeen, and at 21
ln(l - x ) = [1 - -l] -∣[ l-χ -l] 2 + ⅜ [ l - χ - I] 2 - -
published his first important
work, Geometrica organica. At (-x )2 (-x )3 (-x )4
27 he became deputy, or assist = (“ A) - 2 3 4
ant, to the professor of mathe
matics at the University of r3 x4
Edinburgh. There was some dif 2 3 4
ficulty in obtaining a salary to By subtracting series we find
cover his assistantship, and
Newton offered to bear the cost ∕( x ) = ln(l + x) - ln(l - x)
personally so that the university I 2 3
— I __ v
-A- ι v
-A- γ-4
___ -A I x 2 _ xZ xi
could secure the services o f so
v
L 2 3 4 2 3 4
outstanding a young man. In
time Maclaurin succeeded the = 2 [A- + < + ⅝ + -]
man he assisted. His treatise on
fluxions appeared when he was
44; this was the first logical and 1I ∖ ∙ ι
Next, by solving the equation = 2, we find x = ∙⅛ > so we are looking
systematic exposition of ι ∖ 1 √L ^ '~
Newton’s method of fluxions. (ι1 + x
X∕
) where x = Ij . It follows that
If we approximate this by a Taylor polynomial we find
ln 2 = 2 + 2fl∖ 3 + 2fl∖ 5 2 fQ 2n + l
m 3 3^3/ 5∖3j 2ΛZ ÷ 1∖3∕
where -R,,(∣ ) is the remainder term. To estimate the remainder, we note that
-R,,(∣
) is the tail of the infinite series for In 2. Thus,
2 ∕T ) 2n+3 2 ∕1 Y B+ S ,
2« + 3∖3 ∕ 2w + 5∖3 ∕
2 (l∖ 2n+i , 2 Λ lV lY n+3
2n + 3∖ 3j ^t^ 2n + 3∖9Λ3J
2
2/7 + 3
∣ 9
Because the geometric series in brackets converges to ----- 7 = ⅜, we find that
1_ 1 1 0
9
√ j‰ 2 ∣ W lY n+3
<3j 2n + 3 W W
In particular, to achieve five-place accuracy we must make sure the term on
the right is less than 0.000005 (six places to account for round-off error). With
a calculator we can see that if /7 = 4, we have
w+3(t)(j),' " o∙0000012

Thus, we approximate In 2 with n = 4:
In 2 ≈ T4(x) = j + j ÷ τ(3 ) ^"~τ(⅛) 5(⅜) "5 θ -693 146 0
which is correct with an error of no more than 0.000001 2; therefore
0.693 1460-0.0000012 < In 2 < 0.693 146 + 0.0000012

0.693 1448 < In 2 < 0.693 1473
Rounded to five-place accuracy, In 2 = 0.693 15.

We can check with calculator accuracy: In 2 ≈ 0.693 147 180.
You might wonder “why bother with the accuracy part of Example 7? I can
just press In 2 and obtain better accuracy anyway.” The answer to this might be
that the accuracy in the design of calculators and computers and programs is of
utmost importance. Indeed, the way that engineers ensure accuracy in the design
is to use series and arguments similar to the one shown in Example 7. Calculators
and computers have not changed the need for accuracy consideration, but have
simply pushed out the limits of these accuracy arguments. It is only for conve
nience that we ask for limits of accuracy that are less than what we can obtain with
a calculator (or any available machine), but the method is just as valid if we ask for
100- or 1,000-place accuracy.
E X A M P L E 8 Maclaurin series by rewriting given function in terms

of a geometric series
Graph ∕(x ) = and the 5 —2x
Find the Maclaurin series for f(x ) =
first five terms of the
Maclaurin series. Note that Solution The direct approach, namely, finding the successive derivatives o ff is
close to x = 0 the graphs seem not a very pleasant prospect. Instead, we recall that a geometric series
to coincide.
a
∑ ark converges to the sum if ∣r∣ < 1
1- r
Our approach with this example is to rewrite f in the form ∣ a~ so that we can
refer to the geometric series.
8
5 - 2x _ _ i1 +
. 8 _ _ 1 , 3
3 + 2x 3 + 2x l-( -jx )
Vι S<5-2X>z(3+2X)
VzB -l+8×3-16Xz9+ The related series converges for
32X≡z27-64X^3z81
Xnin=-1.5 Vmin='3 jx < 1 or -∣ < x < j
Xmax=2 Vna×=8
Xscl=.5 7scl=2
Thus, for an interval of convergence (-j> ∣) the geometric series is
∑ ark = . a where a = Ij and r = -⅞x

k=o ι —r
Thus,
5 - 2x _ _ 1 + V 8 F _ 2 j= _ 1 + 8 _ 16 32 2 _ M X 3 + ...
1 x 9x 27x 81x
3 + 2x fc 0 3l 3 J
⅛ 3
EXAM PLE 9 Using a trigonometric identity with a known series
Find the Maclaurin series for a. cosx 2 b. cos 2 x
Solution a. In Example 1, we found that
COSM = 1 - ¾τ + ⅛r - ⅜τ + ∙∙∙ for all u

2! 4! o!
Therefore, by substituting u = x 2, we obtain
, 1 (x 2)2 (x 2)4 (x 2)6
cosx - 1 2! + 4! 6! + "
v -4 γ ∙8 v 12
= 1 - ⅛∙ + ¾- - ⅞r + ■" for all x
2! 4! 0!
b. For cos2 x we could use the definition of a Maclaurin series, but
instead we will use a double angle trigonometric identity.
cos2 x = j + j cos2x
1 if (2x)2 (2x)4 (2x)6 I r ot - 0

rχ 1 2∣ 2! ' 4! 6! ' J Let u —^.x.
_ 1 , 1 _ 2x 2 , 2 3x 4 _ 2 5x 6 ,
2 2 2 !4 ! 6!
= 1 —x 2 + ∣ x 4 - ⅛x 6 + ∙ ∙ ∙ for all x
EXAMPLE 10 Maclaurin series by adding two power series

Historical Note
5+ γ
Find the Maclaurin series for ——— - —≈.
2 - x - x 2-
Solution Using partial fractions, we can obtain the decomposition (without
showing the steps):
5+ x _ 2 + 1
2 - x - x2 1- x 2+ x
Both of the terms on the right can be expressed as the sum of a geometric
series:
1 X
1
Y = 2LLI 2Xy∆l = 2[1 + x + x 2 + x 3 + •■■] = 2 ∑ x k
k=Q
on ( - 1, 1)
J OHANN BERNOULLI
(1667-1748)
The remarkable Bernoulli fam
ily of Switzerland produced at
least eight noted mathemati
cians over three generations. Because we are dealing with two power series, we need to be careful about
Two Bernoulli brothers, Jacob finding the interval of convergence. It is the intersection of the individual
(1654-1705) and Johann, were intervals of convergence, so we note that
bitter rivals. These brothers ( -1 , l ) ∩ ( - 2 , 2) = ( - 1 , 1)
were extremely influential ad
vocates of the newly born cal Thus,
culus. Johann was the most
5 + x
- 2 = 2 ∑ x k + i ∑ ( - 1 )⅛ Y = ∑ [2x fe + ( - 1) ⅛ Y 1
prolific of the clan and was re 7
sponsible for the discovery of 2- x - xλ k=o 2 ⅛=ov ∖2∕ j=o L 2∖2∕ J
S_____ v _____ √ J ∖_____________ _____________/
c _
Bernoulli numbers, Bernoulli
Interval of convergence: ( -1 , 1) ∩ ( - 2 ,2 ) = (-1 , 1)
polynomials, the lemniscate of
Bernoulli, the Bernoulli equa = ∑o (2 + = (2 + I) + (2 - ⅛)x + (2 + ∣)x 2 + ( 2 - ⅛)x3 + ∙∙∙
tion, the Bernoulli theorem,
and the Bernoulli distribution.
He did a great deal of work = ∣ + ^x + ^yx2 + γ⅛x3 + ∙∙∙ converges on ( -1 , 1)
with differential equations.
Johann was jealous and cantan
kerous; he tossed a son (Daniel) EXAMPLE 11 Maclaurin series by integration
out of the house for winning an 1x.
award he had expected to win Find the Maclaurin series for tan
himself. "X ,7.
J 0 1+ F
From Example 6,
= 1 _ Z2 + /4 _ t6+ ... = J θ ( - 1)A7 2∕<
Thus, γ
tan ^1x = I , 1 7 dt
Jo 1 + t~
[1 - t2 + t4 - t6 + ••■
] dt = f Γ ( - l ) z∙72λ' dt
k=0 Jo
-3
γA_ v 5 __ A
γ -7
_ 1∖kγ2k÷ 1
I -A
3 5 7 2 F ∙M for ∣x∣ ≤ 1
fc=0
No. o f Gregory's Series This series for ta n ^ 1 x is called Gregory’s series after the British m athem ati
terms term partial sut cian James Gregory (1638-1675), who developed it in 1671. It can be shown that
0 4 the Maclaurin series for tan^^1x found in Example 11 also represents ta n - 1 x at
1 -1.3333333333333 2.666666667
2 0.8 3.466666667 x = 1 and x = - 1. In particular, at x = 1, we have
3 -0.57142857142857 2.895238095
4 0.44444444444444 3.33968254
5 -0.36363636363636 2.976046176
6 0.30769230769231 3.283738484 4 - tan 1 1- 1 ∣+ ∣ 7 + 9
7 -0.26666666666667 3.017071817
8 0.23529411764706 3.252365935
9 -0.21052631578947 3.041839619 7r
10 0.19047619047619 3.232315809 ~ 4 (1 3+ 5 7+ 9
11 -0.17391304347826 3.058402766
12 0.16 3.218402766
13 —0.14814814814815 3.070254618
14 0.13793103448276 3.208185652 The convergence of this series is very slow (see the computer output in the margin)
15 -0.12903225806452 3.079153394
16 0.12121212121212 3.200365515 in the sense that it takes a relatively large num ber of terms to achieve a reasonable
17 -0.11428571428571 3.086079801 approxim ation for τr. Look at the partial sums for the 49th and 50th terms, and
18 0.10810810810811 3.194187909
19 -0.1025641025641 3.091623807 note that these are not very close to the limit (which is π).
20 0.097560975609756 3.189184782
21 -0.09302325581395 3.096161526 The m athematician Srinivasa Ram anujan (1887-1920) discovered another
22 0.088888888888889 3.185050415
23 -0.08510638297872 3.099944032 series that converges much faster. Ram anujan was a very clever m athematician
24 0.081632653061224 3.181576685 and used the trigonometric identity
25 -0.07843137254902 3.103145313
26 0.075471698113208 3.178617011 No. o f Ramanujan’s Series
27 -0.07272727272727 3.105889738
28 0.070175438596491 3.176065177 terms term partial sum
29 -0.06779661016949 3.108268567 tan a + tan β 0 3.1832635983264
30 0.065573770491803 3.173842337 tan (α + β) =
31 -0.06349206349206 3.110350274 1 - tan a tan β 1 -0.0426665690003 3.140597029
32 0.061538461538462 3.171888735 2 0.001023999998974 3.141621029
33 -0.05970149253731 3.112187243 3 -2.9257142857E-05 3.141591772
34 0.057971014492754 3.170158257
to show that 4 9.10222222222E-07 3.141592682
35 -0.05633802816901 3.113820229 5 -2.9789090909E-08 3.141592653
36 0.054794520547945 3.16861475 ta n - 1 1 = 4 ta n ^ 1 ∣- ta n - 1 ^ 6 1.00824615385E-09 3.141592654
37 -0.05333333333333 3.115281416 7 -3.4952533333E-11 3.141592654
38 0.051948051948052 3.167229468
39 -0.05063291139241 3.1 16596557 Using this identity, we find another series for ττ.
40 0.049382716049383 3.165979273
41 -0.04819277108434 3.117786502
42 0.047058823529412 3.164845325
43 -0.04597701149425 3.118868314
44 0.044943820224719 3.163812134
45 -0.04395604395604 3.11985609
46 0.043010752688172 3.162866843
47 -0.04210526315789 3.12076158 You can compare the convergence of this spreadsheet approxim ation with the first
48 0.041237113402062 3.161998693
49 -0.04040404040404 3.121594653 one and see that it approximates ττ much more quickly. (Note that the seventh
50 0.03960396039604 3.161 198613
term is fairly close to the limit π.)
Approximation of ττ using Some of the more common power series that we have derived are given in
Gregory’s series Table 8.2. For a more complete list see Student M athematics Handbook. Before we
present this list, there is one last result we should consider. It is a generalization of
the binomial theorem that was discovered by Isaac Newton while he was still a
student at Cambridge University.
THEOREM 8.28 Binomial series theorem
The binomial function (1 + x) p is represented by its Maclaurin series
(i + x y> = 1 + +¾ ⅛ + p 0 and p is not an integer; and for - 1 < x ≤ 1 if - 1 < p < 0.
TABLE 8.2 Power Series for Elementary Functions

Name Series Interval of Convergence
Taylor series at x = c follows the Maclaurin series for each function.
Exponential series e"= l + u + ⅛ + ⅛ + ⅞ + -+ ⅛ + - ( —∞ , ∞ )
e< = et' + e'(x - c) + + + - +

e ^x ~ + - ( —∞ , co)
Cosine series ., - 1 _ id
co s Ui _ ... ( " 1 )⅛ 2 t ...
COSM 1 2! ++ 4! 6!
+
+ +
+ (2fc)!
+
+ ( —∞ , ∞ )
, , . (x - c)2 (x - c)3 .
cosx = cosc — (x - c)sιnc - i — 2 , cose + i— 31 sine +
( —∞ , ∞ )
u3 u 5 u1 ( - l ) ⅛ 2 t+ l
Sine series strut w 3∣ + 5∣ 7∣+ + ^2 k + 1 j∣ + ( — co, ∞ )
(x - c)2 . -
(x - c)3
sιn x = sine + (x - c)cosc —-— 2 l sine ~— 3 ∣ ∞ sc + ( —∞ , ∞ )
Geometric series j2u= 1 + u 2 + u 3 + ■■■ + uk + ■■■ ( - 1 , i)
Reciprocal series i = 1 - (tz - 1) + (u - 1)2 - (M - 1)3 + (w - 1)4 ----- (0, 2)
Logarithmic series in u = ( » - 1) - ⅜ ιi! + - - ÷ H ) t ^ly - 1>t + - (0, 2]
In (1 + x) = x - ∣ X 2 + ∣x 3 - ∣x 4 + + —~ η — + ( -1 ,1 ]
Il nn rχ -- , nl n rc ++ X ~ C
c 2~C 2 +
+ 3^c3 "' +1 kc k ~ +, '" (0, 2c]
-1 ιι3 u 5 ιι1 ( - l ) k u2 k + l
Inverse tangent series tan π —u 3 ÷ 5 γ + ∙∙∙ + 2/^- + ∣ ÷ ∙∙∙ [ - 1 , 1]
d n - ∣ τ, _ 1 u 3 , 1 ∙ 3t∕5 1 1 ■3 ∙ 5 ∙ ¾7 1 ∣ 1 ∙ 3 ∙ 5 - (2⅛ - 3)u2*~1

Inverse sine series sin u- u+ 2 .3 ∙ 2 .4 .5 1 2 .4.6 .7 ∣ 1 2∙4 ∙6 (2⅛ - 2)(2A,- - 1) [ - 1 . i]
Binomial series (1 + u)p = 1 + pu + ^w 2 + -'------ — - u 3 + + — k + I) u ∣

c + .ι .
Proof: We begin by showing how the Maclaurin series is found. Let

∕( x ) = ( l + χ p .
∕( x ) = (l + x X /(0) = 1
f ' ( x ) = jp(l + x r 1
∕'( 0 ) = P
∕" ( x ) = jp ( ∕> - 1)(1+ x ) p ^2 f " W = p ( P ~ 1)
f ' " ( x ) = p ( p - !) ( / ? - 2)(1 + x y -i ∕" ' ( 0 ) = jp ( p - l ) ( p - 2 )
f t∙n∖x) = p(p - 1) ■
■■(p - ∏+ 1)(1 + x) p n
and
∕ w (0) = p ( p - 1) ∙∙∙ (/? - n + 1)
which produces the Maclaurin series. We could use the ratio test to show that this
series converges to some function on the interval (—1, 1), but to show that it
converges to (1 + x) p requires solving a differential equation. We leave this proof
to a course on advanced calculus. ■
E X A M P L E 12 Using the binomial series theorem to obtain a Maclaurin

series expansion
Find the Maclaurin series for ∕( x ) = V9 + x and find its radius of conver
gence.
________ / ∖1∕2
Solution Write ∕( x ) = V 9 + x = (9 + x ) ιz2 = 3( 1 + . Thus,
.-------- / γ∖ 1'2
√9 + x = 3( 1 + ⅛)
=3[1+∣(⅛+ ι
M⅛~ )(⅜ y+¾
⅛ - 1)⅛- 2)⅛y+-^∣
L 2∖9∕ 2! '' ∕ 3! ∖ / - y y l
= 3[1 + ⅛ X - 6⅛λ'2 + llj 664∙χ 3 - - ∙∙ ] = 3 + ∣ X - 2⅛∙χ 2 + 3⅛8 χ 3 -
We know from Theorem 8.28 that the series converges when ∣^∣ < 1; that is,
∣x∣ < 9, so the radius of convergence is R = 9. BMB
PROBLEM SET 8 .8
@ 1. ■ What Does This Say? Compare and contrast 19. ex + sinx 20. sinx + cosx
Maclaurin and Taylor series. 1
21. 22. -τ - - — > a ≠ 0
2. ■ What Does This Say? Discuss the binomial series 1 + 4x 1 - ax
theorem. l 1
23. - - > a ≠ 0 24. -,2 - - 2√ a ≠ 0
a+ x α + x
Find the Maclaurin series for the functions given in Problems
3-30. Assume that a is any constant. 25. ln(3 + x) 26. log(l + x)
2
3. e 2 '^ 4. e x 5. ex 6. eax 27. tan - 1 (2x) 28. V x
2 sin 2x
7. sin x 8. 9. sinαx 10. cos ax . . Γe~χ l if x ≠ 0
29. /r W = ll ifx = 0
11. cos2x 2 12. cosx 3 13. x 2cosx 14. siny
15. x 2 + 2x + 1 16. x 3 - 2x 2 + x — 5 f, λ i f x ≠ 0
30. ∕W = χ
17. xe x 18. e~x + e 2x ,1 if x = 0
Find the first four terms o f the Taylor series o f the functions in
Problems 31-42 at the given value o f c.
31. f ix ) = ex at c = 1 2
v
59' ^ ⅞ - l )
32. f(x ) = In x at c = 3
33. f{x ) = co sx at c = j
60∙ ™ = ( 1 - X) ( 2 -⅛ 7 - X)
34. f ix ) = sin x at c = j
35. f ix ) = tan x at c = 0
61. Use an appropriate identity to find the Maclaurin series
36. f ix ) = x 2 + 2x + 1 at c = 200
for fix ) = sin x cosx.
37. f ix ) = x 3 - 2x 2 + x - 5 a t c = 2
62. a. It can be shown that
38. f ix ) = V x at c = 9
cos3x = 4 cos3 x - 3 cosx
39. fix ) = ς ⅛ at c = 5
Use this identity to find the Maclaurin series for cos3 x.
40. = ⅛ at c = - 2 b. Find the Maclaurin series for sin3 x.
41∙ = ∑ ⅛ τ at c = 2 63. Use the identity
cosx + cosy = 2 cos∣(x + y) cos∣(x - y'ι

42. jf ix
v )’ = ⅛- ∙J --A at c = 2
3x + 2
to find the Maclaurin series for
Expand each function in Problems 43-48 as a binomial series.
Give the interval o f absolute convergence o f the series. ∕∖x) = (cos H cos
43. f ix ) = √ m 44. f ix ) =
64. Find the Maclaurin series for
45. f ix ) = (1 + x) 2z3 46. f ix ) = (4 + x) - ιz 3
ln[(l + 2x)(l + 3x)]
47∙ ∕W = √ T ⅛ 48∙ ∕W =
θ 49. Use term-by-term integration to show that
1 + 2x 1
≈ 1 ∖ ⅛ -l r ⅛ - 3x + 2x 2 J
ln( 1 + x) = ∑ ------j—— for ∣x∣ < 1
fc=l ^-
50. Use the Maclaurin series for ex and e~x to find the
Maclaurin series for ∕( x ) Λ- + j i nx
=
and then use this series to find lim f ix ) .

51. Use the Maclaurin series for ex and e x to find the
Maclaurin series for
ex - 1
sinhx = ---- 2---- S(x) = x
and then use this series to find lim g(x).

52. How many terms of a Maclaurin series expansion of e are
necessary to approximate U e to three-decimal-place In Problems 68- 70, the Maclaurin series o f a function appears
accuracy? with its remainder. Verify that the remainder satisfies the given
53. How many terms of a Maclaurin series expansion o f e are inequality.
necessary to approximate ∖-ze to three-decimal-place n ( - V f - ∖χ 2 k - ∖ ∣X∣2" +1
68. sin x = ∑∣ — _ ∣)∣ l^ ∙^2∏W, I-^2ΛW ∣ S (2« + 1)!
accuracy?
n lxl2n+2
Use partial fractions to fin d the Maclaurin series for the
69. cosx = ∑θ — (2 ty ↑ I^ ^2 Λ+ I W> ∣ - ^ 2 n + ιW ∣ s (2w + 2)!
functions given in Problems 54-60.
54∙ = ⅛ 55. f ix ) = f f ⅞ z j- l / ∣ ∖k γ 2 k + 1
ta n " 1 x = ∑ ⅛+1 1 + R ^ ’
Λ 1 1~r A
70. z κ2
k=0
y -2 n + 1
w ∙ ∕ ω - ¾ ⅛ 5 , ∙ ^ > - √ ⅛ 2
∣ ^ 2 ,ιω ∣ ≤ ⅛ T T for 0 ≤ X ≤ 1
c. Now integrate P ιl twice, applying the initial condi

Computational Window tions, thus obtaining our desired solution. Provide
a graph of the resulting y. Note: Make sure your y
In the following two problems we will always be expanding satisfies y(0) = 0.3, /(0) = -0 .1 .
about x = c = 0, but will refer to the Taylor father than
the Maclaurin) expansion. Also, let P n(x) denote the
Taylor polynomial o f degree n approximating f(x). θ 73. Use term-by-term differentiation of a geometric series to
71. Consider the Taylor approximations to find the Maclaurin series of the function
f{x) = x cos 2x.
a. An easy way to get an expansion for f is the /(X )
following: You know the expansion for cos x; so ( l-χ ) 3
simply replace x by 2x to obtain several terms of
cos 2x. Then multiply by x to obtain an expression 74. Find the Maclaurin series expansion for the function
fo r∕(x ). defined by the integral
b. Suppose you want to approximate∕( x ) by P 5(x) on F(x) = I e~ t ^ dt
[ - ∣ , ∣] and you are concerned about accuracy. Jo
-l
Rather than compute the remainder terms R 5(x),
which is not pleasant, consider the following.
Look at the next Taylor term (involving x 7), and
J 0
x°.2e .v c∕ x as an infinite series.
76. The functions J 0(x) and ∙ ∕ 1(x) are defined as the following
see how large this term could be on the interval
power series:
[ - j , J - Although not fool-proof, this is usually a
good estimate of the error in P fx ) . Use this to J ( χ ) = Σ c ~ υ - 2*
°w ⅛÷o (Zc!)22 2λr
estimate the error, ∕( x ) - P fx ) .
c. As a check on your work in part b, use a computer and
to compute the Taylor P fx ) .
f (~ 1 )⅛ 2k
d. Now, plot the difference f(x) - P{x) on [ - ∣ , ∣] 7 ( v ) =
1W i⅛>k∖(k + l)!2 2fc
and see how the actual error compares with the
estimate you found in part b. a. Show that J r0(x) and J r1(x) both converge for all x.
72. The idea in this problem is to use power series to b. Show that J ' o (x) = J l (x).
solve a simple-looking, but difficult, differential
equation. We want a solution on [ - 1 , 1]: These functions are called Bessel functions of the first
kind. Bessel functions first were used in analyzing
y" = g(x) = e^x sin x Kepler’s laws of planetary motion. These functions play
an important role in physics and engineering.
with initial conditions
77. Show that J o , the Bessel function of the first kind defined
X O ) = O.3, /(0) = -0 .1 in Problem 76, satisfies the differential equation
If you attempt a direct solution, you will have trouble x 2Jθ(x) + χ J ' f x ) + χ lJ fχ)~ θ
integrating y" to obtain y '. Instead, we seek an
approximation to the solution, as follows. 78. For x in the open interval ( - 1, 1), let
a. Try to integrate g(x) using your computer. De
scribe what happens. f ix ) = x + y + y + "■ + y + -
b. Approximate g(x) by a Taylor P (x) until it looks
good on [ - 1 , 1]. If possible, provide a graph. Show that f is a logarithmic function. Hint: Find f ' and
then retrieve f by integrating/'.
C H A P TER 8 R E V IEW
IN TH IS R E V IEW Proficiency examination; supplementary problems

concepts of this chapter, and Problems 31 -40 test your understanding of those
Concept Problems 21. State the absolute convergence test.

1. What is a sequence? 22. What is meant by absolute and conditional conver
2. What is meant by the limit of a sequence? gence?
3. Define the convergence and divergence of a sequence. 23. What is the generalized ratio test?
4. Explain each of the following terms: 24. What is a power series?
a. bounded sequence 25. How do you test the convergence of a power series?
b. monotonic sequence 26. What are the radius of convergence and interval of
c. strictly monotonic sequence convergence for a power series?
d. nonincreasing sequence 27. What is a Taylor polynomial?
5. State the BMCT. 28. State Taylor’s theorem.
6. What is an infinite series? 29. What are Taylor series and Maclaurin series?
7. Compare or contrast the convergence and divergence of 30. State the binomial series theorem.
sequences and series.
8. What is a telescoping series? Practice Problems
9. Describe the harmonic series. Does it converge or di In Problems 31-33, either find the limit o f the given sequence
verge? or show that the sequence diverges.
3∣ ⅛ 1 32 ∣3 M2 - π + 11 33 1Λ + Λ 1
10. Define a geometric series and give its sum. lra!∫ I (1 - 1ri)n J l∖ «/ J
11. State the divergence test.
12. State the integral test. In Problems 34-37, test the given series for convergence.
13. What is a p-series? Tell when it converges and when it ef Λ 3k 2 - k + 1
3■
X4A
' J ∕d 3, -5 ∙ J 1 (l-2 k )k 3, 6λ ' ∑⅛ 1
k ln k
diverges. 1 2
14. State the direct comparison test.

15. State the limit comparison test. 37. 1 - ∣+ I - ⅛ + ⅛ “ ⅛ + ⅛ - ⅛ + ’” + + "'
16. State the zero-infinity limit comparison test. 38. Find the interval of convergence for the series
17. State the ratio test.
1 - 2x + 3x 2 —4x 3 +
18. State the root test.
19. What is the alternating series test? 39. Find the Maclaurin series for f(x ) = sin2.x.
20. How do you make an error estimate for an alternating 40. Find the Taylor series for ∕'(x) = 1 β at c = ∣.
γ
series?
40. v 3 %k
1
D eterm in e whether each sequence in P roblem s 1—17 converges 41. y
⅛÷ι k ⅛ 2 (ln fc ) ιz *-
or diverges. I f it converges, f in d its lim it.
'■{ (-2 )"l 2 '(In w)2 "l ∙l
< Vn J
42. ∑ e - k2 43. y (V fc 3 + ι - V⅛ 3)
n 2 + 1J k≈∖ k=0
H I . 1 e !
0.∖r.
3. (
45. y k ~ k 1
1
.n 5 - I n + 1 44. Y
fc÷ ι ∖f i ( k + 1)
⅛ k2
n + (-1) M
5. ∣ 6. { 3"
1
n ,3 " + 2 ” .
46. ≠∙ fc!
47. y
4 b k 2 ( k + 1)2 1+ Vk
7∙{
5n 4 - n 2 -- 700 1 8 - l ( ι j , e ∖d
3w4 — 10«.2 + 1. n)
v ⅛⅛E
∖∑} 48. 2 . M
9. 1V « + i — 10. (V l 74 + 2'.n2 - - n2} l~ι ft∙∙
11. ∣
In « j
n J
12. { 1 H- ( - ι )”} D eterm in e whether each series given in Problem s 4 9 -5 8 con
( - 4 ) 1 verges conditionally, converges absolutely, or diverges.
13. ∣ 14. {5 2/"}
49 _ L _---- L_ 4— 1--- L_ + ...
1∙ 2 3∙2 5 ■2 7∙2
n
15. ∙[ /73/4 sin n2 nl p ]
16. Σ ____
n + 4 . 2
'k= ∣ n +
50 —— ---- — + —----------— I-
1∙ 2 2 ∙3 3 ∙4 4∙5
1Hint. T his seque nee is related to εin :integr al.
51. 3 4 ,5 6 ,
123,456,789
k_______ 2 3 4 5
17. Σ
k= -123,456.788 370,2170,367 52. 1 -l + l--L + -L-.
3 9 27 81
F in d the su m oj f the g iven convergent g<eorne trie or l elescoping
series; in Proble m s 18 -2 7 . 53. 1 -l l-l+ ...
1 2 + 3 4
s <
V
is. ; 19. i , ∣
k k 54. - 11 ++ J
√ 2__ dL∑ + _
^L4_ .
? .l ) ∙÷ m
20. ;Γ —2 1— 21. f (-i)λ
k-⅛ k - 1 ⅛ 55.
⅛÷1 k [ln(Zc + 1)]2
r 1 ∞ e k + i5 ⅛-ι
22. ] 2 23. 'Σ ∕c + l
30 / 1 ∖
k∙÷ 1 4τt - 1 k =0
6
V (t 1
57. ∑ 1 ')⅛⅛
t dtlα
1n
-1 1
∖2 k + l
25 ∙ ' ∑∣
24. ;E ( - i ) i + 1 (j-
+ ;4 τ ) '_____ I___________ L !
k-i ∖K, rc + 1 ∕ /= 2 ,In(fc + 1)
L
In Zd
< F in d the interval o f convergence fo r each pow er series in
26. ’ L)1 27∙ Σ∞ k Problem s 5 9 - 70.
h I =1 (fc + l)(fc + 2)(fc + 3) ∞ kxk
59. Σ
k
60. ∑ K x - ∖) k
Test ι'he series i n P rol dem s 2 8 -4 8 fo r c<om mergence. k=∖ 7> k=∖
1
28. ]'-k 5k%k 29. f
2
∞
xk In Zc(x2 ) i
k- fc=l V '⅛ + 4 61. Σ 62.
k=[ k (k + 1) ⅛ι Vk
∞
J ____
30. ]Γ ⅛k 31- ∑ 2/c — 1 k ∖ x + 1)2⅛ ≈ ( -l) ⅛
k--70 2 A∙=C 63. Σ 64. J i Vk
k=l 2k
k :1 ∞ ( - l ) t ( 2 x - l ) fc 4 (x + 2 fi
3 2 . :> ke~ k 33. ∑ kl
h ≡b k=0l 65. λv 66.
k=∖ k2 k ln ( k + 1)
34. ] k _____ X k (x - 3) z∙'

35. ∑ y ¾ "
k--7l k
2
k=l (2 k - 1)! 67. Σ 68. ⅛ «!
k=∖ (fc + 3)! π
2 >
36. ]”⅛ ( * 3 k+ i) 2 37. χ
£ ∞ ( - l ) ' ,x ' i
k-- k=0 3*
y
69. z∑
«=1 9 n 2 - 1
1 ( - l ) ∕ c ÷ 1χ 2 ⅛ - l
38. j° 1
39. ∑ x 3 ,μ x 5 _ x 1 4 ,
1 2 70. .. 4- 2:------- 4---------------------L . ..
k--¾ ∕<(ln k)'- k=2 fc( In fc) 3! 5! 7! {2k ■- 1)!
Find the Maclaurin series far each function given in Problems 90. Use a Maclaurin series to find a cubic polynomial that
71-76. approximates e 2x for ∣x∣ < 0.001. Find an upper bound
71. fax) = x 2e 3x 72. fax) = x 3cosx for the error in this approximation.
73. fax) = M 74. fax) = cos3 x 91. In the expansion
5 + 7x llx —1
75. fax) = 76. fax) ln(l - χ ) = - χ - ^ - ^
1 + 2x - 3x 2 2 + x - 3x 2
77. Compute the sum o f the series
show that the remainder term satisfies
l + lflY -± flY γ ∙n+1
2 2!∖2√ 3!∖2√ ∖R ,∣l
1 < — —-r when 0 < x < 1 and
" n + 1
correct to three decimal places.
78. Find a bound on the error incurred by approximating the
sum of the series How many terms of the expansion are sufficient to
compute In 1.2 correct to six decimal places? Estimate the
1 2! + 3! 4! +
error incurred by computing In 1.2 as the sum of the first
three terms in the series. 3
by the sum o f the first eight terms. 92. For what values o f x can sin x be replaced by x — if the
79. Use geometric series to write 12.342132 as a rational allowable error is 0.0005?
number. x2
93. For what values o f x can cos x be replaced by 1 - 4- if the
80. A ball is dropped from a height of A feet. Each time it allowable error is 0.00005?
drops h feet, it rebounds 0.8Λ feet. If it travels a total tanj S .
94. Find the Maclaurin series forj fix, ) —
distance o f 20 feet before coming to a stop, what is A? 1 + tan 2 x
81. Use the Maclaurin series for tanx and then use term-by-
95. a. Find a formula for the sum of the first n terms o f an
term integration to find the first three terms of the
Maclaurin series for ln(cos x) arithmetic progression that has a for its first term and
d for its common difference.
82. Show that
b. Find a formula for the sum of the first n terms of a
(χ — c)2 geometric progression that has a for its first term and r
cosx = cose - (sιnc)(x - c) - (cose)— —
for its common ratio.
z . √ x - c)3
+ (sιnc)-— ------ I- ∙∙∙ a
c. IfII
∣r∣ < 1, show that Δ-V ar" = 1∙, — r
l «=0 1 ,
83. Show that

96. P U T N A M E X A M IN A T IO N P ROBLEM Express
r , x (X - c )2 (x ~ c Y
e λv = ec + e c(x - c) + e c -— y — + ∙∙∙ + e c -— — +
y ____________6k ____________
84. Express the integral j^ V x 3 + 1 dx as an infinite series. ⅛÷ι(3n+1 _ 2i + 1 )(34 - 2 k )
Hint: Use the binomial theorem and then integrate term

as a rational number.
by term.
85. Estimate the value of the integral 97. P U T N A M E X A M IN A T IO N P ROBLEM For
0 < x < 1, express
rθ-4
I dJx
Jo V l + x3
with an error no greater than 0.000001.

86. Show that if a > 0, α ≠ 1, as a rational function o f x.
Λ (In a)k . 98. P U T N A M E X A M IN A T IO N P ROBLEM For positive real x , let
axx = Λ — r>— χk for alln x
k=0 κl
B n(x) — F ÷ 2 1 ÷ 3 x + ∙∙∙ + n x
87. Show that for 0 < x < 2c,
i = 1 - ^ ( x - c) + ^ ( x - c) 2 - j i (x - c)3 + - Test for convergence for the series
88. Use series to find sin 0.2 correct to five decimal places. y ¾ ⅛ 2)
89. Use series to find In 1.05 correct to five decimal places. k=2 (k log, k)2
. .............................. GROUP RESEARCH PROJECT*
"Elastic Tightrope Project"
Suppose a cockroach starts at one end of a 1,000-m tightrope and runs toward the
other end at a speed of 1 m∕s. At the end of every second, the tightrope stretches
uniformly and instantaneously, increasing its length by 1,000 meters each time.
Does the roach ever reach the other end? If so, how long does it take?
concerns. Suppose it is you standing on the elastic rope b meters from the left end
and c meters from the right end. Then suppose the entire rope stretches uniformly,
increasing its length by d meters. How far are you from each end? You might also
try to generalize the roach problem.
The infinite'. No other question has

ever moved so profoundly the spirit
o f man∖ no other idea has so fruit
fully stimulated his intellect', yet no
other concept stands in greater
need o f clarification than that of
the infinite.
David Hilbert
To Infinity and Beyond by Eli Maor
(Burkhauser, 1987), p. vii
There is nothing now which ever

gives me any thought or care in
algebra except divergent series,
which I cannot follow the French in
rejecting.
Augustus De Morgan
Graves' Life o f W. R. Hamilton
(New York: 1882-1889), p. 249
*This group project was obtained courtesy of David Pengelley, New Mexico State University. You
might also check “The Beetle and the Rubber Band,” by Alexander A. Pukhov, Quantum, March/April
1994. pp. 42-45.
9
Polar Coordinates
and the Conic Sections
■ CONTENTS
9.1 The Polar Coordinate System
Plotting points in polar
coordinates; Primary represen PREVIEW
tations of a point; Relationship
between polar and rectangular We begin this chapter by investigating an alternative coordinate system
coordinates; Points on a polar
curve involving polar coordinates. Much of our work will involve representa
9.2 Graphing in Polar Coordinates tion of graphs, but we shall also investigate tangent lines, area, arc
Graphing by plotting points; length, and other issues involving differentiation and integration. We
Cardioids; Symmetry and close the chapter by examining the conic sections: parabolas, ellipses,
rotations; Limacons; Rose
curves; Lemniscates; Summary and hyperbolas.
of polar-form curves
9.3 Area and Tangent Lines in
Polar Coordinates
Intersections of polar-form
curves; Area bounded by polar PERSPECTIVE
graphs; Tangent lines
In the rectangular coordinate system of representation, points in the plane are
9.4 Parametric Representation of
Curves
located by specifying their position in relation to a fixed pair of perpendicular
Parametric equations; lines. However, this is by no means the only way of locating points, and in this
Derivatives; Arc length; Area chapter, we shall study several alternative forms of representation.
9.5 Conic Sections: The Parabola Another topic, the conic sections, has been studied extensively since ancient
Standard-position parabolas; times and has many important applications. For instance, in the early 17th
Translation of parabolas;
Representation of a parabola in century, Johannes Kepler observed that the planets travel in elliptical paths,
polar coordinates; Parabolic and Galileo discovered that in a vacuum, a projectile follows a parabolic path.
reflectors At the end of the 17th century, Newton used the fact that planets follow
9.6 Conic Sections: The Ellipse and elliptical paths as the basis for the inverse-square law of gravitational attraction.
the Hyperbola In more modern times, conic sections have been used in architecture, in the
Ellipses; Hyperbolas;
Eccentricity and polar design of lenses and mirrors, and to study the paths of atomic particles. We
coordinates; Geometric shall examine some of these applications in our study of conic sections. (See
properties Sections 9.5 and 9.6.)
Chapter 9 Review
569
570 Ch. 9 Polar Coordinates and the Conic Sections
9 .1 THE POLAR COORDINATE SYSTEM
IN THIS SECTION Plotting points in polar coordinates, primary

representations of a point, relationship between polar and rectangular
coordinates, points on a polar curve
Up to now, the only coordinate system we have used is a Cartesian coordinate
system. In this section, we introduce the polar coordinate system, in which
points are located by specifying their distance from a fixed point and their
direction in relation to a fixed line.
We use the term rectangular co ■ PLOTTING POINTS IN POLAR COORDINATES

ordinates to refer to Cartesian
coordinates to help you remem In a polar coordinate system we fix a point O, called the pole, and draw a ray (half
ber that Cartesian coordinates line), called the radial axis, from the point O. Then we represent each point in the
are plotted in a rectangular plane by an ordered pair P(r, θ), where θ (the polar angle) measures the angle from
fashion. the radial axis and r (the radial distance) measures the directed distance from the
pole to the point P. Both r and θ can be any real numbers. When plotting points
rotate the radial axis through an angle θ, as shown in Figure 9.1. You might find it
helpful to rotate your pencil as the axis—the tip points in the positive direction
and the eraser, in the negative direction. If θ is measured in a counterclockwise
direction, it is positive, and if θ is measured in a clockwise direction, it is negative.
Next, plot r on the radial axis (the pencil). Notice that any real number can be
plotted on this real number line in the direction of the tip if the number is positive
and in the direction of the eraser if it is negative. Plotting points seems easy, but it
is necessary that you completely understand this process. Study each part of
Example 1 and make sure you understand how each point is plotted. We assume
an orientation provided by an x-axis and a y-axis with the pole at the origin, and
Figure 9.1 Polar-form points the rotation of the radial axis is measured from the positive x-axis. We consider
that points can be plotted using either rectangular or polar coordinates, and that
the system we are using in a particular example must be either specified or
obvious from the context.
EXAM PLE 1 Plotting polar-form points
Plot each of the following polar-form points: A(4, J ), 5 (-4 , j), C(3, -f ) ,
Z>(-3, - f ) , £ ( - 3 , 3), £ ( - 3 , - 3 ), (7 (-4 , -2 ), //(5, ⅜r ), / ( - 5 , f), j(5 , - f ) .
Solution Points A, B, C, and D illustrate the basic ideas of plotting polar-form

points.
Plot C(3, - y ) ; negative θ.
0
Rotate 0 = - y
6
Sec. 9.1 The Polar Coordinate System 571
Points E, F, G, H, I, and / illustrate common situations that can sometimes be

confusing. Make sure you take time with each example.
Note that the point (-r, θ) can be obtained by reflecting (r, θ) through the origin,
whereas the point (r, —0) can be obtained by reflecting (r, θ) in the x-axis. These
features of the polar coordinate system are illustrated in Figure 9.2.
a. The point (-r, θ) is obtained b. The point (r, - θ ) is

by reflecting (r, θ) through obtained by reflecting
the origin. (r, θ) in the x-axis.
Figure 9.2 Plotting points using reflections
■ PRIMARY REPRESENTATIONS OF A POINT
Notice from Example 1 that ordered pairs in polar form are not associated in a
one-to-one fashion with points in the plane. Indeed, given any point in the plane,
there are infinitely many ordered pairs of polar coordinates associated with that
point. If you are given a point P(r, θ) other than the pole in polar coordinates, then
(- r, θ + τr) also represents P. In addition, there are infinitely many other polar
representations of P; namely, (r, θ + nπ) for n an integer. We call (r, θ) and
0 Remember to simplify so the
(- r, θ + π) the primary representations of the point if the angles θ and θ + π are
second component represents a
between 0 and 2π.
nonnegative angle less than one
revolution. 0
Primary Representations of a Every nonzero point in polar form has two primary representations:
Point in Polar Form
(r, θ) where 0 ≤ θ < 2τr and
(- r , π + 0) where 0 ≤ ττ + θ < 2π
57 2 Ch. 9 Polar Coordinates and the Conic Sections
EXAM PLE 2 Primary representations of a point
Give both primary representations of the points

a. (3, f) b. (5, ⅞) c. (-6 , -⅛ ) d. (9, 5) e. (9, 7)
Solution a. (3, f) is primary; the other is ( - 3 , j¾r)

Change the sign o f r and add ιr to 0; -%+ π.
b. (5, j¾γ) is primary; the other is ( - 5 , f).

∙¾r + 77 = ¾r ' but ( -5 , ¾r) is not a primary representation because ¾τ > 2τr.
c. ( -6 , - =yj has primary representations ( -6 , ⅛fr l and [6, f).

d. (9, 5) has primary representations (9, 5) and ( - 9 , 5 - TT); a point like
( -9 , 5 - rr) may be approximated by ( -9 , 1.86).
e. (9, 7) has primary representations (9, 7 - 2ττ), or (9, 0.72), and ( -9 , 7 - ττ),
or ( - 9 , 3.86). —
■ RELATIONSHIP BETWEEN POLAR AND RECTANGULAR

COORDINATES
The relationship between polar and rectangular coordinates can be found by using
the definition of trigonometric functions (see Figure 9.3).
Relationship Between 1. To change from polar to rectangular-.

Rectangular and Polar
x = r cos θ y = r sin θ
Coordinates
2. To change from rectangular to polar-.
r - V Λ^2 + ρ 2 θ = tan~1 - L x ≠ 0
IX∣
where 0 is the reference angle** for θ. Place θ in the proper quadrant by noting
the signs of x and y. If x = 0, then ~θ = f .
EXAM PLE 3 Converting from polar to rectangular coordinates

P(χ,y)
Change the polar coordinates ( -3 , ¾r) to rectangular coordinates.
Solution x = - 3 cos¾r = -3(^-⅛r) = 2 ⅛=-
y = - 3 sin¾r = - 3 ^ - ^ = -=⅛p-
∕3 V z2 3V2∖
The rectangular coordinates are (~⅜~, )∙
Figure 9.3 Relationship between

rectangular and polar coordinates
EXAM PLE 4 Converting from rectangular coordinates to polar coordinates
Write both primary representations of the polar-form coordinates for the

point with rectangular coordinates (⅛m, —1).
*The reference angle 0 for a standard position angle 0 is defined to be the smallest positive angle
that the angle 0 makes with the Λ-axis. Reference angles are discussed in trigonometry and are reviewed
in the Student Mathematics Handbook fo r C A L C U L U S .
Computational Window o n
Solution ⅞ ÷⅛ =5
Check the owner’s manual for
your calculator. Many have Note that θ is in Quadrant IV because x is positive and y is negative.
keys for converting between
polar and rectangular coordi thus, θ = 1 ∣ 2λ (Quadrant IV)
nates.
The polar-form coordinates are (5, ~5~) and (-5 ,
■ POINTS ON A POLAR CURVE
The graph of an equation in polar coordinates is the set of all points P whose polar
coordinates (r, 0) satisfy the given equation. Circles, lines through the origin, and
rays emanating from the origin have particularly simple equations in polar
coordinates.
EXAM PLE 5 Graphing circles, lines, and rays
Graph: a. r = 6; b. θ = f ’ r ≥ 0; and c. θ = %.
Solution
a. The graph is the set b. The graph is the c. This is the line
of all points (r, 0) such closed half-line (ray) that through the origin that
that the first component emanates from the ori makes an angle of f with
is 6 for any angle θ. This gin and makes an an the positive x-axis.
is a circle with radius 6 gle of f with the positive
centered at the origin. x-axis.
As with other equations, we begin graphing polar-form curves by plotting

some points. However, you must first be able to recognize whether a point in polar
form satisfies a given equation.
EXAM PLE 6 Verifying that polar coordinates satisfy an equation
Show that each of the given points lies on the polar graph whose equation is
1 - cos θ
a. (2 ,f) b. (-2 ,⅜ ) c. ( - l ,2 π )
Solution Begin by substituting the given coordinates into the equation.
a 2 ± ___ 2___ _ _ 2 _ _ 22∙ '

∙ 1 -c o sf 1 -0
(2, ⅞) is on the curve, because it satisfies the equation.
b. - 2 ⅛ ----- 2 _ _ = = 2 This equation is not true.

1 - cos ⅛r 1 0
Although the equation is not satisfied, we cannot say that the point is not
on the curve. Indeed, we see from part a that it is on the curve, because
( - 2 , ⅛ and (2, name the same point! So even if one prim ary
representation of a point does not satisfy the equation, we must still check
the other prim ary representation of the point. An ordered pair representing
a polar-form point (other than the pole) satisfies an equation involving a
function of nθ (for an integer ri) if at least one of its prim ary representa
tions satisfies the given equation.
c. - 1 - τ----
1 —cos- —≈y-
2ττ = ι----
1 —r1’ which is undefined.
Check the other representation of the same point— namely, (1, π)'.
_L 2 = -
1 - cos ri 1 -(-1 )
Thus, the point is on the curve.
We will discuss the graphing of polar-form curves in the next section, but some
polar-form equations can be graphed by changing them to rectangular form.
EXAM PLE 7 Polar-form graphing by changing to rectangular form
Graph the given polar-form curves by changing to rectangular form.
a. r = 3 cos θ b. r cos θ = 4
2 sin 0 + cos θ
Solution a. r = 3 cos θ Given equation
r2 - 3r cos θ Multiply by r.
x 2 + y 2 = 3x Because x = r cos θ and r2 = x 2 + y 2
[x 2 - 3x + g ) 2] + y 2 = I Complete the square.
W 2 ÷ τM
We see this is a circle with center at (j, θ) and radius j. The graph is shown
in Figure 9.4a.
b. Because x = r cos θ, we see that the given equation can be written x = 4,
which is a vertical line (Figure 9.4b).
e. r= Λ cos0 G iv e n e q u a t i o n
2 s in
2r sin 0 + r cos θ = 6
2y + x = 6
We see that this is a line with y-intercept 3 and slope - ∣(Figure 9.4c).
PROBLEM SET 9 .1
θ 1. ■ What Does This Say? Illustrate the procedure for θ Write each equation given in Problems 2 3 -3 0 in rectangular
plotting points in polar form. coordinates.
2. ■ What Does This Say? Show the derivation o f the 23. r = 4 sin θ 24. r = 16
polar-rectangular-form conversion equations. 25. r = 1 - sin 0 26. r = 2 cos θ
27. r = sec θ 28. r = 4 tan θ
2 2
29. Γ2 = 30. Γ2 =
1 + sin 2 θ 3 cos 2 θ — 1
In Problems 3 -1 1 , plot each o f the given polar-form points. Sketch the graph o f each equation given in Problems 31-38.
Give both prim ary representations and the rectangular coordi 31. r = ∣ 32. r = 4> 0 ≤ θ <: 2
nates o f the point. 33. r = √ 2 , 0 ≤ θ < 2 34. r = 4
3 .( 4 ,f ) 4 .( 6 4 ) 5 ∙( 5 ,⅛ ) 35. β = 1 36. θ = 1, r ≥ 0
6- (3. - f ) 7. (2, -⅝ ) 8. ( - 4 , 4) 37. Φ I q
q Λ 0 38. θ = l
σ
9. (l,3 τr) 10. ( - 2 , -¾ j) 11. ( 0 ,- 3 )
In Problems 3 9-44, tell whether each o f the given points lies on
the curve
1 - sin θ
Plot the rectangular-form points in Problems 12-20 and give 39. (IO ,?) 4 θ∙ (5, ⅛ 41. ( -1 0 , ⅝)
both prim ary representations in polar form . 42. ( - ⅛ , ⅝ ) 43. (2 0 + 1 0 √ 3 ,f) 44. ( -1 0 , f)
12. (5, 5) 13. ( - 1 , √3) 14. (2, -2 √ 3 )
15. ( - 2 , - 2 ) 16. (3, -3 ) 17. (3, 7) In Problems 45-52, tel! whether each o f the given points lies on
18. (3, -3 √ 3 ) 19. (√ 3 , - 1 ) 20. ( - 3 , 0) the curve r = 2( 1 - cos θ).
21. Name the country or island in Figure 9.5 in which each o f 45. ( I ,? ) 46. ( I , - ? )
the named points is located. 48. ( - 2 4 )
47. ( - I , ? )
a. (5.5, 260°) b. (7.5, 78°) c. ( - 2 , 140°) d. ( - 4 , 80°)
49. (2 + √ 2 , f ) 50. ( - 2 - √ 2 , f)
22. Name possible co
ordinates for each 51. (θ, f) 52. (θ, -⅜ )
o f the following cit
ies (see Figure 9.5). F in d three distinct ordered pairs satisfying each o f the equa
a. M iam i, Florida tions in Problems 53-60. Give both primary representations fo r
b. Los Angeles, each point.
California 53. r 2 = 9 cos θ 54. r 2 = 9 cos 2θ
c. M exico City, 55. r = 3θ 56. r = 5θ
Mexico
57. r = 2 - 3 sin θ 58. r = 2(1 + cos θ)
d. London,
59. ----- '------ = 2 60. r = ______ 8______
England Figure 9.5 1 - sin θ 1 - 2 cos θ
© 6 1 . a. What is the distance between the polar-form points 62. Use the result of Problem 61b to find an equation for a
(3, -j) and (7, 5)? Explain why you cannot use the circle of radius a and polar-form center (R, a).
distance formula for these ordered pairs. 63. Show that the graph of the polar equation
r = a sin θ + b cos 0 is a circle. Find its center and radius.
b. What is the distance between the polar-form points
(r1, 01) and (r2, 0,)?
9 .2 GRAPHING IN POLAR COORDINATES
IN THIS SECTION Graphing by plotting points, cardioids, symmetry and

rotations, limaςons, rose curves, lemniscates, summary of polar-form curves
In the last section we found that the graph of the polar equation r = a is a circle
centered at the origin with radius a, and θ = a is a line passing through the
origin. In this section we discuss techniques for graphing polar-form equations
and categorize several important types of polar-form curves.
■ GRAPHING BY PLOTTING POINTS
We have examined polar forms for lines and circles, and in this section we shall
examine curves that are more easily represented in polar coordinates than in
rectangular coordinates. We begin with a simple spiral.
EXAM PLE 1 A spiral by plotting points
Graph r = θ for θ ≥ 0.
Solution Set up a table of values.
θ r
0 0
1 1
2 2 Choose a θ, and then find
3 3 a corresponding rso that
(r, θ) satisfies the equation.
4 Plot each of these points
5 5 a d connect them, as shown
∏
6 6 in Figure 9.6.*
Figure 9.6 Graph of r = θ
Notice that as θ increases, r must also increase. M M
In Example 1, the polar equation r = θ, we see that there is exactly one value
of r for each value of θ. Thus, the relationship given by r = θ is a function of θ, and
we can write the polar form r = f(θ), where /(0) = θ. A function of the form
r = f(β)> where 0 is a polar angle and r is the corresponding radial distance, is
called a polar function.
* Many books use what is called polar graph paper, but such paper is not really necessary. It also
obscures the fact that polar curves and rectangular curves are both plotted on a Cartesian coordinate
system, only with a different meaning attached to the ordered pairs. You can estimate the angles as
necessary without polar graph paper. Remember, just as when graphing rectangular curves, the key is
not in plotting many points but in recognizing the type of curve and then plotting a few key points.
Sec. 9.2 Graphing in Polar Coordinates 577
Computational Window Q B i ■ CARDIOIDS
Check your owner’s manual. Next, we examine a class of polar curves called cardioids because of their heartlike
Some graphing calculators will shape.
draw polar-form graphs.
EXAMPLE 2 A cardioid by plotting points
Graph r = 2(1 - cos θ).

Solution Construct a table of values by choosing values for θ and approximating
the corresponding values for r.
θ r
r ιB 2 < l - c o s θ )
0 0
8n i r ∣
=0
8rnax=6.2831853... 1 0.919395
8 s te p = . 1308996... 2 2.832294
Xr ∣i n=- 6
Xr∙tax=2 3 3.979 985
4 3.307287
V n in=^3
S∙,nax=3 5 1.432676
6 0.079659
The points are connected as

shown in Figure 9.7. Figure 9.7 Graph of r = 2(1 - cos#) '∙
The general form for a cardioid in standard position is given in the following
box.
Standard-Position Cardioid r = a( 1 - cos θ)
In general, a cardioid in standard position can be completely determined by

plotting four particular points:
θ r = a(l - cos θ)
0 r = a(l - cos0) = α(l - 1) = 0
2 r = tz(l - cos J) = a(l - 0) = a
ττ r = α( 1 - cos π) = α( 1 + 1) = 2«
3ττ
2 r = α(l - cos⅛p) = α(l - 0) = a
These reference points are all you need when graphing other standard-position
cardioids, because they will all have the same shape as the one shown in Figure
9.7.
■ SYMMETRY AND ROTATIONS
When sketching a polar graph, it is often useful to determine whether the graph
has been rotated or if it has any symmetry.
If an angle a is subtracted from 0 in a polar-form equation, it has the effect of

rotating the curve (see Problem 73). As before, we note that
Rotation of Polar-Form Graphs The polar graph of r = /( 0 - α) is the same as the polar graph of r = /(0) only
rotated through an angle a. If a is positive, the rotation is counterclockwise, and
if a is negative, then the rotation is clockwise.
EXAM PLE 3 Rotated cardioid
Graph r = 3 - 3 cos (0 - f).
Solution Recognize this as a cardioid with a = 3 and a rotation of f. Plot the four
points shown in Figure 9.8 and draw the cardioid. H M
If the rotation is 180o , the equation simplifies considerably. Consider

r = 3 —3 cos(0 — ττ) Cardioid with 180o rotation
= 3 —3[cos 0 cos 77 + sin 0 sin ττ] cos (α - β) = cos a cos β + sin a sin β
= 3 - 3[cos 0 ( - 1 ) + sin 0 (0)]
= 3 - 3 (-l)c o s 0
Figure 9.8 Graph of
r = 3 - 3 cos(θ - f) = 3 + 3 cos 0
Compare this with Example 3 and you will see that the only difference is a 180o
rotation instead of a 30o rotation. This means that, whenever you graph an
equation of the form
r = α(l + cos 0) Note the plus sign.
instead of r = α(l - cos0), it is a standard-form cardioid with a 180° rotation.
Similarly,
r = α(l - sin 0) is a standard-form cardioid with a 90° rotation.
r = α(l + sin 0) is a standard-form cardioid with a 270° rotation.
These curves, along with other polar-form curves, are graphed and summarized in
Table 9.1 on page 583.
The cardioid is only one of the polar-form curves we will consider. Before
sketching other curves, we will consider symmetry. There are three im portant
kinds of polar symmetry, which are described in the following box and are
demonstrated in Figure 9.9.
Symmetry in the Graph of the A polar-form graph r = /(0) is sym . . . if the equation r = /(0) is un
Polar Function r = /(0 ) m etric with respect to . . . changed when (r, 0) is replaced by . . .
x-axis (r, - 0 )
j-axis (r, 77—0)
origin ( - r , 0)
Sec. 9.2 Graphing in Polar Coordinates 57 9
Figure 9.9 Symmetry in polar form
■ LIMACONS
We will illustrate symmetry in polar-form curves by graphing a curve called a

limaςon.
EXAM PLE 4 Graphing a limaςon using symmetry
Graph r = 3 + 2 cos θ.
Solution Let /(0) = 3 + 2 cos θ.
Symmetry with respect to the x-axis:

/( - 0 ) = 3 + 2 cos ( - 0) = 3 + 2 cos 0 = /(0)
Yes; it is symmetric, so it is enough to graph / for 0 between 0 and π.
Symmetry with respect to the y-axis:

∕(ττ —0) = 3 + 2 cos (ττ - 0)
= 3 + 2[cos ττ cos 0 + sin 7τ sin 0]
= 3 - 2 cos 0
After checking the other primary representation, we find that it is not
symmetric with respect to the y-axis.
Symmetry with respect to the origin:

- r Ψ /(0) and r ≠ /(0 + ττ), so the graph is not symmetric with respect to the
origin.
The graph is shown in Figure 9.10; note that we sketch the top half of the
graph (for 0 ≤ 0 ≤ τr) by plotting points and then complete the sketch by
reflecting the graph in the x-axis. Because cos 0 steadily decreases from its
largest value of 1 at 0 = 0 to its smallest value - 1 at 0 = ττ, the radial distance
r = 3 + 2 cos 0 will also steadily decrease as 0 increases from 0 to τr. The
largest value of r is r = 3 + 2 (l) = 5 at 0 = 0, and its smallest value is
1——■
I—Bottom — 4+ d —fτt^ ∖ —l-
half—r
is sketched using x-axis r = 3 + 2 (-1 ) = 1 at 0= 7τ. ■—
symmetry.
The graph of any polar equation of the general form
r = 3 + 2 cos θ r - b ± a cos θ or r - b ± a sin θ
is called a limaςon (derived from the Latin word “limax,” which means “slug”).
The special case where a = b is the cardioid. Figure 9.11 shows four different
kinds of limaςons. Note how the appearance of the graph depends on the ratio alb.
We have discussed cases II and III in Examples 2 and 4. Case I (the “inner loop”
case) and case IV (the “convex” case) are examined in the problem set.
Case I Case II Case ΠI Case IV

(r = l- 2 c o s θ ) (r = 2 - 2 c o s θ ) (r= 3 -2 c o sθ ) ( ∕'= 3 - c o s θ )
Figure 9.11 Limaςonsι r = b ± a cos θ or r = b ± a sin θ
Just as with the cardioid, we designate a standard-form limaςon and consider the
others as rotations:
r= b- a cos θ Standard form
r = b- a sin θ 90o rotation
r = b+ a cos θ 180° rotation
r = b+ a sin θ 270° rotation
Standard-Position Limaςon r = b - a cos θ
■ ROSE CURVES
There are several polar-form curves known as rose curves, which consist of several
loops, called leaves or petals.
EXAM PLE 5 Graphing a four-leaved rose
Graph r = 4 cos 20.

Solution Let f(θ) = 4 cos 20.

f ( - θ ) = 4 cos 2 ( - θ) = 4 cos 2θ = f(θ∖ , Yes, symmetric

∕(ττ - θ) = 4 cos 2(τr - 0) = 4 cos (2τr - 20) = 4 cos (-20) = 4 cos 20 = /(0);
Yes, symmetric
Because of this symmetry, we shall sketch the graph of r = /(0) for 0

between 0 and f ’ and then use symmetry to complete the graph.
θ
When θ = 0, r = 4, and as θ increases from
f( β )
0 to f ’ the radial distance r decreases from 4 to
0 4
0. Then, as θ increases from f to γ r becomes
0.2 3.684244
0.4 2.786827 negative and decreases from 0 to -4 . A table of
0.6 1.449431 values is given in the margin.
0.8 -0.116798
1 -1.664587
1.2 -2.949575
1.4 -3.768889
The next part of the graph
is obtained by first reflecting The last part of the graph is found
Keep plotting points until you in the y-axis. by reflecting in the x-axis.
are satisfied that you have a
reasonable representation for the
graph. After working through
Example 6, you will be able to
do other rose curves more easily.
In general, r = a cosn0 is the equation of a rose curve in which each petal has
length a. If n is an even number, the rose has 2n petals; if n is odd, the number of
petals is n. The tips of the petals are equally spaced on a circle of radius of a.
Equations of the form r = a sinn0 are handled as rotations.
Standard-Position Rose Curve r = a cosnθ
EXAM PLE 6 Graphing a rose curve with 8 petals and a rotation
Graph r = 5 sin 40.

Solution We begin by finding the amount of rotation.
r = 5 sin 40
= 5 cos - 40) Cofunctions o f complementary angles
= 5 cos (40 - f) Remember, cos ( - 0) = cos θ.

Figure 9.12 Graph of r = 5 sin 4θ
= 5 cos 4(0 - f)
Recognize this as a rose curve rotated f. There are 2(4) = 8 petals of length 5.
The petals are a distance of f (one revolution = 2ττ divided by the number of
petals) apart. The graph is shown in Figure 9.12. ≡≡≡
■ LEMNISCATES
The last general type of polar-form curve we will consider is called a lemniscate.
Standard-Position Lemniscate r 2 = rz2cos20
EXAMPLE 7 Graphing a lemniscate
Graph r2 = 9 cos 2θ.

Solution As before, when graphing a curve for the first time, begin by checking
symmetry and plotting points. For this example, note that you obtain two
values for r when solving this quadratic equation. For example, if θ = 0, then
cos 2θ = 1 and r2 = 9, so r = 3 or -3 .
9 cos [2(-0)] = 9 cos 20, so r 2 = 9 cos 20 is not affected when 0 is replaced
by - 0; yes, symmetric with respect to the x-axis.
9 cos [2(ττ - 0)] = 9 cos (2π - 20) = 9 cos (-20) = 9 cos 20; yes; symmetric
with respect to the y-axis.
Symmetry with respect to the origin:
(-∕∙) 2 = r 2, so r2 = 9 cos 20 is not affected when r is replaced by -r ; yes;
symmetric with respect to the origin.
Note that because r2 ≥ 0, the function cos 20 is defined only for those values
of 0 for which cos 20 ≥ 0; that is
_ ττ . r∖ . 7T. 3τr .- a . 5τr.
4 ≡= - 4’ 4 - t ' - 4 ,...
We begin by restricting our attention to

the interval 0 ≤ 0 ≤ f . Note that cos 20
decreases steadily from 3 to 0 as 0 varies from
0 to f.
A second step is to use symmetry Finally, obtain the rest of the graph
to reflect the curve in the x-axis. by reflecting the curve in the y-axis.
■ SUMMARY OF POLAR-FORM CURVES
We conclude this section by summarizing the special types of polar-form curves

we have examined. There are many others, some of which are represented in the
problems; others are found in Chapter 5 of Student Mathematics Handbook.
TABLE 9.1 Directory of Polar-Form Curves
CARDIOIDS r = α(l ± cos θ) and r = α(l ± sin θ)
r = a - a cos θ r = a + a cos θ r = a - a sin θ, r = a + a sin θ,

standard form π rotation κ ⅛
2 rotation -=- rotation
ROSE CURVES LEMNISCATES

r = a cos nθ or r = a sin nθ r 2 = α 2cos2θ or r 2 = α 2sin20
Two loops
r 2 = a 2 cos 2θ
r = a sin θ standard form
π
2 rotation,
one (circular) petal
r = a cos 3 θ r = a sin 3θ r = a cos 2θ

standard form, 71rotation, standard form,
zo
three petals four petals
three petals
PROBLEM SET 9 .2
θ 1. ■ What Does This Say? Describe a procedure for 41. /(0) = 2∣cos 0∣, 0 ≤ 0 ≤ 2τ7
graphing polar-form curves. 42. /(0) = 4∣sin 0∣, 0 ≤ 0 ≤ 2τ7
2. ■ What Does This Say? Discuss symmetry in the graph 43. /(0) = V∣cos 0∣, 0 ≤ 0 ≤ 2 TT; lazy eight
of a polar-form function.
44. /(0) = Vcos 20, 0 ≤ 0 ≤ 2 77; lemniscate
3. ■ What Does This Say? Compare and contrast the
forms of the equation forlimaςons, cardioids, rose curves, Graph the set o f points (r, θ) so that the inequalities in Problems
and lemniscates. 45-54 are satisfied.
4. Identify each of the curves as a cardioid, rose curve (state 45. 0 ≤ r ∙ I, 0 ≤ 0 < 2∙π
number of petals), lemniscate, limaςon, circle, line, or 46. 2 ≤ r ≤ 3, 0 ≤ 0 < 2-τr
none of the above. 47. 0 ≤ r < 4, 0 ≤ 0 ≤ J
a. r2 = 9 cos20 b. 7 = 2 sin
48. 0 ≤ r ≤ 4, 0 ≤ 0 ≤ 77
c. r = 3 sin30 d. r = 30 49. r > I, 0 ≤ 0 < 2 T7
e. r = 2 - 2 cos 0 f. 0 = f 50. r ≥ 2, f ≤ 0 ≤ 77
g. r 2 = sin 20 h. r - 2 = 4 cos 0
51. 0 ≤ 0 ≤ f > r ≥ 0
5. Identify each of the curves as a cardioid, rose curve (state
52. —7 7 ≤ 0 < 7 7 , r ≥ 0
number of petals), lemniscate, limaςon, circle, line, or
53. 0 ≤ 0 ≤ f> 1 ≤ r ≤ 2
none of the above.
a. r = 2 sin20 b. r2 = 2 cos20 54. 0 ≤ 0 ≤ f > r > 1
c. r = 5 cos60o d. r = 5 sin 80 55. Show that the polar equations
r = cos 0 + 1 and r = cos 0 - 1
e. rθ = 3 f. r2 = 9 cos(20 -
have the same graph in the xy-plane.
g. r = sin 3(0 + f) h. cos 0 = 1 - r
56. Spirals are interesting mathematical curves. There are
6. Identify each of the curves as a cardioid, rose curve (state three special types of spirals:
number of petals), lemniscate, limaςon, circle, line, or
a. A spiral of Archimedes has the form r = aθ', graph
none of the above.
r = 20.
a. r = 2 cos 20 b. r = 4 sin 30°
b. A hyperbolic spiral has the form rθ = a; graph rθ = 2.
c. r + 2 = 3 sin 0 d. r + 3 = 3 sin 0
c. A logarithmic spiral has the form r = a kθ-, graph r = 2 θ.
e. 0 = 4 f. 0 = tan f 57. The strophoid is a curve o f the form
g. r = 3 cos 50 lι. r cos 0 = 2 r = a cos 20 sec 0
Graph the polar-form curves given in Problems 7-38.
Graph this curve where a = 2.
7. r = 3, 0 ≤ 0 ≤ ■ 727 8. 7 = —1 , O ≤ 0 ≤ 7 7
58. The bifolium has the form
9. 0 = - f ’ 0 ≤ r ≤ 3 10. 0 = f 1 ≤ 7 ≤ 2
11. r = 0 + 1, 0 ≤ 0 ≤ 77 12. 7 = 0 - 1, 0 ≤ 0 ≤ r = a sin 0 cos2 0
13. r = 20, 0 ≥ 0 14. r = f > 0 ≥ 0
Graph this curve where a = 1.
15. r = 2 cos 20 16. r = 3 cos 30
59. The folium of Descartes has the form
17. r = 5 sin 30 18. r2 = 9 cos 20
19. r2 = 16 cos 20 20. 72 = 9 sin 20 r _ 3a sin 0 cos 0
21. r = 3 cos 3(0 - 3) 22. r = 2 cos 2(0 + sin 3 0 + cos3 0
23. r = 5 cos 3(0 - f) 24. r = sin 3(0 + Graph the curve where a = 2.
60. The oval of Cassini has the form
25. r = sin (20 + 26. 7 = cos (20 +
r4 + ⅛4 - 2⅛2r2 cos 20 = k 4
27. r 2 = 16 cos 2(0 - f ) 28. r2 = 9 cos (20 - j)
29. r = 2 + cos 0 30. 7 = 3 + sin 0 Graph the curve where b = 2, k = 3.
31. r = l + sin 0 32. 7 = 1 + cos 0
In Problems 61-66, graph the given pair o f curves on the same
33. r cos 0 = 2 34. 7 sin 0 = 3
coordinate axes. The first equation uses (x, y) as rectangular
35. r = l + 3 cos 0 36. 7 = 1 + 2 sin 0 coordinates and the second uses (r, θ) as polar coordinates.
37. 7 = - 2 sin 0 38. 72 = -co s 20 61. y = cosx and r = cos 0
© Sketch the graph o f the polar function given in Problems 39-44. 62. y = sinx and r = sin 0
39. /(0) = sin 20, 0 ≤ 0 ≤ f ; rose petal 63. y = tan x and r = tan 0
40. /(0) = ∣sin20∣, 0 ≤ 0 ≤ 277; four-leaf rose 64. y = secx and r = sec 0
Sec. 9.3 Area and Tangent Lines in Polar Coordinates 5 85
65. y = cscx and r = esc θ

required to get the entire graph? Attempt to ex
66. y = cotx and r = cot θ
plain this. Hand in your favorite graph.
b. Now do a study for m = 1, 2, . . . and n = 1, 2,
Computational Window . . . and attempt to generalize the situation regard
ing the period of the sine functions, the necessary
Exploring the periodicity o f polar plots II is surprisingly 6 interval, and the number of loops in the graphs.
difficult to predict the periodic behavior o f polar plots. For Hand in your favorite graph.
example, here you are to investigate the behavior o f the c. A challenge Attempt to carefully explain the
graphs o f the form necessary 6 interval in part b.
r = f(θ) = sin mθ
for m a positive integer. θ 69. Show that the curve r = f(θ) is symmetric with respect to
the Λ^-axis if the equation is unaffected when r is replaced
67. Consider r = /(6) = sin 26.
by - r and 6 is replaced by π - 6.
a. Since sin 26 has period ττ, one would think that
70. Show that the curve r = f(θ) is symmetric with respect to
surely the graph repeats itself for 6 > τr. Graphi
the y-axis if the equation is unaffected when r is replaced
cally show that this is not the case. Verify that it is
by - r and 6 is replaced by -6 .
necessary for 0 < 6 < 2τr to obtain the entire
graph. Describe the graph. 71. Show that the curve r = /(6) is symmetric in the line
y = x if the equation is unaffected when 6 is replaced by
b. Explain why the interval 0 < 6 < π is not
f - 6.
enough.
72. State and prove a theorem regarding the relationship of
c. Now, systematically study r = f(θ) = sin m x for
the graph of r = /(6) to that of r = /(6 — 60) for
m a positive integer. In particular, describe why m
being even or odd makes a fundamental differ θ κ~ ¾ < 5 ’
ence. Summarize the number of leaves on the 73. a. Show that if the polar curve r = f(θ) is rotated about
roses relative to m. the pole through an angle a, the equation for the new
curve is r = f(θ - a).
68. Continuing the theme of Problem 67, consider
b. Use a rotation to sketch
r - ∕ ( f> ) - s in(!≡⅛)
r = 2 sec (6 -
where m and n are positive integers (and relatively 74. Sketch the graph of
prime). Let P denote the period of the sine.
a. First set m = 1 and study the behavior of the
graphs for n = 1, 2, 3, . . . . Clearly, the period of In particular, show that the graph has a vertical asymptote
the sine function is P = 2nπ. But what 6 interval is at 6 = f ∙
9 .3 AREA AND TANGENT LINES IN POLAR COORDINATES
IN THIS SECTION Intersections of polar-form curves, area bounded by

polar graphs, tangent lines
We have introduced the common polar curves and have considered symmetry
and rotations as aids to graphing. In this section we consider the intersection of
polar-form curves and the area of the region bounded by polar curves. We
conclude by considering tangent lines to polar curves.
■ INTERSECTION OF POLAR-FORM CURVES
In order to find the points of intersection of graphs in rectangular form, you need
only find the simultaneous solution of the equations that define those graphs. It is
not even necessary to draw the graphs, because there is a one-to-one correspon
dence between ordered pairs satisfying an equation and points on its graph.
However, in polar form, this one-to-one property is lost, so that without drawing
the graphs you may fail to find all the points of intersection. For this reason, our
method for finding the intersection of polar-form curves will include sketching the
graphs.
E X A M P L E 1 Finding the intersection of two polar-form curves
Find the points of intersection of the circles r = 2 cos θ and r = 2 sin θ for
0 ≤ θ < 2 77.
Solution First, consider the simultaneous solution of the system of equations.
2 cos θ = 2 sin θ
1 _ 2 sin θ
2 cos θ
1 = tan θ
Then find r using either of the two given equations:

r = 2 cos f = V2 and r = 2 cos ¾r = - V z2
This gives the points (VΣ, f) and ( - V z2, ¾r). We see that these ordered pairs
are primary representations of the same point. Thus, the simultaneous
solution yields one point of intersection.
Figure 9.13 Graphs of r = 2 cosθ Next, consider the graphs of these circles, shown in Figure 9.13. It looks as
and r = 2 sin θ though (0, 0) is also a point of intersection. Check this point in each of the
given equations:
If r = 0 in r = 2 cos θ, then θ = 7.
If r = 0 in r = 2 sin θ, then θ = 77.
At first it does not seem that the pole satisfies the equation, because r = 0 gives
(θ, ⅞) and (0, 77), respectively. Notice that these coordinates are different and
do not satisfy the equations simultaneously. But, if you plot the point with the
coordinates, you will see that (θ, f), (0, 77), and (0, 0) are all the same point.
It’s as if two ants were crawling along the curves and came to the origin at
different times— they would pass each other without colliding. M M
The pole is often a solution for a system of polar equations, even though it
may not satisfy the equations simultaneously. This is because when r = 0, all
values of θ will yield the same point— namely, the pole. For this reason, it is
necessary to check separately to see if the pole lies on the given graph.
Graphical Solution Step 1. Find the simultaneous solution of the given system of equations.
of the Intersection Step 2. Determine whether the pole lies on the two graphs.
of Polar Curves Step 3. Graph the curves to look for other points of intersection.
E X A M P L E 2 Intersection of polar form curves
Find the points of intersection of the curves r = ∣ - cos θ and θ = 3 r.

Solution Step 1: Solve the system by substitution:
r = ∣-c o s ⅛ = j - ( 4 ) = 2
Sec. 9.3 Area and Tangent Lines in Polar Coordinates 587
The solution is (2, y ) .

Step 2: If r = 0, the first equation has no solutions because
0 = 2 - cos θ or cos θ = ⅜
and a cosine cannot be larger than 1.

Step 3: Now look at the graphs, as shown in Figure 9.14. From the graphs we
see that (- l, ⅛r) may also be a point of intersection. It satisfies the equation
θ = y ’ but what about r = ∣ - cos 0? Check ( - 1, y ) :
-1 - 5 - ∞ s ( τ )
Figure 9.14 Graphs of
c* = 3_ / n
n
∣rq - cos θ and θ = 2f 2 1 2)
= 2 Not satisfied
However, if you check the alternative representation ( -1 , y I = (1, y ) :
1 ⅛j - cos(¾r)
- 3 _ 1
- 2 2
Satisfied
Be sure to check for points of intersection that you may have missed,
■ AREA BOUNDED BY POLAR GRAPHS
In order to find the area of a region bounded by a polar graph, we use Riemann
sums in much the same way as when we developed the integral formula for the
area of a region described in rectangular form. However, instead of using
rectangles as the basic unit being summed in polar form, we sum the area of sectors
of a circle. Recall from trigonometry the formula for the area of such a sector.
Area of a Sector The area of a circular sector of radius r is given by
A = ⅛r2 θ
where θ is the central angle of the sector measured in radians.
This formula gives the following proportions:
Angle Area Figure

Using this formula for the area of a sector, we now state a theorem that gives
us a formula for finding the area enclosed by a polar curve.
THEOREM 9.1 Area in polar coordinates
Let r = f(θ) define a polar curve, where/is continuous and∕(0) ≥ 0 on the closed
interval a ≤ θ ≤ β. Then the region bounded by the curve r = /(0) and the rays
θ = a and θ = β has area
. rβ 2 ∣ rt>
A = ∣ r dθ = ± ∖ [f(θ)] 2 dθ
^Ja “ Ja
a. The region bounded by the Proof: The region is shown in Figure 9.15. To find the area bounded by the
polar curve r= f(θ ) and
the rays θ = a and θ= β graphs of the polar functions, partition the region between θ = a and θ = β by a
collection of rays, say θ0, 01, . . . , θn . Let a = θ0 and β = θn .
Pick any ray θ = θ'k , with 0jt-1 ≤ θk ≤ θk . Then the area ΔΛ fc of the circular
sector is approximately the same as the area of the region bounded by the graph of
/ and the lines θ = θk ~ l and θ = θk . Because this circular sector has radius∕(0 * ) and
central angle ∆θk , its area is
ΔA i = ∣(radius)2(central angle) = ∣[∕(0⅛)]2Δ0A
The sum ∑ ΔA, is an approximation to the total area A bounded by the polar
k= 1
curve, and by taking the limit as n → ∞, we obtain
n 1 n ι (,β
b. The area A can be estimated A = lim ∑ Δ +k= 4 1 i m ∑ [∕(6U)] 2Δ6>. = / [∕(0)] 2 dθ _
by adding the area of ''small" n→ ≈ ⅛=1 z → ∞ ⅛=∣ κ κ Z Ja ■
n
(pink) circular sectors.
Figure 9.15 Area in polar form The process is familiar, and you will find that the most difficult part of the
problem is deciding on the limits of integration. A decent sketch of the region will
help you do this.
EXAM PLE 3 Finding area enclosed by part of a cardioid
Find the area under the top half (0 ≤ θ ≤ ττ) of the cardioid r = 1 + cos θ.
Solution The cardioid is shown in Figure 9.16. Note that the top half of the graph
lies between the rays θ = 0 and θ = τr. Hence the required area is given by
A = A [ (1 + cos 0)2 dθ = X ∣ (1 + 2 cos θ + cos2 θ) dθ
Jo 2 Jo
= ^ 0 + 2 sin θ + 5 + ≡ 2 θ Integration table, formula 317
Figure 9.16 Area under the top = ∣ [ 1r + 2 ( θ ) + f + ∣ - θ ] = ⅜ _

half of the cardioid / - = 1 + cos θ
EXAM PLE 4 Finding area enclosed by a four-leaved rose
Find the area enclosed by the four-leaved rose r = cos 20.

Solution The rose curve is shown in Figure 9.17. We will find the area of the top
half of the right loop (shaded portion) in order to make sure/(0) ≥ 0. We see
this corresponds to angles with measures from θ = 0 to θ = f . By symmetry,
the entire area enclosed by the four-leaved rose is 8 times the shaded region.
Thus, the required area is given by
Γι f π74 2 * 1
A=8 5 cos 20 dθ
Jo -I
r . -ι π∕4
= 4 ^ + s^1g J Integration table, formula 317
= 4⅛ + O -θ l
_o J
Figure 9.17 Area enclosed by the

four-leaved rose r = cos 20 EXAMPLE 5 Finding the area of a region defined by two polar curves
Find the area of the region common to the circles r = a cos θ and r = a sin θ.
Solution The circles are shown in Figure 9.18. Note that they intersect at θ = f
and at the pole (both by inspection). In order to set up the integrals properly,
remember to think in terms of polar coordinates and not in terms of
rectangular coordinates. Specifically, we must scan radially. This means that
we need to find the area of intersection by finding the sum of the areas of the
regions marked R l and R 2. Region R l is bounded by the circle r = a sin θ and
the rays θ = 0, θ = f. Region R, is bounded by the circle r = a cos θ and 0 = f ,
θ = -j. Note that the rays θ = 0 and θ = J are not necessary as geometric
boundaries for R l and T?2, but they are necessary to describe which part of
each circle is being defined.
A = AREA OF R 1 + AREA OF R 2
rττ∕4 . rπ ∕2
α2sin2 θ dθ + 24 α2cos2 0r∕0
∣0 Jπ ∕4
Figure 9.18 Area enclosed by r -ι ττ∕4 -ι
2
_ α 0 _ sin 20 2r
, α 0 , sin20
r = a cos θ and r = a sin θ Integration table, formulas 348 and 317.
2 [2 4 J 0 2 |_2 4 J
_ α2Γ7r _ 1
2 Lδ 4
= ⅛a2(π - 2)
EXAMPLE 6 Finding the area between a circle and a limaςon
Find the area between the circle r = 5 cos 0 and the limaςon r = 2 + cos 0.
Round your answer to the nearest hundredth of a square unit.
Solution As usual, begin by drawing the graphs, as shown in Figure 9.19. We see
that because both the lima90n and circle are symmetric with respect to the x-
axis, we can find the area in the first quadrant and multiply by 2.
Next, we need to find the points of intersection. We see that the curves do
not intersect at the pole. Now solve
5 cos 0 = 2 + cos 0
cos 0 = y
Figure 9.19 Area between
0= Solution in the first quadrant
r = 5 cos θ and r = 2 + cos θ
In order to find the area, divide the region into two parts, along the ray θ = j .
The right part (shaded red) is bounded by the limaςon r = 2 + cos θ and the
rays θ = 0 and θ = j . The left part (shaded blue) is bounded by the circle
r = 5 cos θ and the rays θ = j and θ = f . Using this preliminary information
we can now set up and evaluate the required integral.
A = 2[AREA OF RIGHT PART + AREA OF LEFT PART]
rπ∕3 . rπ∕2
= 2 -
2 L2 (2 + cos θ)2 dθ + 2⅛ (5 cos 0)2 dθ
Io Jπ ∕3
rττ∕3 rπ ∕2
(4 + 4 cos θ + cos2 θ) dθ + 25 cos2 θ dθ
∣0 Jτr∕3
. _ -∣ ττ∕2
= [40 + 4 sin 0 + $ + θλ , sin 20 I
2+ 4 J ∣ 7, 3
∣^4-π∙ , 4V z3 +, -77+, V 3 - 0 ∩1 , ∏cΓ-7τ,∩ π

-⅛ + ~ 6 T J + 214 + 0 - 6
_ 9ττ , 17√3 , 25τr _ 25√3 = 43π

6 8 12 8 12
■ TANGENT LINES
0 The slope at (r, 0) of r = /(0) In rectangular coordinates, the slope of the tangent line to a graph of y = g(x) at
is N O T ∕'(0 ). 0 χ = x fl is g'(∙⅞)> and we might expect the slope of the tangent line to the graph of
r = /($) at 0 = 0o to be ∕'( 0 o ). A simple counterexample shows that this is false.
The graph of the polar function r = c (c > 0, a constant) is a circle centered at the
origin. Because f is a constant, we have∕'( 0 ) = 0 for all 0, so that if∕'( 0 ) were the
slope of the tangent line to this circle, every tangent line would be horizontal,
which is clearly false. It turns out that the slope of the tangent line to a polar curve
is given by the following more complicated formula.
T H E O R E M 9.2 Slope of polar curves
If /(0) is a differentiable function of 0, then the slope of the tangent line to the
polar curve r = /(0) at the point P(r0, 0o) is given by
∕( 0 o ) cos0 o + ∕'( 0 o) sin 0o

/77 = -----------------------------------
- / ( ¾) sin0 o + ∕( 0 o) cos0o
whenever the denominator is not zero.

Proof: Because r = /(0) and x = r cos 0, y = r sin 0, we have
x = /(0) cos 0 and y = /(0) sin 0
Using the chain rule, we find that = o r > equivalently,
dy
dy = dθ
dx dx
dθ
Because x = /(θ) cos θ and y = / (θ) sin θ, it follows that

⅛ = ∕(θ ) ⅛ (c°s 0) + cos e < ⅛ =/(0) ⅛(si∏ 0) + sin 0 ⅛
= -/(0) sin 0 + ∕'( 0 ) cos 0 = /(0) cos 0 + ∕'( 0 ) sin 0

and the slope of the tangent line is given by:
dy
_ dy_ _ dθ, _ /(0) cos 0 + ∕'( 0 ) sin 0
dx dx -/(0) sin 0 + ∕'( 0 ) cos 0
dθ
The theorem for slope in polar form is not easy to remember, and it may be
easier for you to remember dy
dy = dθ
dx dx
dθ
where x = /(0) cos 0 and y = ∕(θ)sinθ. With this form, notice that horizontal
tangents (dy∕dx = 0) occur when = 0 and ⅛ ≠ 0, and vertical tangents occur
when = 0 and ≠ 0. If ~ = 0, then no conclusions can be drawn

without further analysis.
EXAM PLE 7 Computing the slope of a cardioid at a given point
Find the slope of the tangent line to the cardioid r = 1 + cos 0 at the point
V 2 ’ 6/
Solution First find x and y as functions of 0:
x = /(0) cos 0 = (1 + cos 0)cos 0 = cos 0 + cos2 0
y = ∕(0)sin 0 = (1 + cos 0)sin 0 = sin 0 + cos 0 sin 0 = sin 0 + ∣ sin20
Then find the derivatives and

dy
= -s in 0 - 2 cos 0 sin 0 = cos 0 + cos 20
= -s in 0 —sin 20
At 0 = we have
dx _ - s i n f - s i n f = - 1 - ^
dθ 5 2
dy _ C O S^ + COS∙J = + y =
~tθ~
so the slope is
. ⅜ √3 + l
m _d y _d O _ _ .
dx dx - ι - √3
Figure 9.20 Graph of dθ 2
r = 1 + cos θ and line passing
/ Λ ∕+ τr∖ The graph of the cardioid, and the tangent line at 0 = / are shown in Figure
through I 1 + - y - ’ I with slope -1
9.20. — ∣
EXAM PLE 8 Finding tangents to a spiral
Find all numbers θ with 0 ≤ θ ≤ π so that the tangent line to the spiral r = θ
is horizontal.
Solution We first look at the graph (see Figure 9.21) so that we know what to
expect. We now carry out the analytic work:
x = f(θ) cos θ = θ cos θ and y = f(β) sin θ = θ sin θ, so that
dx _ - θ sin θ + cos θ dv
and ~dQ- θ cθs 0 + sin 0
dθ
Thus,
dy
-2 -1 1 _ dθ _ θ cos 0 + sin 0
dx - 0 sin 0 + cos 0
dθ
Figure 9.21 By drawing a spiral The tangent line is horizontal when m = 0, which is equivalent to
with 0 ≤ θ ≤ ττ, it looks like
there is a horizontal tangent at the
pole and at one other point. 0 cos 0 + sin 0 = 0
0 cos 0 = -sin 0
θ= cos 0 cos 0 ≠ 0
0 = -ta n 0
- 0 = tan 0
This is not an equation we are used to solving, so we can turn to methods of

approximation. Either Newton’s method (Section 2.9) or a graphical solution
is possible, as shown in Figure 9.22.
a. Newton’s method for finding b. On a graphing calculator,

the solution o f the equation graph y = tan θ and
- θ - tan 0; the approximate y =-θ', use ∣TR ACE∣
solution is 0 = 2.0288. to approximate the solutions:
θ =2.0503026 and 0.
Figure 9.22 Approximating the roots of - 0 = tan 0
Figure 9.23 Graph of r = θ ≥ 0 We now find the points on the spiral corresponding to the solution shown in
and the two horizontal tangents Figure 9.22, namely, 0 = 0 and approximately 0 = 2.03, as shown in Figure
for 0 ≤ θ ≤ π 9.23. —b
Sec. 9 .3 Area and T angent L ines in Polar C oordinates 593
PROBLEM SET 9 .3
© 1. ■ W hat D oes T his Say? Discuss a procedure for finding
36. /(0) = γ 0 ≤ 0 ≤ 2ττ
the intersection of polar-form curves.
«2
2. ■ W hat D oes T his Say? Discuss a procedure for finding 37. /(0) = —’ 0 ≤ 0 ≤ 2τr
77
the area enclosed by a polar curve.
In Problems 38-47, find the slope o f the tangent line to the
3. ■ W hat D oes T h is Say? Discuss why the slope of the graph o f the polar function curves at the given point.
tangent line of the polar function f(θ) is not ∕'(0 ). 38. /(0) = 1 - cos 0 at f)
Find the points o f intersection o f the curves given in Problems
4-29. You need to give only one primary representation for 39. /(0) = 4 cos 0 + 2 at the pole
each point o f intersection (give the one with a positive r and 40. /(0) = Vcos 20 at the pole
with 0 ≤ 0 < 2π). 41. /(0) = 2 at (2, f)
4. j[r = 4 cos 0 lr = 8 cos 0
5. < 42. r = 2 sec 0, where 0 = f
Lr = 4 sin 0 lr = 8 sin 0
43. r = 0, where 0 = J
Ir = 2 cos 0 7. jfr = 4 sin 0 4
6. < 44. r = ~—:—7— 7----7’ where 0 = τr
[r = 1 lr = 2 3 sin 0 - 2 cos θ
3
Γr2 = 9 cos 20
8 ,∣ lr 2 = 4 sin 20 45. r - 7---- „ , -—:—7’ where 0 = π
2 cos 0 + 3 sin 0
. r= 3 I r=2
46. r = 4 sin 0 cos2 0, where 0 = j
‘ » •1 lr = 2(1 + cos 0) lr = 2(1 + sin 0) 47. r = 2 cos 0 sin 2 0, where 0 = f
11.
lr = 2(1 - cos 0) lr = 2(1 —sin 0) θ 48. Find all points on the cardioid r = 1 + sin 0 where the
2
12. |∖r2
= 9 sin20 1 3 .1 Ir = 4 sin 20 tangent line is horizontal.
Ir= 3 I r = 2V2 cos 0 49. Find all points on the cardioid r = a(l + cos 0) where the
tangent line is horizontal.
14. j∣
r2 = 9 cos 20 lr 2 = 4 cos 20
15. 1 50. Find all points on the cardioid r = a(l - cos 0) where the
L r = 3V2 sin 0 I r=2
cn tangent line is vertical.
I
16. ∣ h 17. jlr 2 = cos 20 51. Find all points on the circle r = 2 sin 0 where the tangent
I ∣m
I r = V2 sin 0
line is parallel to the ray 0 = f ∙
18. ∣
fr = 2(1 —cos 0) 19. j[r = 2( 1 + cos 0) 52. Find the area of one loop of the four-leaved rose
lr = 4 sin 0 lr = - 4 sin 0 r = 2 cos 20.
20. jlr = 2(1 - sin 0) lr = 2 cos 0 + 1 53. Find the area enclosed by the three-leaved rose
21. < r = a sin 30.
lr = 4 cos 0 lr = sin 0
5 54. Find the area of the region that is inside the circle
22. jlr = 2 sin 0 + 1 fr
23. - 3 - cos 0 r = 4 cos 0 and outside the circle r = 2.
lr = cos 0 lr = 2 55. Find the area of the region that is inside the circle r = a
2 4
and outside the cardioid r = α(l - cos 0).
(r fr
24. 1 1 + cos 0 25. <j ' 1 - cos 0 56. Find the area of the region that is inside the circle
lr = 2 l r = 2 cos 0 r = sin 0 and outside the cardioid r = 1 - cos 0.
1
57. Find the area of the region that lies inside the circle
(r
27. jIr = 2 cos 0 r = 6 cos 0 and outside the cardioid r = 2(1 + cos 0).
26. < 1 + cos 0
l r = 2(1 - cos 0) lr = sec 0 58. Find the area of the portion of the lemniscate
r2 = 8 cos 20 that lies in the region r ≥ 2.
28. jlr = 2 sin 0 fr sin 0 = 1
29. ∙ 59. Find the area between the inner and outer loops of the
lr = 2 esc 0 lr = 4 sin 0
limaςon r = 2 - 4 sin 0.
Find the area o f each polar region enclosed by f(θ), θ = a, θ = b θ 60. Find an equation in 0 for all polar-form points (r, 0) on the
for a ≤ θ ≤ b in Problems 30-37. spiral r = 02 where the tangent line is horizontal. What
30. f(θ) = sin θ, 0 ≤ θ ≤ f happens at the origin?
31. f(θ) = cos θ, 0 ≤ 0 ≤ f 61. Let a and b be any nonzero numbers. Show that the
32. /(0) = sec0, - f ≤ θ ≤ % tangent lines to the graphs of the cardioids
33. f(θ) = √s⅛ 0, f ≤ θ ≤ f r. ~ a( 1 + sin 0) and r2 = ⅛( 1 - sin 0)
34. f(θ ) = e el2, 0 ≤ θ ≤ 2ττ are perpendicular to each other at all points where they
35. f(θ) = sin 0 + cos θ, 0 ≤ θ ≤ f intersect.
62. Because the formula for the slope in polar-form is compli

cated, it is often more convenient to measure the inclina
tion of the tangent line to a polar curve in terms of the
angle αextending from the radial line to the tangent line,
as shown in Figure 9.24.
Let P be a point on the polar curve r = f(ff), and let a be
the angle extending from the radial line to the tangent
line, as shown in Figure 9.24. Assuming that f'(θ) ≠ 0,
/ ( θ)
show that tan a = r , , ,.
./ W
63. Use the formula in Problem 62 to find tan a for each of
the following:
a. the circle r = a cos θ
b. the cardioid r = 2(1 - cos θ)
c. the logarithmic spiral r = 2e3β
9.4 PARAMETRIC REPRESENTATION OF CURVES
IN THIS SECTION Parametric equations, derivatives, arc length, area

Up to now, each curve we have discussed has been represented by a single
equation. In this section we see how we can consider both x and y as functions
of some other variable called a parameter. After we consider graphing curves
with a parameter, we use calculus to find derivatives, arc length, and area for
curves defined with a parameter.
S¾"⅛SΓ*SSJMrf,'βSJ¾.5. S r∣Λ W ⅛ JV K 3 5 < ⅛ "Λ ,'.∙. . --
Computational Window ~ π
-
■ PARAMETRIC EQUATIONS
Most graphing calculators will It is sometimes useful to define the variables x and y in the ordered pair (x, y) so
sketch curves in parametric that they are each functions of some other variable, say t. That is, let
form. Check your owner’s
manual, but the usual proce x = g(f) and y = h(f)
dure is to change the (MODE) for functions g and h, where the domain of these functions is some interval I. The
from function format to para variable t is called a parameter and x = g(t) and y = h(t) are parametric equations.
metric form.
Parametric Representation Let f and g be continuous functions of t on an interval /; then the equations
for a Curve
x = ∕( 0 and y = g(t)
are called parametric equations for the curve C generated by the set of ordered
pairs (%(Z), y(t))∙
EXAMPLE 1 Graphing a curve from parametric equations
Graph the curve defined by the functions x = 2 cos θ, y = 2 sin θ for

0 ≤ θ < 2ττ.
Solution We could begin by letting θ be the parameter to determine x and y

values. When we draw the graph, the parameter is not shown because we still
graph the ordered pairs (x, y) in the usual fashion (see Figure 9.25).
Sec. 9.4 Parametric Representation of Curves 595
0 X y
0 2 0
1 1.08 1.68
77
? 0 2
2 -0.83 1.82
77 -2 0
3 77
2 0 -2
If you are using a computer or a graphing calculator, plotting points can be an
efficient way of obtaining the graph, but sometimes it is more efficient to
eliminate the parameter.
Figure 9.25 Graph of x = 2 cos θ,
y = 2 sin θ x 2 = 4 cos2 θ
y 2 = 4 sin2 θ
Add: x 2 + y 2 = 4 cos2 0 + 4 sin2 θ = 4(cos2 θ + sin2 θ) = 4
We recognize this as the equation of a circle centered at (0, 0)
with radius 2. ■■■
If we start with the parametric equations x = 2 cos θ, y = 2 sin θ, we see from

Example 1 that after eliminating the parameter we obtain x 2 + y 2 = 4. On the
other hand, we say that the x = 2 cos θ. y = 2 sin 0 is a parametrization of the curve
x 2 + y 2 = 4.
Figure 9.26 Graph of x - R cos θ, y = R sin θ

Notice that the curve in Example 1 is not the graph of a function, but it is
defined parametrically using functions. This points out one of the advantages of
parametric equations—they can be used to represent graphs that are more general
than graphs of functions.
In practical applications it is often possible to obtain useful information about
a parametrized curve by seeing how it is traced out as the parameter varies from
one value to another. For Example 1 we can generalize to say that the parametric
equations
x = Rcosθ y = Rsinθ R > 0 constant
trace out a circle of radius R and center (0, 0) as 0 varies from 0 to 2τr. We can
describe this trace by starting at 0 = 0, the point (7?, 0), and ending at 0 = 2τr, as
shown by the arrows in Figure 9.26. We shall refer to this trace (oriented curve)
as the path of x = f{t~), y = g(t) for a ≤ t ≤ b from the point (∕(α), g(a)) to
(∕(⅛), g(Z>))∙
EXAM PLE 2 Sketching the path of a parametric curve
Sketch the path of the curve x = Z2 - 9, y = ∣Z for - 3 ≤ t ≤ 2.
Solution Values of x and y corresponding to various choices of the parameter t

are shown in the following table:
t x y
- 3 0 -1 (Starting point)
- 2 -5 -j
- 1 -8 -∣
0 - 9 0
1 -8 ⅜
2 -5 j (Endingpoint)
The graph is shown in Figure 9.27. Figure 9.27 Graph of x = t2 - 9,

y = ∣Z, for - 3 ≤ t ≤ 2. Notice the
arrows show the orientation
as t increases from - 3 to 2.
The parametrization of an equation is not unique. For example, the set of

parametric equations
x = 9(9z2 - 1), y = 3z for - ∣ ≤ t ≤ ∣
has the same graph as the one given in Example 2. If we consider t as measuring
time, we see that the curve is traced out much more rapidly with this set of
equations than the ones given in Example 2. When a parameter is used to describe
motion, we call the path a trajectory. In many applications we use a parameter to
represent different speeds at which objects can travel along a given path.
Sometimes a curve defined by parametric equations can be analyzed by the
algebraic process of eliminating the parameter.
EXAM PLE 3 Sketching the path of a parametric curve by eliminating the

parameter
Describe the path x = sinτrt, y = cos 2πt for 0 ≤ t ≤ 0.5.
Solution Using a double-angle identity, we find

cos 2τrZ = 1 - 2 sin2 τrt
so that
y = 1 - 2x 2
We recognize this as the Cartesian equation for a parabola. Because y ' = -4 x ,
we can find the critical value x = 0, which locates the vertex of the parabola at
(0, 1). The parabola is the dashed curve shown in Figure 9.28. Because t is
Figure 9.28 The parabolic arc restricted to the interval 0 ≤ Z ≤ 0.5, the parametric representation involves
x = sin πt, y = cos 2πt for only part of the right side of the parabola y = 1 - 2x 2 . The curve is oriented
0 ≤ t ≤ 0.5. from the point (0, 1), where Z = 0, to the point (1, -1 ), where Z = 0.5. M M
EXAM PLE 4 Adjusting the domain after eliminating a parameter
Sketch the graph of x = 2 z, y = 2 z+1 for t > 0.

Solution Eliminating the parameter, we obtain
- = t = 2 z+1-z = 2 so that y = 2x
x 2
0 This answer is not complete until we examine the domain of t. 0
For the given parametric equations we see that the domain for t must be
positive numbers, because 2' is not defined in our example for t ≤ 0.
Figure 9.29 Graph of x = 2 z, However, the equation y = 2x is defined for all values of x and y, so we must
y = 2 z+1 adjust the domain of this equation. If t > 0, then x > 1 and the equation is
y = 2x, x > 1. The graph is shown in Figure 9.29. c=□
When it is difficult to eliminate the parameter from a given parametric

representation, we can sometimes get a good picture of the parametric curve by
plotting points.
EXAM PLE 5 Describing a spiraling path with a decreasing amplitude
Discuss the path of this curve described by the parametric equations

x = e - z cos t, y = e^zsin t for t ≥ 0
Solution We have no convenient way of eliminating the parameter so we write
out a table of values (Λ', y) that correspond to various values of t. The curve is
obtained by plotting these points in a Cartesian plane and passing a smooth
curve through the plotted points, as shown in Figure 9.30.
Note that for each value of t, the distance from P(x, y) on the curve to the
origin is
Vx 2 + y 1 = V(e^zcost)2 + (e~zsinZ)2 = Ve -2 z (l) = e~t
Because e~t decreases as t increases, it follows that Pgets closer and closer to
the origin as t increases. However, because cos t and sin t vary between -1 and
+ 1, the approach is not direct but takes place along a spiral. M M
EXAM PLE 6 Finding parametric equations for a trochoid
A bicycle wheel has radius a and a reflector is attached at a point P on the

spoke of a bicycle wheel at a fixed distance d from the center of the wheel,
which has radius a. Find parametric equations for the curve described by P as
the wheel rolls along a straight line without slipping. Such a curve is called a
trochoid.
Solution Assume that the wheel rolls along the x-axis and that the center C of the
wheel begins at (0, a) on the y-axis. Further assume that P also starts on the y-
axis, d units below C . Figure 9.31 shows the initial position of the wheel and
its position after turning through an angle θ.
We begin by labeling some points: The point A is on the x-axis directly

beneath C , whereas B is the point where the horizontal line through C meets
the rim of the wheel. Finally, Q is the point on B C directly beneath P, and 5' is
the point where the line through C and P intersects the rim. Let P have
coordinates (x, y).
We need to find representations (in terms of a, d, and θ) for x and y.
x = ∣O4∣ + ∖C Q ∖
= aθ ÷ ∣C(2∣ Because the wheel rolls along the x-axis
without slipping, ]CL4∣ is the same as the
arc length from A to S, so ∣Ct4∣ = aθ.
y = ∖A C ∖+ ∖Q P ∖
= a + ∖Q P ∖
To complete our evaluation of x and y, we need to compute ∣C<2∣ and ∣(2P∣∙

These are sides of Δ P C Q . Note that ∕-PCQ = 0; therefore, by the
definition of cosine and sine, we have
cos (⅞f - θ') = sθ ∣C Q ∖= d cos (¾r - θ) = ~ d sin θ
Similarly, ∣βP∣ = d sin(⅛yr - θ) = - d cosθ.

We can now substitute these values for ∣C(9∣ and ∣(2P∣ into the equations we
derived for x and y.
x = aθ + ∣Cβ∣ = aθ - d sin θ
y = a + ∣ζλP∣ = a - d cos θ
The special case where P is on the rim of the wheel in Example 6 (when d = a)
is a curve called a cycloid. There are several problems involving these and similar
curves in the problem set.
Two variations in the cycloid path discussed in Exam the assistance of a computer or a graphing calculator.
ple 6 may be described as paths traced out by a fixed However, if you have access to a computer you might
point P on a circle. If it rolls inside a larger circle, the look at an article by Florence and Sheldon Gordon,
trace is called a hypocycloid, and as it rolls around the “ M athematics Discovery via Com puter Graphics:
outside of a fixed circle, the trace is called an epicy Hypocycloids and Epicycloids” in The Two-Year Col
cloid. Even though it is not too difficult to derive their lege M athematics Journal, November 1984, p. 441. If
param etric equations (see Problems 62 and 63), it is a is the radius of the fixed circle and R is the radius of
difficult to graph variations of these curves without the rolling circle, then the param etric equations are as
follows.
Hypocycloid: x = (a - R)cos t + R cos a - R t, y = (α - R)sin t - R sin

R
Epicycloid: = (α + R)cos t - R cos u + R t, y = (<2 + R)sin t - R sin
R
Some variations of these curves, which were drawn by computer, are shown below.
Hypocycloid
P(xW),⅛(f))
-2 --
(«, 0)
-4 -- -8 --
a = 5:R= 1 a = T ,R = 2 o = 28;/?= 13
(0 ≤ t ≤ 4π) (0 ≤ t ≤ 26ττ)
Epicycloid
P W ), y(∕))
0,0) (α,0
a = 7; R = 1 o = 8;/? = 3 o = 28;/?= 13
(0 < t < 6π) (0 ≤ ∕< 2 6 π )
■ DERIVATIVES
We begin by finding the derivative of equations of curves defined by parametric

equations.
THEOREM 9.3 Parametric form of the derivative
If a curve is described parametrically by the equations x = x(t), y = y(t), where x

and y are differentiable functions of t, then the derivative ⅛ can be expressed in
7 √y
terms of and by the equation
dy
dy = lt_ = y∖ t) ∕] γ
whenever ≠ 0
dx dx χ'(t)
dt
Proof: This formula follows directly from the chain rule, ~J~t n
The curve has a horizontal tangent when = 0 and ≠ 0, and it has a
vertical tangent where = 0 and ≠ 0. If = 0 and = 0 at the same point,

further analysis is required.
EXAM PLE 7 Finding the derivative of a function in parametric form
A curve C is described parametrically by x = It + 2, y = ∕ 3 - 12/ for all t. Find

at the point where t = 3, and then find all points of C where the tangent line
is horizontal.
Solution We have ^ = 7 and = 3 ∕ 2 — 12, and thus
dy = dy/dt = 3/2 - 12
dx dx!dt 7
When t = 3, we have
dy 3(3)2 - 12 15
dx t=3 7 7
The curve and the line tangent at the point where t = 3, are shown in Figure
9.32. The tangent line is horizontal when = 0;
3 ∕ 2 - 12 = 0 when t = ±2
At / = 2, x = 7 (2 )+ 2 = 16 and y = 23
- 12(2) = - 16, and at / = - 2 ,
x = 7(-2) + 2 = -1 2 and y = ( -2 ) 3 - 12(-2) = 16, so that the tangent line
Figure 9.32 Graph of x = It + 2, is horizontal at the points (1 6 ,-1 6 ) and (-1 2 , 16). Figure 9.32 also shows the
y = t 3 - 12/ with tangent at t = 3 horizontal tangents. r≡z□
To find the second derivative of a function defined parametrically, we set

y' = and substitute into the derivative formula for parametric equations:
dy'
d 2y = _d_(dy\ = _d_( = dyf = ^df_
dχ 2 d x∖dx) dx^y , dx dx
dt
For instance, in Example 7, we found y ' = = 3r 12 so that
7
d~y = ∣(2∕) = 6/
dx 2 — 49
■ ARC LENGTH
Parametric equations can be used to specify a curve, the speed of an object at a
given time, or the distance an object has traveled. Finding the distance a point
moves along a curve defined by parametric equations requires that we find the arc
length.
THEOREM 9.4 Arc length of a curve described parametrically

Let C be the curve described parametrically by the equations x = x(?) and y = y(?)
on an interval [a, /?], where and are continuous. Suppose C does not intersect
itself (such a C is called a simple curve). Then C has arc length given by
Proof: Recall that the formula for the arc length of a curve C given by y = f(x)
over an interval [a, b} is
•b _____________
λ∕l + [∕'(x )] 2 dx
Use this formula to derive the desired equation. The details are left for the
problem set (Problem 62). ■
Computational Window S B EXAMPLE 8 Finding the arc length of a curve defined by parametric
equations
Let C be the curve defined by x = dsin t, y = c zcos t for 0 ≤ t ≤ ττ. Find the
length of C on the interval [0, τr].
Solution We find that
= e zcos t + ezsin t = e z(cos t + sin t)

dy t, . . t t, . .
X iτ BeA T s in T = e'(-sin?) + c'cos? = eqcos? - sin?)
l√ iτB e ^ T c o ≥ T
Tnin=Θ To obtain the length of the curve we first compute

Trιax=π
T s te p = . Θ30Θ996... dx∖ 1 , — ’ = e 2z(cos t + sin t)2 + e2z(cos t - sin t)2
Xr∣ i n= ^15 .dt) dt
Xnax=25 = e2t(cos2 t + 2 cost sin? + sin2 ? + cos2 ? - 2 cost sin? + sin2 ?)
X≡∙cl=5
V n in = - 25 = 2e2z
Ynax=5
Thus,
You can use the graph to make
an estimate of the arc length.
Imagine that you are standing
on the curve at the point
marked t = 0 and you walk to
the point marked t = rr, how
far did you walk? The answer
is 8 units to the right, 8 units to
the left, and 20 down. This is
about 36 units; compare this
with the calculated answer: ■ AREA
V2(e 7r - I) ≈ 3l.3l 16678 We close this section with a theorem and an example that show how to find the
area under a curve when the curve is defined by parametric equations.
Computational Window ~ π T H E O R E M 9.5 Formula for the area under a curve defined parametrically
Suppose y is a monotonic, continuous function of x on the closed interval [a, b]

and that x and y are both functions of a parameter t, for ∕ 1 ≤ t ≤ t2 . Then if x '(∕)
is continuous on [ ∕ 1, f ] , the area under the curve described parametrically by
x = x(t), y = y(t) over [tp ∕ 7] is given by
∕∙ ∂ rt2
A = y dx = y(t) x'(l) dt
Ja J∕,
Xt τ ≡e^T Proof: This proof is straightforward.
ViτB2e^T+l
T∩in=0 X ≥ cl= l
Tnax=l '⅛ir∣=0
Tstep=.04 Y∩ax=10 EXAM PLE 9 Finding the area under a curve defined parametrically
Xnιn=0 Yscl=10
Xnax=5
Find the area under the curve described by
After drawing the graph o f the
parametric equations, we see x = e t , y = rλe t + 1 over the interval 0 ≤ t ≤ 1
that the graph is linear, so cal
Solution Note that x and y are both continuous and that y(∕) > 0 on [0, 1].
culate the area with a single
trapezoid: fe f1 ,
A τ = ( 1.7182813)(3 + 6 - ^ - - 6- - A ~ y(x) dx = y(∕) x'(f) dt
J∣ Jo
≈ 8.10733783
Compare with (2et + 1) e t dt
e>2 + e - 2 ≈ 8.107337927
The reason for the discrepancy = f (2e~t + c z) dt
is round-off error imposed by Jo
calculator evaluation.
= (e2t + ez)∣θ = e~ + e - 2
PROBLEM SET 9 .4
θ 1. ■ What Does This Say? What is a parameter? 18. x = 4 sec 2t, y = 2 tan 2/, 0 ≤ /<
2. ■ What Does This Say? Describe the process o f finding 19. x = 3 t , y = 3 z+ 1 , / ≥ 0
the derivative in parametric form. 20. x = 2 t , y = 2 ' ~ t , / ≥ 0
F in d an explicit relationship between x a n d y in Problems 3 -2 4 21. x = e l , y = e t + ', /≥ 0
by elim inating the parameter. In each case, sketch the path 22. x = et, y = e 1~z, /≥ 0
described by the parametric equations over the prescribed 23. x = ∕3 , y = 3 In /, / > 0
interval. 24. x = e t , y = e~ l , (-∞ . ∞)
3. x = t + 1, y = t - 1, 0 ≤ / ≤ 2
4. x = - t , y = 3 - 2t, 0 ≤ /≤ 1 F in d and -τ ^ in Problems 2 5 -3 0 without elim inating the
aX uX
5. x = 60/, y = 80/ - 16∕2 , 0 ≤ /≤ 3
parameter.
6. x = 30/, y = 60/ - 9 ∕ 2 , -1 ≤ / ≤ 2 25. x = ∕2 , y = ∕ 4 + 1 26. x = t2 + 1, y = P
7. x = /, y = 2 + ∣ ( ∕ - 1), 2 ≤ t ≤ 5 27. x = e f y = sin 2/ 28. x = e t , y = ∕2 e ,
8. x = /, y = 3 - j ( ∕ + 2), -1 ≤ / ≤ 3 29. x = a cos /, y = b sin t 30. x = a cosh /, y = b sinh /
9. x = t 2 + 1, y = ∕ 2 - 1, -1 ≤ / ≤ √2
10. x = 2t 2 , y = ∕ 2 + 2, - 1 ≤ / ≤ λ ∕2 F in d in Problems 3 1 -3 4 and then elim inate the parameter
11. x = ∕3 ,y = ∕2 , /≥ 0 to express your answer in terms o f x and y.
12. x = ∕4 , y = t2 , -1 ≤ / ≤ √2 31. x = ∕ 2 + 1, y = P + 1
13. x = 3 cos θ, y = 3 sin θ, 0 ≤ θ < 2τr 32. x = 2a co st, y = a sin 2 /; a > 0
14. x = 2 cos θ, y = 2 sin θ, 0 ≤ θ < 2π 33. x = 1 - e~ t , y = 1 + e t 34. x = ∕2 , y = In /
15. x = 1 + sin /, y = - 2 + cos /, 0 ≤ t < 2π F in d the area under the curves given in Problems 35-40.
16. x = 1 + sin 2 /, y = —2 + cos /, 0 ≤ / < 2ττ 35. x = ∕ 4 + 1, y = ∕2 , 0 ≤ /≤ 1
17. x = 4 tan 2/, y = 3 sec 2/, 0 ≤ t < π 36. x = ∕4 + t2 + 1, y = (/ + 2)2 , 0 ≤ /≤ 1
37. x = θ — sin θ, y = 1 + cos θ, 0 ≤ θ ≤ f

Problems 58 and 59 explore the work required to move a
38. x = tan θ, y = sec2 θ, 0 ≤ θ ≤ f mass or particle along a planar curve which is described
39. x = tan -1 u, y = u3 , 0 ≤ u ≤ 1 parametrically. Let C be the set o f points (x, y), where
40. χ = sin -1 ιι, y = u, 0 ≤ u ≤ ∣
x = x(t), y = y(f), a ≤ t ≤ b
Find the length o f each curve in Problems 41-45.
41. x = V t , y = ∕,3z4 for 0 ≤ t ≤ 4 Suppose a force F = F(t) is applied to the mass in the
42. x = I t 2, y = t3 for 1 ≤ t ≤ 2 direction o f the curve (that is, the force is applied in the
43. x = I ln(t2 - 1), y = V t 2 - 1 for 3 ≤ t ≤ 7 direction o f the tangent to C).
44. x = t sin t + cos t, y = sin t — t cos t for 0 ≤ t ≤ f 58. a. Argue that on the interval [r, t + dt}, dt very small,
the work done is approximately
45. x = 2 sin - 1 1, y = ln(l - t2) for 0 ≤ t ≤
© 46. Describe the path of curve defined by x = sin πt, dW = F(f) ds = F ( ffix '( f) ∖2 + [√(∕)]2 dt
y = cos2ττt for 0 ≤ t < 1.
47. Let x = 4α sin t, y = b cos2 1. Express y as a function of x. b. Form an appropriate Riemann sum, take a limit
and get the formula for the work done on C:
48. Find the area under one arch of the cycloid
x = 2(0 - sin θ), y = 2(1 - cos θ).
w = F(0√[x'(t)]2 + [y'(χ)]2 dt
49. Find the length o f the ja
curve given by x =
2 cos3 t, y = 2 sin 3 t, c. Suppose that F ,(t) = 2e' and x{t) = 6 cosh t,
0 ≤ t ≤ 27r. This curve y(f) = 6t for 0 ≤ t ≤ 1. Compute the work in
is called an asteroid. volved on the entire curve C. Note-. You may use a
computer, but a computer is not needed to answer
this question.
59. As we have seen, arc length intervals are often
impossible to do without a computer. The same is
true o f work integrals. For example, suppose a parti
cle is moving on an elliptical path described by
,, b x(f) = cos t, y(f) = 3 sin t,

Use the formula
Ja
^ [f(θ )] 2 + [ f '( θ ) ] 2 dθ
0 ≤ t < 2π
which is derived in Problem 65 to fin d the arc length o f the The force applied is F(f) = 5∣j∕(Z)∣, hence proportion
polar form curves given in Problems 50-53. al to the vertical component of velocity.
50. the spiral r = e β, for 0 ≤ θ ≤ J a. Compute the arc length of the ellipse to 3-
51. the circle r = 2 cos θ, for 0 ≤ 0 ≤ f decimal-place accuracy. This is a surprisingly
52. the cardioid r = 2(1 - cos 0) on [0, ττ] tough integral; so make a common-sense check to
make sure your answer is reasonable. Explain your
53. the spiral r = e~ θ for 0 ≥ 0
check.
b. Set up the integral and compute the work involved
in one orbit of the particle to 3-place accurary.
Computational Window c. Suppose a constant force F = 5 is applied to the
spiral path (gravity is working): For t ≥ 0,
54. Write a paper showing the graphs of several hypocy
cloids, along with an analysis of what you did to x(f) = 2e~'l l ° cos t,
graph them. Make some generalizations based on
your graphs. y(t) = 3e^'z ι ° sin t
55. Write a paper describing what happens if the radius
You are to compute the work (so that the relative
o f the inner rolling circle is larger than the radius of
error is less than 0.0005). But in order to obtain a
the outer fixed circle.
rough check on your answer and to determine how
56. Write a paper showing the graphs of several epicy far to integrate, you may wish to solve (without a
cloids, along with an analysis of what you did to computer) first a problem with an easier path
graph them. Make some generalizations based on — for example (for t ≥ 0)
your graphs.
57. Write a paper describing what happens if the radius x(f) - 2.5e - 'z10 cos t,
of the outer rolling circle is larger than the radius of
the inner fixed circle. y(f) = 2.5e - 'z ι ° sin t
60. SPY P ROBLEM Using a special meditation technique, the Show that this epicycloid has parametric equations
spy convinces his captors that their truth drug has killed
x = (α + Λ)cos∕ - R cos ( f7
him (Problem 58, Section 8.2). As soon as they start to
drag his body away, he revives and dispatches all but one
of his tormentors, and with a little persuasion, the survi
y = (a + A)sin t — R sin ^ R jt
vor tells him that Boldfinger has left the castle through an
underground water passage in his private submarine. The 63. A circle of radius R rolls without slipping on the inside of
spy races from the castle to a nearby naval base and a fixed circle of radius a. Find parametric equations for
commandeers a destroyer. Since Boldfinger suspects no the curve traced out by a point P on the circumference of
danger, his submarine is running on the surface, and the the rolling circle of radius R. Let t be the angle measured
spy’s destroyer is able to get within 3 miles of the from the positive x-axis to the ray that passes through the
submarine before it submerges. The spy assumes Bold center of the rolling circle, and assume that the point P
finger will try to escape by traveling at maximum speed in begins on the x-axis (that is, P has coordinates (a, 0) when
one direction, but, of course, he does not know which t = 0. Show that this hypocycloid has parametric equa
direction his adversary will take. If the destroyer’s best tions
speed is twice that of the submarine, what path should the
spy order his destroyer to travel in order to guarantee that x = (a - R)cost + R cos ( a ^ t,
it will eventually pass over the submarine? Hint: The spy
may need the formula for polar arc length given in
y = (α - A)sin t - R sin ( A R )1
Problem 65.
θ 61. Find parametric equations for the polar curve r = f ( 0) in
64. Prove Theorem 9.4.
terms of the parameter θ.
65. Arc Length in Polar Coordinates Suppose the polar func
62. A circle of radius R rolls without slipping on the outside
tion r = /(0) and its derivative∕'( 0 ) are continuous on the
of a fixed circle of radius a. Find parametric equations for
closed interval a ≤ θ ≤ b , { b ≤ a + 2π). Show that the
the curve traced out by a point P on the circumference of
arc length of the graph o f / from 0 = a to 0 = b is given by
the rolling circle of radius R. Assume the fixed circle is
centered at the origin and that the moving point begins at z,√ [∕(θ )] 2 + [∕'(0 )] 2 dθ
(a, 0). Use the angle I measured from the positive x-axis to
the ray from the origin to the center of the rolling circle.
9.5 CONIC SECTIONS: THE PARABOLA
IN THIS SECTION Standard-position parabolas, translation of parabolas,

representation of a parabola in polar coordinates, parabolic reflectors
A second topic in this chapter involves the so-called conic sections: ellipses,
parabolas, and hyperbolas. All of these curves can be obtained by intersecting a
plane with a right circular cone, as shown in Figure 9.33. In this section we look
at one of these conic sections, the parabola.
We first encountered parabolas in our study of quadratic functions in Chapter 1.

At that time, we found that the graph of an equation of the general form
y = ax λ + bx + c (with a ≠ 0) is a parabola that opens upward if a > 0 and
downward if a < 0. The graph of this quadratic equation is a parabola, but not all
parabolas can be represented by this equation, because not all parabolas are
graphs of functions. Consider the general second-degree equation
Aχ- + Bxy + Cy2 + Dx + Ey + F = 0
for any constants A, B, C, D, E, and F. If A = B = C = 0, the equation is not
quadratic but linear (first degree); but if at least one of A, B, or C is not zero, then
the equation is quadratic. Historically, second-degree equations in two variables
Sec. 9.5 Conic Sections: the Parabola 605
were first considered in a geometric context and were called conic sections because
Historical Note
the curves they represent can be described as the intersections of a double-napped
right circular cone and a plane.
HYPATIA (370-415)
Work with conic sections was
done by Apollonius
(2607-200? B.C.), but his work
was not popularized until 600 ■ STANDARD-POSITION PARABOLAS
years later, when Hypatia wrote
several mathematical commen We shall use the following definition for a parabola.
taries, including “On the Conic
Sections.” Hypatia was a
daughter of a professor at the
Parabola
University of Alexandria, where
she also became a professor of A parabola is the set of all points in the plane that are equidistant from a fixed
mathematics, philosophy, and point (called the focus) and a fixed line (called the directrix).
astronomy. See the suggested
book review Hypatia's Heritage
after the guest essay in Chapter The line through the focus perpendicular to the directrix is called the principal
11. axis of the parabola, and the point where the axis intersects the parabola is called
the vertex. The line segment that passes through the focus perpendicular to the
axis and with endpoints on the parabola is called the focal chord. This terminology
is shown in Figure 9.34.
To obtain the equation of a parabola, first consider a special case: a parabola

with focus F(0, c) and directrix y = -c , where c is any positive number. This
parabola must have its vertex at the origin (the vertex is halfway between the focus
and the directrix) and must open upward, as shown in Figure 9.35.
Let (x, j ,) be any point on the parabola. Then, from the definition of a
parabola,
DISTANCE FROM (x , j ) to (0, c) = DISTANCE FROM (x , y) TO DIRECTRIX
V (x - 0 )2 + ( τ - c)2 = y + c
χ 2 + y 2 —2cy + c 2 = y 2 + 2cy + c 2 Square both sides.
x 2 = 4cy
This is the equation of the parabola with vertex (0, 0) and directrix y = - c .
You can repeat this argument (see Problems 56-58) for parabolas that have
their vertex at the origin and open downward, to the left, and to the right to obtain
the results summarized next.
Standard-Form Equations Parabola Focus Directrix Vertex

for Parabolas
Upward, x 2 = 4cy (0, c) y = -c (0, 0)
Downward, x 2 = -4 c y (0, - c ) y = c (0, 0)
Right, y 2 = 4cx (c, 0) x = -c (0,0)
Left: y 2 = - 4 c x ( - c , 0) x -c (0, 0)
The length of the focal chord is the coefficient 4c.
In Chapters 1 and 3 we used calculus to graph parabolas. In this chapter we will

find and plot the vertex, determine c (usually by inspection), and count out c units
from the vertex in the appropriate direction as determined by the form of the
equation. Finally, it is shown in the problem set that the length of the focal chord
is 4c, and we use this number to determine the width of the parabola, as shown in
the following example.
EXAM PLE 1 Graphing a standard-form parabola
Graph 2y 2 - 5x = 0.
Solution First, algebraically change the equation so that it is in standard form by
solving for the second-degree term:
y2 = ⅛
The vertex is (0, 0) and
4c = j so c= ∣
Thus, the parabola opens to the right, the focus is (∣, θ), and the length of the
focal chord is 4c = j , as shown in Figure 9.36. t= ∣
There are two basic types of problems in analytic geometry:
1. Given the equation, draw the graph; this is what we did in Example 1.
2. Given the graph (or information about the graph), write the equation. The
next example is of this type.
EXAM PLE 2 Writing the equation of a parabola
Find an equation of a parabola with focus F(0, -2 ) and directrix y = 2.

Solution This is the curve drawn in Figure

9.37. We see this is a parabola that opens
downward with vertex at the origin. By
inspection, the value of c is c = 2. The
form of the equation is x 2 = -4 c y , so the
desired equation is
x 2 = -8 y
Figure 9.37 Graph of the parabola

with focus (0, -2 ) and directrix
y =2
■ TRANSLATION OF PARABOLAS
If a parabola is not in standard position but its axis is parallel to one of the
coordinate axes, it can be put into standard form by a change of variable of the
form X = x - h , Y = y - k. Such a change in variable is called a translation, and it
has the general effect shown in Figure 9.38. Suppose (h, k) = (4, 5); then the
coordinates of P are (x, y) = (6, 9) and (X, Y) = (2, 4); check
X = x - A= 6- 4= 2
Figure 9.38 Translation of a
parabola; vertex is at (Λ, k). Y = y -k = 9 -5 = 4
Effect of a Translation Replacing x - Λ by X and y - k by Y in an equation has the effect of translating

the graph of an equation
h units horizontally (right if h > 0 and left if h < 0);
k units vertically (up if k > 0 and down if k < 0).
This means that the equations for parabolas with vertex (Λ, ∕<) are
(x - Λ)2 = 4c(y - k), X 2 = 4cY; (y - k)2 = 4c{x - h), Y 2 = 4cX
(x - Λ)2 = -4 c (y - k), X 2 = -4 c Y ; (y - k)2 = -4 c (x - h), Y 2 = -4 c X
EXAM PLE 3 Graphing a parabola using a translation
Sketch the parabola y = x 2 + 2x + 3. Find the vertex, c, and the length of the
focal chord. Also, find the focus and the equation of the directrix.
Solution First, complete the square to obtain y - 2 = (x + 1)2 . Next, plot the
vertex ( - 1 , 2). If we replace y - 2 by Y and x + 1 by X , we have
Y = X 2, which tells us the parabola opens upward, 4c = 1, and c = ∣. Thus,
plot the focus by counting up ∣ unit, and then draw the focal chord with length
1. Because these points are fairly close on the scale we have chosen, we plot an
additional point, say the y-intercept: If x = 0, then y = 02 + 2(0) + 3 = 3. See
Figure 9.39.
Figure 9.39 Graph of Notice that we do not need to know the coordinates of the focus or the
y = x 2 + 2x + 3 directrix to draw the graph. We needed to know only the vertex, the distance c,
and the length of the focal chord 4c. If c is small relative to the vertex, it may
be necessary to plot an additional point.
However, we may need to know the coordinates of the focus and directrix
for further analysis. To find these numbers, we can use the reverse translation
x = X+h, y=Y+k.
AT-coordinates xy-coordinates
Vertex (0, 0) ( -1 , 2)
Focus (M ( -ι> l)
Directrix γ = -- 4 T= 5
EXAM PLE 4 Finding the equation of a translated parabola
Find an equation for the parabola with focus (4, -3 ) and directrix the line
x + 2 = 0.
Solution Sketch the information as shown in Figure 9.40. The vertex is (I, -3 ),
because it must be equidistant from Fand the directrix. Note that c = 3. Thus,
substitute into the equation of a parabola that opens to the right— namely
y 2 = 4cx— and then translate to the point (∕z, k) to obtain the equation
(y - k)2 = 4c(x - h)
Figure 9.40 Graph of a parabola The desired equation, because (A, k) = (1, -3 ), is
with focus (4, -3 ) and directrix
x +2=0 ( y + 3 )-= 1 2 (x - 1)
■ REPRESENTATION OF A PARABOLA IN POLAR

COORDINATES
Next, we shall see how a parabola can be represented in polar coordinates. Let a
parabola be given in the plane. Then place the x-axis along the principal axis of the
parabola, and place the pole at the focus, as shown in Figure 9.41. We refer to this
as a standard polar position for the parabola. Assume the parabola opens to the
right and that the directrix L is the vertical line x = - p , where p > 0 is the
distance from the focus to the directrix.
If P is a point on the parabola with rectangular coordinates (x, y) and polar
coordinates (r, 0), we must have
DISTANCE P TO F = DISTANCE P TO L
∣r∣ = ∖p + r cos 0∣
Figure 9.41 A standard polar r = ± (p + r cos θ)
position for a parabola
P _ . = P
1 1 - cos θ , 1 + cos θ
It can be shown (see Problem 60) that these two equations represent the same
graph, so we shall use the one on the left to represent the given parabola. We can
similarly derive equations for parabolas that open downward, left, or to the right.
These graphs are shown in Figure 9.42.
Standard Polar Equations for Directrix Polar-form

Parabolas Parabola Focus (rectangular-form*) Vertex
Upward-. r = P (0, 0) y = -p
(P 3-τr∖
1 - sin θ ∖2 2 )
= - ..Λ - - ∕P 7τ∖
Downward'. r (0,0) v= P k2 2/
1 + sin θ
Right: r = - P ____
1 - cos θ
(0, 0) x = -p II
Left: r = - ^ - ____ (0, 0) X =p ( f ’ θ)
1 + cos θ
EXAM PLE 5 Graphing a polar-form parabola
Describe and sketch the graph of the equation
3 - 3 cos θ
Solution .
4
1 43
r = . 3=
3 - 3 cos θ 1 1 - cos θ
3
By inspection, you can now see (Figure 9.43) that the parabola opens to the
right and that p = ∣. Thus the vertex is
Plot the vertex and the line x = - ⅜>as shown in the margin. You can plot other
Figure 9.43 Graph of points that are easy to calculate, such as the points where θ = j and θ = π or
1 = 4 θ = 2f . —
3 - 3 cos θ
EXAM PLE 6 Finding a polar-form equation of a parabola
Find a polar-form equation for the parabola with focus at the origin and
vertex at the polar-form point (3, -ττ).
*We could, o f course, state these equations of lines in polar form. For example, y = - p is
r sin 0 = - p , but we prefer writing the equation o f the directrix in rectangular form.
Solution The vertex (3, ττ) is on the x-axis and to the left of the focus (the pole).
Thus, the parabola opens to the right and has a polar-form equation
1 - cos θ
where p is the distance from the focus to the directrix. Because the vertex
(3, ττ) is halfway between the focus and the directrix, we must have p = 6, so
that the required equation is
r = . - 6- ■■
1 - cos θ ∙
■ PARABOLIC REFLECTORS*
Parabolic curves are used in the design of lighting systems, telescopes, and radar
antennas, mainly because of the property illustrated in Figure 9.44 and described
more formally in Theorem 9.6.
THEOREM 9.6 Reflection property of parabolas
Let P be a point on a parabola in the plane, and let T be the tangent line to the
parabola at P. Then the angle between T and the line through P parallel to the
principal axis of the parabola equals the angle between T and the line connecting P
to the focus.
Proof: A proof is outlined in Problem 65. ■
Figure 9.44 The reflection As an illustration of how this property is used, let us examine its application to
property of parabolas reflecting telescopes. The eyepiece of such a telescope is placed at the focus of a
parabolic mirror. Light enters the telescope in rays that are parallel to the axis of
the paraboloid. It is a principle of physics that when light is reflected, the angle of
incidence equals the angle of reflection. Hence, the parallel rays of light strike the
parabolic mirror so that they all reflect through the focus, which means that the
parallel rays are concentrated at the eyepiece located at the focus.
Flashlights and automobile headlights simply reverse the process: A light
source is placed at the focus of a parabolic mirror, the light rays strike the mirror
with angle of incidence equal to the angle of reflection, and each ray is reflected
along a path parallel to the axis, thus emitting a light beam of parallel rays.
Radar utilizes both of these properties. First, a pulse is transmitted from the
focus to a parabolic surface. As with a reflecting telescope, parallel pulses are
transmitted in this way. The reflected pulses then strike the parabolic surface and
are sent back to be received at the focus.
A reflecting telescope: A parabolic flashlight:

Light rays parallel to the A light source at the focus
axis are concentrated at sends out beams of light
the focus. parallel to the axis.
*Since these reflectors are three-dimensional, the precise word is paraboloidal, but the most
common usage is to look at a cross section that is parabolic.
Sec. 9.5 Conic Sections: the Parabola 61 1
PROBLEM SET 9 .5
1. ■ What Does This Say? Discuss a procedure for graph 50. Find the equation of the set of all points with distances
ing a parabola in rectangular form. from (4, 3) that equal their distances from (0, 3).
2. ■ What Does This Say? Discuss a procedure for graph 51. Find the equation of the set of all points with distances
ing a parabola in polar form. from (4, 3) that equal their distances from (—2, 1).
Sketch the curves in Problems 3-24. Label the focus F and the 52. Find an equation for a parabola whose focal chord has
vertex K length 6, if it is known that the parabola has focus (4, -2 )
3. y 2 = 8x 4. y 2 = -1 2 x and its directrix is parallel to the y-axis.
2
5. y = -2 0 x 6. 4x 2 = 1Oy 53. A parabolic archway has the dimensions shown in Figure
7. 3x 2 = - 1 2 j 8. 2x 2 = - 4 y 9.45. Find the equation of the parabolic portion.
2
9. 2x + 5y = 0 10. 5y2 + 15x = 0
2
11. 3y - 15x = 0 12. 4y2 + 3x = 12
2
13. 5x + 4y = 20 14. 4x 2 + 3y = 12
15. (y - 1)2 = 2(x + 2) 16. (y + 3)2 = 3 ( χ - 1)
2
17. (x + 2) = 2(y - 1) 18. ( χ - l) 2 = 3(y + 3)
2
19. y + 4x - 3y + 1 = 0 20. y 2 - 4x + 10.y + 13 = 0
2
21. y + 4y - 10x + 74 = 0 22. x 2 + 9y - 6x + 18 = 0
23. 9x 2 + 6x + 18j-, —23 = 0 24. 9x 2 + 6y + 1 8 x - 23 = 0
Find an equation for each curve in Problems 25-32.
25. Directrix x = 0; focus at (5, 0)
26. Directrix y = 0; focus at (0, —3) Figure 9.45 A parabolic
archway
27. Directrix x - 3 = 0; vertex at (-1 , 2)
28. Directrix y + 4 = 0; vertex at (4, -1 ) 54. Beams of light parallel to the axis of the parabolic mirror
shown in Figure 9.46 strike the mirror and are reflected.
29. Vertex (- 2 , -3 ); focus at (-2 , 3)
Find the distance from the vertex to the point where the
30. Vertex (—3, 4); focus at (1, 4) beams concentrate.
31. Vertex (—3, 2) and passing through (—2, -1 ); axis parallel
to the y-axis
32. Vertex (4, 2) and passing through (-3 , -4 ); axis parallel
to the x-axis
Sketch the graph o f the polar-form parabola in Problems
33-38. Check your work by finding a Cartesian equation. Figure 9.46 A parabolic
33. r = — - 34. r = 7— ^ - a 35. r = 1 ~
9
o
mirror
1 + cos θ 1 - sm θ 1 + sm θ
55. A radar antenna is constructed so that a cross section
36. r = . ~ 2 , 37. r = ~— γ ----- z 38. r = ------ along its axis is a parabola with the receiver at the focus.
1 - cos θ 2 - 2 cos θ 3 + 3 cos θ Find the focus if the antenna is 12 m across and its depth
Find a polar equation for a parabola with its focus at the pole is 4 m. See Figure 9.47.
and with the property given in Problems 39-42.
39. vertex at the polar-form point (4, 0)
40. vertex at the polar-form point (2, τr)
41. directrix at y = - 4 42. directrix at x = 3
Find a polar equation for the parabola with the Cartesian
equation given in Problems 43-46.
43. y 2 = 4x 44. x 2 = - 2 y
2
45. 4x = y - 3 46. x + 1 = 2(y - 3)2
47. Find the point(s) on the parabola y 2 = 9x that is (are)
closest to (2, 0).
48. Find the point(s) on the parabola x 2 = 4cy that is (are)
closest to the focus.
49. Find the equation for the tangent line and the line
perpendicular to the parabola y 2 = 4x at the point ( 1, - 2). Figure 9.47 Dimensions for a radar antenna
θ 56. Derive the equation of a parabola with F(0. - c ) , where

c is a positive number and the directrix is the line y = c.
57. Derive the equation of a parabola with F(c, 0), where c is a
positive number and the directrix is the line x = —c. Use Figure 9.48 to prove the reflection property of the
58. Derive the equation of a parabola with F ( - c , 0), where c is parabola by carrying out the indicated steps.
a positive number and the directrix is the line x = c.
59. Show that the length of the focal chord for the parabola
y 2 = 4cx is 4c (c > 0).
60. Show that the polar equations
P ~P
r = t----------λ and r = — ------- -
1 — cos θ 1 + cos θ
represent the same graph.
61. Show that the vertex is the point on a parabola that is
closest to the focus.
Figure 9.48 Reflection property
62. Show that the tangent to a parabola at the two ends of the
focal chord intersect on the directrix.
63. Find the area of the triangle formed by the focal chord a. Find an equation for the tangent line T to the parabola
a t p (⅞, L )∙
and the two tangent lines at the end of the focal chord of o
the parabola x 2 = 4cy. b. Find the coordinates of the point Q where T crosses
64. Suppose a circle intersects the parabola x 2 = 4cy in four the y-axis.
distinct points (x 1, y 1), (x2, y 2), (x 3, y 3), and (x4 , y 4 ). Show c. If F is the focus of the parabola, show that ∣FF∣ = ∣F(2∣-
that x 1 + x 2 + x 3 + x 4 = 0. Conclude that ∕∖Q F P is isosceles.
65. Reflection property of the parabola Assume that a para d. Let L be a line parallel to the y-axis. Show that the
bola is given by the equation angle between L and T is θ = Z-FPQ.
9.6 CONIC SECTIONS: THE ELLIPSE AND THE HYPERBOLA
IN THIS SECTION Ellipses, hyperbolas, eccentricity and polar coordinates,

geometric properties
In the last section we introduced the geometric definition of the parabola, and
in this section, we shall examine two more conic sections, the ellipse and the
hyperbola.
■ ELLIPSES
Ellipse An ellipse is the set of all points in the plane the sum of whose distances from
two fixed points is a constant.
The fixed points are called the foci (plural of focus). To see what an ellipse
looks like, we will use the special type of graph paper shown in Figure 9.49a, where
F 1 and F 2 are the foci.
Let the constant distance be 12. Plot all the points in the plane so that the sum
of their distances from the foci is 12. If a point is 8 units from Fj, for example,
then it is 4 units from F ,, and you can plot the points F 1 and P,. The completed
graph of this ellipse is shown in Figure 9.49b.
The line passing through F l and F , is called the major axis. The center is the
m idpoint of the segment F i F 1 . The semimajor axis is the distance from the center
to a point of intersection of the ellipse with its major axis. The line passing
through the center perpendicular to the major axis is called the minor axis. The
semiminor axis is the distance from the center to a point of intersection of the
Sec. 9.6 Conic Sections: The Ellipse and the Hyperbola 613
a. Definition of ellipse b. An ellipse: IFF,I + l∕ψ ,l = k,

where k = 12 in this figure
Figure 9.49
minor axis with the ellipse. The ellipse is symmetric with respect to the major and
minor axes. The intercepts on the major axis are called the vertices of the ellipse.
To find the equation of an ellipse, first consider a special case where the center
is at the origin. Let the distance from the center to a focus be the positive number
c; that is, let F 1( - c , 0) and F,(c, 0) be the foci and let the constant sum of distances
be 2α, as shown in Figure 9.50.
If P(x, y) is any point on the ellipse, then by definition,
R I + R ∣ = la
V (x + c)- + (y - 0)2 + V ( Λ' - c)2 + (y - 0)2 = la
Simplifying (the details are not shown), we obtain
Figure 9.50 Developing the equa 2
tion of an ellipse by using the

Let b 1 = a- —c 1 to obtain + = L The graph of this equation is shown in
definition Figure 9.51. Notice that foci are on the major axis, and the intercepts on the minor
axis are (0, ⅛) and (0, -b ) . Also note that a > c and a > b.
A similar derivation applies to the ellipse in standard position with foci on the
j>-axis.
Standard-Form Equations Orientation Equation Foci Constant Center

for Ellipses
Horizontal y2 = 1
x~ + 75 (-c, 0), (c, 0) la (0, 0)
a1 1 b
υ2 r2
Vertical -51 + = 1 (0, c), (0, -c) 2a (0, 0)
a b~
where b2 = d 1 - c2 or c2 = a 2 - b2 with a > b >■0
In order to sketch an ellipse, plot the center, the intercepts ± a on the major axis,
and ± b on the minor axis.
HI What This Says: Write the equation in standard form, so that there is a 1
on the right and the numerator coefficients of the square terms are also 1. The
center is (0, 0); plot the intercepts on the x- and p-axes. For the x-intcrcepts,
plot ± the square root of the number under the x 2-termj for the ^-intercepts,
plot ± the square root of the number under the p 2-term. Finally, draw the
ellipse using these intercepts. The longer axis is called the major axis; if this
larger axis is horizontal, then the ellipse called horizontal, and if the major axis
is vertical, the ellipse is called vertical.
EXAM PLE 1 Graphing an ellipse centered at the origin
Sketch 9x2 + 4y2 = 36. Find the foci.

Solution First, rewrite the equation in standard form by dividing both sides by
36:
4 9 1
Because a~ = 9, b1
= 4, the foci are found by calculating c as follows:
Figure 9.52 Graph of c2 = α2 - ⅛2 = 9 - 4 = 5. Thus, the foci are (0, -V 5), (0, V5). The graph is
9x 2 + 4y 2 = 36 shown in Figure 9.52.
Sec. 9.6 Conic Sections: The Ellipse and the Hyperbola 61 5
As with the parabola, it is easier to use your calculator to graph an ellipse by

parameterizing the equation. This is illustrated in the following example.
EXAM PLE 2 Parameterize the equation of an ellipse
Graph 2x 2 + 5p2 = 10 by using a parameterization.

Solution One way to parameterize an ellipse is to recall the identity
cos2 θ + sin2 0 = 1 .
We begin by dividing both sides of the given equation by 10:
2Λ-2 + 5y2 = 10
γ V
We know that if cos θ = y = and sin θ = ~∕=> then cos2 θ + sin 2 θ = 1. There
are, of course, other choices we could make, but this observation leads us to let
x = V5 cos θ and y = ∖ r2 sin θ. You can set up a table of values or use a
calculator to obtain the graph shown in Figure 9.53. If you use a table of
values, you need only consider values of θ between 0 and because we know
the ellipse is symmetric with respect to both the major and minor axes.
0 X y
0°; 0 2.24 0
150∙ — 2.16 0.37
’ 12
30°; f 1.94 0.71
4 0
4 05 ∙, —
4 1.58 1
60°; 1.12 1.22
7 <°∙ 5IT 0.58 1.37
’ 12
90°; ∣ 0 1.41
Figure 9.53 Graph of x = V5 cos θ, y = V2 sin θ
The parameterization we obtained for the ellipse in Example 2 is not unique.

For example, x = V z5 sin θ ,y = V2 cos θ is the same ellipse, but if you sketch these
parametric equations, you will note that even though the ellipse is the same, the
orientation has reversed.
EXAM PLE 3 Finding the equation of a given ellipse
Find an equation for the ellipse with foci ( -1 , 0) and (1, 0) and vertices
( - 2 , 0) and (2, 0).
Solution By inspection, the center of the ellipse is (0, 0) and the distance to a
vertex is 2, so a = 2; the distance to a focus is 1, so c = 1. Therefore, we find
b 1 = d 1 - c λ = 3. An equation is
A→ -∑∑= 1
4 3 ≡≡
If an ellipse is not in standard position, but its axes are parallel to the
coordinate axes, complete the square to determine the translation. Here is an
example of this procedure.
EXAM PLE 4 Graphing an ellipse by completing the square
Sketch the graph of the equation 9x2 + 4y2 - 18x + 16y - 11 = 0.

Solution Complete the square in both x and y:
9x2 + 4y2 - 18x + 16y - 11 = 0
9(x2 - 2x ) + 4(y2 + 4y ) = 11
9(x2 - 2x + 12 ) + 4(y2 + 4y + 2 2) = 11 + 9 ∙ 12 + 4 ■2 2
9(x - 1)2 + 4(y + 2)2 = 36
(x - 1)2 (y + 2)2 _
4 + 9 1
Thus, the graph may be obtained by translating the graph of the ellipse -⅛ x 2- +
Figure 9.54 Graph of y2
= 1 by 1 unit to the right and 2 units down. This process is shown in Figure
9x2 + 4y2 - 18x + 1 6 y - 11 = 0
9.54. First, plot the center (h, k) = (1, -2); then, from that point, count out a
distance of a = ÷ 3 (the vertices) on the major axis and label those vertices.
Finally, from the center, count out the distance b = ± 2 on the minor axis.
Using those four points on the ellipse, you can sketch the graph.
■ HYPERBOLAS
The last of the conic sections to be considered has a definition similar to that of
the ellipse.
Hyperbola A hyperbola is the set of all points in the plane such that, for each point on the
hyperbola, the difference of its distances from two fixed points is a constant.
^ ∙L , , *r ⅛ I " « S ® M M S Cc, ⅛*"+ ι- U <⅛"f^'C.2Γ'J S2'.∖ , 4S%⅛V*S 'Λ .' M l Λ C'∙Γ' '+
The fixed points are called the foci. A hyperbola with foci at F { and F 2, where
the given constant distance is 8 is shown in Figure 9.55.
The line passing through the foci is called the transverse axis. The center is the
midpoint of the segment connecting the foci. The line passing through the center
perpendicular to the transverse axis is called the conjugate axis. The transverse
axis intersects the hyperbola at points called the vertices, and the conjugate axis
does not intersect the hyperbola. The hyperbola is symmetric with respect to both
the transverse and conjugate axes.
If you use the definition, you can derive the equation for a hyperbola with foci
at (-c, 0) and (c, 0) and constant difference 2a (both c and a are positive). If (x, y)
is any point on the curve, then
IV(x + c)2 + (y - 0)2 - V(x - c)2 + (y - 0)2 = 2a
Figure 9.55 Graph of a hyperbola The procedure for simplifying this expression is left as a problem. After several
from the definition steps you should obtain
>’2
a2 c2 -a 2
If b2 = c2 - a 2, then
= 1
a2 b2
Sec. 9.6 Conic Sections: The Ellipse and the Hyperbola 61 7
which is the standard-form equation. Notice that c2 = a 2 - b 2 for the ellipse and
that c 2 = a 2 + b 2 for the hyperbola. For the ellipse it is necessary that a 2 > b2 , but
for the hyperbola there is no restriction on the relative sizes for a and b (but c is
still less than a for the hyperbola).
Repeat the argument for a hyperbola with foci F 1(0, c) and ∕ ,'2(0. - c ) , and you
will obtain the other standard-form equation for a hyperbola with a vertical
transverse axis.
Standard-Form Equations Orientation Foci Constant Center

for Hyperbolas
2 y2
Horizontal'. ¾i - 751 = 1 ( - G 0), (c, 0) la (0, 0)
a b
V2 r2
Vertical: ⅛1 = 1 (0, c), (0, -c ) 2a (0,0)
ai b
where b 2 = c 2 —a 2 or c2 = a 2 + b 2
As with the other conics, we shall sketch a hyperbola by determining some

information about the curve directly from the equation by inspection. The
vertices are located ± a units from the center. The number la is the length of the
transverse axis. The hyperbola does not intersect the conjugate axis, but if you plot
the points located ÷ b units from the center, you determine a segment on the
conjugate axis with length 26 called the length of the conjugate axis. The endpoints
of this segment are useful in determining the shape of the hyperbola.
EXAM PLE 5 Sketching a hyperbola in standard form
Sketch =¾- - 7j- = 1.
Solution The center of the hyperbola is (0, 0); a = 2 and b = 3. Plot the vertices
at x = ± 2, as shown in the margin. From the form we see that the transverse
axis is along the x-axis and the conjugate axis is along the y-axis. Plot the
length of the conjugate axis by plotting ± 3 units from the origin. We call these
points the pseudovertices, because the curve does not actually pass through
these points.
Next, form a rectangle by drawing lines through the vertices and pseudo
vertices parallel to the axes of the hyperbola. This rectangle is called the
central rectangle. The diagonal lines passing through the corners of the central
rectangle are oblique asymptotes for the hyperbola, as shown in Figure 9.56;
they aid in sketching the hyperbola.
For the general hyperbola given by the equation

x ^ _/ = ι
a 2 b2
the equations of the oblique asymptotes described are found by replacing the
constant term 1 by 0 and then factoring and solving:
y = ∣ * and y = - ∣x
Figure 9.56 Graph of To justify this result, you are asked in Problem 66 to show that the branches of the
x 2 _ >,2 ι
4 9 hyperbola approach y = ± x as ∣x∣ → ∞.
EXAM PLE 6 Completing the square to sketch a hyperbola
Sketch I6x 2 - 9y 2 - 128JV - 18y + 103 = 0.

0 Watch the signs on the sec
Solution Complete the square in both x and y.
ond parentheses. 0
16x 2 - 9 ∕ - 1 2 8 x - 18y + 103 = 0
16(x2 - 8 x ) - 9(y 2 + 2y ) = -1 0 3
16(x2 - 8 x + 4 2 ) - 9 ( y 2 + 2y + 12 ) = -1 0 3 + 16 ∙ 4 2 - 9 ■12
1 6 ( x - 4 ) 2 - 9 ( p + 1)2 = 144
(Λ- - 4 ) 2 ( y + l ) 2 .
9 16 1
The graph is shown in Figure 9.57. M M
We conclude our discussion of hyperbolas by considering an example in which

information about the graph is given and we are asked to find the equation of the
Figure 9.57 Sketch of 16x2 - hyperbola.
9y2 - 128Λ∙ - 18y + 103 = 0
EXAM PLE 7 Equation of a hyperbola given information about the graph
Find the set of points such that, for any point, the difference of its distances
from (6, 2) and (6, - 5 ) is always 3.
Solution From the definition, we see this is a hyperbola with center (6, - j) and
c = 5. Also, 2« = 3, so a = ∣. Because c 2 = a 2 + b 2 , we have
-φ = + ⅛2 so that b 2 = 10
The desired equation is
(T ÷ ¾ 2 (x -6 )2 (-y - 6)2
9 10 io
■ ECCENTRICITY AND POLAR COORDINATES
We defined the parabola as the set of all points P equidistant from a given point F
(the focus) and a given line L (the directrix). In other words, for a parabola
DISTANCE FROM P TO F = ,
DISTANCE FROM P TO L 1
This form of the definition of a parabola is part of the following characterization

of conic sections.
Eccentricity Let F be a point in the plane and let L be a line in the same
plane. Then the set of all points P in the plane that satisfy
DISTANCE FROM F TO F =
DISTANCE FROM P TO L ε
is a conic section, and ε is a fixed number for each conic called the eccentricity
of the conic. The conic is
An ellipse if ε < 1;
A parabola if ε = 1;
A hyperbola if ε > 1.
These criteria for a conic are illustrated in Figure 9.58.
Figure 9.58 Eccentricity characterization of a conic section
Next, we shall examine polar characterizations for the ellipse and the
hyperbola that involve the eccentricity ε. Consider an ellipse with one focus F a t
the origin of a polar coordinate plane. Assume that the corresponding directrix L
is the vertical line x = p (p > 0; or in polar form, r cos 0 = p) and that the ellipse
has eccentricity ε. Then, if P(r, 0) is a polar-form point on the ellipse, we have
_ DISTANCE FROM P TO F _ r
E DISTANCE FROM P TO L p - Γ COS 0
This relationship is shown in Figure 9.59.
Figure 9.59 Polar representation of an ellipse
Solving for r, we find the ellipse has the polar equation

εP
Because 0 ≤ ε < 1 we have ε cos θ ≠ -1 .
1 + ε cos 0
εp a n c ' if the
Similarly, if the directrix is x = - p , the equation is r = γi -z —
ε cos (/
directrix is y = p or y = - p , the corresponding equations are, respectively,
Figure 9.60 Forms for the equation of an ellipse (ε < 1) in standard polar form
EXAM PLE 8 Describing the graph of an equation in polar form
Discuss the graph of the polar-form equation r = ~ _

C O S tf ∕l
Solution Begin by writing the equation in standard form: r = ------:--------.

I - 2 cos 0
This form involves a cosine and has ε = ∣ < 1, so by comparing with the
forms in Figure 9.60, we see that the graph must be a horizontal ellipse. The
form also tells us that εp = 1, so p = 2 and the directrix is x = —2. The focus
F l closer to the directrix is at the pole, and the vertices occur where θ = 0 and
θ = 7τ. For θ = 0, we obtain r = 2, and for θ = π, r = y so the vertices are the
polar points K1(∣> τr) and F 2(2, 0). Since the focus F 1(0, 0) is j unit to the left
of F 1(2, 0), F 2 is the polar point ( j> 0). The center of the ellipse is midway
between the foci, at (j> θ), so the minor axis is the vertical line passing through
this point. This is the line x = j or, in polar form,
2
r cos θ = ⅜ or r = —^-5
3 cos θ
To find the endpoints of the minor axis, we need to solve simultaneously the
equations for the axis with the equation of the ellipse; namely,
2
2 3
2 - cos θ 2 - cos 0 cos θ
2 cos θ = j(2 - cos θ)
I COS θ = y
cos 0 = 2
so that θ = f and ⅛r. Solving for r, we find r = 5, which gives the vertices (∣, f)
and (j, ⅛p). The graph is shown in Figure 9.61. , , -.,, ι
Formulas for hyperbolas in polar coordinates are obtained in essentially the

same way as polar formulas for ellipses. The four different cases that can occur for
hyperbolas in standard form are summarized in Figure 9.62.
Figure 9.62 Forms for the equation o f a hyperbola (ε > 1) in standard polar form
EXAM PLE 9 Describing the graph of a hyperbola in polar form
Discuss the graph of the polar equation r = 3 + 4 s jn Q∙
5
α
Solution The standard form of the equation is r = -------------- . The form tells us
1 + j sin 0
that the eccentricity is ε = j> and because ε > 1, the graph is a hyperbola. We
also see that the transverse axis is the j ,-axis and that because εp = ∣ we have
Figure 9.63 Graph of r = 3 + 4 s jn θ p = ∣. Thus, the graph has one focus F ∣at the pole and directrix y = p = ∣
.
rr. , . τ,. . The correspondingb vertex occurs when θ = ?:

z
Historical Note
3 + 4 sin f 7
so the polar coordinates of this vertex Jz 1 are (7’ 7). The opposite vertex occurs
where 0 = ⅛y and
5
r ~ 3 + 4 s in ⅜ ^ ~ 5
so the point is F, ( - 5 , ⅛f). Because the vertex K1 is located 7 unit above

F 1(0, 0), the other focus F , will be located 5 units above the vertex K1. Thus, F 2
is the polar point (⅛p> f). The graph is shown in Figure 9.63.
BENJAMIN BANNEKER
(1731-1806) ■ GEOMETRIC PROPERTIES
Benjamin Banneker, born in
Maryland, was the first Ameri Like the parabola, the ellipse has some useful reflection properties. Let P be any
can black to be recognized as a point on an ellipse with foci F 1 and F 2 , and let T be the tangent line to the ellipse at
mathematician. P, as shown in Figure 9.64.
“He exhibited unusual tal
ent in mathematics, and with
little formal education, pro
duced an almanac and invented
a clock. Banneker was born free
in a troubled time of black slav
ery. The right to vote was his
by birth. It is important to note
that this right to vote was de
nied him as well as other free
black Americans by 1803 in his
native state of Maryland.”*
*From Virginia K. Newell, Joella
H. Gipson, L. Waldo Rich, and Beaure
Figure 9.64 Reflection property of an ellipse
gard Stubblefield, Black Mathemati
cians and Their Works (Admore:
Dorrance), p. xiv. Then the line segments F l P and F 2P (called the focal radii) make equal angles with
the tangent line T at P (see Problem 69). The reflection property of an ellipse has
the following physical interpretation.
Reflective Property An elliptic m irror has the property that waves emanating from one focus are
of an Ellipse reflected toward the other focus.
The “whispering room ” phenomenon found in many science museums and in

famous buildings such as the old U.S. Capitol in Washington, D.C., is an
application of this principle. Two people stand at each focus of an elliptic dome. If
one person whispers, the other will clearly hear what is said, but anyone not near a
focus will hear nothing. This is especially impressive if the foci are far apart.
The elliptic reflection principle is also used in a procedure for disintegrating
kidney stones. A patient is placed in a tub of water with the shape of an ellipsoid (a
three-dimensional elliptic figure) in such a way that the kidney stone is at one
Whispering room: Only a person Pulse emanating from focus Tj is

standing at focus F1 clearly hears a concentrated on the kidney stone
sound at focus Fj. at focus F,.
focus of the ellipsoid. A pulse generated at the other focus is then concentrated at
the kidney stone.
Figure 9.65 LORAN measures the There is also a useful reflection property of hyperbolas. Suppose an aircraft
differences in the time of arrival has crashed somewhere in the desert. A device in the wreckage emits a beep at
of signals from two sets of sta regular intervals. Two observers, located at listening posts a known distance apart,
tions. The plane’s position at the time a beep. It turns out that the time difference between the two listening posts
intersection lines is charted on a multiplied by the velocity of sound gives the value 2a for a hyperbola on which the
special map based on a hyperbolic
airplane is located. A third listening post will determine two more hyperbolas in a
coordinate system.
similar fashion, and the airplane must be at the intersection of these hyperbolas.
(See Problem 61.)
PROBLEM SET 9 .6
© 1. ■ What Does This Say? Discuss a procedure for graph- 20. x 2 - 4y 2 + 2x + 8y - 7 = 0
ing an ellipse. 21. 4x 2 + y 2 + 8x - 2y + 4 = 0
2. ■ What Does This Say? Discuss a procedure for graph- 22. x 2 + 4y 2 + 2x - 8j' + 4 = 0
ing a hyperbola. 23. 9x 2 + 4y 2 - 8y = 32
3. ■ What Does This Say? Outline a procedure for recog 24. 4x 2 + 9y2 - 8x = 32
nizing a conic section by looking at its rectangular-form
equation. Find the standard-form equation for the given information in
4. ■ What Does This Say? Outline a procedure for recog Problems 25-38.
nizing a conic section by looking at its polar-form equa- 25. an ellipse with vertices at (0, 8) and (0, 2) and c = V5
tion. 26. an ellipse centered at the origin with focus at (0, 3); semi
major axis with length 4
For each equation in Problems 5-24, sketch each curve. 27. an ellipse centered at the origin with focus at (—2, 0);
minor axis with length 4
28. a hyperbola with foci at (0, 3) and (0, -3 ) and one vertex
4 16 1 16 4
¾r at (0, -2 )
xf - ^ = 1 ∞ lσ I I
7. ' 29. a hyperbola with foci at (Vz2, 0) and (—V2, 0) and one
25 9
vertex at (1, 0)
9. y2 - χ 2 = 1 10. x2 - y 2 = 8
30. a hyperbola with vertices at (5, 0) and (-5 , 0) and one
11. 2 2
(x - 1) + 4y> = 64 12. 2x 2 + 3y2 = 12 focus at (—7, 0)
13. x 2 - 2y 2 + 2 = 0 14. 5x 2 - 3y2 + 15 = 0 31. a conic with major axis - 4 ≤ x ≤ 4 and minor axis
(A- ~ 1)2 , (J' + 3)2 . 11 6, (A - i) 2 1, (τ + 3)2 . -3 ≤ y ≤ 3
15. ∙ 16 4 -
4 16 32. a conic with transverse axis - 3 ≤ x ≤ 3 and conjugate
17. 2 2
4 ( χ - 3 ) - 9 ( y + 1) = 36 axis - 4 ≤ y ≤ 4
18. 9(x + 1)2 —4(y —3)2 = 36 33. an ellipse with center at (2, 1), semiminor axis with length
19. 4x 2 - 9p 2 - 8x + 54y - 4 1 = 0 3, and vertices at (2, 6) and (2, -4 )
34. a conic with foci at ( - 1, 0) and (1,0) with major axis with Cartesian equation for this hyperbola. Note'. This is the
length 12 reason sinh t and cosh t are called “ hyperbolic functions.”
35. a conic with foci at (0, 6) and (0, —6) with transverse axis 54. Show that the equations x = x 0 + a sinh/t and
with length 12 y = y 0 + b cosh t are parametric equations for one branch
36. the set o f points such that for any point, the sum of its of a hyperbola. How does this hyperbola differ from the
distances from (4, —3) and (—4, -3 ) is 12 one in Problem 53?
37. the set o f points such that, for any point, the difference of 55. Find the area of the region bounded by the hyperbola
its distances from (4, -3 ) and ( -4 , -3 ) is 6 9x 2 - 4)'2 = 36 and the vertical line x = 4.
38. a hyperbola with vertices (3, 0), ( -3 , 0) and asymptotes 56. An equilateral hyperbola is one with an equation of the
y = 3x and y = - 3 x general form y 2 - x 2 = a 1 . Show that a hyperbola in
standard form is equilateral if and only if its asymptotes
θ Sketch the graph o f each polar-form equation in Problems are perpendicular to each other.
39-42. 57. Let R be the region bounded by the x-axis, the line x = 4,
î<) r = ----- - ----- 40 r = ------ - ------ and the portion of the hyperbola 9x 2 - 4y2 = 36, with
6 + cos θ ' 2 + 3 cos θ y ≥ 0. Find the volume of the solid generated by revolv
41 r = ------ - ------ 42 r = — → —
ing R about
1 - 2 sin θ ' 2 - sin θ a. the x-axis
43. Find an equation for the tangent line to the ellipse b. the y-axis
5x2 + 4y2 = 56 at the point ( -2 , 3). 58. The orbit of a planet is an ellipse whose major axis and
44. Find an equation of an ellipse that is tangent to the minor axis are, respectively, 100 million and 81 million
coordinate axes and the line y = 6. Is there only one such miles long. Find an equation for the ellipse. How far apart
ellipse? are its foci?
x2
45. Find two points on the ellipse + y^ = 1 where the
tangent line also passes through the point (0, -2 ).
46. Find the smallest distance from the point (2, 0) to the
ellipse 3x 2 + 2y 1 + 6x — 3 = 0
47. Find the volume of the solid generated by revolving the
top half ( j ≥ 0) of the ellipse
about the x-axis.

48. Find the volume of the solid generated by revolving the
right half (x ≥ 0) of the ellipse
Figure 9.66 Planetary orbits
59. The orbit of the earth around the sun is elliptical with the
about the y-axis. sun at one focus. The semimajor axis of this orbit is
49. Let R denote the region of the plane that lies inside the 9.3 × 107 mi and the eccentricity is about 0.017. Deter
ellipse mine the greatest and least distance o f the earth from the
sun (correct to two significant digits). Hint'. Use polar
coordinates.
θ 60. Consider a person A who fires a rifle at a distant gong B.
Find the volume of the solid generated by revolving R Assuming that the ground is flat, where must you stand to
about the line y = —2. hear the sound of the gun and the sound of the gong
simultaneously? Hint: To answer this question, let x be the
50. Find an equation for the hyperbola with vertices (3, -1 ),
distance that sound travels in the length of time it takes
( - 1, - 1), and asymptotes y = ∣x - ψ and y = - ∣ x + ∣. the bullet to travel from the gun to the gong. Show that the
51. Find an equation for the hyperbola with vertices (9, 0), person who hears the sounds simultaneously must stand
( -9 , 0) whose asymptotes are perpendicular to each other. on a branch of a hyperbola (the one nearest the target), so
52. Find an equation for a hyperbola in standard position the difference of the distances from A to B is x.
that contains the points (3, ∖z 5∕2) and ( -2 , 0). 61. Three L O R A N stations are located at (4, 0), (0, 0), and
53. Show that x = x 0 + a cosh t, y = y,0 + b sinh t are para (4 , 2f) in a polar coordinate system. Radio signals are sent
metric equations for one branch of a hyperbola. Find a out from all three stations simultaneously. An airplane
receiving the signals notes that the signals from the second 67. Note that a hyperbola never intersects its asymptotes.
and third stations arrive 2/c seconds later than the signal Show that any line parallel to an asymptote will intersect
from the first, where c is the velocity of a radio signal. the hyperbola exactly once.
What is the location (in polar coordinates) of the air
plane? 68. As with the parabola, a focal chord of an ellipse or
hyperbola is a chord through a focus perpendicular to the
62. Derive the equation for the ellipse with foci F 1(-c , 0) and
principal axis.
F 1(c, 0) for c > 0 and constant distance 2α.
a. Show that the focal chords of the ellipse and the
63. Derive the equation for the hyperbola with foci F 1(-c , 0)
hyperbola
and F 2 (c, 0) for c > 0 and constant distance 2α.
64. Find conditions on the coefficients of the equation x 2 + y⅛2 = 1 and x⅜2 - v⅛2 = 1
⅞
Ax2 + Cy2 + Dx + Ey + F = 0 a 2 b2 a 1 b~
with AC > 0 that guarantee the graph of the equation 2b 2 .

each have length —
will be
a. a line b. Show that the tangents at the endpoints of the focal
b. an ellipse chord of an ellipse or hyperbola intersect on the
c. a circle directrix.
d. a point c. Find the length of the focal chord of the graph of the
e. a hyperbola polar equation
f. no graph
65. a. Show that the tangent line to the ellipse εp
1 - ε cos θ
69. Show that at each point P(x 0 , t,0 ) on the ellipse

at the point (x0 , y0) has the equation
⅞* ⅞22
a2 b2
a2 b2
b. Use part a to show that a tangent line to a vertex of an
ellipse in standard form is either vertical or horizontal. the focal radii F {P and F1P make equal angles with the
66. Prove that the lines y = ∩x and j = -⅛ χ are asymptotes tangent line. Hint. Use the result in Problem 40, Section
of the hyperbola 1.3, to show that
x2 = , h2
tan θ,1 = tan 02,= —c
a2 b2 y0
Hint'. Show that the vertical distance D(x) between
where c = ∖ ∕a2 - b 2 . See Figure 9.68.
y = ^ V Λ 2 — a 2 and the line y = tends to 0 as x → ∞, as
Figure 9.67 Problem 66 Figure 9.68 Problem 69

Chapter 9 Review

C o n c e p t P r o b le m s 19. Give a geometric definition of a parabola, and state the

1. What are the primary representations of a point in polar standard-form equations.
form? 20. What is the reflection property of parabolas?
2. What are the formulas for changing from polar to 21. Give a geometric definition of an ellipse, and state the
rectangular form? standard-form equations.
3. What are the formulas for changing from rectangular to 22. Give a geometric definition of a hyperbola, and state the
polar form? standard-form equations.
4. What does a cardioid look like, and what are the 23. What is the reflective property of ellipses?
standard-form equations for cardioids? 24. How do you find a focus for an ellipse? A hyperbola?
5. What is the formula for the rotation of a polar-form 25. Explain or illustrate the meanings of the following
equation? terms: conic sections; eccentricity; principal axis; major
6. How do you check for symmetry with a polar-form axis; minor axis; transverse axis; conjugate axis.
graph? 26. What is the polar-form equation for the conic sections?
7. What does a limaςon look like, and what are the How do you recognize each type of conic section from
standard-form equations for limapons? this general form?
8. What is a rose curve? What are the standard-form
equations for rose curves, and how do you determine the
number of leaves? P r a c tic e P r o b le m s
9. What does a lemniscate look like, and what are the 27. Identify each of the curves whose equations are given,
standard-form equations for lemniscates? a. 3x - 2y 2 - 4y + 7 = 0 b. y 2 = x 2 —1
10. State the formula for finding the slope of a polar-form c. x 2 + y 2 + x - y = 3 d. r = 2 sin ⅞
curve.
11. What is the procedure for finding the intersection of ee'
-4÷ -9= l f —4 - - 9= 1
polar-form curves? g. r = 3 sin 2θ h. r —5 = 5 cos θ
12. What is the polar-form formula for the area of a sector?
13. Outline a procedure for finding the area bounded by a
polar graph. Identify and then sketch the curves whose equations are given
14. What do we mean by parametric equations for a curve? in Problems 28-33.
15. What are a trochoid and a cycloid? 28. C f - 2 ) 2 = 1 6 (χ-3 )
16. How do you find the derivative with parametric equa 29. r = 2 cos θ
tions? 30. 2x 2 + 3y2 - 4x + 12y = - 8
17. What is the parametric formula for arc length? 31. 9x 2 —y 2 + 12y, - 18x = 36
18. State a formula for finding area under a parametrically 32. r2 = cos 20
defined curve. 33. r = 4 - 2 cos θ
34. Find the area of the intersection of the circles
r = 2a cos θ and r = 2a sin θ, where a > 0 is constant.
Name and sketch the graph o f the curves whose equations are lr = 1 + sin 0 r2 = cos 2(0 - 7)
given in Problems 1-22. lr = 1 + cos 0 r2 = cos 20
1. r = 2 sin 0 - 3 cos 0 2. r = -2 0 , 0 ≥ 0 f r cos 0 + 2r sin 0 = 4 r = 2 - 2 cos 0
3. r = - 4 cos20 4. r = 2 cos 0 sin 0 lr = 2 sec 0 r = 2 cos 0
5. x 2 + 8y - 6 = 0 ∫r = 2 sin20 r = a( 1 + cos 0)
6. 9y2 - 25x 2 - 90y - 50x - 25 = 0 lr = 1 for 0 ≤ 0 ≤ 7 r = α(l — sin 0)
7. x 2 + 2x - 2y + 2 = 0 8. 9v2 + 4x 2 - 54y + 45 = 0 lr = α(l + sin 0)
9. r = —esc 0 10. r = sin 20 11. r = cos (0 - y) lr = α(l - sin 0) a
12. r = sin + 0) 13. r = 3 cos 20 14. r 2 = cos 0
Find the area o f the regions described in Problems 59-65.
1ς , 4 _ 6
1 — cos 0 - 4( 1 + sin 0) 59. the region bounded by the polar curve r = 4 cos 20
17 60. the region inside both the circles r = 1 and r = 2 sin 0
2 + 3 sin 0 2 + cos 0 61. the region inside one loop of the lemniscate r2 = cos 20
19. r = 9 cos 0 sin 2 0 20. r = 5 sin 20 esc 0 62. the region inside the rose petal r = sin 20 for 0 ≤ 0 ≤ 7
21. r = 2 sec 0 - 4 22. r = 3 sin (20 - 1)
63. the region inside the circle r = 2a sin 0 and outside the
Sketch the curves in Problems 23-26 and fin d all points where circle r = a
the tangent line is horizontal. 64. the region under the curve given parametrically by
23. r = I - 2 cos θ 24. r = 2 + 4 sin θ
25. 26. r = 3 x = -c o s 4 0, p = sin 0 for 0 ≤ 0 ≤ J
1 — sin θ 4 - 3 cos 0
In Problems 27-38, convert each polar equation to Cartesian 65. the region bounded by the spiral r = e2β and the x-axis for
form and each Cartesian equation to polar form. 0 ≤ 0 < 77
27. 2x + 3y = 4 28. r = - 2 sin θ 66. Find the slope of the tangent line to the spiral r = 20 at
29. x 2 + y 2 - 3x = 5 30. r = cos 20 the point where 0 = j . Find an equation (in either polar
31. x 2 + y 2 = tan~1 θθ 32. r = 1 + cot 0 or Cartesian coordinates) for this tangent line.
33. y = m x 34. x 2 + y 2 + ax = α V x 2 + y 2 67. Find the length of the curve given by x = 1 + cos 2t,
2 2
35. (x - α) + y = a 2
36. (x 2 + y2)2 = a 2(x 2 - y 2) y = sin 2t for - 7 ≤ t ≤ %.
68. Find the length of the curve given by x = y = sin t
37. 36 36
38. for 0 ≤ t ≤ 7.
13 cos2 0 - 4 5 cos2 0 + 4
Find a Cartesian-form equation for each curve described in 69. Find the arc length of the spiral given by x = 7 ’ y = -72 for
Problems 39-48. t ■ t
1 ≤ t ≤ 2.
39. the parabola with focus (0, -5 ) and directrix y = 5
70. A tangent line to the hyperbola 9x 2 —y 2 = 36 intersects
40. the ellipse with vertices (0, 8), (0, -8 ) and foci (0, 4), the y-axis at (0, 6). Find all possible points o f tangency.
(0, -4 )
71. Let C be given by x = a cot 0, y = a sin 2 0. Show that the
41. the hyperbola with foci (8, 0), ( -8 , 0), and vertices (4, 0), Cartesian form of the curve C is y = α 3∕(α 2 + x 2) and
( -4 , 0) sketch C . This curve is called the witch of Agnesi.
42. the ellipse with center (2, 1), focus (1, 1), and semimajor 72. a. Find an equation for the tangent line to the curve C
axis 2 given by x = 4 cos t, y = 3 sin t, at the point where
43. the set of points such that, for each point, the difference of t = τr∕6.
the distances from (3, 0) and (9, 0) is always 4 b. Find an equation for the line perpendicular to the
44. the parabola whose vertex is (4, 3) and directrix is x = 1 curve C in part a at the point where t = τr∕4.
45. the set of points such that, for each point, the difference of 73. Find the equation of the tangent line to the cycloid
distances from ( -3 , 4) and ( -7 , 4) is equal to 2 x = a{θ - sin 0), y = a(l —cos 0) at the point where
46. the set of points such that, for each point, the sum of the a. 0 = J b. 0 = 7r c. 0 = 2 T7
distances from ( -3 , 4) and (—7, 4) is 12
74. a. Show that the tangent line to the cycloid
47. the ellipse with foci (2, 3) and ( -1 , 3) and eccentricity ∣
48. the hyperbola with vertices (0, —3) and (0, 3) and eccen x = a(θ - sin 0), y = α(l - cos 0)
tricity ∣ θa
Find the points o f intersection o f the polar-form curves in has slope c o t y at the point where 0 = 0θ.
Problems 49-58. b. For what values o f 0 is the tangent line to the cycloid
.„ ∫r = 2 co s0 ∫ r = l+ c o s 0 horizontal?
49. 1 , 50. ,
l r = 1 +l cos n0 lr = 1 — cos 0 n
c. For what values o f 0 is the tangent vertical?
75. A curve C is given parametrically by x = 2t 2 + 1, Sketch the graph of an epitrochoid for a = 4, b = 3, and
y = t - 1 for all t. Find all points on C where the tangent k = 5.
line passes through the point (7, 1). 91. Find the length of the hypocycloid
76. Find a polar equation for the line through (r1, 01) and
x = (a - b) cos 0 + b cos θ
(r2 , 02).
77. Find the area under one arch of the cycloid y = (a — b) sin θ — b sin 0
x = a(θ — sin θ), y = a(l — cos θ) where a > 0 is a con
stant. Hint'. One arch corresponds to the arc obtained for 0 ≤ 0 ≤ ~π (assume a > b > 0).
when the point on the rim of the wheel begins on the
92. Two LO R A N stations A and B are 100 mi apart. An
ground and returns to the ground.
airplane is flying a level course 70 mi parallel to the line
78. Find the length of one arch o f the cycloid x = 9(0 — sin θ), between A and B. Signals are sent from A and B at the
y = 9(1 - cos 0). same time, and the signal from A reaches the plane 400 μs
79. Show that the parametric equations before the one from B. Assuming the signals travel at
0.187 mi∕μ,s, locate the position of the plane in relation to
A and B. Hint'. It may help to set up a coordinate system
describe the unit circle x 2 + y 2 = 1. with A at the origin, B at the point (100, 0), and the plane
80. Find all points on the curve C defined by x = 3 - 4 sin /, flying along the line y —70.
y = 4 + 3 cos / where the tangent line is horizontal. 93. Points in a certain city are given coordinates such that
81. Let x = sin l + cos t, y = cos t - sin t. store A is at the origin (0, 0) and store B is at (10, 0), where
a. Show that dy/dx = -χ∕y. the units are in miles. Suppose hamburger costs $3.00 per
pound at stored and $2.75 at store B and that the cost of
b. Find a relationship between x and y.
traveling anywhere in the town is 304 per mile. Let P(x, y)
82. Suppose x = x(t), y —y(t) for a < t < b. Show that be a point in the city where it costs the same to buy one
d 2y _ /d 2y dχ _ d 2χ <M / fd x ∖3 pound of hamburger by traveling to store A as it does to
dx 1 ∖dt2 dt dt 2 dt)/ ∖dt) travel and buy at store B. Find an equation for the
coordinates of P. In other words, find a curve C on which
83. Find the slope of the tangent line to the polar curve it doesn’t matter where you shop. This is sometimes
r = a tan ∣ at the point where θ = ?. called an indifference curve.
84. A particle moves along the circle r = 4 cos θ. If f - = 2, 94. P U T N A M E X A M IN A T IO N P RO BLEM Consider the two mu
find the rate at which r is changing with respect to t when tually tangent parabolas y = x 2 and y = - χ 2 . The upper
0 = f. parabola rolls without slipping around the fixed lower
a parabola. Find the locus (set of possible points) of the
85. A particle moves along the parabola r - γ + cos g∙
focus of the moving parabola, as shown in Figure 9.69.
certain instant, ⅛ = - 3 cm ∕s, and θ = ? . Find
. ., . . . . dt 4 dt
at this instant.
86. Sketch the curve r = sec 0 —2 cos θ for —J < θ < J and
find the area enclosed by its loop. This curve is called a
strophoid.
87. The circle r = 1 of radius 1 centered at the origin has the
property that any “ diameter” (line through the origin)
intersects the circle in a segment of length 2.
a. Show that the cardioid r = 1 + sin 0 has the same
property.
b. Show that the region bounded by the cardioid and the Figure 9.69 Problem 94
circle has different areas. Thus, the area of a region
cannot be determined in terms o f its diameters alone. 95. P U T N A M E X A M IN A T IO N P ROBLEM For a point P on an
88. The major and minor axes of an ellipse have lengths 10 ellipse, let d be the distance from the center o f the ellipse
and 5, respectively. Find the volume obtained by revolv to the line tangent to the ellipse at P. Prove that
ing the elliptic region on one side of the major axis about ∣P F I ∣∣EF 2∣ d 2
its major axis.
89. A straight rod o f fixed length L moves in such a way that is constant as P varies on the ellipse, where ∣PFj∣ and ∖P F 2 ∖
its top always touches the y-axis and its bottom touches are the distances from P to the foci F ∣ and F 2 o f the
the x-axis. Let P(x, y) be the point on the rod that is j the ellipse.
distance from the bottom to the top o f the rod. Describe 96. P U T N A M E X A M IN A T IO N P RO BLEM Find the minimum
the path that P traces out as the rod slides down the y-axis. value of
90. An epitrochoid is a curve with the parametric equations (∕z - v)2 + f√ 2 - u 2 - ≤
x = a cos 0 + b cos kθ, y = a sin 0 + b sin kθ for 0 0.
“Security System Project0
You are designing a security system for a hospital. You must decide how to
program a detector that will be used to watch a 40-ft-long hallway with a door in
the middle. The detector runs on a track and points a beam of light straight ahead
on the opposite wall. The beam reaches from the floor to the ceiling. Design a
security system that will detect an intruder.

a. Think of the hallway as a coordinate line with the middle of the door at the
origin and the hallway to be watched on the interval [-20, 20]. You need to
decide what x(t) is for t, in seconds, where x(f) represents the position of
the beam at time t. For example, x(5) = -1 5 means that the beam is
pointing at the part of the wall 15 ft to the left of the door 5 sec after the
detector starts.
Everything that the greatest minds
o f all times have accomplished to b. Assume that t ranges from 0 to 30 min.
ward the comprehension o f forms c. The door is 3 ft wide, and as long as the beam is hitting any part of the
by means o f concepts is gathered door, it is under surveillance.
into one great science-, mathemat d. The beam must stay on an object for at least 0.1 sec in order to detect that
ics.
object.
J. F. Herbart
concerns. Draw a graph of x versus time for what you think is a good choice for
x(t). You should mention restrictions on this function, if any, and include the
longest time interval that the door will not be under surveillance. Could an
intruder get to the door by walking down the hallway without being detected by
your system? Explain how the intruder could do it and how likely you think it is.
The group project is courtesy of Steve Hilbert, Ithaca College, Ithaca, New York.
______ 10
Vectors in
the Plane
and in Space
■ CONTENTS
10.1 Vectors in the Plane
PREVIEW
Introduction to vectors;
Standard representation of In this chapter, we focus on various algebraic and geometric aspects of
vectors in the plane vector representations. Then in Chapter 11, we see how vectors can be
10.2 Quadric Surfaces and Graph combined with calculus to study motion in space and other applications.
ing in Three Dimensions
Three-dimensional coordinate
system; Graphs in R 3 (planes,
spheres, cylinders, and quadric
surfaces)
PERSPECTIVE
10.3 The Dot Product
Vectors in R 3; Definition of Suppose a child is pulling a sled by a rope across a flat field. Intuitively, we
dot product; Angle between would expect quite different results if the child pulls the rope straight up than if
vectors; Projections; Work as
a dot product the same effort is applied at an angle of π∕4, for instance. In other words, to
10.4 The Cross Product describe the force exerted by the child on the sled, we must specify not only a
Definition of cross product; magnitude but also a direction. In general, a scalar quantity is one that can be
Geometric interpretation of described in terms of magnitude alone, whereas a vector quantity requires both
cross product; Properties of magnitude and direction. Force, velocity, and acceleration are common vector
cross product; Triple scalar
product, volume; Torque quantities, and an important goal of this chapter is to develop methods for
10.5 Lines and Planes in Space dealing with such quantities.
Lines in R 3; Direction cosines;
Planes in R 3
10.6 Vector Methods for Measuring
Distance in R3
Distance from a point to a
plane; Distance from a point
to a line
Chapter 10 Review
630
Sec. 10.1 Vectors in the Plane 631
10.1 VECTORS IN THE PLANE
IN THIS SECTION Introduction to vectors, standard representation of

vectors in the plane
Many applications of mathematics involve quantities that have both magnitude
and direction, such as force, velocity, acceleration, and momentum. Vectors are
an important tool in mathematics, and in this section we introduce you to the
terminology and notation for vector representation.
■ INTRODUCTION TO VECTORS
A vector in a plane can be thought of as a directed line segment, an “ arrow” with

initial point P and terminal point Q. The direction of the vector is that of the arrow,
and its magnitude is represented by the arrow’s length, as shown in Figure 10.1.
We shall indicate such a vector by writing P Q in boldface print, but in your work
you may write an arrow over the designated points: PQ. The order of letters you
write down is important: P Q means that the vector is from P to Q, but Q P means
that the vector is from O to P. The first letter is always the initial point and the
second letter is the terminal point. We shall denote the magnitude (length) of a
vector by ∣ ∣
PQ∣ ∣
. Two vectors are regarded as equal (or equivalent) if they have the
same magnitude and the same direction, even if they do not coincide.
a. The vector PQ has b. Two vectors are equal if

magnitude ∣∣PQ ∣
∣
. they have the same magnitude
and direction.
Figure 10.1 Vectors in a plane
A vector with magnitude 0 is called a zero (or null) vector and is denoted by 0.
The 0 vector has no specific direction, and we shall adopt the convention of
assigning it any direction that may be convenient in a particular problem.
If v is a vector other than 0, then any vector w that is parallel to v is called a
scalar multiple of v and satisfies w = sv for some nonzero number s. A scalar
quantity is one that has only magnitude and in the context of vectors is used to
describe a real number. The scalar multiple sv has length ∣s∣ times that of v; it
points in the same direction as v if s > 0 and the opposite direction if 5 < 0 (see
Figure 10.2). Notice that for any distinct points P and Q, P Q = - Q P . For the zero
Figure 10.2 Some multiples of vector 0, we define sθ = 0 for any scalar 5.
the vector u Physical experiments indicate that force and velocity vectors can be added (or
632 Ch. 10 Vectors in the Plane and in Space
resolved) according to a triangular rule displayed in Figure 10.3a, and we use this
rule as our definition of vector addition. In particular, to add the vector v to the
vector u, we place the end (initial point) of v at the tip (terminal point) of u and
define the sum, also called the resultant u + v, to be the vector that extends from
the initial point of u to the terminal point of v.
Equivalently, u + v is the diagonal of the parallelogram formed with sides u
and v, as shown in Figure 10.3b. The difference u - v is just the vector w that
satisfies v + w = u, and it may be found by placing the initial points of u and v
together and extending a vector from the terminal point of v to the terminal point
of u (see Figure 10.3c).
a. The triangle rule for b. The parallelogram rule c. The difference rule
adding vectors u and v for vector addition
Figure 10.3 Vector addition and subtraction
A vector OQ with initial point at the origin O of a coordinate plane can be

uniquely represented by specifying the coordinates of its terminal point Q. If Q
has coordinates (a, b), we denote the vector OQ by (a, b), where the pointed
brackets ( ) are used to distinguish the vector OQ = (a, b) from the point (a, b).
The vector PQ with initial point P(c, d) and terminal point Q(a, b) can be
denoted in a similar fashion. Using analytic geometry (see Problem 56), it can be
shown that PQ equals vector OR with initial point at the origin (0, 0) and terminal
point R(a - c, b - d), as shown in Figure 10.4. Notice that PQ = (a - c, b - d).
Figure 10.4 Given P(c, d) and Q(a, b), the vector PQ is (a - c, b - d').
Vector operations are easily represented when vectors are given in component
form. In particular, we have
(α 1, ⅛1> = (α2 , ⅛2) if and only if a χ = a1 and b χ = b1
k{a, b) = {ka, kb) for constant k
{a, b) + (c, d) = (α + c, b + d)
{a, b) - (c, d) = (a - c, b - d)
These formulas may be verified by analytic geometry. For instance, the rule for
multiplication by a scalar may be obtained by using the relationships in Figure
10.5a, and Figure 10.5b can be used to obtain the rule for vector addition.
a. Multiplication by a scalar: b.Vector addition

k(a, b) = {ka, kb) (a, b) + (c, d) = (a + c, b + d)
Figure 10.5 Vector operations
EXAM PLE 1 Vector operations
For the vectors u = (2, -3 ) and v = ( - 1, 7) find:
a. u + v b. c. 3u - ∣
v
Solution a. u + v = ( 2 ,-3 ) + ( - 1, 7) = (2 + ( - 1 ) ,-3 + 7) = (1, 4)

b. ∣
u = ¾{2,-3 ) = (^(2), i ( → ) H ⅛ i 2 )
c. 3 u - ∣ v = 3 ( 2 ,- 3 > - ∣ ( - l , 7)
= (6, -9 ) + —2 ,i Scalar multiplication
= (β + —9 — Add vectors.
= l∖T ~ 2T) Simplify.
In general, an expression of the form au + by is called a linear combination of

the vectors u and v. Note that if u = (wp uf) and v = (v1, v,), then
au + by = a{uγ, ιz2) + ⅛(vp v2) = (αzz1 + bvγ, au, + ⅛v2)
Vector addition and multiplication of a vector by a scalar behave very much
like ordinary addition and multiplication. The following theorem lists several
useful properties of these operations.
T H E O R E M 10.1 Properties of vector operations
For any vectors u, v, and w in the plane and scalars 5 and t.
Commutativity of vector addition u + V= V+ u

Associativity of vector addition (u + v) + w = u + (v + w)
Associativity of scalar multiplication (sZ)u = s(Zu)
Identity for addition u+ 0= u
Inverse property for addition u + (-u) = 0
Vector distributivity (5 + Z)u = SU + Zu
Scalar distributivity s(u + v) = SU + SV
Proof: Each vector property can be established by using a corresponding

property of real numbers. For example, to prove associativity of vector addition,
let u = (u 1, u f , v = (v1, v f and w = (w1, w f. Then
(u + v) + w = ((M1, M2) + (v1, v2)) + (w1, w2)
= <M1 + V1 , M2 + V2 ) + <W1 , W2 )
= + v1) + W1, (w2 + v2) + w2)

= (M1 + (v1 + W1), M2 + (v2 + W2)) Associativity of addition for the real numbers
= {u l , u f + (<vp v2) + (w1, w2))
= u + (v + w)
You are asked to prove the other six properties in the problem set. ■
E X A M P L E 2 Vector proof of a geometric property
Show that the line segment joining the midpoints of two sides of a triangle is
parallel to the third side and has half its length.
c Solution Consider Δ A B C and let P and Q be the midpoints of sides A C and B C ,
/ respectively, as shown in Figure 10.6.
A P = 1A C
P∕
Z ^∖
∣ Given-.
BQ = 2 b c A P = ∣A C and BQ = ∣B C
Prove-. P Q is parallel to AB and ∣∣PQ∣∣ = 5∣∣AB∣∣, which means that
A
we must establish the vector equation P Q = ∣A B . Toward this end, we
Figure 10.6 Vector proof o f a
geometric property begin by noting that AB can be expressed as the following vector sum
AB = A P + P Q + Q B
= jA C + P Q - BQ AP = ∣
A C and Q B = - B Q
= 5(AB + BC) + P Q - ⅛ C A C = (AB + BC) and BQ = ∣

BC
= ∣A B + ∣B C + P Q - ⅛ C
= ∣A B + PQ
jA B = PQ Subtract ∣AB from both sides.

a. The vector u = (u 1, u 2 )
has length _______ Therefore, the theorem is proved.
l∣
u∣∣= √ U12 + U22
When a vector u is represented in component form u = (M1, 1∕2), its length is
given by the formula
∣
∣u∣
∣= V M2 + M 2
This is a simple application of the Pythagorean theorem, as shown in Figure 10.7a.
Another important relationship involving the length of vectors is the triangle
inequality
∣
∣
u + v∣
∣≤ ∣
∣
u∣
∣+ ∣
∣
v∣
b. The triangle inequality for any vectors u and v. Equality will occur precisely when u and v are multiples of
∣
∣u + v∣∣< ∣ ∣u∣∣+∣∣v∣∣ one another (that is, when u and v have the same direction). To establish the
Figure 10.7 Geometric inequality, we observe that u and v are two sides of a triangle in the plane, the third
representation o f two vector side has length ∣
∣u + v∣
∣and is “ shorter” than the sum ∣ ∣
u∣∣+ ∣∣
v∣∣of the lengths of the
properties other two sides, as shown in Figure 10.7.
EXAM PLE 3 Using vectors in a velocity problem
A river 4 mi wide flows south with a current o f 5 mi∕hr. What speed and
heading should a motorboat assume in order to travel directly across the river
from east to west in 20 min?
Solution Begin by drawing a diagram, as shown in Figure 10.8.
Let B be the velocity vector of the boat in the direction of the angle θ. If the
river’s current has velocity C, the given information tells us that ∣∣
C∣∣= 5 mi/h
and that C points directly south. Moreover, because the boat is to cross the
river from east to west in 20 min (that is, ∣ hr), its effective velocity after
compensating for the current is a vector V that points west and has magnitude
∣∣v ∣∣ = WIDTH OF THE RIVER = 4∏11 = . ? .,.
11 11 TIME OF CROSSING 1 hr
The effective velocity V is the resultant of B and C; that is, V = B + C.

Because V and C act in perpendicular directions, we can determine B by
referring to the right triangle with sides ∣
∣
V∣∣= 12 and ∣
∣
C∣∣= 5 with hypotenuse
∣
∣B∣∣
. We find that
∣
∣
B∣∣= V ∣
∣
V∣∣
2 + ∣
∣
C∣∣
2 = V 1 2 2 + 52 = 13
The direction of the velocity vector V is given by the angle θ in Figure 10.8,
and we find that
tan θ = ]⅜ so that θ = tan -1 (-∣⅜) ≈ 0.3948
Thus, the boat should travel at 13 mi/hr in a direction of approximately

Figure 10.8 A velocity problem 0.3948 radian. In navigation, it is common to specify direction in degrees
rather than radians; this is 22.60 north of west. ⊂=□
A unit vector is just a vector with length 1, and a direction vector for a given
vector v is a unit vector u that points in the same direction as v. Such a vector can
be found by simply dividing v by its length ||v||; that is,
EXAM PLE 4 Finding a direction vector
Find a direction vector for the vector v = (2, -3 ).

Solution The vector v has length (magnitude) ∣ ∣
v∣∣= V 2 2 + ( -3 ) 2 = √T3. Thus,
the required direction vector is the unit vector
u = AL = (2, -3 ) = _ j_ _ = ∕j =
I∣∣ √∏
v∣ √ IT ’ ' W TTV T3∕ —
■ STANDARD REPRESENTATION OF VECTORS IN THE PLANE
The unit vectors i = (1, 0) and j = (0, 1) point in the directions of the positive x-
and j--axes, respectively, and are called standard basis vectors. Any vector
v = (V], v2) in the plane can be expressed as a linear combination of the vectors i
and j, because
v = <v1, v2) = v1( l, 0) + v2(0, 1) = v1i + v2j
This is called the standard representation of the vector v, and it can be shown that
the representation is unique in the sense that if v = ai + by, then a = v1 and b = v2 .
In this context, the scalars v1 and v2 are called the horizontal and vertical
components of v, respectively. See Figure 10.9.
v = v 1i + v2 j
Figure 10.9 Standard representation of vectors in the plane
EXAM PLE 5 Finding the standard representation of a vector
If u = 3i + 2j, v = - 2 i + 5j, and w = i - 4j, what is the standard representa

tion of the vector 2u + 5v - w?
Solution Using Theorem 10.1, we find that
2u + 5v - w = 2(3i + 2j) + 5 (-2 i + 5j) - (i - 4j)
= [2(3) + 5 (-2 ) - l]i + [2(2) + 5(5) - (-4 )]j
= - 5 i + 33j ≡≡≡
EXAM PLE 6 Finding standard representation of a vector connecting two

points
Find the standard representation of the vector PQ for the points P(3, - 4 ) and
β ( - 2 , 6).
Solution The component form of PQ is
PQ = ( ( - 2 ) - 3 , 6 - ( - 4 ) ) = ( - 5 , 10)
which means that PQ has the standard representation PQ = - 5 i + 10j. ≡≡≡
EXAM PLE 7 Computing a resultant force
Two forces F 1 and F 2 act on the same body. It is known that F 1 has magnitude
3 newtons and acts in the direction o f - i , whereas F 2 has magnitude 2
newtons and acts in the direction of the unit vector
u = ji - f j
W hat additional force F 3 must be applied in order to keep the body at rest?
Solution According to the given information, we have
F, = 3 (-i) = - 3 i and F 2 = 2(∣i - fj) = fi - fj
and we want to find F 3 = ai + bj so that F 1 + F2 + F 3 = 0. Substituting into

this vector equation, we obtain
( -3 i) + (fi - fj) + (ai + bj) = 0i + 0j
By combining terms on the left, we find that
(-3 '+ f + α)i + ( - f + &)j = 0i + 0j
Because the standard representation is unique, we must have
- 3 + f + α = 0 and -f + b= 0
a=l b=l
i+∣
The required force is F 3 = ∣ j. This is a force of magnitude
∣∣F 3∣∣ = 7⅛ 2 + (!)2 = ∣V145 newtons
which acts in the direction of the unit vector
PROBLEM SET 1 0 .1
O Sketch each vector given in Problems 1 -4 assuming that its 28. (y - l)i + yj = (log x)i + [log2 + log(x + 4)]j
initial point is at the origin.
In Problems 29-32, fin d a unit vector u with the given
1. 3i - 4j 2. - 2 i - 3j i+∣
3. - ∣ j 4. - 2 ( - i + 2j)
characteristics.
29. u makes an angle of 30° with the positive x-axis.
The initial point P and terminal point Q o fa vector are given in
Problems 5-8 . Sketch each vector and then write it in compo 30. u has the same direction as the vector 2i - 3j.
nent form. 31. u has the direction opposite that of - 4 i + j.
5. P(3, -1 ), β(7, 2) 6. P(5, -2 ) , β(5, 8) 32. u has the direction of the vector from P ( - l , 5) to
β(7, -3 ).
7. P(3, 4), β ( - 2 , 4) 8. P⅛. 6), β ( - 3 , -2 )
In Problems 33-36, let u = 4i - j, v = i + 2j, and
Express each vector PQ in Problems 9-12 in standard form w = -3 i + 4j.
and also fin d its length. 33. Find a unit vector in the same direction as u + v.
9. P ( - l , -2 ) , β ( l, -2 ) 10. P(5, 7), β(6, 8) 34. Find a vector of length 3 with the same direction as
11. P ( -4 , -3 ) , β(0, -1 ) 12. P(3, -5 ), β(2, 8) u - 2v + 2w.
Find a unit vector that points in the direction o f each o f the 35. Find the terminal point of the vector 5i + 7j if the initial
vectors given in Problems 13-16. point is ( -2 , 3).
13. i + j i+∣
14. ∣ j 36. Find the initial point of the vector —i + 2j if the terminal
point is ( -1 , -2 ).
15. 3i - 4j 16. -4 i + 7j O 37. Use vectors to find the coordinates o f the midpoint of the
Let u = (—3, 4) and v = (1, —1). Find s and t so that line segment joining the points P(—3, -8 ) and β(9, —2).
su + tv = w fo r the given vector in Problems 17-20. What point is located j o f the distance from P to Q I
17. w = <6, 0) 18. w = (0, -3 )
38. ∣
v∣
If ∣ ∣ = 3 and - 3 ≤ r ≤ 1, what are the possible values
19. w = < -2 , 1) 20. w = <8, 11)
of H I?
Suppose u = 3i - 4j, v = 4i - 3j, and w = i + j. Express each 39. Show that v = (cos 0)i + (sin Θ)j is a unit vector for any
o f the expressions in Problems 21-24 in standard form. angle θ.
21. 2u + 3v - w 22. ∣
(u + v) - ∣
w 40. If u and v are nonzero vectors and
23. ∣
∣
v∣
∣ ∣
u∣
u+ ∣∣v ∣
24. ∣∣
u∣∣
∣
v∣
∣
w =w
r
Find all real numbers x and y that satisfy the vector equations
„ ■ II IIO llv ll
what is ∣∣ ∣
rv∣7
given in Problems 25-28. 41. If u and v are nonzero vectors with ∣∣u∣∣ = ∣∣v∣∣, does it follow
25. (x - y - 1)i + (2x + 3y - 12)j = 0 that u = v? Explain.
26. xi - 4y2j = (5 - 3y)i + (10 - 7x)j 42. If u = 2i — 3j and v = xi + yj, describe the set of points in
27. (x2 + y 2)i + yj = 16i + (x + 2)j the plane whose coordinates (x, y) satisfy ∣∣v - u∣∣ ≤ 2.
43. Let u0 = xθi + y 0 j for constants x 0 and y 0 , and let a, b, show that a = b = 0.
u = x i + yj. Describe the set o f all points in the plane b. Show that the standard representation of a vector is
whose coordinates satisfy unique. That is, if the vector u has the representation
a. ∣
∣u - u0∣∣= 1 b. ∣∣
u - u0∣∣≤ 2 u = α l i + ⅛1j and u = α2i + ⅛J is another such repre
44. Let u = 3i —j and v = —6i + 2j. Show that there are no sentation, then <71 = α, and ⅛ = b 1.
numbers a, b for which αu + ⅛v = 2i + 5j. 55. Prove that the medians of a triangle intersect at a single
45. Suppose u and v are a pair o f nonzero, nonparallel point by completing the following argument.
vectors. Find numbers a, b, c such that c
a. Let M and N be the midpoints
au + A(u - v) +
o f sides A C and A B , ∕ ∖∖
c(u + v) = 0.
respectively. Show that / ∖ x∖
∣
u∣
46. Suppose u and v are perpendicular vectors, with ∣∣= 2 1 ∖ D
and ∣∣v∣∣ = 4. Sketch the vector cu + (1 — c)v for the cases C N = ⅛AB - A C and ∖
where c = 0, c = ∣ , c = ∣ , c = ∣ , and c = 1. In general, if BM = ∣ A C - AB Z ≤ -
the initial point o f a vector cu + (1 - c)v with 0 ≤ c ≤ 1
is at the origin, where is its terminal point? A N ~ B
47. Two forces F 1 = 3i + 4j and F , = 3i - 7j act on an object. b. Let P be the point where medians B M and C N inter
What additional force should be applied to keep the body sect. Show that there are constants r and s such that
at rest? C P = r(∣
A B - AC) and
48. Three forces F 1 = i - 2j, F , = 3i — 7j, and F 3 = i + j act
on an object. What additional force F 4 should be applied BP = s(∣
A C - AB)
to keep the body at rest?
49. A rock is thrown into the air at an angle of 40o with the Note that C P + PB = C B . Use this relationship to
horizontal and with an initial velocity of 60 ft∕s. Find the prove that r = s = j . Explain why this shows that any
vertical and horizontal components of the velocity vector. pair of medians meet at a point located ∣ the distance
50. A river 2.1 mi wide flows south with a current of 3.1 mi∕hr. from each vertex to the midpoint o f the opposite side.
What speed and heading should a motorboat assume to Why does this show that all three medians meet at a
travel across the river from east to west in 30 min? single point?
© 51. Four forces act on an object: F 1 has magnitude 10 lb and c. The centroid of a triangle is the point where the
acts at an angle of measured counterclockwise from the medians meet. Show that if a triangle has vertices
positive x-axis; F 2 has magnitude 8 lb and acts in the (x l , y, 1), (x2 , y 2), (x3, y 3), then the centroid has coordi
direction o f the vector j; F 3 has magnitude 5 lb and acts at nates
an angle of 4τr∕3 measured counterclockwise from the
positive x-axis. What must the fourth force F 4 be to keep
the object at rest?
52. Use vector methods to show that the diagonals of a 56. Show that the vector with the initial point P(c, d) and
parallelogram bisect each other. the terminal point Q(a, b) has the component form
P Q = {a - c, b - d). See Figure 10.4.
53. In a triangle, let u, v, and w be the vectors from each
vertex to the midpoint o f the opposite side. Use vector 57. Prove the following parts o f Theorem 10.1.
methods to show that u + v + w = 0. a. commutativity b. identity
54. Two nonzero vectors u and v are said to be linearly c. vector distributivity d. associativity o f scalar
independent in the plane if they are not parallel. e. inverse property of multiplication
a. If u and v have this property and αu = ⅛v for constants addition f. scalar distributivity
1 0 .2 QUADRIC SURFACES AND GRAPHING IN THREE DIMENSIONS
IN THIS SECTION Three-dimensional coordinate system, graphs in R3

(planes, spheres, cylinders, quadric surfaces)
Because we live in a three-dimensional world, it is essential that we be able to
visualize surfaces in three dimensions. To that end we consider ordered triplets
and their representation in three-dimensional space, as well as develop some
knowledge of simple three-dimensional surfaces, namely, the quadric surfaces.
In the next section we will consider vectors in three dimensions in order to
describe motion in space efficiently.
Sec. 10.2 Quadric Surfaces and Graphing in Three Dimensions 639
■ THREE-DIMENSIONAL COORDINATE SYSTEM
Our next goal is to see how analytic geometry and vector methods can be applied
in space. We have already considered ordered pairs and a two-dimensional
coordinate system. We shall denote this two-dimensional system by R2 . Because
we exist in a three-dimensional world, it is also important to consider a three-
dimensional system. We call this three-space and denote it by R3. We introduce a
coordinate system to three-space by choosing three mutually perpendicular axes to
serve as a frame of reference. The orientation of our reference system will be right-
handed in the sense that if you stand at the origin with your arms stretched out in
the direction of the positive x-axis and y-axis, your right arm will be the one along
the x-axis and your left arm will be stretched straight ahead along the y-axis. as
shown in Figure 10.10. Your head will then point in the direction of the positive
z-axis.
In order to orient yourself to a three-dimensional coordinate system, think of
the x-axis and ρ-axis as lying in the plane of the floor and the z-axis as a line
perpendicular to the floor. All the graphs we have drawn in the first nine chapters
of this book would now be drawn on the floor. If you orient yourself in a room
(your classroom, for example) as shown in Figure 10.11, you may notice some
important planes. Assume the room is 25 ft × 30 ft × 8 ft and fix the origin at a
front corner (where the board hangs).
Figure 10.10 A “right-handed”

rectangular coordinate system
Figure 10.11 A typical classroom; assume the dimensions are 25 by 30 ft with an 8-ft
ceiling
Floor: xy-plane; equation is z = 0.

Ceiling: Plane parallel to the xj-plane; equation is z = 8.
Front wall: jz-plane; equation is x = 0.
Back wall: Plane parallel to the yz-plane; equation is x = 30.
Left wall: xz-plane; equation is y = 0.
Right wall: Plane parallel to the xz-plane; equation is y = 25.
The xy-, xz-, and yz-planes are called the coordinate planes. Points in R3 are
located by their position in relation to the three coordinate planes and are given
appropriate coordinates. Specifically, the point P is assigned coordinates (a, b, c)
to indicate that it is a, b, and c units, respectively, from the yz-, xz-, and xjxplanes.
Name the coordinates of several objects in your classroom (or in Figure 10.11).
DRAWING LESSON 2A : PLOTTING POINTS

ASSIGNMENT:
Plot P(3, 4, -5).
a. Sketch x-axis and y-axis, adding tick b. Sketch z-axis, adding tickmarks. Use
marks. Outline the xy-plane. dashed segments for hidden parts.
c. Plot x-distance and y-distance; darken

segments from each along gridlines. d. Plot z-distance, using the unit size from
Colored pencil or highlighter may the z-axis. Lightly sketch a grid on the
help you visualize the figure. xy-plane, using tickmarks as guides.
DRAWING LESSON 2B : DRAWING VERTICAL PLANES

ASSIGNMENT:
Add planes x = 2 and
y = 0 to the first sketch.
a. Draw a segment of the line x = 2 on the b. Shade the plane x = 2 where it is not hid
xy-plane. Through each endpoint, draw a den by the xy-plane. Erase hidden parts of
segment parallel to the z-axis. Then con both planes, and use your eraser to dash
nect the endpoints. hidden parts of the axes.
c. Follow the same procedure to draw and d. Use colored pencils or highlighters to
shade the plane y - 0. Draw the intersec distinguish individual planes.
tion of the two planes.
EXAM PLE 1 Points in three dimensions
Graph the following ordered triplets:
a. (10, 20, 10) b. (-1 2 , 6, 12)

c. ( - 1 2 ,- 1 8 , 6) d. (2 0 ,-1 0 , 18)
Solution
In Example 1, we measured distances in the x-, y-, and z-directions. We will,

however, also need to measure distances between points in R 3. The formula for
distance in R 2 easily extends to R 3.
Distance Formula The distance ∣Pl P 2∣ between P 1(x 1, y l , z 1) and P 2(x,, y ,, z2) is
∣P1P 2∣ = √ (⅞ - ∙χ ι)2 + (τ 2 - >'l )2 + ⅞ - -^1)2
This formula will be derived in Section 10.3. The distance between (10, 20, 10)
and ( - 12, 6, 12) is
d = √ ( - 1 2 - 10)2 + (6 - 20)2 + (12 - 10)2 = √684 = 6√19
■ G R A P H S IN R 3
The graph of an equation in R 3 is the collection of all points (x, y, z) whose

coordinates satisfy a given equation. This graph is called a surface. You are not
expected to spend a great deal of time graphing three-dimensional surfaces, but
the drawing lessons in this section should help. You may also have access to a
computer program to help you look at graphs in three dimensions. We will discuss
lines and planes in Section 10.5, but it is worthwhile to begin with a brief
introduction to certain surfaces in R 3.
P lanes. We shall obtain equations for planes in space after we discuss vectors in
space. However, in beginning to visualize objects in R 3, we do not want to ignore
planes, because they are so common (for example, the walls, ceiling, and floor in
Figure 10.11). In Section 10.5, we will show that the graph o f ax + by + cz = d is a
SM H
plane if a, b, c, and d, are real numbers (not all zero).
EXAM PLE 2 Graphing planes
Graph the planes defined by the given equations.

a. x + 3y + 2z = 6 b. y + z = 5 c. x = 4
Solution To graph a plane, find some ordered triplets satisfying the equation.
The best ones to use are often those that fall on a coordinate axis (the
intercepts).
a. Let x = 0 and y = 0; then z = 3; plot the point (0, 0, 3).
Let x = 0 and z = 0; then y = 2; plot the point (0, 2, 0).
Let y = 0 and z = 0; then x = 6; plot the point (6, 0, 0).
Use these points to drawthe intersection lines(called trace lines) of the
plane you are graphing with each of the coordinate planes. The result is
shown in Figure 10.12a.
b. When one of the variables is missing from the equation of a plane, then
that plane is parallel to the axis corresponding to the missing variable; in
this case it is parallel to the x-axis. Draw the line y + z = 5 on the yz-plane
and then complete the plane, as shown in Figure 10.12b.
c. When two variables are missing, then the plane is parallel to one of the
coordinate planes, as shown in Figure 10.12c.
b. Graph of y + z = 5 c. Graph of x = 4
parallel to the x-axis parallel to the yz-plane
Figure 10.12 Graphs of planes
Spheres. A sphere is defined as the collection of all points located a fixed

distance (the radius) from a fixed point (the center). See Figure 10.13. In
particular, if P(x, y, z) is a point on the sphere with radius r and center C(α, b, c),
then the distance from C to P is r. Thus,
r = 7(x - α)2 + (y - b)2 + (z - c)2
If you square both sides of this equation you can see that it is equivalent to the
equation of a sphere displayed in the following box. Conversely, if the point
Figure 10.13 Graph of a sphere
(x, y, z) satisfies an equation of this form, it must lie on a sphere with center
with center (a, b, c), radius r
(a, b, c) and radius r.
Equation of a Sphere The graph of the equation

(x - α)2 + (y - b)2 + (z - c)2 = r 2
is a sphere with center (a, b, c) and radius r, and any sphere has an equation of ∙
this form. This is called the standard form of the equation of a sphere (or simply
standard-form sphere).
EXAMPLE 3 Center and radius of a sphere from a given equation
Show that the graph of the equation x 2 + y 2 + z 2 + 4x - 6y - 3 = 0 is a

sphere, and find its center and radius.
Solution By completing the square in both variables x and y, we have
(x2 + 4x) + (y 2 - 6y) + z 2 = 3
(x2 + 4x + 2 2) + Ly2 - 6y + ( - 3)2] + z 2 = 3 + 4 + 9
(x + 2)2 + ( y - 3 ) 2 + z 2 = 16
Comparing this equation with the standard form, we see that it is the equation
of a sphere with center (—2, 3, 0) and radius 4. an
Generatrix
A L
Cylinders. A cross section of a surface in space is a curve obtained by intersect
ing the surface with a plane. If parallel planes intersect a given surface in
congruent cross-sectional curves, the surface is called a cylinder. We define a
cylinder with principal cross sections C and generating line L to be the surface
obtained by moving lines parallel to L along the boundary of the curve C, as
shown in Figure 10.14. In this context, the curve C is called a directrix of the
cylinder, and L is the generatrix.
We shall deal primarily with cylinders in which the directrix is a conic section
and the generatrix L is one of the coordinate axes. Such a cylinder is often named
Principal cross
for the type of conic section in its principal cross sections and is described by an
section curve C equation involving only two of the variables x, y, z. In this case, the generating line
(directrix) L is parallel to the coordinate axis of the missing variable. Thus,
x2 + y 2 = 5 is a circular cylinder with L parallel to the z-axis.
y 2 - z2 = 9 is a hyperbolic cylinder with L parallel to the x-axis.
x 2 + 2z 2 = 25 is an elliptic cylinder with L parallel to the y-axis.
The graphs of these cylinders are shown in Figure 10.15.
Figure 10.14 A cylinder with

directrix C and generatrix L
a. A circular cylinder b. A hyperbolic cylinder c. An elliptic cylinder

x2 +y2 = 5 y2 - z 2 = 9 x 2 + 2Z2 = 25
Figure 10.15 Cylinders
Quadric Surfaces. Spheres and elliptic, parabolic, and hyperbolic cylinders are
examples of quadric surfaces. In general, such a surface is the graph of an equation
of the form
Ax2 + By 2 + C∑2 + Dxy + Exz + Fyz + Gx + Hy + Iz + J = 0
Quadric surfaces may be thought of as the generalizations of the conic sections in
R 3 . The trace of a curve is found by setting one of the variables equal to a constant
and then graphing the resulting curve. If x = k (k a constant), the resulting curve is
drawn in the plane x = k, which is parallel to the j/z-plane; and if z = k, the curve
is drawn in the plane z = k, parallel to the xy-plane. Table 10.1 shows the quadric
surfaces.
DRAWING LESSON 3: SKETCHING A SURFACE
ASSIGNMENT:
nGraph1, y = ------12=---- 2y
, ' 1 + Λ' + y
a. Draw the xy-plane in three dimensions, adding

the z-axis.
b. Draw a trace in one of the coordinate planes(in c. Draw a trace in another coordinate plane (in this
this case, the plane x = 0). If necessary, adjust case, the plane y = 0).
the z-scale to show the trace more clearly.
d. Erase all hidden lines. Draw several additonal e. Erase all hidden lines. Use highlighters or pen -
trace curves to reveal the contours of the surface. cils to color the surface and the xy-plane.
TABLE 10.1 Quadric Surfaces

Surface Description Surface Description
The trace in the try-plane The trace in the xy-

is a point; in planes plane is a point; in
parallel to the try-plane planes parallel to the
it is an ellipse. Traces in try-plane it is an ellipse.
the xz- and yz-planes are Traces in the xz- and
intersecting lines; in yz-planes are parabolas.
planes parallel to these
they are hyperbolas:
Hyperboloid o f one sheet The trace in the try-plane Hyperboloid o f There is no trace in the
is an ellipse; in the xz- two sheets. try-plane. In planes
and yz-planes the traces parallel to the try-plane,
are hyperbolas. which intersect the
surface, the traces are
tr 2 , T 2 _ z 2 _ .
ellipses. Traces in the xz-
a2 b2 c2 and .yz-planes are
hyperbolas.
Ellipsoid The traces in the Hyperbolic The trace in the xy-

coordinate planes are paraboloid (also plane is two intersecting
ellipses. called a saddle} lines; in the plane
parallel to the try-plane,
^ + yi + si the traces are
a2 b2 c2 hyperbolas. Traces in
the xz- and j'z-planes
are parabolas.
_ = y∑ _ x÷
b2 a2
Sphere The sphere is a special

kind of ellipsoid for
which a = b = c = r.
x 2 + y 2 + z 2 = r2
EXAMPLE 4 Identifying and sketching a quadric surface

Identify and sketch the surface with equation 9x2 - 16y2 + 144z = 0.
Solution Look at Table 10.1 on page 645 and note that the equation is second
degree in x and y but first degree in z. This means it is an elliptic paraboloid or
a hyperbolic paraboloid. Solve the equation for z;
9x2 - 16y2 + 144z = 0
144z = 16y2 —9x2
z 9 16
We recognize this as a hyperbolic paraboloid.
Next, we take cross sections of 9x2 —16y2 + 144z = 0.
Cross Section Chosen Value Equation Description
xj-planc z=0 9 16 υ Two intersecting lines
Parallel to the xj,-plane z=4 36 64 1 Hyperbola

v 2
xz-plane y=0 z 16 Parabola opens down
Parallel to the xz-plane y = 10 100. x2 Parabola opens down

zτ 9 16
yz-plane x=0 zz = y- Parabola opens up
9
z+ 25=^
Parallel to the pz-plane x=5 - 16 9 Parabola opens up
These traces are also shown in Figure 10.16.
Figure 10.16 G raph of 9x 2 - I6y 2 + 144z = 0

PROBLEM SET 1 0 .2
© 7 M Problems 1-4, plot the points P and Q in R 3, and find the
distance ∖PQ∖.
1. 5(3, - 4 , 5), Q(l, 5, -3 ) 2. 5(0, 3, 0), Q (-2, 5, -7 )
3. 5 (- 3 , - 5 , 8), Q(3, 6, -7 ) 4. 5(0, 5, -3 ) ,β (2 , -1 , 0)
In Problems 5-8, fin d the standard form equation o f the sphere
with the given center C and radius r.
5. C(0, 0, 0), r = 1 6. C (-3, 5, 7), r = 2
7. C(0, 4, —5), r = 3 8. C (-2, 3, -1 ), r = √ 5
Find the center and radius o f each sphere whose equation is

9. x 2 +y 2 + z 2 - 2y + 2z —2 = 0
10. x 2 +y 2 + z 2 + 4x —2z —8 = 0
11. x 2 +y 2 + z 2 —6x + 2j' —2z + 10 = 0
12. x 2 +y 2 + z 2 —2x - 4y + 8z + 17 = 0
In Problems 13-22, match the equation with its graph (A-L).
θ The vertices A, B, and C o f a triangle in space are given in

Problems 23-26. Find the lengths o f the sides o f the triangle
and determine whether it is a right triangle, an isosceles
triangle, or neither.
23. A(3, -1 , 0), B(7, 1, 4), C(l, 3, 4)
24. A (l, 1, 1), B(3, 3, 2), C(3, - 3 , 5)
25. A (l, 2, 3), B (-3 , 2, 4), C(l, - 4 , 3)
26. A(2, 4, 3), 5 (-3 , 2, -4 ), C (-6, 8, -1 0 )
27. ■ What Does This Say? Describe a procedure for
sketching a quadric surface.
28. ■ What Does This Say? Describe a procedure for iden
tifying a quadric surface by looking at its equation.
x l y2 -> v -2
2
In Problems 29-40, sketch the graph of each equation in R3. 43. — -y-— z = 1 44. y - F 2 -- z2 = 1
4 9
29. 2x + ι, + 3“ = 6 30. x = 4
31. x + 2y + 5z = 10 32. x + y + z = 1 45.
z = x2 + £ 46. z
z
=^ -
9
yL
16
33. 3x - 2y - z = 12 34. y = z^l √ ,/
35. x = -1 36. y 1 + z 2 = 1 47. x 1 + 2y2 = 9z2 48. z 2 = 1 +
9 4
37. z = ey 38. y = In x 49. Find an equation for a sphere, given that the endpoints of
39. x + z = 1 40. z = y~' a diameter of the sphere are (1, 2, -3 ) and (-2 , 3, 3).
θ 50. Let P(3, 2, -1), Q(-2, 1, c), and R(c, 1,0) be points in R3.
In Problems 41-48, identify the quadric surface and describe For what values of c (if any) is PQR a right triangle?
the traces. Sketch the graph. 51. Find the point P that lies j of the distance from the point
4 (-1 , 3, 9) to the midpoint of the line segment joining
41. 9x2 + 4y2 + z 2 = 1 42. + y2 + y = 1 points B (-2, 3, 7) and C(4, 1, -3).
1 0 .3 THE DOT PRODUCT
IN THIS SECTION Vectors in R3, definition of dot product, angle between

vectors, projections, work as a dot product
In this section, we extend our development of vectors from R2 to R3 by finding
the magnitude of a vector and the distance between points in three dimensions.
We then introduce a product of vectors that yields a scalar and use it to find the
angle between three-dimensional vectors.
■ VECTORS IN T 3
A vector in R 3 may be thought of as a directed line segment (an “arrow”) in space.

The vector P 1P 2 with initial point P 1(x 1, y l , z l ) and terminal point P2(x2, y 2, z 2) has
the component form
P,P 2 = <χ2 - χ t , y 2 - y l , z 2 - z l )
Vector addition and multiplication of a vector by a scalar are defined for vectors
in R3 in essentially the same way as these operations were defined for vectors in
R2 . In addition, the properties of vector algebra listed in Theorem 10.1 of Section
10.1 apply to vectors in R3 as well as to those in R2.
For example, we observed that each vector in R2 can be expressed as a unique
linear combination of the standard basis vectors i and j. This representation can
be extended to vectors in R3 by adding a component k defined to be the unit
vector in the direction of the positive z-axis. In component form, we have in R3,
a. The standard basis

i= < l,0 , 0> j = (0, 1,0) k = (0, 0,1)
vectors in R3
We call these the standard basis vectors in R3. The standard representation of the
7.
vector with initial point at the origin O and terminal point Q(al , a2, a3) is
OQ = a l i + <2-,j + <23k, as shown in Figure 10.17b.
The vector PQ with initial point P(x0 , y 0, z0) and terminal point Q(x1, y l , z j )
has the standard representation
PQ = (χ 1 - ⅞)i + (T1 - F0)j + (z ι ^ ⅞)k
EXAM PLE 1 Standard representation of a vector in R3
Find the standard representation of the vector PQ with initial point

b. The vector from O to Q(a l , a2 , ap P ( - l , 2, 2) and terminal point <2(3, -2 , 4).
is OQ = α 1i + a 2j + a3k.
Solution We have
Figure 10.17 Standard representa
tion of vectors in R3 PQ = [3 - ( - l)]i + [-2 - 2]j + [4 - 2]k = 4i - 4j + 2k h
Sec. 10.3 The Dot Product 649
By referring to Figure 10.17b, we can also derive a formula for the length of a
vector, which in turn can be used to prove the distance formula (stated in the
previous section) between any two points in R 3. Specifically, note that Δ O B Q in
Figure 10.17b is a right triangle with hypotenuse ∣∣
OQ∣ ∣and legs ∣
∣BQ∣ ∣= ∣α3∣ and
∣
∣OB∣∣= yJa 2 + a 2; by applying the Pythagorean theorem, we conclude that the
vector O Q = α l i + α2j + <23k has length
∣
∣
OQ∣
∣= √∣
∣
OB∣
∣
2 + ∣
∣
BQ∣
∣
2 = ⅛ 2 + α
2) + a
2 = y∣
a 2l + a 2 + a 2
Moreover, if A(a 1, a2 , a3) and B(b ∣

, b2 , bi ) are any two points in R 3, the distance
between them is the length of the vector AB. We find
AB = O B - OA = (⅛1 - α 1)i + (⅛2 - α2)j + (b3 - α3)k
so that the distance between A and B is given by
∣
∣
AB∣
∣= √(61 - α1) 2 + (⅛2 - a2 )2 + (b3 - a3)2
Magnitude of a Vector
EXAM PLE 2 Magnitude of a vector
Find the magnitude of the vector v = 2i - 3j + 5k and the distance between

the points A (l, - 1 , -4 ) and B ( -2 , 3, 8).
Solution ∣
∣∣= √22 + (-3 ) 2 + 52 = √38
v∣ and
I R I = √ (-2 - 1)2 + [3 - ( - l ) ] 2 + [8 - (-4 )] 2 = 13
As in R 2, if v is a given nonzero vector in (R3, then a unit vector u that points in

the same direction as v is
u= A
EXAM PLE 3 Unit vector in a given direction
Find a unit vector that points in the direction of the vector PQ from
P ( - l , 2, 5) to β ( 0 ,- 3 , 7).
Solution PQ = [0 - (—l)]i + [-3 - 2]j + [7 - 5]k = i - 5j + 2k
∣
∣
PQ∣
∣= V l 2 + ( -5 ) 2 + 22 = √30
Thus,
PQ _ i - 5j + 2k _ 1 . 5 . 2 ,
U - i∣
PQii - √30 ^ √ 3 0 1 ~ λ ∕3 0 j + √30k
As in R 2, two vectors in R 3 are parallel if they are multiples of one another,

because parallel vectors point either in the same direction or in opposite
directions. That is, nonzero vectors
u and v are parallel if and only if u = sv for some scalar 5
EXAM PLE 4 Parallel vectors

Historical Note
A vector PQ has initial point P(l, 0, -3 ) and length 3. Find Q so that PQ is
parallel to v = 2i - 3j + 6k.
Solution Let Q have coordinates (al , a2, a3). Then,
PQ = [al ~ l]i + [a2 ~ 0]j + [a3 - ( —3)]k = (α 1 - l)i + a2j + (α3 + 3)k
Because PQ is parallel to v, we have PQ = .s-v for some scalar s; that is,
(<21 - l)i + <s2j + (α3 + 3)k = s(2i - 3j + 6k)
This implies
α 1 - 1 = 2s a2 = -3 s α3 + 3 = 6s
a } = 2s + 1 a3 = 6s - 3
Because PQ has length 3 we have
CHARLOTTE ANGAS SCOTT 3 = TJ(Ω1 —1)2 + u2 2 + (α3 + 3)2

(1858-1931)
Master teachers can be judged = √[(2s + 1 ) - 1]2 + (-3 s) 2 + [(65 - 3 ) + 3]2
not only by their own accom = V4.v2 + 9v2 + 36s2 = V49s2 = 7∣
s∣
plishments, but by the accom
plishments of their students. Thus, 5 = ±y and so
Charlotte Angas Scott set the
direction of the high school flι = 2 (4 ) + 1 fl 2 = ~ 3 (i 7) α 3= 6 (± 7 )-3
mathematics curriculum in the
United States during the early - 13 i _ _9 9 _ _ 3 _39
part of the twentieth century. “ 7’ 7 “ 7’ 7 “ 7’ 7
She also headed the mathemat
ics department at Bryn Mawr
There are two points that satisfy the conditions for the required terminal point
College for 40 years. Some not a (V, 4 4 ) and (i, 2 _
able students of Charlotte Scott
include Virginia Ragsdale
(1870-1945), Louise Cum ■ DEFINITION OF DOT PRODUCT
mings (1870-1947), and Bird
Turner (1877-1926). It has The dot (scalar) product and the cross (vector) product are two important vector
been reported that Scott was in operations. We shall examine the cross product in the next section. The dot
fluenced to switch from lan product is also known as a scalar product because it is a product of vectors that
guage to mathematics by Emmy gives a scalar (that is, real number) as a result. Sometimes the dot product is called
Noether, the greatest woman
the inner product.
mathematician of her time (see
Historical Note, p. 442). Scott
was recognized as a great math Dot Product
ematician by her peers, but to
her death she believed her The dot product of vectors v = α 1i + α2j + α3k and w = ⅛l i + ⅛2j + Z?3k is the
greatest accomplishments rested scalar denoted by v ∙ w and given by
in attracting competent women
v ∙ w = α 1⅛1 + a2b1 + a3b3
to mathematics.
The dot product of two vectors v = α l i + α2j and w = ⅛l i + b2i in a plane is given
by a similar formula with a3 = b3 = 0: v ∙ w = <21Z>1 + a,b,.
EXAM PLE 5 Dot product
Find the dot product of v = —3i + 2j + k and w = 4i - j + 2k.

Solution v ∙w = -3 (4 ) + 2 ( - 1 ) + 1(2) = - 1 2
EXAM PLE 6 Dot product in component form
If v = (4, - 1, 3) and w = ( - 1, -2 , 5), find the dot product, v ∙ w.

Solution v ∙ w = 4(-1) + (-1 )(-2 ) + 3(5) = 13 ∣
=>
Before we can apply the dot product to geometric and physical problems, we
need to know how it behaves algebraically. A number of important general
properties of the dot product are listed in the following theorem.
THEOREM 10.2 Properties of the dot product
If u, v, and w are vectors in [R2 or R 3 and c is a scalar, then:
Magnitude of a vector v∙v= ∣ ∣ ∣

2
v∣
Zero product 0 ∙v = 0
Commutativity v∙w= w∙V
Scalar multiple c(v ∙ w) = (cv) ∙ w = u ∙ (cw)
Distributivity u ∙( v + w) = u ∙v + u ∙w
Proof: Let u = a 1i + α2j + α3k, v = ⅛1i + ⅛2j + Z>3k, and w = c 1i + c2j + c3k.
Magnitude of a vector
∣
∣
v∣∣2 = (v7z2 + a 2 + a∣
λ y = a 2 + α∣+ α2 = v ∙ v
Zero product, commutativity, and scalar multiple can be established in a similar
fashion.
Distributivity
u ∙ (v + w) = (α1i + α2j + α3k) ∙ [(Z>1 + c 1)i + (⅛2 + c2 )j + (b3 + c3 )k]
= rz1(b1 + c 1) + α2(⅛2 + c2) + a3(b3 + c3)
= a i b l + α 1c 1 + a2b2 + a2c2 + a3b3 + a3c3
Also,
u ∙ v + u ∙ w = (a 1⅛1 + a2b2 + a3b3) + (<21c1 + a2c2 + α3c3)
= af>l + a l cl + a2b2 + a2c2 + a3b3 + a3c3
Thus, u ∙ (v + w) = u ∙ v + u ■w. ■
■ ANGLE BETWEEN VECTORS
The angle between two nonzero vectors v and w is defined to be the angle θ with
0 ≤ θ ≤ 77that is formed when the vectors are in standard position (initial points
at the origin), as shown in Figure 10.18.
The angle between vectors plays an important role in certain applications and
may be computed by using the following formula involving the dot product.
THEOREM 10.3 Angle between two vectors
If θ is the angle between the nonzero vectors v and w, then

Figure 10.18 The angle between m s A = V ' W
two vectors ∣
∣
v∣∣
∣∣
w∣∣
Proof: Suppose v = α 1i + α9j + α3k and w = ⅛1i + Z>2j + ⅛3k, and consider
∕∖A O B with vertices at the origin O and the points √4(αl , α,, α3) and B(b l , b ,, b3 ), as
shown in Figure 10.19. Note that sides OA and OB have lengths
∣
∣
v∣
∣= λ∕<a2 + a∖ + a i3 and ∣
∣
w∣= ^b∖ + b f + bj
respectively, and that side A B has length
O v A (α 1, α 2 , α 3 )
∣
∣
v - w∣
∣= √(α1 - ⅛1)2 + (α2 - b2)2 + (α3 - b3)2
Figure 10.19 Finding an angle
between two vectors Next, we use the law of cosines.
a 2 = b 2 + c 2 - 2bc cos θ Law of cosines
∣
∣
v - w∣
∣
2 = ∣
∣
v∣
∣2 + ∣
∣
w∣∣
2 - 2∣
∣
v∣
∣∣
∣
w∣∣
cos θ See Figure 10.19.
I∣v ∣∣2 + ∣∣w ∣∣2 _ ∣∣v _ w ∣∣2
cos θ = Solve for cos θ.
2∣
∣
v∣
∣∣
∣
w∣∣
a 2 + α2 + rz2 + Zt2 + ⅛2 + ⅛2 - [(α1 - ⅛1)2 + (a, - b-f 2 + (α3 - ⅛3)2]
S M
2tzj ⅛1 ÷ 2a,fb~l + 2tz3⅛3 3, . w
2∣∣v∣∣ IHI - ∣∣v∣∣ ∣∣w∣∣
EXAM PLE 7 Angle between two given vectors
Let Δ A B C be the triangle with vertices A ( l, 1, 8), B(4, - 3 , -4 ) , and

C ( -3 , 1, 5). Find the angle formed at A.
Solution Draw ∕∖A B C and label the angle formed at A as a as shown in Figure
10.20. The angle a is the angle between vectors AB and A C , where
AB = (4 - l)i + (-3 - l)j + ( - 4 - 8)k = 3i - 4j - 12k
A C = (-3 - l)i + ( 1 - l)j + ( 5 - 8)k = - 4 i - 3k
Thus,
_ . AB ∙ A C 3(-4) + (-4)(0) + (-1 2 )(-3 )
Figure 10.20 Find an angle of a
∣∣
AB∣
∣ ∣∣
AC∣ ∣ √ 3 2 + ( -4 ) 2 + (-1 2 ) 2 V ( - 4 ) 2 + ( -3 ) 2
triangle 24 _ 24
√169 √25 65
and the required angle is a = cos - 1 ® ) ≈ 1.19. l
The formula for the angle between vectors is often used in conjunction with
the dot product. If we multiply both sides of the formula by ∣
∣
v∣
∣∣∣
w∣∣
, we obtain the
following alternative form for the dot product formula.
Dot Product (Alternative Form) v∙w= ∣

∣
v∣
∣∣
∣
w∣∣
cos θ
where θ is the angle (0 ≤ θ ≤ π) between any two vectors v and w.
Two vectors are said to be perpendicular, or orthogonal, if the angle between

them is θ = -ττ∕2. The following theorem provides a useful criterion for orthogonal
ity-
Sec. 10.3 The Dot Product 6 53
THEOREM 10.4 Orthogonal vectors
Nonzero vectors v and w are orthogonal if and only if

v∙w= 0
Proof: If the vectors are orthogonal, then the angle between them is J;
therefore,
v∙w= ∣ ∣
v∣∣∣
∣
w∣ ∣cos J = 0
Conversely, if v ∙ w = 0, and v and w are nonzero vectors, then cos θ = 0, so that

θ = f and the vectors are orthogonal. b
Determine which (if any) of the following vectors are orthogonal:

u = 3i + 7j - 2k v = 5i —3j —3k w=j- k
u ∙ v = 3(5) + 7(-3) + (-2 ) (-3 ) = 0; orthogonal vectors.
u ∙ w = 3(0) + 7(1) + (-2 )(-1 ) = 9; not orthogonal vectors.
v ∙ w = 5(0) + (-3)(1) + (-3 )(-1 ) = 0; orthogonal vectors.
a PROJECTIONS
Let v and w be two vectors in IR2 drawn so that they have a common initial point,
as shown in Figure 10.21.* If we drop a perpendicular from the tip of v to the line
determined by w, we determine a vector called the vector projection of v onto w,
which we have labeled u in Figure 10.21.
Figure 10.21 Projection of v The scalar projection of v onto w (also called the component of v along w) is the
onto w length of the vector projection, so it is denoted by ∣ ∣
u∣∣
. Let θ be the acute angle
between v and w. Then, by definition of cosine we have
0 Note that ( is not the
∣= ∣
11u∣ ∣∣cos θ = ∣
v∣ ∣∣
v∣∣∣
v∣
∣ ∣
∣
w ∣
MlW
,∙∣= ⅞wγ Cosine is positive because θ is acute.
same as f^)w, and you cannot
“ cancel” the vector w. Remem If θ is obtuse, its cosine is negative and ∣∣ ∣= - ∣∣v∣∣cos θ.
u∣
ber, v ∙ w and w ∙ w are numbers, In order to find a formula for the vector projection, we note that it is the scalar
whereas v/w is not defined. 0 component of v in the direction of w. If θ is acute, then the vector projection has
length ∣∣ ∣
v∣ ∣
cos θ and has direction w ∕∣w∣∣
(the unit vector of w). If the angle is obtuse,
the vector projection has length —∣∣v∣∣cos θ and has direction - w ∕∣∣w∣∣
. In either case,
u = (∣∣v∣∣ c o s 0 ⅛ = Y J w = L∆W
lw
j llwl∣2 w
Scalar and Vector Projections

Scalar projection of v onto w (a number):
of v onto w
Vector projection of v in the direction of w (a vector): (^,'∙ w)w
In R 3 it is important to note that the i, j, and k components of any vector v are the
projections of v onto the basis vector, as shown in Figure 10.22.
Find the scalar and vector projections of v = 2i - 3j + 5k onto

Figure 10.22 The i, j, and k com w = 2i - 2j + k
ponents of v = ai + bj + ck are
projections of v onto i, j, and k. *Even though Figure 10.21 is drawn in R 2, the projection formula applies to R 3 as well.
Solution We first find the vector projection:

Vector projection of v onto w = ( ~ ~ ) w
p(2) + (-3 )(-2 ) + 5(1)∖

(2i - 2j + k)
∖ 22 + (-2 ) 2 + l 2 /
= ^ (2 i - 2j + k) = ⅛i - + jk
To find the scalar projection, we can find the length of the vector projection or
we can use the scalar projection formula (which is usually easier than finding
the length directly):
Scalar projection of v onto w = ∣n∣-7r∣ = ρ r = 5
■ WORK AS A DOT PRODUCT
One important application of the dot product and projections occurs in physics
when calculating the amount of work done by a constant force. In Section 6.3, we
defined the work, W, done by a constant force directed along the line of the
motion of the object to be W = F d (Figure 10.23a). We now consider a force acting
a. If the force F acts along
the line of motion then in some other direction (Figure 10.23b).
W=IIFII IIPQII. Physics experiments indicate that when a force F moves an object along the
line from point P to point Q , the work performed is given by the product
W = [SCALAR COMPONENT OF F ALONG PQ][DISTANCE MOVED BY THE OBJECT]
The distance moved by the object is the length of the displacement vector PQ, and
we find that
Scalar component of F along PQ
b. If F acts at a nonzero angle
with the line of motion, then
W=llproj. ofFontoPQ II IIPQII.
HZ = W ⅛ = F ' PQ
Figure 10.23 Work Wz as a dot Distance moved by the object
product
Work As a Dot Product An object that moves along a line with displacement PQ against a constant
force F performs
W = F ∙ PQ
units of work.
E X A M P L E 10 Work performed by a constant force
Suppose that the wind is blowing with a force F with magnitude 500 lb in the
N direction of N30 oE over a boat’s sail. How much work does the wind perform
in moving the boat in a northerly direction a distance of 100 ft? Give your
answer in foot-pounds.
Solution We see that ∣
∣F∣
∣ = 500 lb and is in the direction of N30°E, as shown in
Figure 10.24. The displacement direction is PQ = 100j, so ∣ ∣
PQ∣ ∣= 100 ft.
Thus, F = 500cos60°i + 500sin60oj = 250i + 250Vz3j. Thus, the work per
formed is
Side View Top View W = F ∙ PQ = 100(250√3) = 25,000V3

Figure 10.24 Work performed Thus, the work is approximately 43,300 ft-lb.
PROBLEM SET 1 0 .3
0 ι. ■ What Does This Say? Discuss how to find a dot 41. v = 2i + j + 2k; w = - i + 2j - k
product, and describe an application of dot product. 42. v = 2j + 3k; w = i + j + k
2. ■ What Does This Say? Describe and contrast scalar 43. Find a vector that points in the direction opposite to
and vector projections. v = 2i + 3j - 2k.
Find the standard representation o f the vector PQ, and then 44. Find a vector that points in the same direction as
fin d ∣
∣
PQ ∣ ∣in Problems 3-6. v = i + 2j - k and has one-third its length.
3. P (l, - 1 , 3), β ( - l , 1, 4) 4. P(0, 2, 3), 2(2, 3, 0) 45. Find x , y, and z that solve
5. P ( l, 1, 1), Q (-3 , - 3 , -3 ) 6. P(3, 0, -4 ), 2(0, - 4 , 3) x(i + j + k) + y (i - j + 2k) + z(i + k) = 2i + k
Find the dot product v ∙ w in Problems 7-10. 46. Find x , y, and z that solve
7. v= (3, - 2 , 4); w = (2, - 1 , -6 ) x(i - k) + y(j + k) + z(i - j) = 5i - k
8. v= (2, - 6 , 0); w = (0, - 3 , 7) 47. Find a number a that guarantees that the vectors
3i - 2j + k and 2i + aj —2ak will be orthogonal.
9. v= 2i + 3j —k; w = - 3 i + 5j + 4k
48. Find x if the vectors v = 3i —xj + 2k and
10. v= 3i - j; w = 2i + 5j
w = xi + j —2k are to be orthogonal.
State whether the given pairs o f vectors in Problems 11-14 are 49. Find the angles between the vector 2i + j - k and each of
orthogonal. the coordinate axes. The cosines of these angles (to the
11. v = i; w = k 12. v = j; w = —k nearest degree) are known as the direction cosines.
13. v = 3i - 2j; w = 6i + 9j 50. Find the cosine o f the angle between the vectors
14. v = 4i - 5j + k; w = 8i + 10j - 2k v = i - j + 2k and w = 2i + j - k. Then find the vector
projection o f v onto w.
Evaluate the expressions given in Problems 15-18.
15. ∣∣
i + j + k∣∣ 16. ∣
i - j + k∣
∣ ∣ 51. Let v = 4i - j + k and w = 2i + 3j — k. Find:
a. v ∙ w
2i + j - 3k ∣
17. ∣
∣ ∣
2 2(i - j + k) - 3(2i + j- k)∣
18. ∣
∣ ∣
2
b. cos θ, where θ is the angle between v and w
Let v = i —2j + 2k and w = 2i + 4j —k; and fin d thevector or
c. a scalar 5 such that v is orthogonal to v - sw
scalar requested in Problems 19-22.
19. 2v - 3w v∣
20. ∣
∣ ∣
w d. a scalar s such that .w + w is orthogonal to w
21. ∣2v - 3w∣
∣ ∣ 22. ∣
∣
v - w∣∣
(v + w) 52. Let v = 2i - 3j + 6k and w = 4i + 3k. Find:
a. v ∙ w
Determine whether each vector in Problems 23-26 is parallel to
b. cos θ, where θ is the angle between v and w
2i - 3j + 5k.
23. u = (4, 0, 10) 24. v = ( -4 , 6 ,-1 0 ) c. a scalar s such that v is orthogonal to v - sw
25. w = 0 , —1, 2^ 26. v = ∣ , 2^ d. a scalar t such that v + tw is orthogonal to w
53. Find the scalar component of the force F = 4i - 2j + 3k
Let v = 3i - 2j + k and w = i + j - k. Evaluate the expres in the direction of the vector v = i - j + 2k.
sions in Problems 27-30.
54. Find the work done by the constant force F = 2i + 3j + k
27. (v + w) ∙ (v - w) 28. (v ∙ w)w
when it moves a particle along the line from P( 1, 0, - 1 ) to
29. (∣
∣∣
v∣ w∣
w) ∙ ( ∣
∣ ∣
v) 30. g⅛ 2 (3 ,1 ,2 ).
55. Find the work performed when a force F = î - γj + ®k is
Find the angle between the vectors given in Problems 31-34. applied to an object moving along the line from
Round to the nearest degree. P ( -3 , - 5 , 4) to 2(4, 9, 11).
31. v = i + j + k;w = i —j + k 56. Fred and his son Sam are pulling a heavy log along flat
32. v = 2i + k; w = j —3k horizontal ground by ropes attached to the front of the
33. v = 2j + k; w = i - 2k log. The ropes are 8 ft long. Fred holds his rope 2 ft above
34. v = 4i - j + k; w = 2i + 3j + 5k the log and 1 ft to the side, and Sam holds his end 1 ft
above the log and 1 ft to the opposite side, as shown in
Find the vector and scalar projections o f v onto w in Problems Figure 10.25.
35-38.
35. v = i + j + k; w = 2k 36. v = i + 2k; w = - 3 j
37. v = 2i - 3j; w = 2j —3k 38. v = i + j - 2k; w = i + j + k
0 Find two distinct unit vectors that are orthogonal to each pair o f
Side View Top View
vectors given in Problems 39-42. We will do this type o f
problem with cross product in the next section. F ig u re 10.25 Problem 56
39. v = i + k; w = i - 2k If Fred exerts a force o f 20 lb and Sam exerts a force of 30
40. v = i + j - k;w = —i + j + k lb, what is the resultant force on the log?
57. Find the force required to keep a 5,000-lb van from

61. In about 1807 the French mathematician J. B. J.
rolling downhill if it is parked on a 10o slope.
Fourier made the remarkable observation that most
important odd functions [that i s ,∕( - x ) = -f( x ) ] can
be approximated by
f(x) ≈ b l sin x + b2 sin l x + ∙∙∙ + bn sin nx = S n (x)
This formula reminds us o f the Taylor approxima
tions; in fact, some call S n(x) a trigonometric polyno
mial. However, this expansion is physically more
meaningful, since it breaks f(x) down into frequency
components, where these components can represent
things like pitch in sound or colors or chemical
58. A block of ice is dragged 20 ft across a floor, using a force elements.
o f 50 lb. Find the work done if the direction o f the force is
a. Assume for now that f(x) is actually equal to S n (x).
inclined θ to the horizontal, where
To find the coefficients bk , for k = 1, 2 , . . . , n,
multiply the above trigonometric polynomial by
sin k x and then integrate over [-ττ, τr]. Use the
orthogonality o f the sines and the trigonometric
identity 2 sin 2α = 1 - cos l a to show that
rπ
sin 2 kx dx = τr
J~π
I fπ
bk = — I ∕( x ) sin kx dx
i Γ
= — f(x) sin kx dx
π Jo
59. Suppose that the wind is blowing with a 1,000-lb magni
tude force F in the direction of N60oW over a boat’s sail. b. Why, in the evaluation of bk , is it only necessary to
How much work does the wind perform in moving the integrate over [0, TT]?
boat in a northerly direction a distance o f 50 ft? Give your c. Now let us do a sine (Fourier) expansion of the
answer in foot-pounds. simple function f(x) = x . Using the work from
part a, show that
( - 1)"
Computational Window x ≈ 1 sin x - sin l x + —2 ------- sin nx
n
d. For f(x) = x(ττ - x)(π + x) compute the sine ex
Orthogonality o f Functions. One o f the big discov pansion of the above trigonometric polynomial.
eries in modern mathematics was that it is very fruitful to Starting with a small n, (that is, w or 3) increase n
extend some o f the geometric notions o f this chapter to until the sine approximation agrees with ∕( x ) at
functions. For example, taking the dot product o f two the graphical level. If possible, turn in a meaning
functions or asking i f two functions are orthogonal, or ful graph. Then increase n until the difference
attempting to project a function onto a second function (or ∣∕(x ) - (Z>1 sin x + b2 sin l x + + bn sin nx)∣ < 0.02
onto a "plane” o f functions). You are introduced to this
Note: If the function is not odd, then similar
important concepts in Problems 60 and 61.
cosine terms are needed; we have ignored this
60. We say that two functions f and g are orthogonal on
possibility for simplicity.
[a, ⅛] if p
f(x)g(x) d x = 0
Jb θ 62. Let u0 = ai + by + ck and let u = x i + yy + zk. Describe
a. Show that the two functions x 2 and x 3 - 5x are the set of points in R 3 defined by
orthogonal on [-⅛, Z>] for any positive b.
b. For positive integers k and n Ψ k, show that sin kx l∣χ0 - u∣l < r
and sin nx are orthogonal on the interval [—π, ττ]. where r > 0 and a, b, c are constants.
You may need the product-to-sum identity. (Iden 63. Suppose the vectors v and w are sides of an equilateral
tity 38 from the Student Mathematics Handbook): triangle with area 25V3. Find v ∙ w.
64. Find the angle (to the nearest hundredth radian) between
2 sin a sin β = cos(α — β) — cos(α + β)
the diagonal of a cube and a diagonal of one of its faces.
That is, the fam ily of functions sin x , sin l x , 65. If v ∙ v = 0 and v ∙ w = 0, what can you conclude about v
sin 3 x ,. . . are mutually orthogonal on [ - π , ττ]. and w?
c. Make a careful sketch of sin x and sin l x on 66. a. Show that (v + w) ∙ (v + w) = ∣ ∣
v∣∣ ∣
2 + ∣ ∣
w∣2 + 2(v ∙ w).
[-τr, τr] and explain why the integral of their
b. Use part a to prove the triangle inequality:
product is 0.
v + w∣
I∣ ∣≤ ∣∣
v∣∣+ ∣∣
w∣∣
Sec. 10.4 The Cross Product 657
Hinf. Note that a. Prove the Cauchy-Schwarz inequality. Hint: Use the
formula for the angle between vectors.
∣
∣
v + w∣
∣
2 = (v + w) ∙ (v + w)
b. Show that equality in the Cauchy-Schwarz inequality
67. The Cauchy-Schwarz inequality in R 3 states that for any occurs if and only if v = Λv for some scalar t.
vectors v and w c. Use the Cauchy-Schwarz inequality to prove the trian
gle inequality:
∣
v w ∣≤ ∣
∣
v∣
∣∣
∣
w∣∣
I∣
v + w∣
∣≤ ∣
∣
v∣
∣+ ∣
∣
w∣∣
10.4 T H E C R O SS P R O D U C T
IN T H IS S E C T IO N Definition of cross product, geometric interpretation of

cross product, properties of cross product, triple scalar product and volume,
torque
We introduce a second type of vector multiplication called cross product, which
turns out to be a vector. We shall find that the cross product is a vector
orthogonal to each of the given vectors. This geometric property will be
particularly useful in the next section.
■ D E F IN IT IO N O F C R O S S P R O D U C T
The cross product is sometimes called the vector product because the result is a
vector. In other applications it is called the outer product. The definition requires a
basis of i, j, and k and therefore, is a definition that makes sense only in R3.
Cross Product If v = α 1i + α2j + α 3k and w = 6 l i + ⅛2j + b,jk, the cross product, written v × w, I
is the vector
v × w = (α2 ⅛3 - α3 ⅛2 )i + (α 3⅛1 - α 1⅛3)j + (fl1⅛2 - α2 ⅛1)k
These terms can be obtained by using a determinant

i j k
v× w= α2 α3
⅛2 b3
A brief discussion of the basic properties of determinants is included in Student

Mathematics Handbook. To verify the determinant formula for the cross product,
we will expand this determinant about the first row:
j k
fl 3 - ∣β l fl 3 «1 a.'2 k
v × w= ι a.'2 fl3 i
α
b ⅛2 653 I 1
b. b3 b x b., 2
bi b,2. 3 A
j is in row 1, column 2, so do not forget negative sign here.

= (' a2 ,b3 , - a,3b,2')' ι - (' a1 ,b3 , - a,3b,1)' ij + (' a1 .b2 , - a21b,1)' k
= (a-,b3 - α3 A,)i + (a3b l - α 1⅛3 )j + (α 1⅛7 - a 2 ⅛1)k
EXAMPLE 1 Cross product

Find v × w, where v = 2i - j + 3k and w = 7j - 4k
Solution ι j k
v× w= 2 -1 3
. Do not forget minus here (j is negative position).
0 7
I
= [(-1)(-4) - 3(7)]i - [2(-4) - 0(3)]j + [2(7) - 0 (- 1)]k
= -17i + 8j + 14k
Properties of determinants can also be used to establish properties of the cross

product. For instance, the following computation shows that the cross product is
not commutative. This property is sometimes called anticommutativity.
i j k j k
v × w= α ι ⅛2 ⅛3 = -(w × v)
z,ι ⅛2 b3 αι α2 f l 3
THEOREM 10.5 Properties of cross product

If u, v, and w are vectors in (R3 and 5 and t are scalars, then
Scalar distributivity (sv) × (tv∕) = 5t(v × w)

Vector distributivity* u × (v + w) = (u × v) + (u × w)
(u + v) × w = (u × w) + (v × w)
Anticommutativity v × w = -(w × v)
Parallel vectors v × v = 0; in particular, if w = sv, then v × w = 0
Zero product v×0=0×v=0
Proof: Let u = <ηi + α9j + α3k, v = ⅛1i + ⅛2j ÷ ⅞k, and w = c 1i + c2j + c,3k.
Scalar distributivity and vector distributivity are proved by using the definition
of cross product and the corresponding properties of real numbers.
Zero factors:
j k
v × v= b-, ⅛3 = 0 Property o f determinants (two rows the same)
⅛2 b 3
v × w = v× cv = c (v × v )= 0
Zero product is obvious.
’Properly, it is the distributive property of vectors for cross product over addition, which, for
convenience, we shorten to vector distributivity.
V× w ■ GEOMETRIC INTERPRETATION OF CROSS PRODUCT
We will show with the following theorem that the vector (v × w) is orthogonal to
both the vectors v and w. The only way this can occur is in a three-dimensional
setting. Any two distinct nonzero vectors in R3 that are not parallel (that is, not
scalar multiples of one another) can be arranged to determine a plane. Then,
Figure 10.26 Vector product of according to the geometric property of cross product (Theorem 10.6) the vector
two vectors product must be orthogonal to this plane, as shown in Figure 10.26.
THEOREM 10 .6 Geometric property of cross product
If v and w are nonzero vectors in R3 that are not multiples of one another, then
v × w is orthogonal to both v and w.
Proof: We will show (v × w) is orthogonal to v and leave the proof that v × w
is orthogonal to w as an exercise. Let v = α 1i + a 2j + <23 k and w = 6μ + Zy + 63k.
Then,
• j k
v× w= Ω 1 tZ2 Cl3 = (a2 b3 - a3b2)i - (α1Z>3 - tz3⅛1)j + (al b2 — α2h l )k
bχ b2 b3
To show this vector is orthogonal to v, we find v ∙ (v × w):
v ∙ (v × w) = α 1(α2⅛3 - α3⅛2) - α2(α1⅛3 - α3⅛1) + α3(<2167 - α7⅛1)
= aâ2b3 - a χa3b2 - a3 a2b3 + a2a3b i + a χa3b2 - aâ3bi
=0
EXAM PLE 2 A vector orthogonal to two given vectors
Find a nonzero vector that is orthogonal to both v = - 2 i + 3j - 7k and

w = 5i + 9k.
Solution The cross product v × w is orthogonal to both v and w.
i j k
v× w= - 2 3 - 7 = (27 + 0 ) i - ( - 1 8 + 35)j + ( 0 - 15)k
5 0 9
= 27i - 17j - 15k
Because both v × w and w × v are orthogonal to the plane determined by v and
w, and because (v × w) = -(w × v), we see that one points up from the given plane
and the other points down. In order to see which is which, we state the right-hand
rule, which is described in Figure 10.27.
Figure 10.27 The right-hand rule: If you place the

palm of your right hand along v with your fingers
pointing in the direction of v, and you curl your
fingers toward w, then your thumb points in the
direction of v × w.
EXAM PLE 3 Right-hand rule
Use the right-hand rule to verify each of the following cross products.
i×j = k j × i = -k i× i = 0 i× k = -j k×i =j
j× j = θ j× k = i k× j = -i k×k = 0
6 60 Ch. 10 Vectors in the Plane and in Space
Solution
i ×j Place the palm of your right hand

along i, pointing in the direction of
i, and curl your fingers toward j.
The answer is k.
j ×i Place the palm of your right hand

along j, pointing in the direction of
j, and curl your fingers toward i.
The answer is —k.
-k
j k
0 0=0
0 0
The other parts are left for you to verify.
■ PROPERTIES OF CROSS PRODUCT
Next, we will find the magnitude of the cross product of two vectors.
THEOREM 10.7 Magnitude of a cross product

If v and w are nonzero vectors in IR3 with θ the angle between v and w (0 ≤ θ ≤ -ττ),
then
∣∣
v × w∣ ∣=∣ ∣
v∣∣ ∣
∣
w∣∣sin θ
Proof: Let v = α1i + α2j + α3k and w = ⅛1i + Z>2j + ⅛3k.
∣
∣∣
v∣∣
∣
w∣∣
sin θ = ∣
∣∣
v∣∣
∣
w∣∣
V l - cos2 θ
=H H √1-[M⅞]2
= ∖∣
(al 2 + a2 2 + α3 2)(⅛12 + ⅛2 2 + b3 2) - (ai bi + a2b2 + a3b3)2
= ∖(a2b3 - ai b2)2 + (a 3Z>1 - a l b3)2 + (a1b2 - a2b l )2

= Ik × w∣∣
There is another interpretation for Theorem 10.7—namely, 11v × w∣∣is equal to

the area of a parallelogram having v and w as adjacent sides, as shown in Figure
10.28. To see this, note that because h = ∣∣
w∣∣sin θ, we have
AREA = (BASE)(HEIGHT) = ∣
∣∣
v∣ ∣
(∣
w∣∣ ∣
sin θ) = ∣∣V × w∣
Figure 10.28 A geometric inter Here is an example that uses this formula. You might also note that even though
pretation of ∣
∣
v × w∣ v × w ≠ w × v, it is true that ∣
∣ ∣
v × w∣=∣∣
w × v∣∣
Find the area of the triangle with vertices P ( -2 , 4, 5), Q(0, 7, -4 ) , and
5, 0).
Draw this triangle as shown in Figure 10.29. Then Δ P Q R has half the
area of the parallelogram determined by the vectors P Q and PR ; that is, the
triangle has area
A = 1∣
∣
PQ × PR∣ ∣
First find
P Q = (0 + 2)i + ( 7 - 4)j + ( - 4 - 5)k = 2i + 3j - 9k
P R = ( - 1 + 2)i + ( 5 - 4)j + (0 - 5)k = i + j - 5k

Figure 10.29 Area of a triangle by and compute the cross product:
cross product A = ^ ∣
∣
v × w∣
i j k
PQ × PR = 2 3 -9
1 1 -5
= ( - 15 + 9)i - ( - 10 + 9)j + ( 2 - 3)k
= -6 i + j - k
Thus, the triangle has area
A = ⅛∣
∣
PQ × PR ∣
∣
= ∣ V ( - 6 ) 2 + 12 + ( - l ) 2
■ TRIPLE SCALAR PRODUCT; VOLUME
The cross product can also be used to compute the volume of a parallelepiped in
R 3. Consider the parallelepiped determined by three nonzero vectors u, v, and w
that do not all lie in the same plane, as shown in Figure 10.30.
u×v
a. The parallelepiped determined b. The parallelepiped has volume

by u, v, and w V = Ah = ∣(u × v) ∙ w ∣
.
Figure 10.30 Computing volume with the triple scalar product
It is known from solid geometry that this parallelogram has volume V = Ah,
where A is the area of the face determined by u and v and h is the altitude from the
tip of w to this face (see Figure 10.30b).
The face determined by u and v is a parallelogram with area A = ∣ ∣ιι × v ∣
∣
, and
we know that the cross-product vector u × v is perpendicular to both u and v and
hence to the face determined by u and v. From the alternative form of the dot
product, we have
(u × v) ∙ w = ∣
∣
u × v∣
∣∣
∣w∣
∣cos θ
where θ is the angle between u × v and w. Thus, the parallelepiped has altitude
(u × v) ∙ w
h= ∣
∣
∣ ∣cos0∣ =
w∣
∣
11« × v∣
and the volume is given as
(u × v) ∙ w
∣
V = Ah = ∣∣u × v∣ = ∣
(u × v) ∙ w∣
l∣
u × v∣∣
The combined operation (u × v) ∙ w is called the triple scalar product of u, v, and w.
These observations are summarized in the following box.
Volume Interpretation Let u, v, and w be nonzero vectors that do not all lie in the same plane. Then the
of the Triple Scalar parallelepiped determined by these vectors has volume
Product
V= ∣
(u × v) ∙ w∣
EXAM PLE 5 Volume of a parallelepiped
Find the volume of the parallelepiped determined by the vectors

u = i - 2j + 3k v = - 4 i + 7j - 1 Ik w = 5i + 9j - k
Solution We first find the cross product.
i j k
U × V= 1 -2 3 = (22 -2 1 )i - ( - 1 1 + 12)j + ( 7 - 8)k
-4 7 -11
= i- j - k
Thus,
V = ∣(u × v) ∙ w∣ = ∣(i - j - k) ∙ (5i + 9j - k)∣ = ∣5 - 9 + 1∣ = 3
The volume of the parallelepiped is 3 cubic units.
THEOREM 10.8 Determinant form for a triple scalar product: volume of a

parallelepiped
If u = α 1i + rz2j + a3k, v = ⅛1i + ⅛2j + ⅛3k, and w = c 1i + c2j + c3k, then the triple
scalar product can be found by evaluating the determinant
al Cl
b 2 b3
(u × v) ∙ w =
bι
cι ci c3
Proof: The proof follows by expanding the determinant (see Problem 49). ■
We can use Theorem 10.8 to rework Example 5.

1 -2 3
(u × v) ∙ w = -4 7 -11 = -3
5 9 -1
Thus, the volume is ∣-3∣ = 3.
In the problem set we have included several exercises involving the triple
scalar product and the triple vector product u × v × w. In case you wonder why we
neglect the product (u ∙ v) × w, notice that such a product makes no sense, because
u ∙ v is a scalar and the cross product is an operation involving only vectors. Thus,
the product u ∙ v × w must mean u ∙ (v × w).
Sec. 10.4 The Cross Product 6 63
■ TORQUE
A useful physical application of the cross product involves torque. Suppose the
T = PQ × F force F is applied to the point Q. Then the torque of F around P is defined as the
cross product of the “ arm” vector PQ with the force F, as shown in Figure 10.31.
Thus, the torque, T, of F at Q about P is
T = PQ × F
The magnitude of the torque, ∣ ∣
T∣∣, provides a measure of the tendency of the
Figure 10.31 Torque given by a
vector arm P Q to rotate counterclockwise about an axis perpendicular to the plane
cross product determined by PQ and F (see Figure 10.31).
EXAM PLE 6 Torque on the hinge of a door
Figure 10.32 shows a half-open door that is 3 ft wide. A horizontal force of 30

lb is applied at the edge of the door. Find the torque of the force about the
hinge on the door.
Solution We represent the force by F = -3 0 i (see Figure 10.32). Because the
door is half open, it makes an angle of f with the horizontal, and we can
represent the “ arm” P Q by the vector
PQ = 3(cosf i + s in f j) = 3 θ ^ j + j) = 3√2 ∣ + 3yT j∙
The torque can now be found:
• J k
Figure 10.32 The torque of a 3√2 3√2
T = PQ × F = 2 2 0 =45√2k
force applied at the edge of a door
about the hinges -3 0 0 0
The magnitude of the torque (45λ∕2 ft-lb) is a measure of the tendency of the
door to rotate about its hinges. M M
PROBLEM SET 1 0 .4
θ Find v × w for the vectors given in Problems 1-10. 19. v = i + j + k; w = 3i + 12j - 4k
1. v = i; w = j 2. v = k; w = k 20. v = 2i —2j + k; w = 4i + 2j —3k
3. v = 3i + 2k; w = 2i + j 4. v = i —3j; w = i + 5k
Find the area o f the parallelogram determined by the vectors in
5. v = 3i - 2j + 4k; w = i + 4j - 7k Problems 21-24.
6. v = 5i —j + 2k; w = 2i + j —3k 21. 3i + 4j and i + j - k
7. v = 3i —j + 2k; w = 2i + 3j —4k 22. 2i - j + 2k and 4i - 3j
8. v = - j + 4k; w = 5i + 6k 23. 4i - j + k and 2i + 3j - k
9. v = i —6j + 10k; w = —i + 5j —6k 24. 2i + 3k and 2j - 3k
10. v = cos 0 i + sin θy, w = -s in θ i + cos θ j
Find the area o f ∕∖ PQR in Problems 25-28.
Find sin θ where θ is the angle between v and w in Problems 25. P(0, 1, 1), <2(1, 1, 0), JR(1, 0, 1)
11-16. 26. P (l, 0, 0), β(2, 1 ,-1 ), JR(0, 1, -2 )
11. v = i + k; w = i + j 12. v = i + j; w = i + j + k 27. P (l, 2, 3), β(2, 3, 1), Λ(3, 1, 2)
13. v = j + k; w = i + k 14. v = i + j; w = j + k 28. P ( - l , - 1 , -1 ), <2(1, - 1 , - 1 ) , P ( - l , 1, -1 )
15. v = i + 2j + 3k; w = 4i + 5j + 6k
Determine whether each product in Problems 29-31 is a scalar
16. v = cos θ i —sin θ j; w = sin θ i - cos θ j or a vector or does not exist. Explain your reasoning.
Find a unit vector that is orthogonal to both v and w in 29. a. u × (v ∙ w) b. u ∙ (v × w)
Problems 17-20. 30. a. u × (v × w) b. u ∙ (v ■w)
17. v = 2i + k; w = i —j - k 18. v = j —3k; w = —i + j + k 31. a. (u × v) ∙ (u × w) b. (u × v) × (u × w)
In Problems 32-35, fin d the volume o f the parallelepiped

determined by vectors u, v, and w.
32. u = i + j; v = j + 2k; w = 3k
33. ιι = j + k; v = 2i + j + 2k; w = 5i
34. u = i + j + k; v = i - j - k; w = 2i + 3k
35. u = 2 i + j - k ; v = 3i + k;w = j + k
θ 36. ■ What Does This Say? Contrast dot and cross prod Figure 10.34 Finding the torque
ucts o f vectors, including a discussion o f some of their
properties. 49. Prove the determinant formula for evaluating a triple
37. ■ What Does This Say? What is the right-hand rule? scalar product.
38. ■ What Does This Say? Describe the meaning of 50. a. Show that the vectors u, v, and w are coplanar (all in
∣(u × v) ■w ∣. the same plane) if
39. T HINK TANK P ROBLEM u ∙ (v x w) = 0 or (u × v) ∙ w = 0
a. If u × w = v × w, does it follow that u = v?
b. If u ∙ w = v ∙ w, does it follow that u = v? b. Verify that the vectors u = i + 3j + k, v = 2i - j - k,
and w = 7j + 3k are coplanar.
c. If both u × w = v × w and u ∙ w = v ∙ w, does it follow
that u = v? 51. Show that the triangle with vertices (x 1, y, l ), (x2 , y 2),
40. Find a number s that guarantees that the vectors i, (x 3, y 3) has area A - ∣Z>, where
i + j + k, and i + 2j + .vk will be parallel to the same 1
T1
plane. 1
T2
41. Find a number t that guarantees the vectors i + j, 1
y3
2i - j + k, and i + j + ∕k will be parallel to the same
plane. 52. Using the properties of determinants, show that
42. Find the angle between the vector 2i —j + k and the u ∙ (v × w) = (u × v) ∙ w
plane determined by the points P ( l, —2, 3), β ( - l , 2, 3),
and Λ ( l, 2, -3 ) for any vectors u, v, and w.
43. Let u = i + j, v = 2i - j + k, and w = 3i. Compute © 53. Let A, B, C , and D be four points that do not lie in the
(u × v) × w and u × (v × w). What does this say about the same plane. It can be shown that the volume of the
associativity of cross product? tetrahedron with vertices A , B, C , and D satisfies
44. Show that (αu) × (⅛v) = ab(y × v) and scalars a and b. VOLUM E OF 1 1 / a r e a o f ∖ / A L T IT U D E FR O M ∖
45. For a given vector v in IR3, find all vectors w such that T E T R A H E D R O N A B C D ∖ = 3^ Δ A B C ) ∖ D T O Δ A B C J
V × w = w.
46. What can be said about vectors v, w if v ∙ w = 0? What can Show that the volume is given by
be said if w × v = 0?
47. One end of a 2-ft lever pivots about the origin in the yz- ∣
V = ∣(AB × AC) ∙ AD ∣
plane, as shown in Figure 10.33.
Figure 10.33 Finding the torque 54. Show that if u, v, and w are vectors in R 3 with
u + v + w = 0, then
If a vertical force o f 40 lb is applied at the end o f the lever,
what is the torque o f the lever about the pivot point (the u×v=v×w =w ×u
origin) when the lever makes an angle of 30o with the xy-
plane? 55. Suppose u, v, and w are nonzero vectors in R 3 with
48. A 3-lb weight hangs from a rope at the end of Q o f a 5-ft u × v = w and u ∙ v = 0. Show that
stick P Q that is held at an angle o f 60o to the horizontal, as v = s(w × u) and u = t(y × w)
shown in Figure 10.34. What is the torque about the point
P due to the weight? for scalars s and t.
Sec. 10.5 Lines and Planes in Space 665
56. Show that (cu) × (c∕v) = α∕(u × v). 60. Let a, b, and c be vectors in space. Show
57. Show that ∣
∣
v × w∣
tan θ = "-------- a × (b × c) = (c ∙ a)b - (b ∙ a)c
v∙w
where θ (0 ≤ θ ≤ J) is the angle between v and w. This is called the “cab - bac” formula.
58. Show that if all three vertices of a triangle have integer
Establish the validity of the equations in Problems 61-64 for
coordinates, the area of the triangle is at least 0.5.
arbitrary vectors u, v, w, and z in R3 .
59. Let u, v, and w be nonzero vectors in R 3, no two of which 61. (u × v) × (w × z) = (u ∙ w × z)v - (v ∙ w × z)u
lie in the same plane. Show that
62. u × (v × w) + v × (w × u) + w × (u × v) = 0
∣
u ∙ (v × w)∣= ∣
v ∙ (u × w)∣ 63. u × v = (u ∙ v × i)i + (u ∙ v × j)j + (u ∙ v × k)k
What other triple scalar products involving u, v, and w 64. u × [u × (u x v )]∙w = —∣ ∣u∣∣
2u ∙ v × w
have the same absolute values?
1 0 .5 LINES AND PLANES IN SPACE
IN THIS SECTION Lines in R3, direction cosines, planes in IR3

We represent lines in either parametric or symmetric forms and then turn our
attention to planes in R3.
■ LINES IN [R3
As in the plane, a line in space is completely determined once we know one of its
points and its direction. We used the concept of slope to measure the direction of a
line in the plane, but in space, it is more convenient to specify direction with
vectors.
Suppose L is a line in space that contains Q(x0 , y 0, z0) and is parallel to the
vector v = Ai + 2?j + Ck. We say that L is aligned with v, as shown in Figure 10.35.
We also say that the line has direction numbers A, B, and C and denote these
Figure 10.35 If L is aligned with v direction numbers by [A, B, C]. If P(x, y, z) is any point on L, then the vector QP is
and contains Q, then P is on L parallel to v and must satisfy the vector equation QP = tv for some number t. If we
whenever QP = tv.
introduce coordinates and use the standard representation, we can rewrite this
vector equation as
(x - x 0 )i + (y - p0)j + (z - z0)k = t[A∖ + Bj + Ck]
By equating components on both sides of this equation, we find that the
coordinates of P must satisfy the linear system
x - x nu = tA y∙ - y∙ u= tB
z z z - z nu = tC
where t is a real number.
Parametric Form of a Line If L is a line that contains the point (x0, y 0, z0) and is aligned with the vector
in R3 v = zli + Bj + Ck, then the point (x, y, z) is on L if and only if its coordinates
satisfy
0 t will be a different value for
each point (x, y, z). 0 x = x 0 + tA y = y 0 + tB z = z0 + tC
for some number t.
Turning things around, if we are given the equation of a line with direction
numbers [A, B, C], then v = ∕li + /?j + Ck is the vector aligned with L.
E X A M P L E 1 Parametric equations of a line in space
Find the parametric equations for the line that contains the point (3, 1, 4) and
is aligned with the vector v = - i + j - 2k. Find where this line passes through
the coordinate planes and sketch the line.
Solution The direction numbers are [ - 1, 1, -2 ] and x 0 = 3 ,y 0 = 1, z0 = 4, so the
line has the parametric form
x = 3 —Z y = 1+ t z = 4 - 2z
This line will intersect the x j ,-plane when z = 0;
0 = 4 - 2z implies t= 2
If t = 2, then x = 3 - 2 = 1 and y = 1 + 2 = 3. This is the point (1 ,3 , 0).
Similarly, the line intersects the xz-plane at (4, 0, 6) and the yz-plane at
(0, 4, -2 ). Plot these points and draw the line, as shown in Figure 10.36. ■■■
Figure 10.36 Graph of the line In the special case where none of the direction numbers A , B, or C is 0, we can
x = 3 ~ t , y= ∖ + t, z = 4 - It solve each of the parametric-form equations for t to obtain the following
symmetric equations for a line.
Symmetric Form of a Line If L is a line that contains the point (x0, y 0 , z0 ) and is aligned with the vector ∣
in R 3 v = Ai + Z3j + Ck (A, B , and C nonzero numbers), then the point (x, y, z) is on
L if and only if its coordinates satisfy
x - x 0 _ y - j.-0 _ z - zp
EXAM PLE 2 Symmetric form of the equation of a line in space
Find symmetric equations for the line L through the points P ( - l , 3, 7) and
(2(4, 2, -1 ). Find the points of intersection with the coordinate planes and
sketch the line.
Solution The required line passes through P and is aligned with the vector
P Q = [ 4 - ( - l) ] i + [2 - 3]j + [-1 - 7]k = 5i - j - 8k
Thus, the direction numbers of the line are [5, - 1 , -8 ], and we can choose
either P or Q as (xθ, y 0 , z0 ). Choosing P, we obtain
x + 1_ T ~ 3 . z - 7
5 -1 -8
Next, we find points of intersection with the coordinate planes:
xy-plane: z = 0, so λ g 1 = $ implies x = and = ∣ implies y = -^ ∙
The point of intersection of line with the xy-plane is (qp ∙y , θ). Similarly, the
other intersections are
xz-plane: (1 4 ,0 ,-1 7 )
x + I , .V~3 _ z - 1
yz-plane: (θ, -y> y )
5 -1 -8
The graph is shown in Figure 10.37.
You might remember that two lines in R 2 must intersect if their slopes are
different (because they cannot be parallel), but two lines in space may have
different direction numbers and still not intersect. In this case, the lines are said to
be skew. The reason why the situation in R 3 is different from that in R 2 is that even
though lines with different direction numbers cannot be parallel, there is still
enough “ room” in space for the lines to lie in parallel planes and be aligned with
vectors that are not parallel. This situation is shown in Figure 10.38.
a. Intersecting lines b. Parallel lines c. Skew lines

Figure 10.38 Lines in space may intersect, be parallel, or skew.
EXAM PLE 3 Skew lines in space
Determine whether the following pair of lines intersect, are parallel, or are
skew.
. x - 1_ y + 1 _ z -2 and L 2 -. = ⅛ =
1∙ 2 1 4
Solution Note that L l has direction numbers [2, 1, 4] (that is, L is aligned with
2i + j + 4k) and L 2 has direction numbers [4, - 3 , 1]. If we solve
<2, 1, 4)=/(4, - 3 , 1)
for t, we find no possible solution for any value of t. This implies that the lines
are not parallel.
Next we determine if the lines intersect or are skew. The lines intersect if
and only if there is a point P that lies on both lines. To determine this, we write
the equations of the lines in parametric form. We use a different parameter for
each line so that points on one line do not depend on the value of the
parameter of the other line:
L l '. x = l + 2 s y = - 1+ 5 z = 2 + 4s
Note that 5,(1, -1 , 2) lies on L 1 L 2. X = - 2 + 4∕ y = -3 > t z = -l + t
and T(-2, 0, -1 ) lies on E,.
The lines intersect if there are numbers 5 and t for which
x = 1 + 2s = - 2 + 4t or 2 5 -4 / = -3 ∖'
y = -1 + 5 = -3/ or 5 + 3/=1
z = 2 + 4 s= -l + t or 4 5 -/ = -3
This is equivalent to the system of linear equations
f 25 - 4 / = - 3
15 + 3/=1
[45 - / = - 3
Any solution of this system must correspond to a point of intersection of L 1

and L 2, and if no solution exists, then L 1 and L 2 are skew. Because this is a
system of three equations with two unknowns, we first solve the first two
equations simultaneously to find 5 = -⅛> / = ∣∙ Because 5 = - 5 and t = ∣ do
not satisfy the third equation, it follows that L χ and L 2 do not intersect, so they
must be skew. M B
EXAM PLE 4 Intersecting lines
Show that the lines

τ. -y ~ 1 - ∙y + 1 - z ~2 a ∏d τ∙ X + 2_ y _ z ~2
-½∙ 2 1 4 2’ 4 -3 -1
intersect and find the point of intersection.
Solution L χ has direction numbers [2, 1, 4] and L 2 has direction numbers
[4, - 3 , —1]. Because there is no / for which [2, 1, 4] = ∕[4, —3, -1 ], the lines
are not parallel. Express the lines in parametric form:
L χ. X = 1+25 y = -l + 5 z = 2 + 4s
L 2 . x = - 2 + 4t y = -3 ∕ z= ∣ - ∕
At an intersection point we must have

1 + 25 = - 2 + 4/ or 2 5 - 4 / = - 3
-1 + 5 = - 3 / or 5 + 3/=1
2 + 45 = 5 - t or 45 + ∕ = - J
Solving the first two equations simultaneously, we find 5 = 4 ’ t = ∣∙ This

solution satisfies the third equation, namely,
4 (4 ) + l = - l
To find the coordinates of the point of intersection, substitute s = -⅜ into the

parametric-form equations for L χ (or substitute t = ⅛into L 2) to obtain
⅞ = l+ 2 ( 4 ) = 0
To = - ι + ( 4 ) = 4
Graph of L 1 and L 2
⅞ = 2 + 4(4 ) = 0
Thus, the lines intersect at P(θ, - ∣ , θ).
■ DIRECTION COSINES
z
Besides using direction numbers, the direction of a nonzero vector
v = α,i + α7j + α3k
can be measured in terms of angles a, β, and γ between v and the coordinate axes,
as shown in Figure 10.39. These angles are called the direction angles of v and their
cosines are known as the direction cosines of v.
Note that because i is a unit vector represented by (1, 0, 0), we have
a determines blue plane.
β determines pink cone, v ∙ i = (α1 ∙ 1 ) + (a, ∙ 0) + (α3 ∙ 0) = α1
/determines red vector.
Therefore,
Figure 10.39 Direction cosines of a.
the vector v aχ = v ∙ i = ∣
∣
v∣
∣∣
∣
i∣
∣cos a so that cos a = O
Similar formulas hold for the other direction cosines. If u is a unit vector in the
same direction as v, we have
v d, CL^ CL->
u = M = M i + M j + M k = c° s α i + cos/3 j + c° s γ k
w 2 = (≡ y + (w )2 + (w ) = c ° s2 α + c ° s 2 / 3 + c ° s2 γ = ι
Find the direction cosines and direction angles (to the nearest degree) of the
vector v = - 2 i + 3j + 5k, and verify the formula cos2 α + cos2 β + cos2 γ = !.
We find ∣∣v∣∣ = V ( - 2 ) 2 + 32 + 52 = V z38. Therefore,
a
cosα _ ||v||ι = ξ ∕ = ≈ -0.3244428 a ≈≈cos~1(-0 .3 2 4 4 28) ≈= !09 o
a 2 = ^ ≈ O . 4 86664 2 1 4 4 0
c o s β ^H β ≈ ≈ co s- (θ. 8666 2) ≈ 6 !
—a"i = -⅜ = ≈ 0 .8 1 HO 7 1 -1 11
y ≈= cos (0 .8 t 07i) ≈ 36
o
cos ^ -∣
∣
v∣
∣ V3o
Graph of v = -2 i + 3j + 5k and cos2 a + cos2 β + cos2 γ ≈ (-0.3244) 2 + (0.4867)2 + (0.8111)2 ≈ 1.
■ PLANES IN + 3
Planes in space can also be characterized by vector methods. In particular, any

plane is completely determined once we know one of its points and its orientation;
that is, the “direction” it faces. A common way to specify the direction of a plane
is by means of a vector N that is orthogonal to every vector in the plane, as shown
Figure 10.40 A plane may be de in Figure 10.40. Such a vector is called a normal to the plane.
scribed by specifying one of its
points and a normal vector N. EXAMPLE 6 Obtain the equation for a plane
Find an equation for the plane that contains the point 2(3, -7 , 2) and is
normal to the vector N = 2i + j - 3k.
Solution The normal vector N is orthogonal to every vector in the plane. In
particular, if ∕ j (x, y, z) is any point in the plane, then N must be orthogonal to
the vector
QP = (x - 3)i + (y + 7)j + ( z - 2)k
Because the dot (or scalar) product of two orthogonal vectors is 0, we have
N ∙ QP = 2(x - 3) + (l)(y + 7) + (-3)(z - 2) = 0
Graph of plane in Example 6 2x —6 + y + 7 - 3z + 6 = 0

2x + y - 3 z + 7 = 0
Therefore, 2x + y —3z + 7 = 0 is the equation of the plane. ■■■
By generalizing the approach illustrated in Example 6, we can show that the

plane that contains the point (x0 , y 0 , z0) and has normal vector N = Ai + 5j + Ck
must have the Cartesian equation
N = Ai + Bj + Ck A(x - x 0) + B(y - y0) + C(z - zθ) = 0
This is called the point-normal form of the equation of a plane. By rearranging
terms, we can rewrite this equation in the form Ax + By + Cz + D = 0. This is
called the standard form of the equation of a plane. The numbers [A, B, C] are
called the attitude numbers of the plane (see Figure 10.41).
Notice from Figure 10.41 that the attitude numbers o f a plane are the same as
Figure 10.41 The graph of a plane the direction numbers o f the normal line. This means that you can find normal
with attitude numbers [A, B, C] vectors to a plane by inspecting the equation o f the plane.
EXAMPLE 7 Relationship between normal vectors and planes
Find normal vectors to the planes

a. 5x + 1y - 3z = 0; b. x - 5y + V2z = 6; c. 3x - 7z = 10
Solution a. A normal to the plane 5x + 1y - 3z = 0 is N = 5i + 7j - 3k.
b. For x - 5y + V^2z = 6 the normal is N = i - 5j + V2k.
c. For 3x - 7z = 10 it is N = 3i - 7k. ≡M
Forms for the Equation A plane with normal N = Ai + B] + Ck that contains the point (x0, y 0, z0) has
of a Plane the following equations:
Point-normal form: A(x —xθ) + B(y —y0) + C(z - zθ) = 0
Standard form: Ax + By + Cz + D = 0
where A, B, C, and D are constants.
THEOREM 10.9 Existence of a plane
Let A, B, C, and D be constants with A, B, and C not all zero. Then the graph of the
equation
Ax + By + Cz + D = 0
is the equation of a plane with normal vector N = Ai + Bj + Ck.
Proof: Suppose A ≠ 0. Then the equation Ax + By + Cz + D = 0 can be
written as
∕ l ,Lλ' - √
7±J + By + Cz = 0
which is the point-normal form of the plane that passes through the point
{-DIA, 0, 0) with a normal N = Ai + Bj + Ck. A similar argument applies if A = 0
and either B or C is not zero. ■
plane EXAM PLE 8 Equation of a line orthogonal to a given plane
Find an equation of the line that passes through the point Q(2, -1 ,3 ) and is
orthogonal to the plane 3x - 7y + 5z + 55 = 0. Where does the line intersect
the plane?
Solution By inspection of the equation of the plane we see that N = 3i - 7j + 5k
is a normal vector to the plane. Because the required line is also orthogonal
to the plane, it must be parallel to N. Thus, the line contains the point
x -2 = y+ 1 = z -3 Q{2, -1 , 3) and has direction numbers [3, -7 , 5], so that its equation is
3 - 7 5
x -2 T ÷ l . z -3
3 - 7 5
To find the point where this line intersects the plane, we rewrite it in
parametric form:
x = 2 + 3t, y = -1 - I t, and z = 3 + 5t
Now, substitute into the equation of the plane:
3(2 + 3t) - 7 ( - 1 - 7t) + 5(3 + 5t) = -55
6 + 9 t + 7 + 49Z+ 15 + 25t= -55
83t = -83
t = -1
Thus, the point of intersection is found by substituting t = -1 for x, y, and z:
x = 2 + 3(-1) = -1
p = -l-7 (-l) = 6
z = 3 + 5(-1) = - 2
The point of intersection is (-1 , 6, -2). ■■■
EXAM PLE 9 Equation of a plane containing three given points
Find the standard form equation of a plane containing P ( - l , 2, 1),

β(0, -3 , 2), and R{∖ , 1, -4).
Solution Because the normal N to the required plane is orthogonal to the vectors
PQ and PR, we find N by computing the cross product N = PQ × PR.
PQ = (0+ l)i + (-3 - 2 ) j + (2 - l)k = i - 5j + k
PR = (1 + l)i + (1 - 2)j + (- 4 - l)k = 2i - j - 5k
i j k
N = PQ × PR = 1 - 5 I
2 -I -5
= (25 + l)i - (-5 - 2)j + ( - 1 + 10)k
= 26i + 7j + 9k
We can now find the equation of the plane using this normal vector and any
point in the plane. We will use the point P:
Attitude numbers of the plane — from N = 26i + 7j + 9k
2 6 (x + 1 ) + 7 ( v - 2 ) + 9 (z - 1 ) = 0
↑ ↑ ↑
Point on the plane; we are using P ( - l , 2, 1).
Thus, the equation of the plane is
26x + 26 + 7y - 14 + 9z - 9 = 0
26x + 7y + 9z + 3 = 0 M B
EXAMPLE 10 Equation of the line parallel to the intersection of two given

planes
Find the equation of a line passing through ( - 1, 2, 3) that is parallel to the line
of intersection of the planes 3x —2y + z = 4 and x + 2y + 3z = 5.
Solution By inspection, we see that the normals to the given planes are
N 1 = 3i - 2j + k and N 2 = i + 2j + 3k. The desired line is perpendicular to
both of these normals, so the aligned vector is found by computing the cross
product:
= -8 i - 8j + 8k
The direction of this vector is (-8 , -8 , 8) = -8(1, 1, -1). The equation of the
desired line is
x + 1_ y ~ - _ z - 3
1 1 -1
3x-2y + z = 4
Example 10 can also be used to find the equation of the line of intersection of
the two planes. Instead of using the given point (-1 , 2, 3), you will first need to
find a point in the intersection and then proceed, using the steps of Example 10.
We conclude this section by finding the equation of a plane containing two given
(nonparallel) lines.
EXAMPLE 11 Equation of a plane containing two intersecting lines
Find the standard-form equation of the plane determined by the intersecting
lines
x - 2 V + 5 z + 1 AN D A + 1 = JL = Z ^16
3 - 2 4
Sec. 10.5 Lines and Planes in Space 67 3
Solution Proceding as in Example 4, we find that the lines intersect at (-19, 9,

(-1.0. 161 -29). The aligned vectors for these two lines are v 1 = 3i - 2j + 4k and
v2 = 2i - j + 5k and the normal to the desired plane is orthogonal to both
vl and v2 . Thus, we take the normal to be the cross product:
(-5. 2, 6)
i j k
N = v ∣× v2 = 3 -2 4 = ( - 1 0 + 4 ) i - ( 1 5 - 8 ) j + (- 3 + 4)k
2 -1 5
[2 ,1 ,5 ] = —6i —7j + k
(2 ,-5 ,-1 )
[3,-2, 4] The point of intersection P ( - 19, 9, -29) is certainly in the plane, so we can
use (2, -5 , -1), (-1 , 0, 16), or (-19, 9, -29) to obtain the equation of the
plane is
-6 (x - 2 ) - 7(y + 5) + l(z + 1) = 0
-6 x + 12 —7y - 35 + z + 1 = 0
6x + 7y - z + 22 = 0
PROBLEM SET 1 0 .5
© 1. ■ What Does This Say? Contrast the parametric and 19. x = 6 —It, y = 1 + t, z = 3/
symmetric forms of the equation of a line. How do these 20. x = 6 + 3t, y = 2 - t, z = 2t
forms compare with the equation of a plane?
2. ■ What Does This Say? Describe the relationship be In Problems 21-26, tell whether the two lines intersect, are
tween normal vectors and planes. parallel, are skew, or coincide. I f they intersect, give the point o f
Write each equation for a plane given in Problems 3 -6 in intersection.
standard form. 71 x - 4 _ y ~ 6 = z + 2. x _ J, + 2 _ z - 3
3. 4(x + 1 ) - 2(y + 1) + 6(z - 2) = 0 2 -3 5 , 4 -6 10
4. 5(x - 2 ) - 3(y + 2) + 4(z + 3) = 0 22. x = 4 - 2t; y = 61; z = 7 - 41;
5. -3 (x - 4 ) + 2(y + 1) - 2(z + 1) = 0 x = 5 + t; y = 1 - 3t; z = - 3 + 2t
6. -2 (x + 1) + 4 ( y - 3) - 8z = 0 23. x = 3 + 3t, y = 1 - 41; z = - 4 - If,
x = 2 +3t, y = 5 - ⅛ t , z = 3 - 7t
Find the parametric and symmetric equations for the line(s)
passing through the given points with the properties described 24. x = 2 - 41, y = 1 + t, z = 5 + 5t;
in Problems 7-16. x = 3t, y = - 2 - t, z = 4 - 2t
7. (1, -1 , -2 ); parallel to 3i - 2j + 5k y- 1 z- 4 y -3
25. x - 3 x +2 z- 2
8. (1, 0, -1 ); parallel to 3i + 4j 2 -1 1 ’ 3 -1 1
9. (1, -1 , 2); through (2, 1, 3) _ y - 3 _ z - 2. y+ i z- 3
26. x + 1 x + 1
10. (2, 2, 3); through (1, 3, -1 ) 2 -1 1 ’ 2 3 -4
11. (1, - 3 , 6); parallelto = =

Find the direction cosines and the direction angles for the
12. (1, - 1 , 2); parallelto λ ^J 3 =— = vectors given in Problems 27-32.
27. v = 2i - 3j - 5k 28. v = 3i - 2k
13. (0, 4, -3 ); parallelto 1 = = 29. v = 5i - 4j + 3k 30. v = j - 5k
22
31. v = i - 3j + 9k 32. v = i - j + 3k
14. (1, 0, -4 ); parallelto x = - 2 + 31, y = 4 + t, z = 2 + 2t
15. (3, - 1 , 2); parallel to the xy-plane and the yz-plane Find an equation for the plane that contains the point P and
16. ( - 1, 1, 6); perpendicular to 3x + y - 2z = 5 has the normal vector N given in Problems 33-38.
33. P ( - l , 3, 5); N = 2i + 4j - 3k
Find the points o f intersection o f each line in Problems 17-20
34. P(0, -7 , 1); N = - i + k
with each o f the coordinate planes.
17 x ~ 4 _ T + 3 _ z + 2 + 2 _ z - 6 35. P(0, -3 , 0); N = -2 j + 3k
lz ∙ ιs x + 1 „ k
4 3 1 1 2 3 36. P (l, 1, -1 ); N = - i - 2j + 3k
37. P(0, 0, 0); N = k 57. Find the equation of the line that passes through the point
38. P(0, 0, 0); N = i P(0, 1 ,-1 ) and is parallel to the line of intersection of the
39. Find two unit vectors parallel to the line planes 2x + y —2z = 5 and 3x - 6y —2z = 7.
58. Find a vector that is parallel to the line of intersection of
x - 3 _ y ~^ 1 _ z + 1 the planes 2x + 3y = 0 and 3x - y + z = 1.
4 2 1 59. Find the equation of the line of intersection of the planes
40. Find two unit vectors parallel to the line 3x + y - z = 5 and x - 6y —2z = 10.
60. Find the equation of the line of intersection of the planes
x —1 _ F ÷ ~ _ z + 5 2x - y + z = 8 and x + y - z = 5.
2 4 1
61. Let v = 2i + j and w = 2i —j + 3k. Find the direction
41. Find two unit vectors that are perpendicular to the plane cosines and the direction angles of v x w.
2x + 4y - 3z = 4. 62. Find the direction cosines of a vector determined by the
42. Find two unit vectors that are perpendicular to the plane intersection of the planes x + y + z = 3 and 2x + 3y - z
5x - 3_p + 2z = 15. = 4.
θ 43. Show that the vector 3i - 4j + k is orthogonal to the line θ 63. What can be said about the lines
that passes through the points (0, 0, 1) and (2, 1, -1). χ - χ0 y~y0 z ~ z0
44. Show that the vector 7i + 4j + 3k is orthogonal to the line a{ bl c1
passing through the points (-2, 2, 7) and (3, -3 , 2).
and
45. Find two unit vectors that are parallel to the line of χ - χ0 y~y0 ς ~ ZO
intersection of the planes x + y = 1 and x - 2z = 3. a2 b2 c2
46. Find two unit vectors that are parallel to the line of in
tersection of the planes x + y + z = 3 and x —y + z = 1. in the case where a l a2 + ⅛1⅛2 + cl c2 = 0?
47. Find an equation for the line of intersection of the planes 64. In Figure 10.42, N is normal to the plane p and L is a line
x + y - z = 4 and 2x - y + 3z = 1. that intersects p. Assume N = ai + bj + ck and that L is
given by
48. Find an equation for the line of intersection of the planes
x + 2y + z = 3 and x + y —2z = 4. x = x 0 + At, y = y 0 + Bt, z = zθ + Ct
49. Find an equation for the plane that passes through
P(l, —1, 2) and is normal to PQ where Q(2, 1, 3). a. Find the angle θ between L and the plane p.
50. Find an equation for the line that passes through the point b. Find the angle between the plane x + y + z = 10 and
(1, —5, 3) and is orthogonal to the plane 2x — 3y + z = 1. the line
51. Find an equation for the plane that contains the point
(2, 1 ,-1 ) and is orthogonal to the line
χ - 3 = F+ ' _ z
3 5 2
52. Find a plane that passes through the point (1, 2, - 1) and
is parallel to the plane 2x —y + 3z = 1.
53. Show that the line
x —1 _ y ÷ 1 _ z —2
2 3 4
is parallel to the plane x —2y + z = 6.
54. Find the point where the line
65. Show that a plane with x-intercept a, ^-intercept b, and
x - 1= y + 1= z z-intercept c has the equation
2 - 1 3
intersects the plane 3x + 2y - z = 5.
55. The angle between two planes is defined to be the acute
angle between their normal vectors. Find the angle be assuming a, b, and c, are all nonzero.
tween the planes 2x + y - 4z = 3 and x —y + z = 2, 66. Suppose planes p, and p 2 intersect. If v1 and w1 are vectors
rounded to nearest degree. on p 1 and v2 and w, are on plane p 2, then show that
56. Find the equation of the line that passes through the point (v1 × w,) × (v2 × w2)
P(2, 3, 1) and is parallel to the line of intersection of the
planes x + 2y - 3z = 4 and x - 2y + z = 0. is aligned with the line of intersection of the planes.
Sec. 10.6 Vector Methods for Measuring Distance in R3 675
1 0 .6 VECTOR METHODS FOR MEASURING DISTANCE IN R3
IN THIS SECTION Distance from a point to a plane, distance from a point

to a line
We shall see how the dot product and cross product can be used to measure
distances between points, lines, and planes.
■ DISTANCE FROM A POINT TO A PLANE
Figure 10.43 Distance from To prepare for deriving a formula for the distance from a point to a plane, we will
P to L consider a simpler case, namely, the distance from a point to a line in R2 . Let L be
any given line and P any given point not on L. We wish to find the distance from P
to L — that is, the perpendicular distance d, as shown in Figure 10.43.
If L is a vertical line, then the distance from P to L is easy to find (why?). If L
is not vertical, then we let Q be any point on the line and N be normal to L at Q.
Because Q can be any point on the line, we choose a convenient point, say the y-
intercept (see Figure 10.44).
The distance we seek is seen to be the scalar projection of the vector QP onto
N. Thus,
QP ∙ N ∣
QP ∙ N ∣
N∣
l∣ ∣ ^ ∣
∣
N∣∣
Figure 10.44 Procedure for find
ing the distance from a point to a In particular, we will now apply this formula to find the distance from the point
line P(x0 , v0) t 0 ⅛e line
Ax + By + C = 0
Because Q is any point on the line, we choose Q to be the y-intercept. That is,
choose x = 0. Then y = -C !B (because L is not vertical, B ≠ 0) so that
QP = (x 0 - 0)i + (y 0 + C∕B)j
It can be shown that the normal to the line Ax + By + C = 0 is N = ∕li + 5j so
Q P ∙N √Lx0 + By0 + C
N∣
l∣ ∣ VA 2 + B 2
EXAM PLE 1 Distance from a point to a line in Rt2
Find the distance from the point (5, -3 ) to the line 4x + 3y - 15 = 0 using the
following methods: a. as a scalar projection; b. by using the formula.
Solution a. Let Pbe the point (5, -3 ) and Q be the ^-intercept (0, 5) of the line.
Then
QP = 5i - 8j and N = 4i + 3j
QP ∙ N 2 0 -2 4 -4 _ 4
N∣
l∣ ∣ V16 + 9 5 5
b. Note A = 4, B = 3, C = -15, x,l = 5, and y0 = - 3 so that
Ax 0 + By0 + C 4(5) + 3(-3) - 15 -4 _ 4
VA 2 + B2 vW T 5 5
We also use a projection to find the shortest distance from a point to a plane,
as described by the following theorem.
T H E O R E M 10.10 Distance from a point to a plane in R 3
The distance from the point (xθ, y 0 , z0 ) to the plane A x + By + Cz + D = 0 is given

by
A x r, + By n + Cz f. + D
V τ2 + 52 + C2
Proof: If Q is any point in the given plane, the required distance is found by
Plane Ax + By + Cz + D = projecting the vector QP onto a normal N for the plane. Thus, the distance from
the point to the plane is given by
∣
∣
QP∣∣∣∣
N∣∣∣cosβ∣ _ ∣
QP ∙ N ∣
d= ∣∣QP ∣∣∣cos 0∣ =
l∣
N∣ ∣ ∣ ∣
N∣∣
where θ is the (acute) angle between QP and N (see Figure 10.45).
Suppose P has coordinates (xθ, y 0 , zg ) and the given plane has the
standard form A x + By + C z + D = 0. Then N = Ai + Bj + Ck is a normal
to this plane, and if Q (x l , y 1, z 1) is any particular point in the plane, we have
Figure 10.45 The distance from a
A x l + B y j + C z 1 + D = 0 and
point to a plane in R 3
QP = (x0 - x 1)i + (y0 - y 1)j + (z0 - z)k
The dot product of QP with N is given by
QP ∙ N = (x0 - x 1)T + (v0 - y f B + (z0 - z 1)C
= (Ax 0 + By 0 + Cz 0) - (A x l + B y l + C z 1)
= A x 0 + By 0 + Czθ + D
Because the normal vector has length ∣
∣
N∣∣ = V A 2 + B 2 + C 2 , we can substitute into
the formula
QP ∙ N A x 0 + By g + CZQ + D
d =
I∣N ∣ V∕l2 + B 2 + C 2
E X A M P L E 2 Equation of a sphere given a tangent plane
Find an equation for the sphere with center C ( -3 , 1, 5) that is tangent to the
plane 6x - 2y + 3z = 9.
Solution The radius r of the sphere is the distance from the center C to the given
Plane 6x - 2y + 3z = 9 tangent plane, as shown in Figure 10.46.
6(-3) + (-2)(1) + 3 ( 5 ) -9 _ -1 4 _
√ 6 2 + (-2 ) 2 + 32 7
Therefore, an equation of the sphere is
(x + 3)2 + ( y - l) 2 + ( z - 5 ) 2 = 22 m
■ DISTANCE FROM A POINT TO A LINE

Next, we shall derive a formula for the short
Figure 10.46 The sphere with
est distance from a point P to a line L in R 3.
center C(-3, 1, 5) and radius 2
is tangent to the plane Let Q be a point on L and let v be a vector
6x —2y + 3z = 9. aligned with L . Then, as shown in Figure
10.47, the distance from P to L is given by
d= ∣
∣
QP∣
∣∣sin 0∣
where θ is the acute angle between v and the Figure 10.47 The distance from a
vector QP. point to a line in lR3
Sec. 10.6 Vector Methods for Measuring Distance in R3 677
This reminds us of the cross product v × Q P , and because

l k × Q P ∣ - M IIQP∣I ∣si∏o∣
we have
< ∕-∣∣Q P ∣∣ ∣ s in β ∣ = f c ^ 2 S l
THEOREM 10.11 Distance from a point to a line
The shortest distance from the point P to the line L is given by the formula
j J∣y × Q P ∣i
l∣∣
v∣
where v is a vector aligned with L and Q is any point on L .
Proof: A sketch of the proof precedes the statement of the theorem.
Find the distance from the point P(3, - 8 , 1) to the line

x - 3 _ J' + 7 z + 2
3 -1 5
We need to find a point (9 on the line. We see that Q(3, - 7 , -2 ) is on the
line and that
Q P = - j + 3k
The vector v aligned with L is v = 3i - j + 5k. We now find
j k
v × Qp = 3 -1 5 = ( -3 + 5)i - (9 - 0)j + (-3 + 0)k
0 -1 3
= 2i - 9j - 3k
Finally,
∣∣v × QP∣∣ _ √ 2 *2 + (~9)2 + (-3 ) 2 _ /94
INI √ 3 2 + ( - l ) 2 + (5)2 v35
PROBLEM SET 1 0 .6
θ Find the distance between the point and the line in Problems 9. P (l, 1, -1 ); x - y + 2z = 4
1-6. 10. P(2, 1, -2); 3x - 4v + z = -1
1. (4, 5); 3x - 4y + 8 = 0 11. P(a, -a , 2a); 2ax - y + az = 4α, a Ψ 0
2. (9, -3); 3x - 4y + 8 = 0 12. P(α, 2a, 3a); 3x - 2y + z = —Ha, a ≠ 0
3. (4, -3 ); 12x + 5y - 2 = 0
4. (1 ,—6 );x —3y + 15 = 0 Find the distance from the point ( - 1, 2, 1) to each plane given
5. (8, 14); x - 3y + 15 = 0
in Problems 13-16.
13. the plane through the points (0, 0, 0), (1, 2, 4), and
6. (4, 5); 2x — 5y = 0
( - 2 ,- 1 ,1 )
Find the distance between the point and the plane given in 14. the plane through the point (1,0, 1) with normal vector
Problems 7-12. 2i - j + 2k
7. P (l, 0, - l ) ; x + y - z = 1 15. the plane through the point (-3 , 5, 1) with normal vector
8. P(0, 0, 0); 2x - 3y + 5z = 10 3i + j + 5k
30. x * 1 = '1 2 3= - + 2 an j t jl e ∣jn e pa s s i n g through

16. the plane through the points (—1, 1, 1), (4, 3, 7), and
( 3 ,- 1 ,0 ) (1, 3, -2 ) and (0, 1, -1 )
Find the distance from the point P to the line L in Problems 31. x = -1 + t, y = - 2 t , z = 3 and the line passing through
17-22. (0, - 1 , 2) and (1, - 2 , 3)
—7 V + 1 7 — 1 θ 32. Vector methods can also be used to derive a formula for
17. P( 1, 0, —1);r =■— j— =
the distance between two skew lines L l and L 2 in R 3 .
Assume that L i contains the point P, and is aligned with
18. P ( l, 0, i ) ; f = = f
vector v 1, whereas L 2 contains P 2 and is aligned with v2.
Then, the distance d between L 1 and L 2 (see Figure 10.48)
19. P ( l, - 2 , 2)',γ = γ = ^ j∙ is the same as the distance between two parallel planes
2 λ
containing the lines. Because N = v 1 × v, is normal to
20. P(0, 1, -1 ); ~-- = ⅛ =
both planes, it follows that the required distance d is a
scalar multiple of v l × v2. These geometric observations
21. P(a, 0, -a ); = = ~ j l ∖a ≠ 0 are used in the following theorem.
-7-7 (O ∏ x ~a = i = z + 4 fl
z z . rP⅛u, a, 2/’ ] ι 1
23. Find the equation of the sphere with center C ( -2 , 3, 7)
that is tangent to the plane 2x + 3y —6z = 5.
θ 24. a. Show that the line
x - i _ y _ z + ι
3 -2 1
and the plane x + 2y + z = 1 are parallel.

b. Find the distance from the line to the plane in part a.
25. Three of the four vertices of a parallelogram in R 3 are
O ( - 1, 3, 5). R(6, - 3 , 2), and 5(2, 4, -3 ). What is the area
of the parallelogram? Figure 10.48 Distance between lines
26. Find an equation for the set of all points P(x, y, z)
such that the distance from P to the point P 0( - l , 2, 4) Assume L l and L , contain the points P l and P , and are
is the same as the distance from P to the plane aligned with the vectors v l and v2, respectively. Then the
2x - 5y + 3z = 7. (Do not expand binomials.) distance between the lines is
27. Find an equation for the set of all points P(x, y, z) such
that the distance from P to the line (v1 × v 2) ∙ P , P 2 _ N ∙ P ,P ,
l∣
v, × v2∣
∣ ∣
∣
N∣∣
x - ! + 1
_ z
4 - 1 3
is 5. (Do not expand trinomials.) Prove this result.
33. a. Is the distance between two parallel planes
Find the (perpendicular) distance between the lines given in A x + By + Cz = D { and A x + By + C z = D 2 given by
Problems 28-31. See Problem 32 for a formula. the formula d = ∣Z> - Z>2 |? If not, what is the correct
7S x + 1 = -l ' ~ 2 = z - 1 n n d * ~ 2 = T + 1 = Z formula?
28. 3 _ 2 1 and 5 1 3
b. Find the distance between the parallel planes
29. x = 2 - t, y = 5 + 2t, z = 3t and x = 2t, y = - 1 - /,
z = 1 + 2t x + y + 2z = 2 and x + y + 2z = 4
Chapter 10 Review

Concept Problems 25. How do you find the volume of a parallelepiped?

1. What is a vector and what is a scalar? 26. What is the parametric form of a line in R3?
2. Give both an algebraic and a geometric interpretation of 27. What is the symmetric form of a line in R3?
scalar multiplication of a vector by a scalar. 28. What are the direction cosines of a vector in R3?
3. What is the triangular rule for vector sums? 29. What is the normal vector of a plane in R3?
4. What is the parallelogram rule for vector sums? 30. What is the point-normal form of a plane in R3?
5. State each of the following properties of vector opera 31. What is the standard form of a plane in R3?
tions:
32. What is the formula for the distance from a point to a
a. commutativity of vector addition plane?
b. associativity of vector addition 33. What is the formula for the distance from a point to a
c. associativity of scalar multiplication line?
d. identity for vector addition
e. inverse property for vector addition
Practice Problems
f. vector distributivity
34. Given v = 2i - 3j + k, w = 3i - 2j. Find each of the
g. scalar distributivity for vectors following vectors.
h. magnitude of a vector
a. 2v + 3w
i. parallel vectors
∣
∣
b.v∣
p- ∣ ∣w∣p
j. commutativity for dot product
vector projection of v onto w
c.
k. scalar multiple for dot product
scalar projection of w onto v
d.
1. distributivity for dot product over addition 35. For v = 2i - 5j + k and w = j - 3k, find
m. scalar distributivity for cross product a. v ∙ w b. v × w
n. zero product for cross product
36. Given u = 2i - 3j + k, v = i + j - 2k, and w = 3i + 5k.
o. anticommutativity for cross product In each of the following cases, either perform the
p. vector distributivity indicated computation or explain why it cannot be
r. parallel vectors property for cross product computed.
6. How do you find the length of a vector? a. (u × v) ∙ w b. (u ∙ v) × w
7. State the triangle inequality. c. (u × v) × w d. (u ∙ v) ■w
8. What are the standard basis vectors in R3? Find the equations for the lines and planes in Problems
9. What is the standard-form equation of a sphere? 37-41.
10. What is a cylinder? 37. The line through the points P ( - l , 4 ,- 3 ) and
11. What is the distance between two points in R3? 2(0, -2 , 1)
12. How do you find a unit vector in a given direction? 38. The plane that contains the point P(l, 1, 3) and is
13. How do you determine if two vectors are parallel? normal to the vector v = 2i + 3k
14. Define dot product. 39. The line of intersection of the planes 2x + 3y + z = 2
15. What is the formula for the angle between two vectors? and y - 3z = 5
40. The plane that contains the points P(0, 2 ,-1 ),
16. What is meant by orthogonal vectors? What is the
algebraic condition for finding orthogonal vectors? 2(1, ~3, 5), and R(3, 0 ,-2 )
41. The line that contains the points (1, 2 ,-1 ) and
17. What is a vector projection? Give a formula for finding
the vector projection of v onto w. (3 ,-1 ,2 )
18. What is a scalar projection? Give a formula for finding 42. Find the direction cosines and the direction angles of
the scalar projection of v onto w. the vector u = —2i + 3j + k. Round to the nearest de
gree.
19. What is the vector formula for work?
43. In each case, determine whether the lines intersect, are
20. Define cross product. parallel, or are skew. If they intersect, find the point of
21. What is the right-hand rule? intersection.
22. Give a geometric interpretation of cross product. a. x = 2t - 3, y = 4 - t, z = 2t and
23. What is the formula for the magnitude of a cross ■x + 2 _ y - 3. .
product? 3 5 ,z J
24. What is the determinant form for the triple scalar

product?
47. Find the distance from the point 5(4, 5, 0) to the line
44. Let u = 2i + j, v = i - j —k, and w = 3i + k.
a. Find the volume of the parallelepiped determined by x -1 _y _ z+ 1
these vectors. 3 5 - 1
b. Find a positive number A that guarantees that the 48. An airplane flies at 200 mph parallel to the ground at an
tetrahedron determined by Au, Av, and w has a altitude of 10,000 ft. If the plane flies due south and the
volume which is twice the volume of the original wind is blowing toward the northeast at 50 mph, what is
tetrahedron. the ground speed of the plane (that is, effective speed)?
49. A girl pulls a sled 50 ft on level ground with a rope
45. Find the distance from 5 (-1 , 1, 4) to l x + 5y - z = 3. inclined at an angle of 30o with the horizontal (the
ground). If she applies 3 lb of tension to the rope, how
46. Find the distance from the line x = t, y = It, z = 3t - 1
much work is performed on the sled?
to the line x = 1 - t, y = t + 2, z = t.
A triangle in R3 has vertices A(0, 2 ,-1 ), 5(1, 1, 3), and 19. the plane passing through (4, 1, 3), (—4, 2, 1) and (1, 0, 2)
C(l, 0, -4). Use this information for Problems 1-4. 20. the plane passing through (4, -1 , 2) and parallel to
1. Find the perimeter of the triangle.
2. Find the area of the triangle. x +1 y ~ 2 z- +- ι X - 2 _ 5 _ 3 _ Z -4
-t - = - ∏ and
3. Find the three vertex angles of the triangle. (Round to the
nearest degree.) For the vectors given in Problems 21-24, find ∣ ∣ ∣
, v —w,
v∣
4. Find a number p such that the points A, B, C, and and I N + 3w.
D(p, p, 0) form a tetrahedron of volume V = 100 cubic 21. v = 3i - 2j + k, w = 4i + j - 3k
units. 22. v = i, w = j
Find the equations, in both parametric and symmetric forms, of 23. v = 5i - 3j + 2k, w = —i + 2j —3k
the lines described in Problems 5-8. Find two additional points 24. v = 5i + 4k, w= j + 3k
on each line. For the vectors given in Problems 25-30, find v ∙w and v × w.
5. passing through √4(1, —2, 3), 5(4, -1 ,2 ) 25. v = i, w = j 26. v = k, w = k
6. passing through A(l, 0, -3), 5 (-4 , 3, —2) 27. v = 3i + 2k, w= 2i + j 28. v = i - 3j, w = i +5k
7. passing through 5(1, 4, 0) with direction numbers [2, 0, 1] 29. v = 2i + 3k, w= i - 2j
8. passing through 5(1, 2 ,-5 ) with direction numbers 30. v = i - j + k, w = i + j - k
31. Given the vectors v = 4i + 2j + k, w = 2i + j - 5k. Find
a. 5v - 3w b. ∣∣
2v - w∣∣
Find the equation of the line passing through P and parallel to
the line of intersection of the planes given in Problems 9-12. c. vector projection of v onto w
9. 5(3, 4, -1); x + ly + l z + 5 = 0; l x + y - 3z - 6 = 0 d. scalar projection of w onto v
32. Find the dot and cross product for v = 2i + 5j and
10. 5 (-1 , 0, —2); x + 3 y - z + 4 - G∖ 3x —y —z + 5 = Q
w = 3i + 2j - 4k.
11. 5(2, 3, - 1);l x + 3y - 4z - 3 = 0; x - 5,y + 2z + 1 = 0 33. Find the direction cosines for v = (2i + j) × (i + j - 3k).
12. 5(5, 1, -3); x - y - z + 5 = 0; 2x + 3y - z - 3 = 0 34. Find the direction angles of w = -5 i + 4j - 3k. Round to
the nearest degree.
Find the equation of the plane satisfying the conditions given in
Problems 13-20. 35. Find two unit vectors that are parallel to the line
13. the xy-plane x = y z- 1
=
14. the plane parallel to the xμ-plane passing through (4, 3, 7) 6 2 6
15. the plane through (1 ,-3 , 4) with attitude numbers 36. Find the (acute) angle, rounded to the nearest degree,
(3, 4 ,-1 ) between the intersecting lines
16. the plane through (-1 , 4, 5) and perpendicular to a line _
with direction numbers (4, 4, -3) X - 1 _ 5 - 3 _ +5 X -1 5 3 +5
~ - ~ ~ ~ 2Z ~ and ~ = = Z
17. the plane through (4, —3, 2) and parallel to the plane
5x - ly + 3z — 10 = 0 37. Find the area of the parallelogram determined by 3i - 4j
18. the plane containing and - i —j + k.
x - 3 _ z - 1. x - 3 ,.5 + 2 z - l 38. Find the center and the radius of the sphere
4 2 ’ y = -2 ; and 3 1 -2 4x 2 + 4y2 + 4z2 + 12j∕ - 4z + 1 = 0
39. Find the points of intersection of the line x = 6 + 3z, i - 2j is the same as the angle between v and 2i + k.
y = 10 —2t, z = St with each of the coordinate planes. 53. Find an equation of a plane that passes through (a, 0, 0),
40. Find the points of intersection of the line (0, a, 0), and (a, 0, a).
x + 1= 7 - 2 = z + 4 54. Find an equation of a plane that passes through (⅛, - b , 0),
3 4 8 (0, b, -b ), and (-⅛, 0, b).
with each of the coordinate planes.
55. Find the area of a triangle with vertices ( 0 ,- 1 , 2),
41. Find the point of intersection of the planes (1, 2, -1 ), and ( 3 ,- 1 , 2).
3x —y + 4z = 15, 2x + y - 3z — 1 = 0. and
56. Find an equation for the set of all points whose distances
x + 3y + 5 z - 2 = 0
from the point (0, 0, 6) are equal to their distance from the
42. Find the point of intersection of the planes xp-plane.
x + 2y - 3z + 8 = 0, 5x - 4y + 5z + 2 = 0, and
57. Find the area of the triangle in which the vectors
3x + 3y - 2z + 5 = 0
v = i —j + k and w = 2i + j —2k form two sides of the
43. Find the equation of the plane determined by triangle.
x + 3 _ F .. z - 7 , x + 6 _ J-' + 5 _ z - 1
3 -2 6 and 1 -3 ~ 2 58. Find an equation for the plane that passes through the
origin and is parallel to the vectors i —2j —3k and
44. Find the distance from the point (1, - 1 , 2) to the plane
- i + j + 2k.
2x + y - z = 4.
59. Find an equation for the plane that passes through the
45. Find the distance from the point (1, - 1 , 2) to the line
origin and whose normal vector is parallel to the line of
x +2 = y - 1_ z intersection of the planes 2x - y + z = 4 and
2 1 4
x + 3y - z = 2.
46. Find the work done by the constant force F = 5i + 4j + k
60. Find a number A such that the planes 2Ax + 3y + z = ∖
when it moves a particle along the line from P(2, 1, - 1 ) to
and x - Ay + 3z = 5 are orthogonal.
2(4, 1,2).
61. The lines L t and L 2 are aligned with the vectors i - j and
47. How much work does it take to move a container 25 m
i - j + 2k, respectively. Find an equation for the line L
along a horizontal loading platform onto a truck using a
that passes through the point ( - 1, 2, 0) and is orthogonal
constant force of 100 newtons at an angle of f from the
to both L 1 and L r
horizontal?
48. SPY P ROBLEM The spy successfully computes the inter 62. A parallelepiped is determined by the vectors
cept path in Problem 60, Section 9.4, but the wily u = i - j + k, v = i + 2j - k, and w = 2i + j + k. Find
Boldfinger escapes by varying his speed and direction the altitude from the tip of w to the side determined by u
instead of traveling at top speed in a fixed direction. Two and v.
days later, the submarine is observed landing offshore 63. In Chapter 11, we show that T = i + 2,,vj is a vector in the
from a desert area, and Boldfinger takes off in a helicop direction of the tangent line at each point P(x, x 2) on the
ter. Suppose the desert is coordinatized (units of thou parabola y = x 1 . Find a unit vector normal to the parabo
sands of feet) and that two observation posts are at points la at the point (3, 9).
√4(7, 0, 0) and β(0, 4, 0). At noon, Boldfinger’s helicopter 64. For any nonzero vectors u, v, and w show that the vector
is observed from A in the direction of the vector (u × v) × (u × w) is parallel to u.
—4i + 2j + 5k and simultaneously from B in the direction 65. Let v = ai + bj + ck and w = Ai + Bj + Ck, where a, b, c,
of 3i - 2j + 5k. One minute later, it is observed from A in A, B, and C are constants. Describe the set of vectors
the direction 6i + 7j + 5.005k and from B in the direc v + w∕, where t is any scalar.
tion 13i + 3j + 5.OO5k. The spy is notified and at 12:10
66. Using vectors, show that if the diagonals of a rectangle
P.M. is flying at the point 2(θ, 0, 1) — that is, 1,000 ft
intersect at right angles, then the rectangle is a square.
above the origin O — in an attack helicopter traveling at
150 mph. Assuming that Boldfinger travels in a fixed 67. Use vectors to show that the sum of the squares of the
direction at a fixed rate this time, in which direction lengths of the sides of a parallelogram equals the sum of
should the spy fly (from Q) in order to intercept Bold the squares of the lengths of the diagonals.
finger? 68. Use vector methods to show that any angle inscribed in a
49. Suppose v and w are nonzero vectors. Show that semicircle is a right angle.
∣∣
v∣∣
w+ ∣ ∣
w∣ ∣
v and ∣
∣v∣
∣w —∣ ∣w∣∣v are orthogonal vectors. 69. Let v and w be nonzero vectors. Show that the vector
50. Let v = cos θ i + sin θ j and w = cos φ i + sin φ j. Find B= ∣ ∣
w∣ ∣
v+ ∣∣v∣
∣
w
v × w. Interpret this cross product geometrically, and use bisects the angle between v and w.
it to derive a well-known trigonometric identity. 70. The vectors u, v, and w are said to be linearly independent
51. Find the three vertex angles (rounded to the nearest in R 3 if the only solution to the equation
degree) of the triangle whose vertices are (1, —2, 3), au + by + cw = 0 is a = b = c = 0. Otherwise, the vectors
(-1 , 2, -3 ) and (2, 1, -3 ). are linearly dependent. Determine whether the vectors
52. Find a relationship between the numbers α 1, b }, and c 1 so u = - i + 2k, v = 2i - j + 3k, w = i + 3j - 2k are linearly
that the angle between the vectors v = α l i + ⅛1j + c 1k and independent or dependent.
71. Figure 10.49 shows a parallelogram A B C D . If M is the

midpoint o f side A B , show that line C M intersects diago
nal B D at a point P located one-third of the distance from
B to D by completing the following steps.
B C
V
Figure 10.50 Area of a triangle
79. T H IN K T A N K P RO BLEM In Figure 10.51, Δ A B C is equi

lateral and the points M , N, and O are located so that
Figure 10.49 Parallelogram A B C D A D AM = ∣

AB BN = ∣
BC
a. Let a and b be scalars such that M P = α M C and It can be shown that ∕∖P Q R
BP = ⅛BD. Show that is also equilateral, and
∣
A B + ⅛[AD - AB] = α[⅛AB + AD] ∣
∣
PM] ∣
= ∣
∣
QN∣∣= ∣
∣RO∣ ∣and
∣
∣
AP∣
∣= ∣∣
BQ∣
∣= ∣
∣CR∣∣
.
b. Use the fact that AB and AD are linearly independent
(see Problem 70) to show that
2 - ⅛—∣α = 0 and a - b = 0
a.Show that P Q = γAN, then show that ∣ ∣
AN∣ ∣
2 = ∣
∣∣
AB∣ ∣
2.
Solve this system o f equations to show that P has the Hint: Use the law of cosines.
required location. b. Show that Δ P Q R has area 1/7 that of Δ A δ C .
72. Show that u = a l i + α,j + α3k, v = ⅛1i + ⅛2j + ⅛3k, and c. Do you think the same result would hold if Δ A B C were
w = c 1i + c2j + c,k are linearly dependent (see Problem not equilateral? Investigate your conjecture.
70) if and only if 80. Gram-Schmidt orthogonalization process Let u, v, and w
be nonzero vectors in IR3 that do not lie on the same plane.
αι a2 a3
b b b3
Define vectors a and β as follows:
ι 2
cι c2 c3
α" ^ [≡ ]u and 0= w ^[¾i⅛, ■
73. Show that the vectors u, v, and w lie in the same plane if
and only if u × w ∙ w = 0. a. Show that u. a. β are mutually orthogonal (any pair are
74. Show that orthogonal).
b. If γ is any vector in R 3, show that
u ι ∙ vι u ∙ V
u' . v ; = (u ι × u
2) ∙ (v ι × v
2) Γγ ’ α ∩ Γ τ - β^rl „
U 2 ∙V1 r ^ Γ[γi ∙fu^
Γ] + ⅛ r + ⅛
Hint: First examine the case where u l and u2 are chosen
from the vectors i, j, and k. 81. P U T N A M E X A M IN A T IO N P ROBLEM Find the equations of
75. Prove the Lagrange identity: ∣ ∣
u × v∣ ∣
2 = ∣∣
u∣∣
2∣∣
v∣∣
2 - (u ∙ v)2 . two straight lines, each of which cuts all four of the
following lines:
76. Let A ( - 2 , 3, 7). B (l, 5, -3 ), C(2, 8, -1 ) be the vertices of
a triangle in R 3 . What are the coordinates of the point M L l '. x = 1, y = 0 L 2. y = 1, z = 0
where the medians of the triangle meet (the centroid)?
77. The medians of a triangle meet at a point (the centroid) L 3: Z = 1 ,X = 0 L 4: X = y = - 6 z
located two-thirds of the distance from each vertex to the 82. P U T N A M E X A M IN A T IO N P RO BLEM Find the equation of
midpoint of the opposite side. Generalize this result by the smallest sphere that is tangent to both the lines
showing that the four lines that join each vertex of a
tetrahedron to the centroid of the opposite face meet at a L { : x = t + 1, y = 2z + 4, z = —3? + 5 and
point located three-fourths of the distance from the vertex
L 2 : x = 4z - 12, y = t + 8, z = t + 17
to the centroid.
78. A triangle in R 3 is determined by the vectors v and w, as 83. P U T N A M E X A M IN A T IO N P RO BLEM The hands of an accu
shown in Figure 10.50. Show that the triangle has area rate clock have lengths 3 cm and 4 cm. Find the distance
between the tips of the hands when the distance is
A = ⅜√∣
∣
V∣
∣2∣
∣
W∣∣
2 - (v ∙ w)2
increasing most rapidly.
"Star Trek Project"
The starship Enterprise has been captured by the evil Romulans and is being held
in orbit by a Romulan tractor beam. The orbit is elliptical with the planet
Romulus at one focus of the ellipse. Repeated efforts to escape have been futile
and have almost exhausted the fuel supplies. Morale is low and the food replicator
is failing.
In searching the ship’s log, Lieutenant Commander Data discovers that the
Enterprise had been captured long ago by a Romulan tractor beam and had
escaped. The key to that escape was to fire the ship’s thrusters at exactly the right
position in the orbit. Captain Picard gives the command to feed the required
information into the computer to find that position. But, alas, a Romulan virus
has rendered the computer all but useless for this task. Everyone turns to you and
asks for your help in solving the problem.
Here is what Data discovered: If F represents the focus of the ellipse and P is the
position of the ship on the ellipse, then the vector V from F to P can be written as a
sum T + N, where T is tangent to the ellipse and N is normal to the ellipse. The
thrusters must be fired when the ratio ∣ ∣T∣∕∣
∣∣
N∣ ∣is equal to the eccentricity of the
ellipse.
Your mission is to save the starship from the evil Romulans.
Time is said to have only one di

mension, and space to have three
dimensions . . . The mathematical
quaternion partakes of both of
these elements; in technical lan
guage it may be said to be “time
plus space,” or “space plus time:”
and in the sense it has, or at least
involves a reference to, four
dimensions . . .
W. R. Hamilton
Graves' Life o f Hamilton (New *MAA Notes 17 (1991). “Priming the Calculus Pump: Innovations and Resources,” by Marcus S.
York, 1882-1889), Vol. 3, p. 635. Cohen, Edward D. Gaughan, R. Arthur Knoebel, Douglas S. Kurtz, and David J. Pengelley.
Vector Calculus
11.1 Introduction to Vector
Functions The marriage of calculus and vector methods forms what is called vector
Vector-valued functions;
Operations with vector calculus. The key to using vector calculus is the concept of a vector
functions; Limits and valued function. In this chapter we introduce such functions and
continuity
examine some of their properties. We shall see that vector-valued
11.2 Differentiation and Integra functions behave very much like the scalar-valued functions studied
tion of Vector Functions
Vector derivatives; Tangent earlier in this text.
vectors; Properties of vector
derivatives; The motion of an
object in space; Vector
integrals
11.3 Modeling Ballistics and PERSPECTIVE _
Planetary Motion
The motion of a projectile in a A car travels down a curved road at a constant speed of 55 m i∕hr. What
vacuum; Kepler’s second law additional information do we need about the car to determine whether it will
11.4 Unit Tangent and Normal stay on the road or skid off as it rounds a particular curve? Can we modify the
Vectors; Curvature road (say, by banking) so an average-sized car can travel at moderate speeds
Unit tangent and principal unit
normal vectors; Arc length as a without skidding? How does a highway departm ent decide what warning sign to
parameter; Curvature install on a particular curve? A soldier fires a howitzer whose muzzle velocity
11.5 Tangential and Normal and angle of elevation are known. If the shell overshoots its target by 40 yd, how
Components of Acceleration should the angle of elevation be changed to ensure a hit on the next shot? If a
Components of acceleration; satellite is 200 mi above the earth, how fast must it travel to remain stationary
Applications
about a particular point on the earth’s surface? These and other similar
Chapter 11 Review questions can be answered using vector calculus.
Guest Essay, Howard Eves
684
Sec. 11.1 Introduction to Vector Functions 685
1 1 .1 INTRODUCTION TO VECTOR FUNCTIONS
IN THIS SECTION Vector-valued functions, operations with vector

functions, limits and continuity
Vector-valued functions are defined and are used to study curves in the plane
and in space. We shall also study limits and continuity of vector-valued
functions.
■ VECTOR-VALUED FUNCTIONS
In Section 9.4 we described a plane curve using param etric equations

x = f(t) and y = g(f)
where f and g are continuous functions of t on some interval. We extend this
definition to three dimensions. A space curve is the set of all ordered triples f i (t),
f 2(t),f i (t) satisfying the param etric equations x = f i (t),y = f 2(t),z = f 3(t), w here/,,
∕ 2 , and ∕ 3 are continuous functions of t on some domain D.
The concept of a vector-valued function is fundamental to the ideas we plan to
explore. Here is a definition of this concept.
Vector-Valued Function A vector-valued function (or, simply, a vector function) F with domain D assigns
a unique vector F(Z) to each scalar t in the set D. The set of all vectors v of the
form v = F(Z) for t in D is the range of F. In this text, we shall be concerned
with vector functions whose range is in R 2 or R 3 . That is,
F(Z) = ∕ l (z)i + ∕ 2(z)j in r 2 (plane)
F(Z) = ∕ 1(Z)i + ∕ 2 (Z)j + ∕ 3(Z)k in (R3 (space)
where∕ 1, ∕ 9 , and ∕ 3 are real-valued (scalar-valued) functions of the real num ber t
defined on the domain set D. In this context ∕ 1, f 2 , and ∕ 3 are called the
components of F.
Let F be a vector function, and suppose the initial point of the vector F(Z) is at
the origin. The graph of F is the curve traced out by the terminal point of the
vector F(Z) as t varies over the domain set D, as shown in Figure 11.1.
/ / /
F(r1y / / ∕∖ √
∕∖
Figure 11.1 T he graph o f the vector function F(Z) is traced ψr

out by the term inal p o in t o f F(Z) as t varies over D. O
EXAMPLE 1 Graph of a vector function
Sketch the graph of the vector function

F(Z) = (3 - t)i + (2Z)j + (3z - 4)k
for all t.
686 Ch. 11 Vector Calculus
Solution The graph is the collection of all points (x, y , z) with

x = 3- t y = 2t z = 3t —4
for all t. We recognize these as the parametric equations for the line in R 3 that
contains the point P{3, 0, -4 ) and is aligned with the vector
v = - i + 2j + 3k
as shown in Figure 11.2. ⊂=>
E X A M P L E 2 Graph of a circular helix
Figure 11.2 The graph of Sketch the graph of the vector function
F(r) = (3 - Z)i + (2Z)j + (3t - 4)k
F(∕) = (2 sint)i - (2 cosz)j + (3∕)k
Solution The graph of F is the collection of all points (x, y, z) in R 3 whose
coordinates satisfy
χ = 2 sin t y = —2 cos t z = 3t for all t
The first two components satisfy
χ 2 + y 1 = (2 sin/)2 + ( -2 cost)2 = 4(sin2 ∕ + cos2 /) = 4
which means that the graph lies on the surface of the right circular cylinder
with radius 2, whose axis of symmetry is the z-axis, as shown in Figure 11.3.
We also know that as t increases, the z-coordinate of the point P(x, y, z) on the
graph of F increases according to the formula z = 3t, which means that the
point (x, y, z) on the graph rises in a spiral on the surface of the cylinder
x 1 + y 2 = 4. The point on the graph of F that corresponds to t = 0 is (0, - 2 , 0),
and the points that correspond to t = f and t = ττ are (2, 0, ⅛f) and (0, 2, 3τr),
respectively. Thus, the graph spirals upward counterclockwise (as viewed from
Figure 11.3 The graph of above). The graph, which is known as a right circular helix, is shown in Figure
F(Z) = (2 sin∕)i - (2 cos∕)j + (3z)k 11.3. ≡≡
A well-known application of a helix is the D N A (deoxyribonucleic acid)

molecule, which has a structure consisting of two intertwined helixes, as shown in
Figure 11.4. Some other computer-generated helixes are also shown.
1953 by James Watson and Francis Crick. Figure 11.4 Examples o f helixes
Examples 1 and 2 illustrate how the graph of a vector function

F(0 = ∕ l (0i + ∕ 2(0j + ∕ 3(0k
can be obtained by examining the parametric equations
× = /,(0 y = ∕ 2(0 z = Λ(0
In Example 3, we turn things around and find a vector function whose graph is a
given curve.
EXAMPLE 3 Find a vector function
Find a vector function F whose graph is the curve of intersection of the

hemisphere z = V4 —x 2 —y 2 and the parabolic cylinder y = x 2, as shown in
Figure 11.5.
Solution Finding the parametric representation is sometimes called parametriz
ing the curve. There are several ways this can be done, but the natural choice is
Figure 11.5 The intersection to let x = t. Then y = t2 (from the equation of the parabola), and by
curve of the hemisphere and the substituting into the equation for the hemisphere, we find
cylinder
z = V4 —x 2 —y 2 = V4 - (t)2 - (t2)2 = V4 —t2 - ∕ 4
We can now state a formula for a vector function of the given graph:
F(Z) = ti + t2j + λ∕4 - t 2 - t4k am
■ OPERATIONS WITH VECTOR FUNCTIONS
It follows from the definition of vector operations that vector functions can be
added, subtracted, multiplied by a scalar function, and multiplied together. We
summarize these operations in the following box.
Vector Function Operations Let F and G be vector functions of the real variable t, and let ∕(Z) be a scalar
function. Then, F + G, F - G ,∕F , and F × G are vector functions, and F ∙ G is
a scalar function. These operations are summarized as follows:
Vector functions:
(F + G)(t) = F(∕) + G(t) (F - G)(t) = F(t) - G(t)
(∕F)(t) = ∕(t)F(t) (F × G)(t) = F(t) × G(t)
Scalar function: (F ∙ G)(t) = F(t) ∙ G(t)
EXAMPLE 4 Vector function operations
Let F(t) = t2i + Zj - (sin∕)k and G(t) = ti + γj + 5k. Find
a. (F + G)(t) b. (ez F)(Z)

c. (F × G)(t) d. (F ∙ G)(t)
Solution a. (F + G)(Z) = F(Z) + G(Z)

= [Z2i + Zj - (sin Z)k] + [ri + γ j + 5k]
= (Z2 + Z)i + (Z + r 1)j + (5 - sinZ)k
b. (ezF)(z) = e zF(Z) = e zZ2i + e t tj - (ezsin Z)k
c. (F × G)(Z) = F(Z) × G(Z)
= [Z2i + Zj - (sin Z)k] × [ri + γ j + 5k]
i j k
_ t2 t -s in Z
t 1 5
= [5/ + 5 ψ ^ ]i - [5Z2 + t sinZ]j + [Z - Z2]k
d. (F ■G)(Z) = F(Z) ∙ G(Z)

= [Z2i + Zj - (sin Z)k] ∙ [zi + γ j + 5k^∣
= Z3 + 1 - 5 sinZ
■ LIMITS AND CONTINUITY
Lim it o f a Vector Function Suppose the components f ∖, f 2 , f i o f the vector function

F(Z) = ∕ 1(Z)i + ∕ 2(Z)j + ∕ 3 (Z)k
all have finite limits as t → Zθ, where Zo is any number or + ∞. Then the limit of
F(Z) as t → Zo is the vector
!⅞f w =[!⅛∕<'>]i + [!⅞Λ<)]j +[}⅛∕3(<)]k
EXAM PLE 5 Limit of a vector function
Find lim F(Z), where F(Z) = (Z2 - 3)i + e l i + (sin 77Z)k.

Solution lim F(Z) = [lim (Z2 - 3)[∣
i + [lim (e z)[∣
j + [lim (sinττZ)[∣k
= li + e 2j + (sin 2τr)k
= i + <?2j
For the most part, vector limits behave like scalar limits. The following
theorem contains some useful general properties o f such limits.
T H E O R E M 11.1 Rules for vector limits
If the vector functions F and G are functions o f a real variable Z and 7z(Z) is a scalar
function such that all three functions have finite limits as Z → Zo , then
Lim it of a sum lim [F(Z) + G(Z)] = lim F(Z) + lim G(Z)
z - * zo t ~*t 0 t ~*t 0
Lim it of a difference lim [F(∕) - G(r)] = lim F(Z) - lim G(r)

t ~*t0 t ~*t 0 t ~*t 0
Lim it o f a scalar multiple lim [Λ(Z)F(Z)] = [lim Λ(Z)]∣[lim F(Z)]

Limit of a dot product lim [F(Z) ∙ G(Z)] = [lim F(Z)] ■[lim G(Z)]
Limit of a cross product lim [F(Z) × G(Z)] = ∣^lim F(Z)1 × [lim G(Z)1
^→zo U - *⅞ J U →⅛ -1
These limit formulas are also valid as t → +∞ or as t → -∞ , assuming all limits
exist.
Proof: We shall establish the formula for the limit o f a dot product and leave
the rest o f the proof as an exercise. Let
F(0 = Λ ( z)i* + ∕ 2 (0j + Λ ( z)k and G(Z) = g 1(Z)i + g 2(Z)j + g 3(Z)k

Apply the limit o f vector function along with the sum rule and product rule for
scalar limits to write
liml [F(Z) ■G(Z)] = lliml [ ∕ l (Z)g1(Z) + ∕ 2 (Z)g2 (Z) + f 3(t)g3(t)]

l o ~→o
EXAM PLE 6 Limit of a cross product of vector functions
Verify that lim [F(Z) × G(Z)] = [lim F(Z)j × [lim G(Z)] for the vector functions
F(Z) = Zi + (1 - Z)j + Z2k and G(Z) = e ti - ( 3 + e t )k

Solution
i j k
F(Z) × G(Z) = t 1 - t t2
et 0 -( 3 + e t)
= [(1 - Z)(-3 - e t ) - 0]i - [-Z(3 + e t ) - t2 e t]) + [0 - e z(l - Z)]k
= (tet + 3t - e t - 3)i + (t2e r + te t + 3z)j + (tet - e t)k
Thus, the limit o f the cross product is
lim [F(Z) × G(Z)] = lim [te t + 3t - e l - 3]i + lim [t2 e t + te t + 3z]j + lim [tet - c z]k
t→∖ t→∖ t→∖ (→1
= (e + 3 - e - 3)i + (e + e + 3)j + (e - e)k = (2e + 3)j
Now we find the cross product o f the limits.
lim F(Z) = [lim z]i + [lim (1 - Z)]j + [lim Z2 ]k = i + k
lim G(Z) = [lim e z]i + [lim ( - 3 - e z)]k = ei + ( - 3 - <?)k
Thus,
^l Γ 1 ' j k
lim F(z) J × [lim G(Z)J = 1 0 1
e 0 -3 -e
= (0 - 0)i - ( - 3 - e - e)j + ( 0 - 0)k = (3 + 2e)j
We see lim [F(Z) × G(Z)] = [lim F(Z)j × [lim G(Z)
Continuity of a Vector Function A vector function F(Z) is said to be continuous at t0 if t0 is in the domain of F
and lim F(Z) = F(Z0 ).
What this says: This is the same as requiring each component of F(Z) to be
continuous at t0 . That is,
F(0 = ∕ 1(0i + ∕ 2ωi + ∕ 3(0k
is continuous at t0 when t0 is in the domain of the component functions ∕ 1(Z),
∕ 2(0> and ∕ 3(Z) and
l im ∕ 1(t) = ∕ 1(t0) lim ∕ 2 (Z) = ∕ 2(Z0) lim ∕ (t) = ∕ 3(t0 )
f→f0 1→ I Q l→l0 3
The rules for vector limits listed in Theorem 11.1 can be used to derive general
properties of vector continuity.
EXAM PLE 7 Continuity of a vector function
For what values of t is F(Z) = (sinZ)i + ( 1 - Z)- 1 j + (In t)k continuous?

Solution The vector function F is continuous where its component functions
∕ 1(t) = sin t ∕ 2(t) = (1 - Z)~1 ∕ 3(t) = lnZ
are continuous. The function ∕ 1 is continuous for all t; f 2 is continuous where
1 - t ≠ 0 (that is, where t ≠ l ) ; / 3 is continuous for t > 0. Thus, F is continu
ous when t is a positive num ber other than 1— that is, t > 0, t ≠ 1. ≡≡
PROBLEM SET 1 1 .1
θ F in d the domain fo r the vector functions given in Problems Sketch the graph in R 3 o f the vector functions given in Problems
1 -8. 9-20.
1. F(Z) = 2Zi - 3zj + ∣ k 9. F(Z) = 2zi + Z2j 10. G(Z) = (1 - Z)i + ∣ j
2. F(Z) = ( 1 - z)i + V t j - ⅛ k
r 11. G(Z) = (sinz)i - (cosz)j 12. F(z) = (2cosz)i + (sinZ)j
13. F(Z) = Zi - 4k 14. G(Z) = e tj + Zk
3. F(Z) = (sin Z)i + (cos Z)j + (tan Z)k
15. F(Z) = (cosZ)i + (sinZ)j + Zk
4. F(Z) = (cosZ)i - (cotZ)j + (cscZ)k
16. F(z) = e , i + e l j + e~ , k
5. A(Z)F(Z) where ⅛(Z) = sin Z and
+
17. G(Z) = (1 - Z)i + Z2j + Zk
F(Z) = ~ ⅛ + Γ¼ k
x ' cosZ sιnZ tanZ 18. F(Z) = ( ^ sin Z)i + ( ^ sin Z)j + (cos2 Z)k
6. F(Z) + G(Z), where
19. F(Z) = Zi + (Z2 + l)j + Z2k
F(r) = 3Zj + Z^1k 20. G(Z) = (2 sin Z)i + (2 cos Z)j + 3k
and
G(Z) = 5zi + V 1 0 - Zj θ 21. ■ What Does This Say? Discuss the concept o f a vector
valued function.
7. F(Z) - G(Z), where
22. ■ What Does This Say? Discuss the concept o f the limit
F(Z) = (In Z)i + 3zj - Z2k o f a vector-valued function.
and 23. Show that the curve given by
G(Z) = i + 5zj - z2 k
R(Z) = (2 sin Z)i + (2 sinZ)j + (V δ cosZ)k
8. F(Z) × G(Z), where
lies on a sphere centered at the origin.
F(Z) = Z2 i - Zj + 2zk
and 24. Sketch the graph o f the curve given by
G(Z) = 7⅛ + (/ + 4)j - V + k
R(Z) = (2 cos Z)i + (sin Z)j - 2zk
Sec. 11.1 Introduction to Vector Functions 6 91
Find a vector function F whose graph is the curve given in 30. The line o f intersection of the planes 2x + y + 3z = 6 and
Problems 25-30. x - y - z= 1
26. x 2 + y 2 = 4; z = —1
Perform the operations indicated in Problems 31-42 with
F(∕) = 2/i - 5j + ∕ 2k, G (∕) = (1 - t)i + ∣ k ,
H(∕) = (sin ∕)i + e t j

31. 2F(t) - 3G(Z) 32. ∕ 2F(∕) - 3H(∕)
33. F(∕) ∙ G(t) 34. F(∕) ■H(∕)
35. G(∕) ∙ H(t) 36. F(t) × G(∕)
37. F(t) × H(∕) 38. G(t) × H(t)
27. x = 2t, y = 1 - /, ’ 39. 2ezF(∕) + ∕G(∕) + 10H(∕)
40. F(∕) ∙ [H(∕) × G(∕)]
41. G(∕) ∙ [H(∕) × F(∕)]
42. H(∕) ∙ [G(∕) × F(∕)]
Find each limit indicated in Problems 43-50.
43. lim 3/i + e2zj + (sin ττ∕)k]
(sin∕)i — ∕k^∣
44. lim L ∕ 2 + t - 1 J
z→ 0
7 ^7-1 ÷ t ∖~ 3 t + -,2 j + G 2 + i)e z^ 1k ∣
45. lim
z→ ι L t - 1 t2 + t - 2 i v , J
z
fe i + ^ z L i l
46. lim
z→ 0 L1 — e l cos r J
s ip l i + L - T o s ∕ j + e ι - ∕ k ]
47. lim
f→ 0
Γsin3∕ l∏(sin∕) ,1
48. lim+ Lsin2∕ + ln(tan∕) j J
z→ 0
49. lim [2/i —3j + e zk]
50. lim [(2i - Zj + ezk) × (Z2i + 4 sin Zj)l
Z→ 2
29. The curve of intersection o f the hemisphere Determine all values o f t for which the vector function given in
z = V 9 - x 2 - J’2 and the parabolic cylinder x = y 2 Problems 51-56 is continuous.
51. F(∕) = ti + 3j - (1 - ∕)k
52. G(∕) = ∕i - ∣k
53. G(,) - i i za
54. F(∕) = (e⅛in∕)i + (ezcos∕)k

55. F(∕) = e t [A + ∣j + 3k]
56. G(∕) = ∣
∣
^∣
j' where u = ∕i + Vtj
57. Given the vector functions Hint. Note that

Δ(F × G)(Z) = F(Z + ∆Z) × G(Z + ∆Z)
F(Z) = Zi + Z2j + Z3k and G(Z) = γi - e lj
- F(Z + ∆Z) × G(Z) + F(Z + ∆Z) × G(Z)
directly verify each of the following limit formulas (that
is, without using Theorem 11.1). - F(Z) × G(Z)
a. lim e , F(t) = Γlim e t l Γlim F(Z)1
∣→o L≈→θ J L'→o 1 59. Prove the limit o f a sum rule of Theorem 11. 1.
b. lim F(Z) ∙ G(Z) = Γlim F(Z)1 ∙ lim G (∕) 60. Prove the limit of a difference rule of Theorem 11.1.
∣→ι L<→ι J √→1 J
c. lim [F(Z) × G(Z)] = Γlim F(Z) × him G(Z) 61. Prove the limit o f a scalar multiple rule of Theorem 11.1.
∣→ι L<∙→ι 62. Prove the limit of a cross product rule of Theorem 11. 1.
θ 58. If H(Z) is a vector function, we define the difference 63. T H IN K T A N K P RO BLEM Let F(z) and G(Z) be vector
operator ΔH by the formula functions that are continuous at Zo , and let h(f) be a scalar
function continuous at Zo . In each of the following cases,
ΔH = H(Z + ∆Z) - H(Z) either show that the given function is continuous at Zo or
where ∆Z is a change in the parameter Z. (Note-. Usually, provide a counterexample to show that it is not continu
∆Z is a small number.) If F(Z) and G(Z) are vector func ous.
tions, show that a. 3F(Z) + 5G(Z) b. F(Z) ∙ G(Z)
Δ(F × G)(Z) = F(Z + ∆Z) × ΔG(Z) + ΔF(Z) × G(Z) c. ∕ι(Z)F(Z) d. F(Z) × G(Z)
1 1 .2 DIFFERENTIATION AND INTEGRATION OF VECTOR FUNCTIONS
IN THIS SECTION Vector derivatives, tangent vectors, properties of vector

derivatives, the motion of an object in space, vector integrals
Our next objective is to introduce the derivative and integral of a vector
function, along with some of their basic properties and applications.
■ VECTOR DERIVATIVES
In Chapter 2, we defined the derivative of the (scalar) function/to be the limit as

∆f
Δ Λ- → 0 of the difference quotient We call this the scalar derivative to
distinguish it from derivatives involving vector functions. The difference quotient
of a vector function F is the vector expression
ΔF = F(Z + ∆Z) - F(Z)
∆r ∆z
and we define the derivative of F as follows.
Derivative of a Vector Function The derivative of the vector function F is the vector function F' determined by
the limit
, ΔF ∙ F(Z + ∆Z)-F(Z)
F v(Zf = hm ^ . = 1hm 't
∆∕→o ΔZ ∆z→o ΔZ
wherever this limit exists. In the Leibniz notation, the derivative of F(Z) is
denoted by
We say that the vector function F is differentiable at t = t0 if F'(Z) is defined

at Zo .
Sec. 11.2 Differentiation and Integration of Vector Functions 693
The following theorem establishes a convenient method for computing the

derivative of a vector function.
T H E O R E M 11.2 Derivative of a vector function
The vector function F(∕) = ∕1(∕)i + f 2(t)j + ∕3(∕)k is differentiable whenever the
component functions f v f 2, and ∕ 3 are all differentiable, and in this case
F'(∕) = ∕ 1'(∕)i + ∕2'(∕)j + ∕'(0k
Proof: We use the definition of the derivative, along with rules of vector limits
(Theorem 11.1), and the fact that the scalar derivatives ∕'( t ) , ∕ 2'(t), and ∕ 3'(∕) all
exist.
κ,. . .. F(∕ + Δ∕) - F(∕)
F (/) = hm —------ r ------—
M →O ∆t
[∕1(∕ + ∆∕)i + ∕2(∕ + ∆∕)j + ∕3(∕ + ∆∕)]k - [∕i 1(∕)i + ∕2(0j + ∕3(0k]
hm ---------------------------------------— -------------- ---------- ---------- ------
∆z→o Δγ ∕
f (t + Δ ∕) - ∕,(Q -∣.
lim l ∆t 1
Δ∕→0
= ∕ i'(0i + ∕2'(0j + ∕3'(0k
E X A M P L E 1 Differentiability of a vector function
For what values of t is G (∕) = ∣∕∣i + (cos∕)j + (/ - 5)k differentiable?

Solution The component functions cos / and t - 5 are differentiable for all /, but
∣∕∣ is not differentiable at / = 0. Thus, the vector function G is differentiable
for all / ≠ 0. ⊂~⊃
E X A M P L E 2 Derivative of a vector function
Find the derivative of the vector function

F(∕) = e ti + (sin∕)j + ( ∕ 3 + 5∕)k
Solution F'(∕) = (er)'i + (sin∕)'j + ( ∕ 3 + 5/)'k = e t i + (cos∕)j + (3 ∕2 + 5)k M
■ TANGENT VECTORS
Recall that the scalar derivative ∕' ( x 0) gives the slope of the tangent line to the
graph of f at the point where x = x (J and thus provides a measure of the graph’s
direction at that point. Our first goal is to extend this interpretation by showing
how the vector derivative can be used to find tangent vectors to curves in space.
Let / be a number in the domain of the vector function F(∕), and let P o be the
point on the graph of F that corresponds to ∕ 0, as shown in Figure 11.6. Then, for
any positive number Δ ∕, the difference quotient
ΔF _ F (ZO+ Δ 0 ~ F (⅛)
Δ ∕ ^^ Δ∕
Figure 11.6 The difference quo is a vector that points in the same direction as the secant vector
tient is a multiple of the secant
line vector. P 0Q = F(∕o + Δ∕) - F(∕o)
where Q is the point on the graph of F that corresponds to t = t0 + Δ ∕ (see Figure

11.6). If we choose ∆t < 0, the difference quotient Δ F ∕Δ t points in the opposite
direction to that of the secant vector P 0Q.
Suppose the difference quotient Δ F ∕Δ t has a limit as ∆l → 0 and that
lim ⅛j ≠ 0
∆Z→O ∆t
Then, as ∆t → 0, the direction of the secant P 0Q— and hence that of the difference
quotient ΔF∕∆Z— will approach the direction of the tangent vector at P Q , as shown
Figure 11.7 As Δ∕ → 0, the vector in Figure 11.7. Thus, we expect the tangent vector at P o to be the limit vector
P Q — and hence the difference
quotient ΔF∕∆Z—approaches the
tangent vector at PQ .
which we recognize as the vector derivative F'(Z0). These observations lead to the
Tangent Vector Suppose F(Z) is differentiable at t0 and that F'(tθ) ≠ 0. Then F'(∕θ) is defined to ∣
be a tangent vector to the graph of F(t) at the point where t = t0 .
EXAM PLE 3 Finding a tangent vector
Find a tangent vector at the point where t = 3 on the graph of the vector
function
F(t) = e 2zi + ( ∕ 2 - Z)j + (In Z)k
Solution Because F(3) = e 6i + 6j + (In 3)k, the point of tangency P o has coordi
nates (e6, 6, In 3). The derivative of F is
F'(Z) = 2e2⅛+ (2t — l)j + t^ 1k
The required tangent vector is found by evaluating F'(Z) at t = 3:
F'(3) = 2e6i + 5j + ⅜k β
Once we have the point of tangency and a tangent vector, we can find an
equation for the tangent line to a graph by using the parametric form for a line. For
instance, the tangent line in Example 3 contains the point P 0(e6, 6, In 3) and is
aligned with the tangent vector v = 2e6i + 5j + ∣ k, so that the tangent line at P o
has the parametric form
x = e 6 + 2e6 t y = 6 + 5t z = In 3 + ∣z
where t is a real number (scalar).

We have seen that the graph of F will have a tangent vector F'(Z0) at the point
P o , where t = t0 if F'(7o ) exists and F is not the zero vector. If we also require the
derivative F' to be continuous at t0, then the tangent vector at each point of the
graph of F near P o will be close to the tangent vector at P Q . In this case, the graph
has a continuously turning tangent at P o , and we say that it is smooth there.
Smooth Curve The graph of the vector function defined by F(t) is said to be smooth on any
interval of t where F' is continuous and F'(t) ≠ 0. The graph is piecewise
smooth on an interval that can be subdivided into a finite number of
subintervals on which F is smooth.
As indicated in Figure 11.8, a smooth graph will have no sharply angled points
(cusps). For this reason, vector functions with smooth graphs are “nicely behaved”
and play an especially im portant role in vector calculus.
■ PROPERTIES OF VECTOR DERIVATIVES
Higher derivatives of a vector function F are obtained by successively differentiat

ing the components of F(Z) = ∕ 1(Z)i + f 2(f)j + f 3(f)k. For instance, the second
a. A curve that is smooth derivative of F is the vector function
has a continuously turning
tangent. F''(t) = [F'(Z)]' = ∕ 1"(0i + ∕ 2 "(f)j + ∕ 3 "(0k
whereas the third derivative F"'(Z) is the derivative of F"(Z), and so forth. In the
Leibniz notation, the second vector derivative of F with respect to t is denoted by
√2F
1 Z∕ 3 F
~rτ∙>
2
and the third derivative by ~ if ∙
dt ' dt
EXAM PLE 4 Higher derivatives of a vector function
Find the second and third derivatives of the vector function

F(Z) = e 2t i + ( 1 - Z2 )j + (cos2z)k
b. A curve that is not smooth can
Solution F'(Z) = 2e 2 t i + (-2Z)j + ( - 2 sin2z)k
have “sharp” points. Note that
this graph is piecewise smooth. F"(Z) = 4e 2zi - 2 j - ( 4 cos2z)k
Figure 11.8 Smooth and piecewise
F"'(Z) = 8e 2zi + (8 sin2z)k ≡≡
smooth curves
Several rules for computing derivatives of vector functions are listed in the
following theorem, which can be proved by applying rules for limits of vector
functions to appropriate theorems for scalar derivatives (or even for difference
quotients).
THEOREM 11.3 Rules for differentiating vector functions
If the vector functions F and G and the scalar function h are differentiable at Z,
then so are αF + bG, hF, F ∙ G, and F × G, and
Linearity rule (αF + ⅛G)'(Z) = αF'(Z) + Z>G'(Z) for constants a, b

Scalar multiple ( W ( t ) = 7z, (Z)F(Z) + ∕z(Z)F'(Z)
Dot product rule (F∙G)'(Z) = (F'∙G)(Z) + (F∙G')(Z)
Cross-product rule* (F × G)'(Z) = (F' × G)(Z) + (F × G')(Z)
Chain rule [F(A(Z))]' = h ' ( t ) F ' m
Proof: We shall prove the linearity rule and leave the rest of the proof as an
exercise. Note that if F and G are vector functions differentiable at Zand a, b are
constants, then the linear combination αF + bG has the difference quotient
Δ(aF + ⅛G) a ∆F , ⅛ΔG
∆Z ∆Z ∆Z
*The order is important in the cross-product rule, because the cross product of vectors is not
commutative.
We can now find the derivative:
Δ(αF + b G ) αΔF , ⅛ΔG^∣

(αF + 6G)'(Z) = firn = lim ∆Z ∆Z J
St Δz→ O
= a lim A v + b lim = αF'(Z) + ⅛G'(Z)

Δι→0 ΔZ ∆Z→O ΔZ
EXAM PLE 5 Derivative of a cross product
Let F(Z) = i + Zj + Z2k and G(Z) = Zi + e , ∖ + 3k. Verify that

(F × G)'(Z) = (F' × G)(Z) + (F × G')(Z)
Solution First find the derivative o f the cross product:
i j k
(F × G)(Z) = 1 t Z2 = (3Z - Z2e z)i - ( 3 - Z3)j + (ez - Z2)k
t et 3
so that (F × G)'(Z) = (3 - 2te t - Z2e z)i + (3z2)j + (<?z - 2z)k.
Next, find (F' × G)(Z) + (F × G')(Z) by first finding the derivatives o f F
and G and then the appropriate cross products.
F'(Z) = j + 2Zk and G'(Z) = i + e tj
j k
F' × G = 0 1 2t = (3 - 2Zez)i - ( -2 z *2)j + (-Z)k
t e, 3
i j k
F × G' = 1 t t 2 = (-Z⅛ z)i - ( -Z 2 )j + (e t - Z)k
1 e' 0
Finally, add these vector functions:
(F' × G) + (F × G ') = [(3 - 2te t) - t 2e t ]i + [2t2 + Z2]j + ∖- t + e t - Z]k
= (3 - 2te t - t2 e t )i + (3z2 )j + (e t - 2z)k
Thus, (F × G )' = (F' × G) + (F × G ').
EXAM PLE 6 Derivatives of vector function expressions
Let F(Z) = i + e tj + Z2k and G(Z) = 3Z2i + e~ tj - 2zk. Find
a. ⅛ 2F(Z) + Z3G(Z)] b. ⅛ [F(Z)∙G(Z)]
Solution Begin by finding and
= e tj + 2zk and = 6zi - e zj —2k
a. ⅛ 2F(Z) + ,≈G(Z)] - 2 ⅞ + [ 1 ≈ f + ⅛ = ) G(<)]
= 2(e'j + 2zk) + Z3(6zi - e~ tj - 2k) + 3z2(3z2i + e~ tj - 2zk)

= (6z4 + 9z4 )i + (2e t - t 3e~ t + 3t2e~ t)} + (4z - 2z3 - 6z3)k
= 15Z4i + [2e t + e~ t t 2(3 - Z)]j - 4z(2z2 - l)k
b. ⅛[F(Z)∙G(Z)] = f ∙G(Z) + F ( Z ) ∙⅛
= (etj + 2zk) ∙ (3z2i + e~ tj - 2tk) + (i + e tj + Z2k) ∙ (6zi - e~lj - 2k)

= [0(3z2) + e t (e~t ) + 2z(-2z)] + [l(6z) + e t( - e ~ t ) + Z2(-2)]
= (1 -4 Z 2) + (6Z - 1 - 2 z 2)
= —6Z2 + 6z h b
We shall use Theorem 11.3 in both applied and theoretical problems. In the
following theorem, we establish an important geometric property of vector
functions.
T H E O R E M 11.4 Orthogonality of a derivative
If the nonzero vector function F(Z) is differentiable and has constant length, then
F(Z) is orthogonal to the derivative vector F'(Z)
Proof: We are given that ∣∣F(Z)∣∣ = r for some constant r and all t. To prove that
F and F' are orthogonal, we shall show that F(Z) ∙ F'(Z) = 0 for all Z. First, note that
r2 = ∣∣F(Z)∣∣2 = F(Z) ∙ F(Z)
for all Z. We take the derivative of both sides (remember the derivative of r2 is zero
Tangent because it is a constant):
[r2]' = [F(Z)∙F(Z)]'
0 = F'(Z) ∙ F(Z) + F(Z) ■F'(Z)
0 = 2F(Z) ∙ F'(Z)
0 = F(Z) ∙ F'(Z)
Thus, F and F' are orthogonal. ■
You may recall from plane geometry that the tangent line to a circle is always
perpendicular to the radial line from the center of a circle to the point of tangency
(see Figure 11.9). Theorem 11.4 can be used to extend this property to spheres in
Figure 11.9 The tangent R'(Z0) at R 3. Suppose P is a point on the sphere
Po is orthogonal to the radius vec
tor R(Z0) from the center of the x 2 + y 2 + z 2 = r2
sphere to Po . Let R(Z) = x(Z)i + ,v(Z)j + z(Z)k be a vector function whose graph lies entirely on
the surface of the sphere and contains the point PQ . Suppose P o corresponds to
Z = Zo . Then R(Z0) is the vector from the center of the sphere to PQ , and R'(Z0) is
a tangent vector to the graph of R(Z) at P o , which means that R'(Z0) is also tangent
to the sphere at P o . Because R(Z) has constant length,
∣∣R(Z)∣∣ = V x 2(z) + y 2(f) + z 2(Z) = r for all Z
it follows from Theorem 11.4 that R is orthogonal to R'. Thus, the tangent vector
R'(Z0) at P o is orthogonal to the vector R(Z0) from the center of the sphere to the
point of tangency.
■ THE MOTION OF AN OBJECT IN SPACE
Recall from Chapter 2 that the derivative of an object’s position is the velocity,
and the derivative of the velocity is the acceleration. We frequently know the
acceleration and can use integration to find the velocity and the position. We will
now express these concepts in terms of vector functions.
An object that moves in such a way that its position at time t is given by the
vector function R(z) is said to have
Position Vector Position vector (or displacement) R(t) and

Velocity Velocity V = = R'(t)
and
. . .. t∕ 2R
4 = aV = -^
Acceleration Acceleration A 2^
At any time t,
the speed is ∣
∣
V∣∣, the magnitude of velocity, and
V
the direction of motion is ∏τ7∏∙
≡≡w≡β≡≡≡≡≡≡≡≡
The graph of the position vector R(t) is called the trajectory of the object’s
motion. According to the results obtained earlier in this section, the velocity
V(t) = R'(Z) is a tangent vector to the trajectory at any point where V(t) exists and
V(t) ≠ 0 (see Figure 11.10). Appropriately, whenever V(t) = 0, the object is said to
be stationary.
In practice, the position vector is often represented in the form
R(0 = ∕ 1(0i + ∕ 2 (t)j + ∕ 3(t)k
and in this case, the velocity and the acceleration vectors are given by
Figure 11.10 The velocity vector V(∕) = R'(t) = ∕,'(t)i + ∕ 2'(0j + ∕ 3'(0k
is tangent to the trajectory of the
and
object’s motion.
A(z) = V'(t) = R"(t) = ∕ ' 1'(t)i + ∕ ' 2'(t)j + ∕ ' 3'(t)k
EXAM PLE 7 Speed and direction of a particle
A particle’s position at time t is determined by the vector

R(Z) = (cos ∕)i + (sin t)j + t3k
Analyze the particle’s motion. In particular, find the particle’s velocity, speed,
direction, and acceleration at t = 2.
Solution V= = (-sin t)i + (cos t)j + 3Z2k
A = = (-cosZ)i - (sinZ)j + 6zk
The velocity at time t = 2 is

(-sin 2 )i + (cos2)j + 3(2)2k ≈ —0.9li - 0.42j + 12k
The acceleration at t = 2 is
(-cos2)i - (sin2)j + 6(2)k ≈ 0.42i - 0.9lj + 12k
The speed is ∣∣V∣∣ = V ( -s in t) 2 + (cost)2 + (3 ∕2)2
0 Notice V = ∣
∣
V∣∣
∙^ = V l + 9t4
= (SPEED) ∙ (D IRECTION). 0 At time t = 2, the speed is V l + 144 ≈ 12.04
The direction of motion is = [(- sin t)i + (cosZ)j + 3Z2k]

At time t = 2, the direction is
1 [(-sin2)i +
(cos2)j + 12k] ≈ -0 .0 8 i - 0.03j + k
V145
■ VECTOR INTEGRALS
Suppose we know the velocity of a particle and wish to know its position. We need
to integrate the velocity vector to obtain its displacement. This can be accom
plished by using the following definition.
Let F(Z) = ∕ l (Z)i + f 1{t)] + ∕ 3(Z)k, where ∕ 1, f 2 , and ∕ 3 are continuous on the
Indefinite Integral closed interval a ≤ t ≤ b. Then the indefinite integral of F(Z) is the vector
function
Definite Integral The definite integral of F(Z) on a ≤ t ≤ b is the vector

■b
EXAM PLE 8 Integrals of vector functions
Find ∣ [Zi + 3j - (sin Z)k] dt.
Solution [Zi + 3j - (sinZ)k] dt = t dt i + sin t dt k

∣o
π
(tΛ i + (3O∣oJ ^^ (-∞ st)∣ πk
∖2 j 0 o
= ∣τr2i + 3ττj + (cos 7τ - cos 0)]k
= 57r2i + 3τrj - 2k
EXAM PLE 9 Position of a particle given its velocity
The velocity of a particle moving in space is

V(Z) = e t i + Z2j + (cos2t)k
Find the particle’s position as a function of t if R(0) = 2i + j - k.
Solution We need to solve the initial value problem that consists of:
the differential equation V(Z) = = e ti + Z2j + (cos2z)k
the initial condition R = 2i + j - k
We integrate both sides of the differential equation with respect to t:
e t dt i + t2 dt j + cos2z dt k
R(Z) = (ez + C l )i + (∣Z3 + C 7)j + (∣ sin 2t + C 3)k
= e ti + ∣Z3j + (∣ sin 2z)k + C 1i + C ,j + C 3k

~C~
Now use the initial condition to find C:
2i + j - k = e0i + ∣(0) 3j + (5 sin θ)k + C = i + C

R(0)
so that
i+j - k= C
Thus, the particle’s position at any time t is
R(Z) = e t∖ + ∣Z3j + (j sin2∕)k + i + j - k
= (et + l)i + (∣Z3 + l)j + (| sin2/ - l)k
PROBLEM SET 1 1 .2
θ F in d the vector derivative F' in Problems 1 -4 . F in d the tangent vector to the graph o f the given vector function
1. F(Z) = ti + Z2j + (Z + Z3)k F at the points indicated in Problems 19-24.
2. F(.s, ) = (si + s 2j + s 2k) + (2s2i - sj + 3k) 19. F(Z) = Z2i + 2Zj + (Z3 + Z2 )k; Z = 0, Z = 1, Z = - 1
3. F(s) = In s[si + 5j — e 1 k] 20. F(Z) = Z^3i + Z^2j + Γ 'k -. Z = 1, Z = - 1
4. F(0) = cos 0[i + (ta∏ θ)j + 3k]
M . F<O-!1±'1+⅛,-O, 1-2
F in d F' and F" fo r the vector functions given in Problems 5 -8 . 22. F(z) = z2i + (cos z)j + (z2cos z)k; t = 0, Z = J
5. F(z) = Z2 i + Γ 1j + e 2zk
23. F(Z) = (sinZ)i + (cosZ)j + atk∖ t = fy t = ττ
6. F(s) = ( 1 - 2s 2)i + (5 cos.v)j - sk
24. F(Z) = (e⅛in ττZ)i + (ezcos πZ)j + (sin πt + cos irt)k; Z = 0,
7. F(s) = (sins)i + (cos.sjj + s 2 k
Z = 1, Z = 2
8. F(0) = (sin 2 0)i + (cos20)j + 02k
F in d the indefinite vector integrals in Problem s 25-30.
Differentiate the scalar functions in Problems 9-12. 25. ∫ [Zi - e 3zj + 3k] dt
9. f ( x ) = [xi + (x + l)j] ∙ [(2x)i - (3x 2)j]
26. ∫ [(cos Z)i + (sin Z)j - (2z)k] dt
10. ∕( x ) = [(cosx)i + x j - xk] ∙ [(secx)i - x 2j + 2xk]
11. g(x) = ∣∣(sinx)i — 2xj + (cosx)k∣∣ 27. ∫[(ln Z)(i - Zj + 3k)] dt
12. ∕( x ) = ∣∣[xi + x 2j - 2k] + [(1 - x)i - exj]∣∣ 28. ∫ e^ z[3i + zj + (sin z)k] dt
29. ∫ [(Z In Z)i - sin (1 - Z)j + Zk] dt
In Problems 13-18, R is the position vector fo r a particle in
space at tim e t. F in d the particle's velocity and acceleration 30. ∫ [(sinh Z)i - 3j + (cosh Z)k] dt
vector and then fin d the speed and direction o f motion fo r the © 31. ■ What Does This Say? Describe the concept o f a
given value o f t. smooth curve.
13. R(∕) = Zi + Z2j + 2zk at Z = 1 32. ■ What Does This Say? Compare or contrast the no
14. R(Z) = ( 1 - 2z)i - Z2j + e , k at Z = 0 tions o f speed and acceleration o f a vector function.
15. R(Z) = (cos Z)i + (sin Z)j + 3zk at Z = J
Let v = 2i - j + 5k and w = i + 2j — 3k. F in d the derivatives
16. R(Z) = (2 cos Z)i + Z2j + (2 sin Z)k atZ = J in Problems 33-36.
17. R(Z) = e ti + e~ tj + e 2 , k at Z = In 2 ∞ .½ ( y + tvP) 3 4 .^ ( y - t ⅜ ' )
18. R(Z) = (In Z)i + ∣Z 3j - Zk at Z = 1
Sec. 11.3 Modeling Ballistics and Planetary Motion 701
*π∕2
35. + *2∣
H∣) 36. J t {ty× t2w)
In Problems 37-38, verify the indicated equation for the vector

J 0
F(Z) dt, where
F(Z) = (cos Z)i + (sin Z)j + (sin Z cos Z)k

functions.
F(Z) = (3 + Z2)i - (cos3z)j + Z^1k In Problems 43-44, show that F(z) and F''(Z) are parallel for all
and t with constant k.
G(t) = sin (2 - Z)i - e 2tk 43. F(Z) = ekt i + e~kt) 44. F(Z) = (cos∕cZ)i + (sin∕<Z)j
© 45. Let F(Z) = w(t)i + v(z)j + w(z)k, where u, v, and w are
37. (3F - 2G)'(Z) = 3F'(Z) - 2G'(Z)
differentiable scalar functions of Z. Show that
38. (F ∙ G)'(Z) = (F' ∙ G)(Z) + (F ■G')(Z)
39. The velocity of a particle moving in space is F(z) ∙ F'(z) = ∣∣F(Z)∣∣ (∣∣F(z)∣∣)'
v(z) = ^ Γ = + ⅛ 46. Let F(z) be a differentiable vector function and let ⅛(Z) be a
differentiable scalar function of t. Show that
Find the particle’s position as a function of t if [A(Z)F(Z)]' = ⅛(Z)F'(Z) + A'(Z)F(Z)

R(0) = i - j.
47. If F and G are differentiable vector functions of Z, prove
40. The acceleration of a particle moving in space is
(F ∙ G)'(Z) = (F' ∙ G)(Z) + (F ∙ G')(Z)
2
A(Z) = dN _ d R2 = 24z2i + 4j
dt dt 48. If F and G are differentiable vector functions o f Z, prove
Find the particle’s position as a function of t if (F × G)'(Z) = (F' × G)(Z) + (F × G')(Z)

R(0) = i + 2j and V(0) = 0.
49. If F is a differentiable vector function such that F(Z) ≠ 0,
f1 show that
41. Find F(Z) dt, where
F(z) = ( z √ Γ T 7 ) i + ( τ ⅛ ) j d F(Z) = F'(Z) _ [F(Z) ■F'(Z)]F(Z)
dt ∣∣F(Z)∣∣ ∣∣F(Z)∣∣ ∣∣F(Z)∣∣3
1 1 .3 MODELING BALLISTICS AND PLANETARY MOTION
IN THIS SECTION The motion of a projectile in a vacuum, Kepler’s second

law
Historically, the study of projectile planes and planetary motion were two of
the earliest applications of calculus. In this section, we explore these applica
tions using the methods of vector calculus.
■ THE MOTION OF A PROJECTILE IN A VACUUM
In general, it can be quite difficult to analyze the motion of a projectile, but the
problem becomes manageable if we assume that the acceleration due to gravity is
constant and that the projectile travels in a vacuum (that is, there is no air
resistance). A model of the motion in the real world (with air resistance, etc.)
based on these assumptions is fairly realistic as long as the projectile is reasonably
heavy, travels at relatively low speed, and stays close to the surface of the earth.
Figure 11.11 shows a projectile that is fired from a point P and travels in a
vertical plane coordinatized so that Pis directly above the origin O and the impact
point I is on the x-axis at ground level. We let s0 be the height of P above O, v0 be
the initial speed of the projectile (the muzzle velocity), and a be the angle of
elevation of the “gun” that fires the projectile.
Figure 11.11 The motion of a Because our projectile is to travel in a vacuum, we shall assume that the only
projectile force acting on the projectile at any given time is the force due to gravity. This
force F is directed downward. The magnitude of this force is the projectile’s

weight. If m is the mass of the projectile and g is the “free-fall” acceleration due to
gravity (g is approximately 32 ft∕s 2 or 9.8 m ∕s 2), then the magnitude of F is mg.
Thus, F = (—mg)}. Also, according to Newton’s second law of motion, the sum of
all the forces acting on the projectile must equal mA((), where A(t) is the
acceleration of the projectile at time t. Because F = -mg] is the only force acting
on the projectile, it follows that
—mgj = F = znA(Z)
A(t) = -g j for all t ≥ 0
The velocity V(Z) of the projectile can now be obtained by integrating A(t).
Specifically, we have
V(t) = J A(Z) dt = ∣(-gj) dt = -gt] + C 1
where C 1 = V(0) is the velocity when t = 0 (that is, the initial velocity). Because the
projectile is fired with initial speed v0 at an angle of elevation a, the initial velocity
must be
C 1 = V(0) = (vθcosα)i + (v0sinα)j
(as shown in the detail of Figure 11.11), and by substitution we find
Detail of Figure 11.11
V(t) = -g ∕j + (v0cos α)i + (v0sin α)j
= (v0cos α)i + (v0sin a - g∕)j
Next, by integrating V with respect to t, we find that the projectile has displace
ment
R(Z) = J V(t) dt = J [(v0cos α)i + (v0sin a - gt)j] dt
= (vθcosα)fi + [(vθsinα)Z - ⅛g∕2]j + C2
where C2 = R(0) is the displacement at time t = 0. Because the projectile begins its
flight at P, the initial displacement is R(0) = s0j, and by substituting this
expression for C2, we find that the displacement at time t is
R(t) = [(v0cosα)t]i + [(v0sinα)t - ∖gt 2 + s j j
This vector gives the position of the projectile at any time t after it is fired and
before it hits the ground. If we let x(t) and y(t) denote the horizontal and vertical
components of R(t), respectively, we can make the following statement.
Motion of a Projectile Consider a projectile that travels in a vacuum in a coordinate plane, with the x-
in a Vacuum axis along level ground. If the projectile is fired from a height of s0 with initial
speed vθ and angle of elevation α, then at time t (t ≥ 0) it will be at the point
(x(t), y(t)), where
x(t) = (v0cosα)Z and y(Z) = ~⅛gt2 + (v0sin a)t + .sθ
The parametric equations for the motion of a projectile in a vacuum provide

useful general information about a projectile’s motion. For instance, note that if
a ≠ 90o, we can eliminate the parameter t by solving the first equation, obtaining
t = — -—
vθcos a
By substituting into the second equation, we find

y = ~⅛ ∣ — ~— I + (v∩sin α)∣— —— ] + .s∩
∙f 2 6 Lv cosαJ
0
v 0 'Lv0cosαJ θ
= x 2 + (tan a)x + snu

L2(v0cosα) 2J
This is the Cartesian equation for the trajectory of the projectile, and because the
equation has the general form y = ax 2 + bx + c, with a < 0, the trajectory must
be part of a downward-opening parabola. That is, i f a projectile is not fired
O Range I vertically (so that tan a is defined), its trajectory will be part o f a downward-opening
Figure 11.12 The motion of a parabolic arc, as shown in Figure 11.12. The time o f flight T f of a projectile is the
projectile. The trajectory is part of elapsed time between launch and impact, and the range is the total horizontal
a parabolic arc. distance R traveled by the projectile during its flight. Because y = 0 at impact, it
follows that Tf must satisfy the quadratic equation
- ⅛ 7 }2 + (v os i n α )7 } + ⅛ = 0
Then, because the flight begins at x = 0 when t = 0 and ends at x = R when t = T f

the range R can be computed by evaluating x(Z) at t = Tf -.
E X A M P L E 1 Path of a projectile
A boy standing at the edge of a cliff throws a ball upward at a 30° angle with an
initial speed of 64 ft∕s. Suppose that when the ball leaves the boy’s hand, it is
48 ft above the ground at the base of the cliff.
a. What are the time of flight of the ball and its range?
b. What are the velocity of the ball and its speed at impact?
c. What is the highest point reached by the ball during its flight?
Solution: We know that g = 32 ft∕s 2, and we are given s0 = 48 ft, v0 = 64 ft∕s, and
a = 30°. We can now write the parametric equations for the parabola:
x(Z) = (v0cosα)Z = (64 cos30°)Z = 32Vz3 t 166.277 ft
T(0 = ^ ⅛ 2 + (⅞sin a)t + .sθ
= -⅛(32)Z2 + (64 sin 30o)Z + 48
= -16Z 2 + 32Z + 48
a. The ball hits the ground when y = 0, and by solving the equation
- 16z2 + 32z + 48 = 0 for Z ≥ 0, we find that Z = 3, so the time of flight is
T f = 3 sec. The range is
x(3) = 32√3(3) ≈ 166.27688
That is, the ball hits the ground about 166 ft from the base of the cliff.
b. We find that x'(Z) = 32Vz3 and y'(f) = -2>2t + 32, so the velocity at time Z
is
V(Z) = 32√3 i + (-32z + 32)j
Thus, at impact (when t = 3), the velocity is V(3) = 32Vz3 i - 64j, and its
speed is
∣∣V(3)∣∣ = λ ∕(32√3) 2 + (-6 4 ) 2 ≈ 84.664042
That is, the speed at impact is about 85 ft∕s.

c. The ball attains its maximum height when the upward (vertical) compo
nent of its velocity V(∕) is 0— that is, when y'(t~) = 0. Solving the equation
-32/ + 32 = 0, we find that this occurs when t = 1; therefore, the maxi
mum height attained by the ball is
y m = y (l) = - 16(1)2 + 32(1) + 48 = 64
x m = x (l) = 32√3(1) ≈ 55.425626
Thus, the highest point reached by the ball has coordinates (rounded to the
nearest unit) of (55, 64). M M
We shall now carry out the steps illustrated by Example 1 in a general way to
derive a formula. If the projectile is fired from ground level, then sθ = 0. In this
case, the time of flight satisfies
- ⅛ 7}2 + ( *v os i n a Y
τ f= 0
and we obtain (solving for 77 and rejecting the solution Tf = 0):

rr } = -2v sιnα
7 ∙
0
By substituting / = 7jr in the equation for x , we can find the range:

∕2vθsin α∖ v 2 v2
R - x(T f ) = (v0 cosα) = — (2 sin a cos a) = — sin 2a
∖ g / g g
Notice that for a given initial speed v0, the range assumes its largest value
R m = ⅛v o2 w hen sin2α = 1— that is, when a = 450 = f .
o
Time of Flight and Range A projectile fired from ground level has time of flight 7jand range R given by
the equations
2 ∙ v o2
7 y = ⅜ 0sιnα and Λ=-⅛~sm2α
o o
v2
The maximal range is R m = -⅛-> and it occurs when α = f .
<5
E X A M P L E 2 Flight time and range for the flight of a projectile
A projectile is fired from ground level at an angle of 40° with muzzle velocity
of 110 ft∕s. Find the time of flight and the range.
Solution With g = 32 ft∕s 2, we see that the flight time 7) is
Tf = ∣v 0sin a = ⅛ ( 110)(sin 40°) ≈ 4.4191648
The range R is
v2 /11 ∩∖2
R = -^-sin2α = k ^ y -(s in 80°) ≈ 372.38043
<5
That is, the projectile travels about 372 ft horizontally before it hits the
ground, and it takes a little more than 4 sec. ■■■
■ KEPLER’S SECOND LAW
In the 17th century, the German astronomer Johannes Kepler (1571-1630)

formulated three useful laws for describing planetary motion. The guest essay at
the end of this chapter discusses the discovery of these laws.
Kepler’s Laws 1. The planets move about the sun in elliptical orbits, with the sun at one
focus.
2. The radius vector joining a planet to the sun sweeps over equal areas in
equal intervals of time.
3. The square of the time of one complete revolution of a planet about its
orbit is proportional to the cube of the orbit’s semimajor axis.
We will show how to use vector methods to prove Kepler’s second law, and the
other two can be established in a similar fashion. However, before we begin we
need to prove a result about the velocity and acceleration of a planet.
We shall describe the motion of the planet by means of a polar coordinate
system with the sun at the origin. The position, velocity, and acceleration can be
expressed in terms of the moving unit vectors
u. = (cos 0)i + (sin 0)j and uβ = (-sin 0)i + (cos 0)j
which are shown in Figure 11.13. The “particle” might be a planet in orbit around
the sun, a satellite moving around a planet, or a comet revolving about a star. The
vector u has unit length and points in the direction of OP. The vector u0 has unit
Figure 11.13 Describing the length, points in the direction of increasing θ, and is perpendicular to u. Notice
motion of a particle along a curve that because u = (cos 0)i + (sin 0)j, we have
du
-jρ = (-sιn0)i + (cos 0)j = ue
<Λι
Similarly, you can show that = - u r.
We now find the velocity:
,/p
V(Z) = Definition of velocity
= (rur)' R(t) = rur for r = ∣

∣ ∣
R(z)∣because uf
and OP point in the same direction.
dr cû r
= + Scalar multiple {Theorem 11.3)
dr du a
= d i υ >+ r l θ d i C h a in r u le
To date, there are more than 4,000 You can find a similar formula for acceleration (see Problem 34). We summarize
earth satellites. these formulas in the following box.
Polar Formulas for Velocity xτtt∖ _ dr , l.dθ.,

and Acceleration
ur + dlθ + 2 drdθ∖ a
dt dt2 \.r dt2 dt dt∖ e
EXAMPLE 3 Find the velocity of a moving body
The displacement of a moving body is R(Z) = 2ti - t2j for t > 0. Express R
and the velocity vector V(t) in terms of uf, and uβ∙
Solution We note that on the trajectory, x = It, y = - t 2, and the displacement R
has length
r=∣
∣ ∣
R(t)∣= V(2t)2 + ( - t 2)2 = √ 4 t 2 + t4 = tVt2 + 4
We see R = rur = tV t 2 + 4ur. Because
v (z>= ⅛ur + r ‰ weflnd S and f
⅜ = √ F T 4 + <⅛)(Z2 + 4)” ιz2(2∕) =
Because the polar angle satisfies θ = tan 1 (y∕x) for all points (x, y) on the
curve (except where x = 0), we have θ = tan -1 ( ) = tan^ l I -⅛)
dθ 1 ( Λ —7
d, ^ (-ιn " ^ —2----- Theorem 5.18, page 367
1 + t +4
Thus,
+ rd t 2 + 4 -2 j (2t2 + 4)u, - 2 ∕u β
V(∕) = - ⅜2 = u
Vt + 4 r .t2 + 4 ∕ β Vt 2 + 4
We now have the tools we need to establish Kepler’s second law.

THEOREM 11.5 Kepler’s second law
The radius vector from the sun to a planet in its orbit sweeps over equal areas in
Proof: The situation described by Kepler’s second law is illustrated in Figure
Time T 11.14. We shall assume that the only force acting on a planet is the gravitational
attraction of the sun. According to the universal law of gravitation, the force of
attraction is given by
F = - G " ψ1 u r
r
Figure 11.14 The radial line where G is a physical constant and m and M are the masses of the planet and the
sweeps out equal area in equal sun, respectively. Because this is the only force acting on the planet, Newton’s
time second law of motion tells us that F = mA, where A is the acceleration of the
planet in its orbit. By equating these two expressions for the force F, we find
mA = -G ^ u r
A
λA = ≡ ≡Q^ u11
r2 r
This says that the acceleration o f a planet in its orbit has only a radial component.
Note how this conclusion depends on our assumption that the only force on the
Planet
planet is the sun’s gravitational attraction, which acts along radial lines from the
sun to the planet.
This means that the u 0 component of the planet’s acceleration is 0, and by
examining the formula for acceleration we see that this condition is equivalent to
the differential equation
Γt∙ d 2 θ , 9 dr dθ∖
Lr dt 2 dt dt∖ = 0
2
If we set z = we obtain d θ2 _ dz so that
dt dt ’
r, dz +,
di 20 z, dr _ 0∩
d t~
Figure 11.15 The radial line
z7
-∖ dz _ ~ l7 rr -∖ dr
sweeps out area at the rate d t~ dt
⅛4 _ 1 7 dθ
dt 2 d t' z~ , dz = J ( - 2 r ~ , )dr
In ∣
z∣+ Cj = - 2 In ∣
r∣+ C2
In z = In (Cr^ 2) since r > 0 and z > 0
z = Cr~2
de = Gr
dt
ade _
dt
This formula is equivalent to Kepler’s second law by the following argument. Let 5
be the area swept out by the line from the sun to the planet as time varies from t0 to
t, as shown in Figure 11.15. Then from Theorem 9.1 (area in polar form),
1 f 0w 2 dS 1 2 de 1
S(t) = 52 r dθ so that dt
=
2 r dt
=
2
∙W 0)
Integrating, we find that 5 = ¼Ct + K. In Problem 38 you are asked to show that
because the area lS,(∕) swept out between times t0 and t is a linear function of t, it
follows that equal area is swept out in equal time. This verifies Kepler’s second
law. ■
An object is said to move in a central force field if it is subject to a force that is

always directed toward a single point. Our derivation of Kepler’s second law is
based on the assumption that planetary motion occurs in a central force field. In
particular, the formula
A=^ ≡ u
shows that a planet’s acceleration has only a radial component, and this turns out
to be true for an object moving in a plane in any central force field.
Much of what we know about the bodies in our solar system is a result of the
m athematics associated with Kepler’s laws. Table 11.1 provides some of these
data.
TABLE 11.1 Data on Planetary Orbits

Mean Orbit Period Comparative M ass
Planet (in millions of miles) Eccentricity (in days) (earth = 1.00)
Mercury 36.0 0.205635 88.0 0.06

Venus 67.3 0.006761 224.7 0.81
Earth 93.0 0.016678 365.265 1
Actual mass: 5.975 × 10 z4 kg
Mars 141.7 0.093455 687.0 0.12
Jupiter 484.0 0.048207 4,332.1 317.83
Saturn 887.0 0.055328 10,825.9 95.16
Uranus 1,787.0 0.047694 30,676.1 14.5
Neptune 2,797.0 0.010034 59,911.1 17.2
Pluto 3,675.0 0.248646 90,824.2 0.0025
PROBLEM SET 11.3

θ In Problems 1-8, fin d the time o f flight Tf (to the nearest tenth
second) and the range R (to the nearest unit) o f a projectile fired
at the given angle a with the indicated initial speed v0 . Assume
that g = 32 ft∕s 2 or g = 9.8 m∕s~.
1. a = 35°, v0 = 128 ft/s 2. a = 45°, v0 = 80 ft/s
3. a = 48.5°, v0 = 850 m/s 4. a = 43.5°, v0 = 185 m/s
5. a = 23.74°, v0 = 23.3 m/s 6. a = 31.04°, vθ = 38.14 m/s
7. α = 14.11°, v0 = 100 ft/s 8. a = 78.09°, v0 = 88 ft/s 20. A baseball hit in the Astrodome at a 24° angle from 3 ft
above the ground just goes over the 9-ft fence 400 ft from
In Problems 9-14, an object moves along the given curve in the home plate. About how fast was the ball traveling, and
plane (described in either polar or parametric form). Find its how long did it take the ball to reach the wall?
velocity and acceleration in terms o f the unit polar vectors ur
and uσ „. 21. One of the shortest possible home runs in major league
9. x = 2t, y = t 10. x = sin/, y = cost baseball is over the right field fence at the Kingdome in
11. r = sin θ, θ = 2/ β
12. r = e~ , 0 = 1 - / Seattle. If the distance is 312 ft and if the fence is 8 ft high,
how long will it take for a ball to leave the playing field at
13. r = 5(1 + cos θ),, θ = 2t + 1 14. r = ---- ----- 3’ θ = t that point if it is struck 3∣ ft above the ground at a 32°
' ' 1 - cos θ
angle and we assume that the ball just barely clears the
θ 15. A shell fired from ground level at an angle of 45° hits the
fence?
ground 2000 m away. What is the muzzle velocity of the
shell? Note-, g ≈ 9.8 m ∕s . 2 22. Steve Young, a quarterback for the San Francisco , 49ers,
16. A gun at ground level has muzzle velocity of 300 ft/s. throws a pass at a 45° angle with a speed of 50 ft/s. A
What angle o f elevation (to the nearest degree) should be receiver, Jerry Rice, running at a constant speed of 32 ft/s,
used to hit an object 1500 ft away? catches the ball at a height of 6 ft. If Steve and Jerry both
17. At what angle (to the nearest tenth of a degree) should a line up on the 50-yard line at the start of the play, how far
projectile be fired from ground level if its muzzle velocity does Steve fade back from the line o f scrimmage before he
is 167.1 ft/s and the desired range is 600 ft? releases the ball?
18. A shell is fired at ground level with a muzzle velocity of
280 ft/s and at an elevation of 45° from ground level.
a. Find the maximum height attained by the shell.
b. Find the time of flight and the range o f the shell.
c. Find the speed of the shell at impact.
19. Jeff Bagwell hit a baseball at a 30° angle with a speed of 90
ft∕sec. If the ball is 4 ft above ground level when it is hit,
what is the maximum height reached by the ball? How far
will it travel from home plate before it lands? If it just
barely clears a five-foot wall in the outfield, how far (to
the nearest foot) is the wall from home plate?
23. A golf ball is hit from the tee to a green with an initial 32. A 3-oz paddleball attached to a string is swung in a
speed of 125 ft/s at an angle of elevation of 450. How long circular path with a 1-ft radius. If the string will break
will it take the ball to hit the green? under a force of 2 lb, find the maximum velocity the ball
24. Basketball player Shaquille O’Neal could hit a shot while can attain without breaking the string. Hint'. Note that the
standing 20 ft from the basket. If “The Shaq” shot from a ball is 3 oz, and this is -⅛ lb. Thus, use F = mA, where
height of 7 ft and the ball reached a maximum height of m = (⅛)∕32 = 3∕2 9.
12 ft before passing through the 10-ft-high basket, what
was the initial speed of the ball?
33. A child, running along level ground at the top of a 30-ft-

20 ft high vertical cliff at a speed of 15 ft/s throws a rock over
25. If a shotputter throws a shot from a height of 5 ft with an the cliff into the sea below. Suppose the child’s arm is 3 ft
angle of 46o and initial speed of 25 ft/s, what is the above the ground and her arm speed is 25 ft/s. If the rock
horizontal distance of the throw? is released 10 ft from the edge of the cliff at an angle of
30o, how long does it take for the rock to hit the water?
26. In 1974 Evel Knievel attempted a skycycle ride across the
How far from the base of the cliff does it hit? You may
Snake River, and at the time there was a great deal of hype
assume that the effective initial speed of the rock is
about “will he make it?” If the angle of the launching
v0 = 15 + 25 = 40 ft/s.
ramp was 45o and if the horizontal distance the skycycle
needed to travel was 4,700 ft, at what speed did Evel have
to leave the ramp in order to make it across the Snake
River Canyon?
θ 34. Use the formula

Hint. Assume that the opposite edges of the canyon are
at the same height.
27. A particle moves along the polar path (r, ff), where
/■(/) = 3 + 2 sin t, θ(t) = t i . Find V(t) and Λ(Z) in terms of
ur and u„u. to verify the acceleration formula
In Problems 28-31, consider a particle moving on a circular A(t) dV(f) £R
= =
path o f radius a described by the equation dt dt
r(t) = [a cos ω∕)i + (a sin ωt)j
where ω = dθ∕dt is the constant angular velocity.
28. Find the velocity vector and show that it is orthogonal to
r(∕). Hint'. Use the product rule to find A(t) = V'(∕). In the
29. Find the speed of the particle. computation of A(∕), use the chain rule to show that
30. Find the acceleration vector and show that its direction is
always toward the center of the circle. ⅛ _ ¾ dθ dθ
dt dθ dt β dt
31. Find the magnitude of the acceleration vector.
and 36. Using Problem 35, show that

⅛
dt
= ¾ dθ
dθ dt
= _
r
dθ
dt ' i,
35. Using the acceleration formula given in Problem 34, show
37. Use Problem 36 to show that if F β(Z) = 0, then
that the force F on an object is given by
F(Z) = mA(Z) = F ur + F βuβ
where
is constant.
38. Use Problem 37 to show that in a central force field, the

radius vector R(t) sweeps out area at a constant rate.
and Hint: Note that the rate at which R sweeps out area is
p (t∖, = m r d^θ + 2m — — 1 2 dθ
W- dt- dt dt 2 dt
1 1 .4 UNIT TANGENT AND NORMAL VECTORS; CURVATURE
IN THIS SECTION Unit tangent and principal unit normal vectors, arc
length as a parameter, curvature
In physical problems, it is often useful to describe the motion of an object in
terms of unit vectors that are tangent and normal at each point on the
trajectory. In this section, we see how unit tangent and normal vectors can be
computed; then in the next section, we examine physical applications of such
vectors. We also introduce the idea of curvature, which measures the amount of
bend in a curve.
■ UNIT TANGENT AND PRINCIPAL UNIT NORMAL VECTORS
From Section 11.2, we know that if R(Z) is a vector function with a smooth graph,
then the derivative R'(Z) is a tangent vector to the graph at the point corresponding
to R(Z). Because the graph is smooth, we have R'(Z) ≠ 0, and a unit tangent vector
T(Z) can be obtained by dividing R'(Z) by its length; that is,
,
T(t) = R ω
l∣
R'(0ll
A unit tangent vector at a point P of a curve is shown in Figure 11.16.
Based on our experience with normal lines to curves in the plane, we would
expect a normal vector to be orthogonal to the tangent vector T(t) at each point on
the graph of R. To obtain such a vector, we recall (Theorem 11.4) that any vector
function with constant length is always orthogonal to its derivative R'(Z). There
fore, because T(Z) has length 1 for all t, it follows that T(Z) is orthogonal to T'(Z) for
all t, so that
N(Z) = T (t)
() l∣
τ '( ∕) ∣
∣
Figure 11.16 Unit tangent vector is a unit normal vector. Actually, there are infinitely many vectors orthogonal to
and vectors orthogonal to a tan T(Z)— see Figure 11.16—and we shall refer to N(Z) as the principal unit normal
gent vector on a given curve whenever it is necessary to distinguish it from other normal vectors.
Sec. 11.4 Unit Tangent and Normal Vectors; Curvature 711
Unit Tangent Vector If R(7) is a vector function that defines a smooth graph, then at each point a
unit tangent is
T(Λ RW
( ) = ι∣
R, ωιι
Principal Unit Normal Vector and the principal unit normal vector is
N(t)
( }
=- ⅛
ll≡ )ll
What this says: For each number t0 in the domain of R(Z), we have
constructed a pair of unit vectors, T(∕o) and N(t0), with T tangent to the graph
of R at t = t0 and with N orthogonal to T at t = t0 .
EXAMPLE 1 Unit tangent and principal unit normal

Find the unit tangent vector T(t) and the principal unit normal vector N(t) at
each point on the graph of the vector function
R(t) = et i + e~tj + V2t k
In particular, find T(l) and N(l).
Solution The derivative of R(z) with respect to t is the vector function

R'(t) = e t i - e~tj + Vz2 k, which has length
∣
∣
R'(OII = ⅛ z) 2 + (-e~ t )2 + (√2) 2
= Ve 2z + <? 2z + 2 = V,(et + e t )2 = e t + e t
Thus,
R'(t) et i - e~rj + √2 k
∣
∣
R'(z)∣∣ e t + e ~t
( √ ⅛ + ( ÷ ⅛ ∙+ V2 k
∖et + e v ∖et + e t ∕ .et + e~t ,
So
r>i — p H + ∖ ∕) Ik
T(l) = ~ ----e + e-ι ≈0∙88i - 0∙12j' + 0'46k
To compute the principal unit normal N, we first find T'(t):
(et + e t )et — e t (et - e t ) .

~ (et + e~t )2 1
(et + e~t )(e~t ) + e~t (et - e~t ) .

+ (et + e~t Y j
- V2(ez + e~t )~2(et - e^z )k

= —-t—* _t 2 Γ(e2z + 1 - e2t + l)i + 2j - V2(ez - e^z )kl
(β T ) L
_ 2i + 2j - V2(ez - e~t )k
(et + e~t )2
and
√ 2 2 + 2 2 + 2(ez - e~ t ) 2 ∖J l e 2t + 4 + 2e~ 2t = + e t )2 = √2
(et + e~t )2 (et + e~t )2 (et + e - z )2 et + e ~t
So, the unit normal is
2i + 2j - V z2(ez - e z )k
________(ez + e~z )2________
≡ =
l∣τ'(0∣∣ V2
e t + e~ t
_ 2i + 2j - V2(e z - e z )k _ V2i + V z2j - (el - e z )k
√2(e z + e, - z ) e t + e~ t
In particular,
n (1) - 2i + ⅜ ^ - V ^ ' ,k “ 0 ∙4 6 i + 0 ∙ 4 6 i - ° 7 6 t
Remember, because T and N are to be orthogonal, we can check our work by

noting that
T ∙ N = (O.88i - 0.12j + 0.46k) ∙ (0.46i + 0.46j - 0.76k) = 0 H ≡
We continue to remind you that if you use some software packages, your
answers may take on slightly different, but yet equivalent, forms. For
Example 1 we obtain
τ 'M - 2e2z ■ , 2e2z . , V 2 e z(l - ez)(ez + 1)
i ω (e2 z +l ) * (e + l ) 2 j
2 2 z (e2 z + l ) 2
Continuing with the software program, we find
∣fr '(')∣∣ = ^⅜τ

Finally,
N (z ) = .T'(0 = V2ez i + ∖-⅛' j + f 1 - e2z ∖

’ !> ∣
∣
T '(≡ e 2t + 1 e2t + l j ∖e 2t + 1/
N (l) = 0.458243571484i + 0.458243571484j - 0.761594155955k

Checking T ∙ N for t = 1, we obtain -1.49398426240 10~6, which is scientif
ic notation for a number very close (but not equal to) zero. Note that the
rounded check in Example 1 showed that this was exactly 0, but that result
was, by coincidence, due to rounding. The best you can do with most
approximate results is to show that they are approximately equal.
■ ARC LENGTH AS A PARAMETER
In Section 11.2, we found that time t is the most natural parameter for studying
motion along a curve. However, if we are primarily interested in the geometric
features of a curve, it may be more convenient to use arc length 5 as a parameter,
because arc length is a natural feature of the curve and is independent of

coordinate systems. The arc length parameter is defined in terms of the following
integral function.
Arc Length Function Let R(∕) be a vector function whose graph is a smooth curve C on the closed
interval [a, ⅛]. Then the arc length function on [a, b] is defined by
ft
s(t) = ∣∣R'(M)∣∣ du a ≤ t ≤ b
Ja
Notice that 5(/) increases in the direction of increasing t. That is, arc length 5
increases in the same direction as the parameter t along the curve C (see Figure
11.17).
In the case where R(∕) = x(∕)i + y(∕)j + z(∕)k, we have
R'(w) = x'(w)i + j , '(z∕)j + z'(w)k
/ Direction o f ∕∖ ∖
and
' increasing t / J_]_
∣
∣
R'(w)∣
∣= V[x'(u)] 2 + [>, '(u)]2 + [z'(w)]2
Figure 11.17 Arc length 5 increas Thus, on the interval [a, b∖, the trajectory of R(∕) has arc length
es in the same direction as t along
the graph of R. L = s(b) = ∣ R'(w)ll du = f V[x'(w)]2 + [p'( u )]2 + IZ (M )]2 du
Ja Ja
This gives a formula for the arc length of a curve in space that is a direct
generalization of the formula obtained for the arc length of a planar curve in
Chapter 9 (Theorem 9.4).
T H E O R E M 11.6 Arc length of a space curve
If C is a smooth curve defined by R(t) = x(∕)i + y(∕)j + z(∕)k on an interval [a, Z>],
then the arc length of C is given by
5 = [ ∣∣R'(∕)∣∣ dt = f V [χ '(∕)] 2 + [ ∕( ∕) ] 2 + [z'(∕)] 2 dt

Ja Ja
rb rb ι rb rb
Proof: s = ∖ ds = % d t = ∖ ∣∣R'(∕)∣∣ dt = V [x '(∕)] 2 + [y'(∕)] 2 + [z'(∕)] 2 √∕
Ja J a UL Ja Ja
Note that 5 is also the length of the trajectory between times t = a and t = b. ■
E X A M P L E 2 Finding the arc length of a space curve
Find the arc length of the curve defined by

R(∕) = 12/i + (5 cos∕)j + ( 3 - 5 sin∕)k
from t = 0 to t = 2.
Solution We have x(t) = 12/, y(t) = 5 cos/, and z(∕) = 3 - 5 sin/, so that
x '(∕) = 12, y'(t) = - 5 sin/, and z'(∕) = - 5 cost. We now use Theorem 11.6.
5 = f V[x'(t)] 2 + [y'(∕)] 2 + [z'(∕)] 2 √t
J (7
r2 ___________________________________ r2
= V 1 2 2 + ( - 5 s in ∕) 2 + (-5cos^) 2 dt = 13d⅛ = 26
Jo Jo
By applying the second fundamental theorem of calculus, we can differentiate

the arc length function on [a, b] to obtain
⅛ = I∣
R'(C ∣
I
Recall from Section 11.2 that the speed of an object moving on its trajectory is
given by ∣∣
V(t)∣
∣= ∣
∣R'(r)∣∣
, where V(t) is the velocity vector of the object. Thus, the
speed, in terms of arc length, is
l∣
V(0 ∣
∣= ∣
∣
R'(0 ∣
∣=⅞
For future reference, we summarize this result in the following theorem.
T H E O R E M 11.7 Speed in terms of arc length

Suppose an object moves with displacement R(Z), where R'(t) is continuous on the
interval [a, ⅛]. Then the object has speed
∣
∣
V(0 ∣
∣= ∣
∣
R , (O∣
I= ⅛ for a ≤ t ≤ b
Proof: An outline of the proof precedes the statement of the theorem. ■

EXAM PLE 3 Speed and distance traveled by an object moving in space
An object moves with displacement

R(t) = e ti + (V2 t + 3)j + e~ t k
Find the speed of the object at time t and compute the distance the object
travels between times t = 0 and t = 1.
Solution By differentiating R(t) with respect to t, we find the velocity:
V(t) = R'(t) = e ti + √2 j - e~ t k
Therefore, the speed at time t is
⅞ = l∣
R'(0ll = ⅛ T + (√2)2 + ( - e - φ
= ∖e 2t + 2 + e~2t = S∣(et + e~ t )2 = e , + e~ t
The distance traveled by the object between times t = 0 and t = 1 is the arc
length and is given by
t f z z 1
s = f (e + e ) dt = [e - e ]θ = e - e - 1 + 1 ≈ 2.350 402 4
∣o
C O M M E N T : Incidentally, notice that the object is at the point P ( l, 3, 1) when

t = 0 and is at Q∖e, V z2 + 3, e - l ) when t = 1. The straight-line distance
between P and Q is
d = √(e - 1)2 + (√2 + 3 - 3)2 + (e~ l - 1)2 ≈ 2.313 453 9

which, as might be expected, is slightly less than the distance measured from P to
Q along the trajectory.
When R is represented as R(s) in terms of the arc length parameter 5, the unit
tangent vector T can be represented as T = dPJds. To see why this is true, recall
from Theorem 11.7 that ∣ ∣
R'(Z)∣∣= ds∕dt, and because
R'(<) = ⅞
it follows that
<√R
τ ... R (0 .. dt dR
R'(Z)∣
∣
∣ ∣ ds ds
dt
Note that because 5 increases with t, we have > 0, so the unit tangent vector
dR
T = must point in the direction of increasing s.
Next, to obtain a formula for the principal unit normal vector
N =X ⅛
in terms of the arc length parameter s, we first note that

T 7 A - d T _ d T ds
' ' dt ds dt
Because > 0, it follows that the vector derivatives and point in the same
direction, and N can be computed by finding and dividing by its length. The
scalar function κ = ∣∣^^∣∣ is called the curvature of the graph (defined below).
Thus, we have
dτ
N = - 1 dT
■1∣√T∣∣ K ds
ι∣6⅛ll
We summarize these observations in the following box.
Formulas for Unit Tangent If R(Z) has a smooth graph and is represented as R(.sj in terms of the arc length
and Principal Unit Normal parameter s, then the unit tangent T and the principal normal N at each point
in Terms of the Arc Uength satisfy
Parameter
T = and N = - ⅞e
ds κ ds
where κ = ∣∣~ι~∣∣ is the curvature of the graph.
We also note that
= (CURVATURE)(PRINCIPAL U N IT NORMAL) = K N
■ CURVATURE
Curvature plays an important role in certain applications, some of which we shall

explore in the next section. For a plane curve, the curvature is the absolute value of
the rate of change of the angle of inclination of the tangent line with respect to the
distance along the curve.
Thus, curvature of the graph of R(Z) may be computed by finding T and then
substituting into the formula
For instance, in Example 1 we found that R(Z) = e t i + e zj + λ∕2zk, and

e ei - e zj + V2k
T(t) = ’ so that
e t + e~ t
⅛ = l∣R'(0ll = et + e t and ∣kr∣∣ = t V2

Idt 11 e + e t
Thus, the curvature in this case is the scalar function

il√τ∣l √2
II z/Z II _ e t + e , _ V2
ds e t + e~ t (e t + e~ t )2
dt
Curvature for
R(∕) = e'i + e ''j + V 2 ∕ k It can be shown (see Problem 27) that a line has curvature 0. Because the curvature
t κ is nonnegative, smaller curvature indicates a gentler curve. For Example 1, we see
that if t = 0, the curvature is about 0.35, whereas if t = 5 the curvature is
0.0 0.353 553 390 59 0.0000642 (almost flat). Some curvature calculations for this curve are shown by
0.5 0.278 051 262 51 the table in the margin (which was done using a spreadsheet program).
1.0 0.148 483 352 44 In the case where the graph of R(Z) lies entirely in a plane, the curvature κ(Z)
1.5 0.063 889 444 89 has a nice geometric interpretation. Note that in this case, the unit tangent vector
2.0 0.024 978 838 68 T at a point P can be expressed as
2.5 0.009 401 771 91
3.0 0.003 488 170 89 T = (cos φ)i + (sin φ)j
3.5 0.001 287 247 14 where φ is the angle of inclination of the tangent line at P, as shown in Figure
4.0 0.000 474 097 66 11.19. Then, differentiating both sides of the equation for T, we have
4.5 0.000 174 484 75
5.0 0.000 064 199 37 ⅞p ~ f - sin φ ⅛^')i ÷ fcos φ
as ∖ a s) ∖ ds )∙,
and because the curvature is the length of this vector, we find
«=∣∣f ∣∣ - ∣ f |v(-sinψ)≈+(∞s ⅛ - ∣ f ∣
This curvature is interpreted geometrically in Figure 11.19b.
a. In a plane, the unit tangent b . T h e curvature κ = ∣ -■■■1measures

Figure 11.19 The curvature vector is T = (cos φ)i + (sin 0)j.
of a graph the rate at w hich the curve bends
aw ay from the tangent.
By looking at the formula κ = ∣-^ ∙, we see that at each point P on a planar curve,
the curvature κ measures the rate at which the curve bends away from the tangent
line.
The preceding curvature formula is not particularly easy to use, and because
the graph C of the vector function is frequently defined by R(Z), it is often useful to
have the following formula for κ in terms of R:
Curvature Formula The curvature of a graph C defined by a vector function R(Z) can be found by
k. ; . l∣R' × R"ll
∣∣R'∣∣3
M B .>V■ » β ≡ V 8S≡≡8≡M !*∙⅛W .W .l β »≡ SB≡ M ≡ ∙ H M M I S » Λ ≡ ∙ * B β ≡ ≡ ∙ ≡ 8»≡ ≡ ≡ ≡ ≡ ≡ ≡ β β ≡ ≡ S S ≡ ≡
This formula is derived in Problem 51 and is illustrated in Example 4.
EXAM PLE 4 Curvature of a helix
Given R(∕) = (a cos Z)i + (a sin ∕)j + btk with a and b both nonnegative,
express the curvature of R(Z) in terms of a and b (a > 0).
Solution We begin by finding V and A and the cross product V × A:

R'(Z) = ( - a sinZ)i + (a cosZ)j + bk
R"(Z) = (- α cost)i + ( - α sin Z)j
i j k
R × R" = - a sin t a cost b = (ab sin Z)i - (ab cos Z)j + α2k
- a cost - a sinZ 0
We now need to find the magnitude of this vector, as well as that of V:

∣
∣
R' × R " ∣
∣= ∖, (ab sinZ)2 + (ab cost) 2 + a 4
= V a 2⅛2(sin2 Z + cos2 Z) + α4
= V<a2Zι2 + Ω4
and
∣
∣
R '∣
∣= V α 2sin 2 Z + α2cos2 Z + b 2 = V α 2 + b 2
Then we have
∣
∣
R' × R " ∣
∣_ V α 2Z)2 + α4 ______ a ∖∕a2 + b 2 _ a
κ ~ ∣∣R'∣∣3 (Vα 2 + ⅛2)3 (a2 + ⅛2)V α 2 + b 2 a 2 + b2
If we increase b, then for a fixed a the curvature decreases. If b = 0, the helix
reduces to a circle of radius a. If a = 0, the helix flattens out along the
z-axis. e≡zzi
In the plane, where a curve is given by R(Z) = x(Z)i + y(Z)j, the curvature may
be computed by the formula
= k > " ~ y 'χ " ∖

κ [(v')2 + (y') 2] 3z2
A derivation of this formula is outlined in Problem 52. We will use this formula to
show that a circle has constant curvature.
EXAM PLE 5 Curvature of a circle
Show that a circle of radius r has curvature κ = y at each point.
Solution Suppose a circle has radius r and center (∕z, k). Then its Cartesian
equation is (x - h)2 + (y - k)2 = r2 , and we can obtain a parametrization of
the circle by setting
x - h = r cos t and y - k = r sin t
In other words, the circle is the graph of the vector function
R(z) = x(Z)i + y(Z)j, where
x(t) = h + rc o st and y(t) = k + r sint
Differentiating with respect to t, we obtain
x'(Z) = -r s in Z y'(Z) = rcosZ
x"(Z) = - r cost y"(t) = - r sin I
and by substituting these into the curvature formula, we find:
∣x'y" - y 'x " ∖ _ (- r sin t)(- r sin Z) - (r cos t)(- r cos t)
κ [(x') 2 + ( ∕ ) 2] 3z2 [(-r sin t)2 + (r cos Z)2] 3z2
_ r2sin 2 1 + r2cos2 1 _ r2 _ 1
[r2(sin2 Z + cos2 Z)]3z2 r3 r
If κ Ψ 0 at the point P on the graph of R(Z), the reciprocal p = -κ is called the

radius of curvature at P. The osculating (kissing) circle at P is the circle in the plane
of T and N whose shape best approximates that of the graph of R(Z) in the vicinity
of P (see Figure 11.20). This follows from the definition, which says that the
osculating circle is the circle that is tangent to the graph at P and whose radius is
the radius of curvature p = l∕κ.
Osculating circle Osculating circle Osculating circle
Note that a gentle curve has a

small κând therefore a large 1.
Figure 11.20 The osculating circle best fits the shape of the curve at P.
In the special case where the graph of R(l) is a circle C, the radius of curvature is
the radius of C, and the osculating circle coincides with C.
We conclude this section by providing a summary of the various curvature
formulas (Table 11.2).
TABLE 11.2 Curvature Formulas

Type Given Information Formula to Use Where Derived
Arc length parameter R(5)

llfll Page 715
l∣R '× R"ll

Vector derivative form R(r) l∣R'll3 Page 717
l∣v × A ∣
Velocity/acceleration form position R(∕) Problem 51
l∣v∣∣3
∣Λ->" - y , x " ∖
Parametric equation form x = x(Z), y = y(t) Problem 52
[(Λ-')2 + ( ∕) 2]3'2
l/'l
Functional form y = ∕M Problem 53
(1 + y ' 2 ) 3' 2
∣r2 + 2 r '2- rr"∖
Polar form r = f(θ) Problem 54
(r 2 + r ' 2) 3'2
PROBLEM SET 1 1 .4
θ In Problems 1-8, fin d the unit tangent vector T(r) and the In Problems 9-14, sketch the given curve and fin d its length
principal unit normal vector N (∕) for the curve given by R(Z). over the given interval.
1. R(∕) = t 2i + Z3j, t ≠ 0 2. R(r) = r 2 i + V t j, t > 0 9. R(Z) = 2zi + Zj, over [0, 4]
3. R(∕) = (<?'cos r)i + (e⅛in ∕)j 4. R(r) = (Z cos Z)i + (Z sin Z)j 10. R(Z) = Zi + 3rj, over [0, 4]
5. R(Z) = (cos Z)i + (sin Z)j + Zk 11. R(Z) = 3Zi + 3 co st j + 3 sinZ k, over [θ, 5]
6. R(Z) = (sinZ)i + (cosZ)j + ∕k 12. R(Z) = Zi + 2zj + 3rk, over [0, 2]
7. R(Z) = (In Z)i + Z2 k 13. R(Z) = 4 cos ri + 4 sin Zj + 5rk, over [0, π]
8. R(Z) = (e~zsinZ)i + e~'j + (e - z cosz)k 14. R(Z) = cos 3 Z i + cos 2 Z k, over [0, 2τr]
Find the curvature o f the plane curves at the points indicated in 39. R(Z) = Zi + Z2j + Z3k (velocity/acceleration form)
Problems 15-26. 40. R(Z) = (Z —cosZ)i + (sinZ)j + 3k (velocity/acceleration form)
15. y = 4x - 2 at x = 2 16. y = m x + b at x = a 41. y = x 2 (functional form)
17. y = x - ∣ x 2 , at x = 3 18. y = 2x 2 + 1, at x = 1 42. y = x 3 (functional form)
2
19. y = ax + bx, at x = c 20. y = x + x - l , at x = 1 43. y = x ~ ', x > 0 (functional form)
2
21. y = V 4 - x . at x = 1 22. y = V r 2 —x 2, at x = 0 44. y = sinx (functional form)
23. y = sinx, at x = f 24. y = cosx, at x = J 45. The spiral r = e θ (polar form)
25. y = In x , at x = 1 26. y = ex , at x = 0 46. The cardioid r = 1 + cos θ, for 0 ≤ θ ≤ 2ττ (polar form)
O 27. Let u and v be constant, nonzero vectors. Show that the 47. ■ What Does This Say? Describe what is meant by
line given by R(Z) = u + vt has curvature 0 at each point. using arc length as a parameter.
28. Find the unit tangent T and principal normal N for 48. ■ What Does This Say? Discuss the various curvature
R(Z) = (coshZ)i + (sinhz)j at the point where t = 0. formulas given in Table 11.2.
29. Find the unit tangent T and principal normal N for
R(Z) = [ln(sin Z)]i + [ln(cosZ)]j at the point where t = j .
30. A curve C in the plane is given parametrically by x = 32Z,
y = 16Z2 —4.
a. Sketch the graph of the curve. Computational Window
.. , L⅛
⅛⅛¾⅛
∙g⅝
⅜⅜^<∙'t⅛ ⅛'⅜4⅛+⅜^⅜⅛.,.'∙il⅜
∙L¾⅛ ⅜⅜r⅝
∣
b. Find the unit tangent vector when Z = 3.
c. Find the radius o f curvature of the point P on C where 49. Projections onto a plane in R 3
Z = 3. A very important notion in applied mathematics is
that of projecting a vector onto a subspace of some
31. For the curve given by
kind, for example, a plane. Later in your career, the
R(Z) = (sinZ)i + (cosZ)j + Zk “ vectors” can also be functions in which a difficult
a. Find a unit tangent vector T at the point on the curve function is approximated by “ nice” functions. In this
where Z = ττ. problem, our work is limited to R 3 .
b. Find the curvature when Z = ττ. a. Suppose that u and v are unit vectors and that they
are orthogonal. Show that the “ linear combina
c. Find the length o f the curve from Z = 0 to Z = ττ.
tion” au + ⅛v generate a plane, through (0, 0, 0), as
32. Let C be the curve given by a and b take on all real values.
R(Z) = (Z — sinZ)i + ( 1 - cosZ)j + (4 sin∣)k b. Let w be a third vector, not the plane described in
part a. We want to project w onto the plane, so we
a. Find the unit tangent vector T(Z) to C.
write w = p + n = (αu + ⅛v) + n, where p is the
b. Find and the curvature κ(Z). desired projection and n is the normal component
ds v ’
of w. To find a, take the dot product o f w and u
33. Find the point (or points) where the ellipse and show that p is orthogonal to n (hence p is the
9x 2 + 4y2 = 36 has maximum curvature. orthogonal projection).
34. Find the point (or points) where the curve y = e 2x has c. For the three following vectors, find the orthogo
maximum curvature. nal projection o f w onto the plane generated by u
35. Find the radius o f curvature at each relative extremum of and v:
the graph o f y = x 6 —3x 2 .
“ = ⅛ i + j + k)
36. Find the curvature of the curve given by x = Z — sin Z,
y = 1 - cos Z. Sketch the curve on 0 ≤ Z ≤ 2τr and
sketch the osculating circle at the point where Z = J . v = ⅛ (2 i - j - k)
Vo
37. The tangent line at a point P on a curve C in space is the
w = j + 2k
line that passes through P and is aligned with the tangent
vector T to C at P. Find parametric equations for the d. A projection is especially simple if we have u = i
tangent line to R(Z) = 2zi - Zj + Z2k at P(2, - 1 , 1). and v = j. Explain the situation here for an
38. The tangent line at a point P on a curve C in space is the arbitrary vector w = ai + bj + ck.
line that passes through P and is aligned with the tangent e. Show that for a given vector w = ai + bj + ck, its
vector T to C at P. Find parametric equations for the c
slope relative to the xy-plane is m = , 2. =∙ 2
tangent line to R(Z) = e ti - 3j + (1 — Z)k at F ( l, - 3 , 1). ∖a + b
50. This problem will make some use o f Problem 49 and
There are many different formulas for finding curvature, as
addresses the arc length o f a curve in R 3 and the work
summarized in Table 11.2. In Problems 39-46, fin d the curva
involved in moving along its path. Suppose that a
ture o f each o f the given curves using the indicated formula.
Sec. 11.5 Tangential and Normal Components of Acceleration 721
54. Let f be a twice differentiable function. Show that, in

mining company has built a road up a hill that
polar coordinates, the curvature of the curve given by
roughly satisfies the following parametric definition
r — f(θ) satisfies
(0 ≤ t ≤ 2π).
x(t) — e~" i cos 3/

y(∕) = e - 'z3 sin 3/
Hint: Use the formula in Problem 51 with
x = /(θ) cos θ ,y = f(θ) sin θ.
55. Use the formula in Problem 51 to show that the curve C
a. Generate a rough graph of this curve; then some given by R(t) has curvature
how (use your imagination!) get a decent estimate
[∣
∣
V∣∣
2∣
∣
A∣∣
2 —V ∙ A ] 1/2
of its length, say within 10%.
b. Use your computer to compute its length; then l∣
v∣∣
3
check with your previous estimate.

56. Let P(a, b) be a point on the graph C o f the vector
c. Let T(∕) denote the tangent to the curve at that
function R(t).
point and assume the force
a. Describe a general procedure for finding an equation
___________ 8z'(∕) for the osculating circle to C at P.
F (° " V[x'(Z)]2 + [√(r)]2
b. Find an equation for the osculating circle at the point
P(32, 12) on the curve C defined by
Compute the work required to move a cart from
the bottom to the top o f the hill. (Recall, x = 32/, y = 16∕2 - 4
work = ∣F ds).
57. If T and N are the unit tangent and normal vectors,
respectively, on the trajectory of a moving body, then the
cross-product vector B = T × N is called the unit
θ 51. One o f the most fundamental forms for curvature is the
binormal o f the trajectory. Three planes determined by T,
velocity/acceleration form. Let C be the trajectory of an
N, and B are shown in Figure 11.21.
object, where position is given by the vector function R(∕).
Show that the curvature o f C is given by
_ IIV × All
l∣
v∣
∣3
Hint: V =⅛ and A = + T⅛
at dt dt dι i
52. Use the formula in Problem 51 to verify the formula
_ ∣x'√' - y'x''∣
κ ~ [(x') 2 + ( ∕ ) 2] 3z2 plane
Jo
Hint: Let R(Z) = x(∕)i + y(∕)j. a. Show that T is orthogonal to ⅛ -.
Hint: Differentiate B ∙ T. 5
53. A curve in the plane is given by y = ∕( x ) . Show that the
functional form for curvature is given by b. Show that B is orthogonal to ⅛ -.
Hint: Differentiate B ∙ B. 5
IΓ(-v)l J D
{i + [ ∕'( χ ) ] 2}3z2 c. Show that ⅛ - = —τN for some constant τ. Note: τ is

ds
called the torsion o f the trajectory. It provides a
Hint: Use the formula given in Problem 51, with measure of the amount o f twisting at each point on the
R(,v) = xi + ∕(x ) j. trajectory.
1 1 . 5 TANGENTIAL A N D NO RM AL C O M PO NEN TS OF ACCELERATION
IN TH IS SECTIO N Components of acceleration, applications

We see how the velocity and acceleration vectors of a moving body can be
decomposed into tangential and normal components. Then we examine a few
applications in which this decomposition is especially useful.
■ COMPONENTS OF ACCELERATION
When a body is caused to accelerate or brakes are applied, it is of interest to know

how much of the acceleration acts in the direction of the body’s motion, as
indicated by the unit tangent vector T. This question is answered by the following
theorem.
THEOREM 11.8 Tangential and normal components of acceleration
An object moving along a smooth curve (with T' ≠ 0) has velocity V and
acceleration A , where
v ≈ (⅞ )τ and A = ⅛ )τ + X(⅛)N
and 5 is the arc length along the trajectory.
Proof: An object moving with displacement R has unit tangent T = and
unit normal N = ± We use the chain rule to write the velocity vector V as
follows:
y = dR = ds - y ds
dt ds dt 1 dt
lr^Γ'
Differentiate both sides of this equation with respect to t and substitute =KN:
Definition o f A
_ -≤L∣^τ d s ∖ v ■⅛)τ
d t[ l dt]
_ d 2s τ ds d rT Product rule
dt2 dt dt
_ d 2s . + ⅛Γ√T d s ∖
Chain rule
dt2 ^t ^ dt\_ ds dt]
_ d 2s τ + 7ds∖2 dT
dt2 ∖dtj ds
- S T + ⅛)⅛N> Since KN =
ds
I What this says: At each point on the trajectory of a moving object, the
velocity V points in the direction of the unit tangent T, but the acceleration A
has both a tangential and normal component. The trajectory may twist and
turn, but the acceleration is always in the plane determined by T and the unit
normal N.
The two components of acceleration have special names.

Tangential and Normal The acceleration A of a moving object can be written as

Components of Acceleration
A = A-f T + /1ΛN
where
/72 r>
Aτ = is the tangential component.
AΛ, = κl I is the normal component.
Figure 11.22 Components of velocity and acceleration
It is usually fairly easy to find A τ , but it may be more difficult to find AN ,

because computing the curvature κ is often a messy process. Fortunately, there is a
way to compute AN without first finding κ. We expand the dot product of
A = A 1-T + AyvN with itself and use the fact that T and N are orthogonal unit
vectors:
∣
∣
A∣∣
2 = A ∙ A Theorem 11.8 A = A rT + Av N
= (Ar τ + ΛJVN ) ∙ (Aτ τ + AΛ,N )
= A 2 (T ∙ T) + 2Λr Λ √ N ∙ T) + Λ 2 (N ∙ N)
= Λ2 (1) + 2AΓ TΛ<0) + Λ 2 (1) T ∙ T = 1, N ∙ T = 0, andN ∙ N = 1
= A~+ A?
Thus, once we know A and A τ , we can compute AN by applying the following
formula.
Computation of Normal The normal component AN can be found using the formula
Component A∣
AN = √I∣∣
2 - A2
Γ
EXAM PLE 1 Finding tangential and normal components of acceleration
Find the tangential and normal components of an object that moves with
displacement
R(Z) = t3i + Z2j + Zk
Solution V= = 3z2i + 2zj + k and A = = 6zi + 2j
= ∣
∣
V∣∣= V(3Z2)2 + (2∕)2 + (1)2 = √9z 4 + 4t2 + 1
A = ⅞ = j ( ’ ' 4 ÷ -*<≈ + l)-'=(36<5 + 8<) = √ ⅛ ⅞ ⅛

This is the tangential component of acceleration.

[√ 3 6 ∕2 + 4] 2 - I 18∕3 + 4 ∕ 12
4 v = V∣∣A∣∣2 - Λ r
2 = L√ 9∕4 + 4 ∕ 2 + 1J
4 ∕ 2(9 ∕2 + 2)2
9 ∕4 + 4 ∕2 + 1
∣36t4 + 3 6 ∕2 +~4 There are several
V 9 ∕4 + 4 ∕2 + 1 simplification steps
that are not shown.
9 9t4 + 9 ∕ 2 + 1
V 9 ∕4 + 4 ∕ 2 + 1
This is the normal component of acceleration.
EXAM PLE 2 Finding tangential and normal components on a helix
An object moves along the helix with position vector

R(∕) = (cos ∕)i + (sin ∕)j + ∕k
Find the tangential and normal components of acceleration.
Solution V = = ( -s in ∕) i + (cos∕)j + k and A = = (-c o s ∕) i + ( -s in ∕) j
= IM∣ = V sin 2 ∕ + cos2 / + 1 = V z2 and A τ = = 0
AN —V∣∣A∣∣2 - A τ 2 = √(Vcos2 ∕ + sin 2 /)2 - 0 2 = 1

The tangential and normal components of acceleration are 0 and 1, respective
ly. This means that the acceleration satisfies
A = A r T + Λ y N = (0)T + (1)N = N
That is, the acceleration vector is the principal unit normal N , and the
acceleration is always normal to the trajectory of the helix.
■ APPLICATIONS
Now that we know how to compute the tangential and normal components of
acceleration, A τ and AN , we shall examine some applications. First, according to
Newton’s second law of motion, the total force acting on a moving object of mass
m satisfies F = mA, where A is the acceleration of the object. Because A = √4r T
+ -4Λ∙N, we have
F = »iA = (mΛr )T + (m^λ,)N = F 7T + Fv N
where
F τ = m ~ 4 and F = mκj ⅜ ]
τ dtλ n ∖atj
For instance, experience leads us to expect a car to skid if it makes a sharp turn
at moderate speed or even a gradual turn at high speed. Mathematically, a “ sharp
turn” occurs when the radius of curvature p = l ∕κ is small (that is, when κ is large),
and “ high speed” means that ds/dt is large. In either case,
wκ (l)^
Figure 11.23 Tendency to skid
will be relatively large, and the car will stay on the road (its trajectory) only if there
is a correspondingly large frictional force between the tires of the car and the
surface of the road (see Figure 11.23).
When forces, masses, and accelerations are involved, it is customary to
express the mass m in slugs of an object whose weight W has been given in
pounds.* From the definition of weight,
where g is the acceleration of gravity (g ≈ 32.2 ft∕s 2)
EXAM PLE 3 Tendency of a vehicle to skid
A car weighing 2,700 lb makes a turn on a flat road while traveling at 56 ft/s
(about 38 mi∕hr). If the radius of the turn is 21 ft, how much frictional force is
required to keep the car from skidding?
∕√ s,y
Solution The required frictional force is F N = m κ γ j t ∖ , where m is the mass of
the car and κ is the curvature of the road. We know that
⅛ - 56 ft/s
and because the car weighs W = 2,700 lb, its mass is m = ~~ = ry r v ≈ 83.85
slugs. Because the turn radius is 21 ft, we have κ = so that
/2,700 s .l u gJ ∖ ∕ 1 V 5g 46 ft∖2 _ 19 t∙ 9 0 lb ∙ s2 1 ft 2 _ 1 9 4r9 9 ιr

rN lv 32.2 A21 f t Λ s) 12,522 f t f t g2 12,522 1b
There are certain important applications in which an object moves along its
trajectory with constant speed ds∕dt, and when this occurs, the acceleration A can
have only a normal component, because d 2s∕dt2 = 0.
THEOREM 11.9 Acceleration of an object with constant speed
The acceleration of an object moving with constant speed is always orthogonal to

the direction of motion.
Proof: Notice that we really do not need to know anything about components
of acceleration to prove this result. Saying that the object has constant speed
means that ∣∣R'(t)∣∣ is constant, and by the orthogonality of a derivative theorem
(Theorem 11.4), we conclude that R'(t) is orthogonal to its derivative R"(t) = A(Z).
But R'(t) points in the direction of the object’s motion along its trajectory, which
means that the acceleration A is orthogonal to the direction of motion. ■
As an application of this result, note that when an object moves with constant
speed v0 along a circular path of radius R (so that κ = 1∕R), its acceleration is
directed toward the center of the path and has magnitude
of motion
∣2 = 1 v 2
If the speed is vθ and the path has 1 R vo
radius R, the acceleration is
(See Figure 11.24).
*A slug is a unit of measurement defined as the unit of mass that receives an acceleration of 1 ft∕s2
Figure 11.24 An object moving when a force of 1 lb is applied to it. That is,
with constant speed on a circular 1 slug = 1 lb = lb ■s 2
path 1 ft∕s 2 ft
EXAM PLE 4 Time for a satellite to orbit
An artificial satellite travels at constant speed in a stable circular orbit 20,000

km above the earth’s surface. How long does it take for the satellite to make
one complete circuit of the earth?
Solution Let m denote the satellite’s mass and v denote its speed. We shall
assume that the earth is a sphere of radius 6,440 km, so the curvature of the
path is κ = 1∕Λ, where
Height
R = 6,440 + 20,000 = 26,440 km
Radius of earth
is the distance of the satellite from the center of the earth. The centripetal
∣
{center-seeking) force F on the object satisfies ∣ = mAN = mv 2∕R. This force
F [∣
is balanced by a centrifugal {center-fleeing) force Fc, and according to
Newton’s law of universal gravitation, ∣Fc ∣= GmM∕R∖ where M is the mass of
Figure 11.25 A satellite in a stable the earth and G is the gravitational constant. It can be shown that the product
orbit; centrifugal force F r equals GM is approximately equal to
the force F due to gravity. GM = 398,600 km 3∕s 2
Since the centrifugal and centripetal forces on the object must be equal for
it to maintain its stable orbit, we have
l∣
FJI _ _ GmM _ ∣
1FJ∣
Solving for v, we obtain
v ^ λ ^ F ∙ √ ¾ w " 3 -8s273653
Finally, suppose that T is the time required for the satellite to make one
complete circuit of the earth (called the period of the satellite). In each period,
the satellite travels a distance equal to the circumference of a circle of radius
R = 26,440 km, and because it travels at v km∕s, we must have vT = 2ττR, so
that
T = 2 jτ^ = 2 ⅞ 6247306 k5 ^ ≈ 4 2 '7 8 6 ∙1 6 8 5 2
V 3 .o o 2 / 5b J 5
or approximately 713 minutes (11 hr 53 min).
PROBLEM SET 1 1 .5
O /« Problems 1-8, an object moves with displacement R{t). Find tion, and the normal and tangential components of accel
the tangential and normal components o f the object's accelera eration of the object at time Z.
tion.
1. R(Z) = fi + Z2j 2. R(Z) = t∖ + e lj 10. An object moves with constant angular velocity ω around
3. R(∕) = (t sin Z)i + {t cos Z)j the circle x 2 + y 2 = r2 in the xy-plane. The displacement
vector is
4. R(Z) = (3 cos Z)i + (2 sin Z)j
5. R{t) = ti + Z2j + Zk
R(Z) = {r cos ωZ)i + {r sin ωZ)j
6. R(Z) = (sin Z)i + (cos Z)j + (sin Z)k
7. R(Z) = (4 cos Z)i + (sin Z)k a. Find the tangential and normal components of acceler
8. R(Z) = (∙j⅜ cosz)i + γj(l - cosZ)j + sinZk ation.
© 9. The position of an object at time t is {x, y), where b. Show that the curvature is κ = 1/r at each point on the
x = 1 + cos2z, y = sin2z. Find the velocity, the accelera circle.
Sec. 11.5 Tangential and Normal Components of Acceleration 7 27
11. Find the tangential and normal components of the accel 17. A Ferris wheel with radius 15 ft rotates in a vertical plane
eration of an object that moves along the parabolic path at ω revolutions per minute. What is the maximum value
y = 4x 2 at the instant the speed is ds/dt = 20. of ω for a Ferris wheel carrying a person of weight W r>
.
12. A pail attached to a rope 1 yd long is twirled at the rate of
1 rev∕s. Find the tangential and normal components of the
pail’s acceleration. Hinf. Assume the rope is twirled from
a position that only rotates.
13. A boy holds onto a pail of water weighing 3 lb and swings
it in a vertical circle with a radius of 2 ft (the length of his
arm). If the pail travels at ω rpm, what is the pressure of
the water on the bottom of the pail when it is at its highest
point? What is the smallest value of ω required to keep the
water from spilling from the pail? Assume the pail is held
by a handle so that its bottom is straight up when it is at
its highest point.
14. A car weighing 2,700 lb ≈ 1.35 tons moves along the Use the formula in Problem 24 to find A τ and A N for the given
elliptic path 9x 2 + 4y 2 = 36 (units in kilometers). If the position vector R(t) in Problems 18-21.
car travels at the constant speed 100 km/h, how much 18. R(∕) = ∕ 3i + ∕ 2j + ∕k
frictional force is required to keep it from skidding as it
19. R(Z) = ti + 2/j + Z2k
turns the “corner” at (2, 0)? What about the corner (0, 3)?
20. R(/) = (cos ∕)i + (sin t)j + k
15. A curve on a road should be banked so that there is no
lateral friction force exerted on the tires as the vehicle 21. R(∕) = (e z cos∕)i + (el sin∕)j + e z k
rounds the curve. Suppose the road is a level curve that is θ 22. Let T be the unit tangent vector, N be the principal unit
a circle of radius 200 ft, and the average speed on this normal, and B be the unit binormal vector (see Problem
curve is 55 mi∕hr. Also assume that the vehicle weighs 57, Section 11.4) to a given curve C. Show
3,500 lb.
dB = τ, y z/N
ds ds
23. An amusement park ride consists of a large (25-ft radius),
flat wheel on which volunteers pay to sit while it is at rest.
The wheel then begins to rotate until all the passengers are
thrown from the wheel. The purpose of this problem is to
discover how fast the wheel can rotate without losing its
passengers.
a. Suppose the wheel rotates at ω revolutions per minute
a. Find the force between the tires and the road so that and that a volunteer weighing Hz pounds stands at the
the car stays on the circular path and does not skid. rim, 15 feet from the center of the wheel. Find the
Hint'. Use F = mA, where m — 3500/32.2. Also, frictional force mA, required to keep the volunteer
1 mi = 5,280 ft. from sliding off, where m = W/g is the volunteer’s
b. Find the angle the roadway should be banked so that mass.
no lateral friction force is exerted on the tires of the
car.
16. What is the smallest radius that should be used for a
circular highway curve if the normal component of the
acceleration of a car traveling at 45 mi/h is not to exceed
2.4 ft/s2?
b. If the frictional force of the volunteer in part a needed

to stay on the wheel is 0.12 PF, find the largest value of
ω that will allow the volunteer to stay in place. Hint:
Set 0.12 IP'equal to your answer in part a and solve for
ω', also note that the weight of the volunteer is irrele
vant.
24. Suppose the position vector of a moving object is R(t). 29. An artificial satellite travels at constant speed in a stable
Show that the tangential and normal components of the circular orbit R km above the earth’s surface (as in
object’s acceleration may be computed by the formulas Example 4).
a. Show that the satellite’s period is given by the formula
= IL LT ∣
∣ ∣
R ' × R "∣
A
Λ T ∣
∣
R'∣
∣ aAND
nα λA " = ∣
∣ ∣
R'∣ (A + R ) 3/2
T = 2ττ----------—
25. Suppose an object moves in the plane along the curve
y = f(x). Use the formulas obtained in Theorem 11.8 to where R e is the radius of the earth.
show that b. A satellite is said to be in geosynchronous orbit if it
completes one orbit every sidereal day (23 hr, 56 min).
∕W "W J∕"W∣ How high above the earth should the satellite be in
τ n
Vι + [∕ W V I +[∕'(χ)]2 order to achieve such an orbit? Assume the radius of
the earth is approximately 6,440 km.
26. If the graph of the function of f has an inflection point at 30. SPY PROBLEM After finally disposing of Boldfinger
x = a and f" (a ) exists, show that the graph of f has (Problem 48, Chapter 10 Supplementary Problems), the
curvature 0 at (α.∕(α)). spy finds himself in hot pursuit of a gang of international
27. A projectile is fired from ground level with an angle of terrorists who are holding his superior’s innocent daugh
elevation a and muzzle velocity vθ. Find formulas for the ter prisoner. The terrorists are hiding in a bunker on the
tangential and normal components of the projectile’s side of a hill inclined at an angle of 15o to the horizontal.
acceleration at time t. What are A τ and AN at the time the The spy flies toward the bunker in a helicopter traveling at
projectile is at its maximum height? 200 ft/s at an altitude of 10,000 ft. Just as the helicopter
flies over the base of the hill, the spy sights a ventilation
28. An object connected to a string of length r is spun
hole in the top of the bunker at an angle of 20°. He decides
counterclockwise in a circular path in a horizontal plane.
to drop a canister of knockout gas into the bunker. How
Let ω be the constant angular velocity of the object.
long should he wait before releasing the canister?
a. Show that the displacement vector of the object is
R(∕) = (r cos ωZ)i + (r sin ωZ)j
b. Find the normal component of the object’s accelera
tion.
c. If the angular velocity ω is doubled, what is the effect
on the normal component of acceleration? What is the
effect on A if ω is unchanged but the length of the
string is doubled? What happens to AN if ω is doubled
and r is halved?
C H A P T E R 11 R E V IEW
IN TH IS R E V IE W Proficiency examination; supplementary problems

TABLE 11.3 Summary of Velocity, Acceleration, and Curvature

CURVES
Position vector R 2 (plane) R(t) = x(t)i + y(∕)j

R 3 (space) R(t) = x(t)i + y(t)j + z(t)k
The graph is called the trajectory o f the object’s motion.
Velocity vector V(t) = ^ = R'(t)
∣
∣
V∣∣= ⅛ This is called the speed.
This is the direction o f V or direction o f motion.
Acceleration vector W = ^ = -C ^ = A T^ +
where
d^^s Vr * A
Aτ = = A ■T = ∣∣
y∣ ∣ This is the tangential component.
f v
= κ (⅛ ) = A ■ =N v ∣
∣∣
a ∣- V
2
This is the normal component.
Unit tangent τ _ v _ R '(9 _ t/R

l∣
v∣∣ ∣
∣
R '(∕) ∣
∣ ds
and normal vectors
This is the unit tangent vector in the direction
o f motion.
rJ^, , /∕∖ ∣ ∕ rI’
N = ., = - , T his is the unit normal vector,
ι∣
τ ,ω ∣ ι * *
ll<zrll ⅛ I∣V × A ∣∣. k>" - A "l
Curvature κ k ds , ∣
∣
v∣∣
3 lf>
t (√)2 +
( v ')2]3∕2
dt
Torsion
-Ilf 1
Concept Problems 9. State the orthogonality of a derivative theorem.

1. What is a vector-valued function? 10. What are the displacement, velocity, and acceleration
2. What are the components of a vector function? vectors?
3. Describe what is meant by the graph of a vector func 11. What is the speed of a particle moving on a curve C?
tion. 12. What are the formulas for the motion of a projectile in a
4. Define the limit of a vector function. vacuum?
5. Define the derivative of a vector function. 13. What are the formulas for time of flight and range of a
6. Define the integral of a vector function. projectile?
7. What is a smooth curve? 14. State Kepler’s laws.
8. Complete the following rules for differentiating vector 15. What are the unit polar vectors, ur and ufJ?
functions: 16. What are the unit tangent and normal vectors?
a. Linearity rule 17. What is speed in terms of arc length?
b. Scalar rule 18. What is the arc length of a space curve?
c. Dot product rule 19. What is the curvature of a graph?
d. Cross-product rule 20. What is the formula for curvature in terms of the
e. Chain rule velocity and acceleration vectors?
21. What are the formulas for unit tangent and principal 27. Find a vector function F such that F"(Z) = e ti — Z2j + 3k
unit normal in terms o f the arc length parameter? and F(0) = i - 2j, F'(0) = 3k.
22. W hat is the radius o f curvature? 28. Find the velocity V , the speed ds∕dt, and the acceleration
23. What are the tangential and normal components o f A for the body with displacement vector
acceleration? R(Z) = t∖ + 2zj + te l k.
29. Find T , N , the tangential and normal components o f
Practice Problems acceleration A τ , and A N ), and κ (curvature) for an object
24. Sketch the graph o f R(Z) = (3cosZ)i + (3 sinZ)j + Zkand with displacement R(Z) = Z2i + 3Zj - 3zk.
find the length o f this curve from t = 0 to t = 2π.
30. A projectile is fired from ground level with initial
25. If F(Z) = ( j ψ ~ j i + ( 5 y 1 ) j + (∞sZ)k, find F'(Z) and
velocity 50 ft/s at an angle o f elevation o f a.
a. I f a = 30°, what is the maximum height reached by
F "(0 .
the projectile?
26. Evaluate [3zi + 3k] × [(In Z)j - Z2k] dt. b. I f a = 30°, what is the time o f flight and the range?
F in d the vector lim its in Problems 1-6.
27. J [te , i — (sin2z)j + Z2k] dt
1. lim [z 2i - 3Zczj + ≤ ⅜ ¾ 1 2. lim Γi +-z∙ i ÷ f ' k l
∣→o L t J t→o L 1- 1 J
3. lim [Zi + 5k] ∙ [(sinZ)i + 3zj - (1 - Z)k] 28. ∣
e 2 , [2i - Zj + (sin Z)k] dt
∕ →0
4. lim [(1 + Z)i - 3j] × [Z2j + (cos πZ)k]
5. lim Γ f 1 + γ Y i - - Zkl t[e , ∖ + (In Z)j + 3k] dt
r→0 L∖ t) ∖ t ) J
6. lim f- 1-~ c o s z ∖ + 4j + f l + - k]
[ezi + 2j - Zk] ∙ [e zi — zj] dt
j
t →o L∖ Z J ∖ t) J
F in d F'(t) and F " (t) fo r the vector functions in Problems 7-12. In Problems 3 1-34, R is the displacement vector o f a moving
7. F(Z) = te t ∖ + Z2j 8. F(Z) = (Z In 2z)i + Z3z2k body. F in d the velocity V , the speed ds!dt, and the accelera
1 , tion A .
9. F(Z) = 2z~ i - 2zj + te~ k
31. R(z) = zi + (3 - z)j + 2k
10. F(Z) = [zi - (1 - Z)j] × [Z2j + e^ zk]
32. R(Z) = (sin2z)i + 2j - (cos2z)k
11. F(Z) = (Z2 + e a t )i + (Ze- °z)j + (e°z + l )k, a a constant
33. R(z) = (z sinZ)i + (Ze^z)j - ( 1 - z)k
12. F(Z) = ( 1 - Z)^1i + (sin2z)j + (cos2 Z)k
34. R(Z) = (In Z)i + e zj - (tanZ)k
Sketch the graph o f the vector functions in Problems 13-18.
F in d rT a n d 1S f o r the vector functions given in Problems 35-38.
13. F(Z) = te t ∖ + Z2j 14. F(Z) = Z2i - 3zj
35. R(Z) = Zi - Zj
15. F(Z) = 2zi + (1 - Z)j + Z2k
36. R(Z) = (3 cos Z)i - (3 sin Z)j
16. F(Z) = ( 1 - cosZ)i + (sinZ)k
37. R(Z) = (4 cos Z)i — 3zj + (4 sin Z)k
17. F(Z) = (3 cosZ)i + (3 sinZ)j + Zk
38. R(Z) = (ez sin Z)i + e lj + (e , cos Z)k
18. F(Z) = (2 sinZ)i + (2 cosZ)j + 5Zk
In Problems 3 9 -4 2, fin d the tangential and norm al compo
ΛetF(Z) = ∕ 1(Z)i + f f f ) j + f f f ) k and fin d the indicated deriva
nents o f acceleration, and the curvature o f a moving object with
tive in terms o f V in Problems 19-24. displacement R(Z).
19. ⅛∣
F(Z)∣
∣ 20. ^ [ ∣
∣
F(z)∣
∣
F(z)] 39. R(Z) = Z2 i + 2zj + e l k
40. R(Z) = Z2 i - 2Zj + (Z2 - Z)k
2i∙⅛[∣∣iy ≈∙⅛FW∙FW] 41. R(Z) = (4 sin Z)i + (4 cosZ)j + 4zk
42. R(Z) = (a sin3z)i + (a + a cos3Z)j + (3fl sin Z)k,
23. [F(Z) × F(Z)] 24. ^ F ( e z)
for constant a ≠ 0
Evaluate the definite and indefinite vector integrals in Prob A polar curve C is given by r = f(θ). I f f"(θ ) exists (see Problem
lem s 25-30. 54, Section 11.4), then the curvature can be fo u n d by the
form ula
25. (e^zi + Z3j + 3k) dt 26. [(1 - z)i - z 1j + e z k] dt ∣[∕( g ) ] 2 + 2 [ ∕ , (e)]2 - ∕( g ) ∕" ( 0 l
κ { [ > ) ] 2 + [ ∕'( θ ) ] 2} 3z2
Find the curvature at the given point on each o f the polar curves
43. r = 4 cos θ, where θ = ? 44. r = 02 , where θ = 2
45. r = e~θ, where θ = 1 46. r = 1 + cos θ, where θ = J
47. r = 4 cos 30, where 0 = f
48. r = 1 - 2 sin θ, where θ = f
49. For what values of t is the following vector function
continuous?
F(t) = (2t - l)i + ( 7 Ξ 7 ) j + 4k
50. Find the length of the graph of the vector function

(2t + l) 3z2. 57. The position of an object moving in space is given by
R(t) = 3 j
R(z) = (e^⅛osZ)i + (e - ⅛in∕)j + e~t k
from t = 0 to t = 6.
a. Find the velocity, speed, and acceleration of the object
51. Find a vector function F such that F(0) = F'(0) = i, at arbitrary time t.
F"(0) = 2j + k, and
b. Determine the curvature of the trajectory at time t.
F"'(t) = (cos∕)i + (sin∕)j + ⅛ 58. Find the point or points on the curve y = e ax (a > 0)
where the radius of curvature is maximized.
59. The position vector of an object in space is
52. An object moves in space with acceleration
R(f) = (a cos ωf)i + (a sin ω∕)j + ω2Zk
A(t) = ( - t)i + 2j + (2 - Z)k
Find the object’s tangential and normal components of
When t = 0, it is known that the object is at the point acceleration.
(1, 0, 0) and that it has velocity V(0) = 2i - 4j.
60. A particle moves along a path given in polar coordinates
a. Find the velocity V(Z) and the position R(Z). (r, 0), where r(Z) = 1 + cosat and Θ(f) = e~at (for positive
b. What is the speed and location of the object when constant a). Find the velocity and acceleration of the
t = 1? particle in terms of the unit polar vectors uj, and u,l .
c. When is the object stationary and what is its displace 61. A car weighing 3,000 lb travels at a constant speed of 60
ment at that time? mi/hr on a flat road and then makes a circular turn on an
53. The position vector for a curve is given in terms of arc interchange. If the radius of the turn is 40 ft, what
length s by frictional force is needed to keep the car from skidding?
R(∙s,) = ∖a c θ s ^)> ÷ ( α s *n β)j + 2sk
for 0 ≤ s ≤ 2τra and a > 0. Find the unit tangent

vector T(.sj, and the principal unit normal N(s).
54. Find the radius of curvature of the curve given by
y = 1 + sin x at the points where x is
aa ∙ —
6 bb∙ —
4 c„∙ 3τr
T
55. Find the radius of curvature of the ellipse given by
62. A nozzle discharges a stream of water with an initial
R(Z) = (a cos Z)i + (b sin Z)j velocity of v0 of 50 ft/s into the end of a horizontal pipe of
inside diameter d = 5 ft. What is the largest distance that
where a > 0, b > 0, a ≠ b, and 0 ≤ t ≤ 2ττ at the the stream can reach?
points where t = 0 and τr∕2.
56. A stunt pilot flying at an altitude of 4,000 ft with a speed
of 180 mi/hr drops a weighted marker, attempting to hit a
target on the ground below, as shown in the figure in the
margin. How far away from the target (measured horizon
tally) should the pilot be when he releases the marker?
You may neglect air resistance.
63. Recall that the tangent line at a point P on a curve in space In Problems 69- 70, fin d F '(%).
is the line that passes through P and is aligned with the
69. F(x) = f [(sint)i - (cos2∕)j + e~t k,] dt
tangent vector T at P. Find parametric equations for the
Ji
tangent line to the curve given by
70. F(x) = [ [Z2i + (sece - z )j - (tane 2z)k] dt
R(t) = 1+ t3 ) + ∖1 + t 37j
at the point P(0, 0). Sketch the curve. This curve is called
the folium o f Descartes. In Problems 71-74, solve the initial value problems for F as a
vector function o f t.
71. Differential equation: F'(Z) = ti + Zj - Zk;
initial condition: F(0) = i + 2j - 3k
72. Differential equation:
F'(t) = ∣ V t + 1 i + (f + l)~1j + e , k
initial condition: F(0) = j + 2k

F'(r) = (sin2z)i + (ezcosZ)j - L 1jk
initial condition: F(0) = i - 3k

⅞τ=
1
~32i
dt
initial conditions:
F = 50j and = 5i + 5k at t = 0
A shell strikes an air
plane flying at a height h above the ground. It is known
that the shell was fired from a gun on the ground with
muzzle velocity v0 , but the position of the gun and its
angle of elevation are both unknown. Deduce that the gun
is situated within a circle whose center lies directly below
65. A bicycle travels at the constant speed of 10 m ∕s. Find the the airplane and whose radius is
maximum velocity of a point on the bicycle tire located 25
cm from the axle. Assume the radius o f the tire is greater jV v 2 - 2gh
0
than 25 cm. o
66. A D N A molecule has the shape of a double helix (see Neglect resistance of the atmosphere.
Figure 11.4). The radius of each helix is about 10^8 cm. 76. P U T N A M E X A M IN A T IO N P RO BLEM A coast artillery gun
Each helix rises about 3 × 10^8 cm during each complete can fire at any angle o f elevation between 0° and 90o in a
turn and there are about 3 × 108 complete turns. Estimate fixed vertical plane. If air resistance is neglected and the
the length o f each helix. muzzle velocity is constant (v = v0), determine the set H
67. Show that the tangential component of acceleration of a of points in the plane that can be hit. Consider only those
moving object is 0 if the object has constant speed. Is the points above the horizontal.
converse statement also true? That is, if A τ = 0, can we 77. P U T N A M E X A M IN A T IO N P ROBLEM A particle moves on a
conclude that the speed is constant? circle with center O, starting from rest at a point P and
68. The path of a particle P is an Archimedean spiral. The coming to rest again at a point Q, without coming to rest
motion of the particle is described by the polar coordi at any intermediate point. Prove that the acceleration
nates r = 10z, and θ = 2ττt, where r is expressed in inches vector o f the particle does not vanish at any point between
and t is in seconds. Determine the velocity of the particle P and Q and that, at some point R between P and Q , the
(in terms of ur and uβ) when acceleration vector points in along the radius R O .
a. t = 0 b. t = 0.25 sec
GUEST ESSAY
For Further Study— The Simulation of Science by Howard Eves
The mighty Antaeus was the giant son of Neptune (god of the sea) and Ge
(goddess of the earth), and his strength was invincible so long as he remained in
contact with his Mother Earth. Strangers who came to his country were forced to
wrestle to the death with him, and so it chanced one day that Hercules and
Antaeus came to grips with one another. But Hercules, aware of the source of
Antaeus’ great strength, lifted and held the giant from the earth and crushed him
in the air.
There is a parable here for mathematicians. For just as Antaeus was born of and
nurtured by his Mother Earth, history has shown us that all significant and lasting
mathematics is born of and nurtured by the real world. As in the case of Antaeus,
Howard Eves was born in Paterson, so long as mathematics maintains its contact with the real world, it will remain
New Jersey, in 1911 and is profes
powerful. But should it be lifted too long from the solid ground of its birth into the
sor emeritus of mathematics at the
filmy air of pure abstraction, it runs the risk of weakening. It must of necessity
University of Maine. This guest es
say first appeared in Great M o
return, at least occasionally, to the real world for renewed strength.
ments in Mathematics Before 1650. Such a rejuvenation of mathematics occurred in the seventeenth century,
Professor Eves reminds us, “ It must following discoveries made by two eminent mathematician-scientists—Galileo
be remembered that a moment in
Galilei (1564-1642) and Johannes Kepler (1571-1630). Galileo, through a
history is sometimes an inspired
sequence of experiments started before his 25th birthday, discovered a number of
flash and sometimes an evolution
extending over a long period o f
basic facts concerning the motion of bodies in the earth’s gravitational field, and
time.” Howard Eves' lectures {from Kepler, by 1619, had induced all three of his famous laws of planetary motion.
which this guest essay was taken) These achievements proved to be so influential on the development of so much of
are so renowned that each college subsequent mathematics that they must be ranked as two of the GREAT MOMENTS
and university at which he taught IN MATHEMATICS. Galileo’s discoveries led to the creation of the modern science of
awarded him, at one time or anoth dynamics and Kepler’s, to the creation of modern celestial mechanics; and each of
er, every available honor for distin these studies, in turn, required, for their development, the creation of a new
guished teaching. Howard Eves is a mathematical tool—the calculus—capable of dealing with change, flux, and
prolific and successful textbook au motion.
thor with a real love for the history o f
mathematics. H e now lives with his Galileo was born in Pisa in 1564 as the son of an impoverished Florentine
wife al a retirement retreat in Lubec,nobleman. After a disinterested start as a medical student, Galileo obtained
Maine. parental permission to change his studies to science and mathematics, fields in
which he possessed a strong natural talent. While still a medical student at the
The die is cast-, I have written my University of Pisa, he made his historically famous observation that the great
book-, it will be read either in the pendulous lamp in the cathedral there oscillated to and fro with a period
present age or by posterity, it mat
independent of the size of the arc of oscillation.* Later, he showed that the period
ters not which-, it may well await a
reader, since God has waited six
of a pendulum is also independent of the weight of the pendulum’s bob. When he
thousand years for an interpreter o f was 25, he accepted an appointment as professor of mathematics at the University
his words. of Pisa. It was during this appointment that he is alleged to have performed
Johannes Kepler experiments from the leaning tower of Pisa, showing that, contrary to the teaching
From James R. Newman, The
World o f Mathematics, Volume I (New *This is only approximately true, the approximation being very close in the case of small
York: Simon and Schuster, 1952), p. 220. amplitudes of oscillation.
of Aristotle, heavy bodies do not fall faster than light ones. By rolling balls down
Historical Note
inclined planes, he arrived at the law that the distance a body falls is proportional
to the square of the time of falling, in accordance with the now-familiar formula
5 = ⅛ 2.
Unpleasant local controversies caused Galileo to resign his chair at Pisa in 1591,
and the following year he accepted a professorship in mathematics at the
University of Padua, where there reigned an atmosphere more friendly to
scientific pursuits. Here at Padua, for nearly 18 years, Galileo continued his
experiments and his teaching, achieving a widespread fame. While at Padua, he
heard of the discovery, in about 1607, of the telescope by the Dutch lens-grinder
Johann Lipersheim, and he set about making instruments of his own, producing a
telescope with a magnifying power of more than 30 diameters. With this telescope
he observed sunspots (contradicting Aristotle’s assertion that the sun is without
blemish), saw mountains on the moon, and noticed the phases of Venus, Saturn’s
JOHANNES KEPLER rings, and the four bright satellites of Jupiter (all three of these lending credence to
(1571-1630) the Copernican theory of the solar system). Galileo’s discoveries roused the
Copernicus revolutionized the opposition of the Church, and finally, in the year 1633, he was summoned to
world with his heliocentric the appear before the Inquisition and there forced to recant his scientific findings. Not
ory, and Galileo confirmed this many years later the great scientist became blind. He died, a prisoner in his own
theory with his telescope, but home, in 1642, the year Isaac Newton was born.
Johannes Kepler is remembered
not only for his work in astron Johannes Kepler was born near Stuttgart, Germany in 1571 and commenced his
omy, but also for his mathe studies at the University of Tubingen with the intention of becoming a Lutheran
matical discoveries. minister. Like Galileo, he found his first choice of an occupation far less congenial
than his deep interest in science, particularly astronomy, and he accordingly
Kepler has been described as changed his plans. In 1594, when in his early twenties, he accepted a lectureship at
“number-intoxicated.” He was
the University of Gratz in Austria. Five years later he became assistant to the
looking for mathematical har
monies in the physical universe.
famous Danish-Swedish astronomer Tycho Brahe, who had moved to Prague to
He is quoted in The World of serve as the court astronomer to Kaiser Rudolph II. Shortly after, in 1601, Brahe
Mathematics'. “Nothing holds suddenly died, and Kepler inherited both his master’s position and his vast
me; I will indulge my sacred collection of very accurate data on the positions of the planets as they moved
fury; I will triumph over man about the sky. With amazing perseverance, Kepler set out to find, from Brahe’s
kind by the honest confession enormous mass of observational data, just how the planets move in space.
that I have stolen the golden
vases of the Egyptians to build It has often been remarked that almost any problem can be solved if one but
up a tabernacle for my God far continuously worries over it and works at it a sufficiently long time. As Thomas
away from the confines of Edison said, genius is 1% inspiration and 99% perspiration. Perhaps nowhere in
Egypt.” the history of science is this more clearly demonstrated than in Kepler’s incredible
pertinacity in solving the problem of the motion of the planets about the sun.
The three laws of Kepler de
scribed in this chapter forever
Thoroughly convinced of the Copernican theory that the planets revolve in orbits
changed the history of science. about the central sun, Kepler strenuously sought to determine the nature and
position of those orbits and the manner in which the planets travel in their orbits.
With Brahe’s great set of observational recordings at hand, the problem became
this: to obtain a pattern of motion of the planets that would exactly agree with
Brahe’s observations. So dependable were Brahe’s recordings that any solution
that should differ from Brahe’s observed positions by even as little as a quarter of
the moon’s apparent diameter must be discarded as incorrect. Kepler had, then,
first to guess with his imagination some plausible solution and then, with painful
perseverance, to endure the mountain of tedious calculations needed to confirm or
reject his guess. He made hundreds of fruitless attempts and performed reams and
reams of calculations, laboring with undiminished zeal and patience for many
years. Finally he solved his problem, in the form of his three famous laws of
planetary motion, the first two around 1609 and the third one 10 years later in
1619:
I. The planets move about the sun in elliptical orbits with the sun at one
focus.
II. The radius vector joining a planet to the sun sweeps over equal areas in
III. The square of the time of one complete revolution of a planet about its
orbit is proportional to the cube of the orbit’s semimajor axis.
The empirical discovery of these laws from Brahe’s mass of data constitutes one
of the most remarkable inductions ever made in science. With justifiable pride,
Kepler prefaced his Harmony o f the Worlds of 1619 with the following outburst:
I am writing a book for my contemporaries or—it does not matter—for
posterity. It may be that my book will wait for a hundred years for a
reader. Has not God waited 6000 years for an observer?
Kepler’s laws of planetary motion are landmarks in the history of astronomy and
mathematics, for in the effort to justify them Isaac Newton was let to create
modern celestial mechanics. It is very interesting that 1800 years after the Greeks
had developed the properties of the conic sections, there should occur such an
illuminating practical application of them. One never knows when a piece o f pure
mathematics may receive an unexpected application.
In order to compute the areas involved in his second law, Kepler had to resort to a
crude form of the integral calculus, making him one of the precursors of that
calculus. Also, in his Stereometria doliorum vinorum {Solid Geometry o f Wine
Barrels, 1615), he applied crude integration procedures to the finding of the
volumes of 93 different solids obtained by revolving arcs of conic sections about
axes in their planes. Among these solids were the torus and two that he called the
apple and the lemon, these latter being obtained by revolving a major and a minor
arc, respectively, of a circle about the arc’s chord as an axis. Kepler’s interest in
these matters arose when he observed some of the poor methods in use by the wine
gaugers of the time.
It was Kepler who introduced the word focus (Latin for “hearth”) into the
geometry of conic sections. He approximated the perimeter of an ellipse of
semiaxes a and b by the formula τr{a + ⅛). He also laid down a so-called principle
of continuity, which postulates the existence in a plane of certain ideal points and
an ideal line having many of the properties of ordinary points and lines, lying at
infinity. Thus he explained that a straight line can be considered as closed at
infinity and that a parabola may be regarded as the limiting case of either an
ellipse or a hyperbola in which one of the foci has retreated to infinity. The ideas
were extended by later geometers.
Kepler was a confirmed Pythagorean, with the result that his work is often a blend
of the fancifully mystical and the carefully scientific. It is sad that his personal life
was made almost unendurable by a multiplicity of worldly misfortunes. An
infection from smallpox when he was four years old left his eyesight much
impaired. In addition to his general lifelong weakness, he spent a joyless youth; his
marriage was a constant source of unhappiness; his favorite child died of
smallpox; his wife went mad and died; he was expelled from his lectureship at the
University of Graz when the city fell to the Catholics; his mother was charged and
imprisoned for witchcraft, and for almost a year he desperately tried to save her
T5f) Ch. 11 Vector Calculus
from the torture chamber; he himself very narrowly escaped condemnation for
heterodoxy; and his stipend was always in arrears. One report says that his second
marriage was even less fortunate than his first, although he took the precaution to
analyze carefully the merits and demerits of 11 women before choosing the wrong
one. He was forced to augment his income by casting horoscopes, and he died of a
fever in 1630 at the age of 59 while on a journey to try to collect some of his long
overdue salary.
— MATHEMATICAL ESSAYS
1. Write a 500-word essay on the life and mathematics of 8. Two hypothetical planets are moving about the sun in
Galileo Galilei. elliptical orbits having equal major axes. The minor axis of
2. Write a 500-word essay on the life and mathematics of one, however, is half that o f the other. How do the periods
Johannes Kepler. of the two planets compare? Support your response.
3. Explain the remark in Galileo’s The Two New Sciences of 9. a. Kepler divided a circle of radius r and circumference C
1638 that “ neither is the number of squares less than the into a large number of very thin sectors. By regarding
totality of all numbers, nor the latter greater than the each sector as a thin isosceles triangle of altitude r and
former.” base equal to the arc of the sector, he heuristically
4. Explain the following geometric paradox considered by arrived at the formula A = ∣ r C for the area o f a circle.
Galileo in his The Two New Sciences: Show how this was done.
Suppose the larger circle of Figure 11.26 has made one b. Use Kepler’s type of reasoning to obtain a volume V of
revolution in rolling along the straight line from A to B, so a sphere o f radius r and surface area S as V = ∣ ΛS,.
that ∣√4B∣ is equal to the circumference o f the large circle. c. Use Kepler’s type of reasoning to obtain the area A of
Then the small circle, fixed to the large one, has also made the ellipse o f semimajor axis a and semiminor axis b as
one revolution, so that ∣CD∣ is equal to the circumference A = τrab.
o f the small circle. It follows that the two circles have equal Book report “ Science is that body of knowledge that
circumferences.
describes, defines, and where possible, explains the
universe . . . we think of the history of science as a history
o f men. . . . [History] is the story of thousands of people
who contributed to the knowledge and theories that consti
tuted the science of their eras and made the ‘great leaps’
possible. Many o f these people were women.” So begins a
history o f women in science entitled Hypatia's Heritage by
Figure 11.26 Aristotle’s wheel Margaret Alic (Boston: Beacon Press, 1986). Read this
book and write a book report.
This paradox had been earlier described by Aristotle and is 11. Make up a word problem involving the calculus o f vector
therefore sometimes referred to as Aristotle's wheel. valued functions. Send your problem to:
5. Prove Kepler’s first law. You may search the literature for
help with this proof. Mathematics Editor,
RE: Bradley and Smith
6. Prove Kepler’s third law. You may search the literature for
Calculus
help with this proof.
Prentice Hall
7. Where is a planet in its orbit when its speed is the greatest?
113 Sylvan Ave., Route 9W
Support your response.
Englewood Cliffs, N J 07632
(along with credit to the problem poser).
CUMULATIVE REVIEW FOR CHAPTERS 7 - 1 1
1. ■ What Does This Say? In your own words, outline a 3. ■ What Does This Say? What is a quadric surface? In
procedure for integrating a given function. your own words, discuss classifying and sketching
2. ■ What Does This Say? In your own words, outline a quadric surfaces.
procedure for deciding if a given series converges or 4. ■ What Does This Say? What is a vector? What is a
diverges. vector function? In your own words, discuss what is
meant by vector calculus. that the partial sum

5. If y = x cosh~ 1 x, find y '. , ,__
- 1 -
dv <? = y '
6. If x sinh y + y tanh ' x = 0, find v~k
Find the integrals in Problems 7-15. approximates the entire sum 5 with 4-decimal-place
rχ accuracy.
7. ι x In f dx 8. sin 2x cos⅛ dx
33. a. Find the Maclaurin series representation for
.∕'(x) = x 2e~χ2 .
9. I1 tan 2 x secx dx 10. 1 x V 16 - x dx
b. Use the representation in a to approximate
-ι
ι ι . IΓ cosh + dx
1 1 + sinh 2x
12. i(1 dx
1 + cosx o Jx 2e~χ2 dx
with 3-decimal-place accuracy.

1 'I i∣ dx 14 / Vχdx ,. [ dx
34. Solve the first-order linear differential equation
1 x 2(x 2 + 5) -ffx ' J V2x - x 2
~ + 2y 2
Find the equation o f the line or plane satisfying the condi dx x = x
tions given in Problems 16-18.
subject to the condition y = 2 when x = 0.
16. The line through (-1 , 2, 5) and perpendicular to the
plane 2x - 3y + z = 11. 35. Sketch the given polar curves.
a
17. The plane passing through (5, 1, 2), (3, 1, -2 ), and a. 2r == 2 - sin θ b. r = ~ , π------ -
2 + 2 cos θ
(3, 2, 5).
36. Find the area of the region inside the circle r = 4 and to
18. The plane that passes through (-2 , - 1 ,4 ) and perpen
the right of the line r = 2 sec θ.
dicular to the line
37. Find an equation in terms of x and y for the line tangent
x T+ 1 3- z to the curve given by
2 5 2
Test the series in Problems 19-25 for convergence. x = t 2 - 2t - 1, y = t 4 - 4 ∕ 2 + 2
20. ∑ 1 at the point (-1 , 2).

19∙ ∑0 F T 2 fc=l
k ∙ 4 fc 38. A particle moves in space with position vector
∞
2i. y - j -
1
-.-. y~, 3⅛2 —7k + 2 R(t) = (sin t)i - (cos ∕)j + k
‰ ⅛ la⅛ ⅛ (2⅛ - l)(fc + 3)
a. Find the velocity and acceleration vectors for the
k'
„ ( - l ∕∙ ÷ ⅛ particle and find its speed.
23 y 24∙ ξ + + < = - l
' ⅛ 2 fc ■k t b. Find the Cartesian equation for the particle’s trajec
tory.
z7 C
3 ∙ 1
1 -τ4 -81 ---1
______L -Γ
27 64
_∣---- 1____∣---- 1----- ---- -r_L τ4. . . .
1 2 5 -t- 2 l 6
c. Find the curvature κ and the tangential and normal
26. In each case, find the sum of the convergent series components of acceleration for the particle’s mo
∞ 2 k- 1 x | tion.
a ' ∑ 5≡ b' ∑ i (3fc - l)(3fc + 2)
1 39. At what speed must a satellite travel in order to main
tain a circular orbit 1,000 mi above the surface of the
27. In each case, determine whether the given improper earth? Assume earth is a sphere of radius 4,000 mi and
integral converges, and if it does, find its value. that GM in Newton’s law of universal gravitation is
Γ , x , . f2 dx approximately 9.56 × 104 m i 3∕s 2 .
a. x~e ax b. z/
Ji Jo V4 - x 20 40. A calculus textbook (other than the one you are present
ly reading) is hurled downward from the top of a 120-ft-
28. Find v ∙ w and v × w for the vectors v = 3i - 2j + 5k
high building at an angle of 30o with the horizontal.
and w = i —3j - k.
Assuming the initial
29. Find F' and F" for F(t) = 2ti + e~3ti + t4 k. speed of the book is ,∙
30. Find J [ezi —j - tk] ∙ [e~t i + tj — k] dt. 8 ft∕s, how far from
the base of the build
31. Find T and N for ing will the textbook
R(t) = 2(sin2t)i + (2 + 2 cos2t)j + 6⅛. land? i2o ft
32. Show that the alternating series
converges and determine what N must be to guarantee

12
■ CONTENTS
12.1 Functions of Several Variables
Basic concepts; Level curves
Partial Differentiation
and surfaces; The graph of a
function of two variables
12.2 Limits and Continuity
Limit of a function of two
variables; Properties of limits;
Continuity; Formal definition
of limit
12.3 Partial Derivatives
Partial differentiation; PREVIEW
Geometric interpretation;
Partial derivative as a rate; The goal of this chapter is to extend the methods of single-variable
Higher partial derivatives
differential calculus to functions of several variables. The vector meth
12.4 Tangent Planes, Approxima ods developed in Chapters 10 and 11 play an important role in our work.
tions, and Differentiability
Tangent planes; Incremental Indeed, we shall find that the closest analog of the single variable
approximations; The total derivative in higher dimensions is the vector function called the gradi
differential; Differentiability
ent, which will be used, among other things, to find tangent planes to
12.5 Chain Rules
Chain rule for one parameter; surfaces in space. We conclude this chapter by solving rate and optimi
Chain rule for two parameters zation problems involving functions of several variables.
12.6 Directional Derivatives and
the Gradient
The directional derivative; The
gradient; Normal property of
the gradient; Tangent planes PERSPECTIVE
and normal lines
In many practical situations, the value of one quantity may depend on the
12.7 Extrema of Functions of Two
Variables values of two or more others. For example, the amount of water in a reservoir
Relative extrema; Second may depend on the amount of rainfall and on the amount of water consumed by
partials test; Extreme value local residents. The current in an electrical circuit may vary with the electromo
theorem
tive force, the capacitance, the resistance, and the impedance in the circuit. The
12.8 Lagrange Multipliers
Method of Lagrange flow of blood from an artery into a small capillary may depend on the diameter
multipliers; Constrained opti of the capillary and the pressure in both the artery and the capillary. The
mization problems; Lagrange output of a factory may depend on the amount of capital invested in the plant
multipliers with two
parameters; A geometric and on the size of the labor force. We will analyze such situations using
interpretation for Lagrange’s functions of several variables. Graphs will play an especially important role in
theorem our work. For instance, the computer graphics of Mt. St. Helens shown here was
Chapter 12 Review created by considering a function of two variables.
Mount St. Helens, Washington
738
Sec. 12.1 Functions of Several Variables 739
1 2 .1 FUNCTIONS OF SEVERAL VARIABLES
IN THIS SECTION Basic concepts, level curves and surfaces, the graph of
a function of two variables
The basic terminology for functions of several variables is introduced, along
with the notion of level curves and surfaces. We also examine various methods
for sketching graphs of functions of two variables.
In the real world, physical quantities often depend on two or more variables (see
the Perspective box at the beginning of this chapter). For example, we might be
concerned with the temperature on a metal plate at various points at time t.
Locations on the plate are designated as ordered pairs (x, y), so that the
temperature T can be considered as a function of two location variables, x and y,
as well as a time variable, t. Extending the notation for a function of a single
variable, we might write this as T(x, y, z). We begin our study of functions of
several variables by examining this notation and a few other basic concepts.
■ BASIC CONCEPTS
We begin our investigation of functions of several variables by introducing

notation and terminology and examining a few basic concepts.
Function of Two Variables A function of two variables is a rule that assigns to each ordered pair (x, j) in a
set D a unique number∕(x , y). The set D is called the domain of the function,
and the corresponding values o f∕(x , J,) constitute the range of f.
Functions of three or more variables can be defined in a similar fashion. For

simplicity, we shall focus most of our attention on functions of two or three
variables.
When dealing with a function of two variables f we may write z = fix , y) and
refer to x and y as the independent variables and to z as the dependent variable.
Often, the functional “rule” will be given as a formula, and unless otherwise
stated, we shall assume that the domain is the largest set o f points in the plane (or in
0 Note the convention in this space) for which the functional formula is defined. These definitions and conven
book for the domain of a func tions are illustrated in Example 1 for a function of two variables.
tion of two variables. 0
EXAMPLE 1 Evaluating a function of two variables and finding the domain
and range
Let fix , y) = V 1 - x + j.
a. Evaluate/(2, 1 ),/(-4 , 3), and ∕(2 t, ∕) .
b. Describe the domain and range of/
Solution a. /(2, 1) = V l - 2 + 1 = 0
/ ( - 4 , 3) = V l - (-4 ) + 3 = √ 8 = 2√2
∕(2 t, / ) = V l - (2t) + (/) = V ∕ - 1)2 = ∖t - 11
740 Ch. 12 Partial Differentiation
b. The domain off is the set of all ordered pairs (x, y) for which ∖^1 - x + y is
defined. We must have 1 - x + y ≥ 0, or, equivalently, y ≥ x - 1, in
order for the square root to be defined. Thus, the domain off is the shaded
set shown in Figure 12.1.
a. Graph of domain in R2 . This is the way b. Graph of domain in R3 . Notice that the
the graph looks in two dimensions. xy-plane is now a horizontal plane.
Figure 12.1 The domain of f(x , y) = V 1 - x + y
Because z = f(x, y) = V 1 - x + y, we see that z must be nonnegative, and the

range o f/is all z ≥ 0 (because square roots are nonnegative).
Functions of several variables can be combined just like functions of a single

variable.
Operations with Functions If ∕(x , y) and g(x, y) are functions of two variables with domain D, then
of Two Variables
Sum ( / + g)(x, y) = ∕(x , y) + g(x, y)
Difference ( / - g)(x, y) = f(x , y) - g(x, y)
Product (J⅛)(x, y) = ∕(x , y)g(x, y)
Quotient = 'g (x^) ψ 0
A polynomial function in x and y is the sum of functions of the form

Cxmy n
with nonnegative integers m and n; for instance,
3x5y 3 - 7x2y + 2x - 3y + 11
is a polynomial in x and y. A rational function is a quotient of two polynomial
functions. Similar notation and terminology apply to functions of three or more
variables.
■ LEVEL CURVES AND SURFACES
By analogy with the single-variable case, we define the graph of the function f(x ,y )
to be the collection of all 3-tuples (ordered triplets) (x, y, z) such that (x, y) is in the
domain of/and z = f(x, y). The graph of∕( x , y) is a surface in R3 whose projection
onto the xy-plane is the domain D.
It is usually not easy to sketch the graph of a function of two variables. One
way to proceed is illustrated in Figure 12.2.
Figure 12.2 Graph of a function of two variables. Computer programs for three-
dimensional sketching are common; see the Technology Manual.
Notice that when the plane z = C intersects the surface z = f(x, p), the result is the
space curve with the equation /(x, p) = C. Such an intersection is called the trace
of the graph of f in the plane z = C. The set of points (x, p) in the xp-plane that
satisfy f(x , p) = C is called the level curve of f at C, and an entire family of level
curves (or contour curves) is generated as C varies over the range off We can think
of a level curve as a “slice” of the surface at a particular location. By sketching
members of this family on the xp-plane, we obtain a useful topographical map of
the surface z = f(x, p).
For instance, imagine that the surface z = f(x, p) is a “mountain,” and further
assume that we wish to draw a two-dimensional “graph” of the mountain. To draw
l500 such a graph, we indicate the paths of constant elevation by sketching the family of
level curves in the plane and pinning a “flag” to each curve to show the elevation
z= ιooo to which it corresponds, as shown in Figure I2.3b. Notice that regions in the map
Figure 12.3 Level curves of a surface

74 2 Ch. 12 Partial Differentiation
where paths are crowded together correspond to the steeper portions of the
mountain. An actual topographic map of M ount Rainier is shown in Figure 12.3c.
You probably have seen level curves on the weather report in the newspaper or
on the evening news, where level curves of equal tem perature are called isotherms.
Other common uses of level curves involve curves of equal pressure (called
isobars) or curves of equal electric potential (called equipotential lines). See Figure
12.4.
Figure 12.4 Isotherms
EXAM PLE 2 Level curves
Sketch some level curves of the function f ( x , y) = xy.

Solution The level curves have the form xy = C for constant C. For C ≠ 0, these
curves are hyperbolas of the form y = C ∕x, and for C = 0 the level curves are
the coordinate axes, as shown in Figure 12.5.
Figure 12.5 Some level curves of f(x, y) = xy

■ THE GRAPH OF A FUNCTION OF TWO VARIABLES
The level curves of a function f(x , y) provide information about the cross sections
of the surface z = ∕(x , y) perpendicular to the z-axis. However, a more complete
picture of the surface can often be obtained by examining the cross sections in
other directions as well. This procedure is used to graph a function in Example 3.
EXAMPLE 3 Level curves
Discuss the level curves of the function f(x, j>) = x 2 + y 2 and sketch the graph
o f/
Solution If C is a real number, the level curve f(x , j) = C has the equation
x 2 + y 2 = C, and it follows that C ≥ 0, because x 2 + y 2 ≥ 0 for all x and y.
The level curve x 2 + y 2 = 0 (that is, C = 0) is the point (0, 0), and for C > 0,
the level curve x 2 + y 2 = C is the circle with center (0, 0) and radius VC
(Figure 12.6a).
b. The surface z = x 2 + y 2 .
Cross sections perpendicular to
the Λ - and y-axes shown in green
are parabolas.
Figure 12.6 The graph of the function ∕( x , y) = x 2 + y 2
We can gain additional information about the appearance of the surface

by examining cross sections perpendicular to the other two principal direc
tions. Cross-sectional planes perpendicular to the x-axis have the form x = A
and intersect the surface z = x 2 + y 2 in parabolas of the form z = A 2 + y 2,
x = A. Similarly, cross-sectional planes perpendicular to the j ,-axis have the
form y = B and intersect the surface in parabolas of the form z = x 2 + B 2,
y = B.
To summarize, the surface z = x 2 + y 2 has cross sections that are circular
SS M
in planes perpendicular to the z-axis and parabolic in the other two principal
directions. For this reason, the surface is called a circular paraboloid
or a paraboloid of revolution. The graph of the surface is shown in
Figure 12.6b. M B
The concept of level curve can be generalized to apply to functions of more

than two variables. In particular, i f / is a function of the n variables x 1, x 2 , . . . , x
and C is a number in the range of / then the solution set of the equation
∕( x l , x2, . . . , x n) = C is a region of tt-space called the level surface o f/a t C. The
level surfaces of a function of three variables are surfaces in R3 and can provide
some insight into the nature of the function, but level surfaces of functions of four
or more variables are very difficult to visualize.
DRAWING LESSON 5: SKETCHING SURFACES WITH LEVEL CURVES
a. Draw the three coordinate axes and the xy-plane. b. Draw a trace of the surface in the xz-plane.
Draw a trace of the surface in the yz-pkιne.
c. Draw a trace in the plane z = constant. Drop dashed d. Lightly erase all planes except the xy-plane. Use
segment from the intercepts of this trace down to the highlighters or colored pencils to shade the surface.
xy-plane. Then draw the level curve in the xy-plane.
Finish outlining the surface.
EXAM PLE 4 Isothermal surface
Suppose a region of R is heated so that its temperature T at each point (x, y, z)

is given by Γ(x, y, z) = 1 0 0 - x 2 - y 2 - z 2 degrees Celsius. Describe the
isothermal surfaces for T > 0.
Solution The isothermal surfaces are given by T(x, y, z) = k for constant k,
Zc > 0; that is, x 2 + y- + z 2 = 100 - k. If 100 - k > 0, the graph of
√ ιo o -c l
T(x, y, z) = C1
= √ !0 0 - C2
T(Λ∙, y, z) = C2
= √Tbδ - c
T(x, y, z) = C
Spheres with larger radii
correspond to cooler temperatures.
Figure 12.7 Isothermal surfaces for

T (x ,y ,z ) = 1 0 0 - χ 2 - j y 2 - z 2
x 2 + y 2 + z 2 = 100 - k is a sphere of radius V 100 - Ze and center (0, 0, 0).

When k = 100, the graph is a single point (the origin), and T(0, 0, 0) = 100oC.
As the tem perature drops, the constant k gets smaller, and the radius
V100 - Zz of the sphere gets larger. Hence, the isothermal surfaces are spheres,
and the larger the radius, the cooler the surface. This situation is illustrated in
Figure 12.7. h b
As you can see by the examples in this section, you will need to recall the
graphs of quadric surfaces. It will also help if you can recognize the surface by
looking at its equation. What you will need to remember is summarized in Table
10.1, page 644.
PROBLEM SET 1 2 .1
θ 1. Let ∕( x , y) = x 2y + x y 2. If Zis a real number, find: < ∙⅛ Λ w ) e ∙ ⅛ ∕d . Λ l > f ∙ ⅛ ∕( l . l , z ≈ )
a ./(0 ,0 ) b ./ ( - 1 , 0 ) c. / ( 0 , - 1 ) d. /(1 ,1 )
4. Let f( x , y, z) = x sin y + y cosz. Find
e . / ( 2 , 4) f.∕(Z,Z) g∙∕(Z,Z ) 2 h .∕( l - Z ,Z ) a ./(0 . 0,0) b . ∕( l , f > π ) C . ∕( l , π , f )
∖2
(1 —- I . If Zis a nonzero real number, find: d. ⅛ f( x , x, x) e. ⅛ ∕( x , 2x, 3x) f. ∕( y , y, 0)

a ./(0 , 1) b ./(5 , 5) c ,∕( 6 , 1) d. /(1 , 2)
5. ■ What Does This Say? Discuss the notion of a func
e.∕(Z,Z) f.∕(5Z,Z) g∙∕(Z,2z) h . ∕( l + Z,Z) tion of two variables. Your discussion should include
3. Let ∕( x , y, z) = x 1ye 2x + (x + y - z)2 . Find: additional examples of a function of two variables.
a ./(0 , 0,0) b ./ ( 1 , - 1 ,1 ) c. / ( - 1 , 1 ,-1 ) 6. ■ What Does This Say? How do you think that a
function of three variables differs from a function of two
variables?
7. ■ What Does This Say? Discuss level curves and ampli

fy with some examples (different from those in this text).
Find the domain and range for each function given in Problems
8-17.
8. f(x , y) = V x - y
11. ∕( x , y) = y f -
10. f(ιι, v) = V MV
12. ∕( x , y) = ln(y - x)__________ 13. ∕(w , v) = √w sin v

14. f( x , y) = V(x + 3)2 + (y - 1)2 15. ∕( x , y) = e(*+1)zb,- 2)
17. ∕( x , y) = 1 ,- -,
16’ >≈ ⅛ =
V9 - x 2 - y 2
Sketch the level curves o f the functions given in Problems
18-23.
18. f(x , y) = 2x - 3y 19. ∕( x , y) = x 2 - y 2
20. ∕( x , y) = x 3 - y 21∙ g(x, y) = x 2 - y
22. h(u, v) = M2 + 23. ∕( x , t) = f
In Problems 24-29, sketch the level surface f(x, y, z) = C for

the given value o f C.
24. ∕( x , y, z)= y 2 + z 2 for C = 1
25. ∕( x , y, z)= x 2 + z 2 for C = 1
26. f(x , y, z)= x + y - z for C = 1
27. ∕( x , y, z)= x + y —z for C = 0
28. ∕( x , y, z) = (x + 1)2 + (y - 2)2 + (z - 3)2 for C = 4
29. ∕( x , y, z) = 2x 2 + 2y2 - z for C = 1
In Problems 30-37, describe the traces o f the given quadric

surface. You might wish to review Section 10.2 (page 644).
Sketch the graph o f the quadric and identify it.
30. 9x 2 + 4y 2 + z 2 = 1 31∙ ⅝ + τ 2 + f = 1
2 γ2
r
32. ⅜ + ⅜- - z 2 = 1 33. ⅜— y 2 - z 2 = 1
D.
34. z = x 2 + J35□. z = — - —
9 16
36. x 2 + 2y2 = 9z 2 37. z 2 = 1 + f +C
Match each level curve given in Problems 38-43 with one o f the
surfaces labeled A-F.
38. f(x , y) = x 2 - y 2 39. ∕( x , y) = e , γ2 y2
Sec. 12.2 Limits and Continuity 7 47
© Sketch the graph of each function given in Problems 44-55. 59. At a certain factory, the daily output is Q = CKrLf~ r
44. fix , y) = - 4 45. f(x , y) = x units, where K denotes capital investment, L is the size of
46. f{x, y) = y 2 + 1 47. fix , y) = x 3 - 1 the labor force, and C and r are constants, with
48. f(x , y ) = 2 x - 3y 49. f(x , y) = x 2 - y 0 < r < 1 (this is called a Cobb-Douglas production
2
50. f(x , y)= 2x + y 2 51. f(x , y) = x 2 - y 2
function). What happens to Q if K and L are both
doubled? What if both are tripled?
52∙ f(χ , y) = y 53∙ ∕(χ> y) = ^ χ + y 60. Sketch the graph of fix , y) = xy in the first octant
lx, y, and z are all positive).
54. f{x, y) = x2 + y2
+2 55. fix , y) = V 1 - x 2 - y 2
61. The ideal gas law says that PV = kT, where P is the
56. In physics, the lens equation states pressure of a confined gas, V is its volume, T is its
temperature (Kelvin), and k is a constant. Express P as a
ι ι =ι
αθ + d. L function of Iz and T and describe the level curves of this
function.
where d0 is the distance of an object from a thin, spherical 62. A publishing house has found that in a certain city each of
lens, di is the distance of its image on the other side of the its salespeople will sell approximately
lens, and L is the focal length of the lens. Describe the
level curves for the constant focal length. r2 +
s2 5
2,000p 100 ~
units per month, where 5 denotes the total number of

focal point salespeople employed, p is the price per unit, and r is the
amount of money spent each month on local advertising.
object
J Express the total revenue P as a function of p, r, and s.
63. A manufacturer with exclusive rights to a sophisticated
—
H new industrial machine is planning to sell a limited
number of the machines to both foreign and domestic
57. The EZGRO agricultural company estimates that when firms. The price the manufacturer can expect to receive
100x worker-hours of labor are employed on y acres of for the machines will depend on the number of machines
land, the number of bushels of wheat produced is made available. It is estimated that if the manufacturer
fix , y) = A x ay b, where A, a, and b are nonnegative supplies x machines to the domestic market and y ma
constants. Suppose the company decides to double the chines to the foreign market, the machines will sell for
production factors x and y. Determine how this decision
affects the production of wheat in each of these cases:
a. a + b > ∖ b. a + b < ∖ c. a + b = 1
thousand dollars apiece at home and
58. Suppose that when x machines and y worker-hours are
used each day, a certain factory will produce
Q(x, y) = 10xy mobile phones. Describe the relationship
between the “inputs” x and y that result in an “output” of
1,000 phones each day. Note'. You are finding a level curve thousand dollars apiece abroad. Express the revenue P as
of the production function Q. a function of x and y.
1 2 . 2 LIM ITS A N D CO NTINUITY
IN TH IS SECTIO N Limit of a function of two variables, properties of

limits, continuity, formal definition of limit
As in the single-variable case, the concept of limit plays a fundamental role in
the calculus of functions of several variables. In this section we define the limit
of a function of two variables and examine some of its properties. We shall also
consider what it means for such a function to be continuous. Our focus will be
on functions of two variables, but the notions discussed also apply to functions
of three or more variables. We will include some functions of three or more
variables in some of the exercises.
■ LIMIT OF A FUNCTION OF TWO VARIABLES
For functions of a single variable (in two dimensions) we first discussed the
intuitive notion of a limit and then stated the formal definition of a limit. We will
do the same in this section.
Recall from Chapter 1 that the limit of a function of a single variable requires
that we find a δ-interval on the x-axis (δ > 0) that forces the function to be within
an ε-interval of some value, called the limit. In particular, the limit statement
lim ∕( x ) = L
x →c
means that the functional values of ∕( x ) can be made arbitrarily close to L by
choosing x sufficiently close to c but not equal to c. That is, to force/to be within
an ε-interval of L on the y-axis, we must choose x within some δ-interval on the x-
axis (see Figure 12.8).
Find a δ-interval such that For a function of two variables, the situation is similar but the domain has two
∣x + 1 is close to L = 2.5 when dimensions. Instead of a δ-interval around c on the x-axis we consider a δ-
neighborhood about a point (x0, j 0) in the plane. This neighborhood is the interior
ever x is in the δ-interval.
of a circle centered at (x0 , >,0) with radius δ > 0; that is,
All points (x, >,) satisfying V (x - x 0)2 + (y - j' 0)2 < δ
This neighborhood (the circle and its interior) is open if we use < and closed if we
use ≤ . The neighborhood is called a disk. A punctured disk is a disk whose center
(x0, jμ0) has been excluded. This disk in the xy-plane is shown in Figure 12.9a. We
will now give an informal definition of a limit of a function of two variables.
Limit of a Function of Two Variables (Informal Definition)

The notation
lim ∕( x , y) = L
(χ, r)→(⅞,j⅛)
means that the functional values ∕( x , y) can be made arbitrarily close to L by
choosing a point (x, v) sufficiently close (but not equal) to the point (x0, j 0).
a. A punctured disk
We need to determine what we mean by “ arbitrarily close.” Given some ε > 0, we

wish to find a δ > 0 so that for any point (x, y) in the punctured disk of radius δ
centered at (x0, y 0), the functional value∕( x , y) lies between L + ε and L - ε. This
is illustrated in Figure 12.9b.
When considering the limit of a function of a single variable we need to
examine (or test) the approach of x to c from two directions (the left- and right
hand limits). However, for a function of two variables, we write (x, y) → (xθ, v,0) to
mean that the point (x, >,) is allowed to approach (x0, y0) along any curve that
Circle with radius δ and passes through (x0 , y 0). If the value of
with deleted center (xθ, >,θ)
lim ∕( x , j )
b. lim f(x, y) = L (x, ,v) → (x0, y 0)
(x, y)→(x0, y0)
is not the same for all approaches, or paths, then the limit does not exist.
Figure 12.9 Limit of a function of
two variables EXAM PLE 1 Limit of a function of two variables
a. Show
,. 2xy
hm —-—≈λ
(x,>>)→(0, 0) χ 2 + y
does not exist by evaluating this limit along the x-axis (r = 0), the y-axis
(x = 0), and along the line y = x and noting that these values are not the
same.
Sec. 12.2 Limits and Continuity 749
b. Evaluate
lim 2xy
(x ,y )→(l, 2)
along the paths y = 2, x = 1, and y = 2x, and show that these values are all
the same.
Solution a. First note that the denominator is zero at (0, 0) so /(0, 0) is not
defined. If we approach the origin along the x-axis (where y = 0) we find that
Functional values
2 2'λ (θλ = 0 as (x, y) → (0, 0) along y = 0 (and x Ψ 0)
along Λ-axis or
y-axis are 0. x2 + y 2 x 2 + 02 v ’
We find a similar result if we approach the origin along the y-axis (where
x = 0); see Figure 12.10. However, along the line y = x, the functional
values are
∕(* , y) = f(×, *) = χ y-* χ 2 = 1 for x ≠ 0

Functional values along
y —x are 1.
so that as (x, y) → (0, 0) along y = x
2xv
Figure 12.10 Graph of y = * 2 + '
and limits as (x, yj → (0, 0)

Because ∕(x , F) tends toward different numbers as (x, v) → (0, 0) from
different directions, it follows th at/h as no limit at the origin.
b. If we approach the point (1,2) along the line y = 2 (see Figure 12.1 la), we
have
2xy = 4x → 4 as (x, y) → (1, 2) along y = 2
Similarly, along the line x = 1 (Figure 12.1 lb)
2xy = 2y → 4 as (x, y) → (1, 2) along x = 1
Finally, along the line y = 2x (Figure 12.1 lc)
2xy = 2x(2x) = 4x 2 → 4 as (x, y) → (1, 2) along y = 2x
0 As in Example la, the two-path procedure is enough to show that a

limit does NOT exist. However, just showing that the limits along different
paths in Example lb are the same is not enough to conclude that a limit
DOES exist, because we have not evaluated the limit along every
pathway. 0
■ PROPERTIES OF LIMITS
The various rules for manipulating limits of a function of one variable all have
counterparts for limits of two variables.
Basic Formulas and Rules Sup1 1pose that lim

(χ,r)→ (Λ⅛r0)
∕( x , y) = L and lim
(⅛r)→ (⅞⅞)
g(x, y) = M , Then
for Limits of a Function
of Two Variables Scalar multiple rule
lim [af](x, y) = a-L for constant a
(x. y)→ (x0, y0)
Sum rule
lim [∕+⅛ '](x, v) = [ bm f ( x , y)l + [ lim g(x, y)l = L + M
(x ,y)→(χ0,y0>'-j δ jκ L(-v,y)→⅛ r 0) 1 L⅛r)→(⅞.>⅛) J
Product rule
lim [ f g](x, y) = Γ lim f( x , v)^∣Γ lim g(x,y')] = L M
(χ, y)→(χ0, y0) u 6 jv λ ' L(χp)→⅛ √ ’ Jb⅛r)→(⅞j⅛) J
Quotient rule j. r, ,
lim Γ{^l(x, y) = ,y = ⅛ if M ≠ 0
(χ,r)→(⅞⅜) g ’ hm g(x, y) M
(x, y)→(χ0, r0)
EXAM PLE 2 Finding lim its using properties of limits
Assuming each limit exists, evaluate:

2χy
a. lim (x 2 + xy + zy 2) b. lim —5------x
(x,y)→ (3, —4) ' ( x ,∕) → (i,2) χ 1 + y 2-
Solution a. lim (x 2 + xy + y 2) = (3)2 + (3)(-4) + (-4 ) 2 = 13

(x,y)→ (3 ,-4)
A graph is shown in Figure 12.12a.
Figure 12.12 Graphs showing limits
2χy ⅛ ⅛ Λ 2(1 )(2) 4

h lim = = =
(χ, ρ → (i, 2) 2
χ + y2 lim (x 2 + y 2) 12 + 22 5
( x ,j)→ ( l ,2 ) v
The graph is shown in Figure 12.12b.

Historical Note ■ CONTINUITY
Recall the definition of continuity for a function of a single variable: f is

continuous at x = c if
1. ∕(c ) is defined 2. lim ∕(x ) exists 3. lim f(x) = ∕(c).

x →c
Using the definition of the limit of a function of two variables, we can now define
the continuity of a function of two variables in an analogous fashion.
Continuity of a Function of Two Variables

The function f(x , y) is continuous at the point (x0 , y 0) if and only if
1. ∕(xθ, y 0) is defined
2. lim f(x , y) exists
M ARY F AIRFAX SOMERVILLE (x, r ) →(¾, j⅛)
3∙ , lim √ ( A , y) = f(x υ0 , y 0υ ).
(1780-1872) (%, r ) →(⅞, x0)
Mary Somerville’s life was de A lso,/is continuous on a set 5 in its domain if it is continuous at each point
voted to popularizing science in S.
and mathematics, and during
her long life she wrote several ■ What this says-. The function / i s continuous at (x0 , y0) if the functional
textbooks. Somerville College value of∕(x , y) is close to∕( x 0 , y 0) whenever (x, y) is sufficiently close to (x0, y0).
was named after her in recogni Geometrically, this means that / is continuous if the surface z = ∕(x , y) has no
tion of her contributions to sci “gaps” or “holes.”
ence. She did not study mathe
matics seriously until she was
27, and she purchased her first
mathematics books when she
was 30. The reason for her late
start was due to the early influ
ence of males in her life. She
received no education and
taught herself reading, writing,
arithmetic, and Latin. At the
age of fourteen she saw an alge
bra problem in a fashion maga
zine and asked her brother’s tu
tor to let her see an algebra
book. Her father, however, op is discontinuous at (0,0); a hole
posed this study and refused
her the opportunity. Likewise,
she married a man who also
opposed the education of wom
en, and it was not until her EXAMPLE 3 Testing for continuity
husband’s death in 1807 that
she became financially indepen Test the continuity of the functions whose graphs are shown here, namely,
dent and able to begin her
study of mathematics. She went Λ*, y) - ⅛ p b∙ Λ¾y) = y ⅛
on to a distinguished career
and received a Victoria gold
medal in 1869. Also, she was Solution a. Because x - y and x 2 + y 2 are both polynomial functions in x and y,
one of the first two women / i s a rational function and the only place where it might not be continuous is
elected honorary members of when the denominator is zero. Because x 2 + y 2 = 0 only when both x and y are
the Royal Astronomical 0, the only possible point of discontinuity is at (0, 0). Because /(0, 0) is not
Society. defined, we see/is discontinuous at (0, 0).
b. Once again we need to check when the denominator is 0:

y - x~ = 0 when y = x 2
We can, therefore, conclude that the function is continuous at all points
except those lying on the parabola y = x 2. B M
The basic properties of limits can be used to show that if f and g are both
continuous on the set S, then so are the sum/ + g. the multiple a f the product fg,
the quotient fig (whenever g Ψ 0), and the root Vf wherever it is defined. Also, if F
is a continuous function of two variables at (xθ, y0) and G is a function of one
variable that is continuous at F(x0, y0), it can be shown that the composite
function G o F is continuous at (x0 , y 0 ).
Many common functions of two variables are continuous wherever they are
defined. For instance, a polynomial in two variables, such as x 3y2 + 3xy3 —lx
+ 2, is continuous throughout the plane, and a rational function in two variables is
continuous wherever the denominator polynomial is not zero. In this text all of the
“standard” functions of two or more variables are continuous over their domains.
FORMAL DEFINITION OF LIMIT*
Next, we examine a formal ε-δ definition of limit. The informal definition of limit
we have given is sufficient for our purposes in this course, but many advanced
applications require precise definitions.
A point P0(x0 , y0) is said to be an interior point of a set 5 in the plane if some
open disk centered at PQ is contained entirely within S, as shown in Figure I2.13a.
a. Po is an interior point. If 5, is the empty set or if every point of >Sis an interior point, then 5 is an open set.
The point Po is called a boundary point of S if every open disk centered at Po
contains both points that belong to 5 and points that do not (see Figure I2.13b.)
The collection of all boundary points of S is called the boundary of S. Finally, the
set 5 is said to be closed if it contains its boundary—that is, if every boundary
point of S is also a point in .S' (see Figure 12.14). If S has no boundary points, then
it automatically contains “all of them” (beca, se there aren’t any) and is therefore
closed. Finally, a two-dimensional set 5' is said to be bounded if it can be enclosed
in a circle (or a rectangle), as shown in Figure 12.13c.
b. Po is a boundary point. An open set A closed set
The boundary is The boundary

not part of the set. is in the set.
c. S is bounded. Figure 12.14 Open and closed sets in the plane
Figure 12.13 Topology of the
plane We now turn to a formal definition of limit of a function of two variables.
The material of this section is not required for subsequent topics.

Limit of a Function of Two Suppose f( x , y) is a function defined everywhere on an open disk centered at
Variables (Formal Definition) the point P0(x0, y 0) except possibly at P o itself. Then the number L is the limit
o f / a t P if for every ε > 0 there exists a δ > 0 such that
∣∕( x , y) - L∖ < ε whenever 0 < V (x - x 0)2 + (y - y 0)2 < <5

In this case, we write
lim ∕( x ,y ) = L
(¾j')→(⅞,y0)
What this says: If L is the limit of f at
P0(x0, y0), then for each point P(x, y)
in the puncturec disk of radius δ centered at
Po , the number f(x , y) lies between L - ε and
L + ε.
⅞ g ≡ ⅜ ∣S g ⅞
EXAM PLE 4 Continuity using the limit definition
Show that/is continuous at (0, 0), where
j s in x x ≠ 0
∕(λ ', y) =
.0 x = 0
Solution The graph is shown in Figure 12.15. It appears to be continuous. In
order to prove this, we must show that for any ε > 0, there exists a δ > 0
such that
∣ ∕( x , y) - 0∣ = y sin∣j < ε whenever 0 < x 2 + y 2 < δ2
(We use x 2 + y 2 and δ2 here instead of V x 2 + y 2 and δ for convenience.) Note

that ∣y sin^∣ ≤ ∣y∣ for all x ≠ 0, because ∣sin⅛∣ ≤ 1 for x ≠ 0. Let δ = ε. Then if
(x, y) satisfies x 2 + y 2 < δ2 = ε 2, we have ∣y∣ < ε so that
∣7 (x, y) - 0∣ ≤ ∣y∣ < δ = ε whenever ∣y∣ < δ
Front view Side view

Figure 12.15 Graph of the function/in Example 4
PROBLEM SET 1 2 .2
@ 1. ■ What Does This Say? Describe the notion of a limit 25. lim exy 26. lim (x + y)exy
(x ,y )→ (0. 0) (x ,y )→ ( l, 1)
of function of two variables.
2. ■ What Does This Say? Describe the basic formulas
and rules for limits of a function of two variables.
In Problems 3-20, find the given limit, assuming it exists.
3. lim (xy 2 + x 2y + 5)
(x ,y )→ ( - l , 0 ) '
4. lim (5x 2 - 2xy + y 2 + 3)
(x ,y )→ (0,0)
e, ,∙ χ + >, x .∙ χ - y
5. hm ------y 6. lim , 2. ,-
2
( x , r t → ( l, 3) x - (x, y)→ (3, 4) V x + y
7. lim exy 8. lim (x + y)e xy

⅛ y ) → (1.0) 2+ v (x,,1 ) → ( l,0 )
9. hm [ex ln(ey2)] 10. lim ln(x2 + y 2)
⅛ γ ) → (0, i) , u , j w t ', 0 '
,. x 12 - 2xy +l v 2 ι (x 2 - l ) ( y 2 - 4 )
11. h m --------- :-----— 12. (χ.y)‰ ) ( x - l ) ( y - 2 )
(Λ∙,y)→(0,0) X~y
,. e vtan -1 y ,. extan - 1 y
13. h m -------- - 14. h m --------- -
(x ,y )→ (0, 0) x
lim ⅛ ± 2 5
15. 16. lim (sinx + cosy)
(x ,y )→ (0, 0)
x4 - y 4
17. hm —— — ■
x4 - y4
18. lι m —— —;L a is a constant,
(x ,y )→ (α, α) X~ ~ y
x 2 - 4y2
19. lim π
(x ,y )→(2, 1) X - 2y
l- hm , , , = ----- sin (x 2 + y 2)
(x ,y )→ (o. o) V x2 + y2 + 4 - 2 29. lim
(x, y )→ (0, 0) x2 + y2
In Problems 21-29, evaluate the indicated limit, i f it exists.
21. (x ,ylim (xy 2 + x 3y) 22. lim (5x 2 - 2xy + y 2)
)→(2, 1) ■ (x ,y )→ ( l , 2 ) v
© In Problems 30-33, show that lim f(x , y) does not exist.

Sec. 12.2 Limits and Continuity 7 55
38. Find the constant A for which the function
r∣ ∖ J X■ >+ ∖ for (x >f) ≠ (θ> θ)

∕( x , y) = < x 2 + y 2 v ’ ∙ ’ v ’
1^4 for (x, y) = (0, 0)
is continuous at the origin.

39. Find the constant B for which the function
'3 Λ -3 _ 3 V 3
∖ --- 22----- 2~ f° r (X >jθ ≠ (O’ 0)
∕(zγx , L) = χ - v2 y ’
J3 for (x, y) = (0, 0)
is continuous at the origin.
40. Let
r 2 2
∕( x , y ) = ½ ⅛ for (x, y) ≠ (0, 0)
34. Let f be the functions defined by
2 ,0 for (0, 0)
{ 0
f o r ( * ,D ≠ ( 0 ,0 )
for (x, y) = (0, 0) 41.
Is f continuous at (0, 0)? Explain.
Assuming that the limit exists, show that
a. Show that ∕( x , y) → 0 as (x, y) → (0, 0) along any χ z
rhm —;------y5--------- r = ∏
0
straight line path o f the form y = kx or along either (x,y, z)→ (0, 0, 0) x 2 + y 2 + z 2
coordinate axis.
b. Show that f ( x , y) is not continuous at the origin. θ Use the definition o f limit to verify the limit statements given in
Hint. Consider the limit as (x, y) → (0, 0) along the Problems 42-45.
parabolic path y 2 = x. 42. lim (2x 2 + 3y2) = 0
(*,r )→ (o, o)
35. Let f be the functions defined by 43. lim (x + y 2) = 0
(x,y)→ (0,0) ∖
' 3 x 2 - v2
44. lim — -j - = 0
, ∖ —
f ( x , y) = χ 2 + yΛ6
R f° r (χκ >, yx>, ≠ (vθ' ’θ) ’ (x,y)→(0,0) x + y
S> for (0, 0) χ2 - y 2
45. lim --------1---- = 2
⅛ y )→ ( i , - i ) x + y
a. Show that f ( x , y) → 0 as (x, y) → (0, 0) along any
straight line path o f the form y = kx or along either 46. Prove that if / is continuous and f( a , b) > 0, then there
coordinate axis. exists a δ-neighborhood about (a, b) such that f ( x , y) > 0
for every point (x, y) in the neighborhood.
b. Show that f ( x , y) is not continuous at the origin.
Hinf. Consider the limit as (x, y) → (0, 0) along the 47. Prove the scalar multiple rule:
path y 3 = x. lim [af](x, y) = a lim f ( x , y) = aL
(X, >) → (x0, ^0) (x, y) → (x0, y0)
36. Let f be the function defined by
48. Prove the sum rule:
2 2
f ( x >>’> = χ 2 + y2 f ° r (*’ ≠ ( ° ’ °) lim [ ∕+ g ] ( x , y) = L + M
(X, y)→(x0,y0)
where
a. Find lim f ( x , y).
(X ,7 ) → (2, 1)
b. Prove that f has no limit at (0, 0) by showing L = lim f( x , y) and M = lim g(x, y).
(χ,y)→(χ0,y,j ) (χ.r)→(χ0,⅞)
that f ( x , y) tends toward different numbers as
49. A function o f two variables f ( x , y) may be continuous in
(x, y) → (0, 0) along each coordinate axis.
each separate variable at x = x 0 and y = y 0 without being
37. Let f be the function defined by itself continuous at (xθ, y 0 ). Let f ( x , y) be defined by
γ2 -4- " 2
∕U , = fθ r (x >f) ≠ (°’ °)
"χ l + "v 2 ∕( x , j ,) = k ¾ 2 for (x, y) ≠ (0, 0)
a. Find lim f( x , y). J) for (0,0)

(x,y)→ (3, 1)
b. Prove that f has no limit at (0, 0) by showing Let g(x) = f( x , 0) and ∕z(y) = /(0, y). Show that both g(x)
that f ( x , y) tends toward different numbers as and A(y) are continuous at 0, but that f ( x , y) is not
(x, y) → (0, 0) along each coordinate axis. continuous at (0, 0).
1 2 .3 PARTIAL DERIVATIVES
IN THIS SECTION Partial differentiation, geometric interpretation, partial

derivative as a rate, higher partial derivatives
In many problems involving functions of several variables, the goal is to find
the derivative of the function with respect to one of its variables when all the
others are held constant. In this section we consider this concept and shall see
how it can be used to find slopes and rates of change.
■ PARTIAL DIFFERENTIATION
The process of differentiating a function of several variables with respect to one of

its variables while keeping the other variable(s) fixed is called partial differentia
tion, and the resulting derivative is a partial derivative of the function.
Recall that the derivative of a function of a single variable /is defined to be the
limit of a difference quotient, namely,
∕'(x )
= f(l ÷ Λ⅜ - >>
∆%→o Ax-
Partial derivatives with respect to x or y are defined similarly.
Partial Derivative of a Function If z = ∕(x , y), then the (first) partial derivatives off with respect to x and y are
of Two Variables the functions/, and / respectively, defined by
f . λ .∙ ∕( * + ∆x, y) - ∕(x , y)
⅛x ∙ >■>=⅛ ------ to
f , = .∙hm —
∕(∙V
— -y--+---∆yy)—- —
∕(x—, y—)
lfX(x, y)
λ
y, Δy→ 0 ∆y
provided the limits exist.

What this says-. For the partial differentiation of a function of two
variables, z = f(x, y), we find the partial derivative with respect to x by
regarding y as constant while differentiating the function with respect to x.
Similarly, the partial derivative with respect to y is found by regarding x as
constant while differentiating with respect to y.
EXAM PLE 1 Partial derivatives
If∕(x , y) = χ3y + x ),2 - find a . / and b ./ .

Solution a. F o r /, hold y constant and find the derivative with respect to x:
∕( x , y) = 3x2y + y 2
b. For f hold x constant and find the derivative with respect to y.
f y (χ , y) = x 3 + 2xy
Several different symbols are used to denote partial derivatives, as indicated
in the following box.
Alternative Notation for First For z = ∕(x , y), the partial derivatives / an d /, are denoted by
Partial Derivatives
∙y ) = 1 = 8 = ⅛ ∕( x > = = β v(∕)
Sec. 12.3 Partial Derivatives 757
and
f y (χ, τ) = f = = ⅜ ∕( Λ, y) = s- = ⅞ ( ∕)
The values of the partial derivatives of∕( x , y) at the point (a, ⅛) are denoted by
⅛ = and f (α,6) = fy ^ , ⅛
E X A M P L E 2 Finding and evaluating a partial derivative
Let z = x 2sin (3x + y 3*).
a. Evaluate ∣ ∣ at (j> θ). b. Evaluate ∣p at (j> θ).
Solution a. ∣∣= 2x sin(3x + y 3) + x 2cos(3x + y 3) ∙ (3)

= 2x sin(3x + y 3) + 3x 2cos(3x + y 3)
Thus,
⅛ l(fo ) = 2 (f) s1 n 7 r+ 3 ( f ) cosτr = y ( 0 ) + y ( - l ) = - y
b. ∣
∣= x 2cos(3x + y 3) ∙ (3y2) = 3x 2y 2cos(3x + y 3), so that
a⅛r3⅛y<°>-
EXAM PLE 3 Partial derivative of a function of three variables
Let ∕( x , y , z) = x 2 + 2xy- + y z 3; find: a. f χ b .f c .f z

Solution a. For f , think of f as a function of x alone with y and z treated as
constants: ∕ v(x, y, z) = 2x + 2y 2 .
b. ∕ v(x, y , z) = 4xy + z 3 c. ∕ z (x, y , z) = 3yz2 «
EXAM PLE 4 Partial derivative of an implicitly defined function
Let z be defined implicitly as a function of x and y by the equation

x 2 + 5xy 2 + z 3 = 1
Find a. ⅛ b. ⅛-.
∂x ∂y
Solution a. Differentiate implicitly with respect to x, treating y as a constant:

2x + 5y2 + 3z2 1⅞ = 0 so that 2~ = ~X 2 ^^^*l
-, ∂x ∂x 3z2
b. For this part, differentiate implicitly with respect to y.
1Oxy + 3z2 ∣ z = 0 so that ∣7 =

∙7 r)12 r∩ 3 <'7∙4 n-
The following example uses implicit differentiation to find the slope of the
tangent line to a level curve at a particular point.
EXAM PLE 5 Slope of a level curve
A certain level curve of the surface z = x 2 + 3xy + y 2 contains the point P,

where x = 1 and y = 1. Find the slope of the tangent line to this curve at P.
Solution The required level curve has the form F(x, y) = C , where
F(x, y) = x 2 + 3xy + y 2
We find the slope, implicitly:
x 2 + 3xy + y 2 = C
- ^ x 2+ 3xy + y>) = ⅛ (C )
2X + 3P + 3X ⅛ + 2 J ⅛ = 0
(3x + 2 y ) ∣ = - 2 x - 3y
dy _ - ( 2 x + 3y)
The slope of the line tangent to a d x ~ 3x + 2y
level curve o f a surface
This provides the slope at each point on the level curve where 3x + 2y ≠ 0. In
particular, at the point where x = 1 and y = 1, we have
dy _ ~[2 (1) + 3(1)] _ - 5 _
tZx(M) 3(1) + 2 (1) 5 1
so that the required line has slope - 1.
■ GEOMETRIC INTERPRETATION
A useful geometric interpretation of partial derivatives is indicated in Figure

12 ∙16. In Figure l 2 ∙16a, the plane y = y 0 intersects the surface z = f( x , y) in a
curve C parallel to the xz-plane. That is, C is the trace of the surface in the plane
y = y Q. An equation for this curve is z = f l x , y 0 ), and because j ,0 is fixed, the
function depends only on x . Because C lies in the plane y = y 0, we can compute the
slope of the tangent line to C at the point P(X Q , y 0 , z0 ) in the plane y = y 0 by
differentiating/^, y 0) with respect to x and evaluating the derivative at x = x 0 .
That is, the slope is f x (x 0 , y 0), the value of the partial derivative f χ at (x0, y 0).
There is a similar interpretation f o r ∕( x 0 , y 0), as shown in Figure l 2 ∙16b.
Figure 12.16 Slope interpretation of the partial derivative

Partial Derivative as the Slope The line parallel to the xz-plane and tangent to the surface z = ∕(x , y) at the
of a Tangent Line point Pθ(x0 , y 0, z0) has slope f x (x0 , >,0)∙ Likewise, the tangent line to the surface at
Po that is parallel to the yz-plane has slope ∕( x 0 , y0).
EXAMPLE 6 Slope of a line parallel to the xz-plane
Find the slope of the line that is parallel to the xz-plane and tangent to the
surface z = xVx + y at the point P(l, 3, 2).
Solution If'∕( x ,p) = xVx + y = x(x + y) ιz2, then the required slope i s ∕ γ(l, 3). We
find that
∕ γ(x, y) xQ)(x + y) ιz2(l +0) + (l)(x + y) ιz2 - + Vx + y
Thus,
/.(1, 3) = ,1 + √1 + 3 = I
λ 2√1 + 3 4
βββ
■ PARTIAL DERIVATIVE AS A RATE
The derivative of a function of one variable can be interpreted as a rate of change,

and the analogous interpretation of partial derivative can be described as follows.
Partial Derivatives as Rates As the point (x, y) moves from the fixed point P 0(x0, y 0), the function ∕(x , y)
of Change changes at a rate given b y ∕ γ(x0, y0) in the direction of the positive x-axis and by
∕,(x 0 , y0) in the direction of the positive y-axis.
EXAMPLE 7 Partial derivatives as a rate of change

In an electrical circuit with electromotive force (EMF) of E volts and
resistance R ohms, the current is I = E/R amperes. Find the partial deriva-
oT'∖T
tives and at the instant when E = 120 and R = 15 and interpret these
derivatives as rates.
Solution We find that because I = ER~ l
y i = R ~ l and - ^ = - E R ~ 2
∂E ∂R
and thus, when E = 120 and R = 15, we have
i l = i s - ' ≈ o.O667 and ⅛ = —(120)(15)^2 ≈ -0.5333
This means that if the resistance is fixed at 15 ohms, the current is increasing
(because the slope is positive) with respect to voltage at the rate of 0.0667
amperes per volt when the EMF is 120 volts. It also means that if the EMF is
fixed at 120 volts, the current is decreasing (because the slope is negative) with
respect to resistance at the rate of 0.5333 amperes per ohm when the
resistance is 15 ohms. BBM
■ HIGHER PARTIAL DERIVATIVES
The partial derivative of a function is a function, and often it is possible to take the
partial derivative of a partial derivative. This is very much like taking the second
derivative of a function of one variable if we take two consecutive partial
derivatives with respect to the same variable, and the resulting derivative is called
the second partial derivative with respect to that variable. However, we can also
take the partial derivative with respect to one variable and then take another
partial derivative with respect to a different variable; in this case, the derivative is
called a mixed partial derivative. The higher partial derivatives for a function of
two variables f(x , j) are denoted as indicated in the following box.
Higher Partial Derivatives Given z = f(x, y).

Second partial derivatives
a 2/ _ a⅛ = _a_(dz\ = /π = f
∂χ 2 ∂x2 ∂χ∖ ∂χ} [Jx'x jχ χ
= ∂ z _ ∂ (∂z∖ _ f ∩
d2f 2
= f
∂y2 ∂y2 θy∖ θy∕ ^ y y yy
Mixed partial derivatives

∂2∕ ' = ∂2z = JL(∂ Z ∖ = ( F∖ = f
∂x∂y ∂x∂y ∂x∖ ∂y) χ j y'χ j yχ
∂2f _ 9⅛ ∂ (∂z∖ _ , ∕∙x _ r

∂y∂x ∂y∂x ∂y∖ ∂x) ^-x 'y ∙2χy
EXAM PLE 8 Higher partial derivatives of a function of two variables
For z = ∕(x , y) = 5x2 - 2xp + 3y3, find the requested higher partial deriva
tives.
2
a. ∂ z c∙ g d∙ ∕√ 3 , 2)
h d z
∂x∂y ∂y∂x
Solution Part of what this example is illustrating is the variety of notation that
can be used for higher partial derivatives.
a. ∣γ = - 2 x + 9y2 First differentiate with respect to y ,

then differentiate with respect to x.
⅛ = M⅛r)
∂x∂y ∂x∖ ∂yj
= τ^(^2x + M = ^2
∂x ■’
b. Differentiate first with respect to x and then with respect to y:
⅛ - = M g = y-(10-r - 2τ)
∂y∂x ∂y∖ ∂x) ∂y ',
= ~2
c. Differentiate with respect to x twice:
∂2z = a ∂z - ⅛ ( 1 0 x - 2 U - 1 0
∂χ 2 ∂x∖ ∂x,
d. Differentiate first with respect to x and then with respect to y (part b), and
evaluate at the point (3, 2): ∕ vι.(3, 2) = -2 . ≡≡
Notice from parts a and b that = 9 y d χ ∙ This equality of mixed partials

does not hold for all functions, but for most functions we shall encounter, it will be
true. The following theorem provides sufficient conditions for this equality to
occur.
T H E O R E M 12.1 Equality of mixed partials
If ∕( x , y) has mixed partial derivatives at (x0, y 0) and the mixed partials are
continuous in an open set containing (x0 , y 0), then
∕ √ ⅞ y 0) = ∕√ ⅛ τ 0)
Proof: This proof requires methods of advanced calculus and is omitted in
this text. ■
EXAM PLE 9 Partial derivatives of functions of two variables

1 'i ∏d.A - ∕ ∖ - ∕ v ,- and where ∕( x , y) = x 2ye∙, '.
r 1
Solution We have the first partial derivatives

f χ = 2xye y f = x 1ey + x 2ye y
The mixed partial derivatives (which must be the same by the previous
theorem) are
f xy = (∕v).r = 7xe '' + 2 x Te v f y x = (.∕,',)λ. = 2xe y + 2xye y
Finally, we compute the second and higher partial derivatives:
j fx x = (×j/x’)' x = 2y' ey and j fx xy = (×j fx x ')y = 2ey + 2yyey M M
∙^∣7~, ")2T
The heat equation = c2 describes the heat in an insulated rod. The
constant c depends on the material in the rod.
E X A M P L E 10 Verifying that a function satisfies the heat equation
Verify that Γ(x, t) = e^ , cos^ satisfies the heat equation.

Solution ⅛y = - e ~ t cos~ and = e - z sin-') = —5 e~t c o s -
∂t c ∂x 2 ∂x∖ c c) c2 c
Thus, T satisfies the heat equation = clff~ y .
Analogous definitions can be made for functions of more than two variables.
For example,
- - aΓ a /∂∕∖^∣ _ a
V _ a Γ a / ∂A^∣
f f
-zz ∂zi azLaz∖az∕J ¼∑ ∂z∂y∂x a∑Lay∖ax∕J
E X A M P L E 11 Higher partial derivatives of a function of several variables
By direct calculation, show that f χyz —f = f zyχ for the function

∕( x , y, z) = xyz + x 2y 3z 4 .
Solution Find the first partials:

f χ (x, y, z) = y z + 2xy 3z 4 ∙, f ( x , y, z) = x z + 3x 2y 2z 4 j f z (x, y, z) = x y + 4 x 2y 3z 3
Next, find the required second mixed partials:
f xy (x, y, z) = (yz + 2xy 3z 4 )v = z + 6xy 2z 4
f z (x, y, z) = (xz + 3x 2y 2z 4 )z = x + I2 x 2y 2 z 3
f zy (x, y, z) = (xy + 4x 2y 3z 3)v = x + 12 x 2y 2z 3
Finally, find the third mixed partials:
∕ v r .(x, y, z) = (z + 6xy 2z 4)2 = 1 + 24xy 2z 3
f (x, y, z) = (x + 12x 2y 2z 3)χ = 1 + 24xy 2z 3
f zyx (x, y, z) = (x + 12x 2y 2z 3)χ = 1 + 24xy 2z 3 MBB
PROBLEM SET 1 2 .3
θ 1. ■ What Does This Say? What is a partial derivative? θ 29. Find f χ and f for
2. ■ What Does This Say? Describe the algebraic process f ( x , y) = J (t 2 + 2t + 1) di
for finding the partial derivative of a function of three
variables. Hint ∖ Review the second fundamental theorem.
Find ∙yfx , ∙jfy , ∙jfxx , and ∙jfyx in Problems 3-8. 30. Find f χ and /', for
3. f(x , y) = x 3 + x 2y + xy 2 + y 3 f( x , y) = (e l + 3t) dt
4. f(x , y) = (x + x y + y) 3
Hint'. Review the second fundamental theorem.
5. f(x , V) = j 6. f( x , y) = xe xy
31. Find the slope of the
7. f(x , y) = ln(2x + 3y) 8. f( x , y) = sinx 2y tangent line to the level
curve at the point
Find f χ and f v in Problems 9-16. P (l, 1, 3) on the surface
9. a. f(x , y) = (sinx 2)cosy b. f( x , y) = sin(x 2cosy) z = x 2 + xy 2 + y 3 .
10. a. f( x , y) = (sinVx)ln y 2 b. f( x , y) = sin (Vx In y 2)
11. f( x ,y ) = √ 3 x 2 + y 4 12. f( x , y) = xy 2ln(x + y)
13. f( x , y) = x 2ex+y cosy 14. f( x , y) = xy 3tan - 1 y
15. f(x , y) = sin - , xy 16. f( x , y) = cos-1 xy
F indf χ , f v , and /' in Problems 17-22.
17. f( x , y, z) = x y 2 + yz 3 + xyz
18. f(x , y, z) = xye z 32. Find the slope of the tangent line to the level curve at the
in n ι x + y2 point P(V π∕2, ∖fiτr!2, ττ∕4) on the surface
19. f( x , y, z) = —y —
xy + yz z = xy + cos (x 2 + y 2)
20. f( x , y, z)
xz
21. f(x , y, z) = ln(x + y 2 + z 3)
22. f(x , y, z) = sin(xy + z)
In Problems 23-28, find ~ ↑and ∣

∕ by differentiating implicitly.
24 3j - +1 = ι 24. 3x 2 + 4y 2 + 2z 2 = 5
9 4 '2
2 3 2
25. 3x y + y z —z = 1 26. x 3 - xy 2 + yz 2 —z 3 = 4
27. Vx + y 2 + sinxz = 2 28. ln(xy + yz + xz) = 5
(x > 0, y > 0, z > 0)
A function f(x, y) is said to be harmonic on the set S (1) if it is a. the mass m.

continuously differentiable, (2) if it satisfies the Laplace equa b. the body temperature T.
tion
c. the fur temperature t.
fj xx + jfyy = θ 44. A gas that gathers on a surface in a condensed layer is said
throughout S, and (3) if its second derivatives are continuous. to be adsorbed in the surface, and the surface is called an
Show that the functions given in Problems 33-36 are harmonic adsorbing surface. It has been determined that the amount
on the prescribed set. You may assume that the functions are of gas adsorbed per unit area on an adsorbing surface is
continuously differentiable and that the second derivatives are given by a formula of the form
continuous.
S(p, T, h) = apeh llh τ ^
33. ∕(x , J ,) = 3x 2y - y 3∙, S is the entire plane.
34. f{x, y) = ln(x2 + y 2); S is the plane with the point (0, 0) where p is the gas pressure, T is the temperature of the gas,
removed. h is the heat of the adsorbed layer of gas, and a and b are
35. f(x, y) = ex siny, S is the entire plane. physical constants. Find the rate of change of S’ with
respect to
36. f(x, j ,) = sinx coshy .S, is the entire plane.
a. p b. h c. T
37. Let f(x , y) = xy 3 ÷ x 3y Find the slope of the tangent line
to the graph o f /a t P(l, -1 , -2 ) in the direction of 45. The ideal gas law says that PV = kT, where P is the
a. the x-axis. b. the yaxis. pressure of a confined gas, V is the volume, T is the
38. Find the slope of the tangent line at the point temperature, and A: is a physical constant.
P(l, -1 , -2 ) on the graph of a. Find
r7 , ∂V ■
b. τ~∙ ∂V
FindJ —
ol ∂P
, .. . ∂P ∂V ∂T ,
c. Show that -rror ■-r=,
aJ ■or = - 1 .
c
in the direction 46. At a certain factory, output is given by <2 = 120K2'3 L 2j5,
a. parallel to the xz-plane. where K denotes the capital investment (in units of
b. parallel to the yz-plane. $1,000) and L measures the size of the labor force (in
39. For∕(x, y) = cosxp2, show / = / worker-hours).
40. For∕(x, y) = (sin2 x)(siny), show ∕ q , = f yχ . a. Find the marginal productivity of capital, ∂Q∕∂K, and
the marginal productivity o f labor, ∂Q!∂L.
41. F in d ∕ χ g , - f yzz , where
b. Determine the signs of the second partial derivatives
∕( x , y, z) = x 2 + y 2 - 2xy cos z ∂2QJ∂L2 and ∂2Q∕∂K2 and give an economic interpreta
tion.
42. It has been determined that the flow (in cm3∕s) of blood Note-. Q is an example of a Cobb-Douglas production
from an artery into a small capillary is given by function.
F(x,y, z) = c- ^ V ^ 47. The temperature at a point (x, y) on a given metal plate in
the xy plane is determined according to the formula
T(x, y) = x 3 + 2xy 2 + y degrees. Find the rate at which
for constant c > 0, where x is the diameter of the the temperature changes with distance if we start at (2, 1)
capillary, y is the pressure in the artery, and z is the and move:
pressure in the capillary. What is the rate of change of the
a. up (parallel to the yaxis).
flow of blood with respect to
a. the diameter of the capillary? b. to the right (parallel to the x-axis).
b. the arterial pressure? 48. In physics, the wave equation is
c. the capillary pressure?
43. Biologists have studied the oxygen consumption of cer ∂2z =
2
2∂ Z
tain furry mammals. They have found that if the ∂t2 ∂x2
mammal’s body temperature is T degrees Celsius, fur and the heat equation is
temperature is t degrees Celsius, and the mammal does
not sweat, than its relative oxygen consumption is approx ∂z = 2
2∂ Z
imated by ∂t ∂x2
C(m, t, T) ≈ σ ( T - t)m ^- 61 In each of the following cases, determine whether z
satisfies the wave equation, the heat equation, or neither,
(kg∕h) where m is the mammal’s mass (in kg) and σ is a a. z = e~z(sin^ + cos2;)
physical constant. Find the rate (rounded to two decimal
places) at which the oxygen consumption changes with b. z = sin 3 ct sin 3x
respect to: c. z = sin 5 ct cos 5x
764 Ch. 12 Partial D ifferentiation
θ 49. T H IN K T A N K P RO BLEM Let tration at position x and time t. Because the number of
molecules in this substance is very large, it is reasonable
Γ ∕x 2 - y 2 '∖
2 if (x, .y) ≠ (0, 0) to assume that C is a continuous function whose partial
∕( x , j>) = ¾ γ + y 2J derivatives exist. One model for the flow yields the
to if( x , y) = (0, 0) diffusion equation in one dimension; namely,
Show that / (0 , y) = —y and f (x, 0) = x , for all x and y. ac= <,a2c
Then show that f (0, 0) = -1 a n d ∕ v(0, 0) = 1. Why does ∂t ∂x 2
this not violate the equality of the mixed partials theor
em? where δ is the diffusion constant.
50. Show that /(0 , 0 ) = 0 but f y (0, 0) does not exist, where a. What relationship must exist among a, b, and δ for a
function of the form
if (x, v) ≠ (0, 0) C(x, f) = e ax+b t
∕( v T) =
if (x, y) = (0, 0) (α and b are constants) to satisfy the diffusion equa
51. Suppose a substance is injected into a tube containing a tion?
liquid solvent. Suppose the tube is placed so that its axis is b. Verify that
parallel to the x-axis as shown in Figure 12.17. C(x, 0 = r l,2 e~ χ 2 ', ^
satisfies the diffusion equation?

52. The area o f a triangle is A = ⅛ab sin γ, where γ is the angle
between sides o f length a and b.
x=0
τ-∙ j ∂A
Find —
∂A , a nj ∂A
a.
∂a >∂b
-√F θ∂y
v^■
b. Suppose a is given as a function of b, A , and γ. What is
Assume that the concentration o f the substance varies
only in the x-direction and let C(x, t) denote the concen ∂b'
1 2 .4 TANGENT PLANES, APPROXIMATIONS, AND DIFFERENTIABILITY
IN THIS SECTION Tangent planes, incremental approximations, the total

differential, differentiability
We now find the equation of a plane tangent to a surface at a given point on the
surface and then use this tangent plane as an approximation for the value of a
corresponding nearby point on the surface. This leads us to introduce the
increment of a function of two variables and to show how this concept can be
used to define differentiability.
■ TANGENT PLANES
Suppose S' is a surface with the equation z = ∕(x , y), where / has continuous first
partial derivatives, and let P(x0, Fo , z0) be a point on S'. Let Cl be the trace on S' of
the intersection of S' and the plane x = x 0 and C7 be the trace on 5 of the
intersection of 5 and the plane y = y0 , as shown in Figure 12.18. Note that C1 and
C2 must intersect at P. Consider the tangent lines / and T2 to the curves C 1 and
C2, respectively, at the point P. Then, if a tangent plane to the surface S' exists at P,
it is reasonable to expect it to contain the tangent lines T χ and T2 to the curves Cl
and C7, respectively.
In Section 12.6, we will see that if C is any other curve that lies on the surface
S' and passes through the point P, then its tangent line will also lie in the tangent
plane at P. Thus, the tangent plane to S' at P consists of all possible tangent lines at
P to curves that lie on S' and pass through P. The tangent plane at P is the plane
Figure 12.18 Tangent plane that most closely approximates the surface S' near the point P.
Sec. 12.4 Tangent Planes, Approximations, and Differentiability 765
In order to find the equation of this tangent plane, recall that the equation of a
plane is
A x ~ ⅞) + A y - To) + c <z - z o) = 0
j D
If C ≠ 0, divide both sides by C and let a = --^∙> and b = to obtain
z - z0 = α(x - x 0) + b(y - y 0)
Note that the intersection of this plane and the plane x = x 0 is the tangent line T { .
Setting x = x 0 , we obtain
z - z 0 = b ( y - y 0)
which we recognize as the equation (in point-slope form) of a line with slope b.
Therefore, b = f ( x 0 , y 0). Similarly, putting y = y 0 , we obtain
z - z0 = a(x - x 0)
which represents the tangent line T ,, so that a = f x (×0 , y 0).
Equation of the Tangent Plane Suppose S' is a surface with the equation z = f( x , y) and let P 0(x0 , y 0, z0) be a I
point on S at which a tangent plane exists. Then the equation of the tangent
plane to S' at P Q is
z ~ z 0 = 4 (⅞ τ 0)(* - ⅞) + f y (χ0 , y 0 )(y - τ 0)
EXAM PLE 1 Equation of a tangent plane for a surface defined by

z = f(χ , y)
Find an equation for the tangent plane to the surface z = tan 1 -. at the point
P 0( l, V 3 , f).
Solution ∕( x , y ) - X 2 ≤ - - 3 = r→ ∕( 1 , √ 3 ) ≈ = V ∣ = = V 5
∣∣/yz ∖z Λ ∣y j
ι ~
= 4 (1∙ v ≡ ) - ⅛ = i
The equation of the tangent plane is

z - f = ( ξ ¾5 ) ( * - i ) + ⅛ y - √ 3 )
Tangent plane for the surface de

or
fined by z = tan - 1 (y∕x)
3Vz3x - 3y + 12z = 4τr
■ INCREMENTAL APPROXIMATIONS
In Chapter 2, we noted that if/ is differentiable at x 0 and ∆x is sufficiently small,

then the increment Δ ∕ = ∕( x 0 + ∆x) - ∕( x 0) can be closely approximated by
∕' ( x 0)∆ x or, equivalently,
∕( x 0 + ∆x) ≈ ∕( x 0) + ∕' ( x 0) ∆x
The analogous approximation formula for a function of two variables is called the
incremental or (linear) approximation and may be stated as follows:
Incremental Approximation If f(x, y) and its partial derivatives f χ and f y are defined in an open region R
of a Function of Two Variables containing the point P(x0, y0) and f χ and f are continuous at P, then
Δ ∕= f(x 0 + ∆x, v0 + ∆jt) - f(x 0 , y 0) ≈ f χ (x0, y 0) ∆x + f y (x0 , y 0) ∆y
so that
f( x 0 + ∆x, y0 + ∆j>) ≈ f(x 0 , y 0) + f χ (x0 , yo)∆x + f y (x0 , y o)∆y
A graphical interpretation of this incremental approximation formula is shown in

Figure 12.19.
Tangent plane
The height of the surface

z = ∕(x , y) above
Q(x 0 + ∆x, y0 + ∆y) is
approximately the same as the
height of the tangent plane at
po ⅛ ⅞ ) a b o v e 6∙
Figure 12.19 Incremental approximation to a function of two variables
The tangent plane to the surface z = f(x, y) has the equation

* - ∕(⅞ , y 0) = f x (χ 0>T0) Δ -Y + 4(∙λ'o> >,0)∆>'
As long as we are near (x0, j 0), the height of the tangent plane is approximately the
same as the height of the surface. Thus, if ∆x and ∆j,> are small, the point
(x0 + ∆x, y0 + ∆y) will be near (xθ, y0) and we have
∕( x 0 + ∆x, y 0 + ∆y) ≈ ∕(xθ, y0) + f x (x0 , J∕o)∆x + f v(x0 , yo)∆jt
Height of z = f(x, .v) Height of the tangent plane above Q
above ζ)(xll + ∆x, y0 + ∆lj
Increments of a function of three variables f(x , y, z) can be defined in a similar

fashion. Suppose/has continuous partial d e riv a tiv e s /,/,/ at and near the point
(x0 , y 0, z0). Then if the numbers ∆x, ∆y, ∆z are all sufficiently small, we have
Δ ∕= ∕( x 0 + ∆x, y0 + ∆y, z0 + ∆z) - ∕(xθ, y 0, z0)
≈ ∕ Λ⅞> FO>⅞)∆x + ∕( x 0, y 0 , ⅞ )∆j + ∕( x 0 , y0 , zθ)∆z
EXAM PLE 2 Using increments to estimate the change of a function
An open box has length 3 ft, width 1 ft, and height 2 ft, and is constructed
from material that costs S2∕ft2 for the sides and $3/ft2 for the bottom.
Compute the cost of constructing the box and then use increments to estimate
the change in cost if the length and width are each increased by 3 in. and the
height is decreased by 4 in.
Solution An open (no top) box with length x, width y, and height z has an area
A = xy + 2xz + 2yz
Bottom Four sides
Sec. 12.4 Tangent Planes, Approximations, and Differentiability 76 7
Because the sides cost $2/ft2 and the bottom S 3 ∕ft 2, the total cost is
C(x, y, z) = 3xy + 2(2xz + 2yz)
The partial derivatives of C are
C v = 3p + 4z C y = 3x + 4z C_ = 4x + 4v
and the dimensions of the box change by
∆x = ⅛ = 0.25 ft ∆y = ⅛ = 0.25 ft ∆z = ⅛ ft
Thus, the change in the total cost is approximated by

Δ C ≈ C γ(3, 1, 2)∆x + C 1,(3, 1, 2)∆y + C.(3, 1, 2)∆z
= [3(1) + 4(2)](0,25) + [3(3) + 4(2)](0.25) + [4(3) + 4(l)](-⅛) ≈ 1.67
That is, the cost increases by approximately $1.67. l w fe l
EXAM PLE 3 Maximum percentage error using differentials
The radius and height of a right circular cone are measured with errors of at
most 3% and 2%, respectively. Use increments to approximate the maximum
possible percentage error in computing the volume of the cone using these
measurements and the formula V = ∣∙π R 2H.
Solution We are given that

∣ ^ ∣ ≤ 0.03 and ≤ 0.02
IΛ ∣ I rl I
The partial derivatives of V are
V βK= i Jπ R H and Vri„ = o∖τrR2
so the change in V is approximated by
Δ V ≈ ^τrRH) ΔR + (∣τrΛ2) Δ H
Dividing by the volume V = y∏R2H , we obtain
Δy ⅛τRH A R + ∣π¾ 2 A H = (±R\ A tf∖

⅛τR2H ∖ R +
) ∖H )
so that
∣¾T∣ ≤ 2 ∣ ^ ∣ + ∣ ^ ∣ = 2(0.03) + (0.02) = 0.08
Thus, the maximum percentage error in computing the volume V is approxi

mately 8%. ≡≡
■ THE TOTAL DIFFERENTIAL
For a function of one variable, y = ∕( x ) , we defined the differential dy to be

dy = f'( x ) dx. For the two-variable case, we make the following analogous
definition.
Total Differential If z = f( x , y) and ∆x and ∆y are increments of x and y, respectively, and if we

let dx = ∆x and dy = ∆y be differentials for x and y, respectively, then the total
differential of f ( x , y) is
d f= lx dx + fry d y = ¼x' dx + ⅛ x , -r ) dy
Similarly, for a function of three variables z = f( x , y, z), with dz = Az, the total
differential is
dJ f = ~ dx + - £ - djy + ≠ dz
∂X ∂y ∂Z
C O M M E N T : The total differential is useful because it is relatively easy to

calculate, and approximates the increment Δ ∕ which generally is not easy to
compute.
EXAM PLE 4 Total differential
Find the total differential of the given functions:
a. ∕( x , y, z) = 2x 3 + 5y4 - 6z b. ∕( x , y) = x 2 ln(3y2 - 2x)
„ . . r , ∂f ∂f , ∂f ,
Solution a. dj = χ - dx + y - dy + √- dz
J ∂X ∂y ■ ∂z
= 6x 2 dx + 20y3 dy - 6 dz
b. dj∣= ÷ - dx + ÷ - dy
∂x ∂y ■
= Γ2x ln(3y2 - 2x) + x 2 2 " ~l dx + Γx2 ^ y 1 dy
= [2x ln(3y2 - 2x) - ? x \ ] dx + 6 λ y dy

L 3y- - 2χJ 3y2 - 2x '
EXAM PLE 5 Application of the total differential
At a certain factory, the daily output is Q = 60X'ιz2L ιz3 units, where K denotes
the capital investment (in units of $ 1,000) and L , the size of the labor force (in
worker-hours). The current capital investment is $900,000, and 1,000 worker
hours of labor are used each day. Estimate the change in output that will result
if capital investment is increased by $1,000 and labor is decreased by 2
worker-hours.
Solution The change in output is estimated by the total differential dQ. We have
K = 900, L = 1,000, dK = Δ∕√ = 1, and dL = Δ L = - 2 . The total differential
of 2(x, y) is
dQ = % d K + dL
∂K ∂L
= 60(∣)A - , Z2L IZ 3 d κ + 60(∣)A ιz2L - 2z3 dL
= 3 0 A - ,z2L ιz3 dK + 2QK',2 L ~ 2' 3 dL

Substituting for K , L , dK, and dL, we obtain
dQ = 30(900)~ιz2(l,000) ιz3(l) + 20(900) ιz2( 1,000)~2z3( - 2) = - 2
Thus, the output decreases by approximately 2 units when the capital
investment is increased by $1,000 and labor is decreased by 2 worker
hours.
EXAM PLE 6 Maximum percentage error in an electrical circuit
When two resistances 7?, and R 2 are connected in parallel, the total resistance
R is given by
RR '2
R l + R-l
If R t is measured as 300 ohms with a maximum error of 2% and R 1 is
measured as 500 Ω with a maximum error of 3%, what is the maximum
percentage error in R r!
Solution We are given that
dR { C!R2
≤ 0.02 and ≤ 0.03
X
I ∕√ p ∣
and we wish to find the maximum value of ∣ ^ l ∣. Because
->D R -2 3D R 2
∂ R Γ (R l + R 2)2 and ∂R~2 = (R l + R 2) 2 Qu o t ι e n t r u l e
it follows that the total differential of R is

dR = dR.1 + ½R- dR,z
c*Kj σ ∕∖ 9
R 12 R t2
~ (R 1 + R 2)2 d R ' + (R 1 + R 2)2 d p ^2
dR
We now find -∏- by dividing both sides by R; however, because
R .R , ι R. + R-,
7? = we note that J =
R l ÷ R2
(R 1 + R 2y d R ' + (jR 1 + R 2)2 dR \
d R 'R ~
dR _ . + * ι . fl^ 2
R R l + R, R, Rl + R2 R2
Finally, apply the triangle inequality (Table 1.1, p. 00) to this relationship:
dR,
X
s 3MΓ⅛δ<0 02>+ 5≡⅛δ<0∙03>’ 0 02375
The maximum percentage is approximately 2.4%.
■ DIFFERENTIABILITY
Recall from Chapter 2 that if f(x) is differentiable at x 0, its increment is

Δ ∕= ∕( ⅞ + ∆x) - ∕( x 0) = ∕'( ⅞ ) ∆ x + ε∆ x
where ε → 0 as ∆ x → 0 and ε = 0 when ∆x = 0. For a function of two variables,
the increment of x is an independent variable denoted by ∆x, the increment o fy is
Δ ∕ = f ( x 0 + ∆x) - f( x 0 ) = f '( x 0 ) ∆ x + ε∆ x an independent variable denoted by ∆y, and the increment off at (x0, y 0) is defined
Error when using
as
d f to estimate Δ ∕ Δ ∕= ∕( ⅞ + ∆x, p0 + ∆y) - ∕( x 0, y 0)
We use this increment representation to define differentiability as follows.
Definition of Differentiability Suppose f(x , y) is defined at each point in a circular disk that is centered at
(x0 , y0) and contains the point (xθ + ∆x, y0 + ∆y). Then f is said to be
differentiable at (x0, y0) if the increment o f/c a n be expressed as
= 4 ( ⅞ To) δ ∙x + Λ ⅛ To ) Δ T + Ε 1Δ * + ε2 ∆y
δ∕
where ε i → 0 and ε, → 0 as both ∆x → 0 and ∆y → 0 (and ε 1 = ε2 = 0 when

∆x = ∆y = 0). Also, f(x , y) is said to be differentiable on the region R of the
plane if / is differentiable at each point in R.
In Section 2.1, we showed that a function of one variable is continuous wherever it

is differentiable. The following theorem establishes the same result for a function
of two variables.
THEOREM 12.2 Differentiability implies continuity

If f(x , y) is differentiable at (x0 , y 0), it is also continuous there.
Proof: We wish to show that ∕(x , y) → ∕( x 0, y0) as (x, y) → (x0, y 0), or,
equivalently, that
. Hm [∕(x , y) - f(x u0 , τ υ0)] = 0
(v)→⅛⅜)
If we set ∆x = x - x0 and ∆p = y - y 0 and let Δ ∕ denote the increment of / at
(x0, TQ), w e have (by substitution)
f(x , T) - f ix 0 , To) = ∕( ⅞ + Δ X >TO + Δ T) - f(x 0, To) = δ ∕
Then, because (∆x, ∆y) → (0, 0) as (x, y) → (x0, y 0), we wish to prove that
lim Δ ∕= 0
(∆x, Δy)→(0,0)
Because/is differentiable at (x0 , p0), we have

4 / = ∕ v(x o>T0) Δ ∙* + Λ∙(⅞> T0) Δ T + Ε Δ
1 -V+ ε2 ∆y
where ε 1 → 0 and ε, → 0 as (∆x, ∆y) → (0, 0). It follows that
(∆Λ∙. f e ( 0 , 0 ) = (∆,v, l ⅛ 0 , 0 ) ‰ *>>δ * + ∙ W *)) Δ T + ≡, Δ * + ¾∆T1
= [∕√⅞ , To)] ■0 + [fy (x0, τ o)] ∙ 0 + 0 + 0 = 0

as required.
0 Be careful about how you use the word differentiable. In a single variable
case, a function is differentiable at a point if its derivative exists there.
However, the word is used differently for a function of two variables. In
particular, the existence of the partial derivatives/ and/, does not guarantee
that the function is differentiable. 0
EXAM PLE 7 Think Tank example: Possible existence of partial derivatives

with a nondifferentiable function
Let
1 if x > 0 and y > 0
0 otherwise
That is, the function f has the value 1 when (x, y) is in the first quadrant and 0
elsewhere. Show that the partial derivatives/, a n d / exist at the origin, but f is
not differentiable there.
Solution We first note that at the origin, the function has the value 0; that is,
In plane z = 1, /(0, 0) = 0. The key to showing that / exists at (0, 0) depends on the fact that
approach is to 1.
f(χ,y ) / ( 0 + ∆x, 0 ) - / ( 0 , 0 )
∆x υ
In ΛV -plane which is true because ∕(x , y) = 0 along the x-axis (not just at the origin). This
approach is
to 0.
means that/(0 , 0) = 0. Similarly,/(0 , 0) = 0, which verifies that the partial
derivatives exist at the origin.
Next, we need to show that ∕(x , y) is not differentiable at the origin. We
recall that differentiability implies continuity (Theorem 12.2), so if we can
show that / is not continuous at (0, 0), we can conclude that it is not
differentiable at this point.
1. /(0, 0) = 0, so /(0, 0) is defined.

2. lim ∕(x , y) = 1 along the line y = x in the first quadrant.
(x,y)→ (0, 0)
T hus,/is not continuous at (0, 0) because

lim ∕( x , y ) ≠ /(0 ,0 )
(x,y)→ (0,0)
even if/ exists at (0, 0). Therefore,/is not differentiable at (0, 0).
Although the existence of partial derivatives at P(x0 , y0) is not enough to

guarantee that ∕(x , y) is differentiable at P, we do have the following sufficient
condition for differentiability.
THEOREM 12.3 Sufficient condition for differentiability
I f /is function of x and y and/ / , and / are continuous in a disk D centered at

(x0 , y0), th e n /is differentiable at (x0, y0).
Proof: The proof is found in advanced calculus. ■
EXAM PLE 8 Establish differentiability
Show that ∕(x , y) = x 2y + xy 3 is differentiable for all (x, y).

772 C h . 12 Partial Differentiation
Solution We shall apply the sufficient condition for differentiability. First,

compute the partial derivatives
∕√ * , = + x y ^ = 2 x y + jp
f y(χ, y) = -^ (χ 2 y + ∙χ τ 3) = χ2 + 3ΛT2
Because f, f χ , and f are all polynomials in x and y, they are continuous
throughout the plane. Therefore, the sufficient condition for the differenti
ability theorem assures us that f must be differentiable throughout the entire
plane. --
PROBLEM SET 1 2 .4
θ In Problems 1-6, fin d standard-form equations for the tangent 15. f i x . y, z) = 3x 3 - 2y4 + 5z
plane to the given surface at the prescribed point P o on the 16. f ( x , y, z) = sin x + siny + cosz
surface. 17. f { x , y, z) = z 2sin(2x - 3y)
1. z = V x 2 + y 2 2. z = 10 —x 2 —y 2
18. f i x , y, z) = 3y2z cosx
a tP 0(3, 1, √Tθ) a t P 0(2, 2, 2)
Show that functions in Problems 19-22 are differentiable for
all (x, y).
19. f i x , y) = xy 3 + 3xy 2 20. f i x , y) = x 2 + 4x - y 2
3
2 1 .∕( x , y) = x y 2 22. f i x , y) = x + 4y - y 2
θ Use an incremental approximation to estimate the functions at
the values given in Problems 23-28. Check by using a calcula
tor and direct calculation.
23. /(1.01, 2.03), where f i x , y) = 3x4 + 2y4
24. /(0.98, 1.03), where∕( x , y) = x 5 - 2y 3
25. ∕ ( f + 0.01, j - 0.01), where∕( x , y) = sin(x + y)
26. f(fi⅛ + 0.01, - 0.01), where ∕( x , y) = sin(xy)

x siny 27. /(1.01, 0.98), where f i x , y) = eλ'l'
3. f i x , y) = x 2 + y 2 + sinxv at 4. f i x , y) = e at
'po = (0, 2, 4) 28. /(1.01, 0.98), where f i x , y) = eχ l y 2
p ( o —, ι ]
0∖ ’ 2 / z 29. Find an equation for each horizontal tangent plane to the
surface
z = 5 - x 2 - y 2 + 4y
30. Find an equation for each horizontal tangent plane to the
surface
z = 4(x - 1)2 + 3(y + 1)2
31. a. Show that if x and y are sufficiently close to zero and /

is differentiable at (0, 0), then
f i x , y) ≈ /(0, 0) + x f fθ , 0) + yfy lθ, 0)

5. z = tan 1-
at P 0n ( 2, 2, ? )
x ∖ ’ ’ 4/ b. Use the approximation formula in part a to show that
6. z = ln∣x + y 2 ∣ at P o( - 3, 2, 0)
Find the total differential o f the functions given in Problems 1+ x - y

7-18.
for small x and y.
7. f i x , y) = 5x 2y 3 8. f i x , y) = 8x 3y 2 - x 4y 5
c. If x and y are sufficiently close to zero, what is the
9. f i x , y) = sinxy 10. f i x , y) = cosx 2y
approximate value of the following expression?
ll .∕ ( x , y ) = y
12. f i x , y) =
___________1___________
13. f i x , y) = ye x 14. f i x , y) = eχ 2 + y (x + 1)2 + (y + 1)2
32. If an electrical circuit with two resistors connected in for a constant c > 0. Estimate the percentage change in R
parallel has a total resistance of R ohms, then that results when x increases by 3% and r is decreased by
2%.
± =l ±
R P+Q 38. In one mole of an ideal gas, the volume V, pressure P, and
absolute temperature T are related by the equation
where P and Q are the resistances (in ohms) of the PV = RT, where R is a certain fixed constant that de
resistors. Suppose P and Q are measured at 6 and 10 pends on the gas. Suppose we know that if T = 400o
ohms, respectively, and that each measurement is accu {absolute) and P = 3,000 lb∕ft 2, then V = 14 ft 3. Approxi
rate to within 0.01 ohms. What is the maximum percent mate the change in pressure if the temperature and
age error involved in the computation? volume are increased to 403o and 14.1 ft 3 , respectively.
33. A closed box is found to have length 2 ft, width 4 ft, and 39. If x gram-moles of sulfuric acid are mixed with y gram
height 3 ft, where the measurement of each dimension is moles of water, the heat liberated is given by
made with a maximum possible error of ±0.02 ft. The top
1.786 xy
of the box is made from material that costs $2/ft2 , and ^ ^ = 1 .7 9 8 x + y c a l
each square foot of the sides and bottom costs only
S1.50∕ft2 . What is the maximum possible error involved Approximately how much additional heat is generated if a
in the computation? mixture of 5 gram-moles of acid and 4 gram-moles of
34. A cylindrical tank is 4 ft high and has a diameter of 2 ft. water is changed to a mixture of 5.1 gram-moles of acid
The walls of the tank are 0.2 in. thick. Approximate ΔI-z and 4.04 gram-moles of water?
assuming the tank has a top and a bottom that are both 40. An editor estimates that if x thousand dollars are spent on
also 0.2 in. thick. development and y thousand dollars, on promotion, ap
35. The Higrade Company sells two brands, X and Y, of a proximately
commercial soap in thousand-pound units. If x units of β(x, y) = 20x 3,2y
brand X and y units of brand Y are sold, the unit price for
brand X is copies of a new book will be sold. Current plans call for
p(x) = 4,000 - 500x the expenditure of $36,000 on development and $25,000
on promotion. Use the total differential of Q to estimate
and that of brand Y is the change in sales that will result if the amount spent on
q(y) = 3,000 - 450y development is increased by $500 and the amount spent
on promotion is decreased by $500.
a. Find an expression for the total revenue R(x, yj ob 41. A grocer’s weekly profit from the sale of two brands of
tained by selling x units of brand X and y units of orange juice is
brand Y.
b. Suppose brand X sells for $500 per unit and brand Y P(x, y) = (x - 30)(70 - 5x + 4j,)
sells for $750 per unit. Estimate the change in total + (y - 40)(80 + 6x - ly)
revenue if the unit prices are increased by $20 for dollars, where x is the price per can in cents of the first
brand X and $ 18 for brand Y. Hint: Find the change in brand and y is the price per can in cents of the second.
x and the change in y that correspond to p(x) increas Currently, the first brand sells for 50Φ per can and the
ing from 500 to 520 and q(y) increasing from 750 to second, for 52¢ per can. Use the total differential to
768. estimate the change in the weekly profit that will result if
36. The output at a certain factory is the grocer raises the price of the first brand by 1Φper can
and raises the price of the second brand by 2Φ per can.
Q = 15OK2' 3 L l ' 3
42. A juice can is 12 cm tall and has a radius of 3 cm. A
where K is the capital investment in units of $ 1,000 and L manufacturer is planning to reduce the height of the can
is the size of the labor force, measured in worker-hours. by 0.2 cm and the radius by 0.3 cm. Use a total differen
The current capital investment is $500,000 and 1,500 tial to estimate how much less the volume will be in each
worker-hours of labor are used. can after the new cans are introduced.
a. Estimate the change in output that results when capital 43. It is known that the period T of a simple pendulum with
investment is increased by $700 and labor is increased small oscillations is given by
by 6 worker-hours. 7' - 2'v⅞
b. What if capital investment is increased by $500 and
labor is decreased by 4 worker-hours?
where L is the length of the pendulum and g is the
37. According to Poiseuille’s law, the resistance to the flow of
acceleration due to gravity. For a certain pendulum, it is
blood offered by a cylindrical blood vessel of radius r and
known that L = 4.03 ft. It is also known that g = 32.2
length x is
ft∕s 2 . What is the approximate error in calculating T by
A(r, x) =
using L = 4 and g = 32?
44. If the weight of an object in the air is x pounds and its 46. Compute the total differentials
weight in water is y pounds, then the specific gravity of
the object is and
Why are these differentials alike?

47. Let A be the area of a triangle with sides a and b separated
For a certain object, x and y are measured to be 1.2 lb and by an angle θ, as shown in Figure 12.20.
0.5 lb, respectively. It is known that the measuring
instrument will not register less than the true weights, but
it could register more than the true weights by as much as
0.01 lb. What is the maximum possible error in the
computation of the specific gravity? b
θ 45. Show that the following function is not differentiable at Figure 12.20 Problem 47
(0, 0): Suppose 0 = f and a is increased by 4% while b is
xy decreased by 3%. Use differentials to estimate the per
if (x, y) ≠ (0, 0) centage change in A.
f( x , ?) = ■ x 2 + y 2
.0 if(x ,y ) = (0, 0) 48. In Problem 47, suppose that θ also changes by no more
than 2%. What is the maximum percentage change in A l
1 2 .5 CHAIN RULES
IN THIS SECTION Chain rule for one parameter, chain rule for two
parameters
Functional composition is as important in several-variable calculus as in the
single-variable case, and now because more than one variable are available, it is
possible to form a “function of a function” in a variety of ways. We shall
examine several derivatives and partial derivatives in each case.
■ CHAIN RULE FOR ONE PARAMETER
We begin with a differentiable function of two variables ∕(x , y). If x = x(t) and
y = y(Z) are, in turn, functions of a single variable t, then z = f[x(f), y(t)] is a
composite function of a parameter t. In this case, the chain rule for finding the
derivative with respect to one independent variable can now be stated.
THEOREM 12.4 The chain rule for one independent parameter

Let ∕(x , y) be a differentiable function of x and y, and let x = x(f) and y = y(t) be
Sec. 12.5 Chain Rules 775
Proof: Recall that because z = ∕( x , y) is differentiable, we can write the

increment ∆z in the following form:
∆z —⅛ ∆ x + ⅛z∆v + ε.1 ∆x + ε, ∆y
2 j
∂x ∂y ∙
where ε 1 → 0 and ε, → 0 as both ∆x → 0 and ∆y → 0 and ε 1 = ε2 = 0 when
∆x = ∆y = 0. Dividing by ∆Z ≠ 0, we obtain
∆z _ az ∆x 3Z ∆y ∆x ∆y
∆F- ax∆ F + a j∆ 7 + ε i∆ F + ε 2∆F
Because x and y are functions of t, we can write their increments as

∆x = x(Z + Δ ∕) - x(Z) and ∆y = y(t + ∆Z) - y(t)
We know that x and y both vary continuously with t (remember, they are
differentiable), and it follows that ∆x → 0 and ∆y → 0 as ∆Z → 0, so that ε, → 0
and ε2 → 0 as ∆t → 0. Therefore, we have
$ = lim = lιm ⅛ ^ x ∂z ∆y ∆x ∆ ∕∣
dt ∆r→o ΔZ ΔI →O
Lax ∆z ay ∆z > ∆z 2 ΔZJ
∂z dx ∂z dy ^ d x dy
= -77 + -77 + 0 - 7- + λ0 - 7 -
ax dt ∂y dt dt dt ■
EXAM PLE 1 Verifying the chain rule explicitly
Let z = x 2 + y 2 , where x = ∣ and y = t2 . Find in two ways:
a. by first expressing z explicitly in terms of t.

b. by using the chain rule.
Solution a. By substituting x = γ and y = t2 , we find that
∕ 1 ∖2
z = x 2 + y 2 = (γ j + (Z2)2 = Γ 2 + Z4 for Z ≠ 0
Thus, ⅛ = -2 Z " 3 + 4Z3.
b. Because z = x 2 + y 2 and x = t~', y = t2 , so

⅛½= 2xj ^ = 2y, ⅛ = - Γ 2- ⅛i = 2t
∂x , ∂y - dt ’ dt
Use the chain rule for one independent parameter:
<⅛_ az dx ∂z~dy
dt - ∂x dt + ∂y dt
= (2x)(-Z~2) + 2y(2Z) Chain rule

= 2(Z- 1 )(-Z^ 2) + 2(Z2)(2Z) Substitute.
= -2Z^ 3 + 4Z3
EXAM PLE 2 Chain rule for one independent parameter
Let z = V x 2 + 2xy, where x = cos θ and y = sin θ. Find

Solution j ∣ = ^(x 2 + 2xp)^ιz2(2x + 2p) and ∣ ∣ = ∣ ( x 2 + 2xy) ιz2(2x)
Also, = -s in θ and = cos θ. Use the chain rule for one independent
parameter to find
dz _ ∂z dx ∂z dy
dθ ~ ∂x dθ + ∂y dθ
= ∣(x 2 + 2xy)~ l,2 (2x + 2y)(-sin θ) + ∣(x 2 + 2xy) ιz2 (2x)(cos θ)
= (x 2 + 2xy)~ l,2 (x cos θ - x sin θ - y sin θ)
EXAM PLE 3 Related rate application using the chain rule
A right circular cylinder is changing in such a way

that its radius r is increasing at the rate of
3 in./min and its height h is decreasing at the rate
of 5 in./min. At what rate is the volume of
the cylinder changing when the radius is 10 in.
and the height is 8 in.?
Solution The volume of the cylinder is V = τrr2h, and we are given = 3 and
= - 5 . We find that
⅛^=ττ(2r)∕z and πr2(l)
By the chain rule for one parameter:

d V _ ∂ V dr , ∂ Jz dh = 2 π r h ⅛ + π r 2 ⅛
dt ∂r dt ^t ^ ∂h dt dt dt
Thus, at the instant when r = 10 and h = 8, we have
^=2 π (1 0)(8 )(3) + ττflO)2(-5 ) = -2 0 ττ≈ -62.83185307
The volume is decreasing at the rate of about 62.8 in.3∕m in.
EXAM PLE 4 Second derivative of a function of two variables
Let z = f( χ , y), where x = at and y = bt for constants a and b. Assuming all

necessary differentiability, find d 2z∕dt2 in terms of the partial derivatives of z.
„ , . , dx dy .
Solution We note that = a and = b, so that
dz ∂z dx ∂z dy ~
d i = ∂ ⅛ d t+ ∂ } d t c ht a m ' u le
Differentiate both sides with respect to t:

d2z _ ∂z d 2x ^d x ∖ ∂ 2z dx ∂2z dy ∂z d 2y dy∖ ∂2z dx ∂2z dy∖
dt2 ∂x dt2 dt Lax2 dt ∂x∂y dt. ∂y dt2 + dt[∂y∂x dt + ∂y2 dt∖
Apply the chain rule again.
=g <°)+“[∂x2 + - ( 0 ) + b ∖ -^- a +⅛
∂x∂y ∂y Layax ∂y2
= fl 2 f⅛2 + 2 ^ ⅛ + ^ ⅛ 2 ∂2z _ ∂2z

Note
∂x ∂x∂y ∂y ∂x∂y ∂y∂x'
■ CHAIN RULE FOR TWO PARAMETERS
Next, we shall consider the kind of composite function that occurs when x and y
are both functions of two variables. Specifically, let z = F(x, y), where x = x(u, v)
and y = y(μ, v) are both functions of two independent parameters u and v. Then
z = F[x(w, v), y(u, v)] is a composite function of u and v, and with suitable
assumptions regarding differentiability, we can find the partial derivatives ∂z∕∂u
and ∂z∕∂v by applying the chain rule obtained in the following theorem.
THEOREM 12.5 The chain rule for two independent parameters

Suppose z = f(x, y) is differentiable at (x, y) and that the partial derivatives
of x = x(μ, v) and y = y(ιι, v) exist at (it, v). Then the composite function
z = f[x(u, v), y(u, v)] is differentiable at (M, v) with
∂z _ ∂z ∂x ∂z ∂y ∂z _ ∂z ∂x ∂z ∂y
∂u ∂x ∂u + ∂y ∂u a n θv ∂x ∂v ^l^ ∂y ∂v
Proof: This version of the chain rule follows immediately from the chain rule
for one independent parameter. For example, if v is fixed, the composite function
z = f[x(u, v), y(u, v)] really depends on u alone, and we have the situation
described in the chain rule of one independent variable. We apply this chain rule
with a partial derivative instead of an “ordinary” derivative (because x and y are
functions of more than one variable):
∂z _ ∂z ∂x ∂z ∂y
∂u ∂x Bit + By Bu
The formula for ⅜2 can be established in a similar fashion. _
∂v ■
EXAMPLE 5 Chain rule for two independent parameters
Let z = 4x - ∙yΛ 2’, whereJ x = uv2 and yB=

U
u 3v. BFind
V
χi and
Solution First find the partial derivatives:

= ^^(v4 'V - zy 2) = 4 = y-(4x - yλ 2) = -2 y2
3x ax ’ ∂y ∂yκ ∙ ’
and
g =⅛("+--1 ta-⅛" 5")- 3"2''
⅛ = ⅛ ( M - 2llV ⅛ = A(,,3V).,,3
Therefore, the chain rule for two independent parameters gives

∂z _ ∂z ∂x ∂z ∂y
∂u ∂x ∂u ∂y ∂u
= (4)(v2) + (-2y)(3w 2v) = 4v2 - 2(W3V)(3M2V) = 4v2 - 6w5v2
and
∂z _ ∂z ∂x ∂z ∂y
∂v ∂x ∂v + ∂y ∂v
= (4)(2MV) + ( - 2 j ,)(w3) = 8MV - 2(M3V)W3 = 8zzv - 2 M6V
The chain rules can be extended to functions of three or more variables. For
instance, if w = f( x , y, z) is a differentiable function of three variables and
x = x(f), y = y(Z), z = z(t) are each differentiable functions of t, then w is a
differentiable composite function of t and
dw _ Tv dx ∂w dy Tv dz
dt ∂x dt + ∂y dt + ∂z dt
In general, if w = ∕( x 1, x 1 , . . . , x w) is a differential function of the n variables x l ,
x 2 , . . . , x , which in turn are differentiable functions of m variables tl , t2, . . . , tm ,
th e n ∂w _ ∂w ∂x i ∂w ∂x2 ∂w ∂xn
∂tl ∂x ∣∂tl + ∂x~l ∂ti + + ∂χ^
∂H' 3ιv ∂x 1 3w ∂x. T v ∂xn

^∂ζ = + w
2 τrn
EXAM PLE 6 Chain rule for a function of several variables
Find ∣ jy if vv = 4x + v2 ÷ z 3 w jι e r e x = e∣
s2^y = ]n /+ L anc ∣z = r s f-
Solution
∂w _ <9vv ∂x ∂w ∂y θw ∂z
∂s ∂x ∂s + 3y ∂s + ∂z ∂s
= 8rsers2 + - ~ - + 3rt2z 2
In terms of r, s, and t, the partial derivative is
⅞ψ= 8rse^ + 7 I iSιn L ±t Λ

OS
+ 3 r 3s 2t 6
PROBLEM SET 1 2 .5
θ 1. ■ What Does This Say? Discuss the various chain rules 20. Find where w = e 2x y + 2 *2 and x = r + s - t,
∂r
and the need for such chain rules.
y = 2r - 3s, z = cos rst.
2. ■ What Does This Say? Discuss the usefulness o f the
schematic representation for the chain rules. 21. Find ~ and ⅞y> where w = ⅜-+ and
j
x = 2rs, y = sin rt,
∂r ∂t 2 —z ’
3. ■ What Does This Say? Write out a chain rule for three z = st 2
independent parameters and two intermediate variables.
θ Find the following higher partial derivatives in Problems
In Problems 4 - 7, the function f(x, y) depends on x and y, which
22-27.
in turn are both functions o f t. In each case, fin d dz/dt two
different ways: a. b. — ∑ c —
a. Express z explicitly in terms o f t. ∂x∂y ∂x 2 ' ∂y2
3 2
22. x + y + z = 5 2
23. xyz = 2
b. By using the chain rule for one parameter.
4. f i x , y) = 2xy + y 2 , where x = - 3 t 2 and y = 1 + ti 24. ln(x + y) = y 2 + 1 25. x ~ i + y~1 + z~' = 3
5. f ( x , y) = (4 + y 2 )x, where x = e 2t and y = e i t 26. x cosy = y + z 27. z 2 + sinx = tany
6. f i x , y) = (1 + x 2 + y 2) lz2, where x = cos 5t and y = sin 5/ 28. Let f i x , y) be a differentiable function o f x and y, and let
x = r cos θ, y = r sin θ for r > 0 and 0 < θ < 2π.
7. f i x , y) = x y 2 , where x = cos 3t and y = tan t
a. If z = f ∖x{r, θ), y(r, 0)], find ∣⅞ and ∣⅞
In Problems 8-11, the function F(x, y) depends on x and y. Let
x = x(u, v) and y = y(u, v) be given functions o f u and v. Let z b. Show that
= F[x(u, v), y(u, v)] and fin d the partial derivatives ∂z∕∂u and
∂z∕∂v in these two ways: ∖2 , /- ∖2 - ∖2 /- ∖2
∂ z∖ . 1 i ∂z1 ∂Z ) , I ∂ Z ]
a. Express z explicitly in terms o f u and v. dr) r2 ∖∂θ) ∂x) ∖∂y)
b. Apply the chain rule for two independent parameters.
Let z = f i x , y), where x = an and y = bv, with a, b
8. F ix , y) = x + y 2 , where x = u + v and y = u - v
constants. Express ∂2z∕∂u2 and ∂2z∕∂v2 in terms of the
9. F ix , y) = x 2 + y 2, where x = u sin v and y = u —2v partial derivatives o f z with respect to x and y. Assume the
10. F ix , y) = ex y , where x = u — v and y = u + v existence and continuity o f all necessary first and second
11. F ix , y) = In xy, where x = e m'2, y = e ,ιv partial derivatives.
Write out the chain ride fo r the functions given in Problems Let (x, y, z) lie on the ellipsoid
12-15.
12. z = f i x , y), where x = x ls, t), y = y(s, t)
13. w = f i x , y, z), where x = x ls, t), y = y(.v, /), z = z(s, t)
14. w = fiu , v), where u = M(X , y, z, w), v = v(x, y, z, w) Without solving for z explicitly in terms of x and y,
compute the higher partial derivatives
15. w = f i x , y, z), where x = x ls, t, u), y = y ls, t, u), z = z(s, t, u)
Find the indicated derivatives or partial derivatives in Problems ∂2z ∂2z
and
16-21. Leave your answers in mixed form (x, y, z, t, ιι, v). ∂x 2 ∂x∂y
16. Find where w = ln(x + 2y - z 2) and x = 2 t - l , 31. The dimensions o f a rectangular box are changing in such
a way that the length ^(t), width wit), and height h f) are
y = γ z = V t.
linear functions of time t. If the length and width are
17. Find ⅛> where w = sin xyz and x = 1 — 3t, y = e i ~t , increasing at 2 in./sec and the height is decreasing at 3
in./sec, find the rates at which the volume V and the
z = 4t. surface area S are changing with respect to time. If
f(0) = 10, w(0) = 8, and ⅛(0) = 20, is V increasing or
18. Find where w = zexy2 and x = sint, y, = co s∕,
decreasing when t = 5 sec? What about S when t = 5?
z = tan2t. 32. Using x hours of skilled labor and y hours of unskilled
19. Find where w = eχ 2 + yz and x = y ’ y = ln(2∕ - 3), labor, a manufacturer can produce f ( x , y) = 10xy 1'2 units.
Currently, the manufacturer has used 30 hr of skilled
z = t2 . labor and 36 h of unskilled labor and is planning to use
1 additional hour of skilled labor. Use calculus to esti 43. Let / and g be twice differentiable functions of one
mate the corresponding change that the manufacturer variable, and let
should make in the level of unskilled labor so that the
u(x, f) = f ( x + ct) + g (x - ct)
total output will remain the same.
In Problems 33-39, assume that all functions have whatever for a constant c. Show that = c2 2
∂t2 ∂x
derivatives or partial derivatives are necessary for the problem
Hint: Let u = x + ct; v = x — ct
to be meaningful.
44. Suppose z = f ( x , y) has continuous second-order partial
33. If z = ∕( u v 2), show that
derivatives. If x = e'cos θ and y = ersin θ, show that
2w F - vF
∂v =°
∂ 2Z , ∂ 2Z _ .-2 rΓ δ ⅛ , ∂ 2 Z ~I
∂u
∂x 2 ∂y2 L∂r2 ∂02J
Hint. Let ιv = uv2 and apply the chain rule.
θ 45. Suppose F is a differentiable function o f x and y and that
34. If z = u + f ( u 2v2), show that
F(Λ', y) = C for some constant C in an open set S where
∂z ∂z y = y(x) is a differentiable function of x. Show that
u - -----v — = u
∂u ∂v
Hint. Let w = ιι2v 2 and apply the chain rule.
35. If z = f{u — v, v — it), show that 46. Let F(x, y, z) be a function o f three variables with
continuous partial derivatives F γ, F y, F . in a certain
∂z ∣ ∂z _ ∩ region where F(x, y, z) = C for some constant C . Use the
∂u ∂ v ~ chain rule for two parameters and ∂y!∂x = 0 to show that
36. If z = u + jf(uv),
v show that u z f - v z f = u. ∂Z ∂X ∑ fFx
’ ∂ιι ∂v = =
∂x ∂F Fz
37. If w = f ( - ---- - ∖ show that r + s 4y- = 0.
j ∖ s J ∂z
∂r ∂s
47. Suppose the system
38. If z = jf ( x 2 - y 2), evaluate y ~ + x z f
- ' - ∂x ∂y xu + yv - uv = 0
39. If z = xy + f ( x 2 + y 2 ), show that yu - xv + uv = 0
can be solved for u and v in terms of x and y , so that
^ - x - ^ = yj 2 - x 2
u = ιι(x, y) and v = v(x, y). Use implicit differentiation to
∂x ∂y
find the partial derivatives ∂u∕∂x and ∂v∕∂x.
40. Let w = f(f) be a differentiable function of t, where 48. Repeat Problem 47 for the system
t = (x 2 + y 2 + z 2) 172. Show that
'xu + yv - uv = 0
yu - xv + uv = 0
for ∂u!∂y and ∂v!∂y.
41. Suppose /’is a twice differentiable function of one varia 49. Suppose that F and G are functions o f three variables and
ble, and let z = f ( x 2 + y 2). Find that it is possible to solve the equations F(x, y, z) = 0 and
G(x, y , ∑) = 0 for y and z in terms o f x , so that y = y(x)
a, a⅛ b. c d⅛
∂x 2 ∂y2 ∂x∂y and z = z(x). Use the chain rule to express dy/dx and
ρ
c z
dz!dx in terms of the partial derivatives of F and G.
42. Find —≡ where/is a twice differentiable function of one Assume these partials are continuous and that
uθ~
variable and z = /(cos θ, sin θ). Hint: Let x = cos θ and ∂F ∂G , ∂F ∂G
y = sin θ. Leave your answer in terms of x and y. ∂y ∂z ∂z ∂y
12.6 DIRECTIONAL DERIVATIVES AND THE GRADIENT
IN THIS SECTION The directional derivative, the gradient, normal

property o f the gradient, tangent planes and normal lines
In Section 12.3, we observed that if PQ is a point on a surface, then the slopes of
the tangent lines through Po on the surface parallel to the xz- or yz-plane are
given b y /. or f respectively, as shown in the illustration.
Sec. 12.6 Directional Derivatives and the Gradient 781
In this section, we seek to find a derivative in other directions, called the

directional derivative. After we find the directional derivative, we will express it
in terms of a vector function called a gradient. We conclude by looking at some
applications of the gradient.
■ THE DIRECTIONAL DERIVATIVE
In Chapter 1 we defined the slope of a curve at a point to be the ratio of the change
in the dependent variable with respect to the change in the independent variable at
the given point. In order to determine the slope of the tangent line at a point
-f,0(x0, To) 0 n a surface defined by z = f(x, y), we need to specify the direction in
which we wish to measure. We do this by using vectors. In Section 12.3 we found
the slope parallel to the xz-plane to be the partial derivative f x (x0, y0). We could
have specified this direction in terms of the vector i (x-direction). Similarly,/ , (x, v)
could have been specified in terms of the j vector. Finally, to measure the slope of
0 u must have unit length. 0 the tangent line in an arbitrary direction, we use a unit vector u = w1i + u2j in that
direction.
To find the desired slope, we look at the intersection of the surface with the
vertical plane passing through the point Po parallel to the vector u, as shown in
Figure 12.21. This vertical plane intersects the surface to form a curve C, and we
define the slope of the surface at PQ in the direction of u to be the slope of the
tangent line to the curve C defined by u at that point.
Figure 12.21 The directional derivative

D ∕( x 0, >,0) is the slope of the tangent
line to the curve on the surface
z = f(x, y) in the direction of the unit
vector u at P0(x0, v0).
We summarize this idea of slope in a particular direction with the following

definition.
Directional Derivative Let f be a function of two variables, and let u = M1i + w2j be a unit vector. The
directional derivative of f at Pθ(xθ, y0) in the direction of u is given by
n ,, . ∕( ⅞ + hu γ, y 0 + hu2) - ∕( x 0, To)
D u f ^r J7θ) = I™ -------------------h-------------------
provided the limit exists.
EXAM PLE 1 Computing partial derivatives as directional derivatives
Use the definition of directional derivative to show that the directional

derivative of ∕'(x, v) in the directions of the positive x- and y-axes a re ∕ v and ∕ ι'j
respectively.
Solution In the direction of the positive X-axis, u = i, so that u l = 1 and w2 = 0.
∕( ⅞ + h, y0) - ∕( x 0, y0)
D∕i√f(∖x nt r, ∙½U'1) = I jlimQ Definition o f directional
z
> h derivative, where w = 1
1
and w2 = 0
= A(⅞> Lo) Definition off χ

In the direction of the positive y-axis. u = j, so that w1 = 0 and M2 = 1.
∕(⅞> y 0 + h) - ∕( x 0, y 0)
Dj f( x 0 , y 0 ) = lim Definition o f directional
h derivative, where u = 0l
and u2 = 1
= ∕√⅞> Lo) Definition o f f ,
The definition of directional derivative is similar to the definition of the

derivative of a function of a single variable. Just as with a single variable, it is
difficult to apply the definition directly. Fortunately, the following theorem allows
us to find directional derivatives more efficiently than by using the definition.
THEOREM 12.6 Directional derivatives using partials
Let f(x , y) be a function that is differentiable at P0(x0 , J o). Then f has a directional
derivative in the direction of any unit vector u = Ml i + w7j. That directional
derivative is given by
1∖
f( χ 0>y 0) = A(⅞> T0)w ι + f y (χ 0 , L0 ) W2
Proof: We define a function F of a single variable h by
F(∕ι) = f(x 0 + hu l , y0 + hιι2), so that
n ,, . .. f ( x 0 + hu i , y 0 + h ι ι f i - f ( x 0 , y 0)
D .,Λ ⅜. y«) = ⅛ ------------------ *-------------------
ιιm ≡ →<o) = F'(0)

A→O n
Apply the chain rule with x = x () + hu γ and y = y 0 + hu2 ,

F '^ = î = d x Jh + d y Jh = ⅛ x >y )u ι + -4(x ’ -^)M 2
When h = 0, we have x = x 0 and y = y 0 , so that

n∕, r)Γ
D u f ( 0> TQ) ~
x (θ) ~ ^Qχ l l ∖ + g y U 2 ~ ∕χ ( x 0, ∙^θ) U \ + ∕y(-λ 05 FQ) U 2
E X A M P L E 2 Finding a directional derivative using partials
Find the directional derivative of∕( x , y) = 3 - 2x 2 + y i at the point P ( l, 2) in

the direction of the unit vector u = ∣ i - X¾.
Solution First find the partial derivatives ∕ v(x, y) = - 4 x and f (x, y) = 3j'2.
Notice that u i = ∣ and u2 = - V 3 ∕2 , so
D a ∕( l , 2) = / / 1 , 2)(⅛)+ / / 1 ,2 ) ( - ^ )
= -4(1)(∣) + 3(2)2( - ^ ) = - 2 - 6√3 ≈ -1 2 .4 _
Notice that the directional derivative is a number. This number can be interpreted
as the slope of a tangent line to z = f( x , y) or as a rate of change of the function
z = Λ × , T)∙
Directional derivative as a slope. Geometrically, we look at the surface

defined by z = 3 - 2x 2 + y 3, as shown in Figure 12.22. The intersection of this
surface with a plane aligned with the vector u = ∣ i- is a curve labeled C , and
the directional derivative is the slope of the tangent line to C at the point on the
surface above P ( l, 2). This directional derivative is shown in Figure 12.22.
a. The point ( 1, 2) in d. Directional derivative

the xy-plane ∕d . 2 )
Figure 12.22 Graph of z = 3 - 2x 2 + y 3 and some derivatives

Directional derivative as a rate. The directional derivative Duf( x 0, y0) can

also be interpreted as the rate at which the function z = f(x , y) changes as a point
moves from P0(x0, y 0) in the direction of the unit vector u. Thus, in Example 2, we
say that the function f(x ,y ) = 3 - 2x2 + y 3 changes at the rate o f - 2 - 6√3 as we
move from P(l, 2) in the direction of the unit vector u = ∣i -
■ THE GRADIENT
The directional derivative Duf(x , y) can be expressed concisely in terms of a

vector function called the gradient, which has many important uses in mathemat
ics. The gradient of a function of two variables may be defined as follows.
Gradient Let / b e a differentiable function at (x, y) and let f(x , y) have partial derivatives
∕ v(χ, y) and∕.(λ', y). Then the gradient of / denoted by V ∕ (pronounced “del
eff”), is a vector given by
vf(χ, y) = f x (χ, y)» + f y (χ, y)j
0 Think of the symbol V as an “operator” on a function that produces a

vector. Another notation for V/is grad f(x , y). 0
EXAMPLE 3 Finding the gradient of a given function

Find V∕(x, y) for ∕(x , y) = x 2y + y 3.
Solution Begin with the partial derivatives:
./U , y) = ⅛(-λ-2y + y 3) = 2xy and ∕.(x, y) = ^ ( x 2y + y 3) = x 2 + 3y2
Then,
V∕(x, y) = 2xyi + (x2 + 3y2)j
The following theorem shows how the directional derivative can be expressed
in terms of the gradient.
THEOREM 12.7 The gradient formula for the directional derivative

If / i s a differentiable function of x and y, then the directional derivative at the
point PQ(X Q, y0) in the direction of the unit vector u is
P>uf(x , y) = V ∕∙ u
Proof: Because V∕(x, y) = f χ (x, y)i + ∕( x , y)j and u = M1i + w2j, then
V∕o ■u = ∕ v(λ-0 , y0)w1 + f f x 0, y 0)u2 = Duf(x 0, y0) Dot product ■
EXAMPLE 4 Using gradient formula to compute a directional derivative

Find the directional derivative of ∕(x , y) = ln(x2 + y 3) at F 0(l, -3 ) in the
direction of v = 2i - 3j.
Solution A-(x >k) = .V22+VJ .Γ SO ./;(!, -3 ) = - ⅛
f y (χ >T) = p ‰ ' so ∕y( 1 =-3 ) = -2 ^
W0 ( l , -3 ) = - ⅜ i - ⅛ j
A unit vector in the direction of v is
=—
∣
∣∣= —
v∣ V2—
2—
+ (— √ T TP i - 3-wi)
-3—) 2 = —
Thus,
0 .0 . -3 ) - ¾ ■1 = (-⅛ )(⅛ ) + ( - g ) ( - ⅛ ) = 2 ⅛ P - 0.82
We know that a differentiable function of one variable ∕(x) has exactly one
derivative ∕ , (x), but a differentiable function of two variables F(x, y) has two
partial derivatives and a multitude of directional derivatives. Is there any single
mathematical entity for functions of several variables that is the analog of the
derivative of a function of a single variable? The answer is that the gradient is
perhaps the best analog, as may be seen by comparing the properties listed in the
following theorem with the properties of the derivative.
THEOREM 12.8 Basic properties of the gradient
Let /'and g be differentiable functions. Then
Constant rule Vc = 0 for any constant c

Linearity rule V(af + bg) = aVf + ⅛Vg for constants a and b
Product rule W g ) = ,∏ g + gV∕'
Quotient rule s. ≠ o
∖ g) g2
Power rule V(P) = n f (n- 1) v /
Proof: Linearity rule:

W + bg) = (α ∕ + bg)χ i + (af + ⅛ ) ι,j = (afχ + bgχ )i + (afy + ⅛,)j
= afxi + bgχ i + afy ] + ⅛ ,j = α ( ∕ γi + f y j) + b(gχ i + g yj)
= aVf+ M g
Power rule:
V∕'" = [ ∕ n].v i + [ f n}y j = n f n ~i f∖ + nf"~' f y j
= n f n ~l [ f j + / , j] = n ∕" - 1V ∕
The other rules are left for the problem set (Problems 58 and 59). ■
In applications, it is often useful to compute the greatest rate of increase (or

decrease) of a given function at a specified point. The direction in which this
occurs is called the direction of steepest ascent (or steepest descent), as shown in
Figure 12.23a.
For example, imagine a person skiing down a mountain. The function z = f( x , y)

gives the altitude of the skier on the slope, and we want to state a theorem that will
give the skier the compass direction of the path of steepest descent. We emphasize
the words “ compass direction” because the gradient gives direction in the xy-
plane and does not itself point up or down the mountain.
T H E O R E M 12.9 Optimal direction property of the gradient
Suppose/is differentiable and let V ∕o denote the gradient at PQ . Then if V ∕ o ≠ 0:

a. The largest value of the directional derivative of D uf is ∣∣V∕0∣∣ and occurs
when the unit vector u points in the direction of V ∕o .
b. The smallest value of D uf is -∣∣V ∕0∣∣ and occurs when u points in the
direction of - V∙7∕u.
Proof: If u is any unit vector, then

The optimal direction property of the gradient
D uf = Vf0 ∙ u = ∣
∣
V ∕ o ∣∣ (∣∣u∣∣ cos θ) = ∣∣V ∕ o ∣∣ cos θ Dot product (alternative form)
Figure 12.23 Steepest ascent or
steepest descent where θ is the angle between V ∕ o and u. But cos0 assumes its largest value 1 at
θ = 0; that is, when u points in the direction V ∕o . Thus, the largest possible value of
7V is
^ ∕= l∣ v ∕ 0∣∣(i) = ∣∣v∕0∣
Statement b may be established in a similar fashion by noting that cos θ assumes
its smallest value -1 when θ = π. This value occurs when u points toward - V ∕ 0,
and in this direction D J = ∣∣V∕0 ∣∣( -l) = ~∣∣V∕0 ∣∣. ■
What this says: The theorem states that at PQ , the function/increases most
rapidly in the direction of the gradient V ∕ and decreases most rapidly in the
opposite direction.
The gradient of / is a vector V ∕o
in the xy-plane. This vector
points in the direction of steep EXAM PLE 5 Optimal rate of increase and decrease
est ascent or descent at a given
In what direction is the function defined by ∕( x , y) = x e 2y~x increasing most
point.
rapidly at the point PQ(2, 1), and what is the optimal rate of increase? In what
direction is / decreasing most rapidly?
Solution We begin by finding the gradient o f/:
V f = f χ i + f y j = [e 2y- χ + x e 2y ~x (~ l)]i + [xe¾'-*(2)]j
= e 2>, ~ ∕( l - x)i + 2xj]
At (2, 1), V ∕ = e 2∞- 2 [(l - 2)i + 2(2)j] = - i + 4j. The most rapid rate of
increase is ∣∣V∕θ∣∣ = V ( - 1 ) 2 + (4)2 = VT7 and it occurs in the direction of
—i + 4j. The most rapid rate of decrease occurs in the direction of
-V∕ 0 = i - 4 j . —
The directional derivative and gradient concepts can easily be extended to

functions of three or more variables. For a function of three variables, f ( x , y, z),
the gradient V/is defined by
V∕o = ∕χ i + f y j + ∕ k
Surface defined by f(x , y) = xe 2y x and the directional derivative D u f o f∕( x , y, z) at P 0 (x 0 , y 0, z0) in the direction of
the unit vector u is given by D f = V∕o ∙ u

where, as before, V/() is the gradient V/evaluated at PQ. The basic properties of the
gradient (Theorem 12.8) are still valid, as are the optimization properties in
Theorem 12.9. Similar definitions and properties are valid for functions of more
than three variables.
EXAM PLE 6 Directional derivative of a function of three variables
Let f(x , y, z) = xy sinxz. Find V∕o at the point Pθ(l, -2 , π) and then compute
the directional derivative of f at Po in the direction of the vector
v = - 2 i + 3j - 5k.
Solution Begin with the partial derivatives:
f x = y sinxz + xy(z cosxz); /.(1, -2 , ττ) = - 2 sin π - 2ττ cos π = 2τr
f y = X sinxz; /(1 , -2 , ττ) = 1 sin π = 0
f z = xy(x cosxz); /(1 , -2 , τr) = (1)(-2)(l)cosτr = 2
Thus, the gradient of f at Po is
V∕n = 2ττi + 2k
To find DU∙7f w
, e need u, the unit vector in the direction of v:
v - 2 i + 3j - 5k 1 z ~.
“ M ^^ V ( - 2 ) 2 + (3)2 + ( - 5 ) 2 - V 38 ( “ 21 + 3j ^ 5k)
Finally,
D /(1, -2 , π) = V∕o ∙ u = ⅛ ( - 4 τ r - 10) -3 .6 6
■ NORMAL PROPERTY OF THE GRADIENT
Suppose A is a level surface of the function defined by ∕(x , y, z); that is,
∕(x , y, z) = K for some constant K. Then if P0(x0, y 0, z0) is a point on S, it can be
shown that the gradient V∕o at Po is a vector that is normal (that is, perpendicular)
to the tangent vector of every smooth curve on .8' that passes through Po (see Figure
12.24).
Figure 12.24 The normal property

of the gradient
THEOREM 12.10 The normal property of the gradient
Suppose the function / is differentiable at the point Po and that the gradient at Po
satisfies V∕o ≠ 0. Then V∕o is orthogonal to the level surface ∕(x , y, z) = K at PQ.
Proof: Let Cbe any smooth curve on the level surface f(x , y, z) = Ai that passes
through P0(x0, y 0, z0), and describe the curve C by the vector function
R(∕) = x(t)i + y(Oj + f° r a ll t ln some interval I. We shall show that the
gradient Vf0 is orthogonal to the tangent vector d'R/dt at PQ.
Because C lies on the level surface, any point P(x(f), y(∕), z(0) on C must
satisfy
∕[x (t), y(t), z(t)] = K
and by applying the chain rule, we obtain
⅛ ∕M 0 , y(z), z(0)] = f f χ , y, + f y (x, y, z ) ^ + f z(x, y, z ) ^
Suppose t = t0 at PQ. Then

^ t [J∖ x(t), y(t), z(t)]\[=to
= ∕χ(*(⅞), M )> z (r θ))⅜ + fy(x (t 0)>y( t 0^ + yô)’ z (t 0^Tt
= ^jr .dR
0 dt
Remember that = ^ i + ^ ∙j + ^rk. We also know that ∕(x(t), y(Z), z(Z)) = K
for all t in I (because the curve Clies on the level surface∕(x , y, z) = K). Thus, we
have
⅛{∕[x(Z), y(t), z(t)]} = ⅛ m = 0
∕∕R
and it follows that V∕o ■ = 0. We are given that V∕o ≠ 0, and dRJdt ≠ 0 because
the curve C is smooth. Thus, V∕o is orthogonal to dRJdt, as required. ■
I What this says: The gradient V∕o at each point P o on the level surface
F(x, y, z) = K is normal to the surface at PQ. That is, V∕o is orthogonal to all
tangent lines to curves on F(x, y, z) = k through PQ.
Here is an example in which/involves only two variables, so∕(x , y) = K is a

level curve in the plane instead of a level surface in space.
EXAMPLE 7 Finding a vector normal to a level curve
Sketch the level curve corresponding to C = 1 for the function
∕(x , y) = x 2 —y 2 and find a normal vector at the point P0(2, V3).
Solution The level curve for C = 1 is a hyperbola given by x 2 - y 2 = 1, as shown
in Figure 12.25 (along with the surface defined by ∕(x , y) = x 2 - y 2),
Plane z = 1
Figure 12.25 The surface f(x , y) = x2 - y2 and the level curve for which C = I
The gradient vector is perpendicular to the level curve, so we have

V ∕ = f χ i + / , j = 2xi - 2yj
At the point (2, λ∕3), V∕o = 4i - 2V 3j is the required normal. This normal
vector and a few others are shown in Figure 12.25. B M
EXAM PLE 8 Finding a vector that is normal to a level surface
Find a vector that is normal to the level surface x 2 + 2xy - y z + 3z 2 = 7 at the

point P 0 (l, 1, - 1).
Solution The gradient vector is perpendicular to the level surface:
V ∕= f χ ∖ + f y i + f : k = (2x + 2y)i + (2x - z)j + (6z - y)k
At the point (1, 1, - 1 ) , V ∕ = 4i + 3j - 7k is the required normal.
EXAM PLE 9 Heat flow application
The set of points (x, y) with 0 ≤ x ≤ 5 and 0 ≤ y ≤ 5 is a square in the first

quadrant of the xy plane. Suppose this square is heated in such a way that
T(x, y) = x 2 + y 2 is the tem perature at the point P(x, y). In what direction will
heat flow from the point P o (3, 4)?
Solution The flow of heat in the region is given by a vector function H(x, y),
whose value at each point (x, y) depends on x and y. From physics it is known
that H(x, y) will be perpendicular to the isothermal curves T(x, y) = C for C
constant. The gradient V7^, and all its multiples point in such a direction.
Therefore, we can express the heat flow as H = -k% T , where k is a positive
constant (called the thermal conductivity) and the negative sign is introduced
to account for the fact that heat flows “ downhill” (that is, toward a decreasing
temperature).
Because T(3, 4) = 25, the point F,o (3, 4) lies on the isotherm T(x, y) = 25,
which is part of the circle x 2 + y 2 = 25, as shown in Figure 12.26. We know
that the heat flow H o at Po will satisfy H o = -∕cV T 0 , where VT0 is the gradient
at Po . Because VT = 2xi + 2yj, we see that VT0 = 6i + 8j. Thus, the heat flow
at P o satisfies
H o = - W T 0 = - k ( 6 i + 8j)
Figure 12.26 The heat flow Po is
in the direction of Because the therm al conductivity k is positive, we can say that heat flows from
-V 7 0 = -6 i - 8j Po in the direction of the unit vector u given by
= - ( 6 i + 8j) = 3 _ 4
V (-6 )2 + (-8 )2 5 5j —
Normal line
■ TANGENT PLANES AND NORMAL LINES
Tangent planes and normal lines to a surface are the natural extensions to IR3 of the
tangent and normal lines we examined in R 2 . Suppose A is a surface and N is a
vector normal to S’at the point Po . We would intuitively expect the normal line and
the tangent plane to A at Po to be, respectively, the line through P o with the
Figure 12.27 Tangent plane and direction of N and the plane perpendicular to the line at P o (see Figure 12.27.)
normal line These observations lead us to the following definition.
Normal Line Tangent Plane Suppose the surface S’ has a nonzero normal vector N at the point PQ. Then the
line through P o parallel to N is called the normal line to S' at P o , and the plane
through P o with normal vector N is the tangent plane to 5 at Po .
By analogy with the single-variable case, we would expect a surface S' to have a
tangent plane precisely where it can be represented by a differentiable function. In
particular, if S has an equation of the form F(x, y, z) = C, where C is a constant
and F is a function differentiable at Po , the normal property of a gradient tells us
that the gradient VF0 at Po is normal to S' (if VF0 ≠ 0) and that S' must therefore
have a tangent plane at Pθ.
EXAM PLE 10 Finding the tangent plane and normal line to a given surface
Find equations for the tangent plane and the normal line at the point
P 0(l, —1, 2) on the surface S' given by x 2y + y 2z + z 2x = 5.
Solution We need to rewrite this problem so that the normal property of the
gradient theorem applies. Let F(x, y, z) = x 2y + y 2z + z 2x, and consider S to
be the level surface F(x, y, z) = 5. The gradient VF is normal to S'.
VF(x, y, z) = (2xy + z 2 )i + (x 2 + 2yz)j + (y 2 + 2 xz)k
Evaluate VF at F 0(l, —1, 2) to obtain the normal to the surface at PQ:
N = VF0 = VF(1, - 1 , 2) = 2i - 3j + 5k
Hence, the required tangent plane is
2(x - 1 ) - 3(y + 1 ) + 5(z - 2) = 0 or 2x - 3y + 5z = 15
The normal line to the surface at P o is
x = 1 + 2t, y = -l-3 t, z = 2 + 5t
By generalizing the procedure illustrated in the preceding example, we are led

to the following formulas for the tangent plane and normal line.
Formulas for the Tangent Plane and Normal Line to a Surface

Suppose S' is a surface with the equation F(x, y, z) = C and let P0 (x 0 , y 0 , z 0 ) be a
point on S' where F is differentiable. Then the equation of the tangent plane to S'
at P o is
7∖ (⅞ y 0 ’ z o)(χ ~ ⅞ ) + F J X O -- y 0 ’ z 0 )(y ~ ⅞) + f -(x 0 , y 0 , ⅞)(^ - ⅞) = θ
and the equation of the normal line to S at Pθ is
x = x 0+ X F z
X( 0’ >0’ ⅞)
Normal line / y = To + p y (χ ^ y 0 >z 0 )t
Figure 12.28 The tangent plane Z = ⅞ + F Z^ y 0 ’ z 0)t
and normal line to a surface provided that F χ , F , and F z are not all zero.
EXAM PLE 11 Equations of the tangent plane and the normal line
Find the equations for the tangent plane and the normal line to the cone
z 2 = x 2 + y 2 at the point where x = 3, y = 4, and z > 0.
Solution If P0 (x 0 , y 0 , ⅞) *s ⅛e point of tangency and x 0 = 3, y 0 = 4, and ZQ > 0,

then
zo= ^⅞ 2 ÷ To 2 = V9 + 16 = 5
If we consider F(x, y, z) = x 2 + y 2 - z 2 , then the cone can be regarded as the
level surface F(x, y, z) = 0. The partial derivatives of F are
Fχ = 2x Fy = 2y F_ = - 2 z
so at P o (3, 4, 5),
F χ (3, 4, 5) = 6, F,(3, 4, 5) = 8, F z (3, 4, 5) = - 1 0
Tangent plane and normal line to Thus the tangent plane is
the cone z 2 = x 2 + y 2 at (3, 4, 5)
6(x - 3 ) + 8(y - 4 ) - 10(z - 5) = 0
or 3x + 4y —5z = 0, and the normal line is
x = 3 + 6t, y = 4 + 8t, z = 5 -1 0 t
PROBLEM SET 1 2 .6
θ Find the gradient o f the functions given in Problems 1-10 23. ze x+2y = 3 at (2, -1 , 3) 24. ze χ 2 - y 2 = 3 at (1, 1,3)
1. f( x , y) = x 2 - 2xy 2. f(x , y) = 3x + 4y 2 Find the direction from Po in which the given function f
3. ∕( 1
ln(x 2 y 2) increases most rapidly and compute the magnitude o f the
j vx , y)’ = x + -y 4. f(x , y) = +
greatest rate o f increase in Problems 25-34.
5. flu , v) = ue 3~'' 6. f(u , v) = e u+v 25. fi x , y) = 3x + 2y — 1
7- f ix , y) = sin (x + 2y) 8. f ix , y, z) = xyz 2 a. P 0( l , - l ) b. P 0( l , l )
9. g(x, y, z) = xe y+3z 10. f ix , y, z) = -- 26. fi x , y) = 1 - x 2 - y 2
a. P 0(l,2 ) b. Pfl(0,0)
Compιιte the directional derivative o f the functions given in 27. f i x , y) = x + y3 3
Problems 11-16 at the point Po in the direction o f the given a. P 0(3, -3 ) b. P o(-3 , 3)
vector v.
28. fi x , y) = ax + by + c; P 0 (α, b)
Function Point Pθ Vector v 29. f(x , y, z) = ax 2 + by2 + cz2 ; P 0(α, b, c)
Assume a 4 + b 4 + c4 ≠ 0.
11. fix , y) = x 2 + xy (1 ,-2 ) •+ j 30. fi x , y) = αx 3 + by 3; P0(a, b)
12. fix , y) = e~x y~ l (2 ,-1 ) -i + j 31. f i x , y) = ln √ x 2 + y 2 ; P 0(l, 2)
13. fi x , y) = ln(x 2 + 3y) (1, i) i+j 32- f i x , y) = sinxy; P 0θ⅞>
14. fi x , y) = ln(3x + y 2) (0, 1) i- j 33. fi x , y, z) = (x + y)2 + (y + z)2 + (z + x) 2; P 0(2, - 1 ,2 )
15. fi x , y) == sec(xy - y 3) (2, 0) - i - 3j 34. fi x , y, z) = z ln(£); P 0(l, e, -1 )
16. fi x , y) = sinxy (Vπ, Vπ) 3τri — τrj θ In Problems 35-38, find a unit vector that is normal to the
given curve or line at the point P0(x 0, y 0) on the line or curves.
Find a unit vector (if possible) that is normal to each surface Assume that a, b, and c are constants.
given in Problems 17-24 at the prescribed point. Also find the 35. The line ax + by = c 36. The circle x 2 + y 2 = a 2
2 v 2 γ 2 y2
equation o f the tangent plane at that point. 37. The ellipseγ ¾2 + 75 = 1 38. The hyperbola ⅛ 2 —752 = 1
17. x 2 + y 2 + z 2 —3 at (1, —1, 1) a b~ a b
39. Find the directional derivative of ∕( x , y) = x 2 + y 2 at the
18. x 4 + y 4 + z 4 = 3 at (1, - 1 , - 1 )
point P o( 1, 1) in the direction of the unit vector u shown
19. cosz = sin(x + y) at (f>
in Figure 12.29.
20. sin(x + y) + tan(y + z) = 1 at (f> 2p - f )
21. Inf — — ) = 0 a t(2 , 5, 3)
∖y - z ∕
22. Inf A ■-j -) = x - z at (1, 0, 1)
∖y + z) v ' Figure 12.29 Problem 39
40. Find the directional derivative of f ( x , y) = x 2 + x y + y 2 52. Let 7'(x, y) = 1 - x 2 - 2y 2 be the temperature at each
at F,0( l, - 1) in the direction toward the origin. point P(x, y) in the plane. A heat-loving bug is placed in
41. Find the directional derivative of f ( x , y) = e'~y 2 at the plane at the point P o (- 1, 1). Find the path that the bug
P 0( l, -1 ) in the direction toward Q(2, 3). should take in order to stay as warm as possible.
42. Let f ( x , y, z) = 2 x 2 - y 2 + 3z2 - 8x - 4y + 201, and let Hint: Assume that at each point on the bug’s path, the
P o be the point (2, —1> ∣) tangent line will point in the direction for which T
a. Find V ∕ o . increases most rapidly.
b. Find cos θ, where θ is the angle between V ∕ o and the 53. Repeat Problem 52 for the temperature function
vector from the origin to Po . T(x, y) = 1 - ax 2 - by 2 and starting at the point (x0 , y 0).
43. Let f ( x , y, z) = xyz, and let u be a unit vector perpendicu Assume a > 0 and b > 0 and (x0 , y 0) Ψ (0, 0).
lar to both v = i —2j + 3k and w = 2i + j —k. Find the
directional derivative o f/ a t P0( l, - 1 , 2) in the direction
of u.
44. Let f ( x , y, z) = y e x + z + zey ~x . At the point P(2, 2, -2 ),
find the unit vector pointing in the direction o f most
rapid increase of f
45. Suppose a box in space given by 0 ≤ x ≤ 2, 0 ≤ y ≤ 2,
0 ≤ z ≤ 2 is temperature-controlled so that the temper 55. S P Y PROBLEM The spy enters the gas filled bunker of the
ature at a point P(x, y, z) in the box is international terrorists (Problem 30, Section 11.5) and
T(x, y, z) = xy + yz + xz. finds himself in a trap. The door slams shut behind him
and he begins to feel very warm. Too late he realizes the
A heat-seeking missile is located at P 0( l, 1, 1)∙ ⅛ what
bunker is a large roasting oven. Fortunately, he is wearing
direction will the missile move in order to increase the
his heat-detector ring, which indicates the direction of
temperature as quickly as possible? What is the maximum
greatest temperature decrease. Suppose the bunker is
rate of change of the temperature at the point //?
coordinatized so that the temperature at each point (x, y)
46. A metal plate that is covering the rectangular region
on the floor o f the bunker is given by
0 ≤ x ≤ 6. 0 ≤ y ≤ 5 is charged electrically in such a
way that the potential at each point (x, y) is inversely T(x, y) = 3(x - 6)2 + 1.5(y - 1)2 + 41
proportional to the square o f its distance from the origin.
If an object is at the point (3, 4), in which direction should degrees Fahrenheit, where x and y are in feet. The spy
it move in order to increase the potential most rapidly? begins at the point (1,5) and stumbles across the room at
47. A hiker is walking on a mountain path when it begins to the rate of 4 ft/min, always moving in the direction of
rain. If the surface o f the mountain is given by maximum temperature decrease. But he can last no more
z = 1 - 3x 2 - jy 2 (where x , y, and z are in miles) and the than 2 minutes under these conditions! Assuming that
rain begins when the hiker is at the point P o ( ∣ , - ∣ , ⅛), in there is an escape hole at the point where the temperature
is minimal, does he make it or is the spy toast at last?
what direction should she turn in order to descend the
Hint: Hallucinating in the heat, the spy remembers the
mountainside most rapidly?
odd fact that f( x , y) = x 2 + y 2 + c is minimized when
48. Let f have continuous partial derivatives, and assume the
x = 0 and y = 0.
maximal directional derivative o f f at (0, 0) is equal to
100. If the maximal value o f the directional derivative is 56. Recall from Chapter 9 (Problem 69, Section 9.6) that an
attained in the direction of the vector (0, 0) toward ellipse is the set of all points P(x, y) such that the sum of
(3, -4 ), find the gradient V ∕ at (0, 0). the distances from P to two fixed points (the foci) is
constant. Let P(x, y) be a point on the ellipse, and let rl
49. Suppose a certain function ∕( x , y) has the directional
and r2 denote the respective distances from P to the two
derivative 8 at the point P o( - 1, 2) in the direction o f the
foci, F and F 2, as shown in Figure 12.30.
unit vector
3∙ 4.
M = 5*^5 J
and the directional derivative 1 in the direction from P o
toward O(l 1, 7). What is the directional derivative o f/at
Pn in the direction of v = 3i - 5j?
50. Suppose f ( x , y) has directional derivatives T>, / = 2 at
P 0 ( l, 5) in the direction of v 1 = 2i + 3j and D v f = - 5 in
the direction of v2 = 3i — 5j. What is the directional
derivative o f f at P o in the direction of v3 = 2i —3j?
51. Let f have continuous partial derivatives and suppose the Figure 12.30 Unit tangent to an ellipse
maximal directional derivative o f/a t P 0( l, 2) has magni
tude 50 and is attained in the direction from P o toward a. Show that T ∙ V(rl + r2) = 0, where T is the unit tan
Q(2>, -4 ). Use this information to find V ∕( l , 2). gent to the ellipse at P.
Sec. 12.7 Extrema of Functions of Two Variables 793
b. Use part a to show that the tangent line to the ellipse at 59. Verify the product rule
P makes equal angles with the lines joining P to the foci
(that is, 01 = θ2 in Figure 12.30.) v (∕g ) = ∕V g + g V ∕
θ 57. A particle P 1 with mass m l is located at the origin, and a for functions of three variables.
particle P 2 with mass 1 unit is located at the point (x, y, z). 60. Find a general formula for the directional derivative D u f
According to Newton’s law o f gravity, the force P 1 exerts o f the function f ( x , y) at the point P(x 0 , y 0) in the
on ∕ j , is given by direction o f the unit vector u = (cos 0)i + (sin 0)j. Apply
- g m 1(xi + yj + zk) your formula to obtain the directional derivative of
∕( x , y) = x y 2ex ~2y at P o( - 1, 3) in the direction of the unit
vector u = (cos^)i + (sin^)j.
The number r is the distance between P 1 and P 2 and g is 61. Suppose that u and v are unit vectors and that f has
the gravitational constant. continuous partial derivatives. Show that
a. Starting from the fact that r 2 = x 2 + y 2 + z 2, show
that D . f = ∣∣ * ∣∣ (£> / + D f )
u+v7 IIu + v ∣
∣v u ∙y γj ’
a Λ∖ -χ ι A (T i= Ξ 2 ,
∂x∖r∕ r3 ∂y∖r] ri ∂z∖r j r2 62. If a is a constant vector and R = xi + yj + zk, show that
V(a R) = a.
b. The function V = —gm J r is called the potential energy 63. Let R = xi + yj + zk, and let
function for the system. Show that F = - V K
58. Verify each of the following properties for functions of r = ∣∣R ∣
∣= V x 2 + y 2 + z 2
two variables. a. Show that W is a unit vector in the direction o f R.
a. V (c∕) = c(V∕') for constant c b. Show that V(r") = nr"^2R, for any positive integer n.
b∙ V ( ∕ + g) = V ∕+ Vg
c∙ V ( ∕∕g ) = g v f ~ ∕v g , ≠
g 0
o
12.7 EXTREMA OF FUNCTIONS OF TWO VARIABLES
IN THIS SECTION Relative extrema, second partials test, extreme value

theorem
In the real world we encounter a variety of problems involving optimization. In
Chapter 3 we considered extrema of functions of a single variable, and in this
section we extend our work to functions of two variables.
There are many practical situations in which it is necessary or useful to know the
largest and smallest values of a function of two variables. For example, if T(x, y) is
the temperature at a point (x, y) in a plate, where are the hottest and coldest points
in the plate and what are these extreme temperatures? This is an example of a
problem in which we are asked to find the extreme (that is, the largest or smallest)
values of a function of two variables. We begin our study of extrema with some
terminology.
Absolute Extrema The function f( x , y) is said to have an absolute maximum at (x0, y 0) if

f( x 0 , y 0) ≥ f( x , y) for all (x, y) in the domain D of f Similarly, f has an I
absolute minimum at (x0, y 0) i f ∕( x 0 , y 0) ≤ f( x , y) for all (x, y) in D. Collectively, ∣
absolute maxima and minima are called absolute extrema.
In Chapter 3, we located absolute extrema of a function of one variable by first

finding relative extrema-, those values o f∕( x ) that are larger or smaller than those
at all “ nearby” points. The relative extrema of a function of two variables may be
defined as follows.
Relative Extrema Let / b e a function defined at (x0 , y0). Then

∕( x 0, y 0) is a relative maximum if∕( x , y) ≤ f(x 0 , y 0) for all (x, y) in an open
disk containing (x0 , y 0).
∕( x 0 , y0) is a relative minimum if∕(x , y) ≥ f(x 0, y0) for all (x, y) in an open
disk containing (xθ, y0).
Collectively, relative maxima and minima are called relative extrema.
We also found in Chapter 3 that an absolute extremum of a function f can

occur either at an endpoint of an interval in the domain o f / or at an interior
Extrema occur on the boundary or critical point where it is a relative extremum—that is, at a point where either∕'( x )
where the tangent plane is hori does not exist o r∕'(x ) = 0. In Section 12.6, we observed that the gradient V/is the
zontal. two-variable analog to the derivative in the single-variable case. Thus, it is
reasonable to expect extrema of a function of two variables to occur in one of the
following situations:
1. either on the boundary of the domain of / or

2. at points in the interior where the gradient V/does not exist, or
3. where Nf = 0; that is, where/. = / = 0.
We begin by considering relative extrema, and then we turn to the more general
question of finding absolute extrema.
■ RELATIVE EXTREMA
In Chapter 3, we observed that relative extrema of the function / correspond to

“peaks and valleys” on the graph of f and the same observation can be made
about relative extrema in the two-variable case, as seen in Figure 12.31.
For a function / of one variable, we found that the relative extrema occur where
∕'( x ) = 0 or f'(x ) does not exist. The following theorem shows that the relative
extrema of a function of two variables can be located in a similar fashion.
THEOREM 12.11 Partial derivative criteria for relative extrema
If /h a s a relative extremum (maximum or minimum) at P0(x0, y0) and partial

derivatives f χ and / both exist at Pθ(x0, y0), then
f S ⅜ τ 0) = f y (χ0, y 0) = o
Figure 12.31 Relative extrema
occur at peaks and valleys.
Proof: Let F(x) = ∕(x , y0). Then F(x) must have a relative extremum at x = x0,
so F'(x 0) = 0, which means that ∕ v(x0 , y0) = 0. Similarly, G(y) = ∕( x 0 , y) has a
relative extremum at y = y0, so C'(y 0) = 0 a n d ∕( x 0 , y0) = 0. Thus, we must have
both f χ (x0 , y0) = 0 and ∕.(x 0, y 0) = 0, as claimed. ■
0 If a point is an extreme point and if the first partial derivatives exist,

then there is a horizontal tangent plane at that point. This does NOT say that
if there is a horizontal tangent plane at a point that point is an extremum.
However, such a point IS a possible location for a relative extremum. 0
In single variable calculus, we referred to a number x0, where ∕'( x 0) does not
exist or∕'( x 0) = 0, as a critical value. This terminology is extended to functions of
two variables, as follows.
Critical Points A critical point of a function f defined on an open set S' is a point (xθ, j 0) in S
where either one of the following is true:
a∙ Λ (⅞ ⅞) = ∕∕⅞ > Lo) = 0 ∙
b. f x (x0 , v0) o r Z(⅞> Vo) does not exist (one or both).
EXAMPLE 1 Distinguishing critical points
Discuss the nature of the critical point (0, 0) for the quadric surfaces
a. z = x 2 + y 2 b. z + x 2 + y 2 = 1 c. z = y 2 - x 2
Solution The graphs of these quadric surfaces are shown in Figure 12.32.
z = -2
z = -l
z = -0.5
z = -0.25
z= 0
z = 0.25
z = 0.5
z= 1
z= 2
Saddle
point
at z = 0
a. z = x 2 + y 2 (paraboloid) and level b. z = 1 - Λ 2 - y 2

(paraboloid) and level c. z = y 2 - x
2 (hyperbolic paraboloid) and
curves; relative minimum at (0, 0, 0) curves; relative maximum at (0, 0, 1) level curves; saddle point at (0, 0, 0)
Figure 12.32 Classification of critical points
Let f(x , y) = x 2 + y 2 , g(x, y) = 1 —x 2 - y 2, and h(x, y) = y 2 —x 2. We find the

critical points:
a. f χ (x, y) = 2x, f y (x, y) = 2y, the critical point is (0, 0). The function f has a
relative minimum at (0, 0) because x 2 and y 2 are both nonnegative yielding
χ 2 + y 2 ≥ 0 for all nonzero x and y.
b. g χ(x, y) = -2 x , gv(x, y) = -2y; critical point (0, 0). The function /h a s a
relative maximum at (0, 0) because z = 1 - x 2 - y 2 and x 2 and y 2 are both
nonnegative, so the largest value of z occurs at (0, 0).
c. h∖ x, y) = -2x, ∕ Z∕ Λ^, y) = 2y; critical point (0, 0). The function f has
neither a relative maximum nor a relative minimum at (0, 0) because when
z = 0, f is positive on the y-axis (where x = 0) and negative on the x-axis
(where y = 0).
If (x 0 , >,0) is a critical point for f( x , y) that is neither a relative maximum nor a

relative minimum, it may be a saddle point. Such a point is a m axim um in one
direction and a m inim um is another direction and so has a saddle shape, as shown
in Figure 12.32c.
■ SECOND PARTIALS TEST
The cases examined in Example 1 point to the need for some sort of a test to
determine the nature of a critical point. In Chapter 3, we developed the second
derivative test as a means for classifying a critical point o f /a s a relative maximum
or m inimum. According to this test, if∕'( c ) = θ, then at x = c, a relative maximum
occurs if ∕" ( c ) < 0 and a relative m inimum occurs iff''( c ) > 0. If∕" ( c ) = 0, the
test is inconclusive. The analogous result for the two-variable case may be stated as
follows.
Second Partials Test Let f ( x , y) have a critical point at P 0(x 0 , y 0) and assume that f has continuous
partial derivatives in a disk centered at (x0 , y 0). Let
d = ∕√ ⅞ > To) ∕√ ⅞ To ) - C∕√⅞> T0 )]2
Summary:
Then,
Critical D f Type
Point A relative maximum occurs at Po if
(⅞ τ l ) Pos. Neg. Rei. max P> > 0 and f χ χ (x 0 , y 0) < 0

(⅞> τ 2) Pos. Pos. Rei. min
(⅞> τ 3) Neg. Saddle pt.
( χ i , τ 4) Zero Inconclusive
A relative minimum occurs at Po if z = x 2 + y 2 (See Ex la)
P> > 0 and f χ χ (x 0 , y 0 ) > 0
A saddle point occurs at Po if D < 0. z = y 2 - χ 2 (See Ex lc)
If D = 0, then the test is inconclusive.
0 The same result holds if f x x {x0 , To) *s replaced by f (xθ, y 0 ). 0

When D(x0, y0) = 0, the critical point (x0, y0) is said to be degenerate; otherwise, it
is nondegenerate. Note that D can be expressed in determinant form as
∩ _ fχ x fχ y
f
j
f
j
xy yy
where all the partials are evaluated at (x0 , y 0). Some students find this the easiest
form of D to remember.
EXAMPLE 2 Second partials test with a relative maximum

Find all critical points on the graph of'f(x, y) = 1 - x 2 - y 2 and use the second
partials test to classify each critical point as a relative extremum or a saddle
point.
Solution f χ (x, y) = - 2 x ,f ( x , y) = -2 y, so the only critical point is at (0, 0).
Aχ(∙*x > = ^^2 ' Λ A T) = ° ’ a n d Λ A J7) = -2 , so
1-2
D= =4
0
Because D = 4 > 0 for all (x, y), and because f xx (0, 0) = - 2 < 0, the second
partials test tells us that a relative maximum occurs at (0, 0).
Graph of f(x ,y ) = ∖ - x 2 - y 2
EXAMPLE 3 Second partials test with a relative minimum and a saddle

point
Find all critical points on the graph of∕(x , y) = 8x3 - 24xy + y 3 and use the
second partials test to classify each point as a relative extremum or a saddle
point.
Solution f χ (x, y) = 24x2 - 24y, f y (x, y) = -2 4 x + 3y2
To find the critical points, solve
124x2 - 24y = 0
l-2 4 x + 3y2 = 0
From the first equation y = x 2; substitute this into the second equation to find
-2 4 x + 3(x2)2 = 0
x 4 - 8x = 0
x(x 3 - 8) = 0
x(x - 2)(x2 + 2x + 4) = 0
x = 0, 2 The solutions of x 2 + 2x + 4 = 0 are not real.
Graph of
∕(x , y) = 8x3 - 24xy + y 3
Summary: If x = 0, then y = 0, and if x = 2, then y = 4, so the critical points are (0, 0),
(2, 4). To obtain D, we first find fχχ (x, y) = 4 8 x ,∕ υ ,(x, y) = -24, and f yy (x, y)
Critical D f χχ Type
Point
= 6y and then compute
48x -2 4
= 288xy- - 576
(0, 0) Neg. Saddle -2 4 6y
(2,4) Pos. Pos. Rei. min.
At (0, 0), D = -576 < 0, so there is a saddle point at (0, 0). At (2, 4),
D = 288(2)(4) - 576 = 1,728 > 0 andfχχ (2,4) = 96 > 0, so there is a relative
minimum at (2, 4). ∙≡ ≡
EXAM PLE 4 Extrema when the second partial test fails
Find all relative extrema and saddle points on the graph of

f(x , j) = x⅞ 4
Solution ∕ v(x, y) = 2xp4 ,∕'(χ , y) = 4x2j 3; we can see the only critical points occur
whenever x = 0 or y = 0; that is, every point on the x- or p-axis is a critical
point. Because f xx (x, τ) = 2y4, ∕ g ,(x, p) = 8∙TP3, f yv (χ , J7) = 12x2y2, we have
8xy3
= 24x2y6 - 64x2y 6 = - 40x2y6
12x 2y 2
For any critical point (x0 , 0) or (0, j 0), D = 0; so the second partials test fails.
But∕(x , y) = 0 for every critical point (because either x = 0 or y = 0, or both),
and because ∕( x , y) = x 2y 4 > 0 when x ≠ 0 and y ≠ 0, it follows that each
critical point must be a relative minimum.
The second partials test provides explicit information about the case where f
is nondegenerate at F,0(xθ, y 0), but if f is degenerate, practically anything can
Graph of f(x , y) = x 2y4 happen, as shown in Problem 47. In that problem you are asked to show that (0, 0)
is a degenerate critical point for each of the following functions:
∕(x , >j = x 4 - y 4 ; g(x, y) = x 2y 2', h(x, y) = x 3 + y 3
The following conclusions can be made:
∕( x , y) has a saddle point at (0, 0).
g(x, y) has a relative minimum at (0, 0).
∕ι(x, y) has neither kind of extremum nor a saddle point at (0, 0).
■ EXTREME VALUE THEOREM
In Chapter 3, we noted that a function of one variable ∕(x ) assumes both a

maximum and a minimum on any closed, bounded interval where it is continu
ous, and that all absolute extrema must occur either at one of the endpoints of the
interval or at an interior critical value. The following is the analogous result for
functions of two variables.
THEOREM 12.12 Extreme value theorem for a function of two variables
A function of two variables ∕( x , y) assumes an absolute extremum on any closed,

bounded set S in the plane where it is continuous. Moreover, all absolute extrema
must occur either on the boundary of S or at a critical point in the interior of 5.
Proof: The proof of this theorem is beyond the scope of this text and is usually
given in an advanced calculus course. ■
■ What this says: Visualize a surface in space z = f(x, y) with f continuous

that projects onto a closed, bounded region D in the xy-plane. Then the surface
must have a highest point (peak) and a lowest point (valley) over D. These
extreme features must occur either on the boundary of D or at a high point
(low point) in its interior.
The extreme value theorem for a function of two variables leads to the
procedure summarized in the following box.
Procedure for Determining Given a function f that is continuous on a closed, bounded set 5:
Absolute Extrema Step 1: Find all critical points o f/in 5.
Step 2: Find all points on the boundary of 5 where absolute extrema can occur
(boundary critical points, endpoints, etc.)
Step 3: Compute the value of∕(xθ, y0) for each of the points (x0 , y0) found in
steps 1 and 2.
Evaluation: The absolute maximum o f / on S’ is the largest of the values
computed in step 3, and the absolute minimum is the smallest of the computed
values.
EXAM PLE 5 Finding absolute extrema

2 2
Find the absolute extrema of the function ∕( x , y) = ex ~y over the disk
x 2 + y 2 ≤ 1.
Solution
Step 1: f χ (x, y) = 2xeχ2 ~y2 and ffx , y) = -2 ye χ2 ~y2 . These partial derivatives
are defined for all (x, y). Because ∕ v(x, y) = ∕,(x, y) = 0 only when x = 0 and
y = 0, it follows that (0, 0) is the only critical point of /
Step 2: Examine the values o f/o n the boundary curve x 2 + y 2 = 1. Because
y 2 = 1 - x 2 on the boundary, we find that
∕(x , y) = eχ2 ~Wχ2χ>= e2*2- 1
on the boundary x 2 + y 2 = 1. Let F(x) = e2χ2-1 for —1 ≤ x ≤ 1. We must
find the points where the extreme values of F(x) occur. Because this is a
function of a single variable, we find
F'(x) = 4xe2χ2~l
for -1 ≤ x ≤ 1, and F'(x) = 0 when
4xe2 χ 2 - 1 = 0
x = 0 or e2χ2~1 = 0
Because the exponential is never 0, we see that F'(x) = 0 only when x = 0.
At x = 0, y 2 = 1 - 0 2 which implies y = ± 1;
this gives the points (0, 1) and (0, -1). The endpoints of F are x = ±1, which
implies y 2 = 1 - 12 = 0; this gives the points (1,0) and ( - 1, 0).
Step 3: Compute the value o f/ fo r the points found in steps 1 and 2:
Points to check Compute f ( x 0 , y 0) = ex « j, θ

(0, 0) /(0, 0) = e 0 = 1
(0,1) /(0, 1) = e~ l ; minimum
(0 ,-1 ) /(0, -1 ) = e~ l ; minimum
(1,0) /(1, 0) = e; maximum
( - 1 ,0 ) / ( - 1 , 0) = e; maximum
Evaluation: As indicated in the preceding table, the absolute maximum

value of / on the given disk is e at (1, 0) and ( -1 , 0), and the absolute
minimum value is e~1 at (0, 1) and (0, -1 ). ■■■
EXAM PLE 6 Minimum distance from a point to a plane
Find the shortest distance from the point (0, 3, 4) to the plane x + 2y + z = 5.
Solution The distance from a point (x, y , z) to (0, 3, 4) is
d = V (x - 0)2 + (y - 3)2 + (z - 4)2
However, because (x, y, z) is on the plane x + 2y + z = 5, we know
z = 5 - x -2 y
so that d = ∖-x 2 + (y - 3)2 + (5 - x - 2y - 4)2. Instead of minimizing d, we
can minimize the simpler expression
<√2 = f{ x , y) = x 2 + (y - 3)2 + (1 - x - 2y)2
To find the critical values, we solve
∕ γ = 2x - 2( 1 - x - 2y) = 4x + 4y - 2 = 0
f y = 2(y - 3) - 4(1 - x - 2y) = 4x + 10y - 10 = 0
The only critical point is (-∣> ^). Also, ∕ γγ = 4, ∕ τ = 10, ∕ γι, = 4, so D > 0,
which means there is a relative minimum at (-j> ∣). Intuitively, we see that
this relative minimum must also be an absolute minimum because there must
be exactly one point on the plane that is closest to the given point. We now
calculate that distance:
Check: You might want to check your work by using the formula for the
distance from a point to a plane in [R3 (Theorem 10.10):
∣Λx 0 + gy 0 + Cz 0 + D _ Q + 2(3) + 4 - 5 _ 5
~ I V'Λ 2 + B- + C 2 √ 1 2 + 2 2 + 12 ^^ V 6
The following example uses calculus to justify a formula used in statistics and
in many applications in the social and physical sciences.
EXAM PLE 7 Least squares approximation of data
Suppose data consisting of n points P v . . . , P n are known, and we wish to

find a function y = ∕( x ) that fits the data reasonably well. In particular,
find a line y = mx + b that “best fits” the data in the sense that the sum
of the squares of the vertical distances from each data point to the line is
minimized.
Solution We wish to find values of m and b that minimize the sum of the squares
of the differences between the y-values on the line y = mx + b. This line we
seek is called the regression line. Suppose that the point Pk has components
(xk , y k ). Now, at this point, the value on the regression line is y = mx k + b and
the value of the data point is y k . The “error” caused by using the point on the
regression line rather than the actual data point is
Λ - (m x k + 6 )
and we wish to minimize the function that represents the sum of the squares of
all these differences: n
F(m, b) = ∑ ∣>fc - (mxk + ∕>)]2
⅛=ι κ κ
The situation is illustrated in Figure 12.33.

Next, we find the critical points of F(m, b) by computing F and Fh and
solving Fm = Fh = 0:
n
Fm (m, Z>) = ∑ 2[y k - (mx k + 6)](-χ jt)
= 2 w ∑ x k + 2b ∑ χ k - 2 ∑ χ
ky k
n k=∖ k=∖ k=∖
b
Fb(m, ) = ∑ 2[yk - (mxk + £>)](—1)
=lm ∑ χ
k +2b ∑ 1 - 2 ∑ yk
Figure 12.33 Least squares k=1 k=1 k=1
approximation of data
=2m ∑ x k+ 2 b n -2 ∑ y
k
k=I k=1
Set each of these partial derivatives equal to 0 to find (after a great deal of
algebra) the critical values:
n
rι ∑ χ
/ n ∖/ n
kyk ~ ∑ χ k ](∑ y k
m= k≈∖ ×k=∖ ' ∖k =∖
and
∏ / n ∖/ n
χ 2 ∑ y - ( Σ ⅞ )
k k=1______
___ k ×fc= 1
X
' ×⅛= 1________ '
(Σ Λ
b=
We leave it to you to verify that these values of m and b yield a mini

mum.
Most applications of the least squares formula stated in Example 7 involve using a
calculator or computer software. The following computational window provides
an example.
Many calculators have special keys to carry out the calculations required by
the least squares approximation theorem. Look at your owner’s manual for
specifics, but most calculators allow you to input data with keys labeled ISTAT∣
and (DATA). After the data are input, the m and b values are given by pressing
the (LinReg) choice. For example, 10 people are given a standard IQ test. Their
scores were then compared with their high school grades:
IQ 117 105 111 96 135 81 103 99 107 109
GPA 3.1 2.8 2.5 2.8 3.4 1.9 2.1 3.2 2.9 2.3
A calculator output shows m - .0224144711 and b - .3173417224

l f, ,
■■∙*,<AΛ⅛⅜-.t*∙.''' - *.-* % , Λ∙ ∙ΛΓ W■I” ” .’ J - - - ” ■’ W⅜¾WA".
PROBLEM SET 1 2 .7
θ 1. ■ What Does This Say? Describe what is meant by a
critical value.
2. ■ What Does This Say? Describe a procedure for deter
mining absolute extrema.
3. ■ What Does This Say? Describe what is meant by a
regression line, as well as describing the procedure for
finding such a line.
Find the critical points in Problems 4-19, and classify each as

a relative maximum, a relative minimum, or a saddle point.
4. 5. f i x , y) = { x - 2)2 + (y - 3)4
f i x , y) = 2x 2 - 4xy + y 3 + 2
9. f i x , y) = x 2 + x y + y 2 10. f i x , y) = x y - x + y
3
11. f i x , y) = - χ + 9x - 4y2 12. f ( x , y) = e~lχ2 + y 2 ,~
13. F(x, y) = (x 2 + 2y 2)e , ~χ 2 ~y2 14. f i x , y) = exy
15. f ( x , y) = x^' + y^ 1
+ 2xy 16. f i x , y) = (x - 4)ln xy
17. f i x , y) = x 3 + y 3 + 3x 2 - 18y 2 + 8 ly + 5
18. f i x , y) = 2x 3 + y 3 + 3x 2 - 3y - 12x - 4
19. f i x , y) = x 2 + y 2 - 6xy + 9x + 5y + 2
y © Find the absolute extrema o f f on the closed bounded set S in

y) = (1 + χ 2 + y 2) e x~χ 2 ~y2 ∣
t le p la n e a s
described in Problems 20-25.
z∣ ^ 20. f i x , y) = 2x 2 - y 2; 5, is the disk x 2 + y 2 ≤ 1.
21. f i x , y) = x 2 + 3y2 - 4x + 2y - 3; .S’ is the square with
vertices (0, 0), (3, 0), (3, -3 ) , and (0, -3 ).
22. f i x , y) = 2 sin x + 5 cosy; 5 is the rectangle with vertices
(0, 0), (2, 0), (2, 5), and (0, 5).
23. f i x , y) = e -P+2χ+y2- $ j s t ]l e r e gj0 n χ 2 + 2χ + y 2 ≤ 0.
24. ∕( x , y) = x 2 + xy + y 2 ; .S' is the square region
—1 ≤ x ≤ 1, and -1 ≤ y ≤ 1.
25. f ( x , y) = x 2 - 4xy + y 3 + 4y; 5 is the square region
0 ≤ x ≤ 2 , 0 ≤ y ≤ 2 .
Find the least squares regression line for the data points given 40 —50x + 40y bottles of California water and
in Problems 26-29. 20 + 60x - 70y bottles of the New York water each day.
26. ( - 2 , - 3 ) , ( - 1 , - 1 ) , (0, 1),(1,3), (3, 5) How should the owner price the bottled water to generate
27. (0, 1), (1, 1.6), (2.2, 3), (3.1, 3.9), (4, 5) the largest possible profit?
39. The telephone company is planning to introduce two new
28. (3, 5.72), (4, 5.31), (6.2, 5.12), (7.52, 5.32), (8.03, 5.67)
types of executive communications systems that it hopes
29. (-4 , 2), (-3 , 1), (0, 0), (1, -3 ), (2, -1 ), (3, -2 ) to sell to its largest commercial customers. It is estimated
30. A rectangular box with no top is to have a fixed volume. that if the first type of system is priced at x hundred
What should its dimensions be if we want to use the least dollars per system and the second type at y hundred
amount of material in its construction? dollars per system, approximately 40 - 8x + 5y consum
31. Show that of all triangles with fixed perimeter P, the ers will buy the first type and 50 + 9x - 7y will buy the
equilateral triangle has the greatest area. second type. If the cost of manufacturing the first type is
$1,000 per system and the cost of manufacturing the
32. Find the smallest perimeter of an isosceles triangle with
second type is $3,000 per system, how should the tele
fixed area A.
phone company price the systems to generate maximum
33. A dairy produces whole milk and skim milk in quantities profit?
x and y gallons, respectively. Suppose the price of whole 40. A manufacturer with exclusive rights to a sophisticated
milk is p(x) = 100 - x and that of skim milk is new industrial machine is planning to sell a limited
q(y) = 100 - y; also assume that C(x, y) = x 2 + xy + y 2 number of the machines to both foreign and domestic
is the joint-cost function of the commodities. Maximize firms. The price the manufacturer can expect to receive
the profit for the machines will depend on the number of machines
F(x, y ) = p x + q y - C(x, y) made available. For example, if only a few of the ma
chines are placed on the market, competitive bidding
34. Repeat Problem 33 for the case where p(x) = 4 - 5x, among prospective purchasers will tend to drive the price
<∕(y) = 4 - 2y, and the joint-cost function is C(x, y) up. It is estimated that if the manufacturer supplies
= 2xy + 4. x machines to the domestic market and y machines
35. A particle of mass m in a rectangular box with dimensions to the foreign market, the machines will sell for
x, y, z has ground state energy 60 - 0.2x + 0.05y thousand dollars apiece at home and
50 - 0.1y + 0.05x thousand dollars apiece abroad. If the
manufacturer can produce the machines at the total cost
E(x, λy, z) =
' ' f>m∖ χ2
+yΛ+Λ)
2z ∕ 2 of $10,000 apiece, how many should be supplied to each
market to generate the largest possible profit?
where k is a physical constant. If the volume of the box is 41. A college admission officer, Dr. Westfall, has compiled
fixed (say Fθ = xyz), find the values of x, y, and z that the following data relating students’ high-school and
minimize the ground state energy. college GPAs:
36. A manufacturer produces two different kinds of graphing
calculators, A and B. in quantities x and y (units of 1,000), HS GPA 2.0 2.5 3.0 3.0 3.5 3.5 4.0 4.0
respectively. If the revenue function (in dollars) is
College GPA 1.5 2.0 2.5 3.5 2.5 3.0 3.0 3.5
2
R(x, y) = - x - 2y2 + lx y + 8x + 5y
Plot the data points on a graph and find the equation of
find the quantities of A and B that should be produced in the least-squares line for this data. Then use the least
order to maximize revenue. squares line to predict the college GPA of a student whose
37. Suppose we wish to construct a rectangular box with high school GPA is 3.75.
volume 32 ft 3. Three different materials will be used in 42. It is known that if an ideal spring is displaced a distance y
the construction. The material for the sides costs $1 from its natural length by a force (weight) x, then y = kx,
per square foot, the material for the bottom costs $3 per where k is the so-called spring constant. In order to
square foot, and the material for the top costs $5 per compute this constant for a particular spring, a scientist
square foot. What are the dimensions of the least expen obtains the following data:
sive such box?
38. A store carries two competing brands of bottled water,
one from California and the other from New York. The X 5.2 7.3 8.4 10.12 12.37
owner of the store can obtain both at a cost of $2 per y (i∏∙) 11.32 15.56 17.44 21.96 26.17
bottle and estimates that if the California water is sold for
x dollars per bottle and the New York water for y
dollars per bottle, consumers will buy approximately Based on these data, what is the “best” choice for k?
Suppose you know that f and g both have a zero

Computational Window near (x0 , y 0) and need to find this zero much more
accurately. Write down the equations of the tan
43. Linearizing nonlinear data gent planes (in R 3 for f and g at the point (x0 , y 0 ,
In this problem, we turn to data that do not tend to z0)). Then, since you want both planes where they
change linearly. Often one can “ linearize” the data intersect in the xy-plane, derive the following pair
by taking the logarithm or exponential o f the data, of linear equations for the unknown intersection
then changing it to a linear fit. point (x 1, y ∕
a. Suppose we have, or suspect, a relationship - + - -
y = k x m . Show that by taking the natural loga ∕√ ⅛ Fo)(¾ ⅞) ∕√ ⅞ T0)(Tι To) = ∕( ⅞ To)
rithm of this equation we obtain a linear relation g f χ 0 , τ 0)(χ 2 ^ ∙⅞)+ gy(χ o> y 0^ y i ~ -1⅛) = τ 0)
ship:
Y = K + mX Solve this pair of equations given the first Newton
iterate (x 1, y 1). Why is this the analog in R 2 of the
Explain the new variables and constant K.
Newton method in one variable?
b. Below are data relating the periods o f revolution t
c. Do one Newton iteration to find (x 1, y 1).
(in days) of the six inner planets and their
semimajor axis a (in 106 km). Kepler conjectured f i x , y) = x 2 - 2x + 2y 2 + 4y - 4
the relationship t =ka"', which is very accurate for
the correct k and a. You are to “ transform” the lx + 2y∖
g(x, y) = 10 - 15 sιn(^— j — )
data as in part a (obtaining T i = In ti , . . . ); and
do a linear fit to the new data, finding k and m.
Then set up to take the second iteration; that is,
t-data: (87.97, 224.7, 365.26, going from (x 1, y 1) to (x2, y 2). If convenient,
686.98, 4332.59, 10 759.2) actually compute (x2 , y 2).
a-data: (58, 108, 149, 228, 778, 1 426) d. To test your progress in part c, evaluate f and g at
44. Below are data pertaining to the 1988 Olympic (x0 , y 0) and (x 1, y 1). Both functions should be
weight-lifting competition. The x-data are the “ class much smaller at (x l , y 1). Take two or three itera
data” giving seven weight classes from Featherweight tions and discuss the convergence.
to Heavyweight-2 (given in kg). The w,-data are the 46. We now turn to finding the maximum or minimum
combined weights lifted by the winners in each class. of a function, say ∕7(x, y). Recall that the primary
Theoretically, we would expect a relationship conditions are H χ = 0 and H y = 0. Here you are to
w = k x m , where m = (Can you see why?) Once apply the Newton method, as’ developed in Problem
45, to accurately find a zero for H χ and H Consider
again, linearize the data as in Problem 43.
a. Linearize the data as in Problem 43 and curve fit H{x, y) = exp^9x - x 2 - y 2 - y j
to find k and m.
x-data: (56, 60, 67.5, 75, 82.5, 90, 100, 110) a. By completing the square in H , show that the
y-data: (292.5, 342.5, 340, 375, unique maximum is at (4.5, —0.25).
377.5, 412.5, 425, 455) b. Letting H χ and H represent f and g in Problem 45,
b. Comment on the 60-km entry. Do you see why set up to apply Newton’s method to search for the
this participant (N. Suleymanoglu of Turkey) was local maximum of H . Derive equations:
referred to as the strongest man in the world? H (X O>T0)(* I ^ ⅞) + ¾ (⅞ > T0)(T1 - To) = ~ H X(X O>y 0)
XX
In Problems 45-46, you will see how the Newton-Raph- H 4 X V Λ>)(*I “ ⅞) + ⅞ √ ⅜ ⅞)(T1 - To) = -¾(⅞> k 0)
son method in one independent variable can be extended
to more than one variable. Then the Newton-Raphson c. Let (xθ, y 0 ) = (4.6, 0) and do one Newton iteration
method will be applied to an extrema problem in R 3. For on H χ and H in searching for the nearby extrema.
two functions z = ∕( x , y) and z = g(x), we are interested in Check for the accuracy of (x 1, y 1).
finding the “ zeros” — that is, values (x, y) such that z = 0 d. If your software allows, take another iteration
for both functions. Geometrically, this means finding the obtaining (x2, y,). How accurate is this value?
intersection points o f the two level curves in the xy-plane e. Newton’s method is quite sensitive to starting
defined by f i x , y) = 0 and g(x, y) = 0. values when there is more than one independent
45. a. Find the zeros for the functions defined by f i x , y) variable; so great care is needed. To see this, use
= 5x — 3y - 21 and g(x, y) = 2x + 6y + 6. (5, 0) as a starting value and report what happens.
Graph of the level curves and the zero(s) found. In f. Repeat c and d for this function, starting at (1, 0)
the linear case, could there be more than one point
such that both functions are zero? Explain. H (x, y) = exp(2x - x 2 - y 2)
b. Now assume that /and g are nonlinear functions. + 0.1(x 2 + y 2 - 2y + 1)
τr~-:--.....
. . .. ι ∙ ∙ ∙ ■:■■■ ∙.. >.J Λ , J, - - ∙ ∙ ∙ ∖ , X ∙. ■■■>■?-
Sec. 12.8 Lagrange Multipliers 805
θ 47. This problem was designed to show, by example, that if f 50. Prove the second partials test. Hint. Compute the second
is degenerate at the critical point, then almost anything directional derivative of fin the direction of u = Ai + Zcj
can happen. and complete the square.
a. Show that∕( x , y) = x 4 - y 4 has a saddle point at (0, 0). 51. Verify the formulas for m and b associated with the least
b. Show that g(x, y) = x 2y 2 has a relative minimum at squares approximation.
(0, 0). 52. This problem involves a generalization of the least
c. Show that h(x, v) = x 3 + y 3 has no extremum or squares procedure, in which a “least-squares plane” is
saddle point at (0, 0). found to produce the best fit for a given set of data. A
48. T HINK TANK P ROBLEM Consider the function researcher knows that the quantity y is related to x. and x 2
f{x, y) = (y ~ x 2)(y ~ 2x2). Discuss the behavior of this by a formula of the form y = k γx γ + k 2x 2, where k i and k 2
function at (0, 0). are physical constants. To determine these constants, she
conducts a series of experiments, the results of which are
49. T HINK TANK PROBLEM Sometimes the critical points of
tabulated as follows:
a function can be classified by looking at the level curves.
In each case shown in Figure 12.34, determine the nature
of the critical point(s) of f at (0, 0). *1 1.20 0.86 1.03 1.65 -0.95 -1.07
0.43 1.92 1.52 -1.03 1.22 -0.06
y 3.21 5.73 2.22 0.92 -1.1 1 -0.97
Modify the method of least squares to find a “best

approximation” for k. and L,.
b. z = -2
z = -1
z= 1
z=2
1 2 .8 LAGRANGE MULTIPLIERS
IN THIS SECTION Method of Lagrange multipliers, constrained

optimization problems, Lagrange multipliers with two parameters, a geometric
interpretation of Lagrange’s theorem
The general problem of optimizing a function subject to a constraint is the
subject of this section. The procedure we use is called the method o f Lagrange
multipliers, after the French mathematician Joseph Louis Lagrange (see the
historical note on page 812).
■ METHOD OF LAGRANGE MULTIPLIERS
In many applied problems, a function of two variables is to be optimized subject

to a restriction or constraint on the variables. For example, consider a container
heated in such a way that the temperature at the point (x, y, z) in the container is
Figure 12.35 Find the maximum
given by the function T(x, y, z). Suppose that the surface z = f(x , y) lies in the
value on the surface defined by container, and that we wish to find the point on z = f(x , y) where the temperature
f(x , y) = 1 - x 2 - y 2 subject to is the greatest. In other words, What is the maximum value o f T subject to the
the constraint x + y = 1 (x ≥ 0, constraint z = f(x , y), and where does this maximum value occur? (See Figure
y ≥ 0). See Example 1 for a solu 12.35.)
tion. We will use the following theorem to solve such a problem.
THEOREM 12.13 Lagrange’s theorem
Assume that f and g have continuous first partial derivatives and that f has
an extremum at P0(x0 , y0) on the smooth constraint curve g(x, y) = c. If
Vg(x0 , y0) ≠ θ, there is a number A such that
W o>y 0) = λVg(x0, y0)
Proof: Denote the constraint curve g{x, y) = c by C, and note that C is
smooth. We represent this curve by the vector function
R(t) = x (∕)i + y(∕)j
for all t in an open interval I, including t0 corresponding to Po, where x'(f) andy'(t)
exist and are continuous. Let F(t) = f[x(f), y(∕)] for all t in I, and apply the chain
rule to obtain
F '(0 = Λ $ + 4 ⅜ = V M 0 , y (0J ∙ R'(0
Because f(x , y) has an extremum at Po , we know that F(f) has an extremum at t0,
the value of t that corresponds to Po (that is, Po is the point on C where t = ∕ 0).
Therefore, we have F '(t 0) = 0 and
P∖ t0) = V∕[x(r 0), y(rθ)] ∙ R'(t0 ) = 0
If V ∕[x(∕ 0), y(t0)] = 0, then λ = 0, and the condition V ∕= AVgis satisfied (which is
trivial with A = 0). If Vf[x(t0), y(f0)] ≠ 0, then V∕[x(t 0), y(Z0)] is orthogonal to
R'(Z0). Because R'(Z0) is tangent to the constraint curve C, it follows that V∕(x0, y0)
is normal to C. But Vg(x0 , y0) is also normal to C (because C is a level curve of g),
and we conclude that V/and Vg must be parallel at Pθ. Thus, there is a scalar A
such that V∕(x 0 , y0) = AVg(x0, y0) as required. ■
■ CONSTRAINED OPTIMIZATION PROBLEMS
The general procedure for the method of Lagrange multipliers may be described as
follows.
Procedure for the Method of Suppose f and g satisfy the hypotheses of Lagrange’s theorem, and suppose that
Lagrange Multipliers f(x~) has an extremum (minimum and/or a maximum) subject to the constraint
g(x, y) = c. Then to find the extreme values, proceed as follows:
Step 1: Simultaneously solve the following three equations:
f x (x, y) = ¼ (x , y) f y (x, y) = λgy (x, y) g(x, y) = c
Step 2: Evaluate f at all points found in Step 1. The largest of these values is
the maximum value of f and the smallest of these values is the minimum
value of f.
EXAM PLE 1 Optimization with Lagrange multiplier
Find the largest and smallest values of ∕( x , y) = 1 - x 2 - y 2 subject to the

constraint x + y = 1 with x ≥ 0, y ≥ 0.
Solution Because the constraint is x + y = 1, let g(x, y) = x + y - 1.

f x (×, y) = - 2 x f y (x, y) = -2.y gχ (x, y) = 1 gy (x, y) = 1
Form the system
- 2 x = λ(l) Maximum ⅛, ∣ , 5)
- -2 y = λ(l)
fcx + y = 1
The only solution is x = y = ∣.
Figure 12.36 The maximum is the high point

of the curve of intersection of the surface and
the plane.
The endpoints of x + y = 1 where x ≥ 0, y ≥ 0 are at (1, 0) and (0, 1), and we

find that
/(1, 0) = 1 - l 2 - 0 2 = 0
/(0, 1) = 1 - 02 - l 2 = 0
Therefore, the maximum value is ∣ at (j> ∣) and the minimum value is 0 at

either (1,0) and (0, 1). M B
The method of Lagrange multipliers extends naturally to functions of three or

more variables. If a function f ( x , y, z) has an extreme value subject to a constraint
g(x, y , z) = c, then the extremum occurs at a point (x0 , y 0 , z0) such that g(x 0 , y 0 , z0 )
= c and V ∕( x 0 , y 0 , z 0 ) = λXg(x 0 , y 0 , z 0 ) for some number λ. Here is an example.
EXAM PLE 2 Hottest and coldest points on a plate
A container in R 3 has the shape of the cube given b y θ ≤ x ≤ 1, 0 ≤ y ≤ 1,

0 ≤ z ≤ 1. A plate is placed in the container in such a way that it occupies
that portion of the plane x + y + z = 1 that lies in the cubical container. If the
container is heated so that the temperature at each point (%, y, z) is given by
T(x, y, z) = 4 - 2x 2 —y 2 - z~
in hundreds of degrees Celsius, what are the hottest and coldest points on the
plate?
Solution The cube and plate are shown in Figure 12.37. We shall use Lagrange
multipliers to find all critical points in the interior of the plate, and then we
shall examine the plate’s boundary. To apply the method of Lagrange
multipliers, we must solve V T = λVg, where g(x, y, z ) = x + y + z - 1. We
obtain the partial derivatives
T x= -4 x Ty - -2J y Tz = —2z &gx = bgy = bg = 1
z
We must solve the system

'- 4 x = λ
-2 y = λ
Figure 12.37 Find the hottest and
coldest points on the plate inside ∣ —2z = λ
the cube. (v + y + z = 1
The solution of this system is x = y = z= and A = -⅛. The boundary

of the plate is a triangle with vertices H (l, 0, 0), B(0, 1, 0), and C(0, 0, 1). The
temperature along the edges of this triangle may be found as follows:
Edge AC: + z = l,y = 0 T 1(x) = 4 - 2x 2 - (0)2 - (1 - x) 2 = 3 - 3x 2 + 2x, 0 ≤ x ≤ l
= l ,z = 0 2
T 2(x) = 4 - 2X - (1 - X )2 - (0)2 = 3 - 3X 2 + 2X , 0 ≤ x ≤ 1
Edge AB: + j
Edge BC: + z = l ,x = 0 T 3(y) = 4 - 2(0)2 - y 2 - ( 1 - y)2 = 3 + 2y - 2y2, 0 ≤ y ≤ 1
Edge A C : Differentiating, T i '(x) = T ,'(x) = - 6 x + 2, which equals 0 when

x = ∣ . If x = I ’ then z = j (because x + z = l , y = 0on edge A C ), so we have
the critical point (∣> 0, ∣).
Edge AB: Because T 2 = T 1, we see x = ∣ . If x = ∣> then y = j (because x + y

- 1, z = 0 on edge B C ), so we have another critical point ( ∣ , j , θ).
Edge B C : Differentiating, T 3 '(y) = 2 - 4y, which equals 0 when y = ∣∙

Because y + z = 1 and x = 0, we have the critical point (θ, ∣ , ∣).
Endpoints of the edges: (1 ,0 , 0), (0, 1, 0), and (0, 0, 1).
The last step is to evaluate T at the critical points and the endpoints:
< 1 ’ 1) = 3-6;
< 0 , j ) = 3|; TG40) = 3⅛ T ( 0 ,H ) = 3.5;
T ( l,0 , 0 )= 2; T(0, 1,0) = 3; T(0, 0, 1) = 3

Comparing these values (remember the temperature is in hundreds of degrees
Celsius), we see that the highest temperature is 360 0C at (∣> ∣) and the
lowest temperature is 200 0C at (1, 0, 0). c ≡ε s
Notice that, in the end, we do not care about the value of λ. The multiplier is
just used as an intermediary device for finding the critical points. However, in
some problems, we do care about the value of A, because of the interpretation
given in the following theorem.
THEOREM 12.14 Rate of change of the extreme value
Suppose E is an extreme value (maximum or minimum) of f subject to the

constraint g(x, y) = c. Then the Lagrange multiplier A is the rate of change of E
with respect to c; that is, A = dE∕dc.
Proof: Note that at the extreme value (x, j) we have
∕λ- = λ S x , f y = λgj ,, and g(x, y) = c
The coordinates of the optimal ordered pair (x, y) depend on c (because different
constraint levels will generally lead to different optimal combinations of x and y).
Thus,
E = E(x, j j where x and y are functions of c
By the chain rule for partial derivatives:

dE _ ∂ E d x + ∂Edy
de ∂x de "1^ ∂y de
= f^ ∙ + f yTc Because E = f ( x ,y )
λg z√v + λg
dv
and
= s de y ~dc Because f x = ¼ fy = ¼
=λ p d x , p dy
∖‰ de + g y de,
-- λλ ⅛ Chain rule
de
dg
Because = 1 (remember g = c)
This theorem can be interpreted as saying that the multiplier gives the change
in the extreme value E that results when the constraint c is increased by 1 unit.
This interpretation is illustrated in the following example.
EXAM PLE 3 Maximum output for a Cobb-Douglas production function
If x thousand dollars is spent on labor, and y thousand dollars is spent on

equipment, it is estimated that the output of a certain factory will be
Q(x, y) = 50x2z5y 3z5
units. If $ 150,000 is available, how should this capital be allocated between
labor and equipment to generate the largest possible output? How does the
maximum output change if the money available for labor and equipment is
increased by $1,000? In economics, an output function of the general form
Q(x, y) = x ay l ~a is known as a Cobb-Douglas production function.
Solution Because x and y are given in units of $ 1,000, the constraint equation is
x + y = 150. If we set y(x, y) = x + y - 150, we wish to maximize Q subject to
g(x, y) = 0. To apply the method of Lagrange multipliers, we first find
Qx = 20x^3z5y 3z5 Qy = 30x2z5y^ 2z5 gχ = 1 gy = 1
Next, solve the system
'20x^ 3z5y 3z5 = λ
• 30x2z5y^ 2z5 = λ
.x + y - 150 = 0
From the first two equations we have
20x~3z5y 3z5 = 30x2z5y~2z5
20y = 30.x
y = 1.5x
Substitute y = 1.5x into the equation x + y = 150 to find x = 60. This leads to
the solution y = 90, so that the maximum output is
2(60, 90) = 50(60)2z5(90)3z5 ≈ 3,826.273 502 units
We also find that
λ = 20(60)^3z5(90)3z5 ≈ 25.50849001
Thus, the maximum output is about 3,826 units and occurs when $60,000 is
allocated to labor and $90,000 to equipment. We also note that an increase of
$1,000 (1 unit) in the available funds will increase the maximum output by
approximately λ ≈ 25.5 units (from 3,826.27 to 3,851.78 units). ≡≡≡
■ LAGRANGE MULTIPLIERS WITH TWO PARAMETERS
The method of Lagrange multipliers can also be applied in situations with more
than one constraint equation. Suppose we wish to locate an extremum of a
function defined by f(x , y, z) subject to two constraints, g(x, y, z) = c l and
h(x, y, z) = c2. By generalizing Lagrange’s theorem, it can be shown that if
(x0 , j 0, z0) is the desired extremum, then there are numbers λ and μ such that
g(x 0 , y 0’ ⅞) = Cl’ λ (⅞' ^0’ z θ) = ⅛ a n d
v ∕(x o , y 0, Zo) = λVg(x0 , y 0, z0) + μVΛ(x0, y 0 , z0)
provided that the functions/^ g, and h are continuously differentiable and that the
extremum occurs in the interior of a region where neither Vg nor V/z vanishes. As
in the case of one constraint, when searching for the optimum values, we find
values that satisfy this equation by solving a system of equations, and then we
must also check boundary points and points where the constraint gradients
vanish. We close this section with an example that illustrates this procedure.
EXAM PLE 4 Optimization with two constraints
Find the point on the intersection of the plane x + 2y + z = 10 and the

paraboloid z = x 1 + y 2 that is closest to the origin.
Intersection curve z Solution The distance from a point (x, y, z) to the origin is s = Vx 2 + y 2 + z 2,
but instead of minimizing this quantity, it is easier to minimize its square.
That is, we shall minimize f(x , y, z) = x 2 + y 2 + z 2 subject to the joint
constraints
g(x, y, z) = x + 2y + z = 10 and h(x, y, z) = x 2 + y 2 - z = 0
Compute the partial derivatives of f g, and h:
Find the point on the
intersection curve that x + 2y + z = 10 f x = 2x = 2j f = 2z
J Z
is closest to the origin.
gχ = 1 gy =2 gz = 1
Graphic representation of Exam
ple 4 hχ = 2x = 2y h_ = -1
To apply the method of Lagrange multipliers, we use the formula
W (⅛ To >⅞) = AVg(x0 , y0, z0) + μV∕z(x0 , y 0, zθ)
which leads to the following system of equations:
'2x = λ(l) + μ(2x)
2y = λ(2) + μ(2y)
- 2z= λ(l) + μ ( - l)
x + 2y + z = 10
„z = x 2 + y 2
This is not a linear system, so solving it requires ingenuity.
Multiply the first equation by 2 and subtract the second equation to obtain
4x —2y = (4x —2y)μ
(4x - 2y) - (4x - 2y)μ = 0
(4x - 2y)(l - μ) = 0
4x - 2y = 0 or 1- μ = 0
Case I: If 4x - 2y = 0, then y = 2x. Substitute this into the two constraint

equations:
x + 2y + z = 10 x2 + y2 - z = 0
x + 2(2x) + z = 10 x 2 + (2x)2 - z = 0
z = 10 - 5x z = 5x 2
By substitution we have 5x2 = 10 - 5x, which has solutions x = 1 and x = - 2 .
This implies
x = 1 x = —2
y = 2x = 2(l) = 2 y = 2x = 2(-2) = - 4
z = 5x 2 = 5(l) 2 = 5 z = 5x2 = 5 (-2 ) 2 = 20
Thus, the points (1, 2, 5) and ( - 2 , - 4 , 20) are candidates for the minimal
distance.
Case II:. If 1 - μ = 0, then μ = 1, and we look at the system of equations

involving x , y, z, λ, and μ.
f 2x = λ( 1) + μ(2x)
2y = λ(2) + μ(2p)
< 2z = λ(l) + μ ( - l)
x + 2y + z = 10
„z = x 2 + y 2
The top equation becomes 2x = λ + 2x, so that λ = 0. We now find z from the
third equation:
2z = - 1 , or z = -⅜
Now, turn to the constraint equations:
x + 2y + z = 10 x2 + y2 - z = 0
x + 2y - J = 10 x 2 + y2 + ∣ = 0
x + 2y = 10 + I x2 + / = - ∣
There is no solution because x 2 + y 2 cannot equal a negative number.

We check the candidates for the minimal distance:
∕( x , y, z) = x 2 + y 2 + z 2 so that
/(1, 2, 5) = 12 + 2 2 + 52 = 30
/ ( - 2 , - 4 , 20) = (-2 ) 2 + (-4 ) 2 + 202 = 420
Because ∕( x , y, z) represents the square of the distance, the minimal
distance is V30 and the point on the intersection of the two surfaces nearest to
the origin is (1, 2, 5). H M
Historical Note ■ A GEOMETRIC INTERPRETATION FOR LAGRANGE'S

THEOREM
Lagrange’s theorem can be interpreted geometrically. Suppose the constraint

curve g(x, y) = c and the level curves f(x , y) = k are drawn in the xy-plane, as
J OSEPH LAGRANGE
(1736-1813)
Some would call Joseph Louis o f∕(.r, y)
Lagrange one of the two great Figure 12.38 Increasing level curves and the constraint curve
est mathematicians of the 18th
century (Leonhard Euler being
the other). There is a distinct To maximize ∕( x , y) subject to the constraint g(x, y) = c, we must find the
difference in style between La highest level curve o f f that intersects the constraint curve. As the sketch in Figure
grange and Euler. Lagrange has 12.37 suggests, this critical intersection occurs at a point where the constraint
been characterized as the first curve is tangent to a level curve—that is, where the slope of the constraint curve
true analyst in the sense that he g(χ, y) = c is equal to the slope of a level curve ∕(x , y) = k. According to the
attempted to write concisely formula derived in Problem 45, Section 12.5,
and with rigor. On the other —g .
hand, Euler wrote using intui Slope of constraint curve g(x, y) = c is —pA
tion and with an abundance of *v
detail. Lagrange was described • -Λ
by Napoleon Bonaparte as “the Slope of each level curve is -y-.
lofty pyramid of the mathemat
ical sciences.” The condition that the slopes are equal can be expressed by
Lagrange followed Euler as
the court mathematician for
“A 'S x . . .. f x fy
= or, equivalently, — =—
Frederick the Great. He was the Jy Sy Sx Sy
first to use the notation ∕'( x ) Let λ equal this common ratio,
and f"(x ) for derivatives. In
this section we are introduced Λ fy
to the method of Lagrange mul
λλ = — andh λ
Λ =—
Sx Sy
tipliers, which provide a proce
dure for constrained optimiza so that
tion. This method was
contained in a paper on me ∕ x = ¼ . a n d f y = λ Sy
chanics that Lagrange wrote and Nf= f i + f y j = λ(gχ i + g ι,j) = λVg
when he was only 19 years old.
Because the point in question must lie on the constraint curve, we also have
g(x, y) = c. If these equations are satisfied at a certain point (a, b), then/will reach
its constrained maximum at (a, b) if the highest level curve that intersects the
constraint curve does so at this highest point. On the other hand, if the lowest level
curve that intersects the constraint curve does so at (a, b), then f achieves its
constrained minimum at this point.
PROBLEM SET 1 2 .8
θ Use the method o f Lagrange multipliers to fin d the required 25. There are 320 yd of fencing available to enclose a rectan
constrained extrema in Problems 1-12. gular field. How should the fencing be used so that the
1. Maximize ∕( x , y) = xy subject to 2x + 2y = 5. enclosed area is as large as possible?
2. Maximize f(x , y) = xy subject to x + y = 20. 26. Use the fact that 12 fl oz is approximately 6.89π in.3 to
3. Maximize f( x , y) = 16 - x 2 - y 2 subject to x + 2y = 6. find the dimensions of the 12-oz soda can that can be
constructed using the least amount of metal. Compare
4. Minimize f( x , y) = x 2 + y 2 subject to x + y = 24.
your answer with an actual can of Pepsi.
5. Minimize f ( x , y) = x 2 + y 2 subject to xy = 1.
27. A cylindrical can is to hold 4ιr in.3 of orange juice. The
6. Minimize f( x , y) = x 2 — xy + 2y 2 subject to 2x ÷ y = 22. cost per in. 2 of constructing the metal top and bottom is
2 2 2
7. Minimize f(x , y) = x - y subject to x + y = 4. 2 twice the cost per in.2 of constructing the cardboard side.
8. Maximize f( x , y) = x 2 - 2 y - y 2 subject to x 2 + y 2 = 1. What are the dimensions of the least expensive can?
9. Maximize f(x , y) = cosx + cosy subject to y = x + J. 28. Find the volume of the largest rectangular parallelepiped
10. Maximize ∕( x , y, z) = xyz subject to 3x + 2y + z = 6.
that can be inscribed in the ellipsoid
11. Minimize ∕( x , y, z) = x 2 + y 2 + z 2 subject to x - 2y + 3z = 4. 2 2
12. Minimize f( x , y, z) — x 2 + y 2 + z 2 χ2, + y + z
⅞ f = 1
2 2
subject to 4x + 2y + z = 4. 2
θ 13. Find the smallest value of ∕( x , y, z) = 2x 2 + 4y2 + z 2 29. A manufacturer has $8,000 to spend on the development
subject to 4x —8y + 2z = 10. What, if anything, can be and promotion of a new product. It is estimated that if x
said about the largest value off subject to this constraint? thousand dollars are spent on development and y thou
14. Let∕( x , y, z) = x 2y 2z 2 . Show that the maximum value o f / sand are spent on promotion, sales will be approximately
subject to x 2 + y 2 + z 2 = R 2 is R i !2~I. ∕( x , y) = 50x ιz2y 3z2 units. How much money should the
15. Find the maximum value o f∕(x , y, z) = x —y + z on the
manufacturer allocate to development and how much to
2 2 2 promotion to maximize sales?
sphere x + y + z = 100.
16. Find the minimum value o f∕(x , y, z) = x - y + z on the 30. If x thousand dollars are spent on labor and y thousand
2 2 2
sphere x + y + z = 100. dollars are spent on equipment, the output at a certain
factory will be Q(x, y) = 60x ιz3y 2z3 units. Assume that
17. Use Lagrange multipliers to find the distance from the
$120,000 is available.
origin to the plane A x + By + Cz = 0 where at least one
of A, B, C is nonzero. a. How should money be allocated between labor and
equipment to generate the largest possible output?
18. Find the maximum and minimum distance from the
2
origin to the ellipse 5x - 6xy + 5y = 4. 2 b. Use the Lagrange multiplier λ to estimate the change in
19. Find the point on the plane the maximum output of the factory that would result if
the money available for labor and equipment is in
2x + y + z = 1 creased by $1,000.
that is nearest the origin. 31. A university extension agricultural service concludes that
on a particular farm the yield of wheat per acre is a
20. Find the largest product of positive numbers x, y, and z function of water and fertilizer. Let x be the number of
such that their sum is 24. acre-feet of water applied and y be the number of pounds
21. Write the number 12 as the sum of three numbers x, y, z in of fertilizer applied during the growing season. The agri
such a way that the product xy 2z is a maximum. cultural service then concluded that the yield (measured
22. A rectangular box with no top is to be constructed from in bushels), represented by / can be defined by the
96 ft 2 of material. What should be the dimensions of the formula ∕( x , y) = 500 - x 2 —2y2 . Suppose that water
box if it is to enclose maximum volume? costs $20 per acre-foot, fertilizer costs $ 12 per pound, and
23. The temperature T at point (x, y, z) in a region of space is the farmer will invest $236 per acre for water and
given by the formula T = 100 - xy - xz - yz. Find the fertilizer. How much water and fertilizer should the
lowest temperature on the plane x + y + z = 10. farmer buy to maximize the yield?
24. A farmer wishes to fence off a rectangular pasture along 32. How much would the farmer of Problem 31 maximize the
the bank of a river. The area of the pasture is to be 3,200 yield if the amount spent is $100 instead of $236?
2
yd , and no fencing is needed along the river bank. Find 33. Present post office regulations specify that a box (that is, a
the dimensions of the pasture that will require the least package in the form of a rectangular parallelepiped) can
amount of fencing. be mailed parcel post only if the sum of its length and
d. What is the Lagrange multiplier in part c? Your answer

should suggest another method for solving part c.
Solve the problem using this alternative approach.
θ 40. A farmer wants to build a metal silo in the shape of a right
circular cylinder with a right circular cone on the top (the
bottom of the silo will be a concrete slab). What is the
Figure 12.39 Maximum volume least amount of metal that can be used if the silo is to have
for a box mailed by U.S. a fixed volume?
parcel post 41. Find the volume of the largest rectangular parallelepiped
(box) that can be inscribed in the ellipsoid
girth does not exceed 108 inches, as shown in Figure
12.39. Find the maximum volume of such a package. a2 + b 2 + c2 1
34. If x, y, z are the angles of a triangle, what is the maximum
value of the product P(x, y, z~) = sinx siny sinz? What (See Problem 28.)
about Q(x, y, z) = cosx cosy cosz? In Problems 42-45, let Q = f(x, y) be a production function in
Use the method o f Lagrange multipliers in Problems 35-38 to which x and y represent units o f labor and capital, respectively.
find the required extrema for the two given constraints. Ifp and q represent unit costs o f labor and capital, respectively,
35. Find the minimum of∕( x , y, z) = x 2 + y 2 + z 2 subject to then p x + qy represents the total cost o f production.
x + y = 4 and y + z = 6. 42. Use Lagrange multipliers to show that subject to a fixed
production level c, the total cost is smallest when
36. Find the maximum of ∕( x , y, z) = xyz subject to
x 2 + y 2 = 3 and y — 2z.
J = and Q = ∕( x , y) = c
37. Maximize ∕( x , y, z) = xy + xz subject to 2x + 3z = 5 and
xy = 4. (provided V ∕ ≠ 0, and p ≠ 0, q ≠ 0). This is often referred
38. Minimize f( x , y, z) = 2x 2 + 3y2 + 4z 2 subject to to as the minimum cost problem, and its solution is called
x + y + z = 4 and x - 2y + 5z = 3. the least cost combination of inputs.
39. A manufacturer is planning to sell a new product at the 43. Show that the inputs x, y that maximize the production
price of $150 per unit and estimates that if x thousand level Q = f(x , y) subject to a fixed cost k satisfy
dollars is spent on development andy thousand dollars on
promotion, approximately f f
—= — with px + qy = k
p q
320y 160X
y +2 x +4 (assume p Ψ 0, q Ψ 0). This is called a fixed-budget prob
lem.
units of the product will be sold. The cost of manufactur 44. A Cobb-Douglas production function is an output func
ing the product is $50 per unit. tion of the form Q(x, y) = cx ay β , with a + β = 1. Show
a. If the manufacturer has a total of $8,000 to spend on that such a function is maximized with respect to the
development and promotion, how should this money fixed cost px + qy = k when x = ak!p and y = βk∕q.
be allocated to generate the largest possible profit? Where does the maximum occur if we drop the condition
b. Suppose the manufacturer decides to spend $8,100 a + β = 1? How does the maximum output change if k is
instead of $8,000 on the development and promotion increased by 1 unit?
of the new product. Estimate how this change will 45. Show that the cost function C(x, y) = p x + qy is mini
affect the maximum possible profit. mized subject to the fixed production level A x ay fi = k,
c. If unlimited funds are available, how much should the k(aq∖ β , k(βp∖ a
manufacturer spend on development and promotion with a + β = 1 when x = -? — and y, = -? — .
' A∖ βp) - A∖ aqj
to maximize profit?
CHAPTER 12 REVIEW

Concept Problems 29. State Lagrange’s theorem.

1. What is a function of two variables? 30. State the procedure for the method of Lagrange multi
2. What is the domain of a function of two variables? pliers.
3. What is a level curve of the function defined by f(x , y)? 31. State the rate of change of the extreme value theorem.
4. What do we mean by the limit of a function of two
variables? Practice Problems
5. State the following properties of a limit of functions of 32. If∕(x , y) = sin - 1 xy, verify t h a t ∕ ζy = f y χ .
two variables. 33. Let w = x 2y + y 2z, where x = t sin t, y = t cos t, and
a. scalar rule b. sum rule z = 21. Use the chain rule to find dw/dt where t - π.
c. product rule d. quotient rule 34. Let∕(x , y, z) = xy + yz + xz, and let Po denote the point
6. Define the continuity of a function defined by f(x , y) at (1 ,2 ,-1 ).
a point (x0, y0) in its domain. a. Find the gradient of f at PQ.
7. If z = fix , y), define the first partial derivatives off with b. Find the directional derivative of / i n the direction
respect to x and y. from Po toward the point 2 (-1 , 1, -1).
8. What is the slope of a tangent line to the surface defined c. Find the direction from Pn in which the directional
by z = ∕(x , y) that is parallel to the xz plane? derivative has its largest value. What is the magni
9. If z = fix , y) find the second partial derivatives and the tude of the largest directional derivative at //?
mixed partial derivatives. 35. Show that the function defined by
10. If z = fix , y), what are the increments of x, y, and z?
11. What does it mean for a function of two variables to be if (x, y) ≠ (0, 0)
differentiable at (x0, y 0)2 ∕( χ , y) =
if(x,y) = (0, 0)
12. State the incremental approximation of fix , y).
13. Define the total differential of z = ∕(x , y). is not continuous at (0, 0).
14. State the chain rule for one parameter. 36. If∕(x , y) = l n Q , find f χ , f y , f yy , andf χy .
15. State the chain rule for two parameters.
37. Show that if∕(x , y, z) = x 2y + y 2z + z 2x, then
16. Define the directional derivative of a function defined
by z = ∕(x , y). ∂f 'fi ∂f ,
+ + χ z
17. Define the gradient V∕(x, y). y ∂=z = (v + ∙y + )'
∂□x τ∂~
18. State the following basic properties of the gradient.
38. Let ∕(x , y) = (x2 + y 2)2. Find the directional derivative
a. constant rule b. linearity rule of f at (2, —2) in the direction that makes an angle of
c. product rule d. quotient rule 2ττ∕3 with the positive x-axis.
e. power rule 39. Find all critical points of∕(x , y) = (x2 + y 2)e- F-2 +y2).
19. Express the directional derivative in terms of the gradi
ent.
20. State the optimal direction property of the gradient
(that is, the steepest ascent and steepest descent).
21. State the normal property of the gradient.
22. Define a normal line and a tangent plane to a surface S
at a point Pθ.
23. Define the absolute extrema of a function of two varia
bles.
24. Define the relative extrema of a function of two varia
bles.
25. What is a critical point of a function of two variables? Figure 12.40 Graph of∕(x , y) = (x2 + y 2)e (χ 2 +r2)
26. State the second partials test.
27. State the extreme value theorem for a function of two 40. Use the method of Lagrange multipliers to find the
variables. maximum and minimum values of the function
28. What is the least squares approximation of data, and ∕( x , y) = x 2 + 2y2 + 2x + 3 subject to the constraint
what is a regression line? x 2 + y 2 = 4.
Describe the domain o f each function given in Problems 1-4. 26. Find and z f where z = x 2 - y 2, and x = u + 2v,
,_________________ γ 2 _ v 2 ∂ιι ∂v '
1. ∕( x , y) = V 1 6 - x 2 - y 2 2. f( x , y) = ^ _ ~ y = u — 2v.
A y
27. Find ⅜=∙ and τ f where z = x t a n - ’ and x = uv, y = - ∙
3. ∕( x , y) = sin ^ 'x + cos^'y 4. f ( x , y) = ex+k'tan~ 1 ( - ∂u ∂v y - v
Use implicit differentiation to find ~ and γ ∙ in Problems

Find the partial derivatives f χ a n d f v for the function defined in 28-33. ' y
Problems 5-10. 2 2 2
γ∙2 — p2 28. x + 6y + 2z = 5 29. e + ey + e z = 3
x
5. ∕( x , y) = χ + 6. f( x , y) = x 3e 3>'z2x 3 2
30. x + 2xz - y z - z = 1 3
7 ∙∕( x , F) = x 2y + sin^ 8. f( x , y) = Inf— 31. ln(2x —z 2) —(x —y)e 2z = 0

A ∖A * ~y 32. x 3 + y 3 + z 3 = 6 33. x + 2y - 3z = In z
9. f(x , y) = 2x 3y + 3xy 2 +∣ 1 0 .∕'(x ,y ) = xye xy In Problems 34-41, fιn d f χ , f y , f χχ , f y χ .

34. ∕( x , y) = tan ' 1xy '. 35. ∕( x , y) = sin 1 xy
For each function given in Problems 11 —15, describe the level 2 3
36. f( x , y) = x + y - 2xy 2 '. f( x , y) = eχ2+y2
37.
curve or level surface f = c for the given values o f the constant c. 38. ∕( x , y) = x In y . f(x , y, z) = x i e 3y' z2
'39.
11. ∕( x , y) = x 2 - y; c = 2, c = - 2
12. f(x , y) = 6x + 2y; c = 0, c = l , c = 2 40. ∕( x , y) = J sin(cost)Λ
,, , x f V x 2 + y 2 if x ≥ 0 . ,
1 3 . r ∕ ( x ,y ) = ∣w π
■ .f x < ( ) c = 2, c = l,c = - l 41. ∕( x , y, z) = x 3y 2Vz + sin(x 2 + y In z)
14. f(x , y, z) = x 2 + y 2 + z 2; c = 16, c = 0, c = -2 5 Find equations for the tangent plane and normal line to the
y2 z2 surfaces given in Problems 42-44 at the prescribed point.
15. f(x , y, z) = x2 + y + y ; c = 1, c = 2
42. x 2y 3z = 8 at P0(2, - 1 , -2 )
43. x 2 + 2xy 2 - 7x 3 + 3y + 1 = 0 at P 0 (l, 1, 1)
Evaluate the limits in Problems 16-19, assuming they exist.
1.m
lι
xy 44. z = - ---- 2------ 2-, at P flv(l, 1, - 1)
16. . ^ ∕
Λ n v 2 -I- -∣
I √
r 2+ x +y ° ’
x2 - y2 Find all critical points o f f(x , y) in Problems 45-50 and
17. ⅛ , 4
classify each as a relative maximum, a relative minimum, or a
x2 - y 2
saddle point.
18.
45. ∕( x , y)= x 2 - 6x + 2y 2 + 4y - 2
x + ye 46. ∕( x , y)= x 3 + y 3 - 6xy
19. lιm — ---------
(x,v)→ (0,0) 1 + xi 47. ∕( x , y)= (x - l)(y - l)(x + y - 1)
48. ∕( x , y)= x 2 + y 3 + 6xy - 7x - 6y
Show that each limit in Problem 20-23 does not exist. 49. ∕( x , y)= x 3 + y 3 - 3x 2 - 18y2 + 81y + 7
v 3 _ ,,3
20. lim — :— — j 50. ∕( x , y) = sin(x + y) + sinx + siny for 0 < x < π,
(Λ-,y)→(0,0) x i + yi 0 < y < 77 (See Figure 12.41.)
V* - »4
21. lim ,4 + y4
∙2 - 2y2
22. lim - - .-1,----- ,2;
(x,y)→ (0,0) 2x + y
23. lim
Find the derivatives in Problems 24-27 using the chain rule.

You may leave your answers in terms o f x, y, t, u, and v.
24. Find ⅛ where z = —xy + y 3, and x — - 3 t 2 , y = 1 + t i .
dt y ' y
25. Find y where z = xy + y 2 , and x = e l t~ l , y = tan l.

Figure 12.41 Graph o f∕( x , y) = sin(x + y) + sinx + siny
51. Use the chain rule to find ⅛ if z = x 2 —3xy2 ; x = 2t, 70. A plate is heated in such a way that its temperature at a
at λ
point (x, y) measured in centimeters on the plate is given
y = t 2.
in degrees Celsius by
dz
52. Use the chain rule to find if z = x In y; x = 2t, y = e , .
64
T(x, y) =
53. Let z = ue"2 ~v2, where u = 2 x 2 + 3y2 and v = 3x 2 - 2y 2 . x 2 + y2 + 4
Use the chain rule to find ~ and
∂x ∂y a. Find how fast the temperature is changing at the
point (3, 4) in the direction 2i + j.
54. Use implicit
r differentiation to find τ r and where
∂x ∂y b. Find the direction and the magnitude o f the greatest
x , y, and z are related by the equation rate of change of the temperature at the point (3, 4).
x 3 + 2xz —y z 2 - z 3 = 1.
71. The marketing manager for a certain company has
55. Find the slope at the point Pθ(l, 1) on the level curve of compiled the following data relating monthly advertis
∕( x , y) = x 2 + y 2 that passes through PQ . ing expenditure and monthly sales (in units of $1,000).
56. Find the slope at the point P 0(2, 0) on the level curve of
f( x , y) = xe y that passes through PQ .
57. Find the equations for the tangent plane and normal Advertising 3 4 7 9 10
line to the surface z = sin x + ex y + 2y at the point
∕ ,o(O, 1,3). Sales 78 86 138 145 156
58. The electric potential at each point (x, y) in the disk
x 2 + y 2 < 4 is V = 2(4 - x 2 - y 2) - ιz 2 volts. Draw the a. Plot the data on a graph and find the least-squares
equipotential curves V = c for c = 8, 6, 4, and 2. line.
59. Let f ( x , y, z) = x 3y + y 3z + z 3x. Find a function b. Use the least-squares line to predict monthly sales if
r}f rif ∂f the monthly advertising expenditure is $5,000.
g(x, y, z) such that ^ + ^ + ⅛ = χ 3 + T 3 + ^3 + 3g(x, y, z).
72. F i n d ∕ cc + f χ y + f y y , where ∕( x , y) = x 2y 3 + x 3y 2 .
60. Let u = s in - + In - ∙ Show that y τ r + x ~ = 0. 73. F in d / , where ∕( x , y, z) = cos(x 2 + y 3 + z 4 )
y x ■ ∂y ∂x
74. Let z = ∕( x , y), where x = t + cos t and y = e l .
61. Let w = ln(l + x 2 + y 2) - 2tan - l y, wherex = ln(l + t2) a. Suppose f χ ( l, 1) = 4 and /(1, 1) = - 3 . Find ⅜
and y = e t . Use the chain rule to find when t = 0. '
b. Suppose∕v(0, 2) = -1 and∕J.(0, 2) = 3. Find 3z∕∂rand
62. L e t ∕( x , y) = tan -1 '∖ Find the directional derivative o ff
∂z∕∂θ at the point where r = 2, θ = J ’ and x = r cos θ,
at (1, 2) in the direction that makes an angle o f J with y = r sin θ.
the positive x-axis.
75. Suppose / has continuous partial derivatives in some
63. Let ∕( x , y) = y x . Find the directional derivative of f at
(3, 2) in the direction toward the point (1, 1). region in the plane that contains another region D , and
suppose∕( x , y) = 0 for all (x, y) in D . If (1, 2) is in D and
64. Find the directional derivative o f ∕( x , y) = (xy)-, ' at the /(1 , 2) = 4 and/(1, 2) = 6, find dy/dx when x = 1 and
point (2, 2) in the direction of the vector i - j. y = 2.
65. Let ∕( x , y, z) = z(x - y) 5 + xy 2z 3 . 76. Suppose V ∕( x , y, z) is parallel to the vector x i + yj + zk
a. Find the directional derivative of/at (2, 1, - 1) in the for all (x, y, z). Show that/(0, 0, α) = /(0, 0, -a ) for any
direction o f the outward normal to the sphere a.
x 2 + y 2 + z 2 = 6. 'll. Find two unit vectors that are normal to the surface
b. In what direction is the directional derivative at given by ∕( x , y) = sin x + ex y + 2y at the point (0, 1).
(2, 1, -1 ) largest? 78. Let ∕( x , y) = 3(x - 2)2 - 5(y + 1)2 . Find all points on
66. Find positive numbers x and y for which xyz is a the graph of / where the tangent plane is parallel to the
maximum, given that x + y + z = 1. plane 2x + 2y - 1 = 0.
67. Maximize ∕( x , y, z) = x 2yz given that x, y, and z are all 79. Let z be defined implicitly as a function of x and y by the
positive numbers and x + y + z = 12. ∂2z
equation cos(x + y) + cos(x + z) = 1. Find in
terms of x, y, and z. x ?
68. Find the shortest distance from the origin to the surface
y 2 —z 2 = 10. 80. Suppose F and F ' are continuous functions of t and that
69. Find the shortest distance from the origin to the surface F'{t) = 0. Define / b y ∕( x , y) = F(x 2 + y 2).
z 2 = 3 + xy. a. Find V ∕( x , y).
b. Show that the direction of V ∕(α , ⅛) is the same as the This is Laplace’s equation in polar coordinates.
direction of the line joining (a, ⅛) to (0, 0). 89. Suppose the angle and radius of a circular sector are
81. L e t ∕( x , y) = 12x^1 + 18y^1 + xy, where x > 0, y > 0. allowed to vary. Use Lagrange multipliers to find the
Find the relative minimum for x > 0, y > 0, if one angle of the circular sector for which the perimeter o f the
exists. sector is smallest, assuming the area o f the sector is a
82. Let f ( x , j) = 3.r4 - 4.r2y + y 2 . Show that /h as a mini fixed constant.
mum at (0, 0) on every line y = m x that passes through 90. For the production function given by O(x, j>) = x ay b,
the origin,/has a minimum at (0, 0). Then show that / where a > 0 and b > 0, show that
has no relative minimum at (0, 0).
x τx + y τd y ^ a + b^
83. Find the minimum o f x 2 + y 2 + z 2 subject to the con
straint ax + by + cz = 1 (with a ≠ 0, b Ψ 0, c ≠ 0).
84. Suppose 0 < a < 1 and x ≥ 0, y ≥ 0. Find the mini In particular, if b = 1 - a with 0 < a < 1,
mum o f x ay x~a subject to the constraint
ax + (1 - a~)y = 1.
85. The geometric mean o f three positive numbers x , y , z i s G 91. The diameter of the base and the height of a right
= (xyz) 1/3 and the arithmetic mean is A = ∣(x + y + z). circular cylinder are measured, and the measurements
Use the method o f Lagrange multipliers to show that are known to have errors of at most 0.5 cm. If the
G(x, y , z) ≤ A(x, y, z) for all x , y, z. Hint. Maximize diameter and height are taken to be 4 cm and 8 cm,
G(x, y, z) subject to a suitable constraint. respectively, find bounds for the propagated error in:
86. Liquid flows through a tube with length L centimeters a. the volume V of the cylinder.
and internal radius r centimeter. The total volume V of
b. the surface area .S' of the cylinder.
fluid that flows each second is related to the pressure P
p ι 92. A right circular cone is measured and is found to have
and the viscosity a o f the fluid by the formula V = 7r r
base radius R = 40 cm and altitude h = 20 cm. If it is
What is the maximum error that can occur in using this known that each measurement is accurate to within 2%,
formula to compute the viscosity a, if errors of ± 1% can what is the maximum percentage error in the measure
be made in measuring r and L , ±2% in measuring V, ment of the volume?
and ±3% in measuring P ?
93. An elastic cylindrical container is filled with air so that
87. Suppose the functions / and g have continuous partial the radius of the base is 2.02 cm and the height is 6.04
derivatives and satisfy cm. If the container is deflated so that the radius o f the
d f_d g ⅛ base reduces to 2 cm and the height to 6 cm, approxi
n — n αnd
and _ — _
∂x ∂y ∂y ∂x mately how much air has been removed? (Ignore the
thickness of the container.)
This is called the Cauchy-Riemann condition. 94. Suppose / is a differentiable function o f two variables
a. Show that level curves of / and g intersect at right with/, and / also differentiable, and assume t h a t ∕ p and
angles provided that V ∕≠ 0 and Vg ≠ 0. f are continuous. Let u = ai + ⅛j be a unit vector in the
b. Assuming that the second partials of / and g are Λψ-plane. The second directional derivative D u f ( x , y) o f/
continuous, show that/and g satisfy Laplace’s equa at the point (a, ⅛), where a and b are the components o f u
tions in the direction of the unit vector u, is defined by
D a f ( x , y) = D u ∖p j ∖x , y)]. Show that
jf x x + j/yy= 0 and ° gx x + ° gyy = 0.
D 2J ∖x , y) = a 2f χ χ (x, j>) + 2abfxy (x, y) + b 2f yy (x, y)
88. Show that if z = ∕( r , θ), where r and θ are defined
implicitly as functions of x and y by the equations Hint: Use P 0 ^ y ,
x = r cos θ, y = r sin θ, then the equation
95. What can be said about the sum o f the intercepts of a
* ⅛ + θ ⅛ = 0 becomes + + = 0
tangent plane to the surface V x + V y + V z = v k for a
∂x 2 ∂y2 ∂r2 r2 ∂θ2 r dr constant kt
96. A pharmaceutical capsule is a cylinder o f radius r and

length €, capped on each end by a hemisphere. Assume
that the capsule dissolves in the stomach at a rate
proportional to the ratio R = S∕V, where Uis the volume
and .S' is the surface area of the capsule. Show that
⅞ ≤<0 and ∣ f < 0

∂r
97. T H IN K T A N K P RO BLEM Find the minimum distance

from the origin to the paraboloid z = 4 - x 2 - 4y2 . The
graph is shown in Figure 12.42. The distance from
P(x, y, z) to the origin is
Soap bubbles form minimal surfaces. For an interesting

d = V x 2 + y 2 + z2
discussion, see “ The Geometry of Soap Films and Soap
Bubbles,” by Frederick J. Almgren, Jr., and Jean E. Taylor,
This distance will be minimized when d 2 is minimized. Scientific American.
Thus, the function to be minimized (after replacing x 2 by
4 - 4y2 —z) is 98. A minimal surface is a surface that has the least surface
D(x, y) = 4 —4y2 —z + y 2 + z 2 = 4 — 3y2 + z 2 —z. area o f any surface with a given boundary. It can be
shown that if z = f( x , y) is a minimal surface, then
Setting the partial derivatives D v and D_ to 0 gives the
critical pointy = 0, z = 0.5. Solving fo rx on the parabo ( 1 + Z∕) ⅛ - ¾ Y + ( 1 + Z Λ∙2)⅛ = 0
loid gives the points 2 1(V zT5, 0, 0.5) and Q 2( - V T 5 , (A cosy'
0, 0.5). These points are N O T minimal because (0, 1, 0) z = In
\T? cosx
is closer. Explain what is going on here. Our thanks to
Herbert R. Bailey, who presented this problem in The is a minimal surface, find A and B.
College Mathematics Journal, May 1991: “ ‘Hidden’ b. Is it possible to find C and D so that
Boundaries in Constrained M a x -M in Problems,” z = Cln (sinx) + D ln(siny) is a minimal surface?
p. 227. 99. P U T N A M E X A M IN A T IO N PRO BLEM Let f be a
real-valued function having partial derivatives that is
defined for x 2 + y 2 < 1 and satisfies ∖f ( x , y)∣ ≤ 1.
Show that there exists a point (x0 , y 0) in the interior of
the unit circle such that [ ∕ t(x0 , y 0)]2 + [∕,(⅞> J⅛)l2 s 1 6.
100. P U T N A M E X A M IN A T IO N PRO BLEM Find the
smallest volume bounded by the coordinate planes and a
tangent plane to the ellipsoid
x 2 , P2 , z 2 = 1
a 2 2b 2 c
101. P U T N A M E X A M IN A T IO N PRO BLEM Find the
shortest distance between the plane
A x + By + C z + 1 = 0 and the ellipsoid
+ + < ι
Figure 12.42 a2 b2 c2 ~
"Desertification”
A friend of yours named Maria is studying the causes of the continuing expansion
of deserts (a process known as desertification). She is working on a biological index
of vegetation disturbance, which she has defined. By seeing how this index and
other factors change through time, she hopes to discover the role played in
desertification. She is studying a huge tract of land bounded by a rectangle; this
piece of land surrounds a major city but does not include it. She needs to find an
economical way to calculate for this piece of land the important vegetation
disturbance index fix, y). The domain for Jis all (x, y) in the rectangle but not in
the city.
M aria has embarked upon an ingenious approach of combining the results of
photographic and radar images taken during Rights over the area to calculate the
index J. She is assuming that Jis a smooth function. Although the flight data does
not directly reveal the values of the function J, it gives the rate at which the values
of J change as the flights sweep over the landscape surrounding the city. Her staff
has conducted numerous Rights, and from the data she believes she has been able
to find actual formulas for the rates at which J changes in the east-west and north
south directions. She has given these functions the names ΛJ and N. Thus, Λfix, y)
is the rate at which J changes as one sweeps in the positive x-di recti on and N(x, y)
is the corresponding rate in the y-direction. Maria shows you these formulas:
M(x, y) = 3.4eλ'b,- 7 ∙8)2 and N(x, y) = 22 sin (75 - 2xy)
Convince her that these two formulas cannot possibly be correct. Do this by
showing her that there is a condition that the two functions M and N must satisfy
if they are to be the east-west and north-south rates of change of the function J and
that her formulas for M and N do not meet this condition. However, show Maria
that if she can find formulas for M and N that satisfy the condition that you
showed her, it is possible to find a formula for the function J from the formulas for
Mathematics in its pure form, as
arithmetic, algebra, geometry, and M and N.
the applications o f the analytic
method, as well as mathematics
applied to matter and force, or
statics and dynamics, furnishes the
peculiar study that gives to us,
whether as children or as men, the
command o f nature in this its
quantitative aspect, mathematics
furnishes the instrument, the tool
o f thought, which we wield in this
realm.
W. T. Harris
Psychological Foundations o f Edu *Marcus S. Cohen, Edward D. Gaughan, R. Arthur Knoebel, Douglas S. Kurtz, and David J.
cation (New York, 1898), p. 325. Pengelley, “Priming the Calculus Pump: Innovations and Resources,” MAA Notes 17(1991).
Multiple Integration
■ CONTENTS
13.1 Double Integration Over
Rectangular Regions
Definition of double integral;
Volume interpretation; Iterated
integration; An informal
argument for Fubini’s theorem PREVIEW
13.2 Double Integration Over
Nonrectangular Regions In this chapter we introduce the concept of multiple integration, which
Nonrectangular regions; More is used in much the same way as single integration. For instance, just as
on areas and volumes;
Reversing the order of single integrals can be used to compute area, we shall find that double
integration in a double integrals (and triple integrals in a different way) are used to compute
integral; Properties of double volume. In Chapter 4, we found that a single integral can be used to
integrals
“ add” small pieces of a quantity to obtain a definition of the “ total”
13.3 Double Integrals in Polar
Coordinates quantity over a given interval.
Change of variables to polar We are interested in using multiple integrals to define and compute
form; Improper double
integrals in polar coordinates surface area, moments and centroids, and some special quantities in
13.4 Surface Area probability theory, business, and economics. In many cases, we shall
Definition of surface area; find it is more natural and computationally easier to use multiple
Surface area projections; Area integrals to define and compute a quantity that can also be handled by
of a surface defined
parametrically single integrals.
13.5 Triple Integrals
Definition of triple integral;
Iterated integration; Volume
by triple integrals
PERSPECTIVE
13.6 Mass, Moments, and
Probability Density Functions We first learned about integration in Chapter 4, and it has played an important
Mass and center of mass;
Moments of inertia; Joint
role in many subsequent topics in this text. The kind of integral introduced in
probability density functions Chapter 4 has been denoted by
13.7 Cylindrical and Spherical
Coordinates [ ∕( v ) dx
Cylindrical coordinates; Ja
Integration with cylindrical
coordinates; Spherical where the integrand f(x ) is continuous on the interval of integration [a, ⅛].
coordinates; Integration with From now on, we shall refer to this kind of integral as a single integral. In this
spherical coordinates
chapter, we shall introduce a generalization o f the single integral called a
13.8 Jacobians: Change of multiple integral. Such an integral is multiple in the sense that the integrand is a
Variables
Change of variables in a function o f several variables. In particular, a double integral is one in which the
double integral; Change of integrand f( x , y) involves two independent variables, and in a triple integral the
variables in a triple integral
integrand f(x , y, z) involves three independent variables.
Chapter 13 Review
821
822 Ch. 13 Multiple Integration
1 3 .1 DOUBLE INTEGRATION OVER RECTANGULAR REGIONS
IN THIS SECTION Definition of double integral, volume interpretation,

iterated integration, an informal argument for Fubini’s theorem
We introduce the concept of a double integral— that is, an integral with respect
to two independent variables. In this and the next section, we see how double
integrals can be interpreted geometrically in terms of volume.
■ DEFINITION OF DOUBLE INTEGRAL
We begin by defining the double integral f ( x , y) dA, where f ( x , y) is continuous

R
over the rectangle R: a ≤ x ≤ b, c ≤ y ≤ d. The definition requires the ideas
and notation shown in the following three steps.
Step 1: Partition the interval a ≤ x ≤ b into m parts and the interval c ≤ y ≤ d
into n parts. Using these subdivisions, we partition the rectangle R into
N = mn cells (subrectangles), as shown in Figure 13.1. Call this partition P.
Step 2: Choose an arbitrary point {xk , y k ) from each cell in the partition of the
rectangle. Form the sum
J 1M ¾ ) ½
where ΔA k is the area of the ∕rth approximating rectangle. This is called the
Riemann sum of f( x , y) with respect to the partition P.
Figure 13.1 A partition of the rec Step 3: We define the norm ∣∣F, ∣∣ of the partition to be the length of the longest
tangle A into m n cells diagonal of any rectangle in the partition. To refine the partition means to add
more cells in such a way that the norm decreases. When this process is applied
to the Riemann sum and the norm decreases indefinitely, we write
lim ∑ ∕( x * y*) ΔA,k

∣∣∕>∣∣→o ⅛=rv k ' k '
We now use this notation and terminology to define a double integral.
Double Integral I f / is defined on a closed, bounded region R in the xy-plane, then the double
integral of f over R is defined by
∣ ∫ ∕( x , y) dA = j fim o I f ( x fk , y*) Δ Λ ft I
provided this limit exists. If the limit exists, we say that /is integrable over R.
Remarks:
1. The number N of cells depends on the partition P.
2. As ∣∣P∣∣ → 0, it follows that N → ∞.
3. More formally, the phrase “ provided this limit exists” means:
If I is the double integral,
I= l y^dλ,
Sec. 13.1 Double Integration Over Rectangular Regions 823
then for any ε > 0, there exists a δ > 0 such that
∕- ∑ ∕( ⅞ X ) ∆ 4 < ε
fc=l
whenever / f(x*, y*)ΔA, is a Riemann sum whose norm satisfies ∣ ∣

P∣∣< δ.
fc=l
4. It can be shown that if f(x , y) is continuous on R, then it is bounded and
integrable on R.
■ VOLUME INTERPRETATION
h
We know that the single integral f f(x ) dx can be interpreted as the area under the
curve y = f(x) over the interval [a, b], and the double integral f(x , y) dA has a
J RJ
similar interpretation in terms of volume. To see this, note that if f(x , y) ≥ 0 on

the region R and we partition R using rectangles, then the product f(x*k , y⅛) ΔA k is
the volume of a parallelepiped (a box) with height f(x*k , y*k ) and base area ΔA k , as
, * * zv * *∖x
( ⅞ ⅞ ∕( ⅛ y * ) )
Hei⅛⅛t∕(Λ* y*)
AΛrea oΓf
R . is Δ A ,..
Figure 13.2 The approximating parallelepiped has volume ΔVk = f(x*k , y*) ΔA k .
Thus, the Riemann sum
Σ ∕( ⅞ y*k ) ∆Ak
K—1
provides an estimate of the total volume under the surface z - f(x, y) over the
rectangular domain R. I f / i s continuous on R, we expect the approximation to
improve if we take a more refined partition of R (that is, more rectangles with
smaller norm). Thus, it is natural to define the total volume under the surface as
the limit of Riemann sums as the norm tends to 0. That is, the volume under
z = f(x, y) over the domain R is given by
Volume approximated by rectan v
gular parallelepipeds = llli∞ ∑ι ∕( ⅛ >fc) A A k = f / f ( x >k) dA
EXAM PLE 1 Evaluate a double integral by relating it to a volume
Evaluate J J (2 - y) dA, where R is the rectangle in the xy-plane with vertices
(0, 0), (3, 0), (3, 2), and (0, 2).

Solution Because z = 2 —y satisfies z ≥ 0 for all points in R , the value of the

double integral is the same as the volume of the solid bounded above by the
plane z = 2 - y and below by the rectangle R. As shown in Figure 13.3, this
solid is a wedge whose volume can be found by looking at it sideways— that is,
with a triangular base of area B and height h = 3. We use the formula V = Bh.
Because the base is a triangle of side 2 and altitude 2, we have B = ⅛(2)(2) = 2,

and
K = 5 Λ = [∣(2)(2)](3) = 6
Therefore, the value of the integral is also 6; that is,

f f (2 - y) dA = 6
J RJ
■ ITERATED INTEGRATION
Just as we did not find derivatives or integrals using their limit definitions, we see
that it is not practical to evaluate a double integral even over a simple rectangular
region by using the definition. Instead, we shall compute double integrals by a
process called successive partial integration. To be specific, if we hold the y-
variable constant (denoted by y) and integrate ∕'(x, y) with respect to x , we have a
single integral
∫ f( × , y) dx
which we shall refer to as the partial integral of f with respect to x. The dx tells us
that the integration is with respect to x and that y is to be held constant for the
integration, but when the process is complete the result is a function of y.
Similarly, the integral
∕( x , y) dy
is the partial integral of /w ith respect to y, which means that x is held constant
(denoted by x) and integration is with respect to y. This integral is a function of x.
In successive partial integration, we evaluate a double integral by integrating

first with respect to one variable and then again with respect to the other variable.
In partial integration, we work from the inside out, as indicated in the following
notation:
Integrate with respect

to x first, then with
respect to y.
J ]"/(*, y )d y d x = ∫ [ f f ( x , .V) ⅛-,] dx Integrate with respect
to y first, then with
respect to x.
Integrals of this form are said to be iterated integrals. Our next theorem tells us
how to evaluate iterated integrals, which in turn enables us to evaluate double
integrals. The theorem was first proved by the Italian mathematician Guido
Fubini (1879-1943) in 1907.
Historical Note
G UIDO F UBINI (1879-1943)
THEOREM 13.1 Fubini’s theorem over a rectangular region
Guido Fubini taught at the In
stitute for Advanced Study in If fix , y) is continuous over the rectangle R: a ≤ x ≤ b, c ≤ y ≤ d, then the
Princeton. He was nicknamed double integral
the “Little Giant,” because of
his small body but large mind. 1 1 f(x , y) dA
Even though the conclusion of
Fubini’s theorem was known
for a long time and successfully may be evaluated by either iterated integral; that is,
applied in various instances, it
was not satisfactorily proved in f j f ( x , y) dA (d f b
fix , y) dx fix , y) dy dx
a general setting until 1907. 'c Ja
H9 What this says: Instead of using the definition of a double integral,

consider the region R, namely, a ≤ x ≤ b and c ≤ y ≤ d, and evaluate either
of the iterated integrals
Limits of x (outside brackets) Limits of x (variable inside brackets)
√ Γ rb
or
. ↑
Limits of y (variable inside brackets) Limits of y (outside brackets)
Proof: We shall provide an informal, geometric argument at the end of this

section. The formal proof may be found in most advanced calculus textbooks. ■
Let us see how Fubini’s theorem can be used to evaluate double integrals. We
begin by taking another look at Example 1.
EXAMPLE 2 Evaluate a double integral by using Fubini’s theorem
Use iterated integrals to compute J J (2 - y) dA, where R is the rectangle with
vertices (0, 0), (3, 0), (3, 2), and (0, 2).
Solution The region of integration is the rectangle 0 ≤ x ≤ 3 , 0 ≤ y ≤ 2 (see

Example 1 and Figure 13.3). Thus, by Fubini’s theorem, the double integral
can be evaluated as an iterated integral:
•3 ∕∙2
(2 - y) dy dx Integrate inner integral
R
lo) Jo with respect to y.
,2' 2 3
∕∙3
dx = 4 -l 2 dx = l x =6
o 'o o
which is the same as the result obtained geometrically in Example 1.
EXAM PLE 3 Double integral using an iterated integral
Evaluate x 2y 5 dA, where R is the rectangle 1 ≤ x ≤ 2 , 0 ≤ y ≤ 1, using
an iterated integral with: a. y-integration first b. x-integration first
Solution The graph of the surface x 2y 5 over the rectangle is shown in Figure
13.4a. In your work, you would usually sketch only the rectangle, as shown in
Figure 13.4b.
In the preceding examples, the amount of work required to evaluate the

integrals was about the same if we integrated first with respect to either x or y.
However, this is not always the case, as illustrated by the following example.
Figure 13.4 Graph o f z = x 2y 5
and defining rectangle EXAM PLE 4 Choosing the order of integration for a double integral
x cosxy dA for R: 0 ≤ x ≤ f ’ 0 ≤ y ≤ 1.
R
Set up both iterated integrals and decide which would be easier to evaluate.
Limits of y
Limits of x
∙ττ∕2
Solution With respect to y first: x cosxy dy dx
∣o ∣o
Limits of x
Limits of y
'
With respect to x first: x cos xy dx dy
The second of these integrals (with respect to x first) requires integration by

parts, but integrating with respect to y first is quite straightforward:
x cos xy dA cos u du dx
Let u = xy; When y = 0, u = 0

then du = x dy. y = 1, u = x
rπ∕2 ∣χ rπ∕2 7τ∕2

[sinw] 10 dx = (sin A: - sinθ) dx = -co sx o = 1
∣o Jo
I AN INFORMAL ARGUMENT FOR FUBINI’S THEOREM
We can make Fubini’s theorem plausible with a geometric argument in the case
where f(x , y) ≥ 0 on R. If f(χ , y) dA is defined on a rectangle R: a ≤ x ≤ b,
J RJ
c ≤ y ≤ d, it represents the volume of the solid Abounded above by the surface

z = f(x , y) and below by the rectangle R. If A(y*k ) is the cross-sectional area
perpendicular to the y-axis at the point y k , then ^4(y*) ∆yj, represents the volume of
a “slab” that approximates the volume of part of the solid Δ, as shown in Figure
13.5.
Figure 13.5 Cross-sectional volume parallel to the xz-plane
By using a limit to “add up” all such approximating volumes, we obtain an

estimate of the volume of the entire solid Δ. Thus, if we let Fbe the exact volume
of the solid S, we have
[ ∫ ∕U ,> .) ⅛ l- r . ι ∣im λ∣ Λ (y ι ) ∆ y ι
R
The limit on the right is just the integral of 4(y) over the interval c ≤ y ≤ d,
where A(y) is the area of a cross section with fixed y. In Chapter 4, we found that
the area A(y) can be computed by the integral
∕∙Z>
A(y) = f(x , y) dx Integration with respect to x (y a constant)
Ja
We can now make this substitution for A(y) to obtain:
∫ ∫∕(¾ y )< 4 4 -r- 1^mιl ^ ^ 0 θ ∆ ¾
∠l(y) dy
Substitution
A(y)
The fact that
rb rd
f(x , y) dA = I f(x, y) dy dx
Ja Jc
can be justified in a similar fashion (you are asked to do this in Problem 42). Thus,
we have
•d Γ rb -1 r f rb
f(x , y) dx dy = f(x, y) dA = f(x , T) dy dx
Ja JpJ Ja
PROBLEM SET 1 3 .1
O Use an appropriate volume formula to evaluate the double
integral given in Problems 1-6.
Use successive partial integration to compute the double inte (0, 0). Assume that the population density r miles from
grals in Problems 7 - 14 over the specified rectangle. the city center is 12e- 0 ∙07r thousand people per square
7. J J x 2y dA; R: 1 ≤ x ≤ 2 ;0 ≤ y ≤ 1 mile. Express the total population of the region of the city
as a double integral.
31. Suppose R is a rectangular region within the boundary of
8. J J( *x + ~y') dA; -R: 2 ≤ x ≤ 3; —1 ≤ y ≤ 1 a certain county, and let f i x , y) denote the housing
density (in homes per square mile) at the point (x, y). If
the property tax on each home in the county is $1,400 per
9. J J 2xe y dA; R: - 1 ≤ x ≤ 0; 0 ≤ y ≤ In 2 year, express the total property tax collected in the region
of the county as a double integral.
10. J j^ x 2 ex y dA; R; 0 ≤ x ≤ 1; 0 ≤ y ≤ 1 32. Compute j J x V 1 - x 2 e i y dA, where R is the rectangle
J R J
ιi∙∫ ∫ ⅜ T j⅛ o ≤ x ≤ n ι ≤ ,≤ 3 0 ≤ x ≤ l ,0 ≤ y ≤ 2 .
ln V y
12.
R
J j y V l —y 2 dA; R; 1 ≤ x ≤ 5; 0 ≤ y ≤ 1 κ ∏
------- ■ dA, where R is the rectangle
1 ≤ x ≤ 4, 1 ≤ y ≤ e.
34. Compute J j^ ^,∙ dA (correct to the nearest
13. J J sin(x + y) dA; R: 0 ≤ x ≤ 0 ≤ y ≤
J RJ
hundredth), where R is the rectangle 1 ≤ x ≤ 3,
14. J J x sin ∙∖y dA; R: 0 ≤ x ≤ τr, 0 ≤ y ≤ 1 1 ≤ T ≤ 2 ∙ ∣- ∣∙
35. Explain why (4 - x 2 - y 2) dA ≥ 2 where R is the
Find the volume o f the solid bounded below by the rectangle R

rectangular domain in the plane given by 0 ≤ x ≤ 1,
in the xy-plane and above by the graph o f z = f(x , y) in
0 ≤ y ≤ 1.
Problems 15-25. Assume that a, b, and c are positive constants.
15. f ( x , y) = 2x + 3y; R; 0 ≤ x ≤ 1; 0 ≤ y ≤ 2 ≡≡≡≡
16. f i x , y) = 5x + 2y; R; 0 ≤ x ≤ 1; 0 ≤ y ≤ 2
17. f i x , y) = ax + by; R; 0 ≤ x ≤ a; 0 ≤ y ≤ b
18. ∕( x , y) = axy; R; 0 ≤ x ≤ b; 0 ≤ y ≤ c 36. Approximate the volume of the solid lying between
19. f i x , y) = V xy; R; 0 ≤ x ≤ 1; 0 ≤ y ≤ 4 the surface
20. f i x , y) = V x y ; R; 0 ≤ x ≤ a; 0 ≤ y ≤ b
∕( x , y) = 4 - x 2 - y 2
21. f i x , y) = xe y ; R; 0 ≤ x ≤ 2; 0 ≤ y ≤ In 2
22. f i x , y) = xe x y ; R; 0 ≤ x ≤ 1; 0 ≤ y ≤ In 3 and the square region R given by 0 ≤ x ≤ 1,
0 ≤ y ≤ 1. Use a partition made up o f squares
23. f i x , y) = lx + y) 5; R: 0 ≤ x ≤ 1; 0 ≤ y ≤ 1
whose edges have length 0.25.
24. f i x , y) = V x + y; R: 0 ≤ x ≤ 1; 0 ≤ y ≤ 1
37. Approximate the volume of the solid lying between
25. f i x , y) = (x + y) n ; n Ψ - 1 and n Ψ - 2 ; R; 0 ≤ x ≤ 1;
the surface
0 ≤ y ≤ 1
θ 26. ■ What Does This Say? Discuss the definition of dou f i x , y) = x 2 + y 2 + 1
ble integral.
and the square region R given by 0 ≤ x ≤ 1,
27. ■ What Does This Say? Describe the implementation 0 ≤ y ≤ 1. Use a partition made up of squares
of Fubini’s theorem. whose edges have length 0.25.
28. Suppose R is a region within the boundary o f a certain
national forest that contains 600,000 trees per square
mile. Express the number of trees per square mile in the 38. Evaluate j j (3 - V x 2 + y 2) dA, where R is the disk
forest as a double integral. Assume that x and y are J
R
measured in miles.
x 2 + y 2 ≤ 9 in the xy-plane. Hint; Relate the integral to
29. Suppose mass is distributed on a rectangular region R in
the volume of a simple solid.
the xy-plane so that the density (mass per unit area) at the 39. Evaluate J j (1 - V x 2 + z 2) dA, where R is the disk
point (x, y) is f i x , y). Express the total mass as a double
integral.
30. Suppose R is the rectangular region within the boundary x 2 + z 2 ≤ 1 in the xz-plane. Hint; Relate the integral to
of a certain city, and let the city center be at the origin the volume of a simple solid.
40. Evaluate J J (4 - V y 2 + z 2) dA, where R is the disk 42. Let f be a continuous function defined on the rectangle R:
a ≤ x ≤ b , c ≤ y ≤ d . Use a geometric argument to
show that
y 2 + z 2 ≤ 16 in the yz-plane. Hint: Relate the integral
to the volume of a simple solid.
@ 41. Let f be a function with continuous second partial deriv J J /(A v) dA = J £ f(x , y) dy dx
atives on a rectangular domain R with vertices (x p y i ),
(x2 , y 2), (x 1, y 2), and (x2 , y 1), where x, < x 2 and y 1 < y γ
Use the fundamental theorem of calculus to show that
Hint: Modify the argument given in the text by taking
/ / JyL dA = ^ x ' , ∙v ') - ∙y ι) + ∙^∙x 2, ⅛) - ≠(χ ι>
cross-sectional areas perpendicular to the x-axis.
1 3 .2 DOUBLE INTEGRATION OVER NONRECTANGULAR REGIONS
IN THIS SECTION Nonrectangular regions, more on areas and volumes,

reversing the order of integration in a double integral, properties of double
integrals
We show how double integrals can be evaluated over regions that are not
rectangles. We conclude by examining some general properties of double
integrals.
■ NONRECTANGULAR REGIONS
A region is said to be bounded if it is contained in a rectangular region R, as shown

in Figure 13.6.
Suppose D is a bounded region in the xy-plane that is not a rectangle. Define a
function F that coincides w ith/on D and is 0 elsewhere in R. If Fis integrable over
R, then f is said to be integrable over D, and the double integral of f over D is
Figure 13.6 The region D is defined as
bounded by a rectangle R. f f f(x , y )d A = f f F(x, T) dA
J DJ J RJ
In the last section, you saw how the double integral f(x , y) dA could be
J RJ
evaluated by successive partial integration when J? is a rectangle. A modification

of this procedure allows you to evaluate the double integral in the important case
where R is not a rectangle, but is still bounded by vertical (or horizontal) lines on
two opposing sides.
Type I Region (Vertical Strip)

Sec. 13.2 Double Integration Over Nonrectangular Regions 83 1
Type II Region (Horizontal A type II region contains points (x, y) such that
Strip) for each fixed y between constants c and d, x
varies from A1(y) to ∕z,(y), where h l and Λ, are
continuous functions. Think of a horizontal
strip.
Type II region: For a fixed y between c and d, x varies

from A1(y) to ⅛2(y).
By applying Fubini’s theorem for rectangular regions, we can derive the

following theorem, which shows how to evaluate a double integral over a type I or
type II region.
THEOREM 13.2 Fubini’s theorem for nonrectangular regions
If D is a type I region, then
g2W
T Y P E I: (vertical strip): ∕( x , y) dy dx
x-fixed, y varies (form dy dx)
Π '∣W
whenever both integrals exist. Similarly, for a type II region,
rd rh 2(y)
T Y P E II: (horizontal strip): f( x , y) dA = ' ∕( x , y) dx dy
■k Jħ,(y)
y-fixed, x varies (form dx dy)
Proof: The proof is omitted.
When using Fubini’s theorem for nonrectangular regions, it helps to sketch the
region of integration D and to find equations for all boundary curves of D. Such a
sketch often provides the information needed to determine whether D is a type I
or type II region (or neither) and to set up the limits of integration of an iterated
integral.
EXAM PLE 1 Double integral over a triangular region
Let T be the triangular region enclosed by the lines y = 0, y = 2x, and x = 1.

Evaluate the double integral
J J (x + y) dA
τ
using an iterated integral with:
a. y-integration first b. x-integration first
Figure 13.7 For each fixed x Solution a. To set up the limits of integration in the iterated integral, we draw
(0 ≤ x ≤ 1), y varies from y = 0 the graph as shown in Figure 13.7 and note that for fixed x, the variable y
to y = 2x. varies from y = 0 (the x-axis) to the line y = 2x. These are the limits of
integration for the inner integral (with respect to y first). The other limits of
integration are the numerical limits of integration for x — that is, x varies
between x = 0 and x = 1.
y= 2 x
y=0 u λ
C [x(2x) + ∣(2x) 2 - (x(0) + 1(0)2)] d x = [ 4x 2 dx = ⅛x 3 l =0 = ⅛

Jo l J JO J I J
b. Reversing the order of integration, we see from Figure 13.8 that for each y,
the variable x varies (left to right) from the line x = y/2 to the vertical line
x = 1. The outer limits of integration are for y as y varies from y = 0 to
T = 2.
Figure 13.8 For each fixed

y(0 ≤ y ≤ 2), x varies from y/2
to 1.
■ MORE ON AREAS AND VOLUMES
Even though we can find the area between curves with single integrals, it is often
easier to compute area using a double integral. Iff ( x , y) ≥ 0 over a region D in the
xy-plane, then ∫ f f ( x , y) dA gives the volume of the solid bounded above by the
D
surface z = f( x , y) and below by the region of D . In the special case where
f ( x , y) = i, the integral gives the area of D.
The Double Integral as Area The area of the region D in the xy-plane is given by
and Volume
If/is continuous a n d ∕(x , y) ≥ 0 on the region D , the volume of the solid under
the surface z = f( x , y) above the region D is given by
E X A M P L E 2 Area of a region in the xy-plane using a double integral
Find the area of the region D between y = cos x and y = sin x over the interval
0 ≤ x ≤ f using
a. a single integral b. a double integral
Solution a. The graph is shown in Figure 13.9.

r, 7τ∕4 I ττ∕4
Figure 13.9 The area of the region
between y = cos x and y = sin x J0
(cosx - sinx) dx = [sinx + cosx] 1 0
= V2 - 1
Sec. 13.2 Double Integration Over Nonrectangular Regions 833
r~∕4 ∣y=cosx
b. 1 dy dx = I [y]∣ dx
rπ∕4 —
= [cosx - sinx] dx = v2 - 1
Jo
The area is V z2 - 1 ≈ 0.41 square units.
In comparing the single and double integral solutions for area in Example 2,
you might ask, “ Why bother with the double integral, because it reduced to the
single integral case after one step?” The answer is that the double integral will do
most of the work in setting up the proper integral. It is often easier to begin with
the double integral
and then let the evaluation lead to the proper form.
EXAM PLE 3 Volume using a double integral
Find the volume of the solid bounded above by the plane z = y and below in
the xy-planc by the part of the disk x 2 + y 2 ≤ 1 with x ≥ 0, y ≥ 0.
Solution The three-dimensional solid is shown in the margin, but for your work
you need only be concerned with the projection in the xv-plane. which is
Graph of z = y showing the disk

in the xy plane
Figure 13.10 The quarter disk x 2 + y 1 ≤ 1, x ≥ 0, y ≥ 0
We can regard D as either a type I or type II region, and because we worked

with a type I (vertical) region in Example 2, we shall use a type II (horizontal)
region for this example. Accordingly, note that for each fixed number y
between 0 and 1, x varies between x = 0 on the left and x = V l —y 2 on the
right. Thus,
V = f f ∕(x, y) dA = [ f y dx dy f(x, y) = y is given; dA = dx dy.

JJ ' Jo Jo
Let u = 1 —y 2 to integrate.
The volume is ∣ cubic unit.

■ REVERSING THE ORDER OF INTEGRATION IN A DOUBLE

INTEGRAL
It is often useful to be able to reverse the order of integration in a given iterated

integral. This procedure is illustrated in the next two examples.
EXAMPLE 4 Reversing the order of integration
Reverse the order of integration in the iterated integral
Solution Begin by drawing the region D by looking at the limits of integration for
both x and y in the double integral. For this example, we see that the y-
integration comes first, so D is a type I region. The inner limits are
x =0 x= 2
Fixedx
y = ex (top curve) and y = I (bottom curve)
a. D as a type I region; y These are shown in Figure 13.1 la. Next, draw the appropriate limits of
varies from 1 to e x integration for x. Note: These limits should be constants:
D as a type I region; y varies from
1 to ex .
x = 0 (left point) and x = 2 (right point)
These vertical lines are also drawn in Figure 13.1 la.
To reverse the order of integration, we need to regard D as a type II region

(Figure 13.1 lb). Note that the region varies from y = 1 toy = c2 (correspond
ing to where y = ex intersects x = 0 and x = 2, respectively). For each fixed y
(horizontal strip) between 1 and e2, the region extends from the curve x = In y
(that is, y = ex ) on the left to the line x = 2 on the right. Thus, reversing the
order of integration, we find that the given integral becomes
The two different ways of representing the integral are illustrated together for
b. D as a type II region; x
varies from In y to 2 easy reference:
D as a type II region; x varies Type I: x fixed (vertical strip) Type II: y fixed (horizontal strip)
from In y to 2. y-integration first; varies from x-integration first; varies from
Figure 13.11 Type I and Type 2 y= l to y = e v. x = In y to x = 2.
regions ∕-e2 ∣
∙2 ’
dx f(x , y) dx dy
Jl J ln y
↑ ↑
Constant limits for x; varies from 0 to 2. Constant limits for y; varies from 1 to e 2.
EXAMPLE 5 Reversing the order that requires a sum of integrals

Reverse the order of integration in the integral
∕∙2 rx
f(x, y) dy dx
∣
-1 Jx 2 -2
Sec. 13.2 Double Integration Over Nonrectangular Regions 83 5
Solution This is a type I form (because the y-integration comes first). To reverse
the order of integration we begin by drawing the region D, as shown in Figure
13.12a. We draw the bottom and top curves:
y = x 2 - 2 (bottom) and y = x (top)
Then we draw the left and right boundaries:
x = -1 and x = 2
To change this to a type II form, we shift from a vertical strip (x fixed) to a
horizontal strip (y fixed), as shown in Figure 13.12b. We see that for y between
- 2 and -1 the horizontal line is bounded by the left and right branches of the
parabolic arch, whereas for y between -1 and 2 it is bounded on the left by the
line x = y and on the right by the parabola. This means that the integral with
the order of integration reversed must be expressed as the sum of two
integrals.
∕∙2 rx
f(x, y) dy dx For a fix e d x between - 1 and 2,
∣-1 Jx2-2 y varies between y = x 2 - 2 and y = x.
r-l r∙Vy+2 r2 ∕∙√μ+2
; i _ ∕( x , y) dx dy + i ; f(x, y) dx dy
J-2 J-V'f+2 J-ι Jy
∖ / V - √
Part between left and right Part between line and
branches of parabola (red part) parabola (blue part)
For a fix e d y between For a fix e d y between

b. Type II description
- 2 and - 1 , x varies - 1 and 2, x varies
Figure 13.12 between x = - V y + 2 between x = y and
and x = Vy + 2 x = ∖,y + 2
A given iterated integral can often be evaluated in either order, although it

may happen that one order is more convenient than the other. For instance, the
integral in Example 5 is easier to handle by performing y-integration first, because
the reverse order involves two integrations instead of just one. However, there are
iterated integrals that can be evaluated only after reversing the order of integra
tion. Such an integral is featured in the following example.
EXAMPLE 6 Evaluating a double integral by reversing the order
Evaluate ey2 dy dx
Solution We cannot evaluate the integral in the given order (y-integration first)
because the integrand ey has no elementary antiderivative. We shall evaluate
the integral by reversing the order of integration. The region of integration is
sketched in Figure 13.13a. Note that for any fixed x between 0 and 1, y varies
from x to 1.
To reverse the order of integration, observe that for each fixed y between 0
and 1, x varies from 0 to y, as shown in Figure 13.13b.
f1 fl y2 Γ1 ΓV y 2 2lx ~y i C 2 7
e dy dx = e dx dy = ∖ xe y I 0 dy = ye y dy Let u = y 1.
l
b. ^-integration first ∣o Jx ' Jo Io Jo ' Jo
Figure 13.13 The region of inte = [⅛e^2 ] ∣
θ = ⅛(e1 - e 0 ) = ⅛(e - i)
gration for Example 6
■ PROPERTIES OF DOUBLE INTEGRALS
Double integrals have many of the same properties as single integrals. Three of
these properties are contained in the following theorem.
THEOREM 13.3 Properties of double integrals
Assume that all the given integrals exist.
Linearity rule: For constants a and b,
f f [af(x, y)
D
Dominance rule: If f(x , y) ≥ g(x, >,) throughout a region D, then
Subdivision rule: If the region of integration D can be subdivided into two

subregions D i and Z>2, then
f f f ( χ , T) dA
J DJ
f f f(
j o↑
χ , y)d A +
j
[∣A ff(x, y) dA
Proof: To prove this, you can first show that each is true for rectangular
regions of integration. Then, the general case is proved by applying the definition
of a double integral, Fubini’s theorem, and various theorems for single integrals. ■
PROBLEM SET 1 3 .2
O 1. ■ What Does This Say? Describe the process for find ∙2 r sin x "τr∕2 rsinx
ing volume using a double integral.

2. ■ What Does This Say? Describe the process of revers
J o Jo
y cos x dy dx
J o Jo
ey cos, x dy dx
ing the order of integration. Evaluate the double integral given in Problems 15-26 for the
specified region o f integration D.
Sketch the region o f integration in Problems 3-14, and com
pute the double integral (either in the order o f integration given
15. ; D is the triangle with vertices (0, 0),
or with the order reversed).
4.γ D
3.
Jo Jo
xy dy dx
∏
6∙ J J
dy dx
χ y dy dx
16.
(0, 1),(1, 1).
; D is the triangle with vertices (0, 0),

5. dy dx D
Jo J-χ2
(1,0), (0, 2).
-2√3 fV 1 2 -y 2 rl ι∙3y+2
J0
Γ3 Γ4^x
!
J V 2 ∕4
dx dy
ι
8.
J —2 J y 2 +4y
dx dy
°. Jθ' ∣, (x + y 2) dy dx
17.
D
y = Vx.
is the region bounded by y = x 3 and
9. (x + ,r) dy dx
Jo Ji
r1 rx
18. is the region bounded by y = x 2 and
11. ∣ J (x 2 + 2y 2) dy dx 12. (3x + 2y) dy dx D
J-ι J -1 y = 2x.
Sec. 13.2 Double Integration Over Nonrectangular Regions 837
r2 r 6 -x 16∕x
19.
D
; D is the region bounded by y = Vx, y = 2 - x, 41.
J —3 Jχ 2
fix , y) dy dx
∏ fix , y) dy dx
and y = 0. θ Express the area of the region D bounded by the curve given in
20. ; D is the region bounded by y = 4 —x 2, Problems 43-44 as a double integral in two different ways (with
different orders of integration). Evaluate one of these integrals
D
to find the volume.
y = 3x, and x = 0. 43. Let D denote the region in the first quadrant of the
21. is the region in the first quadrant bounded xy-plane that is bounded by the curves y = ~j and
D y = 5 - x 2. x
by y = 4 - x 2, y = 3x, and y = 0. 44. Let D denote the region bounded by the ellipse
^ + ξ = l
22. ; D is the triangle with vertices (-1 , 0), a 2 b2
D
45. Find the volume under the surface z = x + y + 2 above
(1, 0), and (0, 1).
the region D bounded by the curves y = x 2 and y = 2.
23. is the region bounded by x 2y = 1, y = x. 46. Find the volume under the plane z = 4x above the region
D D given by y = x 2, y = 0, and x = 1.
x = 2, and y = 0. 47. Evaluate
24. I I ® *s t 'l e triangle bounded by x = 2y, y = - x , Γ1 Cy 2, , Γ2 f 2 ~y 2

(x + y 2) dx dy + (x + y 2) dx dy
D Io Jo Ji Jo
and y = 2.
by reversing the order of integration.
25. ; D is the region in the first quadrant 48. Let fix , y) be a continuous function for all x and y.
D Reverse the order of integration in
bounded by y = x 2 and y = x.
x3 ri ri
J ∫ y dA∖ D is the region bounded by y = In x, y = 0, and
26.
DJ
∏ fix , y) dy dx + fix , y) dy dx
49. Evaluate J J xy dA, where D is the triangular region in

Sketch the region of integration in Problems 27-34, and then J DJ
compute the integral in two ways: (a) with the given order of the xy-plane with vertices (0, 0), (1, 0), and (4, 1).
integration and (b) with the order of integration reversed.
r4 r 4 -x r4 r4x
27. I I xy dy dx 28. I I dy dx
r4 rVx
29. Jθ1 J 2" ey ~x dy dx 30. 1 1 3x 5 dy dx
re r∖n x
31. J J dy dx 32. xy dy dx
∕∙2√3 rV 1 6 -y 2 rl r3y+2 50. Compute J J (x2 - xy - 1) dA, where D is the triangular

33. dx dy 34. dx dy
Jo J y 2∕6 J -2 J y 2+4y J DJ
region bounded by the lines x —2y + 2 = 0,

Sketch the region o f integration in Problems 35-42, and write x + 3y - 3 = 0, and y = 0.
an equivalent integral with the order of integration reversed. Do
not evaluate.
îy Vx
∏ yy
fix , y) dx dy
Πre 2 r2
f(x , y) dy dx
∏ /2
y A -y
fix , y) dx dy 38.
Jl Jinx
’l r 2 -x
f(x , y) dy dx
Π /3
fix , y) dx dy
Jo Jx
f(x , y) dy dx
51. Find the volume under the surface z = x + 1y + 4 above

f ( x , y) dA ≤ M if dA = 1
the region D bounded by the lines y = 2x, y = 3 - x , and
y = 0.
52. Find the volume under the plane 3x + y - z = 0 above 55. Use the result of Problem 54 to estimate the value o f the
the elliptic region bounded by 4x 2 + 9y2 ≤ 36, with double integral
x ≥ 0 and y ≥ 0.
53. Find the volume under the surface z = x 2 + y 2 above the
square region bounded by ∣x∣ ≤ 1 and ∣y∣ ≤ 1.
54. Show that if f ( x , y) is continuous on a region D and where D is the triangle with vertices ( -1 , 0), (2, 0), and
m ≤ f ( x , y) ≤ M for all points (x, y) in D, then (0, 1).
1 3 .3 DOUBLE INTEGRALS IN POLAR COORDINATES
IN THIS SECTION Change of variables to polar form, improper double

integrals in polar coordinates
In general, changing variables in a double integral is more complicated than in a
single integral. In this section, we focus attention on using polar coordinates in
a double integral, and in Section 13.8, we examine changing variables from a
more general standpoint.
■ CHANGE OF VARIABLES TO POLAR FORM
Polar coordinates are used in double integrals primarily when the integrand or the
region of integration (or both) have relatively simple polar descriptions. As a
preview of the ideas we plan to explore, let us examine the double integral
∫J 2
(x j + y + 1) dA
Graph o f∕( x , y) = x 2 + y 2 + 1 where D is the region (disk) bounded by the circle x 2 + y 2 = 4.

Interpreting D as a type I (vertical strip), we see that for each fixed x
between - 2 and 2, y varies from the lower boundary semicircle with equation
y = - V 4 - x 2 to the upper semicircle y = V4 - x 2, as shown in Figure 13.14a.
a. Type I description: b. Polar description:

For fixed x between - 2 and 2 For fixed θ between 0 and 2π,
y varies from y = - v 4 - x 2 r varies from 0 to 2.
to y = "V4 - x 2
Figure 13.14 Two interpretations of a region D

Sec. 13.3 Double Integrals in Polar Coordinates 839
Using the type I description, we have

r2 fvF P
Uυ
(x 2 + y 2 + 1) dA =
J-2 J~V4-x^
__ (x2 + y 2 + 1) dy dx
The iterated integral on the right is clearly difficult to evaluate, but both the
integrand and the domain of integration can be represented quite simply in terms
of polar coordinates. Specifically, using the polar conversion formulas
x = r cos θ y = r sin θ r = Vx 2 + y 2 tan θ = -
we find that the integrand f(x , j j = x 2 + y 2 + 1 can be rewritten

f(r cos θ, r sin 0) = (r cos θ)2 + (r sin 0)2 + 1 = r2 + 1
and the region of integration D is just the interior of the circle r = 2. Thus, D can
be described as the set of all points (r, 0) so that for each fixed angle 0 between 0
and 2τr, r varies from the origin (r = 0) to the circle r = 2, as shown in Figure
13.14b.
But how is the differential of integration dA changed in polar coordinates?
Can we simply substitute “dr dθ, ' for dA and perform the integration with respect
to r and 0? The answer is no, and the correct formula for expressing a given double
integral in polar form is given in the following theorem.
THEOREM 13.4 Double integral in polar coordinates

If f is continuous in the polar region 1) described by a ≤ r ≤ b (a, b ≥ 0),
α ≤ θ ≤ β (Q ≤ β — a < 2π), then
■ What this says: The procedure for changing from a Cartesian integral
f ∣f(x ,y )d A
R
into a polar integral is done in two steps. First, substitute x = r cos 0 and
y = r sin 0, dx dy = r dr dθ into the Cartesian integral. Then supply the polar
limits for the boundary, where D denotes the region of integration in polar
coordinates. Thus,
J J f(x , y) dA = j J f(r cos 0, r sin 0) r dr dθ
Proof: A region described b y α ≤ r ≤ ⅛ , α ≤ 0 ≤ ∕3 i s called a polar rectan-
b. A typical polar rectangle

rectangle polar coordinates in the partition
To prove the theorem, we subdivide the region of integration into polar

rectangles. Pick an arbitrary polar-form point (r*k , 0^) in each polar rectangle in the
partition and then take the limit of an appropriate Riemann sum
∑ ι , < < p ∆ ∕l 1
where A A k is the area of the kth polar rectangle.

To find the area of a typical polar rectangle, let (7, θ) be the center of the
polar rectangle— that is, the point midway between the arcs and rays that form the
polar rectangle, as shown in Figure 13.15b. If the circular arcs that bound the polar
rectangle are ∆r apart, then the arcs are given by
r 1 = r - ∣ ∆r and r2 = r + ∣ ∆r
It can be shown that a circular section of radius r and central angle θ has area ⅛r2 θ
(see the Student Mathematics Handbook for details). Thus, a typical polar
rectangle has area
Δ ∕( = ∣J(r + ∣∆ r) 2 - ∣(r - ∣∆ r) 2]∆0 = r ∆r Δ0
Radius o f outside arc Radius o f inside arc
Finally, we compute the given double integral in polar form by taking the limit
[ f /(r, 0) dA = lim Y ∕( r ,, 0,) A A = lim Y f(r, θ)r Ar Aθ

JD J ∣f∣→° ⅛ ι * k W →0fc=ι
∕( r , 0)r dr dθ
PREVIEW: It can be shown that under reasonable conditions, the change of

variable x = x(u, v), y = y(u, v) transforms the integral ∫ ∫ ∕( x , y) dA into
∫ ∫∕(w , v) ∣ J(μ , v) ∣ du dv, where
∂x ∂x
∂u ∂v
J(u, v) =
∂y ∂y
∂u ∂v
This determinant is known as the Jacobian of the transformation. In the case
of polar coordinates, we have x = r cos θ and y = r sin θ, so that the Jacobian
is
⅛('∙ cos 0) ⅛(r cos 0)

J(r, 0) =
⅛(f sin 0) ^ ( r sin 0)
cos 0 —r sin 0
= r cos2 0 + r sin 2 0 = r
sin 0 r cos 0
This yields the result of Theorem 13.4.
r dr dθ = f(r cos 0, r sin θ)r dr dθ
We discuss this more completely in Section 13.8.

Sec. 13.3 Double Integrals in Polar Coordinates 841
If you carefully compare the result in the preview box with the result of Theorem
13.4, you will see that they are not exactly the same. The region D in Theorem
13.4 already is in polar coordinates. The more common situation is that we are
given f and a region R in rectangular coordinates that we need to change to polar
coordinates.
We now present the example we promised in the introduction.
EXAM PLE 1 Double integral in polar form
Evaluate J J (x 2 + y 2 + 1) dA, where D is the region inside the circle
x 2 + y 2 = 4.
Solution In this example, the region D is given in rectangular form. We will
describe this as a polar region D *. Earlier, we observed that in D *, for each
fixed angle θ between 0 and 2π, r varies from r = 0 (the pole) to r = 2 (the
circle). Thus,
r2ττ r2
Π
Γ~π C~
υ
,
(x 2 + y 2 + 1) dA =
f( x* >T) 0
c -π Γr4 r
° ∕( r cos
2^∣∣2
(r2 + 1) r dr dθ
0> r s >n θ) dA
f2π
= (r3 + r) dr dθ = - j + 4- L dθ = 6 dθ = 60∣-oιr = 12π
Jo Jo Jo l 4 2 -*1 Jo
EXAM PLE 2 Double integral in polar form
Evaluate j j x dA, where D is the region bounded above by the line y = x and
below by the circle x 2 + y 2 - 2y = 0.

Solution The circle x 2 + y 2 - 2y = 0 and the line y = x are shown in Figure
13.16. We rewrite these equations in polar form:
x 2 + y 2 - 2y = (r cos θ)2 + (r sin θ)2 —2(r sin 0) = 0
r2 —2r sin θ = 0 Because cos2 θ + sin2 θ = 1
r(r - 2 sin θ) = 0
r = 0 or r —2 sin θ Reject r = 0.
The polar-form equation for the circle is r = 2 sin θ, and the line y = x in polar
Figure 13.16 The region D form is
r sin θ = r cos θ
This is true only when tan θ = 1 (divide both sides by r cos θ), which implies
θ = 2f . Thus, the region D may be described by saying that for each fixed angle
θ between 0 and 2ρ r varies from 0 (the pole) to 2 sin θ (the circle). Thus,
EXAM PLE 3 Volume in polar form
Use a polar double integral to show that a sphere of radius a has volume ∣mz 3.
Solution We compute the required volume by doubling the volume of the solid
hemisphere x 2 + y 2 + z 2 ≤ a 2, with z > 0. This hemisphere may be regarded
as a solid bounded below by the circular disk x 2 + y 2 ≤ a 2 and above by the
spherical surface.
We need to change the equation of the hemisphere to polar form:
In rectangular form: z = Vtz2 - x 2 - y 2
The top hemisphere is bounded In polar form: z = V α 2 - r2 Because r2 = x 2 + y 2

above by the sphere and below by
Describing the disk D in polar terms, we see that for each fixed θ between 0
the disk D.
and 2ιτ,r varies from the origin to the circle x 2 + y 2 = a 2 , which has the polar
equation r = a, as shown in Figure 13.17. Thus, the volume is given by the
integral
V= 2 Rectangular form
Let u = a 2 - r 2∙,
du = —2r dr.
3 [(α2 - α2)3z2 - (α2 - 0)3z2] dθ

i Jo
Figure 13.17 The disk D: For θ f2π j I 2π Λ
between 0 and 2π, r varies from 0 a i dθ = ½a3 θ∖0 = ⅞τra3
to a. Io 3 1 3
You will need to be familiar with the graphs of many polar-form curves. These
can be found in Table 9.1 on page 583 and also in the Student Mathematics
Handbook.
Example 3 has constant limits for the distance r. However, the extension to
variable limits is quite natural.
Area as a Double Integral If D is the polar region described by r1(0) ≤ r ≤ r2(θ), and a ≤ θ ≤ β, its area
in Polar Coordinates
This formula is illustrated in the following examples.
EXAM PLE 4 Area between two curves in polar form
Evaluate J J dA, where D is the region in the first quadrant that lies inside
the circle r = 3 cos θ and outside the cardioid r = 1 + cos θ.

Sec. 13.3 Double Integrals in Polar Coordinates 84 3
Solution Begin by sketching the given curves, as shown in Figure 13.18. Next,
find the points of intersection:
3 cos θ = 1 + cos θ
2 cos θ = 1
COS 0 = 7
We see that D is the region such that, for each fixed angle θ between 0 and ττ∕3,
r varies from 1 + cos θ (the cardioid) to 3 cos θ (the circle). This gives us the
limits of integration. Finally, we need to write the integrand in polar form:
1 = 1
x r cos θ
Rectangular form Polar form
Thus,
- dA = I f — L —, r dr dθ
x J J r cos °
r3cos θ
~co~s~λθ dr dθ Write —½ = sec 0; note that
1+cos θ
cos θ
sec θ is a constant with
respect to the variable r.
rπ ∕3
= [3 cos θ sec θ - (1 + cos 0)sec θ] dθ
Jo
rπ ∕3
= I (2 — sec θ) dθ Note'. 2 cos θ sec θ = 2 be-
J° cause cosine and secant are
reciprocal functions.
Iπ∕3
= [20 - ln∣sec 0 + tan 0∣] 0
= ⅛ - 1∏(2 + √3)
■ IMPROPER DOUBLE INTEGRALS IN POLAR COORDINATES
Improper double integrals play a useful role in probability theory and in certain
physical applications. We shall demonstrate the basic approach to this topic by
examining the special case of double integrals that are improper because the
region of integration is the (unbounded) first quadrant. A general treatment of
improper multiple integrals is outside the scope of this text.
Suppose the function f ( x , y) is continuous throughout the first quadrant Q;
then define the improper integral in terms of rectangular coordinates as follows;
Figure 13.19 Quarter circular do where S ιι is the square described by 0 ≤ x ≤ n, 0 ≤ y ≤ n. In terms of polar
main coordinates, let C denote the quarter circular region described by r ≤ n,
0 ≤ θ ≤ τr∕2, as shown in Figure 13.19. We use this quarter circular region to

formulate the following definition.
Improper Double Integral Let Q denote the first quadrant of the Cartesian plane, and let C denote the
in Polar Coordinates quarter circular region described by r ≤ n, 0 ≤ θ ≤ Then the improper
integral f ( x , y) dA is defined in polar coordinates as
JQJ
∏ r v ∕ 2 rn
f( r cos θ, r sin θ)r dr dθ = lim f( r cos θ, r sin θ)r dr dθ
,' →'x Jo Jo
If the limit in this definition exists and is equal to L , we say that the improper
integral converges to L . Otherwise, we say that the improper integral diverges.
E X A M P L E 5 Evaluating a convergent improper double integral
Evaluate e F 2 +r2) t∕q , where Q is the first quadrant of the xy-plane.
Solution To evaluate this integral (improper because Q is unbounded), we

convert to polar coordinates.
= lim ⅛(1 - 6>-"2)[0]∣ = lim ¾ l - e - 2) = 77

w→∞z, M→χ H∙ 4
Thus, the improper integral converges to
PROBLEM SET 1 3 .3
O Evaluate the double integral j J f ( x , y) dA in Problems 1-8, Use a double integral to fin d the area o f the shaded region in
J D J Problems 9-22.
and sketch the region o f integration D.
^ττ∕2 r2 sin θ ^τr rl + sin θ
J o Jo
∣ dr dθ
f π72 f 3 ?
o Jo
l
j
J
^π∕2 r2
dr dθ
3.
Jo Jl
(, π r4
re~r dr dθ
o Ji J
rττ∕2 r3
V 4 - r2 r dr dθ
5. I r2 sin2 θ dr dθ 6. r 2 cos2 θ dr dθ
Jo Jo Jo Jι
', 2τr r4 r, 2π r l-sin 0
J o
I 2r2 cos θ dr dθ
Jo J i
o Jo
cos θ dr dθ
Sec. 13.3 Double Integrals in Polar Coordinates 84 5
20. r = 1 and r = 1 + cos 0
22. r = 3 cos 0 and

r = 2 - cos 0
13. r = 4 cos 30 14. r = 5 sin 20
Use polar coordinates in Problems 23-28 to find J J fix, y) dA,

D
15. r = cos 20 16. r = 1 and r = 2 sin 0 Hint.
where D is the circular disk defined by x 2 + y 2 ≤ a 2 for
consider 0 < 0 < f and
constant a > 0.
< 0 < J separately.
23. fi x , y) = y 2 24. fix , y ) = x 2 + y 2
25. fj ivx , zy)7 = - 5-----⅛----- 5 26. f ix , y) = e~P2+y2'>
’ z7 2 ∣ v 2 i ,,2
27. f i x , y) = 28. f ix , y) = ln(α 2 +

a + fix 2 + y2
Use polar coordinates in Problems 29-34 to evaluate the given
double integral.
J J y dA, where D is the disk x 2 + y 2 ≤ 4
29.
DJ
J J (x 2 + y 2) dA, where D is the region bounded by the x-
30.
DJ
axis, the line y = x, and the circle x 2 + y 2 = 1

17. r = 1 and r = 2 sin 0 18. r = 1 and r = 1 + cos 0
J j" eχ2+y2 dA, where D is the region inside the circle
31.
JDJ
x2 + y2 = 9
32. + y 2 dA, where D is the region inside the circle
D
(x - 1)2 + y 2 = 1 in the first quadrant
33. ln(x 2 + y 2 +. 2) dA, where D is the region inside the
D
circle x 2 + y 2 = 4 in the first quadrant
34. j J sin (x 2 + y 2) dA, where D is the region in the first

J DJ
quadrant bounded by the circles x 2 + y 2 = 1 and x 2 + y 2
= 4 and the lines y = 0 and x = V3 y
35. Find the volume of the solid bounded by the paraboloid
z = x 2 + y 2 and the plane z = 4.
36. Find the volume of the solid bounded above by the cone
z = 6 V x 2 + y 2 and below by the circular region
x 2 + y 2 ≤ a 2 in the xy-plane, where a is a positive
constant. 49. T H IN K T A N K P ROBLEM TO evaluate the integral
θ 37. Use polar coordinates to evaluate ∫ $ x y dA, where D is J j r 2 dr dθ
D
the intersection o f the circular disks r ≤ 4 cos θ and where R is the region in the xy-plane bounded by
r ≤ 4 sin θ. Sketch the region of integration.
r = 2 cos θ, we get
38. Let D be the region formed by intersecting the regions (in
Cπ r 2 cos ft f π r 3∣
2cos0
the xy-plane) described by y ≤ x , y ≥ 0, and x ≤ 1. ∣ I r2 dr dθ = y 0 dθ
a. Express ∫ ∫ dA as an iterated integral in Cartesian Jo Jo Jo -*
D
= ∣ R o s W fl^ fs in fl-⅛ ⅛ = 0
coordinates.
3 Jo 3L 3 J
b. Express ∫ J dA in terms o f polar coordinates.
D Alternatively, we can set up the integral as
39. Example 5 of Section 9.3 used a single integral to find the
area of the region common to the circles r = a cos θ and
r = a sin θ. Rework this example using double integrals.
40. Example 6 of Section 9.3 used a single integral to find the - 8 Γ ' 2 cos3 t) df) = -l∖in fl - ⅛ ^ l M =
area between the circle r = 5 cos θ and the limaςon oI cos tf cιu -» sin u ∩ ι _/•> r∖
3J-W2 3L 3 J--β 9
r = 2 + cos θ. Use double integrals to find the same area.
Both of these answers cannot be correct. Which procedure
41. Use a double integral in polar coordinates to find the
(if either) is correct and why?
volume of the solid that lies inside the sphere
x 2 + y 2 + z 2 = 25 and outside the cylinder x 2 + y 2 = 9. θ 50. In Example 5, we used polar coordinates to show that
Evaluate the improper integrals in Problems 42-46 over Q, the ∣I e - u 2 +r2) dA

first quadrant in the xy-plane.
Q
42. (x 2 + y2 + 1) 3/2 dA where Q is the first quadrant of the xy-plane.

a. Explain why
∣J
Q
e -2(x2 + y2 ) d A e (jc2 +y2) dy =

dx
b. Show that
47. Use polar coordinates to evaluate the double integral

f f (x 2 + y 2) dA over the shaded region. c. Explain why
D
51. Evaluate
{Hint: Use Problem 50.)

52. For constant a where 0 ≤ a ≤ R , the plane z = R - a
cuts off a “ cap” from the sphere
x 2 + y 2 + z2 = R 2
48. Use polar coordinates to evaluate the double integral Use a double integral in polar coordinates to find the
f f (x 2 + y 2) dA over the shaded region. volume of the cap.
D
Sec. 13.4 Surface Area 847
1 3 . 4 SU R F A C E A R E A
IN T H IS S E C T IO N Definition of surface area, surface area projections,

area of a surface defined parametrically
In Chapter 6, we found that if /h a s a continuous derivative on the closed,
bounded interval [a, b∖, then the length L of the portion of the graph of y = ∕(x )
between the points where x = a and x = b may be defined by
L = p V l + [∕'(x )] 2 dx
Ja
The goal of this section is to study an analogous formula for the surface area of
the graph of a differentiable function of two variables.
■ D E F IN IT IO N O F S U R F A C E A R E A
Consider a surface defined by z = ∕(x , >-,) defined over a region R of the xy-plane.
We will find the area of this surface by considering a Riemann sum. We begin by
enclosing the region R in a rectangle, which is then partitioned by a grid with lines
parallel to the coordinate axes, as shown in Figure 13.20a. This creates a number
of cells, and we let R v R v . . . , R n denote those that lie entirely within R.
z Plane T is tangent to
surface z = f(x. vi at A .
Figure 13.20 Surface area
b. The surface area above R is

approximated by the area of the
parallelogram on the tangent plane.
For m = 1, 2, . . . , n, let P,n(x m , yζi) be a corner of the rectangle R m , and let Tm
be the tangent plane above Pm on the surface of z = ∕(x , y). Finally, let ΔS denote
the area of the “patch” of surface that lies directly above R m .
The rectangle R m projects onto a parallelogram ABDC in the tangent plane Tm ,
and if R is “small,” we would expect the area of this parallelogram to
approximate closely the element of surface area Δ,Sm (see Figure 13.20b).
If ∆xm and ∆ym are the lengths of the sides of the rectangle R , the
approximating parallelogram will have sides determined by the vectors
AB = ∆xmi + [ ∕ χ (x*m, ∕ m)∆ x jk and AC = Δ y J + [fy (x*m , y*m )∆ym ]k
In Chapter 10, we showed that such a parallelogram with sides AB and AC has
area
Figure 13.21 An element of sur ∣
∣
AB × AC ∣
∣
face area has area ∣
∣
AB × AC ∣∣
. This is shown in Figure 13.21.
If K is the area of the approximating parallelogram, we have

Xm = ∣
∣
A B × AC ∣
∣
First, we calculate the cross product:
j k
AB ×AC = ∆x m 0 ∕X > y ^ χm
0 ^ym χ
fy( ^ y ^ y m
= -∕^ ∆ x ∆ p i - ∕∣,∆ x∆ pj + ∆x∆pk
Then, we calculate the norm:
κm = ∣
∣
AB × AC II
= v [.Λ (< - ‰)]2 ∆⅛ ∆y2 [f y (x *, p*)] 2 ∆ x 2 ΔjF + ∆ X 2 ∆ y 2
+
= V [ ∕χ , ^ )]2 + [ ∕ χ , O 2 + 1 ∆ χ m ∆y m
Finally, summing over the entire partition, we see that the surface area over R
may be approximated by the sum
∆∆,, = ∑ √ [ < ⅛ 2 + < t f + ∙ ∆4m

m= 1
where A4 m = ∆x m∆pm . This may be regarded as a Riemann sum, and by taking an

appropriate limit (as the partition becomes more and more refined), we find that
surface area, S, satisfies
∆=1⅛ Σ √ [ ∕K > ‰ ) ] 2 + [ ∕ χ , ^ ) ] 2 + 1 ∆ ^
n → 0° m=∖
= f ∣√ [ ∕ v(Λ-, J')] 2 + [fy (x, τ)] 2 + 1 dA
Surface Area as a Double Assume that the fun ction ∕(x, y) has continuous partial derivatives jj c a n d f in a
Integral region R of the xp-plane. Then the portion of the surface z = f( x , y) that lies
over R has the surface area
S = j j √[∑γ(A T)]2 + [ ∕∕x , f)] 2 + 1 dA
H What this says∙. The region R may be regarded as the projection of the
surface z = ∕( x , y) on the xy-plane. If there were a light source with rays
perpendicular to the xy-plane, R would be the “ shadow” of the surface on the
plane. You will notice the shadows drawn in the figures shown in this section.
It is also worthwhile to make the following comparisons:
Length on x-axis: Arc length:
ds = f 7 [∕'(x )] 2 + 1 dx
Area in xji-plane: Surface area:

[ [ dA [ [ d S = [ f y] [ f( x , y)] 2 + [ f i x , y)]2 + 1 dA
Sec. 13.4 Surface Area 84 9
EXAM PLE 1 Surface area of a plane region
Find the surface area of the portion of the plane x + y + z = 1 that lies in the
first octant (where x ≥ 0, y ≥ 0, z ≥ 0).
Solution The plane x + y + z = 1 and the triangular region T that lies beneath
it in the xy-plane are shown in Figure 13.22. We begin by letting
f(x , y) = l - x - y . T h e n ∕(x
Λ ∙
, y) = -1 and f Z (x, y) = -1 , and
5 = ∫ ∫ √ [∕x (χ, y)]2 + [fy (χ, y)]2 + 1 dA
= ∫ ∫ √ ( - l) 2 + ( - l ) 2 + 1 dA
a. The surface of the plane

•r + T + Z = ∣in the first octant
Look at Figure 13.22 to see the region

T and the limits for both x and y.
b. The surface projects onto

a triangle in the xy-plane.
Figure 13.22 Surface area v 3 ≈ 0.866.

The surface area is -5-
EXAM PLE 2 Surface area by changing to polar coordinates
Find the surface area (to the nearest hundredth square unit) of that part of the
paraboloid x 2 + y 2 + z = 5 that lies above the plane z = 1.
Solution Let ∕(x , y) = 5 - x 2 - y 2. Then f χ (x, y) = -2 x , f(x, y) = -2y, and
Figure 13.23 The surface x 2 + y 2 + z = 5 above z = 1 projects onto a disk.

Now we need to determine the limits for the region D and carry out the
integration by using Fubini’s theorem. The paraboloid intersects the plane
z = 1 in the circle x 2 + y 2 = 4. Thus, the part of the paraboloid whose surface
area we seek projects onto the disk x 2 + y 2 ≤ 4 in the xy-plane, as shown in
Figure 13.23 (note the shadow).
It is easier if we convert to polar coordinates by letting x = r cos θ,
y = r sin θ. Because the region is bounded by 0 ≤ r ≤ 2 and 0 ≤ θ ≤ 2π, we
have
S = H V4 X 2 + 4y2 + I dA
J DJ
V4r 2 + 1 r dr dθ Let u = 4r 2 + 1, so
du = 8 r dr.
If r = 2, then u = 17
and if r = 0, u = 1.
= ⅛ (17 3z2 - l)( 2 τ r-0) ≈ 36.18

The surface area is approximately 36.18 square units.
■ SURFACE AREA PROJECTIONS
We have been projecting the given surface on the xy-plane, but we can easily
modify our procedure to handle other cases. For instance, if the surface y = ∕(x , z)
is projected onto the region Q in the xz-plane, the formula for surface area
becomes
S = ∣f >J[fx (x, z)]2 + [∕j(x, z)]2 + 1 dx dz
An analogous formula for projection onto the region T in the yz-plane would be
*s, = f ∫ y∣
[fy (y, z)]2 + lf z(y, ^)]2 ÷ ι dy dz
EXAMPLE 3 Surface area of a cylinder
Find the lateral surface area of the cylinder x 2 + y 2 = 4, 0 ≤ z ≤ 3.

Solution The cylinder is shown in Figure 13.24a. We can find the lateral surface
area using the formula for a rectangle by imagining the cylinder “opened up.”
The length of one side is the height of the cylinder (3 units), and the length of
the other side of the rectangle is the circumference of the bottom, namely,
2ττ(2) = 4π. Thus, the lateral surface area is 12τr.
We shall illustrate our formula by using it to recompute this surface area.
Specifically, we shall compute the surface area of the right half of the cylinder
and then double it to obtain the required area. It does no good to project the
Figure 13.24 Surface area of a half-cylinder onto the xy-plane, because if we do we will not be accounting for
cylinder. the height of the cylinder.
Sec. 13.4 Surface Area 851
Instead, we will project the surface onto the xz-plane to obtain the
rectangle Q bounded by x = 2, x = - 2 , z = 0, and z = 3, as shown in Figure
13.24b. Because we are projecting onto the xz-plane, we solve for y to express
the surface as y = ∕( x , z). Because
y = ∕( x , z) = V 4 - x 2
we have f χ (x, z) = ^, and f z(x, z) = 0, so that

b. The projection of the cylinder
on the xz-plane
[ ∕A < + [f 2(χ, < + 1= 4 ⅛ + 1= j - γ-
Figure 13.24 Surface area of a r
cylinder
Finally, we note the limits of integration:
- 2 ≤ x ≤ 2 and 0 ≤ z ≤ 3
The surface area can now be calculated:
Cπ,2 6(2 cos θ dθ)

∣-π∕2 V 4 - 4 sin 2 θ Let x = 2 sin θ, so
dx = 2 cos θ dθ.
If x = 2, then θ = y
and if x = —2, then θ = - 7 .
The surface area of the whole cylinder is 2(6π) = 12ττ.
■ AREA OF A SURFACE DEFINED PARAMETRICALLY
Suppose a surface .S’ (see Figure 13.25) is defined parametrically by

R(t∕, v) = x(w, v)i + y(u, v)j + z(w, v)k
Figure 13.25 Area of a parametrically defined function.

Tangent plane for parameters u and v. Let D be a region in the x j ,-plane on which x , y, and z, as
well as their partial derivatives with respect to u and v are continuous. The partial
derivatives of R(M, V) are given by
Ru = aR= axi + ⅛ 1i + ¾ R = ffi = ⅛

v + ≥ J + ⅛
∂u ∂u ∂u ∂u 3v ∂v ∂vj 3V
Suppose the region D is subdivided into cells, as shown in Figure 13.26.
Consider a typical rectangle in this partition, with dimensions ∆x and ∆p, where
∆ x and ∆y are small. If we project a typical parallelogram onto the surface R(u, v),
we obtain a curvilinear parallelogram with adjacent sides R ii(w, V)Δ M and
R I.(M, v)∆v. The area of this parallelogram is approximated by
Δ5, = ∣∣RH(M, v)∆w × R V(M, v)∆v∣∣ = ∣∣RH(M, V) × R 1,(M, v)∣∣∆w Δ V

∆y
Figure 13.26 Area of a surface By taking an appropriate limit, we find the surface area to be a double integral.
defined by a vector function
Surface Area Defined Let D be a region in the xy-plane on which x , y, z and their partial derivatives
Parametrically with respect to u and v are continuous. Also, let S be a surface defined by a
vector function
R( M, V) = x(u, v)i + y(u, v)j + Z (M, v)k
Then the surface area is defined by
.S' = f f ∣∣RII(M, V) × R V(M, V)∣∣ du dv
J DJ
We call the quantity R (M, V) × R V(M, V) the fundamental cross product.
EXAM PLE 4 Surface area defined by a vector function
Find the surface area (to the nearest square unit) of the surface given
parametrically by
R( M, V) = (M sinv)i + (» cosv)j + w2k for 0 ≤ u ≤ 3, 0 ≤ v ≤ 2ττ
Solution We find that R (/ = sin vi + cos vj + 2wk
R r = u cosvi + ( - M sinv)j
We begin with the fundamental cross product
j k
R × R .= sin v cos V 2w = (2w2sinv)i + (2w2cosv)j - uk
u cos v - u sin v 0
We find that
∣
∣
Rπ × R j ∣
= V 4z∕4sin 2 v + 4w4cos2 v + u 2 = V 4 z∕4 + u 1 = wV4w2 + 1
and we can now compute the surface area:
f2 ττ r3 ___________ f2 τ r Γ . 1 ∣ M=3
5, = w^√4w2 + 1 du dv = T ⅛(4M2 + 1)3/2 u
dv
Jo Jo Jo Lι z J I"
= ∫q [⅛ (37 3z2 - 1)] dv = -3-7- γ 2 ~ 1 t2 ^ “ °1 ≈ 117.3187007
The surface area is approximately 117 square units.

Sec. 13.4 Surface Area 85 3
Notice that in the special case where the surface under consideration has the
explicit representation z = f(x, y), it can also be represented in the vector form
Rχ = xi + vj + ∕(x , y)k
where x and y are used as parameters (x = u and y = v). With this vector
representation, we find that
-= i + Λ-k and r j =J + 4 k
RΛ
so the fundamental cross product is

i j k
Rv × R1.= 1 0
0 1
Therefore, in the case where z = f(x, y), the surface area over the region D is given
by
n
D
which is the formula obtained at the beginning of this section.
PROBLEM SET 1 3 .4
θ Find the surface area o f each surface given in Problems 1-19. 9. The portion of the surface z = x 2 over the square region
1. The portion of the plane 2x + y + 4z = 8 that lies in the with vertices (0, 0, 0), (0, 4, 0), (4, 0, 0), (4, 4, 0)
first octant 10. The portion of the surface z = 2x + y 2 over the square
2. The portion of the plane 4x + y + z = 9 that lies in the region with vertices (0, 0, 0), (3, 0, 0), (0, 3, 0), (3, 3, 0)
first octant 11. The portion on the paraboloid
3. The portion of the paraboloid
z = x 2 + y2
z =4- x2 - y2
that lies below the plane z = 1
that lies above the xj ,-plane 12. The portion of the sphere x 2 + y 2 + z 2 = 25 that lies
4. The portion of the paraboloid above the plane z = 3
13. The part of the cylinder x 2 + z 2 = 4 that is in the first
z = x2 + y2 - 9 octant and is bounded by the plane y = 2
that lies below the xy-plane 14. The portion of the plane x + y + z = 4 that lies inside the
cylinder x 2 + y 2 = 16
5. The portion of the plane 3x + 6y + 2z = 12 that is above
the triangular region in the plane with vertices (0, 0, 0),
(1, 0, 0), and (1, 1, 0)
6. The portion of the plane 2x + 2y - z = 0 that is above
the square region in the plane with vertices (0, 0, 0),
(1, 0, 0), (0, 1, 0), (1, 1, 0)
7. The portion of the surface x 2 + z = 9 above the square
region in the plane with vertices (0, 0, 0), (2, 0, 0), (0, 2, 0),
(2, 2, 0)
8. The portion of the surface z = x 2 that lies over the
triangular region in the plane with vertices (0, 0, 0),
(0, 1, 0), and (1, 0, 0)
15. The portion of the sphere x 2 + y 2 + z 2 — 4 that lies inside

the cylinder x 2 + y 2 = l y
Problem 26
27. Verify that a sphere of radius a has surface area 4πa 2 .

28. Verify that a cylinder of radius a and height h has surface
area 2ττah.
16. The portion of the sphere x 2 + y 2 + z 2 = 8 that is inside 29. Find a formula for the area of the conical surface
the cone x 2 + y 2 - z 2 = 0. z = V x 2 + y2 between the planes z = 0 and z = h. Ex
press your answer in terms of h and the radius of the base
of the cone.
30. Find a formula for the surface area of the frustum of the
cone z = 4V x 2 + y 2 between the planes z = Λ1 and
z = h2 , h l > h2 .
17. The portion of the surface z = x 2 + y above the square

0 ≤ x ≤ 2 , 0 ≤ y ≤ 5
18. The portion of the surface z = x 2 - y 2 that lies inside the
cylinder x 2 + y 2 = 9
19. The portion of the surface z = 9 - x 2 - y 2 that lies above
the xy-plane
© 20. ■ What Does This Say? Describe the process for find
ing a surface area.
31. Find the surface area of that portion of the sphere
21. ■ What Does This Say? Describe what is meant by a
x 2 + y 2 + z 2 = 9z that lies inside the paraboloid
surface area projection. x 2 + y 2 = 4z.
22. On a given map a city parking lot is shown by a rectangle 32. Find the surface area of that portion of the cylinder
that is 300 ft by 400 ft. However, the parking lot slopes in x 2 + z 2 = 4 that is above the triangle with vertices
the 400-ft direction. It rises uniformly 1 ft for every 5 ft of
(0, 0, 0), (1, 1, 0), and (0, 1, 0).
horizontal displacement. What is the actual surface area
of the parking lot?
23. Find the surface area of the portion of the plane In Problems 33-38, set up (but do not evaluate) the double
A x + By + Cz = D (A, B, C, and D all positive) that lies integral for the surface area o f the given portion o f surface.
in the first octant. 33. The surface given by z = e - x siny over the triangle with
24. Find the portion of the plane x + 2y + 3z = 12 that lies vertices (0, 0, 0), (1, 1, 0), (0, 1, 0)
over the triangular region in the xp-plane with vertices 34. The surface given by z = x 3 —xy + y 2 over the square
(0, 0, 0), (0, a, 0), and (a, a, 0). (0, 0, 0), (2, 0, 0), (0, 2, 0), (2, 2,0)
25. Find the surface area of that portion of the plane 35. The surface given by z = cos (x 2 + y 2) over the disk
x2 + y2 ≤ %
x +y + z = a
36. The surface given by z = e~λ cosy over the disk
that lies between the concentric cylinders
x2 + y2 ≤ 9
2
x 2 + y 2 = ~- and x 2 + y 2 = a 2 (a > 0). 37. The surface given by z = x 2 + 5xy + y 2 over the region in
the xy-plane bounded by the lines x = y, x = 3, and y = 0
26. Find the surface area of the portion of the cylinder 38. The surface given by z = x 2 + 3xy + y 2 over the region in
x 2 + z 2 = 9 that lies inside the cylinder x 2 + y 2 = 9. the xy-plane bounded b y θ ≤ x ≤ 4 , 0 ≤ y ≤ x
Sec. 13.5 Triple Integrals 855
Compute the magnitude o f the fundamental cross product for 46. Suppose a surface is given implicitly by Fix, y, z) = 0. If
the surfaces defined parametrically in Problems 39-42. the surface can be projected onto a region D in the xy-
39. R(u, v) = (2u sinv)i + (2w cosv)j + u 2 k plane, show that the surface area is given by
40. R(M, V) = (4 sin u cosv)i + (4 sinw sin v)j + (5 cosw)k
41. R(w, v) = wi + vj + u 3 k
42. R(M, V) = (2H sinv)i + (2w cosv)j + (w2sin2v)k
43. Find the surface area of the surface given parametrically
by the equation where F, ≠ 0. Use this formula to find the surface area of
R(u, v) = wvi + {u —v)j + (w + v)k a sphere of radius R.
for u 2 + v2 ≤ 1. Hint: Use polar coordinates. 47. Let 5' be the surface defined by f(x , y, z) = C, and let R be
44. A spiral ramp has the vector parametric equation the projection of S on a plane. Show that the surface area
of S can be computed by the integral
R(w, v) = (M cos v)i + (M sin v)j + vk
for 0 ≤ M ≤ l , 0 ≤ v ≤ ιr. Sketch the ramp and find its
surface area.
45. Find the surface area of the torus defined by
R(w, v) = (a + b cosv)coswi
where u is a unit vector normal to the plane containing R
+ (a + b cos v)sin uj + b sin vk
and V ∕∙ u ≠ 0. This is a practical formula which is
for 0 < b < a, 0 ≤ u ≤ 2ιτ, 0 ≤ v ≤ 2τr. sometimes used in calculating the surface area.
1 3 .5 TRIPLE INTEGRALS
IN THIS SECTION Definition of triple integral, iterated integration,

volume by triple integrals
Triple integrals are defined and developed in essentially the same way as double
integrals. We evaluate triple integrals by iterated integration and interpret them
geometrically in terms of volume. Changing the order of integration in a triple
integral can be complicated, and we consider several examples of this proce
dure.
■ DEFINITION OF TRIPLE INTEGRAL
A double integral ∫ ∫ f(x, y) dA is evaluated over a closed, bounded region in the

R
plane, and in essentially the same way, a triple integral ∫ f ∫ f(x , y, z) dV is
s
evaluated over a closed, bounded solid region in R3. Suppose∕(.x, y, z) is defined
on a closed region 5, which in turn is contained in a “box” D in space. Partition D
into a finite number of smaller boxes with planes parallel to the coordinate planes,
as shown in Figure 13.27.
Figure 13.27 The region of integration S' is subdivided

into smaller boxes.
856 Ch. J 3 Multiple Integration
We exclude from consideration any boxes with points outside A. Let ΔJz 1, ΔJz 2, . . . ,
ΔJz π denote the volumes of the boxes that remain, and define the norm ∣∣P∣∣ of the
partition to be the length of the longest diagonal of any box in the partition. Next,
choose a representative point (x k , y*k , z*k ) from each box in the partition and form
the Riemann sum
j∕⅛ ⅛ ⅞ > 4 >'ι

If we repeat the process with more subdivisions, so that the norm approaches zero,
we are led to the following definition.
Triple Integral If/ is a function defined over a closed, bounded solid region A, then the triple
integral of f over S is defined to be the limit
∫ ∫ ∫ Λ × , y, z) d V = iilim ∑ 1 ∕( x * , y*, z*) ΔKfe

s
provided this limit exists.
In a 1vanced calculus, it is shown that the triple integral J J J ∕( ∙ x >¾ z ) d V

s
exists if (Λ', y , z) is continuous on A and S is piecewise smooth in the sense that it
consists of a finite number of pieces, each of which has a continuously turning
tangent plane (that is, the tangent plane varies continuously from point to point).
It can also be shown that triple integrals have the following properties (which are
analogous to those of double integrals listed in Theorem 13.3). In each case,
assume the indicated integrals exist.
Linearity rule. For constants a and b
∣∣∣ [af(x, y, z) + bg(x, y, z)] d V

s
= <1 ∫ ∫ ∫ ∕ ( w ) 1∕ κ + 6 ∫ ∫ ∫ g(x, y, z ) d V
s s
Dominance rule. If f ( x , y, z) ≥ g(x, y, z) on S, then
g(x, y , z) d V
s s
Subdivision rule. If the solid region of integration A can be subdivided into two
solid subregions A 1 and A,, then
∣f ∣f( χ , y, z ) d V = 1 1 1 f( χ , y , ?) d κ + f f ∣f( χ , y, z ) d v
s *1 ¾
■ ITERATED INTEGRATION
As with double integrals, we evaluate triple integrals by iterated integration.

However, it is generally more difficult to set up the limits of integration in a triple
integral, because the region of integration A is a solid. The relatively simple case
Sec. 13.5 Triple Integrals 85 7
where 5 is a rectangular solid (box) may be handled by applying the following

theorem.
THEOREM 13.5 Fubini’s theorem over a parallelepiped in space

If ∕(x , y, z) is continuous over a rectangular solid R: a ≤ x ≤ b, c ≤ y ≤ d,
r ≤ z ≤ s, then the triple integral may be evaluated by the iterated integral
∫J> R
, y, z) dV - f(x , y, z) dx dy dz
The iterated integration can be performed in any order (with appropriate

adjustments to the limits of integration):
dx dy dz dx dz dy dz dx dy
dy dx dz dy dz dx dz dy dx
Proof: This proof is beyond the scope of this course and is given in many
advanced calculus courses. ■
EXAMPLE 1 Evaluating a triple integral using Fubini’s theorem
Evaluate J J J z 2ye x dV, where R is the box given by

R
Solution We shall evaluate the integral in the order dx dy dz.

J J j f(x , y, z) dl7 = J J J z 2yex dx dy dz
R
•2
This means z 2ye x dx dy dz.
∣o
•2
z 2y[e - 1] dy dz
It might be instructive for you to verify that the same result is obtained by
using any other order of integration—for example,
dz dy dx M M
To evaluate a triple integral over a region that is not a box, suppose we have a
solid region S' with an “upper” surface z = u(x, j) and a “lower” surface
z = v(x, j-j defined over a common domain A in the xy-plane. In this case, S' may
be described as the set of all points (x, y, z) such that for each fixed point (x, j) in
A, z varies from v to u, as shown in Figure 13.28.
Figure 13.28 A solid region 5

bounded by the surfaces
z = u(x, y) and z = v(x, y)
The domain A is then described in a double integral. These observations are

summarized in the following theorem.
THEOREM 13.6 Triple integral over a general region
If A is a region in space that is bounded above by the surface z = u(x, y) and below
by z = v(x, y) as (x, y) varies over the planar region A, then
/ / / ^ X’ y, z ^d V = / / / (V). f(x, y, z) dz dA
s a
Proof: This proof is omitted.
If the region of integration 5 has the form described in Theorem 13.6, the
integral may be evaluated by integrating first with respect to z (as z varies from
z = v(x, y) to z = w(x, y)) and then computing a double integral over the planar
region A.
EXAMPLE 2 Evaluating a triple integral over a general region
Evaluate / / / λ w he r e 5, is the solid in the first octant bounded by the

s
cylinder x λ + y 2 = 4 and the plane 2y + z = 4.
Solution The solid is shown in Figure 13.29.
Figure 13.29 The region Abounded

by the plane 2y + z = 4 and the cylinder
x 2 + y 2 = 4 in the first octant a. The solid S b. The region A
The upper boundary surface of S is the plane z = 4 - 2y, and the lower
boundary surface z = 0 is the xy-plane. The projection A of the solid on the
xy-plane is the quarter disk x 2 + y 2 ≤ 4 with x ≥ 0, y ≥ 0 (because 5 lies in
the first octant). This projection may be described in type I form as the set of
all (x, y) such that for each fixed x between 0 and 2, y varies from 0 to V 4 - x 2 .
Thus, we have
ri r V 4 -χ 2 r4 - 2y
∫ ∫ ∫ - ∫ ∫ r - =
S a
Io Jo Jo
x dz dy dx
∕∙2 r ∀ 4 -x 2 ∣
∙2 ι∙V 4 -x 2
= x[(4 - 2y) - 0] dy dx = (4x - 2xχj dy dx
Jo Jo ’ Jo Jo
∕, 2 |r = ^4—.X- f 2r ------------- η
= [ 4 x y - x y 2]l 0 dx= 4 x v 4 - x 2 - x(4 - x 2) I dx
Jo ' 1j Jo l j
= [ - f ( 4 - x 2)3z2 - 2x 2 + ∣ x 4]∣θ = [0 - 8 + 4 + ⅛ + 0 - 0] = f
■ VOLUME BY TRIPLE INTEGRALS
Just as a double integral can be interpreted as the area of the region of integration,
a triple integral may be interpreted as the volume of the solid of integration. That
is, if V is the volume of the solid region S, then
1" - ∫ ∫ ∫ ^
E X A M P L E 3 Volume of a tetrahedron
Find the volume of the tetrahedron T bounded by the part of the plane
2x + y + 3z = 6 in the first octant.
Solution The tetrahedron T is shown in Figure 13.30a.
b. The region projected onto

a. The tetrahedron bounded by the plane the xy-plane is a triangle.
2.r + y + 3z = 6 and the positive coordinate
planes.
Figure 13.30 Volume of a tetrahedron

The upper surface of T is the plane z = ∣(6 - 2x - y) and its lower surface is
z = 0. Note that T projects onto a triangle in the xy-plane, as shown in Figure
13.30b. Described in type I form, this triangle is the set of all (x, y) such that
for each fixed x between 0 and 3, y varies from 0 to 6 - 2x. Thus,
Γ C C r C A (6 -2 x -y ) f 3 f 6 -2 x 4 ( 6 - 2 x - y )
K = J J j dV = J J J dz dA = j J j dz dy dx
r3r n ∣ 0 ^∣∣r=β-2χ
l [ 2 y - ^ χ y - - by 2j∣ w dx
f ^2(6 - 2x) - ∣x(6 - 2x) - g(6 - 2x)2 - θ] dx
f g(6 - 2x)[12 - 4x - ( 6 - 2x)] dx Common factor

∣o
⅜f (6 —2x)(6 - 2x) dx = g f (6 - 2x)2 dx = 6

Jo Jo
The volume of the tetrahedron is 6 cubic units.
Sometimes it is easier to evaluate a triple integral by integrating first with

respect to x or y instead of z. For instance, if the solid region of integration 5, is
bounded in the back by x = x 1(y, z) and in the front by x = x 2(y, z), and the
boundary surfaces project onto a region A in the yz-plane, as shown in Figure
13.3la, then
a. A solid S with "front" surface b. A solid S with "side" surfaces

x = x 2 (y, z) and "back" surface y = y 2 (x, z) and y = y l (x,z)
* = * 1(y, z)
Figure 13.31 Iterated integration with respect to x or y first

On the other hand, if the solid region of integration 5 is bounded on one side
by the surface y = y i (x, z) and on the other by y = y 2(x, z), and the boundary
surfaces project onto a region A in the yz-plane, as shown in Figure 13.3 lb, then
f∕ f∕ ( w ) d r = f f Γ f ( x , y , z) dy dA
s A Y '
Sec. 13.5 Triple Integrals 861
As an illustration, we will now rework Example 3 by projecting the tetrahe

dron S onto the yz-plane.
E X A M P L E 4 Volume of a tetrahedron by changing the order of integration
Find the volume of the tetrahedron V bounded by the part of the plane
2x + y + 3z = 6 in the first octant by projecting V onto the yz-plane.
Solution In Example 3 we projected .S onto the xy-plane to find
■ Π R Π Γ ' - <
s a
where A was the projection in the xy-plane and we found the volume to be 6.
Now, project onto the yz-plane as shown in Figure 13.32.
Plane
a. The tetrahedron bounded by the plane b. The region projected onto

2x + y + 3z = 6 and the positive coordinate the yz-plane is a triangle.
planes.
Figure 13.32 Volume of a tetrahedron; alternative projection
First, we note that V is bounded from “ behind” by the yz-plane and “ in front”
by the plane 2x + y + 3z = 6, which we express this time by solving for x:
x = ∣(6 - y - 3z). Thus,
S A yz
where A is the projection onto the yz-plane, as shown in Figure 13.32b. We

see that this projection is the triangle B bounded by the lines z = 0, y = 0, and
z = ∣(6 - y). Thus, for each fixed y between 0 and 6, z varies from 0 to
∣(6 - y), and we have
r6 f i(6-y) rl(6-y-3z) r6 r∣(6-y)
V= ^ dx dz dy = ^ ⅛(6 - y - 3z) dz dy
Jo Jo Jo Jo Jo
f 6 ∣^(6 - y) - ⅛y(6 - y ) - ⅛(6 - y)2 - 0^∣ dy

Io l
= ⅛ f (6 - y)2 dy = 3^(6 - y) 3 = 0 + ⅛(6)3 = 6

Io
As you would expect, this is the same as the result from projecting on the xy-
plane. ≡≡
EXAMPLE 5 Setting up a triple integral to find a volume
Set up (but do not evaluate) a triple integral for the volume of the solid S that
is bounded above by the sphere x 2 + y 2 + z 2 = 4 and below by the plane
y + z = 2.
Solution First, note that the intersection of the plane and the sphere occurs
above the xy-plane (where z ≥ 0), so the sphere can be represented by the
equation
z - V4 - x 2 ~ y 2
(the upper hemisphere). To find the limits of integration for x and y we
consider the projection of 5, onto the xy-plane. To this end, consider the
intersection of the hemisphere and the plane z = 2 - y:
V4 - x 2 - y 2 = 2 - y
Figure 13.33 The projection of S 4 - x 2 - y 2 = 4 - 4y + y 2
onto the xy-plane
x 2 + 2y2 - 4y = 0
x 2 + 2 ( y - 1)2 = 2
Although this intersection occurs in IR3, its equation does not contain z∖
therefore, the equation serves as a projection on the xy-plane, where z = 0.
This is an ellipse centered at (0, 1). We sketch this curve in the xy-plane, as
We consider this as a type II region, which means we will integrate first

with respect to x, then with respect to y. Because x 2 + 2(y - 1)2 = 2, we see
for fixed y between 0 and 2, x varies from -V 2 - 2(y - 1)2 = -V 4y - 2y2
to V4y - 2y2 . However, we can use symmetry to find twice the integral as
x varies from 0 to V4y - 2y2. This leads us to set up the following triple
integral.
[∙2 r∖ ''4y-2y2 ∕∙V4-x 2- y 2
F=2 dz dx dy
Jo Jo J2- y
PROBLEM SET 1 3 .5
θ 1. ■ What Does This Say? Describe Fubini’s theorem
over a parallelpiped in space.
2. ■ What Does This Say? Set up integrals, with appropri s
ate limits of integration, for the six possible orders of where S is the region described by the inequalities
integration for f f f f ( x , y, z) d V where S' is the solid 0 ≤ x ≤ 1, 0 ≤ y ≤ x , and 0 ≤ z ≤ x + y
.s,
2 2 , f ∣{ y ^ r .
described by
S’: y 2 ≤ x ≤ 4; 0 ≤ y ≤ 2; 0 ≤ z ≤ 2- x
where S' is the solid in the first octant bounded by the
hemisphere x = V 9 - ,v2 - z 2 and the coordinate planes
Compute the iterated triple integrals in Problems 3-14.
Find the volume V o f the solids bounded by the graphs o f the
3 f 5 dx dy dz
∏ equations given in Problems 23-28 by using triple integration.
23. x + y + z = 1 and the
coordinate planes
24. y = 9 - x 2 , z = 0,
5. f J j 8x 2y z 3 dx dy dz z =T
x rx + y
∏Jo xy∑ dz dy dx
yz cosvy dz dx dy
^π
J
fl fl
∣ x 2>, cos xyz dz dy dx
o Jo Jo
*l fy f ln y f3 f2 z rIn y
J
∣
o Jo Jo
’
-4 r2z rV 3x χ - y
i ez + 2 x d z d x d y 12.
J ∣Jo Jo
ye~ x dxdydz
13. —
χ
?— dv dx dz
Ji J- ι Jo +y^ '
fl f.v 2 fv
14. I <x + v) dz dy dx
Jo J x - ι J —x
Evaluate the triple integrals in Problems 15-22.
15.
Ill s
(x 2y + y 2z) dV, where S is the box 1 ≤ x ≤ 3,
-1 ≤ y ≤ l , 2 ≤ z ≤ 4
27. x 2 + 3y2 = z 28. x 2 + y 2 + z 3 = 9, z = 0
16. (xy ÷ 2yz) dV, where S is the box 2 ≤ x ≤ 4, and the cylinder
J J J l ≤ y ≤ 3 , —2 ≤ z ≤ 4 y2 + z = 4
17. xyz dV, where S' is the tetrahedron with vertices

j JJ ' (0, 0, 0), (1, 0, 0), (0, 1, 0), and (0, 0, 1)
18. ∣' i x 2y dV, where S' is the tetrahedron with vertices

J JJ ^ (0, 0, 0), (3, 0, 0), (0, 2, 0), and (0, 0, 1)
19. xyz dV, where S is the region given by

J J J χ 2 + y 2 + z 2 ≤ 1, y ≥ 0, z ≥ 0
20. x dV, where S' is bounded by the paraboloid

J J J z = x 2 + y 2 and the plane z = 1
For each given iterated integral there are five other equivalent 42. Find the volume of the solid region between the parabo
iterated integrals. Find the one with the requested order in loid z = k(x 2 + y 2) and the sphere x 2 + y 2 + z 2 —2k^ l ,
Problems 29-32. where k > 0
43. Let V be the volume of the tetrahedron bounded by the
29. fix , y, z) dz dy dx: change the order to dz dx dy.
plane
- +^+-
a b c
change the order to dy dz dx. (a > 0, b > 0, c > 0) in the first octant. Express V as
six different iterated triple integrals and evaluate one of
V 4 - Λ^2 r∖X-x1-yi these integrals to show that V = ⅛abc.
Π Jo
change the order to dx dy dz.
V 4 -x 2 z∙2√χ≈+^2
fix , y, z) dz dy dx;
44. T HINK TANK P ROBLEM Let B be the box defined by
a ≤ x ≤ b, c ≤ y ≤ a, r ≤ z ≤ s. Is it true that
/ / / d V
Π Jo
fix , y, z) dz dy dx;
change the order to dx dy dz.

B
θ 33. Find the volume of the ellipsoid h(z) dz
1
4 9 16 Either show that this equation is generally true, or find a
counterexample.
34. Find the volume of the region between the two elliptic
paraboloids 45. Change the order of integration to show that
Y2 ι γ2 ∣ ∣fiu) du dv — ∣ (x — t) fit) dt
z = y + y 2 - 4 and z = - y - / + 4 lo Jo Jo
35. Find the volume of the region bounded by the Also show that
paraboloids z = 16 - x 2 —2y2 and z = 3x2 + 2 / .
36. Find the volume of the region that is bounded above by / - 0 2 ∕( 0 dt
the elliptic paraboloid
46. Evaluate the triple integral
z≡⅝ +∕
J J J sin(ιr - z)3 dz dx dy
on the sides by the cylinder
s
γ-2
⅝ + y2 = ι when 5 is the solid region bounded below by the xy-plane,
above by the plane x = z, and laterally by the planes x = y
and below by the xy-plane. and y = ττ. Hint'. See Problem 45.
37. Find the volume of the solid region in the first octant that Higher-dimensional multiple integrals can be defined and
is bounded by the planes z - 8 + 2x + y, y = 3 - 2x, evaluated in essentially the same way as double and triple
and v = x integrals. Evaluate the multiple integrals in Problems 47-48.
38. Use triple integration to find the volume of the Great
Pyramid of Cheops in Egypt, which is approximately 480
ft tall and has a square base of side 755 ft. {{J1 xyz w dx dy dz dw,
2 2
θ 39. Use triple integration to find the volume of a sphere. where H is the four-dimensional “hyperbox” defined by
0 ≤ x ≤ 1, 0 ≤ y ≤ 2, —1 ≤ z ≤ 1,1 ≤ w ≤ 2
40. Use triple integration to find the volume of a right
pyramid with height H and a square base of side 5.
41. Use triple integration to find the volume of the ellipsoid
48.
ππ H
e x 2y+z+w fa fa fa .
where H is the four-dimensional region bounded by the

hyperplane x + y + z+ w = 4 and the coordinate spaces
x = 0, y = 0, z = 0, and w = 0 in the first hyperoctant
(a > 0, b > 0, c > 0). Set up, but do not evaluate. (where x ≥ 0, y ≥ 0, z ≥ 0, w ≥ 0)
Sec. 13.6 Mass, Moments, and Probability Density Functions 865
1 3 .6 MASS, MOMENTS, AND PROBABILITY DENSITY FUNCTIONS
IN THIS SECTION Mass and center of mass, moments of inertia, joint

probability density functions
We investigate applications of double and triple integrals to problems involving
center of mass for both homogeneous and nonhomogeneous laminas. We also
examine moments of inertia and joint probability density functions.
■ MASS AND CENTER OF MASS
Recall (from Section 6.3) that a solid object that is sufficiently “flat” to be
regarded as essentially two-dimensional is called a lamina. It is worthwhile to
recall some of the formulas from Section 6.3. Consider a region between the
graphs of y = ∕(x) and y = g(x) over the interval [a, b] with density p. Then,
The mass of the lamina is m = p[f(x) ~ g(x)] dx.

Ja
The moments are defined as follows:
0 The moment of an object f ft P
about an axis is the product of About the x-axis: M = y[f(x) + g (x ][∕W - ⅛'(λ')] dχ
Ja
its mass and the signed distance
from that axis. 0 rb
About the y-axis: M γ = px[∕(x) - g(x)] dx
Ja
_ m _ M
v
The center of mass (x, y) is given by x = — and y =
Suppose a particular lamina occupies a bounded region R in the xy-plane, and

let p(x, y) be the density (mass per unit area) of the lamina. A homogeneous lamina
has constant density p = m ,A, where m is the mass and A is the area of the lamina.
If the lamina is nonhomogeneous, its density p(x, p) varies from point to point. In
Figure 13.34 Partition of a lamina

in the plane
this case, we can partition the region R into a number of rectangles, as shown in
Figure 13.34. Choose a representative point (xk , y k ) in each rectangle of the
partition, and use the formula
Δ ^ = P(⅞ ¾ ) Δ 4
to approximate the mass of each rectangle. We then approximate the total mass m
by the Riemann sum
n
∑Ι P ( 4 y k )∆A k
so that
m
R
We use these observations as the basis for the following definition.
Mass of a Planar Lamina If p is a continuous density function on the lamina corresponding to a plane
of Variable Density region R , then the mass m of the lamina is given by
m 1 1 ρ(χ, y) dA
EXAM PLE 1 M ass of a planar lamina
Find the mass of the lamina of density p(x, y) = x 2 that occupies the region R
bounded by the parabola y = 2 - x 2 and the line y = x.
Solution Begin by drawing the parabola and the line, and by finding their points
of intersection, as shown in Figure 13.35. By substitution,
x = 2 - x2 (x + 2)(x - 1 ) = 0
2
x + x - 2=0 x = -2 , 1
We see that the region R is the set of all (x, y) such that for each x between
- 2 and 1, y varies from x to 2 - x 2.
m = [ [ x 2 dA = [ f x 2 dy dx = [ x 2(2 - x 2 - x) dx
Figure 13.35 A lamina over the
J Jo∖J J-2 Jx J-2
region R bounded by y = 2 - x 2
and y = x
= [ (2x 2 - x 4 - x 3) dx = l^⅛- - ⅜- - ⅜- I = O = 3.15
∖_____________________________L J *-× ∙ —I ZA√
The moment of an object about an axis is the product of its mass and the
signed distance from the axis. Thus, by partitioning the region R as before (with
mass), we see that the moments M χ and M y of the lamina about the x-axis and y-
axis, respectively, are approximately equal to the Riemann sums
n n
M X ≈ ∑ Λ
κ-~ 1 ↑ P(⅞T⅛) Δ A and M y ≈ ∑ x ↑ p(χ*, y*k ) Δ A k
' κ- 1 ↑
Distance to x-axis Distance to y-axis
By taking the limit as the norm of the partition tends to 0, we can formulate the
Moments and Center of Mass If p is a continuous density function on a lamina corresponding to a plane
of a Variable Density Planar region R , then the moments of mass with respect to the x- and y-axes,
Lamina respectively, are
y p(x, y) dA and Λdy = x p(x, y) dA
R J RJ
Furthermore, if m is the mass of the lamina, the center of mass is (x, y), where
_ m , , _ M
x = - a nd _ v
0 M χ has a factor of y and M y
has a factor of x 0 If the density p is constant, the point (x, y) is called the centroid of the region.
Sec. 13.6 Mass, Moments, and Probability Density Functions 86 7
EXAM PLE 2 Finding a center of mass
Find the center of mass of the lamina of density p(x, y) = x 1 that occupies the
region R bounded by the parabola y = 2 - x 2 and the line y = x. This is the
lamina defined in Example 1.
Solution In Example 1, we found that for each fixed x between - 2 and 1, y varies
from x to 2 - x 2 (see Figure 13.35). Thus, we have
9
7
From Example 1, m = ⅛ρ so the center of mass is (x, y), where
My _1 8 M ~⅛ 9∩
___ 5_ - ∣ ≈ -1 .1 4
m 63 —
m = 6T3 T = ~449F ≈ - θ ∙ 41
20 20
In a completely analogous way, we can use the triple integral to find the mass
and center of mass of a solid in R 3. The density p(x, y, z) at a point in the solid now
refers to mass per unit volume, and the mass m, moments M , M χz , M χv about the
yz-, xz-, and xy-planes, respectively, and coordinates x , y , z of the center of mass
are given by:
Mass
Moments
Distance to the yz-plane
M j κ - ∫ ∫ ∫ y p(χ, y , z) d v
R 1
Distance to the xz-plane
Note the distances to the coordi

nate axes. Distance to the xy-plane
Center of mass (x, y, z) = ⅛,

m
‰
m m
If the density is constant, the center of mass is still called the centroid.
Example 3 illustrates how this point can be found by multiple integration.
EXAM PLE 3 Centroid of a tetrahedron
A solid tetrahedron has vertices (0, 0, 0), (1, 0, 0), (0, 1, 0), and (0, 0, 1) and
constant density p = 6. Find the centroid. (The density must be constant to
talk about the centroid.)
Solution The tetrahedron may be described as the region in the first octant that
lies beneath the plane x + y + z = 1, as shown in Figure 13.36a.
z.,
a. A tetrahedron b. Projection on the xγ-plane

Figure 13.36 The centroid of a tetrahedron
The boundary of the projection of the upper face of the tetrahedron in the xy-
plane is found by solving the equations x + y + z = 1 and z = 0 simultaneous
ly. We find that the projection is the region bounded by the coordinate axes
and the line x + y = 1, as shown in Figure 13.36b. This means that for each
fixed x between 0 and 1, y varies from 0 to 1 - x.
m 6 dz dy dx
R
f1 f∣ x r∣ ∣y=l-x
i j 6(1 - x - y) dy dx = [6y - 6xy - 3y2] , υ dx
Jo Jo ' Jo ∣∙ '
f 3(x - 1)2 dx = [x 3 - 3x 2 + 3x] 0 = 1

∣o
Similarly, we find that
M yv , = fJ fJ [J 6x d V =Jof Jo[ Jo[ 6x dz dy dx = ⅛

R
M
= f [ f 6yd V = Jo[ [ [ 6y dz dy dx = ⅛
J J J Jo Jo
R
M 6z dz dy dx = 4
Io
R
(Verify the details of these integrations.) Thus,

M l M l M *
-JΞ = 4 = 0.25,Jλ>= ^ = ≤ = 0.25, z = = ∣ = 0.25
m l ’- m 1
The centroid is (0.25, 0.25, 0.25).

■ MOMENTS OF INERTIA
The moments M χ and M y of a lamina are sometimes called the first moments
about the x- and y-axes.
M χ = J J y ρ(x, y) dA and My = f J x p(x, y) dA
r Mass r Mass
↑ ↑
Distance to x-axis Distance to y-axis
We now attach meaning to a second moment, also known as the moment of inertia.
The concept of moment of inertia is closely related to the kinetic energy of a
rotating particle. Suppose a particle of mass m rotates in a plane at a distance of r
from a fixed axis L that is perpendicular to the plane of motion, as shown in
Figure 13.37.
Figure 13.37 The particle has mo

ment of inertia I - mr2 about line
L.
If the angular velocity is ω = dθ∕dt, then the kinetic energy of the particle is
defined by K.E. = ∖ lω2, where I = mr2 is defined to be the moment of inertia of the
particle about the axis L. Likewise, it can be shown that a particle of mass m
moving along a straight line with velocity v has kinetic energy ⅛mv2, and by
comparing the formulas for energy, we see that the moment of inertia I is in some
sense the rotational analog of the mass of the particle. Alternatively, the moment
of inertia can be thought of as measuring the tendency of matter to resist a change
in its rotation.
Suppose we now look at a lamina that occupies the region R in the plane.
Consider a small “cell” of the lamina with mass ∆m, and assume that the cell is
small enough for all the mass to be located at a single point (x, y). Then the cell is y
units from the x-axis and x units from the y-axis, and it has moment of inertia
∆7γ = ∆my 2 about the x-axis, and Δ ∕, = ∆mx 2 about the y-axis. If the cell has area
ΔA and its density is p(x, y), then ∆m = ρ(x, y)ΔA. Finally, by partitioning the
lamina into a number of cells and taking the limit as the partition becomes more
and more refined, we expect that the moments of inertia of the entire lamina
about the x-axis and the y-axis will be given by the following integrals:
Moments of Inertia The moments of inertia of a lamina of variable density p about the x- and y-
axes, respectively, are
Λc = J p(x1 y )jA and Iy = J J x 2 p(x, y) dA
0 I χ has a factor of y2 and I has r Increment of mass T r Increment of mass
a factor of x 2 . 0 I ↑ I ↑
Moment about (Distance)2 Moment about (Distance)2
x-axis t 0 t∣
l e χ-axis y-axis t 0 ⅛ e j l. a x is
EXAM PLE 4 Finding the moments of inertia
A lamina occupies the region R in the plane that is bounded by the parabola
y = x 2 and the lines x = 2 and y = 1. The density of the lamina at each point
(x, y) is ρ(x, y) = x 2y. Find the moments of inertia of the lamina about the x-
axis and the y-axis.
Solution The graph of R is shown in Figure 13.38. We see that for each fixed x
between 1 and 2, y varies from 1 to x 2 .
Iχ = J J y 2 dm = J J y 2p(x, }’) dA = j J y 2(∙X2y) dA
f 2 fχ2 2 3 1 f2 1 516
- I x y d y d x = ^ ∣ x 2(x8 - 1) dx = ≈ 45.94
Fixed x
Figure 13.38 Moment of inertia
of a lamina
A simple generalization enables us to compute the moment of inertia of a

solid figure about an axis L outside the figure. Specifically, suppose the solid
occupies a region R and that the density at each point (x, y, z) in R is given by
p(x, y, z). Because the square of the distance of a typical cell in R from the x-axis is
y 2 + z 2, the moment of inertia about the x-axis is
J J ,(T2 + Z 2) P(X , y, z) dV
R ∣ Increment of mass
Square of the distance to the x-axis
Similarly, the moments of inertia of the solid about the y-axis and the z-axis are,
respectively,
■2 + z 2) p(x, y, z) dV
R Φ Increment of mass
Square o f the distance to the y-axis
4 = ∫ ∫ ∫ (χ 2 + y 2)p(χ, y, z ) d v
R ∣ Increment o f mass
Square o f the distance to the z-axis
EXAM PLE 5 Moment of inertia of a solid
Find the moment of inertia about the z-axis of the solid tetrahedron A with
vertices (0, 0, 0), (0, 1, 0), (1, 0, 0), (0, 0, 1) and density p(x, y, z) = x.
Solution In Example 3, we observed that the solid 5 can be described as the set of
all (x, y, z) such that for each fixed x between 0 and 1, y lies between 0 and
1 - x, and 0 ≤ z ≤ 1 —x —y. Thus,
∕, = ∫∫∫(x ≈ + y 2)p(x, y, z) dV = x(x 2 + y 2) dz dy dx

Jo ∣
o Jo
s
Sec. 13.6 Mass, Moments, and Probability Density Functions 87 1
—x —y) dy dx
Γ1 fx 3(l - x ) 2 x (l - x ) 4 ] , _ χ
∣0 L 2 + 12 J a x 90
■ JOINT PROBABILITY DENSITY FUNCTIONS
An important application of integration is the computation of probabilities. A

probability density function of a single variable x is a function f( x ) with f(x ) ≥ 0
and
Γx
∕'(x) dx = 1
Then the probability that x is between a and b is defined by

rb
F, (α ≤ x ≤ b) = f(x) dx
Ja
In geometric terms, the probability P(a ≤ x ≤ b) is the area under the graph o ff
over the interval a ≤ x ≤ b. (See Figure 13.39.)
Figure 13.39 Probability as the To extend this definition to two variables, we define a joint probability density
area under a curve function for the variables x and y to be a function f ( x , y) such that
f ( x , y) ≥ 0 and
∕‰ y ) i n ∕ l ] = f J^f(x,y)dA
Then the probability that the point (x, p) lies in the region R is defined by
P[(x, y) in JR] =
See Figure 13.40. The techniques for constructing joint probability density
functions from experimental data are outside the scope of this textbook and are
discussed in many textbooks on probability and statistics. The use of a double
integral to compute a probability with a given joint density function is illustrated
in the next example.
EXAM PLE 6 Compute a probability
Suppose x denotes the time (in minutes) that a customer at a particular

grocery store spends shopping and y denotes the time the customer spends
standing in the checkout line. A study suggests that the joint probability
function for x and y is
Figure 13.40 Probability that a 1 Z, - X∕1 0 eΛ - V∕20 for x ≥ 0 and y ≥ 0

200 e '
point lies in a region f ( x , T) =
to otherwise
Find the probability that the customer’s total time in the store will be no
greater than 30 min.
Solution The goal is to find the probability that x + y ≤ 30. Stated geometrical
ly, we wish to find the probability that a randomly selected point (x, j) lies in
the region R in the first quadrant that is bounded by the coordinate axes and
the line x + y = 30 (see Figure 13.41). This probability is given by the double
integral
r r r3° r30-x ι
P[(x, y) is in /?] = I I ∕(x , y) dA = I I ^ QQ e x '^ e y,2 ° dy dx
= !U ofo e- x /l0 [e-™30-x) _ 1 ] d x

Figure 13.41 Triangle consisting
of all points (x, y) such that - 1 r e - 3/2e-*/20 e -x7l0 ^l 30
x + y ≤ 30, x ≥ 0, y ≥ 0 - 10 - 1∕20 - 1∕ιo θ
= e~i - 2A ~3Z2 + 1
≈ 0.6035
Thus, it is about 60% likely that the shopper will spend no more than 30
minutes in the store. M M
Many times we use a complementary property of probability that states

P(E,) + P(E) = 1
where E and E together constitute the set of all possible outcomes (that is, they are
complementary probabilities). For Example 6, since the events E = {the shopper
will spend no more than 30 minutes in the store} and E = {the shopper will spend
more than 30 minutes in the store}, we can find the probability that the shopper
will spend more than 30 minutes in the store by using this complementary
property:
1 - 0.6035 = 0.3965
PROBLEM SET 1 3 .6
O 1. ■ What Does This Say? Discuss the procedure for find 10. The part of the spherical solid of density p = 2 described
ing the center of mass of a lamina. by x 2 + y 2 + z 2 ≤ 9, x ≥ 0, y ≥ 0, z ≥ 0
2. ■ What Does This Say? Discuss a procedure for finding 11. The solid tetrahedron of density p = 4 bounded by the
moments in three dimensions. plane x + y + z = 4 in the first octant
,3. ■ What Does This Say? Discuss the terms center of 12. The solid bounded by the surface z = sinx, x = 0, x = ττ,
mass and centroid. y = 0, z = 0, and y + z = 1, where the density is p = 1
4. ■ What Does This Say? Discuss moments of inertia.
Use double integration in Problems 13—18 to find the center of
Find the centroid for the regions described in Problems 5-12. mass of a lamina covering the given region in the plane and
5. A lamina with ρ = 5 over the rectangle with vertices (0, 0), having the specified density p.
(3, 0), (0, 4), (3, 4) 13. p(x, y) = x 2 + y 2 over x 2 + y 2 ≤ 9, y ≥ 0
6. A lamina with p = 4 over the region bounded by the curve 14. p(x, y) = k(x 2 + y 2) over x 2 + y 2 ≤ a 2, y ≥ 0
y = V x and the line x = 4 in the first quadrant. 15. p(x, y) = 7x over the triangle with vertices (0, 0), (6, 5)
7. A lamina with p = 2 over the region between the line and (12, 0).
y = 2x and the parabola y = x 2 16. p(x, y) = 3x over the region bounded by y = 0, y = x 2,
8. A lamina with ρ = 4 over the region bounded by and x = 6
y = sin-jx, x = 0, y = 0, x = ⅜ 17. p(x, y) = x^^1 over the region bounded by y = In x, y = 0,
9. A thin homogeneous plate of density p = 1 with the shape x=2
of the region bounded above by the parabola y = 2 - 3x2 18. p(x, y) = y over the region bounded by y = e~x , x = 0,
and below by the line 3x + 2y = 1 x = 2, y = 0
19. A lamina in the xy-plane has the shape of the semicircular 33. Show that the lamina that covers the circular region
region x 2 + y 2 ≤ a 2 , y ≥ 0. Find the center of mass if x 2 + y 2 = α 2, with mass m, will have moment of inertia
the density at any point in the lamina is mfl2∕4 with respect to the y-axis.
a. directly proportional to the distance of the point from 34. Show that the lamina that covers the circular region
the origin x 2 + j,2 _ a 2, w jt ∣1 m a s s m w i∏ have moment of inertia
b. directly proportional to the polar angle m a 2∕2 with respect to the z-axis.
20. A lamina has the shape o f a semicircular region 35. Show that the lamina that covers the ellipse
x 2 + y 2 ≤ a 2, y ≥ 0. Find the center o f mass o f the
lamina if the density at each point is directly proportional
to the square of the distance from the point to the origin.
21. Find the center o f mass of a lamina that covers the region with mass m and density p = 1 has moment of inertia
bounded by the curve y = In x and the lines x = e 2 , and about the x-axis equal to
y = 0 if the density at each point (x, y) is ρ = 1.
I χ = ab i p = ⅛mb2
22. Find the center of mass o f the solid bounded above by the
note: m = p (πab)
elliptic paraboloid z = x 2 + y 2 , on the sides by the cylin
der x 2 + y 2 = 9 and the plane x = 0, and below by z = 0, area of ellipse
where p(x, z) = x 2 + y 2 + z 2 .
36. Find the center o f mass of the tetrahedron bounded by the
23. Find I χ , the moment of inertia about the x-axis, of the plane
lamina that covers the region bounded by the graph of
y = 1 - x 2 and the x-axis, if the density is p(x, y) = x 2.
24. Find Z., the moment o f inertia about the z-axis, of the
lamina that covers the square in the plane with vertices where a, b, and c are all positive constants in the first
( - 1 ,- 1 ) , ( 1 ,- 1 ) , (1, 1), and ( -1 , 1), if the density is octant. Assume the density is p = x.
p(x, y) = x 2y 2 . 37. A solid has the shape of the sphere x 2 + y 2 + z 2 ≤ a 2 .
θ 25. Find the center of mass of the cardioid r = 1 + sin θ if the Find the centroid of the part o f the solid in the first octant
density at each point (r, θ) is ρ(r, θ) = r. (x ≥ 0, y ≥ 0, z ≥ 0). Assume p = 1.
26. Find the center of mass of the loop o f the lemniscate 38. Suppose the joint probability density function for the
r2 = 2 sin 20 that lies in the first quadrant, for density nonnegative variables x and y is ∕( x , y) = 2e^ 2x e~y . Find
p = 1. the probability that x + y ≤ 1.
27. Find the center of mass (correct to the nearest hundredth) 39. Suppose the joint probability density function for the
o f the part of the large loop of the limaςon r = 1 + 2 cos θ nonnegative variables x and y is f ( x , y) = xe~ x e~y . Find
that does not include the small loop. Assume that ρ = 1. the probability that x + y ≤ 1.
28. Find the center of mass of the lamina that covers the 40. Suppose x is the length of time (in days) that a person
triangular region with vertices (0, 0), (a, 0), (a, ⅛), if a and stays in the hospital after abdominal surgery and y is the
b are both positive and the density at Z, (x, y) is directly length of time (in days) that a person stays in the hospital
proportional to the distance of P from the y-axis. after orthopedic surgery. On Monday, the patient in bed
29. T H IN K T A N K P RO BLEM A solid has the shape of the 107A undergoes an emergency appendectomy (abdominal
rectangular parallelepiped given by —a ≤ x ≤ a, surgery), and the patient’s roommate in bed 107B under
- b ≤ y ≤ b , - c ≤ z ≤ c , and its density is ρ(x, y, z) goes orthopedic surgery for the repair o f torn knee
= x 2y 2z 2. cartilage. If the joint probability density function for x
a. Guess the location of the center of mass and the
and y is
moment of inertia about the z-axis. ∕( x , y) = Ie~ x l2 e^yl3
b. Check your response to part a by direct calculation.
find the probability (to the nearest hundredth) that both
30. A rectangular lamina has vertices (0, 0), (a, 0), (a, ⅛), (0, ⅛) patients will be discharged from the hospital within 3
and its density at any point (x, y) is the product
days.
ρ(x, y) = xy. Find the center o f mass of the plate.
41. A certain appliance consisting of two independent elec
In Problems 31-35, assume that the density in the given tronic components will be usable as long as either one of
lamina is constant. Find the centroid o f the plate. its components is still operating. The appliance carries a
31. The boundary and the interior of the circular region warranty from the manufacturer guaranteeing replace
x 2 + y 2 = a χ fo r m a homogeneous lamina of density ment if the appliance becomes unusable within 1 yr of the
p = 1. Find the moment o f inertia of this circular plate date of purchase. Let x denote the life span (in years) of
about a diameter passing through the center of the lami the first component and y the life span (in years) of the
na. Hint: Use polar coordinates. second, and suppose that the joint probability density
32. Show that the lamina that covers the circular region function for x and y is
χ 2 + y 2 = a 2, and has mass m, will have moment of
inertia mα2∕4 with respect to the x-axis. ∕( x , y) = ∖e~xl2 e~yl2
You purchase one o f these appliances, selected at random Note that i f the entire mass m o f R is located at a distance d
from the manufacturer’s stock. Find the probability that from the axis o f rotation L, then R would have the same
the appliance breaks before the warranty expires. moment o f inertia. Use this definition in Problems 48-50.
42. Suppose x denotes the time (in minutes) that a person 48. A lamina has the shape of the right triangle in the xy-
stands in line at a certain bank and y, the duration (in plane where vertices are (0, 0), (a, 0), and (0, ⅛), a > 0,
minutes) o f a routine transaction at the teller’s window. b > 0. Find the radius of gyration of the lamina about
You arrive at the bank to deposit a check. If the joint the z-axis. Assume ρ = 1.
probability density function for x and y is 49. Find the radius o f gyration about the x-axis of the
semicircular region x 2 + y 2 ≤ a 2, y ≥ 0 given that the
f ( x , y) = | e - x/2e--v/4
density at (x, y) is directly proportional to the distance of
find the probability that you will complete your business the point from the x-axis.
at the bank within 8 min. 50. Let R be the lamina bounded by the parabola y = x 2 and
43. Suppose x denotes the time (in minutes) that a person the lines x = 2 and y = 1, with density ρ(x, y) = x 2y.
spends with an insurance agent choosing a life insurance What is the radius o f gyration about the x-axis?
policy and y the time (in minutes) that the agent spends
doing the paperwork once the client has decided. You
arrange to meet with an insurance agent to buy a life
insurance policy. If the joint probability density function
for x and y is
Often the knowledge o f the center o f mass allows us to
,∕'(-Y D = ⅛ L Λ'30C ,10 greatly simply a problem; but sometimes it can lead us to
an incorrect answer. We explore this incongruity in Prob-
find the probability that the entire transaction will take I lems 51-52.
less than half an hour. 51. Centers of mass in physics
a. Suppose a volume of liquid (or granular solid) is I
The average value o f the continuous function f over R is given by
located so that its center of mass is at the origin. ∣
The material is to be lifted to a height of h units I
One variable: ?----- -π— ------------ —= fix') dx
length of segment R J κ ’ above the center of mass. Starting with a small ∣
element o f volume, ΔF, to be lifted to height, h, I
Two variables: ------- τ4— ■
----- ∏ ∕( x , v) dA derive the formula for the total work done:
area of region R J I
R
Work = ρ
Three variables: y θ l u m e ζJ f s o l i d λ ∫ ∫ j f { x , y, z) d V
where p is the weight (density) o f the material.

Use these definitions in Problems 44-47. b. Explain why the integral in part a simplifies to the
44. Find the average value o f f ( x , y) = ex , where R is the formula
region in the first quadrant bounded by y = x 2, y = 0, and Work = pVh = force × distance
x = 1.
45. Find the average value of where h is the distance between the center of mass
and the level to which the material is lifted.
f i x , y) = ex y - m
c. Picture a cylindrical tank of radius 6 ft and height
where R is the region in the first quadrant bounded by 10 ft positioned so its center of mass is at the
y = x 2 , x = 0, and y = 1. origin. Compute the work required to lift (that is,
46. Find the average value of the function pump) the contents (of density p) to a level of 15 ft
above the center o f mass. Compute this two ways:
f i x , y, z} = x + 2y + 3z by the integral in part a (suggestion— do the dx dy
over the solid region .S' bounded by the tetrahedron with integral by inspection), and also by simply lifting
vertices (0, 0, 0), (1, 0, 0), (0, 1, 0), and (0, 0, 1). the center of mass to the required height. The
47. Find the average value of the function f ( x , y, z) = xyz results o f these computations should agree.
over the solid sphere x 2 + y 2 + z 2 ≤ 1. 52. Newton’s inverse square law
Recall that if two masses m and M are each concen
The radius o f gyration for revolving a region R with mass m trated at (or nearly at) a point and are separated by a
about an axis o f rotation L . with moment o f inertia I, is distance p, the the attractive force is expressed
F = GmM∕p 2. What physicists and others prefer to
do in the case o f real-life masses (which occupy some
Sec. 13.7 Cylindrical and Spherical Coordinates 875
volume) is to use p, the distance between the two with that of part c.
centers of mass. The question you are to explore here e. Repeat the last calculation with h = 200 and com
(and again in Section 13.7 is whether, and when, this pare with the center of mass approximation. Com
simplifying way of computing attracting forces is ment on the role the separating distance p plays.
justified. j
a. Suppose one mass, m, is a point mass located at
(0, 0, Λ) and the second M, has its center of mass
at (0, 0, 0—both have density ρ. We are interested θ 53. Find the moment of inertia of a rectangular lamina with
in the total resultant force that M exerts on m, and dimensions h and I about an axis L through its center of
we assume this to be in the z-direction; hence we mass. Assume the lamina has density p = 1.
will sum the vertical component of the forces 54. Prove the following area theorem of Pappus:
acting on m. Consider, in M, a small element of Let Cbe a curve of length L in the plane. Then the surface
volume ΔK (at x, y, z) and argue that the magni obtained by rotating C about the axis ( in the plane has
tude of the force exerted on m and its vertical area 2πLh, where h is the distance from the centroid of C
component are i to the axis of rotation.
GmΔF 55. A torus (doughnut) has the vector parametric form
δ f =
mag p2
R ( M, V) = (a + b cosv)cost∕i + (a + b cosv)sinwj
+ b sin vk
- G m Δ V (h-3} 2
P3 for 0 < b < a, 0 ≤ u ≤ 2τr, 0 ≤ v ≤ 2π. (see Prob
lem 45, Section 13.4.) Find its surface area by applying
where p 2 = x2 + y2+ (h - z)2.
Note that the term
Pappus’s theorem (see Problem 54).
(h — z)∕p projects the force vector onto the z-axis,
giving the vertical component. Do you see why? 56. Prove the volume theorem of Pappus: Let R be a lamina
in the plane with density p = 1 and area A, and let L be a
b. Form the usual Riemann sum and obtain the
line that does not intersect R. If σ is the distance from the
following force acting between m and M. i
center of mass of R to L, then the volume Iz of the solid
formed by revolving R about L is V = 2ττσA. (See Figure
13.42.)
c. Consider a solid (with density ρ = 5) positioned at

- 2 ≤ x ≤ 2, - 4 ≤ y ≤ 4, and -1 ≤ z ≤ 1.
(Note the center of mass is at the origin.) Then
compute the attracting force between this mass
and a point mass, m, at (0, 0, 8) using the center of
mass formula F = GmM!p2 . i
d. For the masses in part c, compute F using the
integration formula in part b; compare this result
Figure 13.42 Torus
13.7 CYLINDRICAL A N D SPH ER ICAL COORDINATES
IN TH IS SECTION Cylindrical coordinates, integration with cylindrical

coordinates, spherical coordinates, integration with spherical coordinates
We have seen that certain curves have simpler descriptions in polar coordinates
than in rectangular coordinates. In this section, we introduce the cylindrical and
spherical coordinate systems, which are equally useful for describing certain
surfaces and solids in space.
■ CYLINDRICAL COORDINATES
Cylindrical coordinates are a generalization of polar coordinates in iR3. Recall that

the point P with Cartesian coordinates (x, y, z) is located z units above the point
Q(x, y) in the xp-plane (below if z < 0). In cylindrical coordinates we measure the
point in the xy-plane in polar coordinates, with the same z-coordinate as in the
Cartesian coordinate system. These relationships are shown in Figure 13.43.
Cylindrical coordinates are convenient for representing cylindrical surfaces
Figure 13.43 The cylindrical co
and surfaces of revolution for which the z-axis is the axis of symmetry. Some
ordinate system
examples are shown in Figure 13.44.
a. Cylinder b. Cone c. Paraboloid

2 2z
Rectangular equation: x 2 + y 2 = a 2 χ 2 +y
.2
= z ^> X ÷ y = az
r = z r-a z
Cylindrical equation: r=a
Figure 13.44 Surfaces with convenient cylindrical coordinates
We have the following conversion formulas, which follow directly from the
rectangular-polar conversions.
Conversion Formulas Cylindrical to rectangular: x = r cos θ

Rectangular/Cylindrical (r, θ, z) to (x, y ,z) y = r sin θ
Coordinates z=z
Rectangular to cylindrical: r = Vx 2 + y 2
(x, y, z) to (r, θ, z) tan 0 =
z=z
If (r, θ, z) with r > 0 are cylindrical coordinates for any point Q not on the z-axis,
then (r, θ + 2∏ττ, z), for any integer n, will also be cylindrical coordinates for Q.
Hence, if r > 0, points that are not on the z-axis have multiple labelings (the
values of θ differ by a multiple of 2ττ). To avoid this multiplicity, we require that
0 ≤ θ ≤ 2τr. The point (0, 0, 0) is called the pole.
EXAMPLE 1 Converting from rectangular to cylindrical coordinates

Find the cylindrical coordinates of the point whose rectangular coordinates
areP(√3, 1, 2).
Solution Because x - Vz3, y = 1, and z = 2, we have
r = √3 + 1 = 2, ta" 9 =⅛ so 77
6
(√3, 1) or ¢ , 2)
0 This means that the cylindrical coordinates are (2, f , 2). Note in Figure 13.45
Figure 13.45 Cylindrical and rec that we are considering a single point, whether we measure its location in
tangular coordinates of a point rectangular or cylindrical coordinates. a ≡≡
As might be expected, cylindrical coordinates are especially convenient for

describing cylinders.
EXAM PLE 2 Rectangular-form equation converted to cylindrical-form

equation
For a > 0, express the equation of the cylinder x 2 + y 2 = a 2 in terms of

cylindrical coordinates.
Solution Because we are working in three dimensions, but the equation does not
contain z, we recognize that z is arbitrary, and we draw the cylinder as shown
in Figure 13.46. First, obtain the polar coordinates for all points in the xy-
plane that lie on the cylinder. Each such point P lies on the circle x 2 + p 2 = a2 .
Because r2 = x 2 + p 2 , we find
r2 = a 2 or r = a z arbitrary
We keep only the positive value because we require r > 0. ≡≡a
Figure 13.46 The cylinder
x 2 + y 2 = a2 or r = a EXAM PLE 3 Rectangular-form equation converted to cylindrical-form
equation
Find an equation in cylindrical coordinates for the elliptical paraboloid

z = x 2 + 3y2 .
Solution We use the conversion formulas x = r cos θ and y = r sin θ.
z = x 2 + 3y2 = (r cos 0)2 + 3(r sin 0)2
= r 2(cos2 θ + 3 sin2 θ)
= r 2[(l - sin2 0) + 3 sin2 0]
= r 2(l + 2 sin2 0) ≡≡
■ INTEGRATION WITH CYLINDRICAL COORDINATES
In order to do triple integration with cylindrical coordinates, we simply apply our

results from Section 13.3 to convert from rectangular to polar form. Even though
there are six possible orders of integration, we will focus on the one for which the
region is bounded below and above by u(r, 0) ≤ z ≤ v(r, 0). In this case we let D
be the region in the xr-planc described by polar coordinates. This means that the
order of integration we will use is dz dr dθ. In other words, we will first integrate
with respect to z, then with respect to r, and finally with respect to 0. Figure 13.47
illustrates triple integration in polar coordinates. Note that dV is replaced by
r dz dr dθ, which follows from Theorem 13.4.
Figure 13.47 Integration with polar coordinates

Triple Integral in Cylindrical If /is a continuous function of r, θ, and z on the bounded, solid region R , then
Coordinates in cylindrical coordinates the triple integral of f over R is
/ / / ^ r , θ , Z ^ V = / / ∕ '° ^ r, θ , z ) r ^z c ^r
R D ( ,
where D is the region in the xy-plane described by polar coordinates.
EXAM PLE 4 Finding volume in cylindrical coordinates
Find the volume of the solid in the first octant that is bounded by the cylinder
x 2 + y 2 = 2y, the cone z = V x 2 + y 2 , and the xy-plane.
Solution Let A be the region occupied by the solid as shown in Figure 13.48. This
surface is most easily described in cylindrical coordinates.
Cylinder: Cone:
x 2 + y 2 = 2y z = V x2 + y2
r2 = 2 r sin θ z= r
r = 2 sin θ
Figure 13.48 The solid bounded
by x 2 + y 2 = 2y, z = V x 2 + y 2, Thus, we see that the region S has the description
and the xy-plane
0 ≤ z ≤ r 0 ≤ r ≤ 2 sin θ 0 ≤ θ ≤ 7T
2
v(r,β)
V = r dz dr dθ = r dz dr dθ
∣
D J Ju(r,θ)
R
f2sinfl rπ∕2 3 ∣ r = 2 s in e O Cπ ,2
r2 dr dθ = 1
„ dθ = / sin 3 θ dθ
∣o Jo J Jo
77/2
= 8∣^ 0 + co d 0 = 16 Integration table (formula 350)
3 I l∕ I β
o 9
EXAM PLE 5 Centroid in cylindrical coordinates
A homogeneous solid S (p = 1) is bounded below by the xy-plane, on the sides

by the cylinder x 2 + y 2 = a 2 (a > 0), and above by the surface z = x 2 + y 2 .
Find the centroid of the solid.
Solution Let S be the described solid as shown in Figure 13.49. Because the solid
is bounded by a cylinder, we will carry out the integration in cylindrical
coordinates.
Cylinder: Paraboloid:
x 2 + y 2 = a2
r2 = a 2
r= a (a > 0)
Figure 13.49 The solid bounded Let (x, y, z) denote the centroid. By symmetry, we have x = y = 0. Let m
by x 2 + y 2 = a 2 and z = x 2 + y 2 denote the mass of 5. Then
M JJJzp^x,y, dV JJJ
zr dz dr dθ
s_______________
s
z=^ =
m / / / ''p ^x , dz dr d θ
s
∙2π ra
zr dz dr dθ
∣0 Jo 'o _ 6.
r2τr ra -a 4 3
r dz dr dθ 2
Io Jo 'o Verify the details o f the
integration.
The centroid is (0, 0, ⅞- k
■ SPHERICAL COORDINATES
In spherical coordinates we label a point P by a triple (p, θ, φ), where p, θ, and φ are
numbers determined as follows (refer to Figure 13.50):
p = the distance from the origin to the point P; we require p ≥ 0.
θ = the polar angle (as in polar coordinates). In spherical coordinates, this
is called the azimuth of P; we take the principal value of the polar
angle to be 0 ≤ θ < 2ττ.
φ = the angle measured down from the positive z-axis to the ray from the
origin through P. The angle φ is called the colatitude of P; we
require 0 ≤ φ ≤ τr.
Figure 13.50 The spherical co You might recognize spherical coordinates (as well as the associated term inol
ordinate system ogy) as being related to the latitude-longitude system used to identify points on the
surface of the earth. In spherical coordinates, points above the xg-plane satisfy
Dallas-Fort Worth 0 ≤ φ < ττ∕2 and points below satisfy π ∕2 < φ ≤ π, whereas the equation
33o N latitude, N φ = π ∕2 describes the xy-plane. When this coordinate system is used to measure
97° W longitude;
points on the earth, the num ber p is the distance from the center of the earth to the
point with longitude θ from the prime meridian; because latitude is the angle up
from the equator, it is measured in spherical coordinates as τr∕2 - φ.
Spherical coordinates are desirable when representing spheres, cones, or
certain planes. Some examples are shown in Figure 13.51.
Figure 13.51 Surfaces with convenient spherical coordinates
We can use Figure 13.51, as well as the definition of the trigonometric

functions, to state the remaining conversion formulas.
Conversion Formulas Spherical to rectangular: x = p sin φ cos θ

for Coordinate Systems (p, θ, φ) to (x, y, z) y = p sin φ sin θ
z = p cos φ
Rectangular to spherical: ρ = V x 2 + p2 + z 2
(x, y, z) to (p, θ, φ) tan 0 =
, -1 ( z
φ = COS 1 ∕ - s2 - — 2A ....... 72
∖Vx + y + z .
Spherical to cylindrical: r = p sin φ
(p, θ, φ) to (r, θ, z) θ= θ
z = p COS φ
Cylindrical to spherical: p = y∕r2 + z 2
(r, θ, z) to (p, θ, φ) θ= θ
φ, = cos -1 ( > 29z 2, ∖
∖Vr + z )
EXAM PLE 6 Converting rectangular-form equations to spherical-form

equations
Rewrite each of the given equations in spherical form.
a. the sphere x 2 + y 2 + z 2 = a 2 (a > 0)

b. the paraboloid z = x 2 + 3y2 (This is the same elliptic paraboloid we
analyzed in Example 3.)
Solution a. Because p = V x 2 + y 2 + z 2, we see x 2 + y 2 + z 2 = p2 , so we can write

p7z = cr7
p= a Because p ≥ 0
b. z = x + 3y2
2
p cos φ = (p sin φ cos θ)2 + 3(p sin φ sin θ)2

p cos φ = p 2 sin 2 φ cos2 θ + 3p2 sin 2 φ sin 2 θ
0 = p2sin 2 φ cos2 θ + 3p2sin 2 φ sin 2 θ - p cos φ
0 = p(p sin 2 φ cos2 θ + 3p sin 2 φ sin 2 θ - cos φ)
Thus, p = 0 or p sin2 φ cos2 θ + 3p sin 2 φ sin 2 θ —cos φ = 0. Solving this second

equation for p, we obtain
_ cos φ cos φ
P sin 2 φ cos2 θ + 3 sin 2 φ sin 2 θ sin 2 ≠(cos2 # + 3 sin 2 θ)
__ ___________cosφ_______________ cosφ
sin2 φ( 1 - sin 2 θ + 3 sin2 θ) sin 2 ¢(1 + 2 sin2 #)
■ INTEGRATION WITH SPHERICAL COORDINATES
For a solid 5 in spherical coordinates, the fundamental element of volume is a

Figure 13.52 A spherical wedge spherical “ wedge” bounded in such a way that
p l ≤ p ≤ p 1 + ∆p ≠ 1 ≤ φ ≤ φ l + Δ<∕>, θi ≤ θ ≤ 0l + Δ0
This wedge is shown in Figure 13.52. It can be shown that the volume of the wedge
is given by
d V = p 2sin φ dp dφ dθ
Using this formula, we can form partitions and take a limit of a Riemann sum as
the partitions are refined.
Triple Integral in Spherical If/ i s a continuous function of p, θ, and φ on a bounded, solid region S, the
Coordinates triple integral of f over S is, in spherical coordinates,
∫ ∫ ∫ ∕( p , θ, φ) d V = ∫ ∫ ∫ ∕( p ,
θ, φ) p 2sin φ dp dθ dφ
5 S
EXAM PLE 7 Volume of a sphere
Find the volume of the sphere described by x 2 + y- + z 2 = 9.

Solution It seems clear that we should work in spherical coordinates, because in
spherical coordinates the equation of the sphere is p = 3 for 0 ≤ θ ≤ 2ττ and
0 ≤ φ ≤ 77.
r r r rπ r2 π r3
V= √K = p 2sin φ dp dθ dφ
J J J Jo Jo Jo
s
Γ7T ΓΛ7T ∩P ∣P=3 Γ7r f z ~
-γ sin φ „ dθ dφ = 9 sin φ dθ dφ
∣o Jo ∣p Jo Jo
18τr sin φ dφ = 18ττ(-cos φ) = 36τr
EXAM PLE 8 Moment of inertia using spherical coordinates
A toy top of uniform density is constructed from a portion of a solid

hemisphere with a conical base, as shown in Figure 13.53. The center of the
spherical cap is at the point where the top spins, and the height of the conical
base is equal to its radius. Find the moment of inertia of the top about its axis
of symmetry.
Figure 13.53 Moment of inertia for a spinning top
Solution We use a Cartesian coordinate system in which the z-axis is the axis of
symmetry of the top. Suppose that the cap is part of the hemisphere
882 Ch. 13 M ultiple Integration
z = V Λ 2 - x 2 - y 2. Since the height of the conical base is equal to its radius,

the cone makes an angle of rr/4 radians (450) with the z-axis. Let S' denote the
solid region occupied by the top. Because the top is constructed from material
whose density is uniform, we can assume that the density is constant, say
d = √θ. Then the moment of inertia /_ with respect to the z-axis is given by
'--∣
sΠ<-
y 2) dm =JJs
j (x 2 + y 2)d0 d V
The shape of the top suggests that we convert to spherical coordinates, and we
find that .S’ can be described as
0 ≤ θ ≤ 2τr 0 ≤ Φ≤ f 0 ≤ P ≤ R
Also,
x 2 + p2 = p2sin 2 φ cos2 θ + p2sin 2 φ sin 2 θ = p2sin 2 φ
We now evaluate the integral using spherical coordinates:
ζ = ∫ ∫ ∣ (* 2 + y 2 )d0 d V
s
rπ ∕4 r2 π rR rπ ∕4 r2τr rR
= d0 p2sin 2 φ p2sin<∕> dp dθ dφ = d0 p4sin 3 φ dp dθ dφ
Jo Jo Jo 2 — 2'---------- 777-------- ' Jo Jo Jo
x + y2 dV
rfi
fιr∕4 [∙2π∖ p=R R 5d n fπ∕4
= ⅛ 3
~F sin φ∖ dθ dφ = <• ° (2ττ - 0)sin3 φ dφ
Jo Jo j p υ ^, Jo
2 R 5dn πr ∣ -11∙τr∕4
— 5—-∣^-cos φ + ∣c o s j ≠J o
≈ 0.09728⅛Λ5
PROBLEM SET 1 3 .7
@ 1. ■ What Does This Say? Compare and contrast the In Problems 15-20, convert from spherical coordinates to (a)
rectangular, cylindrical, and spherical coordinate sys rectangular and (b) cylindrical.
tems. 15. (2, f ’ ⅜ ) 16. (3, f ’ f )
2. ■ What Does This Say? Suppose you need to find a
1 7 . ( 1 4 ’ 0) 18. ( 2 4 4 )
particular volume. Discuss some criteria about choosing a
coordinate system. 1 9 .( 1 ,2 ,3 ) 20. (77, 77, 77)
In Problems 3-20, give answers correct to the nearest hun
dredth. Convert each equation in Problems 21-26 to cylindrical coordi
In Problems 3-8, convert from rectangular coordinates to (a) nates and sketch its graph in R j .
cylindrical and (b) spherical. 21. z — x 2 - y 2 22. x 2 - y 2 = 1
3. (1, 1, 1) 4. (0, 4, √ 3 )
23. — ⅛- + z 2 = 0 24. z = x 2 + y 2
5. (√ 2, - 2 , √ 3 ) 6. ( - √ 2 , √ 3 , 2)
7. (1, 2, 3) 25. y 2 + z 2 = 9, 0 < x < 2
In Problems 9-14, convert from cylindrical coordinates to (a) 26. x 2 + y 2 = 9, 0 < z < 4
rectangular and (b) spherical.
9. (3, ¾r > - 3 ) 10. (4, f >- 2 ) Convert each equation in Problems 27-32 to spherical coordi
nates and sketch its graph in R j .
11. ( 2 4 ’ 77) 12. ( - 1 4 ’ 3)
27. z 2 = x 2 + y 2 , z ≥ 0 28. x 2 + y 2 = 1
13. (1, 2, 3) 14. (π, ττ, τr) 29. 2 x 2 + 2y 2 + 2z 2 = 1 30. 4z = x 2 + 3y 2
Sec. 13.7 Cylindrical and Spherical Coordinates 88 3
31. x 2 + y 2 + z 2 = 4 32. x 2 + y 2 —4z2 = 1
Convert each equation in Problems 33-38 to rectangular

coordinates and sketch its graph in R 5.
33. z = r2sin20 34. r = sin θ
35. z = r2cos2θ 36. p2sin 2 ψ = 1
2
37. p sin φ cos φ cos θ = 1 38. p = sin φ cos θ
Evaluate the iterated integral in Problems 39-44.

r, 7r r 2 ∕*V z4 - r 2
Jo Jo Jo
'∙τr∕4 rl rV r
r sin 0 dz dr dθ 48. Use cylindrical coordinates to compute the integral
J J J z(x 2 + y 2) 1/2 dx dy dz
Jo
-π ∕2
Jo
r2π c2
Jo
r2sin θ dz dr dθ s
where >8’ is the solid bounded above by the plane z = 2 and
below by the surface 2z = x 2 + y 2 .
Jo
Λ TTI2
Jo
rττ!4
Jo
cos φ sin φ dp dθ dφ
rcos φ
p2sin φ dp dθ dφ
Jo
∙2τr
Jo Jo
∕*4 rl
∣
Jo
rπ ∕3
Jo
zr dz dr dθ
Jo
rs in θ r 4 cos θ
44. r dz dr dθ
J - π ∕4 Jo Jo
45. The point (x, y, z) lies on an ellipsoid if 49. Find the center of mass of the solid bounded by the
surface z = V x 2 + y 2 and the plane z = 9. Assume
x = ar sin φ cos θ
y = br sin φ sin θ P = 1∙
z = cr cos φ
Find an equation for this ellipsoid in rectangular coordi
nates if r is a constant.
46. Use cylindrical coordinates to compute the integral
Iff xy dx dy dz
where R is the cylindrical solid x 2 + y 2 ≤ 1 with

0 ≤ z ≤ 1.
Evaluate J J J V x 2 + y 2 + z 2 dx dy
50. dz, where R is
R
defined by x 2 + y 2 + z 2 ≤ 2.
Evaluate J J J (x 2 + y 2 + z 2) dx dy
51. dz, where R is
R
defined by x 2 + y 2 + z 2 ≤ 2.
47. Use cylindrical coordinates to compute the integral 52. Evaluate J J J z 2 dx dy dz, where 5 is the hemisphere
s
ifRf (x 4 + l x 2y 2 + y 4 ) dx dy dz x 2 + y 2 + z 2 ≤ 1, z ≥ 0.
f f f dx dy dz w
where R is the cylindrical solid 53. Evaluate I I I y -> + ? + ^ here .8’ is the sphere
s
x 2 + y 2 ≤ a2 with 0 ≤ z ≤ —
77 x 2 + y 2 + z 2 ≤ 3.
Find the volume o f the solid S given in Problems 54-59 by

the force exerted on m, and the vertical compo
using integration in any convenient system o f coordinates.
nent, are
54. .S’ is bounded by the surface z = 1 - 4(x 2 + j 2) and the xy- _ Gm(δΔK)
plane. n1ag p2
55. S is bounded above by the paraboloid z = 4 - (x 2 + y 2),
below by the plane z = 0, and laterally by the cylinder Gm(δΔV)(R - pcos φ)
LΛ Γ ∙J
x 2 + y 2 = I. P3
Note that the term (R - p cos φ)∕p projects the
56. S is the intersection of the solid sphere x 2 + y 2 + z 2 ≤ 9
and the solid cylinder x 2 + y 2 ≤ 1. force vector onto the z-axis, giving the vertical
component. Do you see why?
57. .S' is the region bounded laterally by the cylinder r = 2
c. Form the Riemann sum involving Δ F and get the
sin θ, below by the plane z = 0, and above by the parabo
loid z = 4 - r2 . ; following for the force acting between point mass
m and the sphere o f radius a:
58. S' is the region that remains in the spherical solid p ≤ 4
after the solid cone φ ≤ f has been removed. 7e
, f f ΓGmδ(R - P cos φ)dV
jJ J F
59. SPY PR O B L EM The spy crawls through the hole in the
floor of the burning bunker (Problem 55, Section 12.6), _ 2OπG
/-■mδs jΓ
° JΓ
θ p 2∣ (R ~ P cos φ) sin φ
and finds himself in a small beehive-shaped cave, sur r 2 + p 2 _ 2 R p c o s φ y l2 dφ dp
rounded by the terrorists. Gasping for breath, he looks at
his captors and is shocked to find that their leader is none Note: The integrand is independent o f θ.
other than his superior’s daughter, Purity. A terrorist 61. a. For an unspecified radius a and R > a, compute
raises his gun, but Purity orders, “ No! Tie him securely. the attractive force between the masses in the
I want him to think for awhile before he dies. And previous problem using the center of mass calcula
hurry— Scelerat is waiting for us and he is not a patient tion.
man.” The spy is tied and the terrorists leave. Purity is the b. Set a and R to numerical values and compute the
last to go, and just before leaving, pulls a lever by the attractive force by the integral in part c of Prob
door, and the cave begins to flood with water. The spy is lem 60. Try two or three different values of R until
able to pull himself into a sitting position, so that his nose you are satisfied that the center of mass “ approxi
is 4 ft above the floor. In desperation, the spy does some mation” is exact in this case.
quick calculations and decides that the beehive shape c. Compare this result with the corresponding calcu
of the cave can be described as the part of the surface lations of Section 13.6 and conjecture why the
p = 4(1 + cos≠), where p is measured in feet, that lies center of mass argument is sound in the case o f the
above the .∖ψ-plane. If the spy needs at least 10 min to free sphere and not for the rectangular region.
himself, and if water is entering at the rate of 25 ft 3∕m in , . ............ .. - .......
will he drown or will he survive?
θ 62. Let S be a homogeneous solid with density 1 that has the
shape of a right circular cylinder with height h and radius
COMPUTATIONAL WINDOW R . Use cylindrical coordinates to find the moment of
inertia o f 5, about its axis o f symmetry.
60. In the last section (Problems 51 and 52), we said that 63. Find the sum I χ + Z + I_ of the moments of inertia of the
it is not always accurate to compute the attractive solid ellipsoid
force between two masses by simply using the dis
x 2 + y 2 + z2 ≤ 1
tance between their centers o f mass. This approxima
tion was done first in astronomy in which the bodies about the coordinate axes. Assume the density in 1.
were typically spherical in shape. As we shall see, in 64. How much volume remains from a spherical ball of
this case the simple calculation is very appropriate. radius a when a cylindrical hole of radius b (0 < b < a)
Consider a sphere of radius a with density δ centered is bored out o f its center?
at the origin; and a point mass m located at (x, y, z) =
(0, 0, R). In this problem, you are to set up the
integral to compute the attractive force between the
masses; in the next problem, you will do some
computing.
a. Show that the distance p between point (x, y, z) =
(0, 0, R) and a point in the sphere, (p, θ, φ), can be
expressed p 2 = R 2 + p2 - 2Rp cos φ. Hint: Use
the law of cosines.
b. In the sphere, consider a small element of volume
Δ K a t point (p, θ, φ). Argue that the magnitude of
Sec. 13.8 Jacobians: Change of Variables 885
1 3 .8 JACOBIANS: CHANGE OF VARIABLES
IN THIS SECTION Change of variables in a double integral, change of

variables in a triple integral
We have seen how a change of variables to polar coordinates often simplifies the
evaluation of a double integral and, likewise, how cylindrical and spherical
coordinates can be used in evaluating triple integrals. However, these are not
the only substitutions that may be used effectively, and the goal of this section is
to examine some general ideas about changing variables in a multiple integral.
■ CHANGE OF VARIABLES IN A DOUBLE INTEGRAL
The German mathematician Carl Gustav Jacobi (see Historical Note, p. 886) was
one of the first to work with change of variables. In honor of his early work, we
define, in the following theorem, an important tool called a Jacobian, which
makes it easy for us to change variables when working with multiple integrals.
For a single variable, the change of variables is called integration by substitu
tion. When we carry out the process of substitution for a single integral, we see that
the change of variable x = g(u) introduces a new factor into the integrand:
f f(x ) d x = [ f[g(u)]g'(u) du
Ja Jc
where the limits o f integration change so that a = g(c) and b = g(d).

For a change of variables in a double integral, we could expect to transform (or
“ map” ) a region D onto a region D *, with the goal of having D* be simpler than D.
Because we know that changing variables with a single variable introduces a new
factor (namely, g'(w)), by analogy we would expect the change of variable with a
double integral to also introduce a “ mapping factor.” The following theorem
shows that this is indeed the case.
T H E O R E M 13.7 Change of variables in a double integral
Let f be a continuous function on a region D , and let T be a one-to-one

transformation that maps the region D in the wv-plane onto a region D * in the xy-
plane under the change of variable x = g(μ, v), y = h(u, v), where g and h are
continuously differentiable on D *. Then
/[£•(«, v), h(u, v)] ∣J(M, V)∣ du dv
where
∂x ∂x
∂u ∂v _ ∂x ∂y _ fry ∂x
J(u, v) = ∂y ∂y ~ ∂u ∂v ∂u ∂v
∂u ∂v
J(u, v) is a mapping factor called the Jacobian for mapping from x , y variables to
u, v variables. The Jacobian is also denoted by ^ ∣. We assume J(μ, v) ≠ 0.
Proof: A formal proof is a matter for advanced calculus, but a geometric

argument for this theorem is presented in Appendix B. ■
EXAMPLE 1 Finding the Jacobian

Find the Jacobian for the change of variables from rectangular to polar
coordinates, namely, x = r cos θ and y = r sin θ.
Solution In this example, we are changing rectangular form (x, yj to polar form
(r, θ), so we write ∂(x, T)
d(r, ΘY
∂x dx
∂(x, J?) _ dr dθ cos θ - r sin 0
= r cos2 0 + r sin2 0 = r
a(r, 0) - ⅛ ∂y sin θ r cos 0
Historical Note ∂r ∂θ
This result justifies the formula we have previously used:
J If(x , y) dy dx = J J ∕( r cos 0, r sin 0) r dr dθ

*
JD J
where D is the region in the (r, 0) plane that is mapped into D* in the (x, y) plane of
the polar transform x = r cos 0, y = r sin 0, as illustrated in Figure 13.53.
CARL G. J ACOBI (1804-1851)

Carl G. Jacobi is usually re
ferred to as C.G. because his
Figure 13.54 Polar transformation of a region D by T: x = r cos θ, y = r sin θ
brother, Moritz H. Jacobi
(1801-1874), was also famous
in the 1840s. In fact, M. H. was EXAMPLE 2 Calculating a double integral by changing variables
known for something he called
“galvanoplastics,” which turned Compute f f dy dx, where D* is the triangular region bounded by
out to be a fraud, and because
C. G. was “only” a mathemati
JD*
cian, he was not well known the line x + y = 1 and the coordinate axes.
outside of mathematical circles.
Solution This is a rather difficult computation if no substitution is made. We
C. G. Jacobi was a gifted stu
dent and had a wide range of
wish to make a substitution that simplifies the integrand while leaving the
talents, and during his lifetime region of integration relatively uncomplicated. Keep in mind that changing
he distinguished himself as one the region from D to D* is the same as changing the limits of integration on a
of Germany’s greatest mathe single integral.
maticians. When he was solving Let u = x - y and v = x + y, so that the integrand simplifies to (u∕v)4 . Let
the problem of finding the cir D* be the interior of a triangle in the xj-plane with vertices (0, 0), (0, 1), (1,0),
cumference of an ellipse, he be as shown in Figure 13.55a. To determine the effect of this transformation on
came interested in integration. the region of integration D*, we need to solve the system of equations for x
The determinant used in the and y in terms of u and v:
change of variable formula de
veloped in this chapter is x = ∣( M + v) y =∣
(v - u)
named in honor of C. G.
Now transform each boundary of D* to the variables u and v:
Boundary Substitute Simplify
x =0 ⅛(u + v) = 0 v = -u
y= 0 j(v - u) = 0 v= u
x + y= 1 ∣( M + v) + ∣(v - w) = 1 v= 1
We draw the simplified boundary on a wv-axis, as shown in Figure 13.55b, and

a. The region of integration D* then note that D is the triangular region with vertices (0, 0), (1, 1), and
( - 1, 1) in the uv-plane.
D can be described in type II form (fixed v) as the set of all (w, V) such that
for each v between 0 and 1, u varies from the line u = —v to the line u = v.
The last step before setting up the integral is to compute the Jacobian of
the change of variables (x, y) to (M, V):
∂ ( u + v∖ _d_( u + v∖ 1 1
∂(x, y) ∂u∖ 2 / θv∖ 2 / 2 2 _ 2
∂(u, v) _d_/ v ~ u∖ JL ( V ~ U∖
∂u∖ 2 / ∂v∖ 2 /
b. The transformed region, D
Thus,
Figure 13.55 Transformation of
D* to D
In Example 2, the change of variables was chosen in order to simplify the

integrand, but it is also useful to introduce a change of variables that simplifies the
region of integration.
EXAM PLE 3 Change of variable to simplify a region
r 2 υ2
Find the area of the region E bounded by the ellipse + p = 1.
Solution The area is given by the integral
where E is the region shown in Figure 13.56a.
a. The elliptical region E b. The transformed region C is a

circular disk.
Figure 13.56 Transformation o f the ellipse E to the circle C
We will transform the elliptical region E to one that is circular. Because E can
be represented by
we consider the substitution u = -x and v = yy which will map the elliptical

region E onto the circular disk C, where
C .u 2 + v2 ≤ 1
as shown in Figure 13.56b. Solving the two equations for x and y, we obtain
x = an, y = bv, the Jacobian gives us the mapping (change of variable) factor:
∂(x, .V) ∂ u ^ ∂ v^ Ia 0
= ab
∂(u, v) 4~(bv) ∕-(⅛ ) I0 b
∂u' ’ ∂vv ’
Because ab > 0, we have ∖ab∖ = ab, and the area of E is given by
ab du dv = ab du dv = ab [τr( 1)2] = πab

κ__ v__ /
Because C is a circle of radius 1
■ CHANGE OF VARIABLES IN A TRIPLE INTEGRAL
The change of variable formula for triple integrals is similar to the one given for
double integrals. Let Tbe a change of variable that maps a region R* in xyz-space
onto a region R in z∕vιv-space, where
T: x = x(u, v, ιv) y = y(ιι, v, w) z = z(u, v, w)
Then the Jacobian of T is the determinant
∂x ∂x ∂x
∂u ∂v ∂W
∂(x, y, z) ∂y ∂y ∂y
∂(u, v, w) ∂u ∂v ∂w
∂z ∂z ∂z
∂u ∂v c)w
and the change of variable yields
∂(x, y, z)
∫ ∫ ∫ ∕⅛ r , z)dx dy dz = ∂(u, v, n,) du dv dw
R∙
EXAMPLE 4 Formula for integrating with spherical coordinates

Find the formula for integrating with spherical coordinates.
Solution The conversion formulas from rectangular coordinates to spherical
coordinates are:
x = ρ sin φ cos θ y = p sin φ sin θ z = p cos φ

The Jacobian of this transformation is
⅜(p sin φ cos θ) ~^(p sin φ cos θ) ^τ(p sin √> cos 0)
∂(x, y , z)
^ ( p sin φ sin θ) ^ ( p sin <∕> sin 0) θ^(p sin φ sin θ)
∂(p, θ, φ)
'∖
-QP (P cos φ) ⅛ (pcos≠) ^ (p c o s ψ )
sin φ cos θ - p s in ≠ s in 0 p c o s≠ c o s0
sin φ sin θ p sin φ cos θ p cos φ sin θ
cos φ 0 - p sin φ
After many steps, we can evaluate this determinant as - p 2sin≠. Because
0 ≤ φ ≤ τr, we have sin<∕> ≥ 0 and p2 ≥ 0 so that ∣-p 2sinφ∖= p2sinφ.
/ / / ^ x , dX cf y = / I '^ P S 'n C 0 S θ , P S 'n $ S 'n P C 0 S P2s*n d p d θ dΦ
R* r C=∑l
PROBLEM SET 1 3 .8
θ Find the Jacobian o f the change o f variables given in Problems 18. Suppose a wv-plane is mapped onto an xy-plane by the
1-12. equations x = u 2 — v2, y = 2uv. Express dx dy in terms of
1. x = u + v, y = uv 2. x = u 2, y = u + v du dv.
3. x = u - v, y = u + v 4. x = u 2 — v2, y = 2 MV
Let D be the region in the xy-plane that is bounded by the
5. x = u 2v2 , y = v2 - u 2 6. x = ιι cos v, y = u sin v coordinate axes and the line x + y = 1. Use the change o f
+v
7. x = e" , y = e ~ ll v 8. x = e “ sin v, y = e ,' cos v variable u = x - y , v = x + y to compute the integrals given in
9. x = u + v - w; y = 2ιι - v + 3iv; z = —u + 2v — w Problems 19-22.
10. x = 2u - w, y = u + 3v, - = v + 2w
11. x = u cos v, y = it sin v, : = weuv
1,∙ ∫ ∫ (τ+y)i * dx
D'
12. w
u 20. [ [ dydx
J J ∖χ - y )
D if
21. j J (x - y) 5(x + y)3 dy dx

JD J∙
22. j j (x — y)e x ~+,'2 dy dx
θ 23. ■ What Does This Say? Discuss the process of changing

variables in a triple integral.
24. ■ What Does This Say? Discuss the necessity of using
the Jacobian when changing variables in a double integral.
Under the transformation
u = ∣(2x + y) v = ∣(x - 2_v)
the square S in the xy-plane with vertices (0, 0), (1, -2), (2, 1),
(3, —1) is mapped onto the square in the uv-plane that is
bounded by the positive coordinate axes and the lines u = 1 and
v = 1. Use this information to find the integrals in Problems
25-30.
25. f [ f ~λ <+ '* ) dy dx
J J ∖χ - 2 y + 5j -
17. Suppose a wv-plane is mapped onto an xy-plane by the
equations x = u( 1 - v), y = uv. Express dx dy in terms of ∣j (2x + y)(x - 2y)2 dy dx
du dv. sj
∂(x, y)
27. a. Compute the Jacobian — ÷r.
3(w, v)
5
b. Let E denote the ellipse
28. - 2y) dy dx x 2 + xy + y 2 = 3
Use a rotation of J to obtain an integral that is equivalent

29.
to ∫ ∫ y dy dx. Evaluate the transformed integral.
30. cos (2x + y) sin (x - 2y) dy dx
.S' 36. T H IN K T A N K P RO BLEM Let R be the region bounded by
31. Let R be the region in the xy-plane that is bounded by the the parallelogram with vertices (0, 0), (2, 0), (1, 1), and
parallelogram with vertices (0, 0), (1, 1), (2, 1) and (1, 0). (—1, 1)∙ Show that under a suitable change of variable of
Use the linear transformation x = u + v, y = v to com the form x = au + bv, y = cιι + dv, we have
pute J J (2x —y) dy dx.
R J J ∕( x , y) dy dx = jθ f(t) dt
32. Use a suitable linear transformation u = ax + by, J J
R
v = rx + sy to evaluate the integral
where/is continuous on [0, 2].
e (y x , '2 dy dx 37. Find the Jacobian of the cylindrical coordinate
transformation:
where R is the region inside the parallelogram with x = r cos θ, y = r sin θ, z = z

vertices (0, 0), (1, 1), (—1, 1), and (0, 2). 38. Use the change of variable
33. Under the change of variables x = s 2 - t2 , y = 1st, the
quarter circular region in the St-plane given by x = au, y = bv, z = cw
s 2 + t2 ≤ 1, s ≥ 0, t ≥ 0 is mapped onto a certain
to find the volume of the ellipsoid
region .8’ of the xy-plane. Evaluate
dy dx
√ x2 + y2
39. Evaluate J J l n ^ ∣ - dy dx, where R is the triangular
34. Use x = ar cos θ ,y = br sin θ to evaluate R J
region with vertices (1, 0), (4, —3), and (4, 1).
© 40. Show that if ad - be ≠ 0, the linear transformation
u = ax + by and v = ex + dy
where .S’ is the quarter ellipse
can be solved for x and y to obtain
y-2 y 2
2⅛ + ⅛ ≤ l, x≥ 0 , v≥ 0 du ~ bv , = av - cu
a- b~ ad - be - ad - be
35. A rotation o f the xy-plane through the fixed angle θ is 41. T H IN K T A N K P RO BLEM Find the volume of the solid
given by under the surface
x = u cos θ — v sin 6; y = u sin θ + v cos θ
over the region bounded by xy = 1, xy = 5, x = 1, and

x = 5. Hint: Use a change o f variables of the form x = u,
x y = v. See Figure 13.58.
Figure 13.57 Rotational transformation Figure 13.58 Problem 4 1

C H A P T E R 1 3 R E V IE W
IN T H IS R E V IE W Proficiency examination; supplementary problems

C o n c e p t P r o b le m s (r, θ, z) X= 1 r=r P=J_

1. Define a double integral. Cylindrical y = JL, θ= θ θ= ?
2. State Fubini’s theorem over a rectangular region. z— 1 z=z Φ= J ~
3. What is a type I region? State Fubini’s theorem for a
type I region. (p, θ, φ) x= 1 r= ? P= P
4. What is a type II region? State Fubini’s theorem for a Spherical y = _JL θ= ? θ= θ
type II region. z= ? z= ? Φ= Φ
5. State the formula for finding an area using a double
integral. 20. a. State the formula for triple integral in cylindrical
6. State the formula for finding a volume using a double coordinates.
integral. b. State the formula in spherical coordinates.
7. State the following properties of a double integral: 21. Define the Jacobian for the change of variables from xy-
a. linearity rule b. dominance rule coordinates to uv-coordinates, where x = g(u, v),
y = h(u, v).
c. subdivision rule
22. State the formula for the change of variables in a double
8. What is the formula for double integral in polar coordi
integral.
nates?
9. Explain how to evaluate an improper integral in polar P r a c t ic e P r o b le m s
coordinates. "π ∕3 rsin y
10.
11.
State the double integral formula for surface area.
State the formula for surface area defined by a vector
J o
rl
Jo
rz
e - x cosy dx dy.
ry - z
function. 24. Evaluate (x + y - z) dx dy dz.
J-ι Jo Jy
12. State Fubini’s theorem over a parallelepiped in space.
13. Give a triple integral formula for volume. 25. Use a double integral to compute the area of the region
A that is bounded by the x-axis and the parabola
14. Explain how to use a double integral to find the mass of
y = 9 - x 2.
a planar lamina of variable density.
26. Use polar coordinates to evaluate
15. How do you find the moment of mass with respect to the
x-axis? cos (x 2 + y 2) dy dx
16.
17.
What are the formulas for the centroid of a region?
Define moment of inertia. 27.
∏
Find the surface area of the portion of the sphere
18. What is a joint probability density function? x 2 + y2 + z 2 = 4 that lies above the xy-plane.
19. Complete the following table for the conversion of
coordinates.
(x, y, z) (r, θ, z} (p, θ, φ)
To Rectangular Cylindrical Spherical
From
Cw y, z) X =X r= ? P= JL,
Rectangular y =y tan θ = ? tan θ = ?
z =z z= ? Φ= J ~
Figure 13.59 Problem 28, page 892
28. Set up and evaluate a triple integral for the volume of 30. Use a linear change of variables to evaluate the double
the solid region in the first octant that is bounded above integral
by the plane z = 4x and on the sides by the paraboloid
z = x 2 + 2y2. See Figure 13.59, p. 891. (x + y)ex ~ 2y dy dx
29. A solid >S'is bounded above by the plane z = 4 and below R
by the surface z = x 2 + y 2. Its density p(x, y, z) at each
point P is equal to the distance of P from the z-axis. Find where R is the triangular region with vertices (0, 0),
the total mass of A. (2, 0),(1, 1).
In Problems 1-10, sketch the region of integration, exchange

the order, and evaluate the integral using either order of 21. -2 + y 2 + 1) dA, where D is the circular disk
integration. D
1. f Jθ3 ' x 2y 2 dy dx
22. ∣J eχ2+y2 dA, where D is the circular disk x 2 + y 2
~σπ∕2 rπ∕2 DJ
J 0
ι*π∕2
Jo
rVsin x
cos (x + y) dx dy
23.
D
dA, where D is the circular disk
J o
;
Jo
xy dy dx
24.
A 4- 1,2
dA, where D is the region above the line x + y = 1

D
and below the circle x 2 + y 2 = 1
"l rl
Jo
∣
Jx 2
x 3 sin y 3 dy dx
∙2 ∕∙√4 - x 2 ____________
25.
D
—y dA, where D is the circular disk
10. V4 - x 2 - y 2 dy dx
Jo Jo
26. - I dy dx, where D is the triangular region
Evaluate the integrals given in Problems 11-38. D
with vertices (0, 0), (2, 0), (0, 2) Hint'. Make a suitable
11. f [ ey ' dy dx 12. [ f ., v , dy dx change of variables.
Jo k Jo I - (* 2 + F2)3Z2
4 ∣
∙y 27. is the shaded portion of the figure
∏
14. J J J
z dz dy dx
^ 2 2
D
e Vχ +v ljy (/x cjz
∙π∕4 r2π rθ
Jo Jo Jo
16. f f [ x 2y dz dy dx
r2 sin φ dr dθ dφ
17. f f f xe z dz dy dx
Jo Jo Jo Jl Jx Jo
∏ l +x rxy
28. D is the shaded portion of the figure
xz dz dy dx D
- χ Jo
∕∙√3 ∣
∙V3 - y 2 ∣
∙9
19. y 2 dz dx dy
J-V 3 J -V 3 - y 2 J(Λ ≈ + y 2) 2
Hint'. Change to cylindrical coordinates.

20. J J x 2y dA, where D is the circular disk x 2 + y 2 ≤ 4
Chapter 13 Review 89 3
42. Express the integral

29. I I -----τ -> a > 0, m > 1, where P is the
J J (a 2 + x 2 + y 2)m y H r(4 - y)∣
3
p
entire xp-plane. Hint'. Note this is an improper double
integral.
∏ f ( x , v) dx dy + I I f( x , y) dx dy
as a double integral with the order of integration reversed.

30. J J e -(⅛l+*Lv∣) ⅛4j a > 0, b > 0, where P is the entire 43. Express the integral
Vy ι∙3 ∣
∙2 - y
xy-plane. Hint: Use Cartesian coordinates and note this

is an improper double integral.
32∕'
Π f ( x ,y ) d x d y + j f ( x , .y) dx dy
as a double integral with the order o f integration reversed.
∏ R ,
dx dy, where R is the rectangle a ≤ x ≤ b,
∂2f
44. Find numbers a, b, and c so that
c ≤ zy ≤ d and . , is continuous over R with exp[-(αx 2 + bxy + cy 2)] dA = 1

∂x∂y
f( a , c) = 4 ,∕( α , d) = - 3 ,f ( b , c) = 1,∕(⅛, d) = 5
Hint: Use a change of variables o f the general form
32. V x 2 + y 2 + z 2 dV, where S is the portion o f the w = A x + By, v = C x + Dy.
5 45. Find the area inside the circle r = cos 0 and outside the
solid sphere x 2 + y 2 + z 2 ≤ 1 that lies in the first octant cardioid r = 1 — cos 0.
46. Find the area outside the cardioid r = 1 + cos 0 and
33.
∕∫ ∕
s
(x 2 + y 2 + z 2) dV, where S is the portion of the
47.
inside the cardioid r = 1 + sin 0.
Find the area outside the small loop of r = 1 + 2 sin 0 and
solid sphere x 2 + y 2 + z 2 ≤ 1 that lies in the first octant inside the large loop.
34 . J J J z 2 dV, where H 48. Use a double integral to compute the volume of the
is the solid hemisphere
tetrahedral region in the first octant that is bounded by
H the coordinate planes and the plane 3x + y + 2z = 6.
x 2 + y 2 + z 2 ≤ 1 with z ≥ 0 49. Find the volume of the tetrahedron that is bounded by the
/ / f V x 2 + y 2 + z2 coordinate planes and the plane
35. where H is the solid hemisphere
-a + ^b + -c = 1
x 2 + y 2 + z 2 ≤ 1, with z ≥ 0
with a > 0, b > 0, c > 0.
36. where .S' is the solid region bounded 50. Find the volume o f the solid that is bounded above by the
s paraboloid z = x 2 + y 2 and below by the square region
on the sides by the paraboloid 2z = x 2 + y 2 , and above by 0 ≤ x ≤ 1, 0 ≤ y ≤ 1.
the sphere x 2 + y 2 + z 2 = 8 51. Find the volume o f the solid that is bounded above by the
surface z = x y , below by the xy-plane, and on the sides by
37. z(x 2 + y 2)~ l ' 2 dV , where 5, is the solid bounded by the part of the circular disk x 2 + y 2 ≤ a 2 (α > 0) that
s lies in the first quadrant.
the surface 2z = x 2 + y 2 and the plane z = 2 52. Find the volume of the solid that is bounded above by
the paraboloid z = 4 - x 2 - y 2 and below by the plane
38.
/// (x 4 + 2x 2y 2 + y 4) dV, where 5 is the solid cylinder
z = 4 - 2x.
53. Find the surface area of that portion of the surface
given by x 2 + y 2 ≤ a 2, 0 ≤ z ≤ — z = V x 2 + y 2 that is contained in the cylinder x 2 + y 2 = 1.
54. Find the surface area o f the portion o f the paraboloid
39. Rewrite the triple integral f f f f ( x , y, z) dz dy dx as z = x 2 + y 2 that lies below the plane z = 9.
Jo Jo Jo 55. Find the surface area of the portion o f the cylinder
a triple integral in the order dy dx dz. x 2 + z 2 = 4 that is bounded by the planes x = 0, x = 1,
40. Convert the double integral y = 0, and y = 2.
^ττ∕2 r2a cos θ
56. Find the volume of the region bounded by the surface
J 0 Jo
to rectangular coordinates and evaluate.
r sin 20 dr dθ y 2 + z 2 = 2x and the plane x + y = 1.
57. Find the volume of the region bounded above by the
paraboloid z = 4 - x 2 —y 2 and below by the plane
41. Evaluate by reversing the order of integration
z + 2x = 4.
J ∣ eχ P 2 d x d y 58. Find the volume of the region bounded by the ellipsoid
z 2 = 4 - 4r2 and the cylinder r = cos 0.
59. Find the volume of the region bounded above by the 73. Use a change of variables after a transformation of the
hemisphere z = V9 —x 2 —y 2, below by the xy-plane, and form x = Au, y = Bv, z = Cw to find the volume of the
on the sides by the cylinder x 2 + y 2 = 1. region bounded by the surface V x + Vz2y + V3z = 1 and
60. Show that the solid bounded below by the cone the coordinate planes.
z = Vx 2 + y 2 and above by the sphere x 2 + y 2 + z 2 = 2az 74. Find the centroid of a homogeneous lamina (ρ = 1) that
for a > 0 has volume V = πa 3. covers the part of the plane Ax + By + Cz = 1, A > 0,
61. Find the volume of the solid region bounded above by the B > 0, C > 0 in the first quadrant.
sphere given in spherical coordinates by ρ = a (α > 0) 75. A homogeneous plate (p = 1) has the shape of the region
and below by the cone φ = φ0 where 0 < φ0 < J. in the first quadrant of the xy-plane that is bounded by
62. A homogeneous solid 5, is bounded above by the sphere the circle x 2 + y 2 = 1 and the lines y = x and x = 0.
p = a (a > 0) and below by the cone φ = φ0, where Sketch the region and find x, the x-coordinate of the
9 < φ0 < y. Find the moment of inertia of 5 about the centroid.
z-axis. 76. Let R be a lamina covering the region in the xy-plane that
63. The region between the circles x 2 + y 2 = 1 and is bounded by the parabola y = 1 - x 2 and the positive
x 2 + y 2 = 4 with 0 ≤ z ≤ 5 form a “washer.” Find the coordinate axes. Assume that the lamina has density
moment of inertia of the washer about the z-axis if the p(x, y) = xy.
density p is given by a. Find the center of mass of the lamina.
a. p = k (k a constant) b. p(x, y) = x 2 + y 2 b. Find the moment of inertia of the plate about the z-
c. ρ(x, y) = x 2y 2 axis.
64. Find x, the ^-coordinate of the center of mass, of a 77. A solid has the shape of the cylinder x 2 + y 2 ≤ a,
triangular lamina with vertices (0, 0), (1, 0), and (0, 1) if with 0 ≤ z ≤ b. Assume the solid has density
the density is p(x, y) = x 2 + y 2 . p(x, y, z) = x 2 + y 2.
65. Find the mass of a cone of top radius R and height H if the a. Find the mass of the solid.
density at each point P is proportional to the distance of P b. Find the center of mass of the solid.
from the tip of the cone.
78. Use cylindrical coordinates to find the volume of the solid
66. Find the centroid of a homogeneous (p = 1) right circular
bounded by the circular cylinder r = 2a cos θ(a > 0), the
cone with height PI and base radius R.
cone z = r, and the xy-plane.
67. Find the Jacobian ¾ - , V’-~- of the change of variables
θ(x, y, z) 79. Let u be everywhere continuous in the plane, and define
u = 2x —3y + z, v = 2y —z, w = 2z. the functions f and g by
∂(ll, V, w)
68. Find the Jacobian τ ~ ——r of the change b of variables ∕(x ) = w(x, y) dy g(y) = w(x, y) dx
∂(x, y, z)
2 2 2 2 2 2 Jy
u = x + y + z , v = 2y + z , w = 2z .
69. Let u = 2x - y and v = x + 2y. Find the image of the unit Show that
square given by 0 ≤ x ≤ I, 0 ≤ y ≤ 1.
rb rb
70. Find the image under the linear transformation
u = x - y, v = y for the lines of the general form
y = Cx + 1, for constant C.
J a
f(x )d x =
Ja
g{y)dy
71. Suppose u = j(x 2 + y 2) and v = j(x 2 —y 2), with x > 0, Hint'. Show that both integrals equal w(x, y) dA for a
y > 0. certain region D. -'d ∙2
a. Find the Jacobian 77— 7. 80. Find the volume of the solid region bounded by the
<9(x, y) surface
b. Solve for x and y in terms of u and v and find the
2∕3 2∕3 2∕3 2∕3
Jacobian χ + y + z = α
∂(x, y)
∂(u, v) where a is a positive constant.
., .. , ∂(u, v) ∂(x, y) 81. Let f be a probability density function on a closed set R.
c. Verify that 77----r — H- = 1 Recall that a probability density function has the proper
J ∂(X , y) ∂(u, v)
72. Let D be the region given by u ≥ 0, v ≥ 0, ties:
1 ≤ u + v ≤ 2. ∕( x , y) > 0 for all (x, y) in R
a. Show that under the transformation u = x + y,
v = x - y, D is mapped onto the region R such that f f(x , y) dA = 1
Show that the following functions are probability density

D functions:
√-∕ ∖ ,/ < if (∙v*> Fz)7 is i∏ and √4v(R)' is the 87. Find the center o f mass of the solid .8' that lies inside the
a. ∕ (x, y) = ∙! A (nR ) sphere x 2 + y 2 + z 2 = 1 in the first octant x ≥ 0, y ≥ 0,
lθ area of R if (x, j j is not in R z ≥ 0 if the density is ρ(x, y, z) = (x 2 + y 2 + z 2 + I)” 1.
b. f ( x , y) = λμ exp (-λx - μy) for λ > 0, μ > 0 and 88. Suppose we drill a square hole of side c through the center
(Λ^, j j is in the first quadrant. of a sphere o f radius c, as shown in Figure 13.61. What is
82. A cube of side 2 is surmounted by a hemisphere of radius the volume of the solid that remains? Hint. Use spherical
1, as shown in Figure 13.60. Suppose the center of coordinates.
coordinates is at the center of the hemispherical dome. If
the combined solid has density p = 1, where is the center
o f mass?
Figure 13.61 Sphere with a square hole drilled out
89. P U T N A M E X A M IN A T IO N P RO BLEM A circle of radius a is

revolved through 180o about a line in its plane, distant b
from the center of the circle (⅛ > a). For what value of
the ratio b/a does the center of gravity o f the solid thus
Figure 13.60 Center o f mass o f a hemisphere dome generated lie on the surface o f the solid?
90. P U T N A M E X A M IN A T IO N P RO BLEM For ∕( x ) a positive,
83. Set up (but do not evaluate) an integral for the mass o f a
monotone, decreasing function defined in 0 ≤ x ≤ 1,
lamina that covers the region outside the circle r = ∖∏ a
prove that
and inside the lemniscate r2 = 4α2 sin 20.
84. Set up (but do not evaluate) an integral for the mass of a
lamina that covers the region enclosed by the rose
r = cos 30.
85. Find the centroid of the solid common to the cylindrical
solids x 2 + z 2 ≤ 1 and y 2 + z 2 ≤ 1.
86. Show that if z = ∕( r , 0) is the equation o f a surface S' in
polar coordinates, the surface area of S is given by 91. P U T N A M E X A M IN A T IO N P RO BLEM Show that the integral
equation
∕(-V, F) = 1 + f(μ , v) du dv
has at most one solution continuous for 0 ≤ x ≤
where R is the projected region in the r0-plane. 0 ≤ y ≤ 1.
"Space-Capsule Design”
Suppose you are transported back in time to be part of a team of engineers

designing the Apollo space-capsule. The capsule is composed of two parts:
1. A cone with a height of 4 m and a base of radius 3 m.

2. A reentry shield in the shape of a parabola revolved about the axis of the
cone, which is attached to the cone along the edge of the base of the cone.
Its vertex is a distance D below the base of the cone. Find values of the
design parameters D and p so that the capsule will float with the vertex of
the cone pointing up and with the waterline 2 m below the top of the cone,
in order to keep the exit port ∣ m above water.
You can make the following assumptions:
a. The capsule has uniform density p.

b. The center of mass of the capsule should be below the center of mass of the
displaced water because this will give the capsule better stability in heavy
seas.
c. A body floats in a fluid at the level at which the weight of the displaced
The science o f mathematics has
fluid equals the weight of the body (Archimedes’ principle).
grown to such vast proportion that
probably no living mathematician
Y o u r paper is not limited to the following questions but should include these
can claim to have achieved its
mastery as a whole. concerns: Show the project director that the task is impossible; that is, there are no
A. N. Whitehead values of D and p that satisfy the design specifications. However, you can solve
An Introduction to Mathematics this dilemma by incorporating a flotation collar in the shape of a torus. The collar
(New York, 1911), p. 252 will be made by taking hollow plastic tubing with a circular cross section of radius
1 m and wrapping it in a circular ring about the capsule, so that it fits snugly. The
collar is designed to float just submerged with its top tangent to the surface of the
water. Show that this flotation collar makes the capsule plus collar assembly
Consider the mathematics dis satisfy the design specifications. Find the density p needed to make the capsule
covered from 1911 to 1995! float at the 2-meter mark. Assume the weight of the tubing is negligible compared
How many times multiplied is to the weight of the capsule, that the design parameter D is equal to 1 meter, and
the quotation of Whitehead? the density of the water is 1.
’MAA Notes 17 (1991), “Priming the Calculus Pump: Innovations and Resources,” by Marcus S.
Cohen, Edward D. Gaughan, R. Arthur Knoebel, Douglas S. Kurtz, and David J. Pengelley.
14
Vector Analysis
■ CONTENTS
14.1 Properties of a Vector Field:
Divergence and Curl
Definition of a vector field; PREVIEW
Divergence; Curl; A physical
interpretation of curl
This chapter joins together two important concepts—vectors and
14.2 Line Integrals
Definition of a line integral; integration. The key concepts involve a collection of vectors called a
Evaluation of line integrals in vector field and a function called a vector-valued function. We use these
parametric form; Line integrals concepts to define a line integral, which can be described informally as
of vector fields; Computing
work using line integrals; integration along a curve in space. We also examine applications of line
Evaluation of line integrals integrals, such as computing the work performed when an object moves
with respect to arc length
along a curve in a force field. A key result in vector analysis is Green’s
14.3 Independence of Path
Conservative vector fields; theorem in the plane, which allows us to compute a line integral around
Fundamental theorem on line a closed planar curve C. We also consider the concept of surface
integrals
integrals and then close by examining Stokes’s theorem and the diver
14.4 Green's Theorem gence theorem, which are generalizations of Green’s theorem involving
Green’s theorem; Area as a
line integral; Alternative forms surface integrals.
of Green’s theorem
14.5 Surface Integrals
Terminology; Surface integrals;
Surface integrals of vector
fields; Parametrically defined PERSPECTIVE
surfaces
In Chapter 11, we showed how vectors can be used in conjunction with the
14.6 Stokes’s Theorem
Stokes’s theorem; Physical methods of calculus to describe and analyze motion in space. At that time, we
interpretation of Stokes’s were primarily interested in applications involving the derivative, and our
theorem present goal is to see how vector methods may be applied to integration
14.7 Divergence Theorem problems. The topic we shall now study is called vector analysis, although
Divergence theorem; “ vector integration” might be a more appropriate name for what we plan to
Applications of the divergence
theorem explore. The material of this chapter provides the mathematical background
necessary for theoretical work in fluid mechanics, electricity, orbital mechan
Chapter 14 Review
Guest Essay, William F. Lucas ics, and magnetism.
897
898 Ch. 14 Vector Analysis
1 4 .1 PROPERTIES OF A VECTOR FIELD: DIVERGENCE AND CURL
IN THIS SECTION Definition of a vector field; divergence; curl; a physical

interpretation of curl
In order to model properties of electricity, magnetism, ∩uid dynamics, and
other applications, we use the notion of a vector field. After introducing some
basic terminology, we consider two concepts, called the divergence and the curl,
which involve vector differentiation. We conclude by considering applications
of these concepts in the study of fluid motion.
■ DEFINITION OF A VECTOR FIELD
The wind-velocity map in Figure 14.1 shows wind measurements over the Indian
Ocean. Notice that vectors are used to show the wind direction and the length
(magnitude) and color of the vectors indicate the wind speed. In Chapter 11, we
studied vector-valued functions that assign a vector to each num ber in a real-
number domain. However, these simple vector-valued functions by themselves
cannot describe the variety of motion shown here. To describe the behavior at
every point in a given space, it is necessary to use a vector field, which is a function
that assigns a vector to each m em ber of a given set of points in IR2 or R 3 . Here is a
Figure 14.1 A wind-velocity map
of the Indian Ocean
definition for the R 3 case.
Vector Field A vector field is a collection 5 of points in space together with a rule that
assigns to each point (x, y, z) in A exactly one vector V(x, y, z).
For instance, we may represent the velocity V(x, y, z) of a fluid flow by drawing an
appropriate vector at each point (x, y, z) in the domain of the flow, and the
resulting collection of vectors is called a velocity field. In practice, we often express
a given vector field V in the form
V(x, y) = u(x, j ,)i + v(x, >,)j i∏

or
V(x, y, z) = u(x, y, z)i + v(x, y, z)j + w(x, y, z)k in [R3
The functions u, v, and w are scalar functions called the components of V. For
example,
V = 2x 2yi + e yz j + (ta n ^ )k
is a vector field with i-component 2 x 2y, j-component ey z , and k-component t a n f .

To get an idea of what a vector field “ looks like,” it is often helpful to draw
V(x, y, z) as an “ arrow” at selected points in S’. We shall refer to such a diagram as
the graph of V.
EXAM PLE 1 Graph of a vector field
Sketch the graph of the vector field F(x, j ) = yi - xj.

Sec. 14.1 Properties of a Vector Field: Divergence and Curl 899
Solution We shall evaluate F at various points. For example,

F(3, 4 ) = 4 i - 3 j and F ( - 1, 2) = 2i - ( - l)j = 2i + j
We can generate as many such vector values of F as we wish. Several such
vectors are shown in Figure 14.2. ≡≡≡
Figure 14.2 The graph of the vec

tor field F(x, y) = yi - λ'j
The graph of a vector field often yields useful information about the
properties of the field. For instance, suppose F(x, y) represents the velocity of a
compressible fluid (like a gas) at a point (x, y) in the plane. Then F assigns a
velocity vector to each point in the plane, and the graph of F provides a picture of
the fluid flow. Thus, the flow in Figure 14.3a is a constant, whereas Figure 14.3b
suggests a circular flow.
Gravitational, electrical, and magnetic vector fields play an important role in
physical applications. We shall discuss gravitational fields now and electrical and
magnetic fields later in this section. Accordingly, we begin with Newton’s law of
gravitation which says that a point mass (particle) m at the origin exerts on a unit
point mass located at the point P{x, y, z) a force F(x, y, z) given by
p (χ , T, *) = χ 2 + ^T+ 2 u (A τ, z)
z
where G is a constant (the universal gravitational constant) and u is the unit vector
extending from the point P toward the origin. The vector field F(x, y, z) is called
the gravitational field of the point mass m. Because
u (x >T, Z) = ∕ 2 ~y2 7 Ξ (x i + TJ + zk)

λ γ +~
it follows that
F (A T , Z) = ( x 2 + ~ ξ 7+ z 2 )3 ∕ 2 (χ i + τi + zk)
b. A circular flow Note that the gravitational field F always points toward the origin and has the
Figure 14.3 Flow diagrams same magnitude for any point m located r = Vx 2 + y 2 + z 2 units from the origin.
Such a vector field is called a central force field. This force field is shown in Figure
14.4a. Some other examples of force fields are shown in Figures 14.4b and 14.4c.
a. A central force field b. Air flow vector field c. Wind velocity on a map
Figure 14.4 Examples of physical vector fields
■ DIVERGENCE
A vector field V can be differentiated in two different ways, one of which produces
a derivative that is a scalar and the other, a vector. The scalar derivative may be
defined as follows.
Divergence The divergence of a vector field

V(x, y , z) = u(x, y, z)i + v(x, y, z)j + w(x, y , z)k
0 The divergence of a vector is denoted by div V and is defined by
field is a scalar. 0
d iv v = f⅛ x >y ’ z ) + y >z ) + y ^z )
U Λ U K U 4
EXAM PLE 2 Finding div V
Find the divergence of each of the following vector fields.
a. F(x, y) = x 2j ,i + xy 3j b. G(x, y, z) = x i + y 3z 2j + x z 3k
Solution a. div F = ^ ( ∙* 2T) + ^^(∙∖y3) = 2xy + 3xy2
b ∙ d iv g = ∂χW + av f'y i z ^ + ⅛(χ z 3 ) = 1 + 3y2z2 + 3 x z 2 ≡≡
Suppose the vector field

V(x, y , z) = u(x, y, z)i + v(x, y, z)j + w(x, y, z)k
represents the velocity of a compressible fluid (for example, a gas) with density
p(x, y, z) at a point (x, y , z) in a certain region R in R 3. In general, the position of
the fluid will vary with time t, and at each fixed time, the vector field V assigns a
velocity to every point in R. Because the fluid is compressible, the scalar function
ρ varies from one point to another. The vector field pV is called the flux density
and is denoted by D. We can think of D = pV as measuring the “ mass flow” of the

liquid.
Assuming there are no external processes acting on the fluid that would tend
to create or destroy fluid, it can be shown that div D gives the negative time rate
change of the density; that is,
div D =
σt
This is often referred to as the continuity equation of fluid dynamics. (A derivation
is given in Section 14.7.). When div D = 0, D is said to be divergence free or
solenoidal, and the fluid is said to be incompressible. If div D > 0 at a point
(x0, y 0 , z 0), then the point is called a source; if div D < 0, the point is called a sink
(see Figure 14.5). The terms sink, source, and incompressible apply to any vector
field F and are not reserved only for fluid applications.
Div D positive Div D negative Div D = 0
Fluid leaves at Fluid arrives at Fluid is

a source point. a sink point. incompressible.
Figure 14.5 Flow of a fluid across a plane region, D
A useful way to think of the divergence div V is in terms of the del operator
defined by
V = — i + — ji + — k
∂x ∂y ∂z
The del operator operates on a differentiable function f ( x , y, z) of three variables,

and for such a function, we have
_ ∂ f. ∂ f. ∂f
^ iJ —τ - 1 + τ -j J + τ- k
∂X ∂y ∂z
We can also define V/in a totally analogous way for functions of one, two, or more
than three variables. In particular, for a differentiable function f ( x , p) of two
variables, we have
∂ f. ∂ j.
VJ fr = √ - i + √-jj
∂X ∂y
We recognize V ∕as the gradient o ff, which was introduced in Chapter 12. We can
think of div V as the dot product of the operator V and the vector field
V = wi + vj + wk; that is,
d i v V = ^ + ∣H + ⅞w = ( ⅛ + ^ -j + ⅛ ) ∙ (Mi + vj + vvk) = V ∙V
∂x ∂y ∂z ∖∂x ∂yj ∂z ) v j '
■ CURL
The del operator may also be used to describe another derivative operation for
vector fields, called the curl.
Curl The curl of a vector field

V(x, y, z) = ιt(x, y, z)i + v(x, y, z)j + w(x, y, z)k
is denoted by curl V and is defined by
curl V = ⅛ ∂v∖. + (∂u ∂¼∆i , (∂y _ ∂ι∆.
∂z) ∖∂z ∂ x ji ∖∂x ∂yj
This can be remembered in the following useful form:
We saw that V ∙ V = div V, and from this definition of curl, we see that
curl V = ⅛ - ^ ) i + (⅛ - ⅜⅛ + (⅛ - ⅛ 4 = V × V
∖∂y ∂zj ∖∂z ∂x J j ∖∂X ∂y)
This representation for curl and the one for divergence derived earlier are
summarized in the following box.
Del Operator Forms Consider a vector field

for Divergence and Curl
V(x, y, z) = u(x, y, z)i + v(x, y, z)j + w(x, y, z)k
0 div is a scalar, and curl is a The divergence and curl of V are given by
vector. 0
div V = V ∙ V and curl V = V × V
EXAM PLE 3 Curl of a vector field
Find the curl of the vector fields

F = x 2yzi + x y 2zj + xyz 2k and G = (x cosy)i + xp 2j
Solution
i j k
curl f = JL J_ J_
∂x ∂y ∂z
x 2yz x y 2z xyz 2
= (⅜ x ^ 2 - i χ y 2 z ) i - (⅛ x ^ 2 - ⅛X>)J+ ( i χ y 2z ~ ^ χ2yz )k
= (xz 2 - xρ 2)i + (x 2y —y z 2)j + (y 2z - x 2z)k
i j k
= ∂ JL J_
∂x ∂y ∂z
x cosp xy2 0
= [0 - ⅛χd i+ [⅛ x cosj;) ^ 0 ]j + ½ x∙μ2 ^ ⅜(χ cos^] k

= (y 2 + x siny)k ∣
Sec. 14.1 Properties of a Vector Field: Divergence and Curl 9 03
EXAM PLE 4 A constant vector field has divergence and curl zero
Let F be a constant vector field. Show div F = 0 and curl F = 0.

Solution Let F = ai + b} + ck for constants a, b, and c. Then
d iv f - ⅛ ω ÷ ⅜ <t > ÷ ⅛ ω = °
i j k
curl F = a a a = 0i - 0j + Ok = 0
∂x ∂y ∂z
a b c
0 Example 4 shows that the divergence and curl of a constant vector field
are zero, but this does not mean that if the divergence or curl vanishes, then
the associated vector field is a constant. For instance, if
F(x, y, z~) — yz∖ + xj + xy 2 k
then div F = 0. 0
■ A PHYSICAL INTERPRETATION OF CURL
The concepts of the divergence and curl, as well as most of the other terms in this
chapter, are motivated by concepts from physics. For example, we can apply the
curl to the study of fluid motion. If a fluid is moving about a region in the xy-
plane, the curl can be thought of as the circulation of the fluid. A good way to
measure the effect (magnitude and direction) of the circulation is to place a small
paddle wheel into the fluid. The curl measures the rate of the fluid’s rotation at the
point P where the paddle wheel is placed in the direction of the paddle wheel’s
axis. The curl is positive when the rotation is counterclockwise and negative when
it is clockwise. Let V(x, y, z) = u(x, y, z)i + v(x, y, z)j + w(x, y, z)k be the velocity
of an incompressible fluid and suppose a paddle wheel is inserted into the fluid in
such a way that its axis lies along the z-axis, as shown in Figure 14.6. Disregard the
weight of the paddle. The fluid tends to swirl about the z-axis, causing the paddle
to rotate, and we can analyze the fluid’s swirling motion by studying the motion of
the paddle. It turns out that the angular velocity of the liquid
∂y∖.
about the x-axis is proportional to ∂ z) ,
∂H,Y
about the y-axis is proportional to
∂ x ∕,
Figure 14.6 A physical interpreta about the z-axis is proportional to ( / ∂u ∖
tion of curl V
∖ U Λ ∂yp
Thus, the tendency of the fluid to swirl is measured by curl V. In the special case
where curl V = 0, the fluid has no rotational motion and is said to be irrotational.
Certain combinations of the gradient, divergence, and curl appear in many
applications. In particular, note that if f is a differentiable scalar function, its
gradient V /is a vector field and we can compute
div V ∕= ( / - i + / - j + / - k) ⅛ + ⅛ i + ⅛ k
J ∖∂X ∂yi ∂z ) ∂x1 ∂y , az
^ f + d2 f +
dΛ
∂x 2 ∂y 2 ∂z 2
= V -V f
The notation V ∙ V/is usually abbreviated by writing V2∕ and it is called either
“del-squared / ” or the Laplacian of f after the French mathematician Pierre
Laplace (1749-1827). A function f whose first and second partials are continuous
and that satisfies the equation V2∕ = 0 is called harmonic. We summarize these
ideas in the following box.
The Laplacian Operator Let f(x , y, z) define a function with continuous first and second partial
derivatives. Then the Laplacian of f is
V 2f = v ∙ V ∕= - ⅛ + - ^ { + ⅛ = ∕ + / + /
J J ∂X 1 ∂y2 ∂z2 j χ x j yy j z z
The equation V2∕ = 0 is called Laplace’s equation, and a function that satisfies
such an equation in a region D is said to be harmonic in D.
EXAM PLE 5 Showing a function is harmonic
Show that ∕(x , y) = ex cosy is harmonic in the plane.

Historical Note
Solution ∕ γ(x, y) = ex cosy and ∕ , γ(x, y) = e vcosy
f y (x, y) = -β' vsiny and ∕ yy (x, y) = - e λ cosy
The Laplacian o f/is given by
∕( * , y) = f xx (x, y) + f yy (x, y) = e⅛osy - eλcosy = 0
T hus,/is harmonic.
In many ways, the study of electricity and magnetism is analogous to that of

fluid dynamics, and the curl and divergence play an important role in this study.
In electromagnetic theory, it is often convenient to regard interaction between
electrical charges as forces somewhat like the gravitational force between masses
and then to seek quantitative measure of these forces.
One of the great scientific achievements of the 19th century was the discovery
J AMES CLERK MAXWELL of the laws of electromagnetism by the English scientist James Clerk Maxwell.
(1831-1879) These laws have an elegant expression in terms of the divergence and curl. It is
James Maxwell was a Scottish known empirically that the force acting on a charge due to an electromagnetic
mathematician who published a field depends on the position, velocity, and amount of the particular charge and
treatise on electricity and mag not on the number of other charges that may be present or how those other charges
netism that provided the frame are moving. Suppose a charge is located at the point (x, y, z) at time t, and consider
work for modern discoveries in the electric intensity field E(x, y, z, t) and the magnetic intensity field H(x, y, z, t).
this area. He compiled all the Then the behavior of the resulting electromagnetic field is determined by the
facts about electricity and mag
netism that were known at the
following equations:
time. He formulated several
equations basic to the theory of div E = — curl E = —⅛(μ∙H) curl H = J
ε ∂ t r^ ’
electricity and magnetism now
known as Maxwell’s equations. where Q is the electric charge density (charge per unit volume), J is the electric
James Clerk Maxwell was
sometimes referred to as dp/dt
current density (rate at which the charge flows through a unit area per second), μ is
because in thermodynamics a constant called the permeability, and ε is a constant called the permittivity.
dp/dt = JCM (a unit of mea Furthermore, if σ is a constant called the conductivity, then
surement named for James
Clerk Maxwell). (V ∙V)E = μ σ ^ + μ ε ^
Working with these equations and terms is beyond the scope of this course, but if
you are interested there are many references you can consult. One of the best (in
spite of the fact it is 30 years old) is the classic book written by Nobel laureate
Richard Feynman, Robert Leighton, and Matthew Sands, Feynman Lectures in
Physics (Reading, Mass.: Addison-Wesley, 1963).
PROBLEM SET 1 4 .1
@ 1. ■ What Does This Say? Compare and contrast diver 37. w(x, y, z) = (x 2 + y 2 + z 2) -ιz 2
gence and curl. 38. r(x, y, z) = xyz
2. ■ What Does This Say? Discuss the del operator. θ 39. If F(x, y, z) = 2i + 2xj + 3yk and
Sketch several representatives o f the given vector field in G(x, y, z) = xi - yj + zk, find curl (F × G).
Problems 3-8. 40. If F(x, y, z) = xyi + yzj + z 2k and
3. F = xi + yj 4. G = - x i + yj G(x, y, z) = xi + yj - zk, find curl (F × G).
5. V = x 2i + yj 6. H = x 2i + y 2j 41. If F(x, y, z) = 2i + 2xj + 3yk and
G(x, y, z) = xi - yj + zk, find div(F × G).
7. F(x, y) = yi + xj 8. G(x, y) = xyi - j
42. If F(x, y, z) = xyi + yzj + z 2k and
In Problems 9-12, fin d div F and curl F for the given vector G(x, y, z) = xi + yj - zk, find div(F × G).
function.
43. Find div F, given that F = V ∕ where ∕( x , y, z) = xy 3z 2 .
44. Find div F, given that F = V ∕ where f(x , y, z) = x 2yz 3.
10. F(x, y) = i + (x 2 + y 2)j 45. A magnetic field that has zero divergence everywhere
11. F(x, y, z) = 2yj 12. F(x, y, z) = zi - j + 2yk is said to be solenoidal (because such a field can be gen
erated by a solenoid). Show that the field
In Problems 13-18, find div F and curl F for each vectorfield F B = y 2zi + xz 3j + y 2x 2k is solenoidal.
at the given point. 46. A central force field F is one that can be described as
13. F(x, y, z) = i + j + k at (2, - 1, 3) F = ∕(r)R . where R = xi + yj + zk and r = ∣ ∣R∣
∣=
14. F(x, y, z) = xzi + y 2zj + xzk at (1, —1, 2) (x 2 + y 2 + z 2) ιz2. Show that such a field is irrotational
15. F(x, y, z) = xyzi + yj + xk at ( 1, 2, 3) everywhere it is defined and differentiable. That is, show
16. F(x, y, z) = (cosy)i + (siny)j + k at ’ π, θ) curl F = 0.
17. F(x, y, z) = e~x y i + e xz j + ey z k at (3, 2, 0) θ 47. Let A be a constant vector and let R = xi + yj + zk. Show
that div(A × R) = 0.
18. F(x, y, z) = (c- x siny)i + (e - ∙vcosy)j + k at (1, 3, -2 )
48. Let A be a constant vector and let R = xi + yj + zk. Show
Find div F and curl F for each vector field F given in Problems that curl(A × R) = 2A.
19-34. 49. If F(x, y) = u(x, y)i + v(x, y)j, show that curl F = 0 if and
19. F = (sinx)i + (cosy)j 20. F = ( - cosx)i + (siny)j only if
21. F = xi —yj 22. F = - x i + yj
∂u ... ∂v
p _____ x . , V .
V x2 + y2 1 V x2 + y 2 j ∂y ∂x
24. F= x 2i - y 2j 50. If R = xi + yj + zk and F = ∣
∣
^ ∙ show that curl F = 0.
25. F= axi + byj + ck forconstants a, b, c
26. F= (Asin y)i + (Acos y)j + k 51. Consider a rigid body that is rotating about the z-axis
(counterclockwise from above) with constant angular
27. F= x 2 i + y 2j + z 2k 28. F = yi + zj + xk
velocity ω. If P is a point in the body located at
29. F= xyi + yzj + xzk 30. F = 2xzi +2yz2j - k R = xi + yj + zk, the velocity at P is given by the vector
2 2 2
31. F = xyzi + x y z j + y z k 2 3
field V = ω × R.
32. F = - z 3i + 3j + 2yk a. Express V in terms of i, j, and k vectors.
33. F = (x - y)i + (y - z)j + (z - x)k b. Find div V and curl V.
xi + yj + zk 52. T HINK. T ANK P ROBLEM Find an expression for
34. F = , 2 , ' i2 , 2
Vx + y + z curl(curl F) for any vector field F.
Determine whether each scalar function in Problems 35-38 is 53. Which (if any) of the following is the same as div(F × G)
harmonic. for any vector fields F and G?
35. ι∕(x, y, z) = e - v(cos y - sin y) I. (div F)(div G)
36. v(x, y, z) = (x 2 + y 2 + z 2) 1/2 II. (curl F) ∙ G - F ∙ (curl G)
III. F(div G) + (div F)G 58. div(F × G) = (curl F) ∙ G - F ∙ (curl G)

IV. (curl F) ∙ G + F ∙ (curl G) 59. curl ( ∕F ) = f curl F + (V ∕ × F)
In Problems 54-64, prove the given property for the vectorfields 60. d iv (∕F ) = ∕d i v F + (V ∕∙ F)
F and G, scalar c, and scalar function f. Assume that the 61. curl(V∕ + curl F) = curl(V∕) + curl F
required partial derivatives are continuous. 62. div(∕V g) = V ∙ (∕V g)
54. div(cF) = c div F 63. The curl of the gradient of a function is always 0. That is,
55. div(F + G) = div F + div G V × (V ∕) = 0.
56. curl(F + G) = curl F + curl G 64. The divergence of the curl of a vector field is 0. That is,
57. curl(cF) = c curl F div(curl F) = 0.
1 4 .2 LINE INTEGRALS
IN THIS SECTION Definition of a line integral, evaluation of line integrals

in parametric form, line integrals of vector fields, computing work using line
integrals, evaluation of line integrals with respect to arc length
In Section 6.3, we showed that the single Riemann integral can be used to
compute the work done when an object moves along a line segment against a
given force, but what if the object moves along a curve in space? This is one of
many situations that occur in science and engineering in which a more general
definition of integration is required. The line integral introduced in this section
is the first of many such generalizations of integration you will encounter if you
continue your studies in mathematics.
■ DEFINITION OF A LINE INTEGRAL

In order to model certain physical notions, such as work or potential, it is
appropriate to generalize the original concept of integral by considering limits of
sums whose summands depend in a certain way on a curve, called the path o f
integration. This leads us to the concept of line integration, which is really
integration along a curve in space. We begin by defining the line integral of a scalar
function f along a curve C with respect to x. The line integral of a scalar function g
along C with respect to y or the line integral of h with respect to z may be defined
in an analogous fashion.
First, we need to introduce some terminology regarding curves. Recall
(Section 1 1.2) that a curve C with the parametric representation R(t) = x(t)i
+ T(f )j + z(t)k i s said t o be smooth on the interval f < t < t2 if the derivatives
x'(t), y'(t), and z'(t) are all continuous and not all simultaneously zero at any
point t on the interval. More generally, C is piecewise smooth if it can be
Figure 14.7 A partitioned smooth decomposed into a finite number of smooth parts. Also, C is said to be orientable
curve if it is possible to describe direction along the curve for increasing t.
Assume C is a piecewise smooth, orientable curve that begins at P = P o and
ends at Q = P n , as shown in Figure 14.7. Let C b e partitioned into n pieces by the
points P o , P γ, . . . , P n, and let (x fc, yk , z k ) be the coordinates of the point P k . Finally,
for k = 1, 2, . . . , n, let P k(x k , yk , z""k ~) be a point chosen arbitrarily from the arc
joining points P k _ 1 and P k , and let ∆x k = x k - x k _1 , ∆yk = yk - yk _v and
∆zi = z k - z fc - 1 . We shall refer to the largest of the ∆x k as the x-norm of the
partition and shall denote this norm by ∣ ∣
∆x ∣∣
. Similar definitions hold for the y-
norm and the z-norm. For a given scalar function f we form the sum
S - Σ Λ⅛Δ¾
K— 1
Sec. 14.2 Line Integrals 907
and define the line integral f dx over C with respect to x as follows.
Line Integral The line integral f dx of the scalar function f with respect to x along
Jc
the curve C is given by
f f dx = lim ∑ /(A ) ΔΛ .
Jc j M →O fc=ι k k
provided that this limit exists.
The line integrals f dy and f dz are defined similarly.

Jc Jc
THEOREM 14.1 Properties of line integrals
L et/b e a given scalar function defined with respect to x on a piecewise smooth,

orientable curve C. Then, for any constant k:
Constant multiple rule: k f dx = k f dx

'c Jc'
Sum rule: ( / + f 2)d x = I ∕ 1 dx + f 1 dx

Jc ^ Jc Jc
where ∕ 1 and ∕ 2 are scalar functions defined with respect to x on C.
Opposite direction rule: f dx = - f dx

J-c Jc
where - C denotes the curve C traversed in the opposite direction.
Subdivision rule: f dx = f dx + f dx
Jc Jcl Jc2
where C is subdivided into subarcs C1 and C2 (with C = Cj U C2 and Cl ∩ C2
= 0 ). This property generalizes to any finite number of subdivisions.
Similar properties hold for line integrals of the form
g dy or h dz.
Jc Jc
Proof: The proof follows directly from the properties of limits and the
definition of a line integral. ■
■ EVALUATION OF LINE INTEGRALS IN PARAMETRIC FORM
In practice, the line integral f dx is almost never evaluated using the definition.
Jc
Instead, we note that if the integrand f(x , y, z) is continuous on C and if C can be
represented parametrically in vector form by R(t) = x ∕) i + y(t)j + z(t)k, where
the derivative R'(Z) exists and is not zero for a ≤ t ≤ b, then
f ft Px
f dx = Ja f[χ(t), y(t), z(t)] ⅛ dt
Similarly, if g and h are continuous on C ,

[ g d y = [ g[x(t), y(0 , z (t)] ⅜ dt [ h ι dz = [ b[x(t), y(t), z(t)] ⅜ dt
JC Ja tu JC Ja u t
These formulas enable us to convert a line integral into “ ordinary” Riemann

integrals that may be evaluated by the methods developed earlier in this text. The
same answer is obtained independent of the parametrization chosen.
EXAM PLE 1 Evaluating a line integral
Let C be the portion of the parabola y = x 2 from (0, 0) to (2, 4). Evaluate
a. I (x 2 + y) dx b. I (x 2 + y) dy
Jc Jc
Solution First, parametrize the curve (this step is not unique). We let x = t so
that y = t2 for 0 ≤ t ≤ 2.
a. Because = 1, we have
r c2 √ r2 r∏ ^∣2
(x 2 + y) dx = [(∕) 2 + (12)] ⅛
z
d t = 2 ∖ t2 dt = ⅛ 3 0 = y
Jc Jo al Jo L-> J
b. Now we have = 2/, and we find
z7v f“ C~ r ^ι □
[ ( t ) 2 + (^ 2 ) ] -∕t dt = I 2t2 -2 td t = 4 I t3 dt = [ t 4Jθ = 16
EXAM PLE 2 Evaluating a line integral
Evaluate xe yz dz, where C is the curve described parametrically by x = /,

Jc
y = t, and z = —t for 1 ≤ t ≤ 2.
_ ∕*2
s ° lu tio n j c xey=dz = te~t 2 f dt
t
t*,2 y
= (-t)e~ t dt Because -I
Ji c
= ⅛(e~4 - e~ , ) β
■ LINE INTEGRALS OF VECTOR FIELDS
We shall now discuss what it means to compute the line integral of a vector field.
Line Integral of a Vector Field Let F(x, y, z) = f( x , y, z)i + g(x, y, z)j + b(x, y, z)k be a vector field, and let C
be the curve with parametric representation
R(∕) = x(Z)i + y(t)j + z(∕)k for a ≤ t ≤ b
Using //R = dx i + dy j + dz k, we denote the line integral of F over C by
F ∙ //R and define it by

Jc r f
F ∙ ∕ZR = f dx + g dy + h dz
Jc Jc
= ∣a [ ∕M 0 , y(ι), z (t)l + g[-l f (0, τ(0> z (t)l ⅛ + K A t) , y(0 , ^(∕)] ^
c ∖ dt
EXAM PLE 3 Line integral of a vector function
Evaluate F ■tZR, where F = (y2 - z 2)i + (2yz)j - x 2k and C is the curve

Jc
defined parametrically by x = t2, y = 2t, and z = t for 0 ≤ t ≤ 1.
Solution Rewrite F using the parameter:
F = [(2z)2 - (t)2]i + [2(2z)(Z)]j - [(t2)2]k = 3t2i + 4z2j - Z4k. Also, because
R(Z) = Z2i + 2/j + Zk, we have √R = (2t dt)∖ + (2 df)j + dtk, so that
F ■√R = (3Z2)(2Z dt) + (4Z2)(2 df) + (-Z 4)(r∕Z) = (6Z3 + 8z2 - ∕ 4) dt.
F dR =
Jo
(6Z3 + 8Z2 - Z4) dt = ⅛4 + t^3- ⅛5υjo . 119
30
EXAM PLE 4 Line integral of a vector defined function
Compute the line integral F ∙ d R , where F = pi + xj and C is the top half of

Jc
the circle x 2 + y 2 = 4 traversed counterclockwise from (2, 0) to (—2, 0).
Solution First, we parametrize the curve by setting x = 2 cos θ, y = 2 sin θ for
0 ≤ θ ≤ 7r. Thus,
R(∕) = (2 cos 0)i + (2 sin θ)j so that d R = ( - 2 sin θ dθ)i + (2 cos θ dθ)j
∣F ∙ d R = ∣ [(2 sin 0)i + (2 cos 0)j] ∙ [(-2 sin θ)i + (2 cos θ)j] dθ
Jc Jo
= f ( - 4 sin 2 0 + 4 cos2 0) dθ
Jo
= 4 ∣ (cos2 0 - sin 2 0) dθ = 4 ∣ cos 2θ dθ = ∣^2 sin 20^∣∩ = 0
Jo Jo l j ≡β
EXAM PLE 5 Line integrals along different paths
Let F = xy 2i + x 2yj and evaluate the line integral F ∙ dR between the points
Jc
(0, 0) and (2, 4) along the following paths:
a. the line segment connecting the points

b. the parabolic arc y = x 2 connecting the points
Figure 14.8 A line integral along different paths

a. The line joining the given points has equation y = 2x, which may be
parametrized by setting x = t, y = It for 0 ≤ t ≤ 2. Thus,
R(∕) = ∕i + 2zj so that J R = dt i + 2dt j
In terms of t, we find F = 4 ∕ 3i + 2 ∕ 3j and
F ∙ J R = 4t3 dt + 4t3 dt = 8 ∕ 3 dt
f F ∙ J R = f^ 8t3 dt = ∣^2∕4 12 = 32
JC Jθ L ju
b. The parabola y = x 2 can be parametrized by setting x = t, y - t2 for

0 ≤ t ≤ 2. Thus,
R(Z) = ti + Z2j so that J R = dt i + It dt j
In terms of t, F = xy 2i + x 2yj = (Z)(Z2)2i + (0 2(Z2)j = + ^4j and

F ∙ J R = t 5 dt + 2z5 dt = 3t5 dt
’ r2
3Z5 dt = Γ⅛6^]2 = 32
Jc
F ■J R =
Jo l j
In Example 5, we see that the value of the line integral is the same for both
paths. Indeed, it can be shown that for the vector field F = xy 2i + x 2>,j, the line
integral F ■J R along any path Cjoining (0, 0) to (2, 4) has the same value. This,
Jc
of course, is not true for every F, but when it is true, we say that the line integral is
independent of path. Path independence is an important feature of line integration
and will be discussed in detail in the next section.
EXAM PLE 6 Line integral over a path consisting of several parts
Evaluate ( - j dx + x dy), where C is the closed path shown in Figure 14.9.

Figure 14.9 A path C composed Jc
of three parts
Solution Because the given closed path can best be described using three separate
equations (C 1, C 2, and C 3 in Figure 14.9), we consider the integration by
evaluating the line integrals for each part separately. Let F = - j i + xj, so that
[ ( - y dx + x dy) = f F ∙ JR . Then
Jc ' Jc
f F ■J R = [ F ∙ J R + f F ∙ J R + f F ∙ J R
Jc Jc, Jc 2 Jc,
Along C j, a parametrization is x = cos t, y = s in ∕ for 0 ≤ t ≤ so
R(Z) = (cos∕)i + (sin Z)j and J R = (-sinZ dt)∖ + (cost Jz)j. In terms of the
parameter, F = (—sin Z)i + (cos∕)j and
∙ rτr∕2 rπl2
J c,
F ∙ JR =
Jo
(sin2 / dt + cos2/ dt) =
Jo
dt = 2
Along C 7, a parametrization is x = 0, y = l - t for 0 ≤ t ≤ 1, so

R(∕) = ( 1 - ∕)j and J R = (—Jz)j. In terms of the parameter, F = (t - l)i and
F ∙ JR = 0= 0
'c, Jo
Along C 3, a parametrization is x = t, y = 0 for 0 ≤ t ≤ 1. R(t) = ti, dR = dti

and F = tj. Because F ■ dR = 0, we see
F ∙ dR = 0
Thus,
I F ■
dR = f F ∙ dR + ( F ∙ dR + f F ∙ dR = f + 0 + 0 = ?
Jc Jc, Jc, Jc, - λ
■ COMPUTING WORK USING LINE INTEGRALS
One of the most important physical applications of the line integral is in

computing work. Recall from Section 10.3 that if an object moves along a straight
line in a constant force field F with displacement R, the work done is F ∙ R.
The case (shown in Figure 14.10) where the force field F is not constant and
the object moves along a smooth curve C (instead of along a straight line) requires
additional attention. Assume that C is parametrized by R(Z) and is oriented in the
direction of motion. Partition C with subdivision points PQ, P l , P v . . . , P , as
Figure 14.10 If F and R are con

stant, work = F ∙ R. If they are
not, then a line integral is re
quired.
a. Work with constant force b. Work with variable force

in a fixed direction along a smooth curve
Figure 14.11 Computing work
For k = 1, 2, . . . , n, let Q k be a point chosen arbitrarily from the Zcth subarc C k

(with endpoints P k _ 1 and P k ), and let F fc be the value of the force field F at Q k . If the
length of the subarc C k is small, F will be approximately constant with value F fc
over C k , and the object’s displacement along C k is given approximately by the
secant vector ΔR, from P. . to P ,. Then, the work done by the force as the object
moves along C k may be estimated by
/ ΔRi ∖
Δ ^ =K ^ Δ∕
and by adding the contributions along all n subarcs, we obtain
/ Δ R ,∖ / ΔR 7∖ + / ΔR∖ « / ΔR,.∖
V ι ’ ~ ∆ Γ ) ∆t + ( F 2 ∙ '" + (F„ ∙ ^∆Γj)∆∕ - fc∑ ι ^Ffe ∙
as an estimate of the total work done by F as the object moves along C . As ∆t → 0.

,/D
The limiting value of this sum is the line integral of F ∙ ; that is,
-Λ / ΔR∕∖ f ∕τ> f
W = lim ∑ ∣ F, ∙ -τ-pd∆∕ = F ∙ z dt = F dR
ΔZ→O k=∖ ∖ k ∆t j j c at jc
This leads us to the following definition.
Work as a Line Integral Let F be a continuous force field over a domain D. Then the work W done by F
as an object moves along a smooth curve C in D is given by the line integral
W = [ F ∙d R
Jc
EXAM PLE 7 Work as a line integral
An object moves in the force field

F = y 2i + 2(x + l)jj
counterclockwise from the point
(2, 0) along the elliptical path
x 2 + 4y2 = 4 to ( - 2 , 0), and back
to the point (2, 0) along the x-axis.
How much work is done by the
force field on the object?
Solution Let C denote the path of the object. Then the total work W done by F on
the object as it moves along C is given by the line integral

F ∙ √R. We divide
Jc
the curve C into two parts so that C = C l U C 2 . The curve C l may be
parametrized by
x = 2 cos t, y = sin t for 0 ≤ t ≤ ττ
∕r ∖ 2
Because I y I + J’2 = 1
R(∕) = (2 cos∕)i + (sin∕)j and ∕7R = ( - 2 sin/ dt)i + (cost √∕)j

In parametric form,
F = (sin2 ∕)i + (4 cost sin / + 2 sinZ)j
and we have
IT1 = J F dR
= ( -2 sin3/ + 4 cos2/ sin t + 2 sin / cos /) dt

Jo
= ∣ ( -2 sin2 1 + 4 cos2 1 + 2 cos /) sin t dt
Jo
= (6 cos2/ + 2 cos/ - 2) sin / dt Let u = cos/;

J° du = -sin t dt.
∣∙-1 If t = 0, u == 1,
= —I (6w2 ÷ 2w —2) du —0 and if t - π, u = -1.
Ji
The curve C 2 is described by y = 0, so F = 0 and, therefore, Wz2 = 0. Thus,
W = W l + W2 = 0. —
Sec. 14.2 Line Integrals 91 3
■ EVALUATION OF LINE INTEGRALS

WITH RESPECT TO ARC LENGTH
Line integrals of the form F ∙ √R can often be expressed in other forms. For
Jc
example, recall from Chapter 11 that T = dRJds is a unit tangent vector to the
curve C at the point P(x, y, z), where 5 is the arc length parameter. We have
[ F ∙ <∕R = ∣F ■? ds = f F ∙ T ds
'c Jc as Jc
In particular, the work W done by a force field F on an object moving along a
curve C may be expressed as
W = f F ∙ dR = f F ■
T ds
Jc Jc
This integral is called the line integral of the tangential component of F and can
also be written as f( x , y, z) ds. The integral will exist if/ is continuous on the
Jc
curve C and if C itself is piecewise smooth with finite length. A formula for
computing this line integral may be obtained by noting that if R(Z) = x(Z)i + y(t)j
+ z(Z)k, then
f f dt
so that I f ( x , y , z) ds = I ∕( x , y, z) ∕y dt. To summarize:
Evaluation of a Line Integral L et/b e continuous on a smooth curve C . If C is defined by

with Respect to Arc Length R(0 = x(t)i + y(t)j + y(√)k, where a ≤ t ≤ b, then
r (∙b ,_______________________________
f( x , y , z )d s = f[x ( f) , y(t), z(Z)]V[x'(t)]2 + [ ∕( t ) ] 2 + [z'(t)]2 dt
Jc Ja
EXAM PLE 8 Evaluating a line integral with respect to arc length
Find (x + y 2 - z) ds, where C is the curve described parametrically by

Jc
x = t, y = 2t, z = t + 1 for 0 ≤ t ≤ 1.
Solution Because = 1, = 2, and = 1, we have
f (x + y 2 - z) ds = f [t + (2t)2 - (t + 1)] V ( l) 2 + (2)2 + (1)2 dt

Jc Jo
Γl
x = t, y = 2t, Z = t + 1 = V6 (4z2 - 1) dt = ∣ V6
for 0 ≤ t ≤ 1 Jo
PROBLEM SET 1 4 .2
0 1. ■ What Does This Say? Describe a line integral.
8. I [(>,2 ^^ x 2) dx - x dy↑
2. ■ What Does This Say? Discuss the evaluation o f a line Jc
integral. C is the quarter-circle
x 2 + y 2 = 4 from (0, 2) to
Evaluate each line integral given in Problems 3-16.
(2, 0).
3.
C is the parabolic path

y = 4x 2 from (1, 4) to (0, 0).
9. [(x2 + y 2) dx + 2xydy] for these choices of the curve C :

Jc
a. C is the quarter-circle x 2 + y 2 = 1 traversed counter
clockwise from (1, 0) to (0, 1).
b. C is the straight line y = 1 - x from (1, 0) to (0, 1).
4. (- y dx + 3x dy)
Jc 10. [x 2y dx + (x 2 - y 2) dy] for these choices of the curve C :
Jc
C is the parabolic path
y 2 = x from (1, 1) to (9, 3). a. C is the arc of the parabola y = x 2 from (0, 0) to (2, 4).
b. C is the segment o f the line y = 2x for 0 ≤ x ≤ 2.
11. Rework Problem 10 for the path C that consists o f the
horizontal line segment (0, 0) to (2, 0), followed by the
vertical segment from (2, 0) to (2, 4).
12. ( - x τ 2 dx + x 2 dy)
Jc
5. (x dy - y dx)
Jc where C is the path defined
C is the line segment defined by the given figure.
by l x - 4y = 1 as x varies
from 4 to 8.
13. ( - y 2 dx + x 2 dy)
Jc
where C is the path defined
6. [(T - x) dx + x 2y dy] by the given figure.
Jc
C is the curve defined by
y 2 - x 3 from (1, -1 ) to (1, 1).
14. [(x2 —y 2 )dx + x dy}, where C is the closed circular

Jc
path given by x = 2 cos θ, y = 2 sin θ, 0 ≤ θ ≤ lτr.
7. [(x + y) 2dx - (x - y) 2dy]
Jc 15. (x 2y dx - xy dy), where C is the closed path that begins
C is the curve defined by Jc
y = ∣2x∣ from ( -1 , 2) to (1, 2). at (0, 0), goes to (1, 1) along the parabola y = x 2, and then
returns to (0, 0) along the line y = x.
r r x dx - y dy ι w ιe r e
16. I ∙y∕χ 2 + * C ι s t 'η e quarter-circular path
x 2 + y 2 = a 2, traversed from (a, 0) to (0, a).

In Problems 17-19, evaluate F ■ √R, where 29. F ∙ dR, where F = y 2i + x 2j - (x + z)k and C is the
Jo Jc
F = (5x + y)i + x j and C is the specified curve. boundary o f the triangle with vertices (0, 0, 0), ( 1, 0, 0),
17. C is the straight-line segment from (0, 0) to (2, 1). (1, 1,0), traversed once clockwise, as viewed from above
18. C is the curve given by R(z) = 2ri + zj for 0 ≤ t ≤ 1.
19. C is the vertical line from (0, 0) to (0, 1), followed by the 30. ∣ F ■T ds, where F = —3yi + 3xj + 3xk, and C is the
Jc
horizontal line from (0, 1) to (2, 1).
straight-line segment from (0, 0, 1) to (1, 1, 1)
Evaluate the line integrals in Problems 20-35.
31. I F ∙ T ds, where F = - x i + 2k, and C is the boundary of
20. (y dx - x dy + dz)∖ C is the helical path given by: Jc
Jc the trapezoid with vertices (0, 0), (1, 0), (2, 1), (0, 1),
a. x = 3 sin t, y = 3 cos t, z = t for 0 ≤ t ≤ f traversed once clockwise as viewed from above
b. x = a sin t, y = a cos t, z = t for constant a and
o ≤ t ≤ f 32. y ds, where C is the curve given by R(Z) = ri + 2z3j,
Jc
0 ≤ t ≤ 2
21. I (x dx + y dy + z dz), where C is the following path:
Jc
33. (x + j j ds, where C is given by
a. the helix defined by x = cos t, y = sin t, z = t for Jc
0 ≤ t ≤ f R(Z) = (cos2 Z)i + (sin2 Z)j, - J ≤ t ≤ 0
b. the straight-line segment from (1, 0, 0) to ^0, 1,
„ f x 2 + xy + y 2 , , . , , .
34j .------------χ------- ds, where C is the path given by
Jc z
22. ( - y dx + x dy + xz dz), where C is the following path: R(Z) = (cos Z)i + (sin Z)j - k for 0 ≤ Z ≤ 2ττ
Jc
35. Evaluate the line integral
a. the helix defined by x = cos t, y = sin t, z = t for
0 ≤ t ≤ 2τr x dy - y dx
b. the unit circle x 2 + y 2 = 1, z = 0, traversed once coun
terclockwise as viewed from above Jc x2 + y2
where C is the unit circle x 2 + y 2 = 1 traversed once
23. (5 xy dx + 10 yz dy + z dz); C is the following path: counterclockwise.
Jc
θ 36. How much work is done by a constant force F = ai + j
a. the parabolic arc x = y 2 from (0, 0, 0) to ( 1, 1 ,0 )
when a particle moves along the line y = ax from x = a
followed by the line segment given by x = 1, y = 1,
to x = 0?
0 ≤ z ≤ 1
b. the straight-line segment from (0, 0, 0) to ( 1, 1, 1) 37. A force field in the plane is given by
F = (x 2 - y 2)i + 2xyj. Find the total work done by this
„ f d x + dy , ∣ ∣ ∣ ∣ ,
24j . I ∣χ ∣ + ∣y∣ where C is the square ∣x∣ + ∣y∣ = 1, traversed force in moving a point mass counterclockwise around
the square with vertices (0, 0), (2, 0), (2, 2), (0, 2).
once counterclockwise 38. Find the work done by the force field
F = (x 2 + y 2)i + (x + y)j a s a n object moves counter
25∙ [(>, + z) dx + (x + z) dy + (x + y) dz], where C is the clockwise along the circle x 2 + y 2 = 1 from (1, 0) to
Jc
(—1, 0), and then back to (1, 0) along the x-axis.
circle o f radius 1 centered on the z-axis in the plane z = 2,
traversed once counterclockwise as viewed from above 39. A force acting on a point mass located at (x, y) is given by
F = yi + 2xj. Find the work done by this force as the
26. I F ■ <√R, where F = (y —3)i + xj, and C is the curve point mass moves along a straight line from (1, 0) to (0, 1).
Jc
given by R(Z) = (sin Z)i — (cos Z)j for 0 ≤ t ≤ π Find the work done by the force F(x, y, z) on an object moving
along the curve C in Problems 40-43.
27. f F ■ <7R, where F = (y - 2z)i + xj - 2xyk and C is the 40. F = (y 2 —z 2)i + 2yz] —x 2k, and C is the path given by
Jc x(Z) = Z, y(Z) = Z2 , z(Z) = Z3, for 0 ≤ Z ≤ 1.
path given by R(Z) = ri + Z2j - k for 1 ≤ t ≤ 2 41. F = 2xyi + (x 2 + 2)j + yk, and C is the line segment from
(1, 0, 2) to (3, 4, 1).
28. F ∙ <√R, where F = xi + xyj + x 2yzk, and C is the
Jc 42. F = xi + yj + (xz - y)k, and C is the line segment from
(0, 0, 0) to (2, 1, 2 ).'
elliptical path given by x 2 + 4y2 - 8y + 3 = 0 in the xy-
plane, traversed once counterclockwise as viewed from 43. F = xi + yj + (xz - y)k, and C is the path given by
above R(Z) = Z2i + 2zj + 4z3k for 0 ≤ t ≤ 1.
44. A person carries a bag θ 47. Suppose a particle with charge Q and mass m moves with
of sand up a circular velocity V under the influence of an electric field E and a
helical staircase on the outside magnetic field B. Then the total force on the particle is
of a tower 150 ft high and F = <2(E + V × B), called the Lorentz force. Use New
80 ft in diameter. Suppose ton’s second law of motion, F = mA, to show that
the person weighs 150 lb,
the bag weighs 40 lb, and m ■V = G>E ∙
50 ft can be climbed in at at
exactly 5 revolutions. Then find the work done by the Lorentz force as a particle
moves along the curve C parametrized as R(t), for
t i ≤ t ≤ t,.
How much work is done by gravity on the person and the
bag during the climb to the top? 48. Suppose a thin wire fits a smooth curve C in R 3 , and let
p{x, y, z) be the density of the wire. Explain why the mass
45. Repeat Problem 44 assuming that the bag leaks 1 lb of
m of the wire is given by the line integral
sand for every 15 ft of ascent. How much work is done by
gravity during the person’s climb to the top?
46. A 5000-lb satellite orbits the earth in a circular orbit ρ(x, y, z) ds
5000 mi from the center of the earth. How much work is
done on the satellite by gravity during one complete Express the center of mass of the wire in terms of line
revolution? integrals.
1 4 .3 INDEPENDENCE OF PATH
IN THIS SECTION Conservative vector fields, fundamental theorem on line

integrals
In general, the value of the line integral F ∙ √R depends on the path of

Jc
integration C, but in certain cases, the integral will be the same for all paths in a
given region D with the same initial point P and terminal point Q. In this case,
we say the line integral is independent of path in D. In this section, we shall study
path independence and characterize the kind of vector field F for which it
occurs.
A line integral F ∙ <∕R that has
Jc
the same value for any curve C in ■ CONSERVATIVE VECTOR FIELDS
D with given endpoints is said to
be independent of path. The issue of path independence for the line integral F ∙ √R is closely related to
Jc
determining whether or not F is the gradient of some scalar function f. It is useful
to have the following terminology.
Conservative Vector Field A vector field F is said to be conservative in a region D if it can be represented
in D as the gradient of a continuously differentiable function f which is then
called a scalar potential of F. That is,
F = V∕ for (Λ-, y) in D
↑ ↑
Conservative vector field Scalar potential
EXAM PLE 1 Conservative vector field
Verify that the vector field F = 2xyi + x 2j is conservative, with scalar potential
∕ = x 2y.
Sec. 14.3 Independence of Path 91 7
Solution V ∕ = 2xyi + x 2j and this is the same as F, so F is conservative.
As we work this example, two questions might come to mind:
1. Does there exist an easy test to determine if F is conservative?

2. Once such a determination has been made, how can a scalar potential
a. A simply connected region function/be found?
We now answer both these questions.

Before developing a simple criterion for determining whether a vector field is
conservative, we need to introduce some new terminology. Specifically, a region D
is connected if any two points in D can be joined by a path that lies entirely within
D . If, in addition, every closed curve in D encloses only points that are also in Z>,
then D is simply connected. Roughly speaking, a simply connected region is one
b. A region that is not with no “ holes,” as shown in Figure 14.12.
simply connected
Figure 14.12 Connected regions
T H E O R E M 14.2 Cross-partials test for a conservative vector field in the
plane
Consider the vector field F(x, y) = u(x, y)i + v(x, y)j, where u and v have
continuous first partials in the open, simply connected region D . Then F(x, y) is
conservative in D if and only if
= throughout
6 D
∂y ∂x
Proof: The proof is left for Problem Set 14.4, after our discussion of Green’s
theorem. ■
In the next two examples, we use this theorem to show that a given vector field
F is conservative. Then we “ partially integrate” the components of F to obtain a
scalar potential of F such that V ∕= F.
EXAM PLE 2 Finding a scalar potential function
Show that the vector field F = (evsiny - y)i + (ex cosy - x - 2)j is conserva
tive and then find a scalar potential function f for F.
Solution Let w(x, y) = G sin y - y and v(x, y) = ex cosy - x - 2. Then
= e-vcosy - 1 and ∣ x = G co sy - 1
Thus, = and F is conservative. To find a scalar potential function /such

∂y ∂x
that V ∕ = F, we note that f must satisfy u(x, y) = f χ (x, y) and v(x, y) = f y (x, y).
This is the “partial integral” in the sense that

y is held constant while the integration is
performed with respect to x alone.
= ex sin y - y x + c(y)
Where c(y) is a function of y alone— a
“constant” as far as ^-integration is concerned
Because f must also satisfy f y(x, y) = v(x, y), we compute the partial derivative
of this result with respect to y.
f y(χ , τ) = ⅛ [^ s i n >, - yχ + c(τ)] = ~x + ^
de
Set this equal to v = ex cosj , - x — 2 and solve for
de
ex cosy - x + = e-γcosy - x - 2
de __
dy ~
We integrate this to find c(j') = -2 y + C. We now have found the function
f(x, y) = eλ sin y - xy - 2y + C. Any such function is a scalar potential of F,
and for simplicity, we pick C = 0:
f(x , y) = (Asin y - xy - 2y ■—
In Example 2, we began by using the fact that f = ιι. In general, the issue of
whether to start with u = f χ or v = f is a matter of personal taste and is often
determined by which equation leads to the simpler integration.
EXAMPLE 3 Testing for a conservative vector field in the plane
Determine whether or not the vector field F = yexyi + (xexy + x)j is conserva
tive; if it is, find a scalar potential.
Solution We have u(x) = yexy and v(x) = xe xy + x.
~ = xyexy + exy, = xye* , + eχy + 1,
so ≠ and F is not conservative.

∂y ∂x ,
■ FUNDAMENTAL THEOREM ON LINE INTEGRALS
Recall that according to the fundamental theorem of calculus, if the derivative f is

continuous on a ≤ x ≤ b, then
r⅛ rb
d[f(x)] = f ∖x) dx = f(b) - f(μ)
ja Ja
The following is a generalization of this result for line integrals.
THEOREM 14.3 Fundamental theorem on line integrals
Let F be a conservative vector field on the region D and let f be a scalar potential
function for F; that is, V ∕= F. Then, if C is any piecewise smooth curve lying
entirely within D, with initial point P and terminal point Q, we have
( F -d R = f( Q ) - f( P )
Jc
Thus, the line integral I F ∙ dR is independent of path in D.
Jc
Sec. 14.3 Independence of Path 91 9
Proof: We shall prove this theorem for the case where the curve C is smooth in
D, leaving the more general case where C is piecewise smooth as an exercise.
Suppose C is described by the vector function R(t) = x(Z)i + y(Z)j where
a ≤ t ≤ b, and P = R(α), Q = R(⅛). Because
F(Λ , y) = V ∕(x, y) = f x (x, y)i + f y (x, y)j, we have
[ F ∙ r/R = Γ F ∙ ⅜ dt
JC Ja dt
= fa
fb d
= yrt {f∖ x(t), y(Z)]} dt Chain rule (in reverse)
= f[x (b ), X&)] - f[x (a ), y(rz)] Fundamental theorem

of calculus
= ∕[R(Z>)] - ∕[R (α )]
= f(Q )-f(P )
EXAM PLE 4 Evaluating a line integral using the fundamental theorem on

line integrals
Evaluate the line integral F ■√R, where
F = (e*siny - y)i + (e χcosy - x - 2)j and C is the path given by

R(Z) = [t i s in ψ ]i - [ f cos(⅛ + f)]j for 0 ≤ t ≤ 1.
Solution Evaluating this line integral by the param etric method would be both
difficult and tedious. However, we showed in Example 2 that the vector field F
is conservative with scalar potential
∕( x , y) = c'sin y - x y - 2y
and according to the fundamental theorem on line integrals, the value of the
line integral depends only on the value off at the endpoints of the path C. (You
should also verify that the hypotheses of the theorem are satisfied by F and C.)
At the endpoint where t = 0: R(0) = 0i - cos(f)]j = 0i + 0j
/(0 , 0) = e 0 sin 0 - 0 - 0 = 0
At the endpoint where t = 1: R (l) = [s in f ]i - cosπ ] = i + f j
∕ ( l , ∣) = c ' s i n f - 5 - 2 ( f ) = . - ⅝
We now apply the fundamental theorem for line integrals:

∫ c F ■dR = ∕(Q ) - ∕( P ) = /(1 , f) - /(0 , 0)
= (e - ⅞ ) - 0 = e - ⅛ _
EXAM PLE 5 Work along a closed path in a conservative force field
Show that no net work is done when an object moves along a closed path back
to its starting point in a connected domain where the force field is conserva
tive.
Solution In such a force field F, we have V ∕ = F, where/is a scalar potential of F,

and because the path of motion is closed, it begins and ends at the same point
P. Thus, the work is given by
W = f F ∙ dR = f(P) - ∕(P ) = 0 ≡a
Jc
This result appears as part of the proof of the following useful theorem, which
ties together several equivalent conditions for path independence.
THEOREM 14.4 Closed curve theorem for a conservative force field
The continuous vector field F is conservative in the open connected region D if

and only if Jc F ■dR = 0 for every piecewise smooth closed curve C in D.
Proof: If a line integral Jc F ∙ dR is independent of path in the open,

connected region D. then fc F ∙ dR = 0 for any piecewise smooth closed path C in
D. Indeed, if P and Q are two points on such a path, and Cτ is the path from P to Q
along, say, the “top” of C, while CB is the “bottom” path along C from Q to P (see
Figure 14.13), we must have Jcfl F ∙ dR = - J c 7, F ∙ dR and
Figure 14.13 A continuous vector

field F is conservative in the re f F ∙ dR = f F ∙ dR + f F ∙ dR = f F ∙ dR - f F ∙ dR = 0
gion D if and only if Jc F ∙ ι∕R = 0 'c Jcτ JCB JCT JCT
for every closed curve C in D.
Conversely, if Jc F ■dR = 0 for every closed curve C in a region D, it can be shown
(see Problem 47) that F must be conservative in D. ■
We have now stated three equivalent conditions for path independence, which
we summarize in the following box.
Equivalent Conditions for Path Let F(x, y, z) have continuous first partial derivatives in an open connected
Independence region D, and let C be a piecewise smooth curve in D. Then the following
conditions are equivalent.
a. F ∙ dR is independent of path within D.
b. F is conservative; that is, F = V ∕for some function f defined on D.
c. F ∙ dR = 0 for every closed path C enclosing only points of D.

Jc
Here is an example that illustrates how these conditions can be used to

evaluate line integrals.
EXAM PLE 6 Using path independence to evaluate a line integral
Show that F = (xy 2)i + (x2 v)j is conservative and evaluate xy2 dx + x 2y dy
Jc
over each of the given curves:
Sec. 14.3 Independence of Path 921
The purpose of this example is to show that the line integral is 0 if F is

conservative and the path is closed, and if the curve is not closed, then the
value of the integral is independent of the indicated paths.
Solution First, apply the cross-partials test to see if F(x, y) = xy 2i + x 2yj is
conservative.
^ (* T 2) = 2xy -^-(x2y) = 2xy
Thus, F is conservative.
a. Because C is a closed curve and because F is conservative,
xy 2 dx + x 2y dy = 0
Jc
b. Because y = V l - x 2 is the upper portion of the circle x 2 + y 2 = 1, we can

parametrize C by x = cos0, y = sin θ for 0 ≤ θ ≤ %. We represent the
curve as
R(t) = (cos 0)i + (sin 0)j
dR = (-sin θ dθ)i + (cos θ dθ)j
and find that
F = xy2i + x 2yj = (cos θ sin2 0)i + (cos2 0 sin 0)j
We have
xy 2 dx + x 2y dy = F ∙ dR
Jc Jc
rπ∕4
= ⅜sin 20 cos 20 dθ
Jo
1 -> I i t ^δ, 1
=I s>n 20∣ o =⅜
c. Because F is conservative, the line integral is independent of path. Thus,
the result is the same as part b.
J xy 2 dx + x 2y dy = ⅛
It might be instructive to illustrate the power of these equivalent condi

tions by evaluating this same line integral without the benefit of part b. ≡m
In the next section, we develop another criterion for path independence as

part of our study of an important result known as Green’s theorem.
PROBLEM SET 1 4 .3
θ 1. ■ What Does This Say? Discuss conservative vector
fields.
2. ■ What Does This Say? Describe the fundamental the
orem on line integrals.
3. ■ What Does This Say? Compare and contrast the
various equivalent conditions for independence of path.
Determine whether or not each vector field in Problems 4 -9 is
conservative, and i f it is, find a scalar potential.
4. y 2i + 2xyj 5. 2xy3i + 3y2x 2j
γl vl
6. (xe 'siny)i + (e cosy + y)j
7. ( - y + e t siny)i + [(* + 2)e2cosy]j
8. (y - x 2)i + (2x + y 2)j 9. (e2 ⅛iny)i + (e¾ osy)j
Evaluate the line integrals in Problems 10-13 for each o f the

given paths.
10. I (3x + 2y)dx + (2x + 3y)dy
Show that the vector field F in Problems 14-19 is conservative,

find, a scalar potential f for F, and evaluate the line integral fc
F ■<7R, where C is any path connecting (0, 0) to (1, 1).
14. F(x, y) = (x + 2y)i + (2x + y)j
15. F(x, y) = 2xyi + x 2j
16. F(x, y) = (y - x 2)i + (x + y 2)j
17. F(x, y) = (2x - y)i + (y 2 - x)j
18. F(x, y) = [3x 2cosy + 7 7 ⅞ y ] i + [-x ⅛ in y + 1 √ λ
y 2p 2]j
19. F(x, y) = [2 cos2x - e - *(cosxy + y sinxy)]i - [xe~j⅛inxy]j

Sec. 14.3 Independence of Path 923
20. A force field It can be shown that the vector field

F(x, y, z) = M(x, y, z)i + N(x, y, z)j + P(x, y, z)k
F(x, y) = (3x 2 + 6xy2)i + (6x 2y + 4y2)j is conservative i f there exists a scalar function φ(x, y, z) such
acts on an object moving in the plane. Show that F is that F = Vφ. It can be shown that under reasonable conditions
conservative, and find a scalar potential for F. How much F is conservative i f curl F = 0.
work is done as the object moves from (1,0) to (0, 1) along
any path connecting these points? (See also Problem 45.) Use this result to verify that each vector
field F in Problems 32-37 is conservative.
Verify that each line integral in Problems 21-26 is independent 32. yz 2i + x z 2j + 2xyzk
o f path and then find its value. 33. exyyzi + exyxzj + exy k
34. yz~'i + xz - l j — xyz~ 2k
21. f (3x 2 + 2x + y 2) dx + (2xy + y 3) dy
35. (x 2 + y 2 + z 2)(xi + yj + zk)
Jc
where C is any path from (0, 0) to (1, 1) 36. (y sin z)i + (x sin z + 2y)j + (xy cos z)k
37. (xy2 + yz)i + (x 2y + x z + 3y2z)j + (xy + y 3)k
22. (xy cosxy + sinxy) dx + (x 2 cosxy) dy 38. a. Over what region in the xy-plane will the line integral
Jc
where C is any path from (θ, j⅞) to ( 1, f) [(-y x - 2 + x~ ')dx + x ~ λ dy}
Jc
23. f (y - x 2)dx + (x + y 2) dy be independent of path?
Jc b. Evaluate the line integral in a if C is defined by
where C is any path from ( - 1, -1 ) to (0, 3)
R(Z) = (cos3Z)i + (2Vsec Z)j for 0 ≤ t ≤
24. ∣(3x 2y + y 2) dx + (x 3 + 2xy) dy
Jc 39. The gravitational force field F between two particles of
where C is the path given parametrically by masses M and m separated by a distance r is given by
R(∕) = Zi + (t 2 + t —2)j for 0 ≤ t ≤ 2
G(x, y, z) = -K rpfif r
25. I (siny) dx + (3 + x cosy) dy
Jc
where C is the path given parametrically by where R = xi + yj + zk and K is the gravitational con
R(Z) = 2 sin(πZ∕2) cos(πZ)i + (sin- 1 Z)j for 0 ≤ t ≤ 1 stant.
a. Show that F is conservative and find a scalar potential
26. I (ex cosy) dx + ( - ex siny) dy for F. The scalar potential function f is often called the
Jc Newtonian potential.
where C is the path given parametrically by b. Compute the amount of work done against the force
R(Z) = (cos Z)i + (sin Z)j for 0 ≤ t ≤ f field G in moving an object from the point P(α 1, b v c 1)
to Q(av b2 , c2).
Evaluate the line integrals given in Problems 27-31 using the
fundamental theorem o f line integrals. 40. Let F ( x j )—
=vi ≠+ ^xi .
27. f (yi + xj) ∙ √R, where C is any path from (0, 0) to (2, 4)
Jc a. Compute the line integral Jc F ∙ <7R, where C is the
upper semicircle y = V l - x 2 traversed counterclock
28. f (xy2i + x 2yj) ∙ t/R; C is any path from (4, 1) to (0, 0) wise. What is the value of Jc F ■<7R if C is the lower
Jc semicircle y = - V 1 - x 2 also traversed counterclock
wise?
29. I (2y dx + 2x dy), where C is the line segment from (0, 0)
Jc b. Show that if F = Mi + 7Vj, then
to (4, 4) ∂M ∂N
=
∂y ∂x
30. (ex siny dx + e'cos y dy), where C is any smooth curve
Jc but F is not conservative on the unit disk x 2 + y 2 ≤ 1.
from (0, 0) to (0, 2π) 41. A person whirls a bucket filled with water in a circle, at
the rate of 1 revolution per second. If the bucket and
31∙ ∕c [ta n 1 (x ) - x 2 + y 2 ∖d x + [x 2 + y 2 + e ? (1 ^^ ∙y ψ ∙ y water weigh 30 lb and are whirled horizontally with the
center of gravity 3 ft away from the person, how much
where C is the curve given by R(Z) = 1- ', (cos t i - t sin t j) work is done by the force that keeps the bucket moving in
for 0 ≤ t ≤ 1 a circular path?
this case, you can eliminate b and do a one-

Computational Window parameter study.
d. Assume that the path in part c cannot go above
In Problems 42 and 43, experiment with computing work. ; y = 1. You may have discovered the optimal path.
along a path, with and without the benefit o f independence ∣ What is it? If you did the parabolic study suggest
o f path. Suppose that you are to power a boat from point ed in part c, what is the optimal path in this case?
A — (0, 0) to point B = (2, 1J, and the primary considera
tion is the force o f the wind that opposes you. Investigate
the effect o f taking different paths from A to B. θ 44. Show that if the vector field
42. a. Suppose that the wind force is F = (-a , -b}, for a F(x, y, ∑) = M(x, y, ∑)i + N(x, y, z)j + P(x, y, z)k is
and b positive. Compute the work involved along conservative, then curl F = 0; that is,
the straight-line path between A and B; then along ∂P=∂N ∂M =∂P ∂N _ ∂M
a second path, of your choice. Does the path ∂y ∂∑ ∂z ∂x ∂x ∂y
matter here? Why or why not? 45. It can be shown that F = M i + Nj + Pk is conservative
b. Due to the effect of the harbor you are entering, if curl F = 0 whenever M, N, and P have continuous
y
suppose the wind force is {-a, -ae~ '). Again, partial derivatives in a ball, a box, or other “ simply
compute the work along two paths as in part a. connected” regions. (The precise definition of a simply
Does the path matter here? Why or why not? connected region in R 3 is a bit technical and will be
c. Repeat part b for (-α , -ae~ y+xl9 ). omitted.) Show that the vector field defined by
F(x, y, z) = (y2 - 2xz)i + (2xy + z)j + (y - Λ2)k
43. An important type of problem is, “ Can we find the
is conservative and complete the following
optimal path in order to minimize the work?” You
steps to find a scalar potential function f(x , y, z):
are to explore this issue in regard to the wind force in II
a. Note that if V ∕= F, we must have
Problem 42c, F = {-a , -ae~ y+xl9 }.
a. You should have the work computed for two ' f x = y 2 - 2xz,f y = 2xy + z ,f z = y - x 2
different paths from the last problem. By studying Partially integrate f χ with respect to x to express f in
F carefully, note that we will be rewarded (or terms of x, y, z, and a function c(y, z) that acts as a
punished) by changing the path slightly from the “ constant” with respect to .^-integration.
straight-line path. What do your observations b. Find f and set it equal to N = 2xy + z. What can you
suggest regarding trying to minimize the work? conclude about the function c(y, z)? Partially integrate
b. Attempt to find an (approximate) optimal path, with respect to y to express f in terms of x, y, z and a
starting with your observations in part a and function of z alone.
common sense. For example, the path could be c. Find the scalar potential f.
described by a parabola (or higher degree polyno 46. Let f and g be differentiable functions of one variable.
mial) or by some trigonometric function. Find a Show that vector field F = [ ∕( x ) + y]i + [g(y) + x]j is
“ good” path for this purpose. conservative; find the corresponding potential function.
c. Explore the following question: “ Is there a realis 47. Complete the proof of Theorem 14.4 by showing that if
tic optimal path from A to B T , Consider two fc F ∙ <7R = 0 for every closed curve C in a domain D
conditions: first, there is no restriction on the where F is continuous, then F must be conservative in D.
path; second, suppose there is a shoreline at y = 1 Hint: Let C 1 and C 2 be two curves in D with the same
so that one’s path cannot exceed this limit. Hinf. endpoints P and Q. Define a closed curve C so that
consider all parabolic paths of the form
I F ∙ √R = I F ∙ c/R - F ■<r∕R
y = x{b -ax) Jc Jc, Jc2
and conclude that fc F ∙ dR is independent of path, so
where a and b are nonnegative, and y(2) = 1. In
that F is conservative in D.
14.4 GREEN’S THEOREM
IN THIS SECTION Green’s theorem, area as a line integral, alternative

forms of Green’s theorem
r b ∏rf
The fundamental theorem of calculus, ^^χdχ = F(b) ~ Hα), c a n t,e described
'a
as saying that when the derivative is integrated over the closed interval
a ≤ x ≤ b, the result is the same as that obtained by evaluating Fix) at the
“boundary points” a and b and forming the difference F)b) —F(a). An
analogous result for functions of two variables is introduced in this section. It is
called Green’s theorem after George Green (Historical Note, page 925).
Sec. 14.4 Green’s Theorem 925
■ GREEN’S THEOREM
Green’s theorem relates the double integral of a certain differential expression

over a region in the plane to a line integral over a closed boundary curve of the
region. We begin with some terminology. A Jordan curve, named for the French
mathematician Camille Jordan (1838-1922), is a closed curve C in the plane that
does not intersect itself (see Figure 14.14). We can redefine a simply connected
region D in terms of a Jordan curve C. A region is simply connected if it is
connected and the interior of every Jordan curve C in D also lies in D, as shown in
Figure 14.14.
Person is walking in the
direction of increasing t.
A Jordan curve A simply connected region
Not a Jordan curve Not a simply connected region; A positively oriented boundary
a Jordan curve in D whose interior curve. The region is “on the left”
contains points not in the region of the curve.
Figure 14.14 A Jordan curve is a closed curve with no self-intersections. A region is simply
connected if every Jordan curve has all its interior points in D.
Picture yourself as a point moving along a curve. If the region D stays on your left
as you, the point, move along the curve C with increasing t, then C is said to be
positively oriented (see Figure 14.14.)
THEOREM 14.5 Green’s theorem

Let D be a simply connected region with a positively oriented piecewise-smooth
boundary C. Then if the vector field F(x, y) = M(x, y)i + N(x, ρ)j is continuously
differentiable on D, we have
Historical Note ∫c (Λ∕Λ + Λ-Λ) = ∫ J (g-⅜ *)< M
GEORGE GREEN (1793-1841)
George Green was a self-taught
19th-century mathematician M IVhat this says: This theorem expresses an important relationship between
whose ideas were far ahead of a line integral over a simple closed curve in the plane and a double integral
his time. He graduated from
over the region bounded by the curve. It is one of the most important and
Cambridge in 1837 and, as an
undergraduate, wrote a brilliant elegant theorems in calculus. Take special note that D is required to be simply
paper that included many im connected with a positively oriented boundary C. This condition is so impor
portant results in mathematical tant that sometimes the notation
physics. Perhaps because he
was still an undergraduate, the (£ (M dx + N dy) or Λj (M dx + N dy)
paper was largely unnoticed un Jc Jc
til it was discovered by Sir Wil is used to indicate the line integral in order to emphasize the positive (or
liam Thompson (later known as counterclockwise) orientation.
Lord Kelvin).
Proof: A standard region is one in which no vertical or horizontal line can

intersect the boundary curve more than twice (see Figure 14.15). We shall prove
Green’s theorem for the special case where D is a standard region, and then we
shall indicate how to extend the proof to more general regions.
Suppose D is a standard region with boundary curve C. We begin by showing
that
∫ ∫ -d f } d x d y = M dx
'c
÷------------- - ------------1-
a b
Because D is a standard region, as shown in Figure 14.15, the boundary curve C is
composed of a lower portion CL and an upper portion Cu , which are the graphs of
Figure 14.15 Standard region: No functions f l (x) a n d ∕ 2(x), respectively, on a certain interval a ≤ x ≤ b. Then we
vertical or horizontal line inter can evaluate the double integral by iterated integration:
sects the boundary more than
twice. ∣I dx dy = ( [ 'l.'''l' dy dx = [ [ dyx dx
Jn J ∂y ■ J J ∂y x Ja [J/|(x) ∂y -J
rZ? rb C C
M ∖x, f 1(x)] dx - M[x, ∕ 1(,x)] dx = M dx - M dx
la Ja J-Cu JCL
=- f M dx + f M dx
_Jcu JCL
A similar argument shows that j^ j dx dy = [ N dy. Thus,
dx ’ Jc
∂N ∂M ^ d x d y -∣
J ^d x d y
∂x ∂ y,
f N dy - f ( - M ) d x = f (M dx + N dy)
'c Jc Jc
Figure 14.16 General case: The This completes the proof for the standard region. If D is not a standard region, it
region is decomposed into a finite can be decomposed into a number of standard subregions by using horizontal and
number of standard regions by vertical “cuts,” as shown in Figure 14.16. The proof for the standard region is then
cuts. applied to each of these subregions and the results are added. The line integrals
along the cuts cancel in pairs, and after cancellation, the only remaining line
integral is the one along the outer boundary C. Thus,
f c (M d x + N dy) = f ∫ ( ≡ - ⅞ ^ ) dA
R
The case where there is one cut is considered in Problem 38. ■
EXAMPLE 1 Verifying Green’s Theorem
Verify Green’s theorem for the line

integral (- y dx + x dy), where
Jc
C is the closed path shown.
Solution First, evaluate the line integral directly. The curve C consists of the
portion of the x-axis from (-1 , 0) to (1, 0) followed by the semicircular arc
from (1,0) back to ( - 1, 0). The line segment is called C 1 and the semicircle is
called C 2 . We parametrize each of these:
C 1: x = t, y = 0 -1 ≤ t ≤ 1
dx = dt, dy = 0
Cy. x = cos s, y = sin 5 0 ≤ 5≤ π
dx = -s in 5 ds, dy = cos 5 ds
( - y dx + x dy) = ( - y dx + x dy) + ( - y dx + x dy)
Jc Jc, ' Jc,
r1 r∏
= [-0fi⅛ + Z∙0] + [-sin s(-sin 5 ds) + cos s(cos s <⅛)]
J-1 Jo
= | (sin2s + cos⅛) ds = ττ
Jo
Next, we use Green’s theorem to evaluate this integral. Note that the
boundary curve C is simple and M = - y , N = x so that F(x, y) = -p i + xj is
continuously differentiable. The domain of D is given by 0 ≤ y ≤ V I - x 2
for -1 ≤ x ≤ 1. We now apply Green’s theorem:
∫c ( ~ y d x + λ ' d y) = ∫ ∫ ( ⅛ w - = ∫ 1 ∫0
2 dy d x
= 2(AREA OF SEMICIRCLE) = 2[∣77( 1)2 ] = 77
EXAMPLE 2 Computing work with Green’s theorem
A closed path C in the plane is defined

by the given figure. Find the work done
by the force F as the object moves
along C in the force field
F(x, y) = (x + xy 2)i + 2(x2p - y 2 sinp)j
Solution The work done, W, is given by the line integral Jc F ■r∕R. Note that F is
continuously differentiable on I) and that D is a simply connected region with
a positively oriented boundary C; thus, the hypotheses of Green’s theorem are
satisfied. We find that
W = ∫ c F ■√R = ∫ ∫ [ ⅛ 2 χ 2 Jz - 2p2sinp) - ^ ( x + xy 2)] dA
= [ ∣(4xy - 2xy) dA = 2 ∣ xy dy dx = 2 ∣ dx
J DJ
= |o (x ~ χ5) d x = ⅛ 2 - ⅛χ 6 ]0 = |
■ AREA AS A LINE INTEGRAL
A line integral can be used to compute an area of a region in the plane by applying
the following theorem.
THEOREM 14.6 Area as a line integral

Let D be a simply connected region in the plane with piecewise-smooth closed
boundary C. Then the area A of region D is given by the line integral
A = 5z
( - y dx + x dy)
Jc
Proof: Suppose C is a closed curve that encloses a region D, as shown in Figure
14.17.
Let F(x, y) = -y i + xj. Then, if F is continuously differentiable on D, Green’s

theorem applies. We have
∫ c ( - y dx + x dy) = ∫ ∫ [⅛ (x) - y)] dA = 1 1 2dA = 2A
Figure 14.17 Area of region D

so that A = ∣ ( - y dx + x dy).
Jc H
0 This gives us yet another technique for finding the area of a region,
especially when its boundary is specified in parametric form. Do not forget
the factor one-half after you finish the integration. 0
EXAMPLE 3 Area enclosed by an ellipse
x2 y 2
Show that the ellipse —1 + 75 = 1 has area ττab.
a b-
Solution The elliptical path E is given parametrically by x = a cos θ, y = b sin θ
for 0 ≤ θ ≤ 2π. We find dx = - a sin θ dθ, dy = b cos θ dθ. If A is the area of
this ellipse, then
Historical Note A =I (- y dx + x dy)

Jc
Green’s work significantly in c2ττ
fluenced the work of later =2 I [- (^ s ' n 0)(- f l s >n 0 dθ) + (α cos θ)fb cos θ dθ)
mathematical physicists at
Cambridge, including William r2ττ
Thompson (Lord Kelvin), =2 I ab(cos~θ + sin2 0)√0
James Clerk Maxwell, and Ga c2ττ
briel Stokes. The theorem = ∣ ab dθ Because cos2 0 + sin 2 0 = 1
known as Green’s theorem was Jo
proved independently by
Thompson and Mikhail = 2<26(2TΓ - 0) = τrab
Ostrogradsky, a Russian mathe
matician. Green’s work also in ■ ALTERNATIVE FORMS OF GREEN’S THEOREM
fluenced Karl Gauss in the de
velopment of Gauss’s theorem, Green’s theorem can be expressed in a form that generalizes nicely to R3. In
which is also called the diver
particular, note that the curl of the vector field F(x, y) = M(x, y)i + N(x, y)j is
gence theorem.
given by
i j k
curl F = ∂ ∂ ∂
∂x ∂y ∂z
M (x, y) N (x,y) 0
= f-θ2Y∖ f∂A∆∙ Γ∂N ∂M Γ∂N aM ^ ∣
∂y J k Because M and N are
+ +
∂z) ∖ ∂ z) i Lax ay. L∂x
functions o f only x a n d y
Thus, the formula in Green’s theorem can be expressed as
In Section 14.6, when we extend this result to surfaces in R 3, it will be called

Stokes’s theorem.
EXAM PLE 4 Relationship of the double integral of div F to the line

integral of F ∙ N
Suppose F(x, y) = M(x, y)i + N(x, y)j with a piecewise-smooth boundary C.

Show
∣ F ∙ N ds = f ∣ div F dA
Jc J J
D
This is the line integral of F ∙ N, the normal component of F.

Solution Let R(.s,) = x(.v)i + y(.v)j so that a unit tangent vector T to the curve C is
T = R'(v) = x'(s)i + y'(s)j, which means that an outward normal vector is
N = y ' ( s ) i - x , (.v)j. We now apply Green’s theorem to find a representation
using div F.
Γ
F ∙ N ds =
cb
(Mi + Aj) ∙ [y'(s)i - x'(s)j] ds
'c Ja
= Γ TS~ N ls ) ds = f ( ~N dx + M d -d
Ja × ' Jc
= fj ( ^ + f f l d x d y =f j dlvFdA
When we extend this result to IR3 in Section 14.7, it will be called the divergence
theorem.
We repeat these alternative forms of Green’s theorem for easy reference.
Alternative Forms of Green’s Let D be a simply connected region with a positively oriented boundary C.
Theorem Then if the vector field F = M i + Nj is continuously differentiable on D , we
have
F ∙d R = f (M dx + N dy) = f f ( ≡ - ^ ) d A = ∣ f (curl F ∙ k)r∕.4
' p~' ∙>c JD J V ∙7 J J
D
Tangent component of F
,≡
D
Normal component of F
In physics, some important applications of Green’s theorem involve the so-

called normal derivative of a scalar function/ which is defined as the directional
derivative o f/in the direction of the outward normal to some curve or surface.
Normal Derivative The normal derivative o f/is denoted by ∂f∕∂n and is the directional derivative
of/ in the direction of the normal pointing to the exterior of the domain of/
The normal derivative of/ satisfies the property
I= V ∙N
where N is an outer unit normal.
The following example illustrates how Green’s theorem can be used in

connection with the normal derivative. Additional problems are found in the
problem set.
EXAMPLE 5 Green’s formula for the integral of the Laplacian
Suppose / is a scalar function with continuous first and second partial

derivatives in the simply connected region D. If the piecewise-smooth curve C
bounds D, show that
ds
D
'∖z,
where V2∕ = f χχ + f yy is the Laplacian of / and = V ∕∙ N is the normal
derivative.
r)∕, r)f
Solution Let u = ∂y and v = χ~.
∂x Then we have
W θy _ ∂u
j =j f +j f =
xx yy ∂x ∂y
∫∫V≈Λ*<fy =∫∫(g-⅛) dx dy = u dx + v dy Green's theorem

D D
dx +, dy
" * v * ds
∂ f d x + ∂fdy∖
∂y ds ∂x ds)
dy dx
Where N = i - j is an outward unit normal to C
Sec. 14.4 Green's Theorem 9 31
PROBLEM SET 1 4 .4
Evaluate the line integrals in Problem 1-6 by applying Green’s
theorem. Check your answer by direct computation (that is, by
parametrizing the curve C).
13. Use Green’s theorem to find the work done by the force
field F(x, yj = (3y - 4x)i + (4x - v)j when an object
moves once counterclockwise around the ellipse
4x 2 + y 2 = 4.
14. Find the work done when an object moves in the force
field F(x, y) = y 2i + x 2j counterclockwise along the circu
lar path x 2 + y 2 = 2.
Use Theorem 14.6 to find the area enclosed by the regions
described in Problems 15-18 and then check by using an
appropriate formula.
15. Circle x 2 + y 2 = 4
16. Triangle with vertices (0, 0), (1, 1), and (0, 2)
17. Trapezoid with vertices (0, 0), (4, 0), (0, 3), and (1,3)
18. Semicircle x = Vz4 - y 2
Evaluate the line integral in Problems 7-12 by applying 19. Evaluate the line integral
Green’s theorem.
(x 2y dx —y 2 x dy)
where C is the boundary of the region between the x-axis

and the semicircle y = V α 2 - x 2, traversed counterclock
wise.
20. Evaluate the line integral
(3y dx - 2x dy)
where C is the cardioid r = 1 + sin 0, traversed counter

clockwise.
21. Show that 29. Evaluate I ds, where z(x, p) = 2x 2 + 3y2 and C is the
Jc ∂n y j
(5 - xy - p 2) dx — (2xy — x 2) dy = 3x circular path x 2 + p 2 = 16, traversed counterclockwise.
where C is the square 0 ≤ x ≤ 1, 0 ≤ y ≤ 1 traversed Γ r)f

30. Evaluate I — ds, where ∕( x , y) = x 2y — 2xy + y 2 and C
counterclockwise and x is the x-coordinate of the centroid
of the square. is the boundary of the unit square
22. Find the work done by the force field 0 ≤ x ≤ 1, 0 ≤ y ≤ 1, traversed counterclockwise.
F(x, p) = (x + 2p2)j
31. Evaluate j x ds, where C is the boundary of the unit
as an object moves once counterclockwise about the circle
(x - 2) 2 + y 2 = 1. square 0 ≤ x ≤ 1, 0 ≤ y ≤ 1, traversed counterclock
23. Let D be the region bounded by the Jordan curve C, and wise.
let A be the area of D. If (x, p) is the centroid of D, show 32. If C is a closed curve, show that
that
[(x - 3p) dx + (2x - y 2) dy] = 5A
Ax = I I x 2 dy
r
where C is traversed counterclockwise. where A is the area of the region D enclosed by C.
24. Use Theorem 14.6 and the polar transformation formulas 33. SPY PROBLEM After escaping from Purity’s watery trap
x = r cos θ, y = r sin θ to obtain the area formula in polar (Problem 59, Section 13.7), the spy takes a deep gulp of
coordinates, namely, fresh air and enters the next room. The door slams shut
behind him, and he finds himself looking across a large,
1 r¾ 2 ι f β2
A =j r dθ = 5 [g(0)]2 dθ rectangular room at Purity and the arch fiend, Andre
2 Jel Jθl Scelerat. The spy takes a step toward his nemesis, but
finds his movements restricted by a force field that
appears to sap his strength. Scelerat laughs cruelly and
says, “I see you have noticed the Death Force my lovely
assistant has designed for you. I hope you appreciate her
efforts.” Scelerat takes Purity’s arm and leads her from
the room, saying, “Goodbye, old friend, and by the way,
it will do you no good to stay still, because the rays given
off by the Death Force will eventually kill you even if
you never move.” One more peal of laughter and the spy
is left alone. Quickly, he presses the stem on his wrist
watch and sees the equation of the force field appear on
the face:
25. —%■+ - ) dx + - dy , where C is any Jordan
.x 2 x ) x ∙ j
F(x, p) = [peλ'i ' + 2xp 3 ]i + [xexy + 3x 2p 2 + cosp]j
curve that does not touch or cross the p-axis, traversed
counterclockwise. Assuming that the spy is at (0, 0) on the wristwatch’s
coordinate system and that the door (and safety) is at
„_ _ , C x d x + y dy „ . , (5, 4), what path should the spy take in order to minimize
26. Evaluate --χ-2=-----≡2 - , wh1ere C is any Jτo rd an curve
]c +y the work of struggling against the Death Force while
whose interior does not contain the origin, traversed crossing the room in the least possible time?
counterclockwise. © 34. Prove the following theoretical application of Green’s
, , f - y d x + (x - 1) dy , _. , theorem: Let F(x, y) = u(x, p)i + v(x, p)j be continuously
27. τ Evaluate ——------—-2----- s—
2
~ , where C is any Jτordan differentiable on the simply connected region D. Then F
Jc (x -l) + y
is conservative if and only if
curve whose interior does not contain the point (1, 0),
traversed counterclockwise. ∂v _ ∂u
∂x ∂y
- (J2 + 2) dx + (x - 1) dy
τ, . f ,
28. Evaluate ( x _ 1)2 + ( j + 2 ) 2 . where C is any throughout D.
Jordan curve whose interior does not contain the point 35. Recall that a scalar function f with continuous first and
(1, -2 ), traversed counterclockwise. second partial derivatives is said to be harmonic in a
Sec. 14.5 Surface Integrals 933
region D if V 2∕ = 0 (that is, i f f χ χ + f y y = 0). If/ is such a 38. Prove Green’s theorem in the plane for the region D
function, show that shown in Figure 14.18.
J J ( f x 2 + f y 2) d x d y = ^ f ^ d s
where C is the boundary of D.

36. Derive the following result, known as Green’s first
identity.
J ∫ [ ∕V ⅛ - g V 2 ∕ ] dx dy = f ( f d∕ n ~ g ^ ) d s
J DJ
Hint: Start with the line integral. Figure 14.18 Green’s theorem
37. Derive the following result, known as Green’s second
identity. The line L “ cuts” the region into two standard subregions
D χ and //.
[ ∕V ⅛ + V ∕∙ Vg] dx dy = I ∕ ∣ f ds
JC Uri
Hint: Start with the line integral.
1 4 .5 SURFACE INTEGRALS
IN THIS SECTION Terminology, surface integrals, surface integrals of

vector fields, parametrically defined surfaces
A surface integral is a generalization of the line integral in which we integrate
over regions in space bounded by a surface. The rest of this chapter deals with
surface integrals. We introduce the concept in this section, and then in the next
two sections we shall use surface integrals to generalize Green’s theorem to R 3.
■ TERMINOLOGY
Recall that line integrals are defined only along oriented curves that are piecewise
smooth and have no self-intersections, and there are similar restrictions on the
nature of the surfaces used in the study of surface integrals. First, recall that a
surface is smooth if it has a unit normal vector N(x, y, z), that varies continuously
as the point (x, y, z) ranges over the surface, and it is piecewise smooth if it consists
of a finite num ber of smooth pieces. For instance, a sphere is a smooth surface,
and a cube is piecewise smooth because it consists of six smooth faces, with
adjacent faces meeting on an edge where it is not smooth.
A is a surface defined by A surface is orientable if it is possible to choose a unit normal N at each point
z = g(χ, y)∙ on the surface S so that N varies continuously over A. A closed surface is one that
bounds a solid region. The region enclosed by 5 is called its interior and the region
outside A is its exterior. The normal N is an outer normal of A if it points toward
the exterior of A and inner normal if it points toward the interior. These terms are
illustrated in Figure 14.19. Most common surfaces, such as planes, spheres,
ellipses, and paraboloids are orientable.
We are not interested in piecewise-smooth surfaces with only one side, such as
the Mobius strip, which is obtained by twisting a strip of paper and then attaching
the ends together (see Figure 14.19b). Henceforth, we shall use the word surface to
mean an orientable, piecewise smooth surface.
a. Two-sided (orientable) surface: a football. b. One-sided (non-orientable) surface:

This is also a closed surface. a Mobius strip
Figure 14.19 Surfaces in space
A boundary point p of a set of points A in R 3 has the property that any sphere
centered at p (no matter how small) contains points in S and also points not in A.
The boundary of S is the collection of all its boundary points, and S is closed if it
contains its boundary. For example, the solid spherical region x 2 + y 2 + z 2 ≤ r2 is
closed because it contains its boundary x 2 + y 2 + z 2 = r2 . A solid region is
bounded if it can be contained inside a sphere of finite radius.
■ SURFACE INTEGRALS
Now we can define the surface integral of a scalar function g. Suppose g is defined
and continuous on a surface S. Partition A into n subregions, the kth of which has
area ΔS j,., and let P'k be a point chosen arbitrarily from the Zcth subregion, for k = 1,
2, . . . , n. Form the sum
∑ι M ¾ Δ ⅞
and take the limit as the largest of the Δ5 A, tends to 0, just as we did with the
surface area in Section 13.4. If this limit exists, it is called the surface integral o fg
over 5 and is denoted by
f ∣ g(χ, y , z) dS
Definition of a surface integral
Recall from Section 13.4 that when the surface S' projects onto the region 2?vv
in the xj~plane and S has the representation z = ∕( x , y), then
dS = j f 2 + f 2 + l dA χy
x y
where dΛ -∖V is either dx dy or dy dx (or r d-V

r dθ if √4 is described in terms of
polar coordinates). We can now state the formula for evaluating a surface integral.
∖Note∖We write R γp and A χy instead of the usual R and A to help you remember
that these are regions in the xy-plane.)
Surface Integral Let S’be a surface defined by z = f( x , p) and R χv its projection on the xy-planc.
l f f , f x , a n d f are continuous in R χγ and g is continuous on S’, then the surface
integral of g over S’ is
∫ ∫ g(x, y ,z ) d S = ∣ ∣ g(x, y ,f ( x , y))>∕[fx (x, y)]2 + [f y (x, y)]2 + 1 dA x y
$ &xy
EXAM PLE 1 Evaluating a surface integral
Evaluate the surface integral
where g(x, y, z) = xz + 2x 2 - 3xy and A is that portion of the plane

2x - 3y + z = 6 that lies over the unit square.
R . 0 ≤ x ≤ l, 0 ≤ y ≤ 1
Solution First, note that the equation of the plane can be written as z = f(x, y),
where f(x ,y ) = 6 - 2 x + 3y. We have f χ (x, y) = - 2 and f( x , y) = 3, so that
The portion of the plane that lies dS = y∣
f x 2 + f y 2 + l dAχy = √ (-2) 2 + (3)2 + 1 dAχy = √14 dAχy
above the unit square Consequently,
f ∫ g d S = I f (xz + 2x 2 - 3xy)V14 dAχy
JS J JS J
[x(6 - 2x + 3y) + 2x2 - 3xy]V14 dy dx

RJ
-l fl 1
J
o Jo
x dy dx = βV14 I x dx = 3V14
Jo
If the function g defined on 5 is simply g(x, y,z) = 1, then the surface integral
gives the surface area of S.
Surface Area Formula Surface area =
A useful application of surface integrals is to find the center of mass of a thin

curved lamina whose shape is a given surface S, as shown in Figure 14.20.
Figure 14.20 A thin lamina whose shape is the surface S
Suppose ρ(x, y, z) is the density (mass per unit area) at a point (x, y, z) on the
lamina. Then the total mass, m, o f the lamina can also be represented by a surface
integral.
M ass of a Lamina m= ∖ p(x:, y, z) dS
s
and the center of mass of the surface is the point C(x, y, z) where
⅛ ∫ jχ p ( χ , y , z )d S ,
These formulas may be derived by essentially the same approach used in our
previous work with moments and centroids of solid regions. (See, for example,
Sections 6.3 and 10.3.)
E X A M P L E 2 Mass of a curved lamina
Find the mass of a lamina of density ρ(x, y, z) = z in the shape of the

hemisphere z = (a2 - x 2 - j ,2 )ιz2 ∙
Solution We begin by calculating dS.
zχ = ∣(α 2 ~ x 2 ~ y 2)~ l '2 (-2 x ) = - x ( a 2 - x 2 - y 2)-ιz 2
zy = ⅛(fl2 - χ 2 - y 2)~ i,2 (~2y) = -y(<a2 - χ 2 - y 2)~ιz2
d s = y∣z 2 + z 2 + 1 dA = J χ2, ~ 2 + — ∖
v x xy 2 + 1 dA
y ∖ a 2 - x 2 - y 2 a2 - x 2 - y 2 xy
=^ 2 - x 2 - y2 dλχy = a { a ~~ χ λ ~ dA xy
a
The mass of the hemisphere 5 is given by
Figure 14.21 Note that 5 projects

onto the circle R in the xv-plane
with the equation x 2 + y 2 = a 2.
■ SURFACE INTEGRALS OF VECTOR FIELDS
Surface Many applications of surface integrals require the integral of the normal compo
nent of a given vector field F — that is, an integral of the form ∫ f F ∙ N dS, where
N is the outward unit normal to the surface 5. Because F ∙ N is a scalar function,
this type of surface integral is really the same as the kind we have already studied,
but it is still useful to examine a few situations that involve surface integrals of this
form. Consider the following example.
EXAM PLE 3 Surface integral of a vector field
Compute J J F ∙ N dS. where
F = xi —5yj + 4zk, and 5 is the union of

the two squares
S ,: x = 0, 0 ≤ y ≤ 1, 0 ≤ z ≤ 1
S ,√ z = 1, 0 ≤ x ≤ 1, 0 ≤ y ≤ 1
Note the orientation of the axes.

Solution On 5 1, the outward unit normal is N 1 = - i and

F ∙ N l = (xi - 5yj + 4zk) ∙ (-i) = - x = 0
(because x = 0 on 5 l ). Thus,
F ∙ N 1 dSi = 0
On S 2 , N 7 = k, and F ∙ N 2 = F ∙ k = 4z = 4 (because z = 1 on 5'2). Thus,

4 dS 2 = 4(area of ∙S',) = 4
Finally,
=0+4=4
The unit normal of a surface is not always as obvious as it was in Example 3.

For an orientable surface S defined by z = g(x, y), we let G(x, y , z) = z - g(x, y).
Then, N can be found as follows (see Section 12.6):
VG(x, y, z) _ S x i - g y i + k
Outward unit normal
∣∣VG(x, y, z)∣∣ yJg χ ^ + gy 2 + 1
- V G ( x , y, z) _ + ~k
Inward unit normal
∣∣VG(x,y,z)∣∣ √ g ∕ + g ∕+ l
EXAMPLE 4 Surface integral of a vector field
Compute F ∙ N dS, where

JSJ
F = xyi + zj + (x + y)k
and .S' is the triangular region cut off from
the plane x + y + z = 1 by the positive
coordinate axes, as shown in Figure
14.22. Assume N is the unit normal that
points away from the origin.
Solution Let g(x, y) = z = l - x - y . Then

gχ = - 1 , g ι, = - 1 , and an outward unit normal to the plane is
_ ( _ l) i_ ( _ l) j + k J _
N = V ( - l ) 2 + ( - l ) 2 + 1 = √3 j '
so that
i + + k )= ^ xy + z + χ
F ■N = [xyi + zj + (x + y)k] ( ⅛ ⅛ j ⅛ + y)
We can write this as a function of x and y because z = 1 - x - y:
F ∙ N = ^ j ( x y + l - χ - y + χ + y) = - f y χ y + 1 )
We also find that

dS = ^ g x 2 + gy 2 + 1 dA χy = √3 dA χy
The final piece of the puzzle we need before turning to the evaluation of
the integral is to find the projection of 5 onto the xy-plane. We see from Figure
14.23 that this is a triangular region. Let us call this region T, which can be
described as the set of all points (x, y) such that for each x between 0 and 1, y
varies between y = 0 and y = 1 - x (as shown in Figure 14.23). Finally, we
find that
1 f (x 3 - 2x 2 - x + 2) dx = j[ ∣∙X 4 - ∣* 3 - 2χ 2 + 2x ]o = 24
2
■ PARAMETRICALLY DEFINED SURFACES
In Section 13.4, we showed that if a surface S is defined parametrically by the

vector function
R(M, V) = x(w, v)i + y(u, v)j + z(w, v)k
over a region D in the wv-plane, the surface area of S is given by the integral
f f ∣∣R ι × R ι,∣∣ du dv
D
Then if / is continuous on D , the surface integral of f over D is given by
f f Λ × , y ,z ) d S = ∣∫ ∕( R ) ∣∣R,, × R,,∣∣ du dv
J S J J DJ
EXAM PLE 5 Surface integral for a surface defined parametrically
Evaluate (x + y + z) dS where 5 is the surface defined parametrically by

J S J
R( M, V) = (2» + v)i + (w - 2v)j + (w + 3v)k for 0 ≤ u ≤ 1, 0 ≤ v ≤ 2.
Solution I I (x + y + z) dS = I I ∕(R)∣∣R a × R v∣∣ du dv

sj DJ
Before we go on, we need to find the component parts for the integral on the
right. If R = xi + yj + zk, we see for this problem x = 2u + v, y = u —2v, and
z = u + 3v. Because ∕( x , y, z) = x + y + z,
∕( R ) = f(2ιι + v, u - 2v, u + 3v)
= 2u + v + ιι —2v + u + 3v = 4w + 2v
R (( = 2i + j + k and R v = i - 2j + 3k, so
j
RU × RV = 2 1 1 = (3 + 2)i - ( 6 - l)j + ( -4 - l)k = 5i - 5j - 5k
1 -2 3
Thus, ∣∣RH × R v ∣
∣= √ 5 2 + (-5 ) 2 + ( -5 ) 2 = 5√3
We now substitute these values into the surface integral formula:
■U× R JI du dv
5 D
■2 r l -2
(2 + 2v) dv = 5√ 3[2v + / ] = 5√3(8) = 40√ 3

o 0
PROBLEM SET 1 4 .5
θ Let S be the hemisphere x 2 + y 2 + z 2 = 4, with z ≥ 0, in
Problems 1-6. Evaluate the surface integral. 20. Evaluate j is the portion o f the plane
1. 2.
s s
(x - 2y) dS (5 - 2x) dS
s s
∣ j (x 2 + y 2)z dS ∣ j (x 2 + / ) dS
In Problems 7- ∣ J x y dS.
7. 5:
8. S: z 21. Evaluate J J (x 2 + y 2 + z 2) dS, where S is the portion of
9. S':
10. 5: z xi , / the plane z = x + 1 inside the cylinder x 2 + y 2 = 1.
4 1
11. S: z
12. 5: z
13. S: z
14. 5: z
In Problems 15-18, suppose S is the portion o f the paraboloid

z = x 2 + y 2 fo r which z ≤ 4. Evaluate the given surface inte
gral. F ∙ N dS for the vector fields F and surfaces S
15. (4 - z) dS s
given in Problems 22-27. Assume N is the outer normal.
S 5
22. F = xi + 2χj + zk, and 5 is the triangular region bounded
dS by the intersection o f the plane x + 2y + z = 1 and the
17. + 4zdS
V l + 4z positive coordinate axes.
s s
2 23. F = xi + 2yj —3zk, and S is that part o f the plane
19. + y 2) dS, where S is the surface bound-
1 5 x -1 2 y + 3z = 6 that lies above the unit square
s
ed above by the hemisphere z = V 1 —x 2 —j ,2 , and below
24. F = xi + χj + 2zk, and S is the surface of the cube bound
ed by the planes x = 0, x = 1, y = 0, y = 1, z = 0, and
z = 1.
25. F = xi + yj, and S is the hemisphere z = V 1 —x 2 —y 2.
26. F = 2xi - 3yj, and S is the part of the hemisphere given
by x 2 + y 2 + z 2 = 5, for z ≥ 1.
27. F = x 2i + y 2j + z 2k, and S is the portion of the plane 37. S is the triangular surface with vertices (1,0, 0), (0, 1, 0),
z = y + 1 that lies inside the cylinder x 2 + y 2 = 1. and (0, 0, 1).
θ 38. Let S be the spherical shell centered at the origin with
θ 28. Evaluate J j (3x - y + 2z) dS, where S is the surface radius a, and let C b e the right circular cone whose vertex
is at the origin and whose axis of symmetry coincides with
determined by R(w, v) = ui + wj — vk, 0 ≤ u ≤ 1, the z-axis. (This has the shape of an old-fashioned ice
1 ≤ v ≤ 2. cream cone.) Suppose the vertex angle of the cone is φ,
with 0 ≤ φ < ττ. Determine the mass o f that portion of
29. Evaluate J j (x - y 2 + z) dS, where 5 is the surface the sphere that is enclosed in the intersection of .S’ and C.
Assume p(x, y, z) = 1.
defined by R(M, V) = u 2i + vj + uk, 0 ≤ w ≤ l , 0 ≤ v ≤ 1. 39. Use the formula
30. Evaluate J J (tan 1x + y - z 2) dS, where S is the surface ∣ J (χ 2 + y 2M χ , y, z) d s
defined by R(u, v) = wi + v2j - vk, 0 ≤ u ≤ l , 0 ≤ v ≤ l . for the moment of inertia about the z- axis. Show that the
moment o f inertia of a conical shell about its axis is ∖ma 2 ,
31. Evaluate j J (x 2 + y - z) dS, where S’ is the surface where m is the mass and a is the radius. Assume that
p(x, y, z) = 1.
defined by R(w, v) = ιA — u 2j + vk, 0 ≤ u ≤ 2, 0 ≤ v ≤ 1. 40. Use the formula
∕ z = J ∣ (λ'2 + y 2)p(χ, y, z) d s
In Problems 32-37, fin d the mass o f the homogeneous lamina

that has the shape o f the given surface S. Assume ρ(x, y, z) = 1. for the moment of inertia about the y-axis. Show that the
32. S isthe surface z = 4 - x - 2y, with z ≥ 0, x ≥ 0
moment o f inertia of a spherical shell of uniform density
33. S isthe surface z = 10 - 2x - y, with z ≥ 0,y ≥ 0 about its diameter is jm α 2, where m is the mass and a is
34. S isthe surface z = x 2 + y 2 , with z ≤ 1. the radius. Assume p(x, y , z) = 1.
35. .S' isthe surface z = 1 - x 2 - y 2, with z ≥ 0. 41. T H IN K T A N K P RO BLEM Show that a Mobius strip is not
36. S is the surface x 2 + y 2 + z 2 = 5, with z ≥ 1. orientable.
1 4 .6 STOKES'S THEOREM
IN THIS SECTION Stokes’s theorem, physical interpretation of Stokes’s

theorem
In Section 14.4, we observed that Green’s theorem can be written
J F ∙ √R = J J (curl F ∙ k) dA
A
where A is the plane region bounded by the closed curve C. Stokes’s theorem is
a generalization of this result to surfaces in space and their boundaries.
■ STOKES’S THEOREM
Before stating Stokes’s theorem, we need to explain what is meant by a compatible

orientation. We say that the orientation of a closed path C on the surface S is
compatible with the orientation on S if the positive direction is counterclockwise in
relation to the outward normal of the surface (see Figure 14.24). If you point the
thumb of your right hand in the direction of the outward unit normal, your fingers
will curl in the direction of the compatibly oriented curve C.
Sec. 14.6 Stokes’s Theorem 94 1
Historical Note
Figure 14.24 Compatible orientation: The surface .S' is on the left of someone walking in a
counterclockwise direction around the boundary curve C.
GEORGE STOKES (1819-1903) THEOREM 14.7 Stokes’s theorem

Stokes’s theorem is named for Let S be an oriented surface with unit normal vector N , and assume that 5 is
George Gabriel Stokes, an
bounded by a closed, piecewise-smooth curve C whose orientation is compatible
English mathematical physicist
who publicized the theorem in
with that of S. I f F is a vector field that is continuously differentiable on 5, then
1854. According to Stokes, the F ■√R = ∫ J (curl F ∙ N) dS
result was first obtained by one
of his colleagues, William
Thompson, who is also known Proof: The general proof is beyond the scope of this text. However, a proof
as Lord Kelvin. where 5, is the graph of z = z(x, y) and F, 5, and C are “ well behaved” is given in
Appendix B. ■
Notice that the surface 5, in the statement of Stokes’s theorem is not closed.
EXAM PLE 1 Verifying Stokes’s theorem for a particular example
Let F = —∣ y 2i - 2xyj + yzk, where .S' is that part of the surface of the plane
x + y + z = 1 contained within the triangle C with vertices (1,0, 0), (0, 1, 0),
and (0, 0, 1), traversed counterclockwise as viewed from above. Verify Stokes’s
theorem.
Solution The planar surface S and boundary curve C are shown in Figure 14.25.
We shall show that the line integral Jc F ∙ <7R and the surface integral
J ∫ (curl F ∙ N) dS have the same value.
s
I. Evaluation o f F ∙ dR
Jc
The three edges of the boundary triangle C are expressed as

E): x + y = l,z = 0 Ey. y + z = l ,x = 0 E 3ι x + z = l , y = 0
We traverse C in a counterclockwise direction (see Figure 14.25).
Edge E { : Parametrize with x = 1 - t, y = t, z = 0, for 0 ≤ t ≤ 1, so we have

R(Z) = (1 - Z)i + Zj and dR = - d t ∖ + dtj. Finally, in terms of the parameter Z,
Figure 14.25 The triangle C we have F(Z) = - ∣Z 2i - 2(1 - Z)Zj.
bounds the part of the plane
x + y + z = 1 that lies in the first F dR = £ (∣Z2 - 2Z + 2z2) dt = ∫θ' ⅛ 2 - 2z) dt = (jz2 - Z2)∣J = ⅛
octant.
Edge Ey. Parametrize with x = 0, y = 1 - 5, z = s, for 0 ≤ 5 ≤ 1, so

R(.s) = (1 - s)j + vk and dR = -d s j + dsk. In terms of the parameter s, we
have F(,sj = -∣(1 - s) 2i + ( 1 - s)sk.
(1 - s ) s d s = ( ⅞ - f ) ∣ o = ∣
F ∙ √R =
∣
E1 Jθ
Edge Ey. Parametrize with x = r, y = 0, z = 1 - r, for 0 ≤ r ≤ 1, so

R(r) = ri + (1 - r)k and t/R = dr'∖ - drk. In terms of the parameter r, we have
F(r) = 0.
F ∙ <√R = 0
Combining these results, we find
f F ∙ t Z R = f F ∙ dR + f F ∙ dR + [ F ∙tZ R = ⅛+ ∣ + 0 = ⅜
∣C JE1 JE2 JE.
II. Evaluation o f J J (curl F ■N) dS
i j k
∂ ∂ ∂
curl F = ∂x ∂y ∂z = zi + yk
-3 2 ∏ y∑
-y y -W
The triangular region on the surface of the plane x + y + z = 1 has outward

unit normal
N = ^ ( i + j + k)
Because z = 1 - x - y, we find
curl F ∙ N = (zi + yk) ∙ ⅛ ( i + j + k) = ^ ( z + y)
= ⅛ 1 ~ χ ~ y + y) = ⅛ ( 1 - χ )
z
Figure 14.26 The projected region
in the xy-plane is a triangle Finally,
bounded by y = 1 - x and the
positive coordinate axes. dS = V z 2 + z 2 + 1 dA χ y = V ( - l ) 2 + ( - l ) 2 + 1 dA χy = √3 dA χy
Combining these results, we get (see Figure 14.26)
∫ ∫ (curl F ∙ N) dS = ∫ ∫ ^ ∣( 1 - x)√3 dA χy = ∫θ j (1 - x) dy dx
r' (i - x) 3 ∣ι ι
= (1 - x)[(l - x) - 0] dx = -------2— ∣o = 3
We see that for this example, F ■dR = -∣ = J J (curl F ∙ N) dS',, as claimed
by Stokes’s theorem.
Sec. 14.6 Stokes’s Theorem 94 3
EXAMPLE 2 Using Stokes’s theorem to evaluate a line integral
Evaluate (⅛y2 dx + z dy + x dz), where C is the curve of intersection of the

Jc
plane x + z = 1 and the ellipsoid x 2 + 2y2 + z 2 = 1, oriented clockwise as seen
from the origin. The curve C is shown in Figure 14.27.
Figure 14.27 The surface bounded by the curve of intersection

projects onto a disk in the xy-plane.
Solution Let F = ∣ y,2i + zj + xk. From Stokes’s theorem, we can evaluate the
given integral by computing
I I (curl F ∙ N) dS
We choose .S' to be the part of the plane x + z = 1 bounded by C, along with C.

Because S is a portion of the plane x + z = 1, we have z = 1 - x so that
zχ = -1 , zp = 0. We now find the required parts of the surface integral.
i j k
∂ ∂ ∂
curl F = ∂x ∂y ∂z - i - j -y k
l v2 z X
2X
The unit normal to the plane x + z = 1 is N = ^ ( i + k)> s o that
curl F ∙ N = ξ ⅛ ( - l -T )
We also have
d s = 'Jz x 2 + ⅜ 2 + 1 dAχy = √ ( - l) 2 + (0)2 + 1 dAχy = √2 dAχy
Finally, to describe C we substitute z = 1 - x into the equation for the
ellipsoid:
x 2 + 2y2 + z 2 = 1
x 2 + 2y2 + ( 1 - x)2 = 1
x 2 - x + y2 = 0
W 2+∕ =⅛
Thus, the projection of this curve on the xy-plane is the circle 5 vι, shown in
Figure 14.28. Notice that this circle has the polar equation r = cos θ. Note also
that although C is oriented clockwise as viewed from the origin, the projected
curve in the xy-plane is oriented counterclockwise as viewed from above.
We now calculate the desired line integral using Stokes’s theorem.
(∣y 2 dx + z dy + x dz) = F ∙ dR
Jc Jc
Figure 14.28 This is the projected = J J (curl F ∙ N) dS Stokes’s theorem

disk in the xy-plane; call it S .
= 1 ∫ ⅛ (-1 -T )√ 2 < F = - ∫ ∫ ( l+ y ) ¾
$xy $xy
rττ∕2 rcos θ
= — (1 + r sin θ) r dr dθ Change to polar coordinates.
J -π ∕2 Jθ
rπ∕2 Γ I
= - L
⅜cos2 0 + 4 cos 3 0 sin θj ∖dθ = - J4
J —rr/2 ■■
Component o f F in
the direction o f N
■ PHYSICAL INTERPRETATION OF STOKES’S THEOREM
In Section 14.1, we defined the flu x density as the vector field F = pV, where
V(x, y, z) is the velocity of a fluid with density p(x, y, z). The flux density measures
the volume of fluid crossing surface S per unit of time and is also called the flux of
F across 5. Suppose the surface 5’ is located in the region through which the fluid
flows. If N is the unit normal to the surface S, then F ∙ N is the component of flux
Velocity field in the direction normal to S, as shown in Figure 14.29. Then the mass of the fluid
Figure 14.29 Flux density is mea flowing through 5 in unit time in the direction normal to the surface is given by a
sured by F = pV. surface integral, which is called a flux integral.
Flux Integral The flux integral of a vector field F across a surface 5' is given by
If V is the velocity field of a fluid flow, then curl V measures the tendency of
the fluid to rotate, or swirl. Usually, if the fluid flows across a surface S, the
rotational tendency will vary from point to point on the surface, and the surface
integral ∫ ∫ (curl V ∙ N) dS provides a measure of the cumulative rotational
s
tendency over the entire surface.
Stokes’s theorem tells us that this cumulative measure of rotational tendency
equals the line integral Jc V ∙ dR. To interpret this line integral, recall that it can be
f f (curl V ∙ N) dS written Jc V ∙ T ds in terms of the arc length parameter 5 and the unit tangent T to
s
measures the cumulative rotation the curve. Thus, the line integral sums the tangential component of the velocity
al tendency of a fluid to “swirl” field V around the boundary C , and it is reasonable to interpret fc V ∙ T ds as a
across the surface S. measure o f the circulation o f the fluid flow around C.
Sec. 14.6 Stokes's Theorem 94 5
[ V ∙ <√R= f V ∙ T ⅛
Jc Jc
measures the tendency of a fluid In physics and other applied areas, Stokes’s theorem is often used as a device for
to “circulate” about the curve C. establishing general properties. For instance, we can use it to prove that a vector
field F is conservative if and only if curl F = 0.
THEOREM 14.8 Test for a vector field to be conservative

If F and curl F are continuous in the simply connected region D, then F is
conservative in D if and only if curl F = 0 in D.
Proof: If F is conservative, let /b e a scalar potential function, so that V ∕ = F.
Then curl F = V × F = V× (V∕) = 0. (See Problem 63, Section 14.1.)
Conversely, if curl F = 0, let Cbe the boundary curve of the smooth surface S'.
Then Stokes’s theorem gives
F ∙r ∕R = F ∙ N) dS = 0 dS = 0
'c JS J
so that F is independent of path and F must be conservative.
EXAM PLE 3 Showing that a vector field is conservative
Show that the vector field F = yz∖ + xzj + xyk is conservative in R3.
Solution We check to see that the region is simply connected.
i j k
a a γ = (x - x)i - (y - y)j + (z - z)k = 0
curl F = ∂x ay z
.yz xz xy
It follows from Theorem 14.8 that F is conservative.
Sometimes it is possible to exchange a particularly difficult surface integration

over one surface for a less difficult integration over another surface with the same
boundary curve. Suppose two surfaces S,l and . / are bounded by the same curve C
and they induce the same orientation on C. Stokes’s theorem tells us that
F ∙ N 1) dS λ F ∙ dR — (curl F ∙ N2) dS2
for any function F whose components have continuous partial derivatives on both
S,1 and S 2. It is also assumed that the unit normals N 1 and N7 are compatible with
C.
EXAMPLE 4 Using Stokes’s theorem to evaluate a surface integral
Evaluate J J (curl F ∙ N) dS, where F = xi + y 2j + ze xy k and S' is that part of
the surface z = 1 - x 2 - 2y2 with z ≥ 0.
Solution By setting z = 0 in the equation of the surface, we find that the

boundary curve C for 5, is the ellipse x 2 + 2y2 = 1, and the outward unit
normal on S induces a counterclockwise orientation on C, as shown in Figure
14.30.
Let S,* be the elliptical disk defined by x 1 + 2y2 ≤ 1. We see that 5 and 5*
have the same boundary, the same orientation, and the same normal N 1 - k.
Thus,
[ f (curl F ∙ N) dS = f f (curl F ∙ k) dS*
JSJ ∙
JS J
Figure 14.30 Surface

and we use this second integral to calculate the required integral. We obtain
z = 1 —x 2 —2y 1 with z ≥ 0 j k
curl F = ∂ ∂ = (zxexy )i - (zyexy )j
∂y ∂z
y 2 zexy
and because curl F ∙ k = 0, we conclude
Stokes’s theorem has important application in the theory of electromagnet

ism. We will now use Stokes’s theorem to derive one of Maxwell’s equations,
named after James Clerk Maxwell (see Historical Note, page 904).
EXAMPLE 5 Maxwell’s current density equation
From physics, Ampere’s circuit law states
H dR = I
where H is the magnetic intensity, I is the current crossing any surface 5,

bounded by the closed curve C, and R = xi + yj + zk. Use Ampere’s circuit
law to derive Maxwell’s equation curl H = J, where H is the magnetic
intensity and J is the electric current density.
Solution If J is the current density, the total current flowing across S is given by
the surface integral J f J ∙ N dS. When this fact is used in conjunction with
s
Ampere’s law, we have
[ H ∙ dR = I = ∫ ∫ J ∙ N dS
By Stokes’s theorem, H ∙d R = (curl H ∙ N) dS.

Sec. 14.6 Stokes's Theorem 947
Equating the two surface integrals that equal H ∙ t∕R, we have

Jc
s s
or, equivalently,
∫s∫( j - curl H) ∙ N dS = 0
Because this equation holds for any surface 5 bounded by C, it follows that
J - curl H = 0
J = curl H ∙≡
PROBLEM SET 1 4 .6
0 Verify Stokes’s theorem for the vector functions and surfaces
8. (y dx - 2x dy + z dz), where C is the intersection of the
given in Problems 1-5. Jc
1. F = zi + 2xj + 3yk, over the hemisphere
surface z = x 2 + y 2 and the plane x + y + z = 1 traversed
x 2 + y 2 + z 2 = 9 for z ≥ 0
counterclockwise when viewed from the origin
2. F = (y + z)i + xj + (z — x)k, over the triangular region
bounded by the plane x + 2y + z = 3 and the coordinate
axes
3. F = (x + 2z)i + (y - x)j + (z - y)k, over the triangular
region bounded by the plane x + 2y + z = 3 and the
coordinate axes
4. F = 2xyi + z 2k, over the paraboloid y = x 2 + z 2 , with
y ≤ 4
5. F = 2yi - 6zj + 3xk, over the paraboloid z = 4 —x 2 - y 2
and above the xy-plane, with the outward normal
Use Stokes’s theorem to evaluate the line integrals given in 9. [2xy2∑dx + 2x 2yz dy + (x 2y 2 - 2z) dz], where C is the
Problems 6-13. Jc
curve given by x = cos t, y = sin t, z = sin t, 0 ≤ t ≤ 2π,
6. (x 3y 2 dx + dy + z 2 dz), where C is the circle traversed in the direction of increasing t
Jc
x 2 + y 2 = 1 in the plane z = 1 traversed counterclockwise
when viewed from the origin.
7. (z dx + xdy + y dz), where C is the triangle with

Jc
10. {y dx + z dy + y dz), where C is the intersection of the
vertices (3, 0, 0), (0, 0, 2), and (0, 6, 0), traversed in the Jc
given order sphere x 2 ÷ y 2 + z 2 = 4 and the plane x + y + z = 0,
traversed counterclockwise when viewed from above
11. (y dx + z dy + x dz), where C is the intersection of the

Jc
plane x + y = 2 and the surface x 2 + y 2 + z 2 = 2(x + y),
traversed clockwise as viewed from above
12. [(z + cosx) dx + (x + y 2) dy + (y + e z) dz] where C is

Jc
the intersection of the sphere x 2 + y 2 + z 2 = 4 and the z = sin θ, 0 < θ ≤ 2τr, oriented in the direction o f in
cone z = V x 2 + y 2, traversed counterclockwise as viewed creasing θ. Use Stokes’s theorem to evaluate
from above ∣ J (curl F ∙ N) dS
j s j
13. (3j> dx + 2z dy —5x dz), where C is the positively
Jc
θ 25. Let .S’ be the ellipsoid + yy + z 2 = 1, and let F be a
oriented boundary of the hemisphere z = V 1 - x 2 —y 2
θ In Problems 14-19, use Stokes’s theorem to evaluate vector field whose component functions have continuous
∫ ∫ (curl F ∙ N) d S for the vector fields and surfaces. partial derivatives on S’. Use Stokes’s theorem to show
s that . .
14. F = xi + y 2j + xyzk and .S' is the part of the paraboloid (curl F ∙ N) dS = 0
z = 4 - x 1 —y 2 with z ≥ 0. Use the upward unit normal.
15. F = xyi — zj and 5 is the surface o f the cube 0 ≤ x ≤ 1, Does it matter that S is an ellipsoid? State and prove a
0 ≤ y ≤ l , 0 ≤ z ≤ 1 except for the face where z = 0. more general result based on what you have discovered in
Use the outward unit normal. the first part o f this problem.
16. F = yi + zj + xk, and S is the part of the plane x + y + z = 1 26. Faraday’s law of electromagnetism says that if E is the
that lies in the first quadrant. Use the outward normal. electric intensity vector in a system, then
17. F = xyi + x 2j + z 2k, and C is the intersection of the
paraboloid z = x 2 + y 2 and the plane z = y. jfc E ∙ d R = -∂φ
-
18. F = xzi + j ,2j + x 2 k, and C is the intersection of the plane
x + y + z = 3 and the surface around any closed curve C where t is time and φ is the
total magnetic flux directed outward through any surface
S bounded by C . Given that
19. F = 4yi + zj + 2 jk, and C is the intersection o f the φ = J ∣ B ■N dS
sphere x 2 + y 2 + z 2 = 4 with the plane z = 0.
In Problems 20-22, use Stokes’s theorem to evaluate the line where B is the magnetic flux density, show that
integral 1 17 3B
f [(1 + y)z dx + (1 + z)x dy + (1 + x)y dz] curl E = ——
ot
Jc
for the given choice for the path C. Hinf. It can be shown that
20. C is the elliptic path x = 2 cos θ, y = sin θ, z = 1 for
0 ≤ θ ≤ 2τr.
21. C is the triangle with vertices (1,0 , 0), (0, 1, 0), (0, 0. 1).
22. C is any closed path in the plane 2x - 3y + z = 1.
27. According to Ampere’s law H ■√R = I, where H is the
23. Let F = y 2 i + xyj + xzk, and suppose 5 is the hemisphere
x 2 + y 2 + z 2 = 1 with z ≥ 0. Use Stokes’s theorem to magnetic intensity and I is the total current flowing across
express a given surface S’ (in the direction of the outward normal).
J J (curl F ∙ N) dS Given that curl B = k J , where J is the current density and
k is a physical constant, show that
as a line integral, and then evaluate the surface integral by B ∙ √R = k l

evaluating this line integral.
24. Let F = zi + xj + yk, and suppose S is a smooth surface for any closed curve C . B is the magnetic flux density (see
in IR3 whose boundary is given by x = 2 cos θ, y = 3 sin θ, Problem 26).
14. 7 DIVERGENCE TH EO REM
IN TH IS SECTIO N Divergence theorem, applications of the divergence

theorem
In this section, we develop the divergence theorem, one of the great theorems of
elementary calculus. We shall also see that the divergence theorem can be used
to derive theoretical results and to study the properties of fluid dynamics. In
particular, in Example 5 we derive the heat equation, which is important in
physics and engineering.
Sec. 14.7 Divergence Theorem 94 9
Historical Note ■ DIVERGENCE THEOREM
We used Green’s theorem to show that F ∙ N ds = div F dA, where D is a
simple connected domain in the plane with the closed boundary curve C. The
divergence theorem (also known as Gauss’s theorem) is a generalization of this
form of Green’s theorem that relates a closed surface integral to a volume integral.
THEOREM 14.9 The divergence theorem

Let D be a region in space bounded by a smooth, orientable closed surface 5,. If F is
a continuous vector field whose components have continuous partial derivatives
in D, then
∫ J F ∙ N dS = ∫ ∫ ∫ div F dV
KARL GAUSS (1777-1855) s D
Karl Gauss is considered to be where N is an outward unit normal to the surface 5'.
one of the three greatest mathe
maticians of all time, along Proof: The general proof of this theorem is beyond the scope of this course,
with Archimedes and Newton. but an important special case is proved in Appendix B. ■
Gauss graduated from college
at the age of 15 and proved EXAMPLE 1 Verifying the divergence theorem for a particular example
what was to become the funda
mental theorem of algebra for Let F = 2xi - 3vj + 5zk, and let >S' be the hemisphere z = V9 - x 2 - y 2
his doctoral thesis at the age of together with the disk x 2 + y 2 ≤ 9 in the xy-plane. Verify the divergence
22. He published only a small theorem.
portion of the ideas that
seemed to storm his mind, be Solution The solid is shown in Figure 14.31. We shall show that the surface
cause he believed that each integral and the triple integral have the same value.
published result had to be com
plete, concise, polished, and
convincing. His motto was
I. Evaluation o f JJ
F ∙ N dS
“Few, but ripe.” The diver
gence theorem of this section
was proved by Gauss prior to
its publication by the Russian
mathematician Mikhail
Vassilievich Ostrogradsky
(1801-1862), so it is sometimes
known as Gauss’s theorem.
Ostrogradsky was also a first- Figure 14.31 The surface of the hemisphere
rate mathematician but was de
nied a university diploma for
his atheism, so he left Russia S consists of two parts: 5 1, the disk on the bottom of the hemisphere, and S 2,
for Paris in 1822. According to the hemisphere. We will consider these separately because
Carl B. Boyer in A History of
Mathematics, Gauss is some
times described as the last
mathematician to know every
thing in his subject. Such a gen
eralization is bound to be inex Consider S γ∙. Note that the disk x λ + y 1 ≤ 9 on the bottom of the hemi
act, but it does emphasize the sphere has outward unit normal N = —k. (Because k points into the solid, - k
breadth of interest Gauss dis is the outward normal.) We have
played.
∫ ∫ F ∙ N dS χ = ∫ ∫ (2xi - 3yj + 5zk) ■(-k ) √5 = ( - 5z) dS χ = 0
1
Because z = 0 on Sj
Consider Sy. Regard the hemisphere as the level surface

S 2(x, y, z) = 9, where 5,2(x, y, z) = x 2 + y 2 + z 2 and z ≥ 0.
The outward unit normal (for z ≥ 0) is given by
N = ---- —
l∣v⅞∣∣
2xi + 2pj + 2zk D A ,ι, , V df ∙ j. d d ■a. aΛ
∏r∕ = —
Remember that ι + — jj + — k.
V 4 x 2 + 4y2 + 4z2 J ∂X ∂y ∂z
xi + pj + zk
V x 2 + y 2 + Z2
xi + j'j + zk
Because x 2 + y 2 + z 2 = 9 on S 2
√9
= ⅜(*> + yj + -k)
Because F = 2xi - 3pj + 5zk, we find
F ∙ N = ⅜(2x2 - 3y2 + 5z2)

Finally, to find dS 2 we note that z 2 = x 2(9 - x 2 - y 2)^1 and
Z2 _ y2(g _ χ 2 _ s θ t ]l a t
d S 2 = √zv+ z" + 1 d A xy
= ΛI χ2 I y2 11 dA
∖ 9 -x 2 ~y 2 9 - x2 - y 2 x*
= y∣ 9 - x 2 -y 2 dλχy
= M ∖ =≡t dA
V9 - x 2 - y- xy
It follows that
∫ ∫ F ■N ⅛ = ∫ ∫ ∣(2x≈ - 3 / + 5z≈) _ 3 , _ <Uv
v 9 χ y l
s2 ¾
J j^ [2x 2 - 3y 2 + 5(9 - x 2 - y 2)] [9 - x 2 - y 2]^ π2 dA χγ
¾
_ ∫ ∫ (45 - 3x 2 - 8p2)(9 - x 2 - y 2) - ιy2 dA χy
s 2
f 2πf 3 45 - 3(r cos θ)2 - 8(r sin θ)2 j jn See Figure 14.32.
I I Γ7~, Cl θ
Jo Jo V 9 - r2
[135 - 54 cos2 0 - 144 sin 2 0] dθ
= 72ττ
We can now state
f f F ■N dS = f f F ∙ N d S i + f f F ∙ N dS 2 = 0 + 72ττ = 72ττ
js j j Slj J SJ
Sec. 14.7 Divergence Theorem 951
II. Evaluation div F d V

D
div F = ^-(2x) + ∕- ( - 3 y ), + ⅛ Vz ) = 2 - 3 + 5 = 4
∂ x x ' ∂y ' ∂Z '
Therefore, f f f div F d V = f J f 4 dV, but f f f d V is just the volume of the
hemisphere z = V 9 - x 2 - y 2 . A sphere of radius r = 3 has volume
V = ∣ττr3 = jτr(3)3 = 36ττ, so a hemisphere of radius 3 has volume
Figure 14.32 The hemisphere ⅛[5π(3)3] = I8τr.
with boundary z 2 = 9 - x 2 - y 2
projects onto disk x 2 + y 2 ≤ 9, ∣ ∣ ∣ div E d V = ∣∣∣ 4 √ K = 4 K = 4 (l8 τ r ) = 72π
or, in polar form, r ≤ 3
D D
and we have f J f div F d V = f f F ∙ N dS, as required.
EXAM PLE 2 Evaluating a surface integral using the divergence theorem
Evaluate F ■N dS, where

SJJ
F = x 2i + xyj + x 3y3k and 5 is the sur

face of the tetrahedron bounded by the
plane x + y + z = I and the coordinate
planes, with outward unit normal N.
Solution We will use the divergence theo
rem. Let D be the solid bounded by A and
note that
d iv F = ⅛(∙χ 2 ) + ∂y(x y ^ + = 2x + x + ° = 3x
We choose to describe the solid tetrahedron as the set of all (x, y, z) such that
0 ≤ z ≤ I - x - y whenever 0 ≤ y ≤ I - x for all 0 ≤ x ≤ I. The projec
tion of S' on the xy-plane is shown in the margin. Finally, by applying the
divergence theorem, we find that
l- x r l-x -y
F ∙ N dS =
Ilf
D
∏ ∣ Jo
3x dz dy dx
Projected region in the xy-plane f1 ∖~x f Γ -I 1 - x

∣ I 3x(l - x - y) dy dx = 3 I ∣^x(l - x)y - ∣x y 2J 0 dx
= 3 f x( 1 - x) 2 - ∣x( 1 - x) 2 dx = A
Jo °
The divergence theorem applies only to closed surfaces. However, if we wish

to evaluate ∫ ∫ F ∙ N dS where S l is not closed, we may be able to find a closed
surface S that is the union of S 1 and some other surface S,2. Then, if the hypotheses
of the divergence theorem are satisfied by F and S', we have
F ∙ N dS + F ∙ N dS =
III D
div F d V
where D is the solid region bounded by S. Thus, if we can compute f f f div F d V

and ∫ f F ∙ N dS, we can compute ∫ ∫ F ∙ N dS by the equation

*ι
I I I
div F d V - J J F ∙ N dS
D sι
This equation can also be used as a device for trading the evaluation of a difficult
surface integral for that of an easier integral. Here is an example of this procedure.
EXAM PLE 3 Evaluating a surface integral over an open surface
Evaluate J j F ∙ N dS, where F = xyi - z 2k and S' is the surface of the upper
five faces of the unit cube 0 ≤ x ≤ 1, 0 ≤ y ≤ 1, 0 < z ≤ l ,a s shown in

Figure 14.33.
Figure 14.33 Evaluating a

surface integral
Solution Note that the surface 5, is not closed, but we can close it by adding the
missing face S m , thus forming a closed surface 5* that satisfies the conditions
of the divergence theorem. The strategy is to evaluate the surface integral of S'*
and then subtract the surface integral of the added face S m .
F ∙ N dS = f [ f div F d V = [ [ [ (y - 2z) dx dy dz
JJ J Jo Jo Jo
cube
(y - 2z) dy dz = (I - 2z) dz = -⅛
Io Jo Jo v z
Also, because the unit normal to the added face S m is N = - k and z = 0 on

this face, it follows that
∫ ∫ F ∙ N dS = J ∫ (xyi - z 2 k) ∙ ( - k ) dS = ∫ ∫ z 1 dS = 0
5w sm stn
Therefore,
F ■N dS = - ∣ - 0 = - ∣
■ APPLICATIONS OF THE DIVERGENCE THEOREM
Like Stokes’s theorem, the divergence theorem is often used for theoretical
purposes, especially as a tool for deriving general properties in mathematical
physics. The key to some of these applications is the fact that because F(x, y, z) is
Sec. 14.7 Divergence Theorem 95 3
the rate of flow per unit area, the surface integral f J^F ∙ N dS represents the net
s
rate of outward flux per unit volume. This is the reason for the name divergence,
because
F ∙ N dS =
D ,
Flux integral Flow out of small cubes
The integral on the left is a flux integral and thus determines the total fluid flow
across the surface S' per unit of time. On the other hand, the integral on the right
measures the same fluid flow by calculating the flow out of small cubes.
Recall from Section 14.1 that a point is called a source if div F > 0, a sink if
div F < 0, and incompressible if div F = 0. The next example derives an
important property of fluid dynamics.
EXAM PLE 4 Continuity equation of fluid dynamics
Suppose a fluid with density p(x, y , z, t) flows in some region of space with
velocity F(x, y, z, t) at the point (x, y, z) and time t. Assuming there are no
sources or sinks, show that
div pF = - ∣
Solution Let D be an arbitrary solid region of space, with closed boundary

surface S. In physics, it is shown that the amount of fluid flowing out of D
across S in unit time is f f pF ■N dS. For example, suppose p is measured in
s
pounds per cubic foot so that is measured in pounds per cubic foot per
second and d V is measured in pounds per second. Thus, - ∫ f pF ∙ N dS is
the amount of outflow and is equal to the amount of inflow, f f f dV,

D DF
because we are assuming there are no sinks or sources. Setting these two
quantities equal, we have
OUTFLOW INFLOW
- f f f [ p y
s D
By the divergence theorem, we have J'J'p F ∙N √ > S ' = J' J J div p F d V so that
$ D
div pF d V + ∫ ∫ ∫ ^ d V = 0
∕∫ ∫ D
∫ ∫ ∫ [ d iv F ∣]
p + dV = 0
D
This equation must hold for any region D , no matter how small, which means
that the integrand of the integral must be 0. That is,
DIV PF =- I —
In physics it is known that the total heat contained in a body with uniform
density p and specific heat σ is f ∫ ∫ σρT d V where T is the temperature. Thus, the
D
amount of heat leaving D per unit of time is also given by
- ⅛ [ ∫ ∫ R ⅜ ∫ ∫ ∫ - σ p σt^ d V
D D
We will use this result in the following example, which is one of the most
important results in mathematics and physics.
EXAMPLE 5 Derivation of the heat equation
Let T(x, y, z, t) be the temperature at each point (x, y, z) in a solid body D at

time t. Given that the velocity of heat flow in the body is F = -k N T for a
positive constant k (called the thermal conductivity) show that
∂T _ k π 2rr
∂t σp
where σ is the specific heat of the body and p is its density.
Solution Let -S’ be the closed surface that bounds D. Because F is the velocity of
heat flow, the amount of heat leaving D per unit time is ∫ ∫ F ∙ N dS, and the
divergence theorem applies: 's
∫ ∫ ∫ div(-kVT) dV= ∫ ∫ ∫ (-fc V ∙ VT) dV
D D
- k V 2T d V
Because this is the amount of heat leaving D per unit of time, we know this is
equal to the heat integral from physics derived just before this example. Thus,
kV 2 T d V = ∣∫ ∫ - σ p ^ d V
∫∫∫
D D
This equation holds not only for the body as a whole, but for every part of the
body, no matter how small. Thus, we can shrink the body to a single point, and
it then follows that the integrands are equal; that is,
-kV 2 T = —σp⅝y
∂t
∂∑ =
∂t σp
For the concluding example, recall that the normal derivative ∂g∕∂n of a scalar
function g defined on the closed surface A is the directional derivative o f/in the
direction of the outward unit normal N to 5; that is,
⅛ = Vg ■N
∂n 6
We will use this equation in the following example, which is a generalization of a

property we first obtained for R2 (see Section 14.4).
Sec. 14.7 Divergence Theorem 955
EXAMPLE 6 Derivation of Green’s first identity

Show that if f and g are scalar functions such that F = ∕V g is continuously
differentiable in the solid domain D bounded by the closed surface S, then
∫ ∫ ∫ [∕V ⅛ + V∕∙Vg] dV= I f f ^ d s
D S
This is called Green’s first identity.
Solution We shall apply the divergence theorem to the vector field F (note that F
is continuously differentiable).
div F = div(∕Vg)
—V ∙ ( f N g ) Property o f divergence (Problem 62, Section 14.1)
=
⅛ + r a ⅛1 + Γ⅛f⅛r , f a ⅛l +
Γ⅛f⅛g , f d ^s
Laχ ∂x J 3JC2J Lay ay 7
ay2J Laz az ^r 7 az2.
=
ΓM ⅜γ + M ⅛ + ⅛L'⅛≥∣ + A d2S ∣ ∂2g l
a⅛-^l
7
Lax ax ay ay az azj Lax2 ay2 az2J
= (V∕) ∙ (vg ) + ∕v⅛
This calculation gives us the first step in the following derivation.
Iff
D
lf ^ 2g + V ∕∙ V g ]d v = f ∫ ∫ div F dV
Divergence theorem
Because F = f S g
∂g
Because Sg ∙ N = — by
definition
PROBLEM SET 1 4 .7
θ Verify the divergence theorem for the vector function F and 4. F = 3xi + 5yj + 6zk; D is the tetrahedron bounded by the
solid D given in Problems 1-4 . Assume N is the unit normal coordinate planes and the plane 2x + y + z = 4.
pointing away from the origin.
1. F = xzi + y 2j + 2zk; Use the divergence theorem in Problems 5-1 9 to evaluate the
D is the ball x 2 + y 2 + z 2 ≤ 4. surface integral ∫ ∫ F ∙ N d S for the given choice o fF and the
s
2. F = xi —2yj; D is the interior o f the paraboloid boundary surface S . For each closed surface, assume N is the
z = x 2 + y 2 , 0 ≤ z < 9. outward unit normal.
3. F = 2y2j; D is the part of the surface o f the plane 5. F = xi + yj + zk; 5 is the surface of the cube 0 ≤ x ≤ 1,
x + 4y + z = 8 that lies in the first octant. 0 ≤ y ≤ l , 0 ≤ z ≤ 1.
6. F = xyzj; S is the surface of the cylinder x 2 + y 2 = 9, for 22. The moment of inertia about the z-axis of a solid T of
0 ≤ z ≤ 5. Note .S’ is the closed can with its top in the constant density ρ = a is proportional to the volume
plane z = 5 and its bottom in the xy-plane. integral
xz 2 , ∫ ∫ ∫ ( χ ≈ ÷ Λ ^
7. F = (cos yz)i + e j + 3z k; 5 is the surface of the hemi
2 2
sphere z = V4 —x —y together with the disk
x 2 + y 2 ≤ 4 in the xy-plane.
8. F = curl[eλ'zi - 4j + (sinxjz)k]'. 5 is the surface of the Express this integral as a surface integral over the surface
ellipsoid 2x2 + 3.y2 + 7z2 = 1. S that bounds T.
9. F = (x2 + y 2 —z 2)i + x⅛j + 3zk; .S' is the surface of the θ 23. Let u be a scalar function with continuous second partial
five faces of the unit cube 0 ≤ x ≤ 1, 0 ≤ y ≤ 1, derivatives in a region containing the solid region T, with
0 < z ≤ 1. closed boundary surface .S’.
10. ,
F = 2ι i —zj + 3xk; .S’ is the five faces of the unit cube
a. Show that =
0 ≤ x ≤ 1, 0 ≤ y ≤ l , 0 < z ≤ 1. j s j r J
11. F = xi + yj + zk; .S’ is the surface of the paraboloid
z = x 2 + y 2 for 0 ≤ z ≤ 9. b. Let u = x + y + z and v = ⅛(x2 + y 2 + z 2). Use part a
12. F = cιιrl(pi + xj - zk); .S’is the surface of the hemisphere to evaluate
z = V4 —x —y .2 2 I J (wVv) ■N dS
13. F = x 2i + y 2j + z 2k; 5 is the surface of the sphere
x 2 + y 2 + z 2 = 4.
14. F = xyzi + xyzj + xyzk; S is the surface of the box where S is the boundary of the cube 0 ≤ x ≤ 1,
0 ≤ x ≤ 1,0 ≤ y ≤ 2, 0 ≤ z ≤ 3 . 0 ≤ y ≤ l , 0 ≤ z ≤ 1.
15. F = xi + yj + (z2 - l)k; 5 is the surface of a solid bound 24. Let f and g be scalar functions such that F = ∕V g is
ed by the cylinder x 2 + y 2 = 4 and the planes z = 0 and continuously differentiable in the region D, which is
z = 1. bounded by the closed surface 5,. Prove Green, s second
16. F = (x5 + 10xy2z 2)i + (y 5 + 10yx2z 2)j + (z5+ 10zx2y 2)kj identity.
f f f ( f ^ 2g - ^ 2∕ ) d v
.S’ is the closed hemispherical surface z = V l - x 2 - y 2
together with the disk x 2 + y2 ≤ 1 in the xy-plane. D
17. F = xy 2i + yz 2j + 2x 2y ⅛ S is the surface bounded by the
cylinder y 2 + z 2 = 1, the xy-plane, and the planes x = 0 - f [(⅛ - , ‰
J J χ ∂n °i ∂n)
and x = 1.
18. 2 2
F = xy 2i + yz ] + x zk; S is the surface (in spherical
coordinates) with top p = 2, 0 ≤ φ ≤ J ’ 0 ≤ θ ≤ 2ττ, 25. Show that if g is harmonic in the region D, then
and bottom 0 ≤ p ≤ 2, φ =0 ≤ θ ≤ 2π.
19. F = x 3i + y 3j + 3α2zk (constant a > 0); S is the surface
bounded by the cylinder x 2 + y 2 = a2 and the planes
z = 0 and z = 1.
where the closed surface 5 is the boundary of D.
θ 20. Suppose that S is a closed surface that encloses a solid
region D. 26. Show that J J F ∙ N dS = 0 if S is a closed surface and
a. Show that the volume of D is given by
F = curl U throughout the interior of 5 for some vector
K(D) (xi + vj + zk) ∙ N dS field U. A vector field U with this property is said to be a
vector potential for F.
27. In our derivation of the heat equation in this section, we
where N is an outward unit normal to .S’. assumed that the coefficient of thermal conductivity k is
b. Use the formula in part a to find the volume of the constant (no sinks or sources). If k = k(x, y, z) is a
hemisphere z = ∖, R 2 —x 2 —y 2. variable, show that the heat equation becomes
21. Use the divergence theorem to evaluate
∕<V2 T + V⅛ ∙ VT = σp
∣
∣
R∣∣R ∙ N dS 28. An electric charge q located at the origin produces the
electric field
where R = xi + rj + zk and S is the sphere gR
x 2 + y 2 + z 2 = a2, with constant a > 0. 4ττe∣∣R∣∣3
where R = xi + ι j + zk and ε is a physical constant,

called the electric permittivity.
a. Show that V
J J E ∙ N dS = 0 30. Suppose that
curl E = —⅛77 and curl H = σE + ε

∂t at
if the closed surface S’ does not enclose the origin. This
where E is electric intensity; B is magnetic flux density; H
is Gauss’s law.
is magnetic intensity; and σ, ε, and μ are positive
b. Show that constants.
∣J E ∙ N dS = ∣ a. Show that curl(curl E) = -^ ( c u r l B).
b. Given that B = μ H , first show that

in the case where the closed surface S encloses the origin.
Note that the divergence theorem does not apply directly curl(curl E) = - μ (curl H)
to this case.
and then derive the formula
29. Gauss’s law can be expressed as
V(div E) - (V ∙ V)E = - μ ∣ σE + ε j ;
J J D ∙ N dS = q
c. If Q = 0, show that div E = 0, and then derive
Maxwell’s equation for the electric intensity E:
where D = εE is the electric flux density, with electric
intensity E, permittivity ε, and q a constant. Show that ∂E , δ2E
( V ∙V ) E = μ σ ~ + μ ε ^ r
div D = Q , where Q is the charge density; that is,
CHAPTER 14 REVIEW
IN THIS REVIEW Chapter checklist, proficiency examination,

supplementary problems
The checklist can be used to summarize the theory techniques and notation
used in this chapter. Problems 1-22 of the proficiency examination test your
knowledge of the concepts of this chapter, and Problems 23-30 test your
understanding of those concepts by asking you to work sample practice
problems from the chapter. The supplementary problems present questions
relating to the material of this chapter in a wide variety of settings.
■ CHAPTER CHECKLIST
In the following statements, assume that all required conditions are satisfied.
Main results are collected here (without hypotheses) so that you can compare and
contrast various results and conclusions.
Scalar function: f = f( x , y, z)
Vector field: F(x, y, z) = u(x, y, z)i + v(x, y, z)j + w(x, y, z)k
Notation: a. a . a,
Del operator: V = ⅛ι + ⅛ j + ⅛ k
Gradient: V ∕= g i + < j + ∣ k
Laplacian: W f= g + + f χ χ + f yy + f zz
Af
Normal derivative: y^~ = Vj ∕'∙ N
∂n
Derivatives of a vector field:

div F = ∂x + ∂y + ∂z This is a scalar derivative,
= V∙ F
If P is a point (x0 , y 0 , z 0), then P is a
source if div F > 0
sink if div F < 0
F is incompressible if div F = 0.
i j k
curl F a a a This is a vector derivative.
∂x ∂y ∂z
u v w
= V× F
Conservative vector fields:
a. F is conservative if it is the gradient of some scalar function f called the
scalar potential o f F; that is, F = V ∕
b. F is conservative if and only
j
if =⅛ = and =
∂y ∂x ∂z ∂x ∂z ∂y
c. Equivalent conditions:
(i) F ∙ JR is independent of path.

Jc
(ii) F is conservative.
(iii) F ∙ d~R = 0 for every closed path C.
Jc
d. Conservative vector field theorem: F is conservative in D if and only if
curl F = 0.
EVALUATION OF LINE INTEGRALS: f F ∙ JR
Jc
Step 1. Check to see if F is conservative; if it is, then
∣F ∙ JR = 0 if C is closed
Jc
C F ∙ JR = f(Q) - f(P) Initial point P, terminal Q
Jc
Step 2. If F is not conservative and Cis a closed curve bounding the surface 5,, use
Stokes’s theorem (or Green’s theorem in R2) to equate the given integral to a
surface integral, namely,
∫ c F ∙ JR = ∫ ∫ (curl F ■N) dS
Step 3. If F is not conservative and C is an open curve, try to add an arc C1 so that
the curve formed by C and C 1 is closed. Try to choose C1 so that £ F ■JR is
relatively easy to evaluate.
ζ F ∙ JR = j^J (curl F ∙ N ) J 5 - [ F JR
Jc,
Step 4. As a last resort, parametrize F and R.

Let R(∕) = x(t)i + y(∕)j + z(t)k for a ≤ t ≤ b.
cb JY
' f(x , y, z) dx = j a f[χ(Q, y(t), z(t)] dt
Jc
f f(x , y, z) ds = f f[x(t), y(t), c(t)]V[x, (t)]2 + [y'(t)]2 + 1 dt

Jc Ja
EVALUATION OF FLUX INTEGRALS: ∫ ∫ F ∙ N dS
Step 1. If S is a closed surface bounding the solid region D, use the divergence
theorem to write the surface integral as a triple integral.
div F dV
Step 2. If the surface S is open, try to find a supplementary surface 5,1 such that
the surface formed by £ and <S,1 is closed. Try to choose 5 1 so that J f F ∙ N dS
is easy to evaluate. Then ,5i
∫ ∫ F ∙ N dS = ∫ ∫ ∫ div F dV - ∫ ∫ F ∙ N dS {
s D s∖
Step 3. If neither steps 1 or 2 applies, parametrize F, N, and dS. This may be quite
difficult.
In the special case where z = f(x,y), we have
∣f g(x, y ,z ) d S = ∫ ∫ g[x, y ,f(x , y)]∖∣
[fx (x, v)]2 + [fy (x, τ)] 2 + 1 dAχy
S R
■ APPLICATIONS
FLUID MECHANICS
Let V be the velocity of an incompressible fluid:
div V measures the rate of particle flow per unit volume at a point.
curl V measures the tendency of a fluid to swirl.
Flux integral: V ∙ N dS measures the rate of flow across the surface 5 in

unit time.
Circulation: V ■T ds measures the tendency of a fluid to move around C.
Stokes’s theorem says that the cumulative tendency of a fluid to swirl across a
surface S is equal to the circulation of a fluid around the boundary curve C.
The divergence theorem says that the total flow across a surface S’per unit of time
is equal to the flow into (or out of) small cubes. If there are no sources or sinks, it
follows that
div pV = — (continuity equation)
where p (x, y, z, t) is the density of the fluid.

TABLE 14.1 Comparison of important integral theorems

Riemann Integral (Section 4.2) Line Integral (Section 14.2)
fi x , y, z) dx
Subdivisions on a curve C in space

Fundamental Theorem o f Calculus (Section 4.3) Fundamental Theorem for Line Integrals (Section 14.3)
J^b
fix ) dx = F(b) - F(a) F ∙ √R = flQ ) - flP )
a Jc
F is an antiderivative off. / i s a scalar potential of F where F is conservative; that
is, V ∕ = F.
Green’s Theorem (Section 14.4); applies to R2 (two dimensions): F(x, y) = Mix, y)i + Nix, y)j
F ∙ dR = IM dx + N dA
Jc
D
and
div F dA
D D
Double integral (Section 13.1) Surface integral (Section 14.5)
dA dS = .2 + gy 2 + 1 dA
R S κ,
where z = f ( x l y).
S∙.z≈ glχ, y)
lχi , yi , z,∙)
Partition of R into mn cells in the xy-plane
Partition of the surface

S’ into n subregions
Stokes’s Theorem
(Section 14.6; Green’s theorem in three dimensions)
F ∙ dR = dS
'c
s
Divergence Theorem (Section 14.7; also known as

Gauss's theorem-, Green’s theorem in three dimensions)
Π ' ∙ -
Γ N K
ELECTROMAGNETISM
Let E be the electric intensity field and H be the magnetic intensity field:
div E = where Q is the electric charge density and ε is the permittivity.
curl E = where μ is the permeability and 1 the time.
curl H = J where J is the electric current density.
Maxwell’s equation for electric intensity:
(V ∙ V)E = μσ + με ηryy where σ is the conductivity.
Ampere’s circuit law: H ∙ dR = Z, where I is the current crossing any surface 5

Jc
bounded by the closed curve C.
THERMODYNAMICS
The heat equation is: ^y = ^-V 2 7,, where T(x, y, z, t) is the temperature at
(x, y, z) at time t. The constant k is the thermal conduc
tivity, σ is the specific heat of the body, and p is its
density.
WORK: W = [ F ■dR
Jc
MASS: m = J j ρ(x, y, z) dS
Concept Problems 12. What is the scalar potential of a conservative vector

1. What is a vector field? field?
2. What is the divergence of a vector field? 13. What is a Jordan curve?
3. What is the curl of a vector field? 14. State Green’s theorem.
4. What is the del operator? 15. How can you use Green’s theorem to find area as a line
5. What is Laplace’s equation? integral?
6. What is the difference between the Riemann integral 16. What is a normal derivative?
and a line integral? Discuss. 17. Define a surface integral.
7. What is the formula for a line integral of a vector field? 18. What is the formula for a surface integral of a surface
8. How do we find work as a line integral? defined parametrically?
9. What is the formula for a line integral in terms of arc 19. What is a flux integral?
length parameter? 20. State Stokes’s theorem.
10. State the fundamental theorem on line integrals. 21. State the conservative vector field theorem.
11. Define a conservative vector field. 22. State the divergence theorem.
Practice Problems
23. Show that yzi + xz] + xyk is conservative and find a
scalar potential function.
24. Compute div F and curl F for F = x 2yi - e>zj + ∣x k .
25. Use Green’s theorem to evaluate the line integral
f F ∙ √R, where F = (2x + y)i + 3y2j and C is the
Jc
boundary of the triangle with vertices ( -1 , 2), (0, 0),
(1, 2), traversed in that order.
26. Use Stokes’s theorem to evaluate the line integral
I F ∙ √R, where F = 2yi + zj + yk and C is the
Jc Figure 14.34 Problem 28
intersection of the plane z = x + 2 and the sphere
x 2 + y 2 + z 2 = 4z, traversed counterclockwise as viewed 29. An object with mass m travels counterclockwise (as
from above. viewed from above) in the circular orbit x 2 + y 2 = 9,
27. Use the divergence theorem to evaluate the surface z = 2, with angular speed ω. The mass is subject to a
integral J J F ∙ N dS, where F = x 2i + (y + z)j - 2zk, centrifugal force F = mω2R , where R = xi + yj + zk.
Show that F is conservative, and find a scalar potential
function for F.
and .S is the surface of the unit cube 0 ≤ x ≤ 1,
0 ≤ y ≤ l , 0 ≤ z ≤ 1. 30. Find the work done by F from Problem 29 as the object
makes a half-orbit from (3, 0, 2) to ( -3 , 0, 2).
' x dx + y dy , _ . , , ,
J
— j------j - y ’ where C is the path shown in
c (x + y )
Figure 14.34.
In Problems 1-6, determine whether the given vector field is 15. F = -(x i + yj + zk), where r —V x 2 + y 2 + z 2 , r Ψ 0
conservative, and i f it is, fin d a scalar potential function.
1. F = 2i - 3j 2. F = xy - 2 i + x - 2 yj 16. F = (xy sin z)i + (x 2cos yz)j + (z sin xy)k
3. F = y - 3 i 4
+ (-3 xy ^ + cosy)j 4. F = y 2i + (2xy) j Evaluate the line integrals in Problems 17-18 by parametriza
tion.
17. [(sin τry) dx + (cos τrx) dy], where C is the line segment
Jc
from (1, 0) to (π, 0), followed by the line segment from
(τr, 0) to (π, π)
In Problems 7-12, fin d F ∙ <7R, where C is the curve 18. I (z dx - x dy + dz), where C is the arc of the helix
Jc Jc
R(∕) = ri + t2j, 1 ≤ t ≤ 2. Note that these are the same as the
vector fields in Problems 1-6. x = 3 sin t, y = 3 cos t, z = t for 0 ≤ t ≤ f .
7. F = 2i - 3j 8. F = xy 2i + x 2yj
- 3 - 4
In each o f Problems 19-33, evaluate the line integral or the
9. F = y i + ( -3 x y + cosy)j 10. F = y 2i + (2xy)j surface integral. In each case, assume the orientation o f C is
counterclockwise when viewed from above or that N is the unit
outward normal.
12. F = ∣~2xtan 1∩ J - y j i + |_2y tan 1 + x jj
19. [yz dx + xz dy + (xy + 2)] dz
Jc
C is the curve R(Z) = (tan- 1 t)i + t2j - 3∕k, 0 ≤ t ≤ 1.
In Problems 13-16, fin d div F and curl F.
13. F = x i + yj + zk [(x2 + y) dx + xz dy - (y + z) dz]
14. F = (tan - 1 —)i - j3j + z 2k
∖ x C is the curve R(z) = ri + t2j + 2k, 0 ≤ t ≤ 1.
33. J V<⅛■N dS, where φ(x, y) = 2x + 3y

21. J J (3x 2 + y - 2 z ) dS
S is the surface R(M, v) = ui + (w + v)j + uk, 0 ≤ u ≤ 1, where .8’ is the portion of the plane ax + by + cz = 1
0 ≤ v ≤ 1. (a > 0, b > 0, c > 0) that lies in the first octant.
22. J J F ∙ N dS In Problems 34-38, fin d j J F ∙ N dS.
F = 3xi + z 2j —2yk and .8' is the surface of the hemi 34. F = x 2i + y 2j + x 2k, and >8' is the surface o f the unit cube
0 ≤ x ≤ 1, 0 ≤ y ≤ l , 0 ≤ z ≤ 1.
sphere z = V 4 —x 2 —y 2.
35. F = 2yzi + (tan^1 xz)j + e'∙vk, and S is the surface of the
sphere x 2 + y 2 + z 2 = 1.
23. (x dx + x dy - y dz)
Jc 36. F = xi —4j + 3k, and .$ is the paraboloid y = x 2 + z 2
C is the curve R(Z) = ti + t2j + tk, 0 ≤ t ≤ 1. with x 2 + z 2 ≤ 9. The disk x 2 + y 2 = 9 is omitted; that
is, the paraboloid is open on the right.
24. f (x 2 dx - 3y2 dz) 37. F = xyzi + xyzj + xyzk, and .8' is the surface of the
Jc five faces of the unit cube 0 ≤ x ≤ 1, 0 ≤ y ≤ 1,
C is the line segment from (0, 1, 1) to (1, 1,2). 0 < z ≤ 1.
38. F = xyi - 2zj, and 5 is the surface defined parametrically
25. I (x 2dx + y dy) by R(w, v) = ui + vj + uk for 0 ≤ u ≤ 1, 0 ≤ v ≤ 1.
Jc
Find all real numbers c for which each vector field in Problems
C is the curve R(Z) = (t sint)i + (1 - t cost)j,
39-41 is conservative.
0 ≤ t ≤ 2ττ.
39. F(x, y) = (V χ + 3xy)i + (cχ2 + 4y)j
26. (xy dx - x 2 dy)

40. F(X , J ,) = ⅛ - 4 + (±-1∖∙
v ■’ ∖X 2 X 2∕ ∖X2 x)i
Jc
C is the square with vertices (1, 0), (0, 1), ( - 1, 0), (0, -1 ) 41. F(x, y , z) = ey z lx ∖(⅛ ∏ i + ( - ) j + ( - ) k
traversed counterclockwise.
42. If F = (y 2 + x ~ 2ye x ,y )i + (2xy + z —x ~ l ex ,y )j + yk, ei
27. I (y dx + x dy - 2 dz) ther show that F is not conservative, or show that it is and
Jc find a scalar potential.
C is the curve of intersection o f the cylinder x 2 + y 2 = 2x 43. Find the work done when an object moves in the force
and the plane x = z. field F = 2xi - (x + z)j + (y - x)k along the path given
by R(∕j = r2i + (t2 - ∕)j + 3k, 0 ≤ t ≤ 1.
28. ∣[(y + z) dx + (x + z) dy + (x + y) dz] 44. Show that the force field
Jc
C is the curve of intersection of the sphere F = yz 2 '∖ + (xz 2 - l)j + (2xyz - l)k
x 2 + (y - 3)2 + z 2 = 9 and the plane x + 2y + z = 3.
is conservative and determine the work done when an ob
ject moves in the force field from the origin to the point
29. I (-2 y dx + 2x dy + dz) ( 1 ,0,1).
Jc
45. Find a region R in the plane where the vector field
C is the circle x 2 + y 2 = 1 in the plane z = 3.
F = —L ( + j) is conservative. Then evaluate I F ∙ r∕R,
30. J (curl yi) ■N d S x + /
i i, Jc
where C is any path in R from the point P 0(a, b) to
S is the hemisphere z = V 1 - x 2 - y 2 . P 1(c, d).
46. Solve (2y sin x co sx + y 2sinx) dx + (sin2x —2 y cosx) dy =0,
31. j J (2x3i + y 3j + z 3k) ∙ N dS
subject to y(0) = 1.
47. Solve (6xy + 2y 2 - 5) dx + (3x2 + 4xy - 6) dy = 0, sub
S is the surface o f the ellipsoid 2x 2 + y 2 + z 2 = 1. ject to y(0) = 0.
48. If u is a scalar function and F is a continuously differ
32. J j (y 2i + y 2j + yzk) ■N dS
entiable vector field, show that
curl(uF) = u curl F + (Vu × F)
S is the surface of the tetrahedron bounded by the plane
2x + 3y + z = 1 and the coordinate planes, in the first 49. Show that div(F × G) = G ∙ curl F — F ■curl G , for any
octant. continuously differentiable vector fields F and G.
50. A vector field F is solenoidal (or incompressible) in a x j’-plane. Let c be a constant, and let R = .xi + yj + zk.
region D if div F = 0 throughout D. If F and G are both Use Stokes’s theorem to evaluate
conservative vector fields in D , show that F × G is
solenoidal. (ck × R) ∙ t/R
51. If F = curl G , show that F is solenoidal. (See Problem 50.)
52. Suppose F = ∕( x , y, z)A, where A is a constant vector and 62. a. Show that if the scalar function w is harmonic, then
/is a scalar function. Show that curl F is orthogonal to A
and to V ∕ V ∙ (w V w) = ∣∣Vw∣∣2
53. If A is a constant vector and F is a continuously differenti b. Let w = x - y + 2z, and let S, be the surface o f the
able vector field, show that div(A × F) = - A ∙ curl F. sphere x 2 + y 2 + z 2 = 9. Evaluate
' x dx - y dy , .
cJ-- 2— —— ’ where C is any
χ ~^ y
path in the xy-plane that is interior to the region x > 0,
I
s
y < x , y > —x and connects the point (5, 4) to (2, 0).
Hint'. See the normal derivative in Section 14.4.
55. Evaluate I I — dx + - dy ), where C is the closed path 63. A particle moves along a curve C in space that is
Jc '-λ- x ) given parametrically by R(t) = x(f)i + y(t)j + z(t)k for
(x - 2)2 2
+ y = 1, traversed once counterclockwise. a ≤ t ≤ b. A force field F is applied to the particle in
56. T H IN K T A N K P RO BLEM Let w(x, y) and v(x, y) be func such a way that F is always perpendicular to the path C.
tions of two variables with continuous partial derivatives How much work is performed by the force field F when
everywhere in the plane, and suppose that ιι and v satisfy the particle moves from the point where t = a to the point
the equation where t = bl
64. If D is the electric displacement field, then div D = φ,
∂u _ ∂v
where φ is the charge density. A region of space is said to
∂y ∂x
be charge-free if φ = 0 there. Describe the charge-free
for all (x, y). Are u and v necessarily harmonic? Either regions of the electric displacement field
show that they are or find a counterexample. D = 2x 2i + 3y2j - 2z 2k.
57. a. Find a region in the plane where the vector field 65. A satellite weighing 10,000 kg travels in a circular orbit
7,000 km from the center of the earth. How much work is
done by gravity on the satellite during half a revolution?
66. If F is a conservative force field, the scalar function/such
is conservative, and find a scalar potential for F. that F = - V / is called the potential energy. Suppose an
object with mass 10 g moves in the force field in such a
b. Evaluate F ∙ <7R, where C is a path from (1, 1) to way that its speed decreases from 3 cm/s to 2.5 cm ∕s.
Jc What is the corresponding change in the potential energy
(3, 4). Are there any limitations on the path C ? of the object?
Explain.
67. Find the work done by the force field
58. Determine the most general function u(x, y) for which the
vector field F = u(x, y)i + (2y ex + y 2e l v )j will be conser
F = 2xyzi + x 2y + ⅛ tan 1 -)k
vative. . z2 z∕
59. Consider the line integral [ (— + — ), where C is the
in moving an object along the circular helix
closed triangular path formed by the lines y = 2x, R(∕) = (sin πt)i + (cosτrt)j + (2t + l)k
x + 2y = 5, and x = 2, traversed counterclockwise. First
for 0 ≤ t ≤ ∣.
evaluate the line integral directly (by parametrizing C)
and then by using Green’s theorem. 68. Find the work done when an object moves against the
60. If 5, is a closed surface in a region R and F is a force field F = 4xyi + (3x + y)j from (1, 0) to ( -1 , 0)
continuously differentiable vector field on R , show that y2
along the top half of the ellipse x 2 + < j = 1. What value of
J J (curl F ∙ N) dS = 0 k minimizes the work?
69. One of Maxwell’s laws of electromagnetism states that
where N is the outward unit normal to S. div E = - Q, where E is the electric intensity, ε (a con
61. A certain closed path C in the plane 2x + 2y + z = 1 is stant) is the permittivity o f the medium, and Q is the
known to project onto the unit circle x 2 + y 2 = 1 in the charge density (charge per unit volume). Show that
76. T H IN K T A N K P RO BLEMProve or disprove that it does

not matter which corner is removed in Problem 75.
77. Show that a lamina that covers a standard region D in the
where q is the total charge in a solid region R with plane with density p = 1 has moment of inertia
boundary surface .S'. This is called Gauss’s law.
I = X ( - y 3 dx + x 3 dv}
70. Evaluate the line integral i Jc
—y dx + x dy with respect to the z-axis, where C is the boundary curve
x2 + y2 o f D.
where C is the limaςon given in polar coordinates by 78. Use vector analysis to find the centroid of the conical
r = 3 + 2 cos θ, 0 ≤ θ ≤ 2ττ. surface (shell), z = V x 2 + y 2 between z = 0 and z = 3.
71. Suppose f and g are both harmonic in the region R with 79. Having conserved as much energy as possible after his
boundary surface S'. Show that escape from the Death Force Field (Problem 33, Section
14.4), the spy continues his search for Purity and Scelerat.
He rushes into a small room and finds no Scelerat but two
S S Purities struggling ferociously on the floor. As he sepa
rates them, he notices one is dressed in red and the other
in green. The red one cries, “ She’s my evil twin, Rottona.
s D She escaped from prison and is going to take my place.”
QΓ
The green Purity shouts back, “ Liar! Γm the real Purity.
Hint', ~ denotes the normal derivative (see Section 14.4).
Γve been tied up here all along and just got free.” The spy
72. Show that curl(curl F) = V(div F) - V 2F if the compo is better at logic than with colors. He thinks for a moment
nents o f F have continuous second-order partial deriva and then says to the Green Purity, “ I remember not —
tives. you are the one who has been trying to kill me. You’re
73. Evaluate the surface integral going back to prison, but first, Γm going to give you the
same kind of chance to escape you gave me. You choose
an integer N , and then your sister will pick a number x
and you will have
ς w . = 3 exp [U + Y ) 2 + 4 Y + 22]
where S is the surface of the unit sphere x 2 + y 2 + z 2 = 1 ' ’ X) (x + 7V)2(e4° + e8 v )
and f is a scalar field such that ∣∣V∕∣∣2 = 3/ and
d i v ( ∕V ∕) = If. Remember that — denotes the normal minutes head start before I come after you.” The Red
Purity laughs and says, “ Choose carefully, dear sister,
derivative. because I plan to minimize your time!” The Green Purity
74. Evaluate the surface integral J J”dS, where 5 is the torus declares, “ It’s not fair. Γm innocent.” Then she blurts out
s a number. The Red Purity’s smile fades and she cries,
R(w, v) = [(a + b cosv)cosw]i “ Can you believe it! She always was lucky.” What number
+ [(α + b cos v)sin z∕]j + (b sin v)k did the Green Purity choose? How much time will she
get? What happens next?
for 0 < b < a and 0 ≤ u ≤ 2ττ, 0 ≤ v ≤ 2π.
80. P U T N A M E X A M IN A T IO N P RO BLEM A force acts on the
75. Evaluate J J F ∙ N dS. where F = xi + j j + zk and S is a
element ds of a closed plane curve. The magnitude of this
force is r~' ds, where r is the radius of curvature at the
cubic surface with a corner block removed, as shown in point considered and the direction o f the force is perpen
Figure 14.35. dicular to the curve; it points to the convex side. Show
that the system o f such forces acting on all elements of the
curve keeps it in equilibrium.
Figure 14.35 Cube with a corner block removed

GUEST ESSAY
Continuous vs. Discrete Mathematics by William F. Lucas
M athem aticians often distinguish between discrete and continuous mathematics.

The latter is illustrated by the calculus and its many descendants, including the
subject of differential equations. Continuous mathematics deals with “solid”
infinite sets such as the real num ber line R and functions defined on R. One of the
prim ary concerns is with the solutions of differential equations, which are typical
families of such solid curves or surfaces. The many applications of such differen
tial equations to discoveries in the physical sciences and engineering over the past
three centuries has been one of the greatest intellectual success stories of all times.
P rio r to the invention of calculus by Newton and Leibniz, most m athematics was
of a discrete nature. It dealt with sets that had a finite num ber of elements and
with infinite but “countable” sets such as the natural numbers. These sets often
William F. Lucas is a professor include continuous curves such as the conics, which can be characterized by a
of mathematics and departm ent small num ber of conditions. Discrete mathematics also deals with the solutions to
chairman of the departm ent algebraic equations, which are usually discrete sets. The twentieth century has
The Claremont Graduate witnessed the creation of many additional subjects in the discrete direction. This
School. He is an Avery Fellow to development was spurred on by the rapidly increasing use of mathematics in the
Harvey Mudd College and is social, behavioral, decisional, and system sciences, where the items under investi
known for his research in game gation are typically finite in num ber and not readily approximated by some
theory. This modern subject continuous idealization. Moreover, nearly all aspects of the ongoing revolution in
provides a mathematical ap digital computers involve discrete considerations.
proach to the study of conflict, A metal bar at the minute level is composed of discrete molecules, atoms, and
cooperation, and fairness. Dr. elementary particles. Many of its physical properties, however, can be determined
Lucas has also been active in by viewing the bar as a solid continuum and employing the analytical techniques
educational reform efforts with of calculus. One applies the basic laws of physics to express the local (infinitesi
a goal toward introducing re mal) properties of the bar in terms of differential equations. The solutions of these
cently discovered topics into the equations, in turn, provide an excellent description of the observed global
mathematics curriculum. behavior of the bar.
O n e of the greatest m athematical discoveries of the twentieth century in
continuous m athematics is the famous fixed-point theorem, published in 1912 by
the Dutch m athem atician L. E. J. Brouwer (1881-1966). It states that any
continuous function f from a set S (with certain desirable properties) into the
same set 5 has at least one fix e d point x 0 , that i s ,∕( x 0) = xθ. For example, for any
continuous function y = f ix ) from the set S' = {x ε R: 0 ≤ x ≤ 1} into S there is
some num ber x 0 ε S such that the curve y = f(x ) crosses the line y = x at x 0 . A
head of hair must either have a seam (a parting of the hair, a discontinuity) or else
a cowlick (where the hair stands straight up, a fixed point). Fixed points are
realized in many physical and social phenomena such as equilibrium states in
Most hum an heads have a fixed
mechanics or stable prices in economics.
point, in the form of a whorl,
sometimes called a cowlick, Like many of the outstanding results in continuous mathematics, Brouwer’s
from which all the hair radiates. theorem has an analog in discrete mathematics, known as the labeling lemma of
It would be impossible to cover the German m athematician Emanuel Sperner, which appeared in 1928. Given a
a sphere with hair (or with radi triangle with vertices 0, 1, and 2 as shown in Figure 14.36, one can partition the
ating lines) without at least one interior of this triangle into (nonoverlapping) smaller triangles in any possible
such fixed point. way.
Historical Note
W ILLIAM R OWAN H AMILTON

(1805-1865)
It is said that Hamilton is the
most renowned Irish mathema
tician. He was a child prodigy
who read Greek, Hebrew, and
Latin by the time he was five, Figure 14.36 Arbitrary partitioning of a triangle
and by the age of ten he knew
over a dozen languages. Many
mathematical advances are One can then label each of the newly created vertices with any one of the numbers
credited to Hamilton. For ex 0, 1, or 2. There is only one stipulation on how the new vertices on the perimeter
ample, vector methods in ana of the triangle can be labeled: Any vertex on a particular side of Δ012 cannot be
lytic geometry and calculus, as labeled with the number appearing on the main vertex opposite (disjoint from)
well as a system of algebraic this side. For example, the side 01 of Δ 012 cannot contain a vertex with the label
quantities called quaternions,
2. Sperner’s lemma states that there must exist at least one elementary triangle
were all first developed by him.
At first, the role of vectors was
inside Δ012 (or an odd number, in general) whose three vertices are “completely
secondary to a system of qua labeled” in the sense that they have all three labels 0, 1, and 2. There are three such
ternions, but by the early 20th triangles (shown in color) in Figure 14.36.
century, the notation and ter There is an elementary constructive proof of Sperner’s lemma that uses the idea
minology of vectors dominated.
of a “path-following algorithm” published by Daniel I. A. Cohen in 1967. This
Much of the credit for the
eventual emergence of vector
proof is illustrated by the dashed line in Figure 14.36. One can enter into Δ012
methods goes not only to from outside via some elementary triangle T γ whose perimeter edge carries the
Hamilton but also to the scien labels 0 and 1. The third vertex on this elementary triangle T l must be either 2 (in
tists James Maxwell (1831- which case 7j is a completely labeled elementary triangle), or else it is 0 or 1. In
1879), J. Willard Gibbs (1839- the latter cases, one can exit T l via a second 01 edge. One can continue to enter
1903), and Oliver Heaviside and exit subsequent elementary triangles T1, T y . . . through successive 01 edges
(1850-1925). until one finally reaches an elementary triangle Tn that has completely labeled
vertices 0, 1, and 2. (If this path were ever to exit Δ012, then there must be still
another 01 edge on the perimeter of Δ012 where one can reenter the main triangle
and continue along the path, as did occur once in Figure 14.36.)
This path-following method also extends to a proof of Sperner’s labeling lemma
for dimensions higher than two. Furthermore, this approach is used in practical
problems to arrive at a “very small” elementary triangle that can serve as an
approximation to a fixed point, when it is excessively difficult or impossible to
determine the precise location of the fixed point itself. In such cases, the broken
path in Figure 14.36 can be viewed as the analog in discrete mathematics of the
continuous curves one arrives at when solving differential equations.
Around 1945, the great Hungarian-American mathematician John von Neumann

(1903-1957) pointed out the need for approximating paths in a discrete manner
when it proves too difficult to arrive at the exact continuous solutions to certain
ordinary or partial differential equations. Cohen’s constructive proof of Sperner’s
lemma, and work in the late 1960s by the mathematical economist Herbert Scarf
on approximating equilibrium points, have paved the way for one very rich and
extensive discrete theory that provides numerical approximations for continuous
phenomena when the latter problems cannot be solved directly by the many
analytic techniques of continuous mathematics.
MATHEMATICAL ESSAYS
1. If your college offers a course on discrete mathematics, 4. Write a 500-word essay on William Rowan Hamilton.
interview an instructor of the course and, using that 5. Write a 500-word essay on Pierre de Fermat.
interview as a basis, write an essay comparing continuous 6. Write a 500-word essay on Archimedes.
and discrete mathematics.
7. Write a 500-word essay on the history of Green’s theorem,
2. Draw several triangles and attempt to draw a counterex Stokes’s theorem, and the divergence theorem.
ample for Sperner’s Labeling Lemma. In each case, show
8. Write a report on four-dimensional geometry.
how this lemma is satisfied.
3. Consider two sheets of paper a Book Report “We often hear that mathematics consists
containing the numbers 1 to mainly in ‘proving theorems.’ Is a writer’s job mainly that
120, as shown in the drawing. o f ‘writing sentences’? A mathematician’s work is mostly a
tangle of guesswork, analogy, wishful thinking, and frus
tration, and proof, far from being the core of discovery, is
more often than not a way of making sure that our minds
are not playing tricks. Few people, if any, had dared write
this out loud before Davis and Hersh. Theorems are not to
mathematics what successful courses are to a meal. The
nutritional analogy is misleading. To master mathematics
is to master an intangible view . . . .” This quotation
If the top sheet is crumpled and dropped on the bottom comes from the introduction to the book The Mathemati
sheet, the fixed-point theorem tells us that at least one cal Experience by Philip J. Davis and Reuben Hersh
point must still lie directly above its starting point. In the (Boston: Houghton Mifflin, 1981). Read this book and
following drawing, it is a point in the region of the number prepare a book report.
78. Perform this experiment several times in an attempt to 10. Make up a word problem involving vector fields. Send
find a counterexample. In each case, show how the fixed- your problem to:
point theorem is satisfied.
Mathematics Editor,
1 2 3 I 4 * 5i . 1 6 1 7 8 9 10 re: Bradley, Smith: CALCULUS
11 12 L3 χ < 5 ∕ ⅛ ! 1 L »8 19
j
20 Prentice Hall
2 ∣”7 > X N √ ∖' ∕ ⅞ 0 ' 28 29 30
— j
1- ~ - * - r ¾ b X v , × A χ v -----------
-1 U J - A √ ⅛ ⅛ > ⅛ ⅜ > ' ∖ * ι 40
113 Sylvan Ave., Route 9W
50
Englewood Cliffs, NJ 07632
60
t ∙ ,⅛ Λ - ⅛ e t > ∙ '> X f . 59 70
71
i
80 (along with credit to the problem poser).
81 82 83 84 ≡ ⅛ W Γ88 89
90
91 92 93 94 95 96 97 98 99 100
101 102103 I M 105 106 107 108109 110
111 112113 114 115 116 117 118 119 120
CUMULATIVE REVIEW FOR CHAPTERS 12-14
1. ■ What Does This Say? Suppose you tell a fellow

21. (x 3 + y 3) ds, where C is given by
college student that you are about to finish a calculus
course. The student has not had any mathematics R(Z) = (cos⅜)i + (sin3 t)j, for 0 ≤ t ≤ 2ττ
beyond high school and asks you, “ What is calculus?”
Answer this question using your own words. 22. Find the equation for
the tangent plane and
2. ■ What Does This Say? There are three great ideas in
the normal line to
calculus: the idea of a limit, the idea of a derivative, and
z = x 2 + y 2 + sin xy at
the idea o f an integral. In your own words, explain each
P = (0, 2, 4).
of these concepts.
3. ■ What Does This Say? Chapters 12-14 were con 23. What are the dimen
cerned with functions of several variables. This is often sions o f the rectangular
referred to as multivariable caculus. In your own words, box of fixed volume Vo
discuss what is meant by multivariable calculus. that has minimum sur
face area?
F in d f χ , f a n d ∕ vv for the functions whose equations are given 24. Suppose we model a
in Problems 4-9.’ heat-seeking missile by
making the following
4. f{ x , y) = 2x 2 + xy - 5y 5. f( x , y) = x 2 ey lx
assumptions. Suppose the spherical ball in the nose of
v2 - 14
6- f( χ , y) = 7. ∕( x , y) = y sin 2 x + cos xy the missile with equation x 2 + y 2 + z 2 ≤ 100 is heated
so that its temperature (in degrees Celsius) at the point
v2 + 1 ,2
8. f{ x , y) = e x + y (x, y, z) is given by
9∙ f( χ , y) =
T(x, y, z) = γθ (x 2 + y + z 3)
Evaluate the integrals in Problems 10-17.
f l f2 x f4 f W where x , y, and z are measured in kilometers.
10. 11.
a. Find the rate at which the temperature is changing as
(x, y, z) moves from the point P o ( - 2, 9, 1) toward
12. (x + y + z) dx dy dz Q (l, - 3, 5).
b. If a heat-seeking missile is at P o , in what direction
15π r ~ rsin φ will it travel to maximize the rate o f heat increase?
13. ρ 3sin φ dρ dθ dφ
Jo Jo c. What is the maximal rate of increase (in degrees
Celsius per kilometer)?
14. 15. 25. Use double polar integration to find the area inside the
R R cardioid r = 1 - cos θ and outside the circle r = 1.
26. Find the volume of the solid bounded above by the
surface z = x 2 + y 2 + 1 and below by the circular disk
16. 17. ∣ J x sin x y dA; x 2 + y 2 ≤ 1 in the xy-plane.
R 27. Find the surface area o f that portion of the paraboloid
y iθ ≤ y ≤ t z = x 2 + y 2 that lies below the plane z = 16.
Evaluate the line integrals in Problems 18-21. 28. Find the center o f mass o f the lamina whose shape is the
region inside the circle x 2 + y 2 = 4 in the first quad
18. fy 1z dx + 2xyz dy + x y 2 dz), where C is any path rant, given that the density at any point (x, y) is
J c
from (0, 0, 0) to (1, 1, 1)
δ(x, y) = x + y. Hint: Reverse the order of integration.
29. A force field F(x, y) = (x - 2y)i + (y - 2x)j acts on an
object moving in the plane. Show that F is conservative,
19. (5 xy dx + 10 yz dy + z dz), where C is given by and find a scalar potential for F. How much work is
lc
done as the object moves from (1, 0) to (0, 1) along any
c = t2 , y = t, z = 2t3 for 0 ≤ I ≤ 1
path connecting these points?
20. F ∙ √R, where F = yzi - xk, and C is the boundary of 30. A particle of weight w is acted on only by the constant
Jc gravitational force F = — wk. How much work is done
the triangle (1, 1, 1), (1, 0, 1), (0, 0, 1), traversed once in moving the weight along the helical path given by
counterclockwise as viewed from the origin x = cos t, y = sin t, z = t for 0 ≤ t ≤ 2 TT?
A P P E N D IC E S
APPENDIX A Theorems by Chapter

CHAPTER 1
THEOREM 1.1 Distance between two points in the plane
The distance d between the points P(x 1, p 1) and (x2, y 2) in the plane is given by
d = V(∆x)2 + (∆j-j2
THEOREM 1.2 Slope criterion for parallel and perpendicular lines
If €, and / are nonvertical lines with slopes m l and m 1, then
and ∙f, are parallel if and only if w = m 1.
and €2 are perpendicular if and only if m l m 2 = - l o r m 1 = ~ ~
THEOREM 1.3 Limits of trigonometric functions

If c is any number in the domain of the given trigonometric function, then
lim cos x = cos c lim sec x = sec c
x →c
lim sin x = sin c lim esc x = esc c
x →c
lim tan x = tan c lim cot x = cot c
THEOREM 1.4 Continuity theorems

I f / is a polynomial, or a rational function, a power function, or a trigonometric
function then, / is continuous at any number x = c for which ∕(c) is defined.
THEOREM 1.5 Properties of continuous functions

If s is a real number (called a scalar) and / and g are functions which are
continuous at x = c, then the following functions are also continuous at x = c.
Scalar multiple sf
Sum and difference f + 5r a n d ∕ - g
Product fg
Quotient provided g(c) ≠ 0
o
Composition / og provided / i s continuous at g(c)
A -l
A -2 Appendices
THEOREM 1.6 The intermediate value theorem

If f is a continuous function on the closed interval [a, b] and L is any number
between f(a) and ∕(∂), then there exists at least one number c on (a, b) such that
» = L.
THEOREM 1.7 Root location theorem

If f is continuous on the closed interval [a, b] and if /(a) and f(b) have opposite
algebraic signs, then ∕(c) = 0 for at least one number c on (a, b).
THEOREM 1.8 Limit limitation theorem

Suppose lim ∕(x ) exists and ∕(x ) ≥ 0 throughout an open interval containing the
x →c
number c, except possibly at c itself. Then
lim ∕(x ) ≥ 0
THEOREM 1.9 The squeeze theorem

If g(χ) ≤ ∕(χ ) ≤ ∕z(χ) for all x in an open interval containing c (except possibly at
c itself) and if
lim g(x) = lim ∕r(x) = L
x →c x →c
then lim ∕(x ) = L.
CHAPTER 2
THEOREM 2.1 Equation of a line tangent to a curve at a point
I f / is a differentiable function at x0, the graph of y = ∕( x 0) has a tangent line at the
point ∕ j (x, ∕( x 0)) with slope ∕'( x ) and equation
y = ∕'(⅞ )( χ - -⅞) + ∕(⅞ )

If a function f is differentiable at c, then it is also continuous at x = c.
THEOREM 2.3 Constant rule

A constant function∕(x ) = k has a derivative∕'( x ) = 0; that is, in Leibniz notation
≠ (∕c) = 0
dx ’
THEOREM 2.4 Power rule

For any real number n, the power function ∕(x ) = x" has the derivative
∕'( x ) = wx"- 1 ; that is, in Leibniz notation
∕( χ " ) - ∞ " - l
THEOREM 2.5 Basic rules for combining derivatives—Procedural forms

If f and g are differentiable functions at x and a, b, and c are any real numbers,
then so are the functions cf f + g, fg, and f / g (for g(x) ≠ 0, and their derivatives
satisfy the following formulas:
App. A: Theorems by Chapter A -3
Name of Rule Prime Notation Function N otation Leibniz Notation
Constant multiple [c∕(x)]' = cf∖ x)

≠ ( c ∕) = J -
(c ∕)' = cf' dxy j , dx
d_, f , . _ df dg
Sum rule ( f+ g ) ' = f ' + g' [j∖ x)+g(x}∖ = f ∖ x) + g'(x) dx dx dx
Difference rule [ ∕w - ^(χ)]' = ∕' ( χ ) - ∕( χ ) ]
d L i f - fx = V - ⅛
( f ~ g)' = f ' ~ g ' d x 8 > dx dx
The constant multiple, sum, and difference rules can be combined into a single rule called the linearity rule.
Linearity rule ( α ∕ + bg)' = a f' + bg' ∖ af(x) + ⅛ (x)]' = af'(x) + bg'(x) Tx^1'+ bg) = a ^ + b ⅛
Product rule ⅛ w -f⅛ * l⅛

(fg)' = fg' + f 'g [/(*)£(*)]' = ∕(χ)√(χ) + ∕'(χ)g(χ)
/ λ Γ⅛
( f∖ = g f' ~ fg' Γ∕( χ ) ^∣
’ _ g(χ ) ∕ ' ( χ ) ~f(×)g'(χ) √ 1/ 1 _ 8 dx j dx
Quotient rule ∖ g) g' Ls(χ)J [g(χ)]2 dx ∖ gj g2
Corollary to THEOREM 2.5 The extended linearity rule

If∕ 1, f v . . . , f 2 are differentiable functions and α 1, a2, . . . , an are constants, then
√ df df. df
[∙a.1√f1 ÷ a2∙f'2t + ∙∙∙ + an∙>
-dj-χ ∖ f n]*~ a.1 -djx— I- rz2, dx— I- ∙ ∙ ∙ + a» -dj x—
THEOREM 2.6 Special limit theorems for the sine and cosine
lim = 1 lim -c o s ⅞ - 1 =0
Λ→O n ∣
t→ o n
THEOREM 2.7 Derivatives of the cosine and sine functions

The functions cos x and sin x are differentiable for all x and
sin x = cos x cos x = -sin x
THEOREM 2.8 Derivatives of the trigonometric functions

The six basic trigonometric functions, cos x, sin x, tan x, sec x, esc x, and cot x, are
all differentiable wherever they are defined and
d sm
∙ x = cos x ⅛ c o sx = -s in ^
tan x = sec2x cot X = -csc⅛
sec x = sec x tan x esc x = -esc x cot x
THEOREM 2.9 Chain rule

If y = f(u) is a differentiable function of u and u = g(x) is a differentiable function
of x, then y = ,∕[y(.v)] is a differentiable function of x and
dy = r/y du
dx du dx
A -4 Appendices
THEOREM 2.9a The chain rule (alternate form)

If u is differentiable at x and f is differentiable at u(x), then the composite
function f o u is differentiable at x and
( ∕o u)'(x) =f'(u)u'(x) or
THEOREM 2.10 Power rule for rational exponents

If n = is a rational number, then (x n) = nx n~i .
THEOREM 2.11 The Newton-Raphson method

To approximate a root of the equation ftx ) = 0, start with a preliminary estimate
x 0 and generate a sequence of increasingly accurate approximations x l , x 2, x 3, . . .
using the formula
f(x ) λ , 0∩
x n+ l = x n f '( χ n ) λ
f(X n) ≠
This sequence of approximations will either approach a limit that is a root of the
equation or the sequence does not have a limit.
CHAPTER 3
THEOREM 3.1 The extreme value theorem
A continuous function f on a closed, bounded interval [a, ⅛] has an absolute
maximum and an absolute minimum.
THEOREM 3.2 Critical value theorem

If a continuous function f has a relative extremum at c, then c must be a critical
value of f.
THEOREM 3.3 Rolle’s theorem

Suppose f is continuous on the closed interval [a, ⅛] and differentiable on the open
interval (a, b). Then, if∕(α ) = f(b), there exists a number c between a and b such
that ∕'( c ) = 0.
THEOREM 3.4 The mean value theorem (MVT) for derivatives

If f is continuous on the closed interval [a, ⅛] and differentiable on the open
interval (a, b), then there exists at least one number c such that
f(b^) - f( a ) , ,
b, _ a = f (c) for a < c < b
THEOREM 3.5 Zero derivative theorem

Suppose f is a continuous function on the closed interval [a, b} and differentiable
on the open interval (a, b). Then if f'(x ) = 0 for all x in {a, b), the function
f is constant on [a, b].
THEOREM 3.6 Constant difference theorem

Suppose the functions f and g are continuous on the closed interval [a, b∖ and
differentiable on the open interval (a, b). Then if f'(x ) = g'(x) on (a, b), there
exists a constant C such that ∕(x ) = g(x) + C for all x on [a, b].
THEOREM 3.7 Increasing and decreasing function theorem

Let f be differentiable on the open interval (a, b). Then the function f is strictly
increasing on (a, b) if
f'(x ) > 0 for a < x < b
a n d ,/is strictly decreasing on (a, b) if
f'(x ) < 0 for a < x < b
THEOREM 3.8 Limits to infinity

If n is a positive rational number, and A is any nonzero real number, then
lim ⅛ = 0
χ→+∞ Λ
Furthermore, if x n is defined when x < 0, then
lim -2⅜
x→-∞ Xn
THEOREM 3.9 l, H6pitaΓs rule

Let f and g be differentiable functions on an open interval containing c (except
possibly at c itself). If
lim produces an indeterminate form θ or then
∕(x ) ∕'( x )
hm = lim ~ ÷ ∕
x→c g(Λ) x →c g∖ χ)
provided the limit on the right side exists (or is infinite).
THEOREM 3.10 Antiderivatives differ by a constant

If Eis an antiderivative of the continuous function/ then any other antiderivative
of /m u st have the form
G(x) = E(x) + C
THEOREM 3.11 Basic integration rules
PROCEDURAL RULES Differentiation Formulas Integration Formulas
lu = C1 f du = c J7(«) du
Constant multiple
( f + s) ∣ [f(u) + ⅛(κ)] du = J /(«) du + J g(u) du

Sum rule
⅛ t <f ~ 8) = fa J [/(«) - #(«)] du = J f(u) du - [ g(u) du

Difference rule
f a ( a f + bg) = o %→ b ^ 1 [af(u) + ⅛g(u)] du = a f(u) du + b ( g(u) du

Linearity rule
A -6 Appendices
BASIC INTEGRATION RULES
Constant rule -r- (c) = 0 ∣ 0 du = c

duy ’ 1
,n +l
Power rule (t∕") = nu"~ l 1 u" du = —
1
— r + C; ∏≠ - 1
n+ I ’
Trigonometric rules (cos w) = -s in w !I sin u du = -c o s u + C
(sin u) = cos u ,∣ cos u du = sin u + C
(tan u) = sec2u ∣ sec⅛ du = tan u + C
(esc u) = -e sc u cot u ;I esc u cot u du = -e sc u + C
(cot u) = - c s c 2w ;I csc 2w du = -c o t u + C
CHAPTER 4
THEOREM 4.1 Basic rules and formulas for sums
For any numbers c and d, and positive integers k, m, and n:
1. Constant term rule £ c = c + c + ∙∙∙ + c = nc

n terms
n n n
2. Sum rule Σ (a k + ⅛) = Σ a k + Σ ⅛
k= 1 k= 1 A :=1
n n
3. Scalar multiple rule Σ cak = c Σ ak
k=∖ k=∖
n n n
4. Linearity rule ∑ (ca + dbk ) = c∑ ak + d ∑ bk
k≈∖ k k=∖ k=l
5. Subtotal rule If 1 < m < n, then
n m n
Σ ak~ ∑ ak + ∑ a
'k
k=∖ k=∖ k=m+∖
6. Dominance rule If ak ≤ bk for k = 1, 2, . . . , n, then
n n
Σ a k ≤ Σ ¾
THEOREM 4.2 Existence of Riemann sum

If f is continuous on an interval [a, b], then f is integrable on [a, ⅛].
THEOREM 4.3 General properties of the definite integral
Linearity rule If f and g are integrable on [a, b], then so is rf + sg for

constants r, s.
rb rb rb
[rf(x) + ⅞z (χ )] dx = r f(x) dx + s g(x) dx
Ja Ja Ja
Dominance rule If/a n d g are integrable on [a, b∖ and ∕(x ) ≤ g(x)

throughout this interval, then
fb fb
f(x) dx ≤ g(x) dx
Ja Ja
Subdivision rule For any number c such that a < c < b

rb rc rb
f(x) dx = f(x) dx + ∕(x ) dx
Ja Ja Jc
assuming all three integrals exist.
THEOREM 4.4 The fundamental theorem of calculus

If f is continuous on the interval [a, ⅛] and F is any function that satisfies
F'(x) = f(x) throughout this interval, then
rb
f(x) dx = F{b) - F(a)
Ja
THEOREM 4.5 Integration by substitution

Let / g, and u be differentiable functions of x such that
∕(x )= ^ (w )⅛
Then
∣f(x) dx = g(u) ~^,dx = G(u) + C
for G an antiderivative of g.
THEOREM 4.6 Substitution with the definite integral

Suppose/ is continuous on the set of values taken on by g. If g' is continuous on
[a, b∖ and if/ has an antiderivative on that interval, then
r⅛ fg ^
∕[g(x)]g (x) dx = /(») dιι u = g(x), du = g (x) dx
Ja jg(a)
provided these integrals exist.
THEOREM 4.7 The mean value theorem for integrals

If/ is continuous on the interval [a, b↑, there is at least one number c between a
and b such that
r⅛
∕(x ) dx = f(c)(b - a)
Ja
THEOREM 4.8 Second fundamental theorem of calculus

Let ∕(Z) be continuous on the interval [a, b] and define the function G by the
integral equation
G(x) = [ f(t) dt
Ja
for a ≤ x ≤ b. Then G is an antiderivative of/ on [a, 6]; that is,

G'(x) = C/X fa dt =
over [a, ⅛].

A -8 Appendices
THEOREM 4.9 Leibniz’s rule

If u(x) and v(x) are differentiable functions of x, then
⅛[Γ≡Λ]-Λ">⅛-∙Z'W ½
THEOREM 4.10 Error in the trapezoidal and Simpson’s rules
Iff has a continuous second derivative on [a, b], then the error E n in approximat
ing ∫*∕(x) dx by the trapezoidal rule is
Trapezoidal error: ∣
E j ∣≤ Where M is the maximum value o f ∣
∕" ( x ) ∣on [a, b]
Moreover, if/h a s a continuous fourth derivative on [a, b], then the error E n in
approximating f%f(x) dx by Simpson’s rule is
(b ~ a)5
Simpson’s error: K∖ ≤ 180w4 Where K is the maximum value o f ∣
∕ ,4 , (x)∣on [a, ⅛]
CHAPTER 5
THEOREM 5.1 Bracketing theorem for exponents
Suppose b is a real number greater than 1. Then for any real number x there is a
unique real number bx . Moreover, if p and q are any two rational numbers such
that p < x < q, then
bp < bx < bq
THEOREM 5.2 Properties of exponential functions

Let x, y be real numbers, and a and b positive real numbers.
Equality rule If x = y and bx > 0, then bx = by

Inequality rules If x > y and b > 1, then bx > by
If x > y and 0 1,
lim bx = 0 and lim bx = +∞
A'→ - ∞ A'→ + ∞
lim x lim b~x = 0

χ-→-∞ b~ = +∞ and χ → +∞
THEOREM 5.3 Existence of an inverse

Let f be a function that is continuous and strictly monotonic on an interval I.
Then f ~ 1 exists and is monotone on / (increasing if/is increasing and decreasing if
/ i s decreasing).
THEOREM 5.4 Continuity and differentiability of inverse functions

Let f be a one-to-one function so that it possesses an inverse.
1. If f is continuous on a domain I, then ∕ ^ 1 is continuous on its domain.
2. If f is differentiable at c, and ∕'( c ) ≠ 0, then / -1 is differentiable a t∕( c ).
THEOREM 5.5 Derivative of an inverse function

Suppose f is strictly monotonic and differentiable for all x in its domain. Then
(/ ) W - ∕ v - 1(*)]
THEOREM 5.6 Basic properties of logarithmic functions

If b > 0 and b ≠ 1, then
Equality rule If x = y , then logftx = ∣ogzj ,
Inequality rules If x > y and b > 1, then logft.r > logfry
If x > y and 0 < b < 1, then logi x < logi p
Product rule logi (xy) = logftx + logfry
Quotient rule io g ⅛(j) = lo g ⅛
x - l o g *j "
'
Power rule logy vp = P lθg⅛x for a ∏y real number p
THEOREM 5.7 Change-of-base theorem

log,*
⅛ " = ⅛ Z > > 0 , ⅛≠1
THEOREM 5.8 Basic properties of the natural logarithm

a. In 1 = 0
b. In e = 1
c. elπ v = x for all x > 0
d. In ey = y for all y
e. logftx = for any b > 0 (⅛≠ 1)
f. bx = ex ln b for any b > 0 (b ≠ 1)
THEOREM 5.9 Derivative of In u

If x > 0, then ^ ( l n x) = ^∙
If u is a differentiable function of x , then ^^(1∏ w) = ' for u > 0.
THEOREM 5.10 Derivative of ln ∣

u∣
I f f( x ) = In ∣x∣, x ≠ 0, then∕'( x ) =
Also, if u is a differentiable function of x , then In ∣M∣ = /
THEOREM 5.11 Derivative of e u
T x^ =
If u is a differential function of x , then = eu

A -10 Appendices
THEOREM 5.12 Derivative of an exponential with base b

For base b (b > 0, b ≠ 1),
^ ⅛ - = (ln b)b×
and if u is a differentiable function of x , then
b u - (In b)b" for all u
THEOREM 5.13 Derivative of a natural logarithm with base b

For base b (b > 0, b ≠ 1)
4 - (log,x) = 7-,— U — for all x > 0

a x v 6 * ’ (In b) x
and u is a differentiable function of x , then
4 - (log, u) = 77—U — Φ for all u > 0

dx 6 * ’ (In b) u ax
THEOREM 5.14 Limits involving natural logarithms and exponentials

If k and n are positive integers, then
lim = -∞ lim = 0
x→o+ x n x→+∞ x n
lim = +∞ lim x ne~kx = 0
THEOREM 5.15 Integration of ex

ex dx = ex + C
THEOREM 5.16 Integration of 1/x
THEOREM 5.17 Integral of tangent and cotangent

tan x dx = -ln∣cos x∣ + C and cot x dx = ln∣sin x∣ + C
THEOREM 5.18 Differentiation formulas for six inverse trigonometric functions

If u is a differentiable function of x , then
⅛ sιn' l≈ '^√ τ⅛ ⅛ ⅛*cos^'"1 '√ f ⅛ ⅛

⅛<,an""> "τ⅛ ⅛ ⅛(∞t-'≈o-τ ⅛ ⅛
≡ ,sec^l ", ^M t⅛ T⅛
THEOREM 5.19 Integrals involving the inverse trigonometric functions
dx = sin 'x for ∣x∣ < 1
Vl -x 2
f dx = tan l . for all x

1 -I- -⅛-2
dx = sec 'x for ∣x∣ > 1

xVx2 - 1
THEOREM 5.20 Properties of the logarithm as defined by In x = J y
Let x > 0 and y > 0 be positive numbers. Then
a. In 1 = 0
b. In xy = In x + In y
c. In - = In x }- In v
y
d. In x p = p In x for all rational numbers p
THEOREM 5.21 Properties of the exponential as defined by E(x)

For any numbers x and y,
a. E(x + y) = E(x)E(y)
b. E(x - y ) = E ( y j
c. [∕f(x)p = E(px)
THEOREM 5.22 Derivative of E(x)

The function defined by E(x) is differentiable and
f =^ )
CHAPTER 6
THEOREM 6.1 Volume theorem of Pappus
The solid generated by revolving a region R about a line outside its boundary (but
in the same plane) has volume V = As, where A is the area of R and 5 is the
distance traveled by the centroid of R.
THEOREM 6.2 General solution of a first-order linear differential equation

The general solution of the first-order linear differential equation
+ yP(x) = Q(x)
is given by
y = 7 ⅛ j QW dx+ c
where Z(x) is the integrating factor ∕(x) = elpfχ Pχ .
CHAPTER 7
THEOREM 7.1 Properties of the hyperbolic functions
cosh2x - sinh2 x = 1
sinh(-x) = -sinh x (sinh is odd)
cosh(-x) = cosh x (cosh iseven)
tanh(-x) = -tan h x (tanh is odd)
sinh(x + y) = sinh x cosh y + cosh x sinh y
cosh(x + y) = cosh x cosh y + sinh x sinh y
A -1 2 Appendices
THEOREM 7.2 Rules for differentiating hyperbolic functions

Let u be a differentiable function of x . Then
⅛(sinhM) = c o s h a g ⅛ (c°sh “ ) = s in h M S
⅛(tanh u) = sech⅛ g ⅛ (∞ th u) = -csch⅛ g
g (s e c h u) = -sech u tanh u g ( c s c h w) = -csch u coth u
THEOREM 7.3 Integration formulas involving the hyperbolic functions

J sinh x dx = cosh x + C j^ cosh x dx = sinh x + C
f sech2x dx = tanh x + C J csch2x dx = -co th x + C
sech x tanh x dx = -sech x + C csch x coth x dx = -csch x + C
THEOREM 7.4 Logarithmic formulas for inverse hyperbolic functions

+ v i ∣^l x ≠ 0
sinh 1x = ln(x + V x 2 + 1), all x csch 1x = Inf—
∣χ∣ /
cosh 'x = ln(x + V x 2 - 1), x ≥ 1 sech~1x = ln f l + V j - χ - ∖ 0 < x ≤ 1
tanh 1x = 2 In γ-3- 1 coth l x = X In —+ √> ∣x∣ > 1

2 x - 1
THEOREM 7.5 Differentiation and integration formulas for the inverse hyperbolic functions
4- (sinh-1 a) = - , g du = sinh 1a + C
V l + u 1 dx V l + a2
⅛ (∞ s h " u) = √ ⅛ ⅛ , ifw > 1

, = cosh 1u + C
J V u 29 - 1
⅛ -(tanh-'20 = τ ⅛ ⅛ if∣W∣ < 1 f du _ ∣∣

ta n 1- w + (j if ∣u ∣ 1
CHAPTER 8
THEOREM 8.1 Limit theorem for sequences

If lim a = L and lim b n = M , then the following limits exist and:
Linearity rule lim (ra + sbn ) = rL + sM

n→∞ n11
Product rule lim (a b ) = L M
Quotient rule Um⅛ = ⅛ provided M ≠ 0

n →∞bnM
Root rule lιm V α '', = ∖ L
n→ ∞
provided VrΓ is defined
for all n and v L exists
T H E O R E M 8.2 Limit of a sequence

Suppose/is a function such that an = f(n) for n = 1, 2, . . . . If lim ∕( x ) exists and
lim ∕( x ) = L , the sequence {a } converges and lim a = L . ""^
T H E O R E M 8.3 Squeeze theorem for sequences

If α , ≤ b ≤ c for all n > N , and lim a = lim c = L , then lim b = L
T H EO R EM 8.4 BM CT; the bounded, monotonic, convergence theorem

A monotone sequence {an} converges if it is bounded and diverges otherwise.
T H E O R E M 8.5 Convergence of a power sequence

If r is a number such that r < 1, then lim rn = 0.
∕7→∞
T H E O R E M 8.6 Linearity of infinite series

If ⅛ak and ~∑
tbk are convergent series, then so is ⅛(cak + dbk ) for constants c, d, and
X(c¾ + db^ = c ^∑
tak + d Xb k
T H E O R E M 8.7 Divergence of the sum of a convergent and a divergent series

If either X ak or X bk diverges and the other converges, then the series X (ak + bk)
must diverge.
T H E O R E M 8.8 Geometric series theorem

∞
The geometric series / ark with a ≠ 0 diverges if ∣r∣ ≥ 1 and converges if ∣r∣ < 1
with sum fc=0
Σ ark = . a
k=o 1- r
T H E O R E M 8.9 The divergence test

If the series t a k converges, then the sequence {απ} converges and lim ak = 0. Thus,
if lim a,κ ≠ 0, then the series X a,,κ must diverge.
∕7→∞
T H E O R E M 8.10 Convergence criterion for nonnegative terms

A series X ak with ak ≥ 0 for all k converges if its sequence of partial sums is
bounded from above and diverges otherwise.
T H E O R E M 8.11 The integral test

If ak = ∕(Zc) for k = 1, 2, . . . where/is positive, continuous, and decreasing for
x ≥ 1, then
∞ r∞
∑ a
j ι k and 1 ∕( x ) dx
either both converge or both diverge.

A -1 4 Appendices
THEOREM 8.12 The /^-series test

∞ 1
The p-series ∑ converges if p > 1 and diverges if p ≤ 1.
THEOREM 8.13 Direct comparison test

Let 0 ≤ a k ≤ ck for all k ≥ N for some N.
∞ ∞
If ∑ ck converges, then ∑ a k also converges.
k=1 k=1
Let 0 ≤ d k ≤ ak for all k.
∞ ∞
If ∑ dk diverges, then ∑ ck also diverges.
k=1 k=1
THEOREM 8.14 Limit comparison test

Suppose ak > 0 and bk > 0 for all sufficiently large k and that
. ak
lιm -r- = L
k→∞ bk
where L is finite and positive (0 < L < ∞). Then ⅛ak and ~Zbk either both
converge or both diverge.
THEOREM 8.15 The zero-infinity limit comparison test

Suppose ak > 0 and bk > 0 for all sufficiently large k. Then
ak
If lim -r- = 0 and X⅛,
κ
converges, the series Xα,
κ
converges.
k→∞ Ofc
ak
If lim τ - = ∞ and X⅛,
κ
diverges, the series Xα.
κ
diverges.
k→∞ "k
THEOREM 8.16 The ratio test
<L.+ 1
Given the series Xαfc with a k > 0 and suppose that lim =L
The ratio test states the following:

If L < 1, then Xαj, converges.
If L > 1 or if L is infinite, then X¾ diverges.
If L = 1, the ratio test is inconclusive.
THEOREM 8.17 The root test

k∕—
Given the series ⅛ak with a k ≥ 0 and suppose that lim v a k = L.
The root test states the following:
If L < 1, then Xαi , converges.
If L > 1 or if L is infinite, then X¾ diverges.
If L = 1, the ratio test is inconclusive.
THEOREM 8.18 Alternating series test

If ak > 0, then the alternating series
∑ι ( - l ) ⅛ and ∑ ι ( - l ) fc+1¾
converge if both of the following conditions are satisfied:
1. lim a, = 0.
Zc→∞ *
2. {ajJ is a decreasing sequence; that is, αfc+l < α t for all k.
THEOREM 8.19 The error estimate for an alternating series

Suppose the alternating series
∞ ∞
Σ ( - l ) ⅛ κ and ∑ (~ l) t + 1 ¾lc
⅛=ι k=l
satisfy the conditions of the alternating series test, namely,
lim ak = 0 and {ak} is a decreasing sequence (ak + i < ak )
If the series has sum S, then
lχ - ¾l < ⅛÷1
where S n is the nth partial sum of the series.
THEOREM 8.20 The absolute convergence test

A series of real numbers ~Zak must converge if the related absolute value series ∑∣αfc∣
converges.
THEOREM 8.21 The generalized ratio test

Suppose ak + 0 for k ≥ 1 and that
V∞ /7 + l
⅛
where L is a real number or ∞. Then

If L < 1, the series ∑afc converges absolutely and hence converges.
If L > 1 or if L is infinite the series X¾ diverges.
THEOREM 8.22 Convergence of a power series

∞
For a power series ∑ ak x k , exactly one of the following is true:
k= 1
1. The series converges for all x.

2. The series converges only for x = 0.
3. The series converges absolutely for all x in an open interval ( - R , R) and
diverges for ∣x∣ > R.
The series should be checked separately at the endpoints, because it could

converge absolutely, or converge conditionally, or diverge at x = R and x = - R .
THEOREM 8.23 The convergence set of a power series

Let ⅛ak uk be a power series, and let
a k+∖
L = lim
k→ ∞ a k
Then If L = ∞, the power series converges only at u = 0.

IfO < L < ∞, let 7? = 1∕L. Then the power series converges absolutely for
∣M∣ < R (or - R R . It may either con
verge or diverge at the endpoints u = - R and u = R.
A -16 Appendices
T H E O R E M 8.24 Root test for the radius of convergence

Let ~Zak ιιk be a power series, and let
L = lim k ∖a. ∣
∕<→∞ ^
Then If L = ∞, the power series converges only at 0.
If 0 < L < ∞, let R = 1 ∕L. Then the power series converges absolutely for
∣M∣ < R and diverges for ∣w∣ > R. It may either converge or diverge at the
endpoints u = - R and u = R.
T H EO R EM 8.25 Term-by-term differentiation and integration of a power series

∞
A power series Y a k u k with radius of convergence R > 0 can be differentiated or
⅛=o
integrated term by term on its interval of absolute convergence. More specifically,
00
i f f(μ ) = Y a,u k for ∣M∣ < R , then for ∣M∣ < R we have
⅛=o
∞
f'(u) = Σ kaκ,u1 k ~ x2 = a, + 2aj 1 u + 3α,w2 + 4Ω, M3 +
k=∖
∕∙x rx / ∞ ∖ ∞ / rx ∖ ∞ Cl,
f(u) du = ∣ y ak u k ∖du = Y I a.u k du] = ∑ > , 1 uk + l
Jo ' Jo ∖fc=o / ⅛=o ∖ Jo ) k≈0 κ + l
T H E O R E M 8.26 The uniqueness theorem for power series representation

Suppose a function/is known to have the power series representation
∞
∕(-r) = ∑θ ak(x - c)k
for - R < x - c < R. Then there is exactly one such representation, and the
coefficients ak must satisfy
__ ∕W (c)
a k ~ k∖
for k = 0, 1, 2, . . . .
T H EO R EM 8.27 Taylor’s theorem

I f/ and all its derivatives exist in an open interval I containing c, then for each x
in I
∕W = ∕( c ) + ^ ↑↑ ∖ x c) + ^ ^ ∖ x c)2 + ∙∙∙ + ∖x c ) + R n(x)
where the remainder function R t[x) is given by

f("+ D(z )
⅛ ) = (« + l ) ! ^ c )" +1
for some zn that depends on x and lies between c and x.
T H E O R E M 8.28 Binomial series theorem

The binomial function (1 + x) p is represented by its Maclaurin series
d + x γ . 1+ px + r ¾ ⅛ + f ⅛ - i ) ⅛ - ¾ 3 + ... _ f ρ y
2! 3! k= 0 ∖
k q]
∕n∖ n∖
where / 1 = for all x if p is a nonnegative integer; for 1 < x < 1 if
p ≤ -1 ; - l ≤ x ≤ l i f ∕7 > 0 and p is not an integer; and for -1 < x ≤ 1 if
-1 < p < 0.
CHAPTER 9
THEOREM 9.1 Area in polar coordinates
Let r = f(θ) define apolar curve, where/is continuous and/(0) ≥ 0 on the closed
interval a ≤ θ ≤ β. Then the region bounded by the curve r = /(0) and the rays
0 = a and 0 = /3 has area
1 rP 1 fβ
A =~ r2 dx = ½ [∕(*)] 2 dx
Ja Ja
THEOREM 9.2 Slope in polar coordinates

If /(0) is a differentiable function of 0, then the slope of the tangent line to the
polar curve r = /(0) at the point P(rθ, 0o) is given by
/ ( 0 ) cos 0 + ∕ '( 0 ) sin 0
/77 = ---- --------- --------- o------- —
- / ( θ0) sin 0o + / (0o) cos 0o
whenever the denominator is not zero.
THEOREM 9.3 Parametric form of the derivative

If a curve is described parametrically by the equations x = x(t), y = y(f), where x
and y are differentiable functions of t, then the derivative can be expressed in
terms of and by the equation
dy
dv g~dt y'(t) dy , dx ,
-±r = r = ⅛ h = 777 whenever ¾ 7≠ 0
d x d x x (t) dt dt
dt
THEOREM 9.4 Arc length for parametric equations

Let Cbe the curve described parametrically by the equations x = x(t) andy = y(f)
on an interval [a, b∖, where and are continuous, such that C does not
intersect itself on the interval. Then C has arc length
H √ ( ⅞ H ⅛ P
THEOREM 9.5 Formula for the area under a curve defined parametrically
Suppose y is a monotone continuous function of x on the closed interval [a, b] and
that x and y are both functions of a parameter t, for t l ≤ t ≤ / . Then if x'(t) is
continuous on [∕ 1, t2], the area under the curve described parametrically by
x = x(t), y = y(t) over [ ∕ l , t2], is given by
r⅛ rt 2
A= y dx = /( /) x (/) dt
Ja Jt,
THEOREM 9.6 Reflection property of parabolas

Let P be a point on a parabola in the plane, and let T be the tangent line to the
parabola at P. Then the angle between T and the line through P parallel to the
principal axis of the parabola equals the angle between T and the line connecting P
to the focus.
A-18 Appendices
CHAPTER 10
T H E O R E M 10.1 Properties vector operations
For any vectors u, v, and w in the plane and scalars s and t:
Commutativity of vector addition u + v= v+ u

Associativity of vector addition (u + v) + w = u + (v + w)
Associativity of scalar multiplication (st)u = s(Zu)
Identity for addition u+ 0= u
Inverse property for addition u + (-u ) = 0
Vector distributivity (s + Z)u = .su + Zu
Scalar distributivity s(u + v) = su + sv
T H E O R E M 10.2 Properties of the dot product

If u, v, and w are vectors (in R 2 or R 3) and c is a scalar, then
Magnitude of a vector v ∙ v = ∣v 2
Zero product 0 ■v = 0
Commutativity v ∙ w = w∙ v
Associativity c(v ∙ w) = (cv) ∙ w = u ■(cw)
Distributivity u ∙( v + w) = u ∙ v + u ∙w
T H E O R E M 10.3 Angle between two vectors

If θ is the angle between the nonzero vectors v and w, then
PN S A_ v , w
M H
T H E O R E M 10.4 Orthogonal vectors

Nonzero vectors v and w are orthogonal if and only if
v∙w= 0
T H E O R E M 10.5 Properties of the cross product

If u, v, and w are vectors in R 3 and .v and t are scalars, then
Scalar distributivity (sv) × (Zw) = st(y × w)

Vector distributivity u × (v + w) = (u × v) + (u × w)
(u + v) × w = (u × w) + (v × w)
Anticommutativity v × w = —(w × v)
Parallel vectors v × v = 0; in particular, if w = sv then v × w = 0
Zero product v×0 = 0×v = 0
T H E O R E M 10.6 Geometric property of cross product

If v and w are nonzero vectors in R 3, that are not multiples of one another, then
v × w is orthogonal to both v and w.
THEOREM 10.7 Magnitude of a cross product

For v and w nonzero vectors in R 3 with θ the angle between v and w (0 ≤ θ ≤ ττ),
then
∣
∣
v × w ∣
∣= ∣
∣∣∣
v∣∣ ∣sin θ
w∣
THEOREM 10.8 Determinant form for a triple scalar product: volume of a parallelepiped
If u = α,i + tz,i + <2,k, v = b,i + bΛ + 6,k, and w = c.i + c,j + = c.k, then the
triple scalar product can be found by evaluating the determinant
Cl1 ^”2
(u × v) ∙ w = b χ b 2 b3
c ι cι C3
THEOREM 10.9 Existence of a plane

Let A , B, C , and D be constants with A , B, and C not all zero. Then the graph of
any equation with form
A x + By + C z + D = 0
is the equation of a plane with normal vector N = ∕l i + Bj + Ck.
THEOREM 10.10 Distance from a point to a plane in R 3

The distance from the point (x0, y 0, z0 ) to the plane A x + By + Cz + D = 0 is given
by
_ A x 0 + By 0 + Cz 0 + D
∖ rA 2 + B 2 + C^
THEOREM 10.11 Distance from a point to a line

The shortest distance from the point P to the line L is given by the formula
√ -∣
∣
v ×Q p ∣
∣
l∣
v∣
l
where v is a vector aligned with L and Q is any point on L .
CHAPTER 11
THEOREM 11.1 Rules for vector limits
If the vector functions F and G are functions of a real variable t and h(t) is a scalar
function such that all have finite limits as t → t0 , then
Limit of a sum lim [F(∕) + G(t)] = lim F(Z) + tlimt G(Z)

Z→Z0 ~* o
Limit of a difference lim [F(Z) —G(Z)] = lim F(∕) - lim G(Z)
f→√0 t ~*t o t ~*t o
Limit of a scalar multiple lim [∕z(7)F(t)] = [lim A(Z)Î [lim F(Z)1
Limit of a dot product lim [F(t) ∙ G(Z)] = [lim F(Z)] ∙ [lim G(Z)[∣
Limit of a cross product lim [F(Z) × G(Z)] = [lim F(t)l × [lim G(Z)1
These limit formulas are also valid as t → +∞ or as t → -∞ , assuming all limits

exist.
A -2 0 Appendices
T H E O R E M 11.2 Derivative of a vector function

The vector function F(7) = ∕ l (Z)i + ∕2(t)j + f 3(t)k is differentiable whenever the
component functions f v f 2, and ∕ 3 are all differentiable, and in this case
F'(0 = f ' l (t)i + ∕'(t)j + Λ ( 0 k
T H E O R E M 11.3 Rules for differentiating vector functions

If the vector functions F and G and the scalar function h are differentiable at t,
then so are αF + b G , hF, F ∙ G , and F × G , and
Linearity rule (αF + bG)'(f) = aF'(t) + bG'(f) for constants a, b

Scalar rule (hF)'(t) = h'(t)F(t) + h(f)F'(f)
Dot product rule (F ∙ G)'(Z) = (F' ∙ G)(Z) + (F ∙ G')(Z)
Cross product rule (F × G)'(Z) = (F' × G)(Z) + (F × G , )(Z)
Chain rule [F(Zz(Z))]' = Zz'(Z)F'(A(Z))
T H E O R E M 11.4 Orthogonality of a derivative

If the nonzero vector function F(Z) is differentiable and has constant length, then
F(Z) is orthogonal to the derivative vector F'(Z).
T H E O R E M 11.5 Kepler’s second law

The radius vector from the sun to a planet in its orbit sweeps over equal areas in
T H E O R E M 11.6 Arc length of a space curve

If C is a smooth curve defined by R(t) = -x(Z)i + J , (z)j + z(t)k, on an interval [a, Z>],
then the arc length of C is given by
∙5, = l∣R'(0∣l dt = y∣
[x'(t)]2 + [ ∕( t ) ] 2 + [√(Z)]2 dt
Ja Ja
T H E O R E M 11.7 Speed in terms of arc length

Suppose an object moves with displacement R(t), where R'(Z) is continuous. Then
the object has speed
l∣V(0ll = l∣R'(0ll = ⅛ for a ≤ t ≤ b
T H E O R E M 11.8 Tangential and normal components of acceleration

An object moving along a smooth curve with T ' ≠ 0 has velocity V and
acceleration A , where
v = (⅛ )τ ≈>"d a = (S ) τ + « ($ )"
and where 5 is the arc length along the trajectory.
T H E O R E M 11.9 Acceleration of an object with constant speed

The acceleration of an object moving with constant speed is always orthogonal to
the direction of motion.
CHAPTER 12
THEOREM 12.1 Equality of mixed partials
If f( x , y) has mixed partial derivatives at (x0, y 0) and the mixed partials are
continuous in an open set containing (x0, y 0), then
∕√ ⅞ > To) = ∕√ ⅞ To )

If f ( x , y) is differentiable at (x0, y 0), it is also continuous there.
THEOREM 12.3 Sufficient condition for differentiability

If/is a function of x and y where//., and / are continuous in a disk D centered at
(x0, y 0), then/is differentiable at (x0 , y 0). ’
THEOREM 12.4 The chain rule for one independent parameter

Let f ( x , y) be a differentiable function of x and y , and let x = x(t) and y = y(t) be
differentiable functions of t. Then z = ∕( x , y) is a differentiable function of t, and
dz _ ∂z dx , ∂z dy
dt ∂x dt ^t^ ∂y dt
THEOREM 12.5 The chain rule for two independent variables

Suppose z = f ( x , y) is differentiable at (x, y) and that the partial derivatives
of x = x(ιι, v) and y = y(μ, v) exist at (u, v). Then the composite function
z = f[x ( u , v), y(μ, v)] is differentiable at (u, v) with
d Z _ _ d Z _ d X .d Z .ty J dz _ dz ∂X , ∂Z ty
du ∂x ∂u ^r ∂y ∂u ∂v ∂x ∂v ∂y ∂v
THEOREM 12.6 Directional derivatives using partials

A function ∕( x , y) that is differentiable at ∕ l(x0 , y 0) â s a directional derivative at
P o in every direction, and if u = u 1i + w2j, is a unit vector, the directional
derivative in the direction given by
D uf(X 0’ To) = 4(⅞> T0)w1 + ∕ y (⅞, T0)M 2
THEOREM 12.7 The gradient formula for the directional derivative

If/ is a differential function of x and y, then the directional derivative at the point
P 0(x0, y 0) in the direction of the unit vector u is
D uf{ x , y') = N f-u
THEOREM 12.8 Basic properties of the gradient

Let / and g be differentiable functions; then
Constant rule Vc = 0 for any constant c

Linearity rule V ( α ∕+ bg~) = a⅛f+ bVg for constant a and b
Product rule V(J⅛) = ∕(V g ) + g V ∕
Quotient rule ,≠ 0
∖6 / g
Power rule V (/”) = n f n ~vV f
A-22 Appendices
THEOREM 12.9 Optimal direction property of the gradient

Suppose/is differentiable and let V∕o denote the gradient at PQ. Then, if V∕o ≠ 0,
a. The largest value of the directional derivative of Duf i s ∣

∣ ∣
V∕θ∣and it occurs
when the unit vector u points in the direction of V∕θ.
b. The smallest value of Duf is - ∣ ∣V∕0∣
∣
, and it occurs when u points in the
direction of -V ∕ 0 .
THEOREM 12.10 The normal property of the gradient

Suppose the function/is differentiable at the point Po and that the gradient at Po
satisfies V∕o ≠ 0. Then V∕o is orthogonal to the level surface f(x , y, z) = K at Pθ.
THEOREM 12.11 Partial derivative criteria for relative extrema

If / has a relative extremum (maximum or minimum) at P0(x0 , y0) and the
first-order partial derivatives of f χ and f y both exist at P0(x0, τ 0)> then
∕√ ⅛ To) = f y (×0, To) = θ∙
THEOREM 12.12 Extreme value theorem for a function of two variables

A function of two variables f(x , y) assumes its absolute extremum on any closed,
bounded set A in the plane where it is continuous. Moreover, all absolute extrema
must occur either on the boundary of S or at a critical point in the interior of S’.
THEOREM 12.13 Lagrange’s theorem

Assume that / and g have continuous first partial derivatives such that / has an
extremum at P0(x0, To) on the smooth constraint curve g(x, τ) = c. If
V<g(x0, To) ≠ θ, then there is a number λ such that
V∕(x 0, To) = ΛVg(x0, τ 0)
THEOREM 12.14 Rate of change of the extreme value

Suppose E is the extreme value (maximum or minimum) of / subject to the
constraint g(x, y) = c. Then the Lagrange multiplier λ is the rate of change of E
with respect to c; that is, λ = dE∕dc.
CHAPTER 13
THEOREM 13.1 Fubini’s theorem over a rectangular region
If ∕(x , τ) is continuous over the rectangle R∙. a ≤ x ≤ b, c ≤ y ≤ d, then the
double integral
f f f(χ , y)dA
is evaluated by either iterated integral; that is,

b rb rd
∏' f(x, y) dx dy =
Ja Jc
f(x, τ) dy dx
THEOREM 13.2 Fubini’s theorem for nonrectangular regions

If D is a type I region, then
[ [ f(x , y)d A = f f k ( } f(x, y) dy dx

J1p J“ JglM
whenever both integrals exist. Similarly, for a type II region,

∏ r d r∕ι 2(y)
f(x , y) dA = f(x , y) dx dy
Jc Jh 1(y)
THEOREM 13.3 Properties of double integrals

Assume that all the given integrals exist.
Linearity rule For constants a and b

[ [ [af(×, y) + bg(x, y)]dA = a f ∣f(x, y) dA + b f f g(x, y) dA
JDJ JDJ JDJ
Dominance rule If f(x , y) ≥ g(x, y) throughout a region D, then
D D
Subdivision rule If the region of integration D can be subdivided into two

subregions D l and D1, then
D D;
THEOREM 13.4 Double integral in polar coordinates

If f is continuous in the polar region described by a ≤ r ≤ b (a, b ≥ 0),
a ≤ 0 ≤ jβ ( O ≤ ∕3 - a < 2ττ), then
rβ rb
f(r cos θ, r sin θ) r dr dθ
'α Ja
THEOREM 13.5 Fubini’s theorem over a parallelepiped in space

Iff(x , y, z) is continuous over a box R ∖a ≤ x ≤ b, c ≤ y ≤ d, r ≤ z ≤ s, then
the triple integral is evaluated by the iterated integral
∣f f f ( χ , y, z) d v= f(χ , y, z) dx dy dz
R
The iterated integral can be performed in any order:
dx dy dz dx dz dy dz dx dy
dy dx dz dy dz dx dz dy dx
THEOREM 13.6 Triple integral over a general region

If A is a region in space that is bounded above by the surface z = u(x, y) and below
by z = v(x, y) as (x, y) varies over the planar region R, then
∫ ∫ ∫ ∕< w ) Λ ' - ∫ ∫ ∫ * ' f(x , y, z) dz dA
s r
THEOREM 13.7 Change of variable in a double integral

Let f be a continuous function on a region D, and let T be a one-to-one
transformation that maps the region D in the uv-plane onto a region D* in the xy-
A -2 4 Appendices
plane under the change of variable x —g(u, v), y = h(u, v) where g and h are
continuously differentiable on D*. Then
f ( x , y) dy dx = f[g (u , v), h(u, v)] ∖J(μ, v)∣ dv du

∙
JD J J DJ
∂x ∂x
where ./v(«, V) = v = " ≠ θ
’ ’ ∂y ∂y7 ∂u ∂v ∂u ∂v
∂u ∂v
is a mapping factor called the Jacobian for the mapping from x , y variables to u, v
variables. The Jacobian is also denoted by S u .
CHAPTER 14
T H E O R E M 14.1 Properties of line integrals
Let f be a given scalar function defined with respect to JC on a piecewise smooth,
orientable curve C . Then for any constant k:
Constant multiple k f dx = k f dx
Jc ’ Jc
Sum rule (7, + f 2 ) d x = ∖ f i d x + ∖ ∕2 dx

Jc Jc Jc
where f l and f 2 are scalar functions defined with respect to x on C.
Opposite direction rule f dx = - ∖ f dx

J-c Jc'
where - C denotes the curve C traversed in the opposite direction.
Subdivision rule ∖ f dx = f dx + f dx
Jc Jc, Jc 2
where C is subdivided into subarcs C 1 U C 2 (where C l ∩ C 7 = 0 ).
This property generalizes to any finite number of subdivisions.
Similar properties hold for line integrals of the form g dy or h dz.

Jc ’ Jc
T H E O R E M 14.2 Cross-partials test for a conservative vector field in the plane

Consider the vector field F(x, y) = u(x, jji + v(x, y)j, where u and v have
continuous first partials in the open, simply connected region D. Then F(JC, y) is
conservative on D if and only if
= g Throughout D
T H EO R EM 14.3 Fundamental theorem on line integrals

Let F be a conservative vector field on the region D and let f be a scalar potential
function for F. That is, V ∕ = F. Then, if C is any piecewise smooth curve lying
entirely within D , with initial point P and terminal point Q, we have
f F ∙ d R = ∕( β ) - ∕( P )
Jc
the line integral ∫c F ∙ c/R is independent of path in D.
App. A: Theorems by Chapter A -2 5
THEOREM 14.4 Closed-curve theorem for a conservative vector field

The continuous vector field F is conservative in the open connected region D if
and only if f c F ■dR = 0 for every piecewise-smooth closed curve C in D.
THEOREM 14.5 Green’s theorem

Let D be a simply connected region with a positively oriented, piecewise-smooth
boundary C. Then if the vector field F(x, ρ) = M(x, p)i + N(x, y)j is continuously
differentiable on D, we have
∫ c (M A + Λ '⅛ ∙ ) ≈ ∫ J ( g - ⅛ ) ^
THEOREM 14.6 Area as a line integral

Let D be a simply connected region in the plane with piecewise-smooth, closed
boundary C. Then the area A of region D is given by the line integral
A = 1 1 (-y dx + x dy)
z Jc
THEOREM 14.7 Stokes’s theorem

Let 5 be an oriented surface with unit normal vector N, and assume that A is
with that of A. If F is a vector field that is continuously differentiable on S, then
[ F ∙ dR = ( [ (curl F ∙ N) dS
Jc Js J
THEOREM 14.8 Test for a vector to be conservative

If F and curl F are continuous in the simply connected region D, then F is
conservative in D if and only if curl F = 0 in D.
THEOREM 14.9 The divergence theorem

Let D be a region in space bounded by a smooth, orientable, closed surface 5. If F
is a continuous vector field whose components have continuous partial derivatives
in D, then
∫ j F ∙ N dS = ∫ ∫ ∫ div F dV
s D
where N is an outward unit normal to the surface A.
APPENDIX B Selected Proofs

■ CHAIN RULE (Chapter 2)
Suppose/is a differentiable function of u and u is a differentiable function of x.
Then,
d f_ d fd u
dx du dx
where ∕[u(x)].
A -2 6 Appendices
Proof: Define an auxiliary function g defined by
t ,,,J ⅛ )÷ l ] - ∕M . ⅜
ι- Cl Lt
lf z + 0 and j,ω . 0 if ,= 0
You can verify that g is continuous at t = 0. Notice that for t = ∆u and t ≠ 0,
f[u(x) + ∆z∕] - ∕ [ M(X)] _ d f
g(∆w) = ∆u du
< ∕M *) + A »] ~ ∕W ) ]
' + du = ∆u
g{û) + ⅞ ] δ m = ∕ W ) + M ^ ∕IX∙*)]
We now use the definition of derivative for f

d f = } - f[u (x + ∆x)] - ∕ [ M(X)]
dx ΔΛ→O ∆x
_ 1∙ ∕ [ M(∙X) + ΔM)] - f[u(x)]
Where ∆ιι = u(x + ∆,v) - u(x)
∆Λ∙→0 ∆x
G ^ + ^ J ΔM
lim Substitution
∆x
= 2≡0 [^ ΔM>+ S Δ^
= lim [g(∆zz) + η r lim —■

∆Λ-→
OL du] Δ,V→
O∆x
= lim g(∆ιι) + lim / lim
L∆.t →0 ' ∆Λ∙→0 du] Δ.r→0 ∆x
Because g is continuous at t = 0
d f du
du dx
■ CAUCHY'S GENERALIZED MEAN VALUE THEOREM (Chapter 3)
Let f and g be functions that are continuous on the closed interval [a, ⅛] and
differentiable on the open interval (a, b). If g(f) ≠ g(d) and g'(x) ≠ 0 on (a, b),
th e n ∕( ⅛ ) ~ Λ 0 = £ £ )
g(,b) - g(a) g'(c )
for at least one number c between a and b.
Proof: We begin by defining a special function, just as in the proof of the MVT
as presented in Section 3.2. Specifically, let
F(x) = ∕(x ) - ∕(α ) - [g(x) - g(α)]
for all x in the closed interval [a, ⅛]. In the proof of the MVT in Section 3.2, we
show that F satisfies the hypotheses of Rolle’s theorem, which means F'(c) = 0 for
at least one number c in (a, /?). For this number c we have
0 = F'(c) = ∕'( c ) ∕( ^ ) ~ f(a)

g(b) - g(a) g'(c)
and the result follows from this equation.
App. B: Selected Proofs A -27
■ L'HOPITAL’S THEOREM* (Chapter 3)
For any number a let f and g be functions that are differentiable on a closed
interval [a, b], where g'(x) ≠ 0. Then if
lim fix ) = 0, lim+ g(x) = 0 and lim ^ 7 exists, then

x →α+ v
x→a ° x →a g ( χ )
Proof: First, define auxiliary functions F and G by

F(x) = ∕(x ) for a < x ≤ b and F(α) = 0
G{x) = g(x) for a < x ≤ b and G(a) = 0
These definitions guarantee that Fix) = f(x) and G(x) = g(x) for a < x ≤ b and
that F(a) = G(a) = 0. Thus, if w is any number between a and b, the functions F
and G are continuous on the closed interval [a, ¼,] and differentiable on the open
interval (a, w). According to the Cauchy generalized mean value theorem, there
exists a number t between a and w for which
F(w) - F(a) F'(f)
G(w) - G(a) G'(t)
F(w) _ F'(f)
Because F(a) = G(a) = 0
G(w) ^ G'(t)
∕ ( H,) _ ∕'( t )
g (w ) '( t )
Because F(x) = f(x) and G(x) = g(x)
g
l. ∕W ,. ∕'( t )
w→a + g (W ) w →a
i g (t)
,. / ( " ') r ∕ '( 0

hm —r x = hm —— Because t is “trapped' between a and w
w→α+ g(H’) t →a* g (f)
rlim ∕W = r lim ∕'( χ )

z x
x→α÷ g(x) x →a + g (x) ■
■ CONTINUITY AND DIFFERENTIABILITY OF INVERSE

FUNCTIONS (Chapter 5)
L e t/b e a one-to-one function so that it possesses an inverse.
1. I f /is continuous on a domain I, then / -1 is continuous on ∕( ∕) .
2. I f /is differentiable at c, and ∕'( c ) ≠ 0, then / -1 is differentiable at ∕(c).
Proof: 1. Recall y = fix) if and only if x = ∕ - 1 (j). Let / be the open interval
(a, b), and let y0 be in the open interval (√~(α), ∕(⅛)) so that x0 is in the
open interval (a, b). To prove continuity we must show that
lim ∕ ^ 1(T) = ∕^^ 1 (LO) = ∙⅞υ
y →y 0
*This is a special case of ΓHδpitaΓs rule. The other cases can be found in most advanced calculus
textbooks.
A -2 8 Appendices
Consider any interval (x0 - ε, X Q + ε) for ε > 0. We must find an

interval (y0 - δ, y 0 + δ) such that whenever y is in (y0 - δ, y 0 + δ),
/ -1 (y) is in (x0 - ε, X Q + ε).
Let δ1 = y 0 - f( x 0 - ε) and δ2 = ∕( x 0 + ε) - y 0 . If δ is the small
er of δ1 and δ2, it follows that if y is in (y0 - δ, y 0 + δ), then f ~ l (y) is
in (xθ - ε, x 0 + ε), which is what we wanted to prove.
2. Let g = f 1 and f ' [g(α)] ≠ 0. We want to show that g is differentiable.

/(«) - ∣im ≡ ∕ ∕ From the definition o f derivative
y -b
~ ⅛ (τ )-> ) I f y = g(x) then x = fiy ), and
ifb = g(a) then a = fib).
,. y - b
Because f is differentiable, it is continuous,
~ y →b f i y ) ~ f i b )
so g = f~ i is
continuous by part (1). Thus, i f x → a,
then g(x) → g(a) so that y → b.
= lιm
y→bf ( y ) - f(b )
y -b
1
∕( v ) ~ »
lim
y →b y -b
= _J_
f'(b )
Because/is differentiable, we see that g is also differentiable.
■ LIMIT COMPARISON TEST (Chapter 8)
Suppose ak > 0 and bl , > 0 for all sufficiently large k and that
Cl,
lim -r- = L where L is finite and positive (0 < L < ∞)
fc→∞Dk
Then ∑afc and XZ>i either both converge or both diverge.
a I
Proof: Assume that lim -r-k - L , where L > 0. Using ε = =⅛in the definition of
k →∞ D∣i 2
the limit of a sequence, we see that there exists a number N so that
0., I τ
b— ∖< Σ whenever k > N
_ L <. ⅜ _ Lτ <
' L
2 bt 2
L . ak 3L
bk < 2
fb k <a k < ^2- bk bk > 0
This is true for all k > N. Now we can complete the proof by using the direct
App. B: Selected Proofs A -2 9
comparison test. Suppose ⅛bk converges. Then the series Xy⅛ j, a ^s o c o n v e r ge s,
and the inequality
a k < ~2 bk
tells us that the series Xαjt must also converge since it is dominated by a convergent
series.
Similarly, if 'lt bk diverges, the inequality
0 < ⅜ bk < a k
tells us that X¾ dominates the divergent series Xy bk , and it follows that Xαjt also
diverges.
Thus X¾ and X⅛z, either both converge or both diverge. ■
■ TAYLOR'S THEOREM (Chapter 8)
If/an d all its derivatives exist in an open interval I containing c, then for each x
in I
∕W = ∕(e ) + - c) + ⅛ r ⅛ - c)2 + ∙∙∙+ ^ p ( x - C)n + R n(x)
where the remainder function R,l(x) is given by

Λ"+1>(z )
for some zn which depends on x and lies between c and x.

Proof: We shall prove Taylor’s theorem by showing that if f and its first n + 1
derivatives are defined in an open interval I containing c, then for each fixed x in I
∕(*) = » + i⅛ ⅛ - ¢) + (χ - c )2 + ' " + ⅛ ∏ ( χ ^ c >''+ + *ι)! (χ ~ 6')"+1
where z is some number between x and c. In our proof, we shall apply Cauchy’s
generalized mean value theorem to the auxiliary functions Fand G defined for all
t in I as follows:
F(∕) = ∕(x ) - ∕(t) - — (x - t ) -----

n∖ (x - ty
. (χ -tr +1
r ι
(n + 1)!
Note that F(x) = G(x) = 0, and thus Cauchy’s generalized mean value theorem
tells us
F'(z) = F(x) - F(c) _ F(c)
G'(z) ^ G(x) - G(c) ^ G(c)
for some number z between x and c. Rearranging the sides of this equation and
finding the derivatives gives
A -3 0 Appendices
F(c) _ F∖ z)
G(c) G'(z)
-β n+1∖ z)
n∖ (x - z)n
- ( x - z)n
n∖
=β n+1 ∖ z)
F(c) = ∕ ( n ÷1>(z) G(c)

^ P ( x - c Γ = ∕( " + 1 V ) (x - c)"+ 1 '
Λ×) ~ ∕( c ) - ∕h '( x - c) ----- (tt + 1)!
Rearranging these terms, we obtain the required equation.
∕( x ) = ∕( c ) + ⅛ > ( x - c) + - + - c)n + ∕< " +1 V ) ' ∕ ∏ j>
■ SUFFICIENT CONDITION FOR DIFFERENTIABILITY (Chap. 12)

I f / is a function of x and y a n d / f χ , and f are continuous in a disk D centered at
(χ 0 , y 0 ), then / is differentiable at (xθ, y 0).
Proof: If (χ, y) is a point in D, we have
f ( χ , y) - ∕( ⅞ , y 0) = Λ× , y) - ∕ ( ⅞ y) + ∕ ( ⅞ y) - Λ ⅛ y 0)
The function ∕( x 0 , j∕) with y fixed satisfies the conditions of the mean value
theorem, so that
f ( x , v) - ∕( x 0 , y) = ∕ ( x 1, y)(x - x 0)
for some num ber x 1 between x and x 0 , and similarly, there is a num ber y 1 between
y and y 0 such that
∕( ⅞ , To) - f ( X 0’ To) = f f X O’ T1)(y - k 0 )
Substituting these expressions, we obtain:
f ( χ , y) - ∕( ⅞ , y 0 ) = t ∕( χ >τ) - ∕( ⅞ , y)] + [ ∕( ⅞ y) - ∕ ( ⅛ y 0)l
= ∖fx (,χv y)(χ - ⅞)] + [ ∕ ∕ ⅞ y i )(y - y 0)]
= lf x (χ i , y)( χ - ⅞)] + ∕ c(⅞, y 0 )(χ ~ ⅞ ) ~ f x (χ 0 ^ y 0 )(χ ~ ⅞ )
This is zero.
+ [fy (χ 0 , y 1)(y - τ 0 )] + f y (χ o>y 0 )(τ ~ τ 0 ) ~ f y (χ 0 >y 0 )(τ ~ τ 0 )
This is also zero.
= A-(⅞> ∕>)( χ ~⅞ )+ 4 (x o> y 0)(τ - τ 0)
+ [ ∕ x ( * p τ) - A-(⅞, y 0 )lCχ ^^ ⅞)
+ [fy (⅞, Tj) - f y ( ⅞ y 0)l (y ~ τ 0)
Let ε 1(x, y) and ε2(x, y) be the functions
≡1(^ y) = f x (x v y) - ∕ γ (⅞, y 0) and ε2 (x, y) = f y (x 0 , ^ 1) - ∕ ∕ x 0 , y 0 )
Then because x 1 is between x and xθ, and y l is between y and y 0 , and the partial
derivatives/, and / are continuous at (x0 , j>0 ), we have
, lim ε (x, y) = lim [ ∕ ∕ x 1, y0) - ∕ γ(x0, y0)] = 0

fo>)→(W 0) 11 (χ,y)→(χ 0 ,y0 ) x
lim ε (x, y) = lim [ f(x 0, y 1) - f7 (x0, y0)] = 0

(x1y)→( V 0) 22 (V )→(W 0) ∙v
so th a t/is differentiable at (x0, y0), as required.
■ CHANGE-OF-VARIABLE FORMULA FOR MULTIPLE

INTEGRATION (Chapter 13)
Suppose/is a continuous function on a region D, and let D* be the image of the

domain D under the change of variable x = g(u, v), y = h(u, v), where g and h are
continuously differentiable on D*. Then
∂(Λ-, y)
∕(x , y) dy dx = f[g(u, v), ∕ Z(M, V)] ∂(ιι, v) dv du
Absolute value of Jacobian
Proof: A proof of this theorem is found in advanced calculus, but we can

provide a geometric argument that makes this formula plausible in the special case
where ∕(x , y) = 1. In particular, we shall show that in order to find the area of a
region D* in the xy-plane using the change of variable x = X(μ, v) and y = Y(ιι, v),
it is reasonable to use the formula for area, A:
∂(χ, y) dv du
∂(u, v)
where D* is the region in the wv-plane that corresponds to D.

Suppose the given change of variable has an inverse u = M(X , y), v = v(x, y)
which transforms the region D* in the xy-plane into a region D in the z/v-plane. To
find the area of D in the wv-coordinate system, it is natural to use a rectangular
grid, with vertical lines u = constant and horizontal lines v = constant, as shown
in Figure B. la. In the xy-plane, the equations u = constant and v = constant will
be families of parallel curves, which provide a curvilinear grid for the region D*
(Figure B.lb).
a. A rectangular grid in the uv-plane b. Corresponding grid in the xy-plane

is curvilinear.
Figure B.1
Next, let R be a typical rectangular cell in the wv-grid that covers D, and let R
be the corresponding set in the xy-plane (that is, R* is the image of R under the
given change of variable). Then, as shown in Figure B.2, if R has vertices A(u, v),
A -3 2 Appendices
B(u + ΔM, V), C(U, v + Δv), and D(u + ∆w, v + ∆v), the set R* will be the interior of
a curvilinear rectangle with vertices
A[X(u,v), Y(u, v)]
B[X(u + ΔM, V), Y(u + ΔM, V)]
C[X(u, v + ∆v), T(w, v + ∆v)]
D[X(u + ΔM, v + ∆v), Y(u + Δw, v + ∆v)]
C'(u, v + ∆v) D'( U + ΔM, v + ∆v)
A'(u, v)
a. 7? is a typical rectangular cell b. The image of 7? under the change of

in the grid covering D*. variable is a curvilinear rectangle 7?*.
Figure B.2
Note that the curved side of R joining A to B can be approximated by the

secant vector
AB = [X(u + ∆w, V) - X(u, v)]i + [Y(u + ∆w, v) - Y(u, v)]j
and by applying the mean value theorem, we find that
AB = H r x(a, v) ∆w i + H r (⅛, v) ∆w^∣j
\_du , J Lθwv ’ J
for some numbers a, b between u and u + ∆zz. If ΔM is very small, a and b are
approximately the same as w, and we can approximate AB by the vector
'∂y j.
∆w i
where the partials are evaluated at the point (M, v). Similarly, the curved side of R
joining A and C can be approximated by the vector
β = Γ∣~ ΔV Î* + Hr ΔV ^∣
J
' L3V J L∂v Jj
The area of the curvilinear rectangle 7?* is approximately the same as that of
the parallelogram determined by a and /?; that is
i j k
~
∂u ΔW y^∆w
Δ
∂u 0
11« × β∖∖ =
^ ∆V ^ ∆ v
Δ o
∂v ∂v
j k
∂x ∂y 0
∂u ∂u ΔVΔM
∂x ∂y
0
∂v ∂v
App. B: Selected Proofs A -3 3
=∣⅛⅛Δ"Δ"
By adding the contributions of all cells in the partition of D, we can approximate
the area of D as follows:
APPROXIMATE AREA OF D = ∑ APPROXIMATE AREA OF CURVILINEAR RECTANGLES
∂'(x, y) . .
=ς ⅛V)
v δvδ"
Finally, using a limit to “smooth out” the approximation, we find
∂(χ, y) ∆v∆w
∂(u, v)
f [ ∂(χ, y) dv du
JD J ∂(U, V)
■ STOKES’S THEOREM (Chapter 14)
Let 5 be an oriented surface with unit normal vector N, and assume that S is
with that of .S'. If F is a continuous vector field whose components have continuous
partial derivatives on an open region containing 5, and C, then
f F -d R = f f (curl F ∙ N) dS
c jsj
Proof: The general proof cannot be considered until advanced calculus.

However, a proof where 5, is a graph and F, S, and C are “well behaved” can be
given. Let Δ be given by z = z(x, y), where (x, y) is in a region D of the xy-plane.
Assume g has continuous second-order partial derivatives. Let C1 be the projec
tion of C in the xy-plane shown in Figure B.3. Also, let F(x, y, z) = ∕(x , y, z)i
+ g(x, y, z)j + h(x, y, z)k, where the partial derivatives of f g, and h are
continuous.
Figure B.3
We will evaluate each side of Stokes’s theorem separately and show that the
results for each are the same. If x = x(t), y = y(t) and z = z(t) for a ≤ t ≤ b and
R(t) = x(t)i + y(t)j + z(t)k.
A -3 4 Appendices
Chain rule
Green's theorem
f θ ∕ + df ∂z , ∂h ∂z , ∂h ∂z ∂z , , ∂2z ∖^∣ j J
∖3y ^r ∂z ∂y ^1^ δy ∂x ^t ^ ∂z ∂y ∂x ^f ^', ∂y∂xj J c
∕⅛ t, ⅛ az , a/13z ⅞∕^ ⅞∕3z a/z3z) J .∣
∖3x ^^∂z ∂x ∂x ∂y ∂y ∂z ∂y ∂y ∂x)
We now start over by evaluating the other side of Stokes’s theorem.
( ∂h ∂z . ⅛ ∂z df ∂z . ∂h ∂z , ∂g ∂ f∖
∖ ay ∂x ^1^ ∂z ∂x ∂z ∂y ^r^ ∂x ∂y ∂x ∂y /
Because these results are the same, we have

∫c F ∙ d R = ∫ ∫ (curl F ∙ N) dS
■ DIVERGENCE THEOREM (Chapter 14)
Let D be a region in space bounded by a smooth, orientable closed surface S. If F is

a continuous vector field whose components have continuous partial derivatives
in D, then
div F d V
where N is an outward unit normal to the surface S.

Proof: Let F(x, y, z) = ∕( χ , y, z)i + g(x, y, z)j + h(x, y, z)k. If we state the
divergence theorem using this notation for F, we have
∫ ∫ [Λ i∙N ) + g ( j∙N )
5
+
ft(k∙N)]⅛S=∫∫∫(g + ∣ + f)rfK
D
∫ J ∕ ( i -N ) dS +
∫∫^(J∙N)<iS+ ∫∫Λ(k∙N)Λ-∫∫∫frfK+ ∫ ∫∫g^+∫ ∫∫ g ^
5 5 D D D
This result can be verified by proving
—
∂x dV
5 D
¥ dV
∂y
s D
^dV
∂z
s D
Because the proof of each of these is virtually identical, if we show one, the other
two can be done in a similar fashion. We will show the verification for the last of
these three. We will evaluate this third integral by separately evaluating the left
and right sides to show they are the same.
We will restrict our proof to a “standard region.” The complete proof can then
be completed by decomposing the general surface .S' into a finite number of
“standard regions.”
The standard solid region we shall consider has a top surface S r with equation
z = u(x, y) and a bottom surface S β with equation z = v(x, y). We assume that both
S τ and S β project onto the region R in the xy-plane. The lateral surface S l of the
region is the set of all (x, y, z) such that v(x, y) ≤ z ≤ u(x, p) on the boundary of
R, as shown in Figure B.4.
We know that the outward unit normal (directed up) to the top surface S τ is
—M i - MJ' + k
N rτ = —r-2- j. = - and dS= ydιι2 + u 2 + 1 dA
Figure B.4 A standard solid y∣
U + U2 + 1 x y xy
region in R 3
Thus,
Similarly, the outward unit normal N β to the bottom surface SB is directed down,
so that
vi - v j- k
NΓ = ∙ ZU- -
\ <+ < + I
and
h[x, y, v(x, v)] dA

A -3 6 Appendices
Because the outward unit normal NL is horizontal on the lateral surface SL , it is

perpendicular to k, and
∫ ∫ ∕z(k ∙ N J dS = 0
We now add for the surface 5:
Next, we start over by looking at the triple integral on the right side. Notice that
we can describe the solid A as the set of all (x, y, z) for (x, y) in R and
v(x, y) ≤ z ≤ u(x, y). Thus,
D r
= ∫ ∫ {h[x, y, u(x, y)] - h[x, y, v(x, y)]} dAχy

J RJ
We see that the left- and right-hand sides are the same, so
We can now conclude that (by similar arguments for the other two parts)
F ∙ N dS = div F dV
APPENDIX C Significant Digits
Throughout this book, various computational windows appear and, in the answers
to the problems, you will frequently find approximate (decimal) answers. Some
times your answer may not exactly agree with the answer found in the back of the
book. This does not necessarily mean that your answer is incorrect, particularly if
your answer is very close to the given answer.
To use your calculator intelligently and efficiently, you should become
familiar with its functions and practice doing problems with the same calculator,
whenever possible. Read the computational windows provided throughout this
text, and consult the Owner’s Manual for your particular calculator when you have
questions. In addition, there are Technology Manuals accompanying this text that
are available for TI and HP Graphic Calculators, as well as for MATLAB and for
Maple.
App. C: Significant Digits A -37
■ SIGNIFICANT DIGITS
Applications involving measurements can never be exact. It is particularly

important that you pay attention to the accuracy of your measurements when you
use a computer or calculator, because using available technology can give you a
false sense of security about the accuracy in a particular problem. For example, if
you measure a triangle and find that the sides are approximately 1.2 and 3.4 and
then find the ratio of 1.2/3.4 ≈ 0.35294117, it appears that the result is more
accurate than the original measurements! Some discussion about accuracy and
significant digits is necessary.
The digits known to be correct in a number obtained by a measurement are
called significant digits. The digits 1, 2, 3, 4, 5, 6, 7, 8, and 9 are always significant,
whereas the digit 0 may or may not be significant.
1. Zeros that come between two other digits are significant, as in 203 or 10.04.
2. If the zero’s only function is to place the decimal point, it is not significant,
as in
0.00000 23 or 23,000
Placeholders Placeholders
If the decimal does more than fix the decimal point, it is not significant, as
in
0.00230 or 23,000.01
↑ ' '
This digit is significant. These are significant.
This second rule can, of course, result in certain ambiguities, such as 23,000
(measured to the exact unit). To avoid such confusion, we use scientific notation
in this case:
2.3 × 104 has two significant digits
2.3000 × 104 has five significant digits
Numbers that come about by counting are considered to be exact and are correct
to any number of significant digits.
When you compute an answer using a calculator, the answer may have 10 or
more digits. In the computational windows, we generally show the 10 or 12 digits
that result from the numerical calculation, but frequently the number in the
answer section will have only 5 or 6 digits in the answer. It seems clear that if the
first 3 or 4 nonzero digits of the answer coincide, you probably have the correct
method of doing the problem.
However, you might ask why are there discrepancies and how many digits
should you use when you write down your final answer? Roughly speaking, the
significant digits in a number are the digits that have meaning. In order to clarify
the concept, we must for the moment assume that we know the exact answer. We
then assume that we have been able to compute an approximation to this exact
answer. Usually, we do this by some sort of iterative process, in which the answers
are getting closer and close to the exact answer. In such a process, we hope that the
number of significant digits in our approximate answer are increasing at each trial.
If our approximate answer is, say 6 digits long (some of those digits might even be
zero), and the difference between our answer and the exact answer is 4 units or less
in the last place, then the first 5 digits are significant.
For example, if the exact answer is 3.14159 and or approximate answer is
3.14162, then our answer has 5 significant digits. Note that saying that our answer
is correct to 5 significant digits does not guarantee that all of those 5 digits exactly
A -3 8 Appendices
is correct to 5 significant digits does not guarantee that all of those 5 digits exactly
match the first 5 digits of the exact answer. In fact, if an exact answer is 6.001 and
our computed answer is 5.997, then our answer has 3 significant digits and not one
of them matches the digits in the exact answer. Also note that it may be necessary
for an approximation to have more digits that are actually significant for it to have
a certain num ber of significant digits. For example, if the exact answer is 6.003
and our approxim ation is 5.998, then it has 3 significant digits, but only if we
consider the total 4-digit num ber and do not strip off the last nonsignificant digit.
Again, suppose you know that all digits are significant in the num ber 3.456;
then you know that the exact num ber is at most 3.4564 or at least 3.4555. Some
people may say that the num ber 3.456 is correct to 3 decimal places. This is the
same as saying that it has 4 significant digits.
Why bother with significant digits? If you multiply (or divide) two numbers
with the same num ber of significant digits, then the product will generally be at
least twice as long, but will have roughly the same num ber of significant digits as
the original factors. You can then dispense with the unneeded digits. In fact, to
keep them would be misleading about the accuracy of the result.
Frequently, we can make an educated guess of the num ber of significant digits
in an answer. For example, if we compute an iterative approxim ation such as
2.3123, 2.3125, 2.3126, 2.31261, 2.31262, . . .
we would generally conclude that the answer is 2.3126 to 5 signficant digits. O f
course, we may very well be wrong, and if we went on iterating the answer might
end up as 2.4.
■ ROUNDING AND RULES OF COMPUTATION USED IN THIS

BOOK
In hand and calculator computations, rounding a num ber is done to reduce the
num ber of digits displayed and make the num ber easier to comprehend. In
addition, if you suspect that the digit in the last place is not significant, then you
might be tem pted to round and remove this last digit. This can lead to error. For
example, if the computed value is 0.64 and the true value is known to be between
0.61 and 0.67, then the computed value has only 1 significant digit. However, if we
round it to 0.6 and the true value is really 0.66, then 0.6 is not correct to even 1
significant digit. In the interest of making the text easier to read, we have used the
following rounding procedure.
ROUNDING PROCEDURES To round off numbers:
1. Increase the last retained digit by 1 if the rem ainder is greater than or equal
to 5; or
2. Retain the last digit unchanged if the rem ainder is less than 5.
Elaborate rules for computation of approximate data can be developed when

needed (in chemistry, for example), but in this text we will use three simple rules:
RULES FOR Addition-subtraction: Add or subtract in the usual fashion, and then round off the
SIGNIFICANT DIGITS result so that the last digit retained is in the column farthest to the right in which
both given numbers have significant digits.
Multiplication-Division: Multiply or divide in the usual fashion, and then round
off the results to the smaller num ber of significant digits found in either of the
given numbers.
App. C: Significant Digits A3 9
Counting numbers'. Numbers used to count or whole numbers used as exponents

are considered to be correct to any num ber of significant digits.
ROUNDING RULE We use the following rounding procedure in problems requiring rounding by
involving several steps: Round only once, at the end. That is, do not work with
rounded results, since round-off errors can accumulate.
■ CALCULATOR EXPERIMENTS
You should be aware that you are much better than your calculator at performing
certain computations. For example, almost all calculators will fail to give the
correct answer to
(10.0 EE + 50.0) + 911.0 - (10.0 EE + 50.0)
Calculators will return the value of 0, while you know at a glance that the answer is
911.0. We must reckon with this poor behavior on the part of calculators, which is
called loss o f accuracy due to catastrophic cancellation. In this case, it is easy to
catch the error immediately, but what if the computation is so complicated (or
hidden by other computations) that we do not see the error?
First, we want to point out that the order in which you perform computations
can be very im portant. For example, most calculators will correctly conclude that
(10.0 EE + 5 0 .0 )- ( 1 0 .0 EE + 50.0) + 911.0 = 911
There are other cases besides catastrophic cancellation where the order in which a
com putation is performed will substantially affect the result. For example, you
may not be able to calculate
(10.0 EE + 50.0)*(911.0 EE + 73.0)/(20.0 EE + 60.0)
but by rearranging the factors as
((10.0 EE + 50.0)/(20.0 EE + 60.0))*(911.0 EE + 73.0)
should provide the correct answer of 4.555 EE 65. So, for what do we need to
watch? Try not to subtract two numbers that are close to each other in magnitude.
If you must, then be aware that you may obtain an inaccurate result. When you
have a sequence of multiplications and divisions in a string, try to arrange the
factors so that in each partial result up to that point, the interm ediate answer stays
as close to 1.0 as possible.
Second, since all computations a calculator performs are done with a finite
num ber of digits, it is unable to do exact computations involving nonterm inating
decimals. This enables us to see how many digits the calculator actually uses when
it computes a result. For example,
(7.0∕17.0)*(I7.0) - 7.0
should give the result, but on most calculators it does not. The size of the answer
gives an indication of how many digits “ Accuracy” the calculator uses internally.
That is, the calculator may display decimal numbers that have 10 digits, but use 12
digits internally. If the answer to the above computation is something similar to
1.0 EE— 12, then the calculator is using 12 digits internally.
■ TRIGONOMETRIC EVALUATIONS
In many problems you will be asked to compute the values of trigonometric

functions such as the sine, cosine, or tangent. In calculus, trigonometric arguments
A -4 0 Appendices
are usually assumed to be measured in radians. You must make sure the calculator
is in radian mode. If it is in radian mode, then the sine of a small number will
almost be equal to that number. For example,
sin(0.00001) = 0.00001
If not, then you are not using radian mode—make sure you know how to put your
calculator in radian mode.
■ GRAPHING BLUNDERS
When you are using the graphing features, you must always be careful to choose
reasonable scales for the domain (horizontal scale) and range (vertical scale). If the
scale is too large, you may not see important wiggles, if the scale is too small, you
may not see important behavior elsewhere in the plane. Of course, knowing the
techniques of graphing discussed in Chapter 3 will prevent you from making such
blunders. Some calculators may have trouble with curves that jump suddenly at a
point. An example of such a curve would be
which jumps at the origin. Try plotting this curve with your calculator using
different horizontal and vertical scales, making sure that you understand how
your calculator handles such graphs.
App. D: Short Table of Integrals A -41
APPENDIX D Short Table of Integrals

Each formula is numbered for easy reference. The numbers in this short table are not sequential,
because this short table is truncated from the table of integrals found in the Student Mathematics Handbook.
BASIC FORMULAS 1. (Constant rule 0 du = c

1 ,n+ ∖
n +। ι1,
2. Power rule U (IV l / ≠ -1
u n du = In ∖u; n = -1
3. Exponential rule e'l du = e"
4. Logarithmic rule M∣
ln∣ du = u In z∕ - u
Trigonometric rules 5. sin ιι du = -cos u 6. cos u du = sin ιι
7. tan u du = -In sec w∣ 8. sin M∣

cot it du = ln∣
9. sec u du = In sec u + tan u∖ 10. csc u - cot u∖

esc u du = ln∣
11. sec2w du = tan u 12. CSC2 M du = -cot u
13. sec u tan u du = sec w 14. esc u cot ιι du = -esc u
Exponential rule 15. a" du = i—

In a a > 0, a ≠ 1
Hyperbolic rules 16. cosh u du = sinh u 17. sinh u du = cosh u
18. tanh u du = In cosh u 19. coth ιι du = In sinh u∖
20. sech u du = tan^ , (sinh w) 21. csch u du = In tanh
22. f du βu⅛-ιM 23.

du =cosh^1^
Inverse rules √ fl2 - u 2 in a Vw2 - a 2
± tanh^ 1^ if u < 1
du 1 __ -∖iι ->c du a a a
24. *
α2 + id ∣
α∣LUii 6 a2 - u2 1 u
-c o th ^ -a if a > 1
{a
26. f du
/—χ---- χ2— a1sec
, tel 27∙ j du = - - s e c h 11-
MV M2 —a uyJa 2 - u2 a 1a∖
du cinħ- ½y 29.
du = -csch~ 1∣
M∣
28.
VT + u2 u∖ , 1 + u2
A -42 Appendices
INTEGRALS INVOLVING au + b
", (aιι + b y +l
30. (vau + b’)'du = (— n+—-lp).a—
f , , / v , 7 (a i/ + by + 2 b(au + by + 1
31. v , (n + 2)a2 (n + l)a 2
7. (au + b)n+3 2b(au + b)n+2 b2(au + b)',+ '
32.
(n + 3 α 3 (n + 2)a3 (n + l)<23
r w"+'(au + by ___ fl b___f m , + b γ ,- l d
m+n+1 m + n + 1 J ll 3aι1 °’ a u
u m (au + b)n+γ
m h f _ , ,
33. u m (au + b)"du = (m + n + l)a (m + n + l)a J li m lzâ u + xt7d u
~ u m + ∖au + b y ^ m + n + 2 [ m( + b γ+ xd
(n + V>b (n+Y)b J ( °’
34. f —⅛ - τ = ~ ln lα u + b∖
j au + b a 1 1
35. Γ u du _ lj - ^ ∖n ∖au + b∖
∣an + b
36. f u 2 du _ (au + b f _ 2 l^ u + b1 + *= +
∣au + b 2ai ai ai
Γ ιι3 du _ (M <+4 b)3 _ 2>b(au + b)2 3b2(au + b) _ 62 , ∣ ,,
37. 3α 2α 4 α4 α4 1 1
' au + b
INTEGRALS INVOLVING u 2 + α2
L! DU , = ⅛1Π(M2 + a 2)
56.
ιι2 + a2 2 '
58. f u 3du j _= u-52----a~2- l.n(/z„√2- ÷, az,2∖
)
J u~ + a 1 2 2
60. f dll = __ 21____L t a n -lM
J u 2(u2 + a 2) a u a3 a
INTEGRALS INVOLVING u 2 - a 2, u 2 > a 2
74. du 4 - In u - a or - - c oth^' - 75. [ _ ^d u _ 1 ln |M 2 _ fl2 |

IT7 —CΓ7 2a ιι + a a a J u~ - a 2 2 '
76. u2 du = u + ⅛ In a 77. ∫ ⅛ ⅜ = τ + T 1"∣≈<2 - o≈∣
u2 - a2 2 u+a
78. du 1 1In —u2 - a 2 79. [ du _ 1 + 1 , u —a
ιι(u2 —a 2) 2a u2 J u2(u2 - a2) a2 u 2a3 u + a
du _ 1_____ 1 , u2
3 2 2
ιι (u - a ) 2a u 2 2 2a 4 ιι — a 2
2
I N T E G R A L S IN V O L V I N G a 2 - u 2, u 2 < a 2
93.
f
1
du 1 ι ∕V + - ) or -ta n h ^ 1- 94.
f u du -5 In∖a 2 - u 2 ∖
1 a 2 - u 2 ' 2a m ∖a -
M∕ a a [ a- - u~
95. .Γ V du = _ + a a + u∖ 96.
f » 3 du _ _ ~ In∖a 2 - w2
∣ a2 - u2 2 l1n ∣a - u∖ 9—
, CΓ 7 u~
07 1f du 1 , I u2 1 OS
Vo.
f
1 du I + I ∣t ⅛ + m
' ιι(a2 - a 2) 2a 2 L !2 - M2 I u 2(a2 - w2) a 2u 2a 3 la - u∖
QQ 1Γ du 1 -+ 1 In 2/2
1 ∩β
1U U.
f du u , I ∣ ∖a + u∖
2 2a 4 m ∣a2 - u 2 ∖ 2a∖ a 2 - u 2) 4a 3 a - u∖
u3(a 2
- a 2) 2a 2 u {a2 - u 2)2
1ι u
∩1ι . If u du 1 11 ∩ 7
u‰.
f u 2 du
,
u I j ∖a + u∖
' (a2 - M2)2 2(a2 — ι<d) ∣ (a 2 - ιι2)2 2(a2 - w2) 4a a - u∖
11U J. IΓ w3 du a2 1 1 1„I„2 ,.2l 1 ∩ ∕i

1U*4.
f du 1 + 1 In “ 2
∣ (a 2 - ιι2)2 2(a2 - ι ∕ 2) ' 2 1"'~ “ 1 ∣ ιι(a2 - M2)2 2a 2(a2 - a 2) 2a 4 ∖a 2 — u 2 ∖
105.
Γ du _ -1 , u , 3 1 a + u∖
∣ w2(a 2 - u2)2 a 4 u 2a 4(a2 - a 2) 4a 5 m ∣a - u∖
106.
f du _ -1 + _______1_______ X ∣-
' a 3(a 2 - u 2y 2a 4ιι2 2a 4(a 2 - M2)+ a 6l n la 2 - a 2 l
I N T E G R A L S I N V O L V IN G Vακ + b
2 {a u Z,)
du 2V an + b 136.
u du = ? VΓ+⅛
Vau + b a Vau + b 3a 2
1 , Vau + b - V b
∕~Γ
f u 2 du = 2(3a⅛ 2 - 4abu + 8⅛2) , du _ Vb Vau + b + V b
1r7 138. 1
J Vau + b Y5ai ι∆⅛ι√ + b 2 t ∖au + b
130 [ du _ V an + b _ a f du 2V(aa + b y
140.
J w2Vaw + b b^lt J uVau + b 3a
, f /--------r , 2(3aw - 2b) -----
141. a V a a + b du = ——7— 25— - λVγ(a ι ι + ⅛)3
J 15a
f ?2 /------- τ 7 2(15a2ιι2 - ∖2a3bu + 862) . r -a-u-- — =-
142. a V a a + b du = ------------- 105a ---------- ’ ^ ( + ^π)3
I N T E G R A L S I N V O L V IN G V u 2 + a 2
168. J V u 2 + a 2 du = u^ u ^
2
+ a ~ + V 1∏∣M + V u 2 + a 2∣
f ∕~^— 7 J (y 2 + a ^y i2
169. uVu~ + a 2 du = ----------j ------
f , —------= , u(u2 + a 2) 2'2 a 2u V u 2 + a 2 4 , ∣ , ^ -■ —
170. u 2V u 2 + a 2 du = ---------- $------------------- $------------f gl - ln∣a + λVλu 2γ + a 52l∣
f ------ (a 2 + a 2)572 a ∖ u 2 + a 2)3z2
171. w3V a 2 + a 2 d5u = ---------- y - 2--------- -— 5— —
172. f , c} u - = ln∣M + V u 2 + a 2 ∣ or sinh^ 1^

J V u 2 + a2 a
A -4 4 Appendices
173. ∣
I √ U^duw ' v /, r 2
Λ +ι ^2
174.
175.
[ _ ⅛ L ÷ ° i )m - a = √ j τ τ 7
, Vw2
+ a2
3
176.
1 du _ _1. ∣ a + ∖⅛ 2 + a 2 ∣
2
' w√^w + α 2 a ∣ u
177. f
∣ du _ V M2 + a 1
u2∖∕u 2 + a2 a 2u
178. f
∣ du _ ∖∕u 2 + a 2 + 1 ■ α + V M2 + a 2 ∖
u3∖ , u2 + a2 2a2 u2 2a 3 ∣ w
INTEGRALS INVOLVING Vu 2 - a2
196. ∣ w + Vw2 - α2 ∣
= ln∣
1 Vw2 - a 2 1
197. f u d u = Vw2 - a
2
I Vw2 - a 2
198. ∣ = U2∕,U2 - a
2
+ cd l ψ + v ^2 _ fl2|
∣V M- - a 2 2 2
199. 1 V M2 —a 2 i
200. f du _ 1 Ji/
∣wVw2 - a 2 a <2
201. f du = Vz∕2 - a 2
u2y∕u 2 - a2 a 2u
202. If ___ du = √ a 2 - a2 , 1 ∣w
2 u V u - a2
2 2a2u 2 2α3 lαl
203. ∣√ u 2 _ fl2 d u = ιΛ ∕u-^- a- _ cr 1 ji ∣
m + V z;2 - fl2|
f ,_____ / ,,2 _ ,2∖3∕2

204. ∣WV M2 - a 2 du = ----- 3—-—
z
205. I w2√ a 2 - a 2 du = φ , a ) +M u 2 -a 2 _ ad ln | l / + V y2 _ fl2|
206. ∣
∖ 3√ π ^ = - ^ 7 2)5z2÷ ^ 2^ 2)3z2
INTEGRALS INVOLVING Vα2 - u 1
224. - , d u , = sin~1-
' Vd 2 - u2 a
225. 1f "du y ∕a 2 ^ u 2
' ∖, a 2
- w2
App. D: Short Table of Integrals A -4 5
u2 du dVa 22 - u2 .∣α■->
2 s i n -ιM
2 olll a
226.
∖∕ a - u
2 2
——— I" u2 aduu (β2 U2j )3'2 2 -f /— J y

227. , = =- ----- a ∖∕a 2 - u2
J ∖'a 1 - u2 3r
228. du . - - In a + V a 2 - u 2
uV, a 2- u 2 a u
229. du Vα2 - u2
u 2yda 2 - u2 a 2u
230. du Vα2 - u 2 2 i 3 ln a + Vα 2 - u2
u 3∖ , a 2 - u2 2a2u 2 u
- u 2 du = u∖ a 2 - u + ^ s i n - 'â
, 2 2
231.
(α2 - w2∖3∕2
232. = 3~^
2 - M2)3Z2!
w2Vα 2 - u 2 du= — --- 4— — ^+, a u ∖∕a8 - u +, α8 sc ι∙n -,wa
----- 2 J w(α 2 2 2 4
233.
FL2 - M2) 5Z2 a<2(a2 - w2)3z2
234. u i3.v∕∙~2-----
a 1 - u 22 du
J
= (----- $—--------
3
INTEGRALS INVOLVING cos au
311. I cos au du = s,ιna a u
312. (∖ cos a udu = ^

∣ ^ +^ ^
a2 a
I∣u cos au du = ⅛2 cos au + ( - — -¾)
2 sin au
313.
' a i ∖a a )
314. ∣w3 cos au du = (⅛ - - A ) cos au + ∖- — ⅛r) sin au
, ∖ a2 a4 J ∖a a 3 /
315. I un cos au du = u n s *n a u
u" 1 sin au du
a aJ
316. , = un sin au +, πu, n ~^ cos au n (n ~, 1)∖fun „-■
If un„ cos au du 2> J
cos au du
a 2 a 2 J
a
317. f cos2 au du = + s *n4 2<2M
INTEGRALS INVOLVING sin au
342. f sin az⅛u⅛1 az7u⅛.=

_ --c--o-s--a--u
J a
343. f u s i n a udu = ^ - ^ ^
J a2 a
344. 2
f u sin au du = ⅛ sin au + (-⅝ - — J cos au
J a2 ∖a i a )
A -4 6 Appendices
345. ∣ M3 sin au du = ∖- 1— A I sin au + p⅛ - — ) cos au

∣ ∖ a1 a V ∖a i a)
346. [ u n sin au du = - fγ c θs au + J M"- 1 COS au du
[ u" sin au du = cos au + to "~^ s9i n au - ( M"- 2 sin au du

347.
a ai a1 J
348. [ sin 2 au du = % -
1 2 4α
349. fI u
,, b
<,iinll 2 a
mu. √
au,, _ « ! _ u sin4β2 au _ cos 2au
350. f sin3(2Wdu = + ≤ ⅛
<2 3<2
INTEGRALS INVOLVING sin au and cos au
' 2 -- f ■ , cos(p - q)u cos(∕7 + q)u

∫ sin au cos au du = s ιn - a i i
2a 3174λ . I sin pu cos qu du = — -------- +
∕24" 1
375. I sin" au cos au du = √ι n , .^ u
J (n + l)a ∫ cos" au sin au du = - cos
(z n
,
+ l)a
377. I sin 2 au cos2 au du = ⅛- sιl2
J 8 32a
INTEGRALS INVOLVING tan au
403. tan au du = - - ln∣cos αz2∣l = - ln∣sec au∖ 404. tan 2 au du = ta n a u - u

J <2 l <2 l l J a
405. ∣ tan 3 au du = cdlλ t-- + - ln∣cos aιι∖ 406. f tan" au du = — ∏ - - ∣ tan"- 2 au du
J la a ' ' J (n — l)a J
407. f tan" au sec2 <21/ du =
tan , Anι
J (zn + l)a
INTEGRALS INVOLVING INVERSE TRIGONOMETRIC FUNCTIONS
445. cos 1^ du = u cos 1 - V α 2 - w2
446. u cos - 1 - du = — ⅛-) cos- 1 - - - ∕V a 2 - u 2

a ∖2 4) a 4
f ? ill . u3 ιu (u~ + 2α2 V a 2 - M2
447. u- cos - 1 - du = ⅜- cos - 1 - + ---------- ⅛-------------
a 3 a 9
' COS - 1 (M∕<2) . 7τ1 ∣ ∣ f sin - 1 (w∕<2) .
448. du = ~ ln∣M∣ du
U I u
' cos - 1 (w∕α) _ cos - 1 (ι∕∕<2) + 1j a + V<22 - u 2
449.
u2 u a u
f ( 1∖ 2 I Λ2
450. cos 1- du = M cos l1- - 2M - 2iy∕a2 - u 2 cos^ ∖U
∖ a) ∖ a) a
451. sin l - du = u sin l - + V<22 - u 2

a a
452. r u sin - l ^ du = ⅛ - ⅞) sin - l ^ + u^ a ∖~ u "

a> ∖ ∣t j CL ∣
, . . , u 3 . -1 u (u 2 + 2a 2) V a 2 - M2
453. M2 s m " 1 -u du = ⅞- sin + 2------------⅛------------
a 5 a 9
sin , (u∕a) (μ∕a)3 1 ∙ 3(M∕<J )5 1 ∙ 3 ∙ 5(w∕a)7
u +
454. u c iu ~ d 2∙3∙3 + 2 ∙ 4 ∙ 5 ∙5 + 2 ∙ 4 ∙ 6 ■7 ∙ 7 + ''
f sin '{u∕a) sin l (¾∕α) _ j a + V α 2 - u2

=
455.
J u2 u a u
∖ 2 / ∖ 2
456.
∏ sin - 1 - ) du = M( sin - 1 ^ J - 2w + 2 V a 2 - w2 sin - 1 ^
457. ∣ tan -1 ~ d u = u tan -1 ~ - ≈ ln(w2 + a 2)

I Ct Ct X*
458. [ u tan -1 ~ du - ⅛(w2 + a 2) tan - 1 τz - ⅛r

) Ct Λ* Ct -
τ,
459. ∫ u 2 tan -1 - du = ⅛- tan -1 - -
a 3 a 6
7 3
+ ⅜ ln(u
6 v
2 + a 2)
’
INTEGRALS INVOLVING e m
483. e au du = —
p auf
484. ue au du = — 1u
aj
^> 7 P aU 2)
485. u 2e au du = — [ u 2 - — + -
a a
7 e au / 1, n _ nu n ~' , »(» ~ 2 (—1)"∕¾!∖
486. u n e au du = ----- g u n ~' eau d u = (U a ' 2 if n = positive integer
a an /
‰
t_ l
__ ≡ I c ^
≈
Γ
+ au , (au)~ , ( M 3 .
487. 1 ∙ 1! 1 2 ∙2 ! 1 3 ∙3 !
_
' e au J _ - e au a f enau t dJ u
488. un u (« - 1)M"^ 1 n -- J u ~
du u 1 1*-J ^t âu∖
489. 1 yc
p + qeau p ap
du _ u 1 1 1-.-.1n r , n au∖
490.
(p + qea ")2 p ap(p + qe au ) ap 2 kr l
—7 = tan 1
du a∖, pq
491.
pe au + qe~au e au - ∖∕ -q lp
----- l n
∖,2 d ∖∕ -p q e au + ∖ , -q∕p
e a,l (a sin bu - b cos bu)
492. e au sin bu du =
a 2 + b2
eaιl(a cos bu + b sin bu)
493. e au cos bu du =
a 2 + b2
INTEGRALS INVOLVING In|«|

499. J ln∣w∣ du = w ln∣w∣ - u 500. J (ln∣w∣)2 du = u (ln]w∣)2 - 2w ln∣w∣ + 2w
501. f (ln∣u∣)" du = u (ln∣w∣)" — n t (ln∣w∣)"^1 du 502. f u ln∣w∣ du = (ln∣w∣ - τθ

APPENDIX E Answers to Selected Problems
CHAPTER 1
PROBLEM SET 1.1, Page 8
Many problems in this book are labeled ■ What Does This

Say? These problems solicit answers in your own words or a
statement for you to rephrase as given statement in your own
words. For this reason, it seems inappropriate to include the
answers to these questions in the answer section.
5. ⅜ 7. 77 9. 2 11. 0

3. 2x + y - 5 = 0 5. 2y - 1 = 0
7. x + 2 = 0 9. 8x - 1y - 56 = 0
11. 3 x + y - 5 = O 13. 8x - 3y - 17 = 0
15. 4x + y + 3 = 0
15. 4ττ 17. ⅛⅞≈ 0.3652
19. m = y; (2, 0), (0, -3 )
1. a. ( -3 , 4) b. 3 ≤ x ≤ 5 c. - 2 ≤ x < 1 d. (2, 7]
3. a. ι ∣ ι ∣ ι ∣ * ∣ ι ∣ ι ⅜ι ∣ ι ∣
» b. ∣ I ∣ I ∣ I * I ∣ I * I ∣ I ∣ I*
- 6 - 4 - 2 0 2 4 6 8 - 6 - 4 - 2 0 2 4 6 8
23. no slope (slope
c. I I I I I I ι <H -⅜ i ∣ i ∣ ι ∣> d. |i
| i ⅛i | o | i |
i i | i !■
: undefined); ( -3 , 0)
- 6 - 4 - 2 0 2 4 6 8 - 6 - 4 - 2 0 2 4 6 8
,. χ.
4--
5. a. d = 4 - M = (3, 0) b. d = 2√5j M = (0, 4)
7. a. d = 5-, M = (1, - ∣ ) b. d = ∖l2∙, M = ( -j> - ∣ ) ---- 1---- 1---- H H — ■
-8 -4
■— I---- 1---- i---- 1— ►
4 8 v
9. {0, 1} 11. { - 2 , 7} u . { -⅛ i V⅛l-+-l-⅛] -4 --
15. { 6 ,-10 } 17. { - 2 , 5} 19. { }; no solution 25. Point (6, 6) with 5x + 2y - 11 = 0; 8x - 3>, + 1 = 0;
8x - 3y — 30 = 0; 5x + 2y - 42 = 0. Answer is not
<71 ∫7τr, 1 1τr 1 ∫ 3ττ, 5τr. π ⅞} 25∙ ⅛ }
z ι ∙ I 6 6 J z j, t 4 4 3 unique. Another solution can be generated using the point
27. (-∞> - f ) 29. (-f> θ) 31. (2, 7] (2, 16).
27. a. -3 8 .2 b. c . _ 40
33. [ - 1 ,3 ] 35. (-∞ > 0 )U (∣> 5) 37. [7.999, 8.001]
29 dollars
39. a. f b. f 41. a. f or 60o b. f or 30o
43. Length is 4 cm.
45. a. b. - √ 2 c. - √ 3
47. a. 1.5574 b. -0.0584 c. 0.9656

49. (x + 1)2 + (y - 2)2 = 9 51. x2 + (y - 1.5)2 = 0.0625
A-49
A -5 0 Appendices
31. Let t = time (in years) and V = value (in dollars). Consid 53. R(a, g(a)), S(x 0 , g(x 0))
er ordered pairs (/, Γ).
V = 200,000 - 19,000 f, in 4 years, V = 124,000.
33. a. (3, 1), (1, 11), and ( -3 , -5 ); also draw parallelograms
35. 289 vehicles

1. D ι(-∞ , + » ) ;/ ( -2 ) = -1 ;/(1 ) = 5 ;/(0) = 3
3. Λ ( - ∞ , +∞)∙,∕(1) = 6 ;/(0) = - 2 ; / (-2 ) = 0
5. Λ ( - ∞ , -3 ) U ( -3 , +∞)5 /(2) = 0; /(0) = - 2 ; / (-3 ) is
undefined b. ∣ or 1.2 candles; 15 ∙ 2 3 or 0.46875 candles
7. D .(-∞ , -2 ] U [0, +∞)∙, / (-1 ) is undefined; ∕Q y =
59. a. (-∞ , 0) U (0, +∞)
/(1) = V3 b. [1, /?] for some positive integer p determined by the
9. 2λ(-∞ , +oo);/ (-1 ) = s in 3 ≈ 0 .1 4 1 1 ;/ ^ = 0 ; rat’s ability to do the experiment.
c. 7 min d. n = 12
/(1) = -sin 1 ≈ -0.8415
e. Mathematically, yes, because the limit is 3, but practi
11. D-.(-∞, +∞ )j∕(3) = 4 ; /(1) = 2;/(0) = 4 cally, no, because n has an upper bound.
1 3 .9 15. 10x + 5⅛ 1 7 .- 1 19. . } ,, 61. a. 625/ + 25/ + 900 b. $6,600 c. 4 hours
x(x + ∕z)
21. a. Not equal b. Equal 23. a. Not equal b. Equal
25. a. Neither b. Odd 27. a. Even b. Even
29. (∕o g)(x) = sin(l - x 2) 1. 0 3. 6 5. 7 7. 6
(g o ∕) ( x ) = 1 - sin 2x = cos2x 11. 2
9. 2 13. 15 15. 10
31. ( ∕° g)(w) = w; (g o ∕)(w ) = M 17. 8 19. 2 21. -1
33. (f o g)(x) = cot x; (g o ∕) ( x ) = tan(l∕x)
25. 1.00 27. 5.00 29. -0 .1 7 31. Does not exist
35. a. w(x) = tan x; g(w) = u 2 b. !∕(x) = x 2; g(u) = tan it
33. -0 .3 2 35. 0.00 37. 0.64 39. Does not exist
41. 1.00 43. 3.14 45. 0.00 47. 0.67
49. -0 .0 6 51. 3.00 53. 1.00 55. 2.00
57. 0.00
59. a. -32/ + 40 b. 40 ft/s
c. After 3 seconds; -5 6 ft/s d. / = 1.25 sec
61. 228 63. 0
65. a. -1 b. 0
c. Oscillates widely at the origin; cannot determine the
limit.
1. - 9 3. - 8 s. 4 7. 2 9.
V3
9
11. 4 13. -1 15. ∣ 1

17. 19. 1
2
21. 2 23. 1 25. - ∣ 27. 0 29. 0
35. -1 37. 1 39. -1 41. 0 1

51.
6
53. y 55. Does not exist 57. 4 59. 8 61. 3 63. - 3x^2
App. E: Answers to Selected Problems A -5 1
PROBLEM SET 1.7, Page 66 7. ∣∕(x ) - L ∖= ∣(x + 3) - 5∣

= ∣x - 2∣
1. Continuous; domain 0 ≤ t < 24, where I is the number
< δ = ε if ∣x - 2∣ < δ
of hours after midnight
Choose δ = ε.
3. Not continuous; p ≥ 0, where p is the selling price of
9. ∣∕(x ) - L ∖= ∣(3x + 7) - 1∣
ATT stock; p quoted in eighth of a dollar
= ∣3x + 6∣
5. Not continuous; domain m ≥ 0, where m is the number = 3∣x + 2∣
of miles traveled on the trip < 3δ = ε if ∣x + 2∣ < δ
7. No suspicious points; no points of discontinuity
Choose δ = ⅜.
9. Suspicious points: x = 0, x = 1;
discontinuous at x = 0, x = 1 11. ∣∕(x ) - L ∖= ∣(x2 + 2) - 6∣
11. Suspicious point: x = 0; = ∣x2 - 4∣
discontinuous at x = 0 = ∣x - 2∣∣x + 2∣
< 4δ = ε
13. Suspicious points: t = 0, t = -1 ;
If δ < 1, then —1 < x —2 < 1, so
discontinuous at t = 0, t = —1
1 < x < 3; thus, ∣x + 2∣ < 5.
15. Suspicious point: x = 1; Choose δ = m in^l, ∙j^, then ∣x - 2∣ < y
no points of discontinuity
17. Suspicious point: x = ? + nιr, and ∣x + 2∣ < 5, so ∣∕(x ) - L ∖ < 5 ∙ ∣ = ε.
discontinuous at x = / + nττ
19. Suspicious points: t = nτr, CHAPTER 1 REVIEW

discontinuous at t = nπ
Each chapter review begins with a proficiency examination.
21. /(2) = 3 23. /(2) = 7τ
Calculus cannot be reduced to formulas and vocabulary, but
25. No value possible 27. Yes your study o f calculus should begin by understanding its basic
29. Yes 31. No tools; that is, you should know the basic concept before solving
problems. It is suggested that you write out each o f the answers
33. Yes 49. a = 1; b = - y to the concepts problems in your own hand before continuing
51. a = 1; b = 5 53. a = 5; b = 5 with the practice problems. We provide all the answers to the
practice problems to help you with your study o f the chapter.
55. x = 1.618 57. x = -0.414
Practice Problems, Page 75
59. x = 1.116 61. - 1
22. a. 6x + 8y - 37 = 0
b. 3x + 10y - 41 = 0
PROBLEM SET 1.8, Page 74 c. 3 x - 2 8 y - 12 = 0
1. ∣∕(x ) —L∣ = ∣(2x - 5) — (-3)∣ d. 2x + 5y - 24 = 0
= ∣2x - 2∣
e. 4x - 2>y + 8 = 0
= 2 ∣ x - 1∣
< 2δ = ε if ∣x - 1∣ < δ
Choose δ = y.
3. Doubter wins;
∣∕(x ) - L ∖= ∣(3x + 1 ) - 5∣
= ∣3x - 4∣
Choose ε = ∣ and no δ can be found.
5. ∣∕(x) - L ∖= ∣(x2 + 2) - 6∣
= ∣x2 - 4∣
= ∣x - 2∣∣x + 2∣
Note: If ∣x - 2∣ < 1 (δ < 1), then 1 < x < 3, so we
note that ∣x + 2∣ < 5. Thus,
∣∕(x ) - L ∖ < 5∣x - 2∣
≤ 5 δ= ε i f ∣ χ - 2 ∣ < δ
Choose δ = mini 1, ∙∣1.
A -52 Appendices
21. A = - 1 and B = 1
23. a. 15 in. b. 21in. c. 29 in. d. 19 in.
25. a. (-∞ , 300) U (300, +∞) b. 0 ≤ x ≤ 100
c. 120 d. 300 e. 60%
27. a. 9π b. 6
29. a. V = 16TT; 5 = 24τr b. K = 15TT; S = 3V34τr + 6τr
33. c = - p x-intercepts are (2, 0), ( -2 , 0)

25. Algebraically they are the same, but they are not equal 35. a. Between 7.5 o C and 5.8 °C; probably about 7.2 0C or
because they do not have the same domain. 7.3 0C.
Domain of/: (-∞ , 0] U (1, +∞)
Domain o f g: (1, +∞) b. About 1.8 or 1.9 seconds
26. ( ∕o ∕)(x ) = sin(Vl - x 2); 27. C = 2 5 v ( ^ ^ ) 37. 5 = V 67t 2 —35t + 49; after 4 hours, s ≈ 31 km
( ^ ° ∕) W = l∞s x∣ 39. They will catch the spy in 5 hr. In 5 hr, the spy will have
28. a. ∣ b. ∣ c. - ∣ d. 0
gone 36 km further, so the spy just makes it.
29. A = - 1 ,B = 1 41. Take it on the 12th day.
45. a. Publisher B b. Publisher B
30. j v∕( x’ ) = x + sin x -- -
∖J x + 3
c. D = (profit A - profit B)
/(0) = < 0and∕(τr) > 0; there is at least one root by C 0.6x for x ≤ 4,000
= 35,600 - 0.3x for 4,000 ≤ x ≤ 30,000
the root location theorem. I -3,900 - 0.05x for x ≥ 30,000
Supplementary Problems, Page 76 John should go with A when D > 0, go with B when
l . P = 3 0 ;A = 30 3. P = 10 + 2√73∙,Λ = 4 0 D < 0, and break even when D = 0.
5. (x —5)2 + (j , —4)2 = 16 7. 5 χ - 3 j> + 3 = 0 d. Break even is 12,000 copies.
47. When one zooms in very close to x = 2, one sees a sharp
9. a. y - 6 = ∣( x - 5)
peak, or “ cusp,” which is missed by most graphs. There is
y - ι = -⅛ x + 2) no tangent at x = 2.
49. Solutions for P U T N A M E X A M IN A T IO N P R O BLEM S can be
y + 2 = - j( χ - 3) found in the American Mathematical Monthly. We believe
the library research to find the solution to these problems
b. The point of intersection is =Q.
is as valuable as the interpretation o f the solution when it
is found. Consequently, we will simply provide the vol
11. The highest point is 1,024 ft. ume number in which the original problem and solution
13. 1 15. - ∣ 17. 3 19. 2x + 1 appeared. This problem is found in Vol. 67 o f American
Mathematical Monthly.
CHAPTER 2
PROBLEM SET 2 .1 , Page 93

3. 0 5. 2
7. 0 9. 0; all x
11. 3; all t 13. 6r; all r
15. 2x - 1; all x 17. I s —2; all s
19. all x > 0 21. y = 3χ- 7

2x
23. 3s - 4y + 1 = 0 25. x + 25y - 7 = 0
27. x + 5j, —68 = 0 29. 216/ - 6 y - 215 = 0
31. 2 33. 0
35. a. -3 .9 b. - 4
App. E: Answers to Selected Problems A-53
47. y = A x 2 , so y' = 2Ax; at x = c, we have y = A c 2. The 3. (15) 2x - 1; (16) -21; (17) 2s - 2; (18) - ⅛ (19)
point is (c, A c 2),
and the slope of the line is m = 2Ac. The
equation of the tangent line is y - A c 2 = 2Ac(x — c). (20) ±(x + 1)~1'2
The x-intercept (y = 0) is x = 5. The ^-intercept (x = 0) is 5. a. 12x 3 b. - 1 7. a. 3x 2 b. 1

y = -A c 2. 11. -1 4 x ^ 3 + ∣ x ^ ιz3
9. 2z + 2z^3 - 20z~5
49. a. 2x + 3 b. 2x; 3
c∙ f'( x ) = g'(x) + h∖x) 13. 1 —x^ 2 + 14χ- 3 15. - 3 2 x 3 - 12x2 + 2
17
17
___ 22___2 19. 4x 3 + 12x2 + 8x
∙ (x + 9)
21. ∕'( x ) = 5x 4 - 15x2 + l;/ " ( x ) = 20x3 - 30x;
∕'''( x ) = 60x2 - 30;∕< 4>(x) = 120x
23. ∕'( x ) = 4x - 3 j ∕' ' ( x ) = -12x~ 4 ;
∕' " ( x ) = 48x^ 5∙, ∕< 4>(x) = -240x^ 6
25. 18x - 14 27. l x + y + 9 = 0
Xmin=Θ X n in = ^1
Xnax=6π X∩ax=4 29. 6x + y — 6 = 0 31. x - 6j, + 5 = 0
X s c l = l . 5707963... X s c l= l
Y n in = ^5 Y ∩ in= ^15 33. ( J , - ⅛ ) ,( l ,0 )
Ynax=25 Ynax=25
Y s c l= 4
37. (1, —2); derivative does not exist at x = 0.
39. No horizontal tangent line
2(X 3) - 3x 2(2x - 3) - 4 x 3 + 9x 2
x6 x6
_ x 2( -4 x + 9) _ - 4 x + 9
x6 - x4
Y 1B X +≥in X Y 1 B X s ir ∣X b. —3x - 4 (2x - 3 ) + x - 3 (2)
V∑BnDeriM<Vt ,X>X Y ≥ B n D e ri ∣
X Y ι , X, X = - 6 x ^ 3 + 9x^4 + 2x^ 3
)
X n in = 0 X n in = θ = - 4 x ^ 3 + 9x - 4
X nax= 6π i X n a x = l. 5707963... c. -4 x ^ 3 + 9x~4
Xsc 1= 1.5707963... X s c l= . 39269908...
Y n in = ^5 Y n in = ^1 d. Note that the results o f parts a, b, and c are all equal.
Yrιax=25 Ynax=2
Y s c l= 4 Y s c l= .5 43. 2x - y - 2 = 0
45. a. 4x + y - 1 = 0 b. x = 0
47. (0, 0), (4, 64) 49. No 51. Yes
53. k = degree of the polynomial function + 1; that is, the
(k + l)th derivative is 0.
Y 1 B s in <2πcos X)
Y 1BX^3-4XZ+5X-2
Y z B n D e riu < Y ι,X > X 1. 2 3. ∣ 5. ∣ 7. 0
) Y 2 B ∏ D e riv < Y ι,X ,X
X n in = ^ l )
Xnax=4 X n in = 0 9. sin 1 11. 1
X ≥ c l= l X nax= π i
Xsc∙ 1= . 39269908... 13. cos x - sin x 15. 2Z - sin Z
Y n in = ^15
Ymax=25 Yrf∣in = ^8 17. 2x cos x —x 2sin x 19. 2 sin Z cos Z
Y≥c∙l=5 Ynax=8
Y S Q 1=4 21. —x ιz2sin x + i x lz2cos x - x csc2x + cot x
b. The graph is increasing where the derivative is positive
and the graph is decreasing where the derivative is - , x cos x —sin x -- Z sec2Z —tan Z
i i∙
negative. x τ- fl
57. ∕'( x ) does not exist. sec2x - 2x sec2x + 2 tan x
2 λ ( l-2 x ) 2
- 0 t cos Z + 2 cos Z - sin Z - 2
zv∙ (Z + 2)2
, 1 1 ,, 2 cos x - sin x - 1
1. (3) 0; (4) 1; (5) 2; (6) 4x; (7) -2 x ; (8) - 2 x cos x - 1 (2 - cos x) 2
A -5 4 Appendices
-}5 ----------- =---------2 37. -sin x 25. 136 units

(sin x - cos x)
27. a. 9.8 m/min b. 9.8 m
39. -s in θ 41. 2 sec2 0 tan θ
29. a. 64 ft/s b. 336 ft
43. sec3 0 + sec θ tan 2 0 45. -s in x - cos x c. -32/ + 64 d. -1 6 0 ft/s
47. csc3y - 2 csc2 y cot y + esc y cot2 y
31. 30ft 33.144 ft
The graphs are 35. v0 = 24 ft/s; height o f cliff is 126 ft
almost the same.
37. a. 200/ + 400 b. 1,400
c. Increase by 1,500
39. a. 0.2 ppm/yr b. 0.15 ppm
c. 0.25 ppm
Y ιB < sin < X + .l)-s
in < X ) ) z .1 41. 91 thousand/hr
YzBco≥ X 43. a. 20 b. 0.39
Xmin=Θ
Xnax=2π 45. 7.5%
X s c l= l. 5707963...
Vmin= ^1. 2 47. a.- ^ z ,3 + l ⅞ ∙ l00 b. Between 2nd and 3rd wk
Vnax=1.2 5 - / + o∕^
V ≥ cl= .5
49. The slower car is traveling at 15Vz3 ≈ 26 mi/h
51. y - 1 = 2 ( x - ∣ )
51. = 3s2; it is 50% of the surface area.
as
55. y = 2x -, dP _ -k T , 2A
d V ( K - B) 2 V3
59. + y = 0 61. A = 0 ; B = - ∣
dx~ ■
3. 6(3x - 2) 5. (x 2 ° a9 )3
1. 1 3. - 3 5. 7. - 1
7. - 2 x sin(x2 + 7)
9. 4 1 1 .- 1 13. - 6
9. a. 5 ∕∕ 4 b. 3 c. 15(3x - 1)4
15. a. v(∕) = 2 / - 2 b. α(∕) = 2
11. a. 15w14 b. 6x + 5 c. 15(6x + 5)(3x2 + 5x - 7)14
c. Advancing on (1, 2]; retreating on [0, 1); total distance
13. a. 2 sin xcos 2x - sin 3x b. - s in 2 0sin x
traveled is 2 units.
15. 4 cos(40 + 2) 17. 2xtan x 2 + 2x 3sec2x 2
d. Always accelerating on [ 1, 2]
19. 2x(cos2x 2 - sin 2x 2)
17. a. v(∕) = 3 ∕ 2 — 18/ + 15
21. 2x(2x 2 + l) 3(x 2 - 2)4(18x2 - 11)
b. α(∕) = 6 / - 18
23. 2∕(cos ∕ 2 — ∕ 2cos t1 — sin ∕ 2)
c. Advancing on [0, 1) U (5, 6]; retreating on (1, 5); total
distance traveled is 46 units. 25. 3(∕2 - 1)2 [2∕ cos(3∕ + 2 ) - ∕ 2sin(3∕ + 2) + sin(3∕ + 2)]
d. Accelerating on (3, 6]; decelerating on [0, 3) 27
Z/
_____________
2
- (2x - l)
2+______
ιz2 2 3z2
(3x + 2)
19. a. v(∕) = —2 ∕^ 2 - 2r 3
29. a. x sin x 2(cos x 2)^3z2 b. ∣

ta n x ∣
b. <z(∕) = 4 ∕^ 3 + 6∕^ 4 z
■ ∣cos x∣
c. Always retreating; total distance traveled is 20/9 units. ,, 2xVx + 1 ,, , 2, ,,
3 ∙ 3√ x(x 2 + 2√Z) 2z3 33∙ y 3 ~ 3 (X 2)
d. Accelerating on [1, 3]
21. a. v(∕) = —3 sin / b. a(f) = —3 cos t 35. y - -i⅛ = 0 37. y = 0
c. Advancing on (π, 2 IT]; retreating on [0, IT); total 16
distance traveled is 12 units. 7
39. x = g 41. x = 1 or x = 7
d. Accelerating on ; decelerating on
43. a. f'( x ) = 2(x + 3)(x - 2)(2x + 1); - 3 , 2, - ∣
[θ, ∣') U ( ~ , 2ιr]
b. ∕" ( x ) = 12x2 + 12x - 22; - 3
23. a. - 6
b. Scores will be steadily decreasing. 45. a. —500π cm 3∕h r b. -200τrcm 2∕h r
47. $4,222.80/hr 23. ⅛ = ∑ ≠ ≈ -0.0249 in./min; = -2 .5 in.2∕m i

dt 64π at
49. - 6 (decreasing demand)
25. ⅛τ = 19.2τr ≈ 60.319 cm 3∕m in
51. a. - b. - - dt
x x
27. 7.2 ft/s 29. 0.13 ft/min 31. 7.5 ft/min
p _J_ d 3 33. -2 0 0 ft/s 35. 30 ft∕s
2x (2x + 1)(1 - x)
37. a. If 5 = V (5 0 ∕) 2 + (250 - 25Z)2, then
53. 1.28 lux/s
cj = ∣[2500z - 50(250 - 25Z)]
t
55. b. 48τr km/min
57. y - 3 = -6 ( x + 3) 59. g'(2) = -2 4 3,125z - 6,250
° r √3,125z2 - 12,500z + 62,500
61. ⅛ ∕[ ∕' ( x ) ] = f ' [ f ' (x)] ■∕''( x )
b. Z = 2 hr
c. 100Vz5 ≈ 224 mi
PROBLEM SET 2.6, Page 134 39. 60 rad/hr or — ≈ 570∕m in 41. ⅞r ≈ 3.927 mi/hr
ττ 4
_x
1. 3. 43. 8.875 knots at 2 P.M.; 10.417 knots at 5 P.M.
y X
2x + 3 j 32
(χ + y) 2 45. The water level is falling at about ∙7-∏— ≈ 0.023 in./min.
5. 7 44 177
3x + 2y (x + y) 2 + l
1 - cos(x + y) 47. a. cot θ = ~ θ b. 0.27 rad/sec
9.
x2 1 + cos(x + y)
1x - y sin xy c. ⅛γ = -⅛ ∙, as v increases, so does ⅛ and it becomes
13. dt 150 dt
x sin xy
hard to see the seals.
h 2x
15. a
a, ~x 0 3(12 —X 2 )2Z3
3y2 ∙
.b ____ 1 2
17. a. y2 ∙ (5 - x) 1. a. 4.123105626 b. 4.125
3. a. 1.967989671 b. 1.97
19. y - 3 = j ( x + 2) 21. y — π = (1 + π)x
5. a. 0.4714045208 b. 0.47
27∙ * - 1 » • -τ⅛ -' ,
23. 0 25. ¾ 7. a. 1.4641 b. 1.4
33. a. - ⅛ b. - ⅜ 9. a. 0.9949874371 b. 0.995
31. —4y^3
b 1 u crv 11. a. -0.00999983333 b. -0.01
35. The line is 4α + 3b = 0; points may vary. 13. a. 0.706399321 b. 0.7063
15. 0.003 17. -0.1 5
19. 0.00325 21. 6x 2 dx
23. x - ιz 2 dx 25. (cos x —x sin x) dx
3x sec2 3x —tan 3x j χ 29. (x 2 - l)^ lz2(2x 2 - 1) dx
27.
2x 2
33. 217.69 approx; 217.7155882 calculator;
error is 0.0255882.
49. θ = ? ( θ = ⅛ is not a solution because 0 ≤ θ ≤ ? . 35. Within approx. 6% 37. Within approx. 3%
J ∖ ∙~) 39. 0.05 ppm 41. -12,000 units
43. 28.3725 in. 3
45. Pulse rate decreases about 2 beats every 3 min.
1. $ = - 3 3. $ = 1 ,0 0 0 ⅜ =15
5. 47. Approx. 2% error
dt dt dt
49. Decrease by about 7 particles/unit area
_ ⅛ _ 4 dx _ 30 d F _ .
z ∙ dt
0 11. dt 3
5 dt 13 51. a. $472.70 b. $468.70
13. -1 0 √ 3 ≈ -5 .7 7 units/s 53. a. ^x + 3 b. $4.00
15. - ≈ 0.637 ft/s 17. $34,000 per year 19. —30 lb ∕in . 2∕s
77 c. $11.80 d. $11.40
A -5 6 Appendices

1. 3.072 3. 1.443 5. 2.006 24x 3 - 6(3x) 1/2 2 x 7(4λ∕3 - x 5z2)
[(3x) ιz2 + 3x^ 2] 5 ° r 9(x 5z2 + √ 3 ) 5
7. -0 .0 0 0 9. 0.453 13. 0.322
15. -2 .5 4 6 8 17. -2 .7 9 6 3 2 19. 0.876726 9. 10x 9(3x 2 + l)(x 2 + 1)9 11. (X 3 + υ
3* ffχ3 + °
21. If X Q = 2, then x l = 0 and it is impossible to continue.
More formally, 13. 8 x 3(x 4 - l) 9(2x 4 + 3)6(17x 4 + 8)
∣∕( x ) ∕" ( x ) ∣ _ ∣(1 - x~ 1)(-2 x ~ 3)∣ j5 —s in V x
I [/'M i 2 I I χ-4 I 4∖zx(x cosV zx ) 1/2
=
2x 2x 4
l ~ l-t ~∣ = ∣~ 2 * + 2∣ = 2 ∣ x - 1∣
17. 5(x lz2 + x ιz3 ) 4 Q x ~ ιz2 + ^ x -2 z 3 ^ or
When x 0 = 2, 2∣x - 1∣ is not less than 1, so Newton’s 5 x 2z3( x lz6 + l) 4 (3x ιz6 + 2)
method fails. 6
2 x ιz2∕ z2 - y i '2
23. Fails because = 0, so we cannot form 19. (cos x)(cos(sin x)) 21.
x ∣ z2
∕⅛ ) ' 23 1 - y cos(xy)
x " +1 X ∕' ⅛ ) x cos(xv) - 1
2∣^3 j 3 + 4 x 2 ^l
25. 1.53981 25. 20x 3 - 60x 2 + 42x - 6 27. 9L / J
27. Find one o f the following roots: —1.16137, 1.27506.
29. y = - 3 31.
29. -2.346407142
31. We need to find one o f the following roots: 0.01672 or 33. x + y - 2 = 0 35. y - 1 = 4 ( x - 1)
1.78105. (Note-, t is not negative.) 37. Tangent: y - 1 = 12(x - 1)
Norm al: y - 1 = - ⅛ (^ - 1)
35. K = ∣
=20 39. (3x 2 — 7)(2 cos t - t sin t) + 6x(sin t + Zcos t)2

37. a
81 41. ∕' ( x ) = 4 x 3 + 4 χ -3
39. They are all the same, because x n+1 0 ∕" ( x ) = 12x2 — 20x^ 6
∕' W f " ( x ) = 24x + 120x^ 7
41. ∙N÷1 = ⅜ - ⅛7 3≈ 0.4245071 ∕< 4>(x) = 24 - 840x^ 8
ς B
43 -----------------------------
(x 4 - 2)4z3(x 4 + l) 2z3
λ - dy_ _ *x + 1 y . d 2y _ 5y 2 - 20xy - 5x 2
CHAPTER 2 REVIEW ^ dx y - 2 x, dx2 (y - 2x) 3
47. Tangent: x - y + 2 = 0; normal: x + y = 0
2 lz2 -4 x 7(√ 3 x 3 - 6) 49. 0.864321
15. 3x + ∣ x - 2 sin 2x 16.
(√ 3 x 3 + 3)5 51. “ The change in tax rate with respect to time” ; or “ the
—x cos(3 - x 2) derivative o f the tax rate”
-y
17. 18. 53. 3 x 2[2(x 3 - 1)2 + 3)] 55. 36x(3x 2 + 1)
[sin(3 - x 2)] lz2 x + 3y 2
18(x - y)(x + 2y) 57. sin ^ ~ ^z cos at x ≠ 0; the derivative does not exist at
19. 0 20.
(x - 4y)3
x = 0.
21. 1 - 6x 22. y - 8 = 14(x - 1)
59. 75 mi/hr 61. - 3 6 cm 2∕h r
+ y - τ1
23. τTangent: >= f ( * - i) 63. 2.4 × 106 m ∕s 2 65. —1 .8 π m m 2∕s
I 1⅛
l 1 A-7 a. C(x)
67. rv ∖ = ^>∖(I —
21 1.500
---- F, x ∖1
1 + 4,800
Normal: y - x
4
= 7,800x 1 + 2x
24. 2ττ^ ft 2∕s 25. 2.89571
b. Cost decreases by approximately $1.16.
69. By im plicit differentiation we obtain a = ⅛ = .
Supplementary Problems, Page 161 / < dt m
1. 4.x3 + 6x - 7 , __________ - 4 x __________ Thus, F = ma = ml — — I = —kx.
(x 2 - l ) lz2(x 2 - 5)3z2 ∖ m ∕
4x - y 2 x 2 - 8x + 5 71. -1 0 0 .9 ft/s 73. - y = - 8.4 in . 3∕m
v or — (7-x------2
5- -x------- 2 —)2
77. See Am erican M athem atical M onthly, Vol. 45.
CHAPTER 3
PROBLEM SET 3.1, Page 175 PROBLEM SET 3.3, Page 190
1. Maximum, 26; minimum, —34

3. Maximum, 0; minimum, - 4
5. Maximum, 1; minimum, - 1
0
9. Maximum, 1; minimum, 0
11. Maximum, minimum, 0.41067 (approx.)

19. Maximum, 20; minimum, -2 0
21. Maximum, 76; minimum, -3 2
23. Maximum, 4; minimum, ⅛
6 o
25. Maximum, 0; minimum, —Λ√Z 108 ≈ -2.1822
27. Not continuous on [0, 2ττ]
31. /(0) = 0 33. g(3) = 3 35. ∕ ( - ∣) = ∣
39. a, b, and d, no such functions exist.

c. Any continuous function will suffice.
43. s'(0) = 60 45. 3 and 6
b. Increasing on (-∞ , —2) U (0, +∞)j
47. Maximum is 1,323 when x = 21 and y = 63. decreasing on (—2, 0)
c. Critical points: (0, 1), relative minimum; (—2, 5),
PROBLEM SET 3.2, Page 181 relative maximum
3. 1 5. 1-5275
7. ∖ ∣.2 5 ≈ 1.0772 9. ∣
11. - 1 + √3 ≈ 0.7321 13. sin^ 1( —) ≈ 0.6901

∖7Γ∕
15. Not differentiable 17. Rolle’s theorem applies.
19. / (-8 ) ≠ /(8); cannot be applied
21. Not continuous 23. Rolle’s theorem applies. 15. a. x = -∣ , x = -2 5
25. The function tan x is continuous on [u, v] and differen b. Increasing on (—∞, —25) U (j, +∞)
tiable on (u, v). This implies that ta n _ *a n --- = sec2α
decreasing on (-2 5 , ≡θ
for some u < a < v. Because sec2α ≥ 1,
∣tan v — tan ιι∖ ≥ ∣v — w∣. c. Critical points: ( j, —9481^, relative minimum;
27. N o ,/ is not differentiable at x = 0 on (—2, 2). (—25, 0), relative maximum
29. N o ,/ is not continuous at .x = 0.
31. b. 0
33. f'( x ) = —3 ’ which is always negative; however,
∕(⅛) ~ f ( a } = - 1 l∙s a l, w a ∙∙ . , ∩ , ,
⅛- g Ys P+ o s ι hve when a < 0 < b, so
there is no such number w.
A -5 8 Appendices
17. a. x = 0, x = 4 c. Critical point: (0, 1), relative minimum

b. Increasing on (—∞, 0) U (4, +∞)j
decreasing on (0, 4)
c. Critical points: (4, -156), relative minimum;
(0, 100), relative maximum
27. a. x = 0, x = 1
b. Increasing on (-∞ , 0) U (1, +∞)j
decreasing on (0, 1)
c. Critical points: (1, -3 ), relative minimum;
(0, 0), relative maximum
19. a. x = - 1, x = 3
b. Increasing on (-1 , 3);
decreasing on (-∞ , -1 ) U (3, +∞)
c. Critical points: (-1 , —1), relative minimum;
b. Increasing on
decreasing on (θ, U 2τr

21. a. t = -1 , t = 3
b. Increasing on (-∞ , -1 ) U (3, +∞)∙,
decreasing on ( - 1, 3)
c. Critical points: (3, —32), relative minimum;
(-1 , 0), relative maximum
23. a. 31. a. x = 0
b. Increasing on ^0, ~
decreasing on 0
c. Critical point: (0, 0), relative minimum
25. a. x = 0 b. Increasing on (0, 2 5 π 2) U (75τr2 , 100τr2)

b. Increasing on (0, +«); decreasing on (-∞ , 0) decreasing on (25π 2 , 75ττ2)
App. E: Answers to Selected Problems A -59
c. Critical points: (75τr2 , -1 ), relative minimum; 9. Critical points: (—3, —11), relative
(25τr2 , 1), relative maximum maximum; (3, 13), relative minimum;
increasing on (-∞ , -3 ) U (3, +∞)j
decreasing on ( -3 , 0) U (0, 3);
concave up on (0, + »);
concave down on (-∞ , 0)
39. Relative minimum at x = 4

41. Relative maximum at x = —3; relative minimum at 11. Critical points: (0, 26), relative
maximum; (—6.38, -852.22),
x = y; neither at x = 1
relative minimum;
(1.88, -6.47), relative minimum;
increasing on (-6 .3 8, 0) U
(1.88, +∞)∙, decreasing on
(-∞ , -6.38) U (0, 1.88);
inflection points: ( -4 , -486),
(1, 9); concave up on (-∞ , -4 )
U (1, +°°); concave down on ( -4 , 1)
51. A = - 16. B = 24, C = 0, D = - 1 5; relative maximum at

x = 0; neither at x = 2 13. Critical points: ( -3 , -9 ) , relative
B 7R maximum; (1, —1), relative minimum;
53. Neither at x = —?; relative minimum at x = --∏ ' increasing on (-∞ , -3 ) U (1, +∞)j__ l
A 5A ’
neither at x = 0 decreasing on ( -3 , -1 ) U ( - 1, 1); ^8
concave up on ( -1 , +∞)j
concave down on (-∞ , —1)
15. Critical point: (0, 0), relative

minimum; increasing on (0, +»);
decreasing on (—∞, 0);
concave up on (-∞ , +∞)
61. True
63. False; counterexamples vary.
65. False; counterexamples vary.
67. False; counterexamples vary. 17. Critical points: (0, 0), point of inflection;
/1 ι ∖ y ∙∙
-τ> , . , - , relative maximum; o.oo4-∙ ^
∖4 Z j θ ∕ ∕∖
PROBLEM SET 3.4, Page 201 0∙002^'
increasing on ^-∞> / ∖
5. Critical point: (-1 8 , -1 ) , relative minimum;
increasing on (-1 8 , +»); decreasing on +«=j; 1-0.2 A ^'' 0'2 1 ∖oz
decreasing on (-∞ , -18);

concave up on (-∞ , +∞) points of inflection: (0, 0), 432/’
concave up on ^0,
concave down on (-∞ , 0) U Q

7. Critical points: ( -3 , 20), relative maximum; (3, -16),
relative minimum; increasing on (-∞ , -3 ) U (3, +°°);
decreasing on ( -3 , 3); inflection point: (0, 2); concave up
19. Critical points: (0, 0), point of inflection;!
on (0, + »); concave down (0.84,—1.8), relative minimum; ∖ 4
on (-∞ , 0) increasing on (0.84, +∞)j ∖
decreasing on (-∞ , 0.84); —÷ lτ ) '
points of inflection: (0, 0), (0.62, 0);
concave up on (-∞ , 0) U (0.62, +<»);
concave down on (0, 0.62)
A -6 0 Appendices
21. Critical points: , relative minimum; c. N'(x) = 0 when x = 2.5, so the optimum time for the
break is 10:30 A . M .
I 1, 2 ), relative maximum; increasing on (—1, 1);
Critical points: (-0 .3 1, -5.21),
decreasing on (-∞ , —1) U (1, +°°); relative minimum; (2.20, -11.32),
points of inflection: (0, 0), relative maximum; (0.36, -4.76),
relative minimum inflection —h
-4
concave up on ( - V z3,
0) U (V3, +∞)', points: (0, -5 ), [5, —
concave down on (-∞ , - V z3) (J (0, V3)
23. Critical points: (0, 0), relative maximum; Critical points: (2.08, -3.12),
(15.4, -444.4), (4.37, -13.29), (5.03, -25.34);
relative minimum; inflection points: (-5 .6 8,
increasing on -28.57), (-2 .5 4 , -2.77),
(-∞ , 0) U (15.4, +»); (0.60, 3.32), (3.74, -10.31)
decreasing on (0, 15.4);
inflection point is (3.9, -140);
concave up on (3.9, +<»);
concave down on (—∞, 3.9)
25. Critical points: (γy> 1.13J , relative maximum; PROBLEM SET 3.5, Page 211
5ττ ι 1. 0 3. 3 5 .- 1 7. - ∣
-p> 0.44 I, relative minimum;
increasing on ( 0, IU 5τr 9. 0 11. +∞ 13. ~ is. o

12’
ττ 5∙π∙).
decreasing on ,12’ 17. ∙ψ 19. -6 4 21. 1 23. 0
12/’
25. +∞ 27. -∞ 29. +∞ 31. +∞
33. 1 35. +∞
concave up on 4’ 4 ’ 37. Asymptotes: x = 7, y = - 3 ;
concave down on I 0, ~ graph rising on (-∞ , 7) U (7, +»);
concave up on (—∞, 7);
concave down on (7, +∞)∙,
∖4 / no critical points;
27. No critical points; no points o f inflection
decreasing on (0, TT);
inflection points: (0.52, 0.65),
(1.57, 0), (2.62, -0.65);
concave up on (0.52, 1.57)
39. Asymptotes: x = 3, y = 6;
U (2.62, 2 TT); concave down on
graph falling on
(0, 0.52) U (1.57, 2.62)
(-∞ , 3) U (3, +∞)j
concave up on (3, +«>);
33. y ' = 2Ax + B and y " = 2A; concave down on (-∞ , 3);
concave up if y " > 0 or 2/1 > 0 or A > 0; no critical points;
concave down if y " < 0 or 2A < 0 or A < 0 no points of inflection
37. Maximum deflection at x = -j-
39. Maximum marginal cost is - 2 at x = 0; 41. Asymptotes: x = 0;

minimum marginal cost is -5 8 at x = 2. graph falling on (-∞ , 0)
41. a. -(2 ) 2 + 6(2)2 + 13(2)=42 U (0, +°o); concave up on
- ∣ ( 2 ) 3 + ∣(2) 2 + 25(2) = 49∣ (—∞, —Λ√Z2) and (0, +»);
3/—
concave down on ( - V 2, 0);
b. N(x) = - χ 3 + 6x 2 + 13x - ∣(4 - x) 3 critical point (1, 3) is a relative
minimum; point of inflection is
+ ∣( 4 - x) 2 + 25(4 - x) ( - ΛVZ2, - 2 ¾ )
43. Asymptotes: x = 2, y = 1;
graph falling on (-∞ , 2)
U (2, +»); concave up on
( - ΛVZ4, 0) and (2, +<»);
concave down on (-∞ , -V z4),
and (2, +∞)j critical point is
(θ, - β points of inflection
45. Asymptotes: x = -4 ,
x = 1, y = 0; graph falling on
(-∞ , -4 ) U (-4 , 1) U (1, +∞)∙,
concave up on (-4 , -1 ),
(1, +“ ); concave down on
(-∞ , -4 ), (-1 , 1); no critical
points; point of inflection
is (-1 , 5)
55. Solution 2 is incorrect; need to divide by the highest

power of x (x 2 in this example).
57. a. +∞ b. +∞ c. —∞ d. —∞ e. +∞
59. a. lim
λ-→+∞ r j ∖ b r-a s) r
b. Vertical asymptote
c. Graph of g Graph of h
For Problems 23-43, in addition to showing the requested
information, as an aid to graphing, we show the critical points
and inflection points with coordinates rounded to the nearest
tenth unit.
23. Relative maximum at x = —2;
relative minimum at x = 0;
inflection point at x = -1 ;
critical points: (-2 , 5), (0, 1);
inflection point: (-1 , 3)

25. Relative maximum at x = 0;
relative minimums at x = ±3;
inflection points at x = ±V3;
critical points: (-3 , 0), (0, 81),
(3, 0); inflection points:
(-1.7, 36), (1.7, 36)
A -6 2 Appendices
27. Relative maximum at x = - 3 ; 41. Inflection point: (0, 0)

relative minimum at x = 3;
critical value at x = 0 (neither;
it is a vertical asymptote);
critical points: ( -3 , -11),
(3, 13); no inflection points

relative minimum at x = —1;
inflection point at x = 1;
critical points: ( -1 , 0), (3, 32);
inflection point: (1, 16)
31. Relative minimum at x = -3/2;

inflection point at x = -2/3, 1;
critical points: (-1 .5 , —32.1),
(1, 7); inflection points: (-0 .6 7,
-16.1)

relative minimum at x = 14/5; 51. False; counterexamples vary.
inflection point at x = -7/5;
critical points: (0, 0), (2.8, -8.3);
inflection point: (-1 .4 , 10.5) PROBLEM SET 3.7, Page 226
3. 8 ft × 8 ft
relative minimum at x = 4; 5. a. 80 ft × 80 ft b. 80 ft × 160 ft
critical value at x = 2 (neither; 7. $360, when x = 30 9. 35 in. × 14 in. × 5 in.
it is a vertical asymptote);
11. r = and h = √∑R 13. r = ψ and h = j
critical points: (0, 0), (4, 8);
no inflection points
17. The largest box is 2 ft × 2 ft × 6 ft.
19. The minimum time is hours or 44.72 minutes. He has

5 minutes and 17 seconds to defuse the bomb.
21. The minimum value for L is 14.283 (when x ≈ 7.395).
23. The maximum area is 64Vz2 (when .s, = 2).
25. r= 27. 4 ft × 4 ft
V 5τr
Md
29 = r^~m - Md
/, M 2 y∕⅛m2 — Λ72
V 1 4m 2
31. R = r 33. d = 2√2 ft ≈ 3 ft
35. 8 ft 37. Minimized at T ≈ 3.96 °C

1. P(x) = - ∣ x 2 + ^ x - 98; x = 26
o Z
. . -2 (3 x 2 - 145x + 450) 15. i 17. 3

3. Pp ^ = 5(x + 3O) ;% = 2° 13. 0
∏ ∙2
5. Maximum average cost o f $25 when x = 4 19. 3 21. 2 23. 0 25. 2
7. a. RD(∕∙xV)∖ = 380.2
xQ x ~. >∕L^,∕(λxz.∖) = C -V _J_
5Λ + 5θ,. 27 .1 29. - ∞ 31. +∞ 33. 0
P(x) = -0 .0 7 x 2 + 14x Ct cos(∕3Q
35. 2 37. 0
b. Maximum profit o f $700 when x = 100 and priced at 2(8
$14
9. Order 400 cases each time.
11. a. x = 8 1. l x + C 3. x λ + 3>x + C
128 5. t4 + ∕ 3 + C 7. 2w3 - 3 sin u + C
b. Increasing if — 2 > 0; this is true on [0, 8).
9. tan θ + C 11. —2 cos θ + C
13. ∣ M5Z2 - ∣w 3z2 - ∣w^ 9 + C
15. j x 3 + ∣ x 5z2 + C 17. - t ~ ' + ∣ Γ 2 - ∣r 3 + C
19. jx 5 + yx 3 + 25x + C
21. - χ ~ ' - ∣x~ 2 + ∣x ^ 3 + C
23. ⅛x3 + ∣ x 2 25. ∣ x 3 - 2x 2 + x + ∣
15. The price should be $41 (assuming only $ 1 price increases

are possible; otherwise the price is $40.83).
17. Lower the fare by $250.
19. Plant 20 additional trees for a total of 80 trees.
21. Plant 62 grapevines (an additional 12 grapevines). Note
that we cannot plant 12.5 additional grapevines (the
maximum). If we check, we obtain the same value for
both 12 and 13, but why plant 13 when 12 gives the same
result?
23. a. Maximum when x = 80 or when p = 20; R is increas
ing on (0, 20) and decreasing on (20, 201√z3). 29. $249 31. 10,128 people
33. k = 20 ft∕s 2 35. 4,167 m i∕hr 2
CHAPTER 3 REVIEW
14. 2 15. - ∣ 16. 0 17. 0
18. Relative maximum: ( -3 , 29);

relative minimum: (1, —3);
inflection point: (—1, 13)
25. θ = I 27. r =
6 V 2ττ
29. The maximum value of V is which occurs at

2
r = 3 v 0∙
19. Relative maximum: (6.8, 38.3);
31. a. 3 machines b. $1,866.67 inflection points: (0, 0),
(-1 3 .5 , -96.4)
3. 8 5. y 7. 10 9. Does not exist

A -6 4 Appendices
20. Relative maximum: (τr, 2); 15. Maximum /(0) = 12;

relative minimum: (0, —2), m inim um ∕(2) = - 4
(2-ττ, -2 ); inflection points:
∕⅛ ( 5ττ
\3’ 4/’ ∖ 3 ’ 4/
21. Relative maximum: ^0, ∣

17. 0 19. - ∣ 21. 0
vertical asymptotes at x - ±2
23. Does not exist 25. 1 27. 9
31. j x 2 - 6x + C 33. -⅛ x~ 2 + C
29. 0
35. The red curve / is the function, the blue curve g is the
derivative.
39. Radius is P lj(ττ + 4)- 1 ; length is P 0(ττ + 4)~'; width is 2.
41. a 2 43. 35 trees; 6,125 peaches
45. 3 weeks; profit is $1,225.
47. Price is $60 (90 units) for a maximum profit of $5,600.
49. 9 machines 53. y = - χ + V 2 57. x =

23. Maximum of 5.00512, where x = 0.4;
minimum o f 4, where x = 1
59. f'{x ) > 0 when x < —j
24. 19 in. × 19 in. × 10 in.
25. The population will increase by 1,200, but the approxi and x > 1;/'(%) < 0 when
mation is ΔP ≈ 1,500. - ∣ < x < 1; f'( x ) = 0 when
Supplementary Problems, Page 252 x = 1, —|; ∕" ( x ) = 0 when

Y ιB X ^ 3 -X * -X + l
Y∑BnDeιîu<Yι ,X ,X
)
V 3 B nD eriu< Y ∑ ,X ,X
)
X n in = ^1. 5 V rιin = ^9
Xnax=2 Vnax=8
X s c l= .5 V ≥ c l= 2
61. f'( x ) > 0 when

x < -0.876123 or
x > 0.876123; f'(x ) < 0
when -0.876123 < x
< 0.876623; ∕'( x ) = 0 when
x = ±0.876123; ∕" ( x ) ≠ 0
Y t BXΛ 3-X+1∕X+1
V∑ BnD eriyC Vι ,X ,X
)
V sB n D e riu < V ≥ ,X ,X
)
X n in = ^ 2 Ψmi r ∣= ^20
Xnax=2 Vnax=2Θ
X s c l= l V s c l= lΘ
63. When x = 0.460355

65. a. 27 mi/hr b. 55 mi/hr
67. a. Estimate r = 310 b. 313.33
69. See American Mathematical Monthly, Vol. 48.
CHAPTER 4
PROBLEM SET 4.1, Page 264 31. l ( f 3z2 + 5)4 + C

0
1. 6 3. 120 5. 225 7. 9,800 33. a. ⅛ (3x 2 - 5)6 + C
JO
9. ∣ 11. 3 13. a. 3.5 b. 3.25
b. ⅛ (3x - 5)7 + ∙⅛(3x - 5)6 + C or
Oj J4
15. a. 2.71875 b. 2.523475 17. 0.795
5⅛(3 X - 5)6(18x + 5) + C
19. 1.269 21. 1.2033 23. 18 25. 68 J /o
27. 2 35. a. ∣(2 x 2 + l) 3z2 + C
29. False; the given area is a trapezoid.
A = i[C α + Cb}(b - d) = ⅛C(b2 - α2) not yC(⅛ - b. y j(2 x + l) 5z2 - ∣(2 x + l) 3z2 + C or
- α)
31. True; the graph represents the top half of a circle with ⅛ ( 3 x - l)(2x + l) 3z2 + C
center at the origin and radius 1. Therefore, the area is
ττ72. 37. a. ∣ ( x 2 + 4)3/2 + C
33. True, since the graph o f an even function is symmetric
with respect to the y-axis. b. j ( x 2 + 4)5z2 - ∣( x 2 + 4)3z2 + C or
37. a. b. g (l) = ∣ g(2) = g(2) - g (l) =
⅛ (3x 2 - 8)(x2 + 4)3z2 + C
c. Yes
39. By the tabular approach, we find the area (for n = 64) to 39. i sin 4 / + C 41. ∣
be 21.83594.
41. By the tabular approach, we find the area (for n = 256) is 43 3- V3 45 — 47 -
2 3 2
0.5990286; from Problem 17 the approximate area is
0.795, so these differ by about 0.2. 49. The integrand (Vx) is undefined on ( -4 , 0) and therefore
can not be integrated.
PROBLEM SET 4.2, Page 275 53. 3πζ + 8 55. 0 57. 0

o
1. 2.25 3. 10.75 5. -0 .8 8 7. 1.18 59. a. T b. T c. F
9. 28.88 11. 1.896 13. - ∣ 15. 1.5 61. a. It turns around when t = 2.
b. It turns around at 5 = —3.
17. - ∣ 19. ∣ 21. f f ( x ) d x = l ∖ f g(x) dx = 1
z ^ J-2 J-2
23. 10 25. 35.5 29. 1.1 31. 3.41364 35. c. 5
PROBLEM SET 4.3, Page 285 PROBLEM SET 4.4, Page 293
1. 140 3. 16 + 8α 7. x 2 + y 2 = C 9. x 3z2 —y 3z2 = C
11. 3y2 + 2(1 - x 2)3z2 = C 13. x 3z2 + 3y 11' 2 = C
9. a. j + π 2 b. 1 + Vτr
15. cos x + sin y = C 17. xy = C
1111. 1—— 13
11.
fz
1135. -4 19. x = Cy
0 15
21. y = kQ, where Q is the number of bacteria at time t

17. a. 1 b. 1 19. a. y b. y
(Q > 0).
.. 45 . 45 c. 45
21. a. y b. y 23. = C (T — TQ), where T is the temperature at time t and
25. ∣( x - 2 7)5/3 + C TQ is the surrounding temperature.

23. γθ(2x + 3)5 + C
25. y = kq(P - O), where Q is the number of people who
27. y - ∣ sin 3x + C 29. cos(4 - x) + C
have caught the disease by time t.
A-66 Appendices
a and d are orthogonal trajectories;

b and e are orthogonal trajectories;
c and f are orthogonal trajectories.
35. About 28 min
37. a. v(1.61) ≈ 98.5 ft/s
b. 5 = 351.56 ft when t ≈ 4.6875 sec
39. a. In 1990 (/ = 0) the population was 33,000.
b. In 1999 (t = 9) the population will be 42,000.
41. Value increases by $2,820 during the next 20 days.
43. It will be 20 years, or, in other words, 10 years from now.
45. No; the stopping distance is 199.9, so the camel is toast.
47. 171.875 calories/s
dh 4-8 b. 39 min
49. a. =
dt 2,304 480

31. Family of curves: x y 2 = C;
orthogonal trajectories: 1. ⅛Λ√Z30 ≈ 1.55, which is between 1 and 2
X 2 - Y 2∕2 = - K 3. V5 ≈ 2.24, which is between 1 and 5
5. Mean value theorem does not apply, because the function
is not continuous.
7. F(x) = 2x 2 + 9x + C; F'(x) = 4x + 9
9. F(x) = ∣ x 5'4 - 4x - i η p + 16; F'(x) = x l ' 4 - 4
l l .F ( x ) = - i ( l - 3 x ) - 3 - ⅛ F ' ( x ) = (r ⅛ 3
13. A = 25; c = 5 15. A = ≡ j c = ~3 *V39 β , 1 θ8
17. A = 1.839; c = ±0.744

19.1.5 21. 4 ~ - ≈ 0.3729
77
23. ∣ 2 5 .- 1 0
27. -25.375 29.
31. cos x 2 33. - ⅛ ⅛
37. 18.7 0C 3 9 . / ( 2 )= ∣

1. Trapezoidal, 2.344; Simpson’s, 2.3333; actual value, ∣
3. a. 0.78279 b. 0.78539
5. a. 2.03787 b. 2.04860
9. 0.79 ± 0.05 11. 0.727
13. 8.5
15. a. 164 b. 18
17. a. 184 b. 22
23. 494 m 2
25. 2φ == 79 17
19. 3.1 21. 578
0 21. 22. 5
3 2
27. a. The error term involves f ^ ∖x), which is zero for a
1
cubic, so the Simpson error is 0. 23. 24. 0
2
b. The difference is that∕ f4, (x) is unbounded near x = 2.
29. 11.25 25. a. 36 b. 26. 0
27. v(0 = t2 + t + 2; s{t) - t l ÷ ⅛ '+ 2t + 4

3
PROBLEM SET 4.7, Page 317 28. a. x = 5 b. 1,250
1. Vertical strip; 1
29. - y = - ∣ cos 3x + C or
7/2
[ ( - x 2 + 6x - 5) ∣ * - ∣ )] dx = 1
' y25 y = 3 sec 3x + C
Z Z∕ ,4∙8
O
30. a. n = 26 b. /7 = 4
3. Horizontal strip;
f [0 - (y 2 - 5y)] dy = - ^
Io υ Supplementary Problems, Page 320
1. 25 3. 6
5. Vertical strip;
'7r r2τr 5. 1,710 7. 6 ~ 32 V ^
sin x dx + (-sinx) dx = 4
9. √ 3 - l 11.0.810063
Should also show a graph in Problems 7-23. 13. ∣ ( x - 1)3 + C 15. x + C
7. 1 9 —
12 x 3(x 2 + l) 3z2 _
17. - ^— 3— ~ + C
n .∣ 13 ≡
5
19. - ∙ β ( l - 5x 2)3z2 + C
1 5 .2 17
12 21. 60
19. ∣ 21 23. 3x 8cos4x 3 - 2x 5 cos4x 2
12
3 25. 27. 9 -2 √ 6 ≈ 4 .1 0 1 0
23. 32
4
29. ⅛ 31. 3 - √3 ≈ 1.2679
25. a. $0.75 b. 0
27. a. $75.00 b. $432 1
33. x = . + C or y = 1
= $5.33 33. - φ ≈ $42.67 i -y x - C
3J1l. 3
35. tan y + cot x = C
35. π 37. a. 9 years b. $12,150 37. y 3 = 3 x + C
39. a. It will be profitable for 12 years.
39. sin y 3 = —∣ cos x 2 + C
b. $1,008
41. a. $20 b. 400 41. 1(√ 2 - 1) ≈ 0.5274
43. a. R(q) = 1(10 - q)2 q-, R'(q) = ∣(10 - <7)(10 - 3q)
43. 1.9541 (exact value is 2)
b. q = 2 45. 0.9089 47. 0.9096
c. p = 16 when q = 2; consumer’s surplus is $8.67 49. 0.2158 (« = 6) 51. 1.0141
8 53. -6.2 5 m ∕s 2 55. $776.48
45. Consumer’s surplus is ∣ units.
57. $2.67 59. x = 27
61. d = 10 ft/s2 ; v0 = 30 ft/s;
.s0 = 45 ft; s 1 = 125 ft
CHAPTER 4 REVIEW
63. a. 5 b. $6,374.26
Practice Problems, Page 319 65. $76.80 increase 67. $20,700
19. ^^5 20. x 5Vcos(2x + 1)
A -6 8 Appendices
CHAPTER 5
5. 31 7. 0.736 806 299 727

9. 13.463 738 035 11. 200.336 809 974
13. 9,783.225 895 91 15. 38,523.625 435 7
17. 2, -1 19. 3
21. 2, - ∣ Z J. 2
25. a. bx = 1 when b. bx = 0 when b = 0,

b = 1; algebraic x > 0; algebraic
y
2-- 2--
—I---- 1---- 1---- 1--------- 1---- 1---- 1---- h- ---- 1---- 1---- 1---- 1— i — 1---- 1 h→- 45. k = -1.498; depth is 3 m.
- 4 - 2 2 4 - 4 - 2 --------------- 2--------4 x
47. 11 years, 202 days
49. First National Bank is better.
51. 5.71%

b. (0, 1); y → +∞ as x → ±∞ 12x3 + 5
1. y' = 2>e3x 3. y'
c. E min. = 0.5 3x 4 + 5x
29. $10,081.44 31. $2,456.19 5. y' = 2x + 2 v ln 2 7. y' = ex + ex e 1
9. y' = 2xe χ 2 11. y' = —4e 4a' —2e3 2λ'

33. a. 30.12% b. 77.69% c. 7.81%
13. f'(t) = (2t + l)e z2+z+s 15. ∕ ' ( M) = e2" + e 2u
35. $659.51 37. 108 0F
PROBLEM SET 5.2, Page 343 17. ∕ , (x) = 2xe~ x - x 2e~x
1. f and g are inverses. 3. Not inverses 21. g'(t) = 2(< √ '- 1)

5. Not inverses
7 . ∕- ' = {(5 , 4), (3, 6 ),(1 ,7), (4, 2)} 23. ∕'(w ) = - l , _ cos x — sin x
j v γ - 25.
’ ztln u sin x + cos x
9. g ^ l (x) = Vx + 5
, _ 2x - ye xy
11. ⅛-'(χ) = j ⅜ ⅛ 27. y' = e~x (l - x) 29. y xe x y
31 dy = 1 - ye xy
13. 0 15. 2 17. - 2 19. 3.5
' ' ~ + ±
21 — 23. 4 25. 0.231 049 06 y
10
10 1 2
27. -1.391 662 509 29. 3 6 = 729 -I , ∙ Y ∙V L2X - 1 2(X - 9 ) X + 3
39. y' = J 6x - - ⅛ + τ^¼
31. 1 33. 2 35. - / ? = =2 r
√1 - x ■L x + 1 4x - 7J
A pp. E : Answers to Selected Problems A -6 9
y d. Concave down on
41. y'
z = -In x
43∙ y = ⅛
x (0, e8z3), concave up on
45. y' = y[2x In x + x] (e8z3, +∞)∙, inflection point
(e 8z3, ∣e~ 4z3)
47. y' = y[ln∣sin x∣ + x cot x]
49. y = 1 51. y = x
53. y = e 55. y = - 2 x + 2 13. a. D = (-∞ , +oo); or the set R
57. y = 1 ~ 3 ln 2 x + ∣ In 2 or v = -O.135x + 0.260 b. x = 0; the graph is rising on (-∞ , 0); falling on (0, +∞)
0
o o c. Relative maximum at (0, 1); intercept (0, 1)
59. J = ⅛(1 - In 4)(x - 4) + d. Concave up on (—∞, -2 ~ ιz2)
and on (2- ι z 2 , +∞)j concave
61. y - 9 = 2 4 0 (x - 1) 63. y = 0 down on (—2^ ιz2, 2 1Z2);
points of inflection (—2~ιz2 ,
e^ιz2), (2" lz2, e - lz 2 )
15. a. D = (-∞ , 0) U (0, 1)

b. None; the graph is rising on (0, 1) and falling on
1. c 3. b (-∞ , 0).
5. a. D = (—∞, +∞)j or the set R c. None; intercepts ( — θ) and (—1 9 0)
b. None; the graph is always rising.
d. Concave down on (-∞ , 0)
c. None; intercept (θ, ⅜ and on (0, 2 - V2); concave
up on (2 - v 2 , 1); point of
d. Concave up everywhere; inflection at x = 2 - V 2
no inflection points -ι— ►
2 t
7. a. D = (0, +∞)
17. a. D = (-∞ , + «); or the set R
b. None; the graph is rising for x > 1 and falling for
0 < x < 1. b. The graph is rising on (0, ∞) and falling on (-∞ , 0).
c. Relative minimum at (1, 1); no intercepts c. Relative minimum at (0, 2); intercept (0, 2)
d. Concave up everywhere; d. Concave up everywhere;
no inflection points no inflection points.
9. a. D = (-∞ , +<»); or the set R

b. x = 1; the graph is rising on (-∞ , 1); falling on (1, +∞). 19. a. D = (0, +∞)
c. Relative maximum at (1, e - 1 ); intercept (0, 0)
b. x = 1; graph is falling on (0, 1) and rising on (1, ∞).
d. Concave up on (2, +»);
c. Relative minimum at (1, 0); intercept (1,0)
concave down on (-∞ , 2);
inflection point at (2, 2e- 2 ) d. Concave up on (0, e),
concave down on (e, ∞)∙,
inflection point at (e, 1).
11. a. D = (0, +∞)

b. x = e 2'. the graph is rising on (0, e2); and falling on 21. a. 5'(x) = ⅛(ex - (-l)e^ *) = C(x)
(e2 , +∞).
c. Relative maximum at (e2, e~ l ); intercept (1,0) C'(x) = l ( e → ( - l ) e - η = S(x)
A -7 0 Appendices

5x + C
1. ⅛ 3. ∣ln ∣2 x + 1∣ + C
v 2
5. y + C 7. - In w + e u + C
1 -χ 2
9. - ∣ e + C
11. —βl∏∣l - ∙*3 ∣ + 2^x ~*~
13. yln(2x 2 + 3) + C
15. l x + 11 ln∣x - 5∣ + C
vv7 + C 0 3 + (i
17. ∣ e 19. -In—2 + c
2
21. ∣(ln x) + C 23. 21n(Vx + 7) + C
25. ln (? + 1) + C 27. yln(5 + 2sin x) + C
29. 2 In sin Vx∣ + C

39. e 31. -ln∣sin x + cos x∣ + C
41. G'(x) = (A B k)exp(-kx - Be~ kx )
33. jin 3 ≈ 0.92 35. 0
G'(x) ≠ 0, so there are no critical points 0
G"(x) = (A B k)exp(-kx - Be~ kx ) { - k + Bke~ kx )

37. e - V e ≈ 1.07 39. yin 2 ≈ 0.3466
G "(x) = 0 when e~kx = 4ρ or x = γ- In B
v ’ B k 1 79833
41. 2.91n(∣ + y ) ≈ ∙
1 3 7 2∕ 7 ≈ 0.2475
43. 45.
4 In 2
-⅛(e 12 - i ) ≈ 10.850
47.
15∖ e)
49. —In 2 ≈ 0.44 51. In In x ∣ + C

77
53. ^ -y -^ ≈ 2 8 .8 55. 4,207
43. Minimum, g(e 1) = 1 — e 59. 8

1 ≈ 0.6321;
57. 1.5 - 21n 2 ≈ 0.11
maximum, g(4) = 8 In 2 + 1 ≈ 6.5452 5
45. a. e° 2≈ 1.22 b. 2%
47. ∕( t ) = Ae~ kt PROBLEM SET 5.6, Page 372
a. ∕'(Z) = -A k e ~ kt
3. λ 77
3
>
b-
77
3 5. a' kb
b. I O O 7 ^ = 1 0 0 ^ ] = -100fc 4 ' 6
∕(Z) L -A ekl J
49. a. 406 copies b. 368 copies 7. a. —y b. τr 9. 11. 3
51. 208 years from now ~2 ^ - ≈ -0.9798 1
13. 15.
" √ -x 2 - X
53. a. v(Z) = x'(Z) =
X 6(sin- 1 2x)2
αw 17. 19.
=x "w = e^{ λ 4 -7Γ4⅛ 0 (2 + x 2) V x 2 + 1 √1 - 4x 2
1 1
b. T = 0.5; Λ-(0.5) ≈ 0.43 21. 23.
(1 + 4x 2)√tan~ 12x cos - 1 x √ 1 - x 2
c. Total distance is 0.49 unit. -1
25. 27. —sin x
55. ∕( π ) = 1 1+ x 2 ∣sin x∣
App. E: Answers to Selected Problems A - 71
1~sin^ ~ b. If x = 1, then
29 τ⅛ 0 = ∕> ∙0 + C , s o C = 0
√T⅛ + tan" x Thus, F(x) = G(x)
In x p = p In x
31. ∣ t a n -, θA + C
9. It is clear that
In 2 < In e < In 3
33. ^ysin~ 1( ⅛ J ^ + C ^ln 2 <-- ^ln e < ^ln 3
2 < e < 3
35. ln(x 2 + 4x + 5) + tan~1(x + 2) + C
37. ∣ln ( x 2 + x + 1 )- ^ = t a n - ' [ ^ - ± - l + c
CHAPTER 5 REVIEW
39. —V 4 —2x —x 2 + sin~l ^x ^ J ^ + C
17. a. 2 x e -^ - ∣ x 3' 2< r ^ 3
41. tan - l (l + x) + C 18' a' V l - (3x + 2p
43. f 45. f - h a n - 1(V)
o o 2 ∖2∕ . In 1.5 bh ∙
x(ln 3x)2 1 +- 4x 2
dy = Γ_________ 2x______________1_ 9 1
dx ∙l L(-V2 - l)ln(x 2 - 1) 3x 1 - 3xJ
51. - - 1- --- 53. 1 57. 7 ⅞ 20. Relative minimum at ( -1 , —2e);
x 12
relative maximum at (3, 6e~3);
59. Relative minimum at x = V a i; points of inflection at
relative maximum at x = - ∖''cιb x = 2 ± V 5 ; intercepts
(0, -3 ), (± √ 3 , 0);
asymptote, y = 0
21. a. ∣ b. d iY ' + C c. -∣ln ∣c o s 2x∣ + C

o 2 2
22. a. e4 b. Does not exist c. 1
23. ∣e 2 - e + ∣ ≈ 1.47624
65. 3√Γ7 ≈ 12.4 ft
24. a. 11 yr 208 days b. 11 yr 180 days
PROBLEM SET 5.7, Page 377 c. 11 yr 166 days
25. $140
1. Let L(x) = ∣ y . Then we have
Jι t
L(xy) = L(x) + L ( j ,) and L (x r) = r£(x) Supplementary Problems, Page 579
In particular, 1. a. 1.504077 b. 16.444647 c. 1.107 149
L(2~jv) = - Λ rL(2) < 0
since ι . a„ . 3
3 h 63
b.
5 65
L(2) = f^ > 0 5. a. V13 b. ⅛

J∣ t
1. a. 16 2, l n 3. ≈ 1.4084
As V → +∞, 2~N → 0 and L(2~ N ) → -∞ . b. - , o
Thus, 2 In 3 - In 2
lim In x = —∞ 9. a. ∕ ^ 1(x) = ln(l + x 2 )
x→0+
3. n = 18 b∙ , Λ 1(V = ⅞ ⅛ τ
5. a. F'(x) = 2 ;
11. j∕ ^ 1(x)
v
= ~— domain1 is1 [χ∣x ≠ -)
’ ex — a c>
G(x) = p In x, so G'(x) = ^ = F'(x).
13. = (4x
v
+ 5)e2χ2+5 * - 3 15. = 32 ^x (l
y
- x ln 3)
Therefore, F(x) = G(x) + C. dx ’ dx '
A -7 2 Appendices
dy y( 1 + xye xy ) 43. ∣ln ( x 2 + 2x + 5) + C

17 = 19. y~ = es'n vcos x
' dx x( 1 - yxe xy ) dx
∙∖⅛t a n '(
21. ⅛ = - π ( l —χ ln 3x) dy _ 2 x 2 + 2xy 2 - 1 45. + C
ax x In 5 v 7 dx 2y - 2 x - 2y 2
25. ⅛ = s in ^l χ ~ x √ Γ ⅛ + - √ - tan - l x ) 47. In 1.5 497. 3
dx ------- ( X X ⅛ ------- x 2V + x 2 )
51. 2ey fx~ + C 53. ex — ln(eλ' + 1) + C
~ιn dy . . . , sin x , χ 55. ln∣cos x + sin x∣ + C
27. -√- = cos xv (sιn 'x) + - , , + cot 1x - 7—— j
dx V 1~ x2 1+ x
57. sin - 1 x + :c + C 59. Xln∣sec x + tan x∣ + C
29. y = 2x - y 31. y = 1
1 3
61. ⅛tan^ x + C 63. tan -1 (In x) + C
33. dy = Γ ln 5 + 3 ( 2 x ^ l j _ 12(2x 2 - 1)]
dx ∙ L x - x 2 x i - 3x J 65. a. c 4 b. 0 67. a. 0 b. 0
35∙ f i - 2¼ = + l >+ ⅛ ] 69.
1 71. ∣ln 2
2 In 2
37. ⅛ = 2,[ta(3*≈ + 2) + 3 ⅛ ]
73. Relative maximum at x =
.∙ ■ ∙ , —υ τ v υ— uru
Domain (0, + »); relative
1
minimum at x = ----------z----------
relative minimum at (e~ γ' 2,
point of inflection at x = e 75. a. $1,075.71 b. $1,070.52
intercept (1, 0);
77. 20π mi/min 79. ‘
defined only for x > 0 1+ x2
81. The body had been in the freezer for 35.5536 hours, so
Siggy had been put into the freezer on Wednesday morn
ing at about 1:30 A.M. Andre was in the slammer, so
Boldfmger must have done it.
41. Domain (—∞, -1 ) U (1, +∞)j 83. First National Bank
no extreme points; 89. 0.64 91. 0.75
no points of inflection; 93. See American Mathematical Monthly, Vol. 46.
no intercepts;
asymptotes y = 0, x = ± 1 95. See American Mathematical Monthly, Vol. 58.
CHAPTER 6
-2
1. 9 3. i 5. 2 7. 36√3 J
o
[(1 + V 4 - j; 2)2 - 1] dy
128√3 √ 1 Λ _ τA or shell: 2ττ ∣ (x + 1)V4 - x 2 dx

0 11 Jo
15 4 V 4/
Γ4 r4
13. a. Disk: π I (4 - x) 2 dx b. Shell: 2π x(4 —x) dx (x —x 4) dx
Jo Jo o
-4
b. Shell: 2ττ f (x 3/2 - x 3) dx
J
o
[(5 - x) 2 — 1] dx
-4
Jo
rl
J o
(x + 1)(4- x) dx
J0
(j7 - y 4) dy
f 2 f------25 ”!
J
Γ2
15. a. Shell: 2ττ y ∖ 4 - y d y b. Disk: π (4 - y 2) dy [(x3 + 2x + 1)2 - 1] dx
Jo Jo o
f4 13. 192 lb 15. 460.8 lb

b. Shell: 2π x(x 3 + lx ) dx
Ji (°’ ^ τ )
17. 9,056√2 lb 19.
21. 144 23. 18 25.
81 , 7 9
10 2 ( 1 1 ∖ /7 11„ ->λ
21. 23. v ’ 2 ~)
∖ln 2 4 In 2/
τr8 ττ7 τr6
29. I 31. 77 33. y + y + y ≈2.555
200τr
25. 27. y77 + 8τr2
37∙
3
35. ⅜ 2ττ 39. Air 41 -X
⅛ l )
3 10 29. 31. 4 ergs
43. f
77 477τ
47. 210
419ττ
v∙ 15 33. 0 ft-lb 35. 345,800 ft-lb
4
16 8ττ 37. 7,920π ft-lb 39. 592.8 lb
51. 28ττ 53. 55.
3 3 41. 278.2 lb 43. 23,400,000 lb
57. 90,000,000 ft 3 59. If ∆x = 0.5, sum is 284.375;
- 64./ - 256 47. ∣√ 6
if ∆x = 0.25, sum is 260.4063; 45. τ* 16’ y 7
if ∆x = 0.1, sum is 246.295
49. a. For V = 500, h = -0 .6 4
About the x-axis is ⅜πab2 ',
61. 10.045 65. b. For V = 800, h = 0.87
about the y-axis is yτrα2A 51. Use method o f disks: V = 77 x 1 dx = ⅛ r 3

Jo
67. V = ∣ ( 2 Λ 3 - 3R 2h + A3) ττL3
53.
4
2
69. V = R h2 π - ∣ τr(2A3 + hi)
1. y = C e 3x 3. y = - I n C - x
PROBLEM SET 6.2, Page 401 5. y = C x 7. y = sin - 1 (sin x + C )
1. 3√Tθ 3. 2√5 5. - ∣+ y √ 5 9. y = Ce x - 10
11. a. ii b. i c. v d. iv e. iii
331 123
7
120
0
32 13. y = 2 x 372 + ∣x + C x ~ ‘1
11. 9.29 (needs numerical integration) 1 -,2x c

15. y = - In x + - 17. y = ½xe -1----- 2x
57
∙ χ X 2 xe
I J3. —
1 3
19. y = - 21. y 2 = x 2 + C
15. a. S = 2τr Γ ( ∣ x 3 + - τ - ) ( ⅛ + x 2 ]dx X
∣1 ∖3 4Λ7∕∖ 4 ) 23. ∣ y = x 2 + C 25. About 11,000 yr old
= 1.505ττ
18 27. $1,902 billion 29. About 8,241,820
2
’ x f x + — ∣ dx = 2τr^20 + ^ln 3^ 31. a. About 70 min b. 94 lb
b. 5 = 2 T7 1
∖ %2; 33. It has no effect. 35. 2,500; (2488.707549)
Ji
37. He will land safely.
17. 12τr√5 19. ¾ τ 7 r ≈ 25.28
o4 _ sm }
,, Γ28√3 16^∣ 43. a. v = + e~ ktlm ∖v„0
23 k ∖ k J
21. 3τr√Tθ ∙ πL 5 15j
, , , mg m ( - ^ ) ( 1 - e - k l " n) + 50
25. 6 units 27. 36 units b. s(t) v0
29. 8.632601 (n = 20) 35. 5 = 4τ72rtf
CHAPTER 6 REVIEW
PROBLEM SET 6.3, Page 413 Practice Problems, Page 429
13. E 14. A, B, F, and G
5. 12,750 ft-lb 7. y ft-lb 15. F and G 16. C and D
9. 6,500 ft-lb 11. 2,000 ft-lb 17. A and F 18. B and G
A -7 4 Appendices
19. , 3 2 b , 2^r c. 128τr 51. 62.4B [∣DΛ + y ] , where h = ∣ V 4 √l2 - B 2

3
13√T3 8 53. 2√2ττ 55. 235.5ιr
20. 21.
27 27 57. ( ∣>∣) 59. a. 51.7% b. 17.44 yr
f 1 [ln∣x + l∣ + l]
22. g'(53z2 - 1) ≈5.33 23. 7
61. Approx. 12,190,000; 35 yr
24. k mi.n = 1°.3 I, b/m
gaaxl; k = 2 Ib/gal; t ≈ 238 min
63. 2π(3 - √5) 65. ∣
67- T⅛ h 2 - ' ) 69. 30 + g l π2
Supplementary Problems, Page 429 CUMULATIVE REVIEW, Page 435

■ 81 7Γ ι 8τΓ
• 2 45 5. ∣ 7. ∣ 9. 0 11. 1
64 7 ( 1 2∖
5. Λ 2 ’ 5) 13. 1 15. 6(2x - l) 2(x 2 + l) 2(3x - 4)
9. 250 in.-lb 11. 24,047 ft-lb 19. - 6 csc2 3xcot 3x

(x + cos x) z
13. ∣ττα 3 15. 2τrιiι 10x + 3
71
21∙ 5x 2 + 3x - 2 23. 5
17. 4,680 ft-lb 19. 30,264 ft-lb
25. √10 - 3
21. 192πlb 23. 9,360 lb
29. 1.812
25. 2,662.4 lb
27. a. mg(s - s0) b. P = mg∖v0 t - ∣gZ 2,
Λl = ∣m(v 0 - gt)2
29. F(Z) = A + (∕>0 - Λ)e* s in '

31. ( 2 , ∣ ) 33. τr⅛2 ( ⅛ - r + j α
∖ ’ 3/ ∖3rz2 3 24√3 - 32√2 + 27
35. 17,500 ft-lb 37. 135 1b
2V2(τr - 4)
F(Z) = ∣Z + j 41. 6,176ττ ft-lb 35. y =
39. TΓ2
7 + Cx 39. y = ln(cos x + e 5 1)
43. 10.34 cm, 11.12 cm, and 5.66 cm 37. y =
1 *1 3 ∙
4 ≡ -τ ⅛ - 0 1 ra d 's
45. x(y) = ye y + C 47. y 2r ft-lb 41. e -
49. 0.6431 45. 79.6 lb salt
CHAPTER 7
PROBLEM SET 7.1, Page 446 19. Formula 174;

y x V x 2 + 1 - ^ln∣x + V x 2 + 1∣ + C
1. ι< / 1 + C 3. l( ln x) 2 + C
16x(x - 1) 2'
21. Formula 173; 4 V 4 x 2 + 1 + C
5. esin v + C 7. ∣ t a n - 1( j J + C
23. Formula 492;
9. ln(x 4 - 2x 2 + 3) + C 11. ln∣x2 + 4x + 3∣ + C
- e - 4 -v(4 sin 5x + 5 cos 5x)
√ r 2 _ „2 41 + C
13. Formula 201; a + C
2
a~x
ln∣⅛x + 1∣ (x + l) 4(4x - 1)
„2 / 1∖ f 1 U- 2Λ ∕ ∙ rs∕-χ ∣
15. Formula 502; y l In ∣x∣ — I + C b 20
29. ⅛ - ⅛ + C 31. s i n - '( ⅛ p λ) + C
oa x l 1∖ 4 1 0 ∖ 3 /
17. Formula 484; — x ---- + C
a ∖ a 33. x ln 3 ∣x∣ - 3xln 2 ∣x∣ + 6xln∣x∣ - 6x + C
35. sin^ 1( j J + C PROBLEM SET 7.3, Page 463
9 ln(V9 + x 2 + x) x√ 9 + x 2 1 — — 4----- ------ 3 3 --

J !∙ ry 4^ ry 4" C. 3x 3(x - 3) x
5 4 8 7 - ± + 3 ______ +
39. √ 4 x 2 + 1 - In 11 + λC4χ2 — I + C ’ x 2x + 1 ’ x2 x (x + 1)^ x + 1
I 2x i
0 ___ 4_ , 17 _ 13 , 10
41. —I sin 5x cos x - ⅛ sin3x cos x + ~ r x - ⅛ sin 2x + C 9x 2 27x 9(x + 3)2 27(x + 3)
0 Z4 10 3Z
. , x(9 - x 2)3z2 J L 27X V 9 ^ T 2 243 . √x∖ „ 11 — ------- 1-------------- 1
------ -— —
2
43. 4-------+ -------- g------- + - r sin >(j J + C 4(1 - χ ) 4(1 + x) 2(1 + x )
13. - ^ ln ∣x ∣ + ∣ l n ∣ x - 3∣ + C
47. ½ + c
15. 3x — ln∣x∣ + C 17. 41n∣x∣ - 41n∣2x + 1∣ + C

aq x _ sin 4x , r sin x c s 3x , sin x cos x , x , r
8 32 4 8 8 19. —-∣— 9 ln∣x∣ + 5 ln∣x - 11+ 6 ln∣x + 11+ C
55.
Γv-l/2
4^¾ - x ιz4 + ln(x lz4 + 1) + C
r
21. x —∣ ln∣x + 2∣ + ∣∙ ln∣x — 1∣ + C
57. 4 x ιz2 - 3 x ιz3 + 3x ιz6 - ∣ln (2 x ιz6 + 1) + C
23. x + i ln∣x — 1∣ - ln∣x + 1∣ - tan - 1 x + C
59. v + x + C 61. square units

25. ln∣x + 1∣ + + C
63. ^ r ( 9 - In 10) ≈ 94.6825
27. - ∣ l n ∣ x ∣ + ∣ ln∣x + 1∣ + ∣ ln∣x - 2∣ + C
65. ∣√ 5 + ∣ln(2 + √5) ≈ 1.4789
29. -ln ∣x + 1∣ - 2 + ln∣x+ 2∣ + C
67. - π^ln(γ∣ + 2) ≈ 3.08097
1o 32 31. 21n∣x + 3∣ + 3 ln∣x - 1∣ + C
33. Let u = x 3 - x 2 + 4x - 4 to obtain
ln∣x 3 - x 2 + 4x - 4∣ + C or use partial fractions to obtain
PROBLEM SET 7.2, Page 453 the equivalent form In ∣x - 11+ ln∣ x 2 + 4 ∣ + C.
c 37. - ∣ln(2ex + 1) + ∣ ln∣ex - 3∣ + C

1. - ¾ 2 - ⅜~ + 3. ⅛ n x - ⅛ + C
2 4 2 4
5. x sin - 1 x + V 1 - x 2 + C 7. V x ln x - 2 V x + C 39. , , 1-----+ C 41. ln∣tan x + 4∣ + C
1 + cos x 1 1
9. =^- + sin 2x + ∣ cos 2x + C 43. - 2 t a n - 1 x ιz4 - ln ∣x ιz4 - 1∣ + ln(x lz4 + 1) + C

45. j ln∣3 tan y + 11 - ∣ ln∣tan - 3 + C
11. In x - + C 13. --⅛(1 - e*)3z2(3e* + 2) + C
47. -ln∣sin x + cos x∣ + C
15. In 2 - ~ ≈ 4.28245 49. ln∣sec x + tan x∣ - ln∣cos x∣ + C or —ln(l - sin x) + C
51. [ta n ( ∣) - 2] 1 + C
17. e - 2 ≈ 0.71828 19. ∣ ( e 2π - 1) ≈ 133.623
23. ∣cos2x - cos2x ln∣cos x ∣ + C

53. [In13 — In x∣ —ln∣ 1 - In x∣ + C
25. -(2 + cos x)ln(2 + cos x) + (2 + cos x) + C 55. In 2 ≈ 0.6931 57. τr(-⅛L + ⅛ n ( ) ≈ o.O826
∖ 1oU 2/ 0/
n+ *
n + 1 In x - (,« +, 1λ)2
γ
27. y(x - sin x cos x) + C 29. + C
- 59. 128ττ In 2 - ^ 7 7 ≈ 10.6484
31. 4 0 0 - l,000e - 3 z 2 ≈ 177 33. 2π(l - 3e~2) ≈ 3.73217
61. t ≈ 6.4 days; the spy is captured.
35. a. πe —2τr b. f ( e 2 + l)
, —x 2cos x + 2x sin x + cos x

37. (0.6774, 1.2715) 3o
9∙ x2 + 1 PROBLEM SET 7.4, Page 467
., , _ z77" . t'o ~ m S l k 3. i In ∣sec 2x 2 + tan 2x 2 ∣ + C
41. / ( π) = - 3 4~
5∙ v ~ ∕< + ek " m
1' 2x 2(x - 1)2
45. 0.07 47. 1.558wAl 5. ln∣sin ex ∣ + C 7. —ln∣cos(ln x)∣ + C

Appendices
9. 3 lnjsec t + tan /1—2 ln∣cos Z∣ + C 79. 4cos 3 1 —4 cos 1 + 4 ≈ 0.09

o 2 3
11. ∣t a n ^ 1e 2x + C 4 1 n (V 2 ÷ l⅜ -
8 χ l l 2 2 l 9
77
13. x + ∣ In (x 2 + 9) - ∣ tan’ 1 + C
8 5 .4 87. ^ y τ τ ≈ 203.58
15. χ - 2 ln∣e∙v - 1∣ + C 17. ∣sin~'t3 + C 89. (0.71, 0.12)

19. x - y ln(e2-v + 1) + C 21. tan - l (x + 1) + C
23. tan~1^~j ^y∣ + C 25. x tan - 1 x — ∣ ln(x 2 + 1) + C
1
3. 5. Diverges 7. 10
x sin 2
27. ^^ x - ê^∙vcos x + C
1 5
9. Diverges 11. 13.
29. xco s^ , ( -x ) + V l - x 2 + C or 10 2
x sin - 1 x + ry 1 + V l - x 2 + C 1
15. 17. Diverges 19. 2e~
9
31. -co s x + c0g x + C 33. 4 cos 5x —⅜cos3x + C 21. 5el ° 23. Diverges 25. 2
27. Diverges 29. Diverges 31. 0
35. - 4 sin 5x c o s x + ⅛ s in 3x c o s x - ⅛ x —∙X ∙sin 2 x + C 5
O 24 10 32 33. 35. 2 37. -1
4
1 2 1
37. - $ cos9x + y cos7x - $ cos5x + C 41. 2
1
39. 1 43.
4
39. 4 tan 6x + 4 tan 8x + C 45. 100,000 millirads 47. 4
o o
41. V l - x 2 - ln ∣ 1 + v
λ
1 ~ —| + C 49. Converges for/? < 1; I =
51. Diverges 53. 8

43. V2x2 - 1 + ln∣V2x 2 - 1 + √2x∣ + C 55. a. 5 b. cos 2t + sin 2t c. -3/ + 2e t
45. ln∣x∣ - ln ∣V x 2 + 1 + 1∣ + C d. 3 cos 2t + 2 s in 2t e' + ^ e - '
47. 5 sin~1f 2 ^ 2J - 2√4x - 2 - x2 + C
57. ____6____4 b. (s + 3)2 + 4 c 5 fe
aa ∙ (5 + 2)
49. ln∣VT + sin 2x + sin x∣ + C 2f(cost
d. 4e — sin t)
51.77 53. 0.2887
0
5V 16
55. J ^t ≈ 0.1 353 57. -=7 9 2 ~ 64λ^ ≈ 113.43
210 1j 1. 3.6269 3. -0.7616 5. 1.1995
59. - ∣ln(√5 - 2√2 + VTθ - 2 ) - ∣√ 2 + √5 ≈ 1.8101 7. 0.9624 9. 1.6667 11. 0.6481
13. 3 cosh 3x
61. ⅛ - ^ ≈ 0.0266 63. ∣ - ∣ ≈ 0 . 1 1 8 7
15. (4x + 3)sinh(2x 2 + 3x)
j3λγ∙2
1 c o s l ∕ 1 "l
65. ln(V l + e 2x + O + C 17 IQ
x2 ∞sh W V I + χ6
67. ξ⅛∣ tan~1^2 ^λ^ ∣ - ln(x 2 + x + 1) + 3 ln∣x∣ + C
21. sec x 23. sec x
V x 2 + 1 c o s h 1.v - V x 2 + 1 sinh - l x
69. 5 ln∣x - 7∣ + + C 71. 3 ln∣x + 11------⅜-r + C Z3.
V x 4 - l(cosh - 1 x) 2
27. cosh x l
73. 3 ln∣x + 2∣ + 2 ln∣x∣ + 1 + C
29. V v 2 — 1 cosh - 1 y - , 2 ι , - e2⅛inh- 1 x
75. — ln∣x + 1∣ + 2 ln∣x + 2 1—⅞ln∣x + 3 ∣ + C L ' vx + 1J
31. —cosh θ^ + C 33. ln∣sinh x∣ + C
5 + tan 7
77. 13 l n 35. ∣ c o s h i β ∏ + C o r ∣ln ( 3 t + V 9 ∕ 2 - 16) + C
1 - 5 tan j
37. sinh l (sin x) + C o r ln(sin x + V 1 + sin 2x) + C 7. 2sin- , (j^) - ^ √ 4 - x 2 + C

39. ∣tanh^ 1x 3 + C or ∣ In Kj + ? + C
3 6 ∣x - 11
9. - ∣+ 1∏∣2L T^∣ + C 11. 4 x ιz4 —4 ln∣1 + x ιz4 ∣ + C
41. i l n ∣
13. i χ 3tan^ , x - ⅛x2 + ⅛ln(x2 + 1) + C
43. cosh^ , e2 -- cosh~l e ≈ 1.0311 or 3 o o
ln(e2 + V e 4 — 1) - ln(e, + V e 2 — 1) 15. 2sin^ 1^ + ∣ e x V 4 - e2x + C
P2 - 1
45. , ι ∖≈ 0.3808
2(e22 + 1)
53. b. v = cosh l x + sinh 2x 17. -⅛∙( 1 + Λ ] ( 2 - -⅞ ) 19. -2 V 1 - sin x + C
15∖ x 2) ∖ χ -)
55. ( e - e ~ , )a 57. f (e2 - e~ 2 + 2) ≈ 17.68
21. ysin(ln x) - ycos(ln x) + C
23. ln(e2 ' + 4e v + 1) - x + C

CHAPTER 7 REVIEW
25. y + y cot - 1 x - ∣ l n ∣ l + x 2 ∣ + C
27. ∣In ∣ x 3 + 6x + 1 ∣+ C
14. a. 0.5493 b. ∣ c. 0.5493
29. 2 cosVx + 2 + 2 V x + 2 sin V x + 2 + C
15. 2 V x 2 + 1 + 3 sinh - 1 x + C or
2 V x 2 + 1 + 3 ln(x + V x 2 + 1) + C 31. iln ( x 4 + 4x 2 + 3) + C
16. - ∣ x cos 2x + sin 2x + C .V 2 -

33. √5 - x 2 + √5 In — + C
17. - 1 cosh(l - 2x) + C 18.

sin-'(f) +c + 2
35. — -3 -(x - 8) + C
2 -1
19. y In x - 1 + i ln(x + 1 ) + ⅛tan x + C
37. T 39 1 _ 2L 41. -tan ~ 1(cos x) + C
2 4
2 2
20. ⅛x + -∣ In ∣x — 11+ C 43. a. True b. True 45. 6
∣ ( l n j + y) ≈ 0.1144 47. For n = 4, L( 1,000) «3 2 1 ; for n = 20, L( 1,000) «201;
21. 61n2 - ∣ ≈ 1.9089 22. and for n = 10,000 (by computer), L( 1,000) ≈ 177
49. (0.647, 1) 51. 2.848
23. y l n ∣ = -0.2027 24. 2 v τj ~ 1 ≈ 0.6095
53. 1.5085 55. ⅜^ π tan '(⅛)
1
25. 26. Diverges
4
πe 2 —~r∑τre 2 + ~rrπ ≈ 23.3832
1
27. Diverges 28. 2
6 '. 3∣α∣ + ⅛ l ∏ 2
1
29. 30. 6τr - 2ττ√5 ≈ 4.7999
(1 - 4x 2)√tanh^ 1 2x x(9 - x 2)5z2 1 5 x ( 9 - x 2) ιz2 405V9 - x 2
65' 6 + -----8-------- + -------- 16------
Supplementary Problems, Page 485
1. 2 71. Diverges
1- x
5. sinh x + x cosh x + (ex - e v)cosh(ex + e λ ) 73. See American Mathematical Monthly, Vol. 87.
CHAPTER 8
PROBLEM SET 8.1, Page 499 11. 1, 3, 13, 183, 33673 13. 5 15. - 7
5 1 i ,l ,i > i 17. 0 19. y 21. 4 23. 0

3. 0, 2, 0, 2, 0
’2 3 4 5
25. 1 27. 1 29. 1 3■M
1 . -2
7z 4 7 z7 13 16 9. 256, 16, 4, 2, √2 33. 1 35. 0
∙ 3’ 4’ > 6 ’ 7
A -7 8 Appendices
37. Bounded below by 0; decreasing 21. Converges 23. Converges 25. Diverges
39. Bounded below by 4; decreasing 27. Converges 29. Converges 31. Converges
41. Bounded below by 0; decreasing 33. Converges 35. Diverges 37. Converges
43. Diverges by oscillation 45. Sequence is unbounded 39. Converges 41. Converges 43. Diverges
45. Converges 47. Converges 49. Converges
47. 6.25%, lθ θ ( ∣) " %
51. Diverges 53. Converges 55. Converges
49. A = 99 51. N = 1,000

3. Converges 5. Diverges 7. Converges
3. 5 5. 3 7. Diverges 9. Converges 11. Diverges 13. Converges
9 ∙⅛ 13∙
11. * 1 ≈ 4 .5 l6 7 -⅜ 15. Converges 17. Converges 19. Diverges
eυ∙2 - I
21. Converges 23. Converges 25. Converges
1 5 .1 17 — 19. 2(2 + √2)
63 The test used in Problems 27-44 may vary.
3√2 + 4
7l 23. I 25. 1 27. Diverges; direct comparison test
2
29. Converges; direct comparison test or ratio test
27. Diverges 29. I 31 —
99 31. Diverges; ratio test
33 — 35. a. A = B = I b .1 33. Converges; ratio test
37
35. Converges; limit comparison test or ratio test
37. 0 39. 9 .14 41 583
120 37. Diverges; direct comparison test
43. - 45. 2 47. 1,500 rev 39. Converges; root test
a
41. Converges; direct comparison test
49. 3 ft 5 l. $10,000 53. 30.8 units
43. Converges; root test
59. b. 3 ------ -
55. 33 trustees λ 1 n 65. 4 45. 0 < x < 1 47. -1 .5 ≤ x ≤ 0.5 49. x > 0
4« 2 + 8« + 3 4
51. 0 < x < -
3. Converges 5. Diverges 7. Converges
9. Diverges 11. Converges 13. Diverges 5. Conditionally convergent 7. Divergent
15. Diverges 17. Diverges 19. Diverges 9. Absolutely convergent 11. Conditionally convergent
21. Converges 23. Converges 25. Converges 13. Absolutely convergent 15. Absolutely convergent
27. Converges 29. Diverges 31. Diverges 17. Absolutely convergent 19. Conditionally convergent
33. Converges 35. Converges 37. Diverges 21. Conditionally convergent 23. Conditionally convergent
39. Diverges 41. Converges 43. Diverges 25. Divergent 27. Conditionally convergent
45. Converges 47. p > 1 49. p > 1 29. Conditionally convergent 31. Absolutely convergent
51. p > 1 33. a. S 4 = ∣ = 0.625; ∣E4 ∣ = ⅛ ≈ 0.00833
53. b. N > 100
b. 0.632
c. N > 100 gives an actual error of 0.009 95 (compared
with a guarantee o f 0.01) 35. a. S 4 = ≈ 0.08230; ∣E4 ∣ = y ∣θ ≈ 0.001372
b. 0.083
37. a. S 4 = - ∣⅞ ∣ ≈ -0.1664; ∣E4 ∣ = y - jL ≈ 0 .000 32
3. Geometric, converges 5. Geometric, diverges
b. -0.167
7. p-series, diverges 9. p-series, converges
39. -1 ≤ x < 1 41. - 1 ≤ x ≤ 1
11. Geometric, diverges 13. Converges
43. -1 < x < 1 45. ∣E ∣ < ⅛
15. Diverges 17. Diverges 19. Converges
47. ∖E , 11 < ⅛ 33. ∕( x ) = cos ? - I x - ? )sin ?

' 3 ∖ 3/ 3
49. Converges (it is a series o f positive terms)
55. a. Diverges b. Converges absolutely
S7. ⅝ = 1
3 5 . ∕( x ) = x + y + ⅛ + ⅜⅛ + -
PROBLEM SET 8.7, Page 547 37. ∕( x ) = (x - 2) 3 + 4(x - 2) 2 + 5(x - 2 ) - 3
1. R = 1; ( - 1 , 1) 3. R = 1; ( - 1 , 1)
'—∞ ox
Γ ∣^ 7 „ _ 4√ 13,
5. R = ∣m 7 ∙ K ~ 3’ ∑ T
9. R = 0, x = 1 11. R = 2; ( - 1 , 3) 41. ∕( x ) = ∑ ( - j)*(-v - 2)*

+ f)
13. R = 3’ 15. R = 7; ( - 7 , 7) k=0 ×
-2 (x -2 )] + [-j( Λ -2 ) ]'
= 1+
17. R = 0 , ; c = 0 19. R = 1; ( - 1 , 1]
21. R = ∞∙, 1(—∞5 ∞) 23. R = 0, x = 0
+ [⅜ -2 ) ] + -
77 R = -^ -. ( ^ L .^ L ∖
25. R = =0; 1( — ∞ 5 co) 5 ∑ 3 ' l5 ^ 3 5 W
43. 1 + i x — 4 χ2 + ⅛χ3 - T⅛χ4 + "4(~ 1, 1)
2 0
29. R = 1; (- 2 , 0) 31. (0, 2)
45. 1 + ⅛ - =∣ 2 4γ^3+ - l ( - l , l )
- + ⅛
33. 35. ∑ k 2 ~ kx k ~'
∖ a a) k=Q
X ∞ 47. X + y + 3 χ ! + 5 x Z + .... / _ [ n
37. ∑ k (k + 2 )x t 1 χ k+1
8 16 ’ , ’ υ
39. ∑ k
k=0 2 (k + 1)
Ay -2k+ 1
fc=0
c i ∖∞y
41∙ Σ k + ∖ ^k+λ - ^ 0 ( 2 ⅛ + l) !
k + 45. ∣x∣ < R',P 55 z w +
1
+ ,
k=0 K ψ 1 ∙ - ∑ 4 ∑ r χ U 2* , P
47. x > 1 49. R = ∞
PROBLEM SET 8.8, Page 562 57-'+>- I ∑ ÷ 1Λ - ∑ ⅛ ÷ 1k

0
“ (2x) jc ∞ x 2fc
3. Z t-ι 5∙ Σ IT cn f ( γ ∖ 4 , 1 1
k=0 κ " k=0 = > .J U ) 3(x + 2) 6(X - 1) 2 ( x + l)
X ( - l ) ( α x ) 2 i+ 1 = I ⅛ ( 4 H ÷ ⅛
∞ ( - i) ⅛ 4t+2
7. 9∙ Σ
4b (2 ^ + 1 )! k=0 (2x + 1)!
” (—l ) λ (2 x 2) 2λ ∞ ( - ∖) k x 2k+2
61 ∙ 2 J ( ( 2 fc + l) ! 63 ∙ 2 ∑√ (2fc)!
11. 13. ∑ o
4b (2⅛)! k=0 (2⅛)!
X
15. x 2 + 2 x + 1 17∙ ∑ ^ T Γ 65. ∑ [ ( - l) * 2 * ÷ 1 + 2 * ÷ 1 + l ] f —
k=0 K ψ 1
k=0
“ Γ.√ , ( - i) ⅛ u + '1 _ Y2 Y4 + «■ ∑o (H⅛ <⅛ sw-'

19. ⅛ Lfc! + (2 fc+ l)! J 1 + 2 x + 2! + 4! "
0
21. ∑ ( - 4 ) ⅛ λ 23. ∑
( - n ⅛ 6,. ∣,82im ∣.
', + il '
∣∞ * y
(2w + 2)!
s j y ς ,
(2/7 + 2)!
f0 r 0 s „ s x
k=Q k=0
I , 7 r 5∖
lr , Q + v ( ~ 1 ) λ' X jc+1 71. a. P 5 ∣ x - 2 x 3 + = ∣ - I
25. 3 3k + l k + 1
b. < 0.000 70 on -⅜ 4
” (—V)k 2 2 k + l x 2k+l ( - l ) * x 2 *∙-
27. 29. ∑
k=0 ^k + ’ k=Q k∖ c. Sam e as above
( * - 1) ∣ e { χ - 1)3 d. T h e actual m axim u m error was 0.000 682 2, so the
31. ∕( x ) = e + e (x 1) +
2! 3! preceding estim ate was excellent.
A -8 0 Appendices
73. ∑ ∣(⅛ + 2)(k + l) x k 75∙⅛ t⅛ ⅛ 41. Diverges 43. Converges 45. Converges
⅛=o 2 47. Diverges 49. Conditionally convergent
51. Diverges 53. Converges conditionally
CHAPTER 8 REVIEW 55. Converges absolutely
57. Converges conditionally
Practice Problems, Page 565 59. ( -3 , 3) 61. [ - 1 ,1 ]
31. 0 32. - ∣ 33. e 63. [-1 - √ 2 , - 1 + √2] 65. [0,1]
67. (—∞, ∞) 69. [ - 1 , 1]
34. Converges to e e - 1 35. Diverges (divergence test)
i+ 2 k k
36. Diverges (integral test) 71. y (-3 )⅛ 73. y (In 3) x
⅛⅛ k∖ k=o k∖
37. Converges absolutely (p-test)
38. Convergence interval is ( -1 , 1) 75. ∑ [ 2 ( - 3 / + 3] x k
k=0
“ ( - l ∕( 2 x ) u+1
⅛ (2 λ '+ l)! 410,993
77. e - ' l2 ≈ 0.6065 79.
33,300
γ3 OY^5
Supplementary Problems, Page 566 81. tan% = x + y + η y + > W < f
1. Diverges 3. Converges to e 2 γ2 γ4 x 6 _
ln(cos x) = —y ~ T2 - 45
5. Converges to 1 7. Converges to ∣
85. 0.396 884 13 89. 0.048 790
9. Converges to 0 11. Converges to 0 91. 8 terms for error < 10 6; with 3 terms the approximate
13. Converges to ea, 15. Converges to 0 error is 0.0004.
e t>ι 232 93. ∣x∣ < 0.187
17 - 19. 23 — — —
3 3- e 77 9(6 - e) __ n(n - 1) , fl(l - rn)
95. a. a n H-------— -d b.
1 1- r
25. - 27∙3
In 2
c. ~~— fo r ∣r∣
1
< 1
29. Diverges 31. Diverges 33. Converges 1- r
35. Converges 37. Converges 39. Converges 97. See American Mathematical Monthly, Vol. 84.
CHAPTER 9
PROBLEM SET 9.1, Page 575 25. x 2 + y 2 = (x 2 + y 2 + y) 2
Polar form Rectangular form

-ι, ∖4 7I ∣, ∣ _ 4 5ττ ∣- (2√2, 2√2)
∙ ∙ p 4 J ∖ ’ 4 /’
f _ 5 5√3 λ∣
∖ 2’ 2 /
( 3 7 π∖ ( _ 3 τr f -3 √ 3 _3 \
\2’ 6 ) ' ∖ 2’ 6., ∖ 4 ’ 4J
9. (l,77), ( -1 ,0 ) ; ( -1 ,0 )
21. a. United States b. India c. Greenland

d. Canada
39. Yes 41. No 43. Yes 45. Yes

47. No 49. No 51. No
Answers vary in Problems 53-59.

, √232 - 42√2 - 42√6 ......
61. a. d = ------------- -------------- ≈ 4.1751
2
b. Vfj + r2 - 2r 1r2cos(02 - 01)
5. a. Rose (4 petals) b. Lemniscate

c. Circle d. Rose ( 16 petals)
e. None f. Lemniscate
A -8 2 Appendices
19. (0, 0), ( y ’ 2 77 - sin’ 1 ⅛) or (3.2, 5.356)
∣ 2 s in ⅛ ⅛ ∣ ≈ (1.85, 1.96)
∖ 0 0 /
j *z,∙ 4 j t 4
57. 477 59. 4π + 12√3
,, . n . 1 - COS θ 1
63. a. -c o t θ b. — ■— 7— c.
sin θ 3

5. (0, 0). (4√ 2, 7. (2, ∣ ) , (2,
9 .( 2 ,^ .( 2 ,^ ) ll.( 0 ,0 ) ,( 2 ,0 ) ,( 2 ,τ τ )
13. (0, 0), (2, ∣ ) 15. (2, 0), (2, 77)
17. (0, 0), ( ^ ∙ ≈ ), (⅛ 2∙ ⅝ )

15. (x - 1)2 + (y + 2)2 = 1
dy _ -, ι2 d 2y
2t ∙ —-2
dx ’ dx
dy _ cos 2t. ⅛ = -b co t t. dy = - b
dx 2e4t ’ dx a ’ dx2 α 2sin 3l
1
d y _ -s in 2 1 - 2 cos 2t
dx 2 4e st
31. ^ = 2 ( x - l ) 33. ^ = e 2t 35. ∣

dx v ’ dx 3
37. 39. ∣ ( 1 - I n 2) 41. ⅛ (223/2 - 8)
o Z Z/
25. y 2 = lθ ( x - ∣
43. 4 + ∣ l n ∣ 45. ln(4√ 3 + 7) 4 7 .- ⅛ 16α2 - x 2)
£ £ 10(2
49. 12 51. y 53∙
59. a. 15.865 b. 161.557 c. 126.645

61. x(θ) = ∕(θ ) cos 0; y(0) = /(0) sin 0
27. (y - 2 ) 2 = - 1 6 ( x + 1) 29. (x + 2)2 = 24(y + 3)
31. (x + 3)2 = - ∣ ( y - 2)
33. y 2 = -1 2 (x - 3) 35.
A -8 4 Appendices
45. 4r 2cos2 θ = rsin θ - 3 47. (0, 0)

49. Tangent: (y + 2) = - ( x - 1)
perpendicular: (y + 2) = (x - 1)
51. 3x + y = 5 53. x 2 = ~6(y + 3)
55. m from the vertex 63. A = 4c2
43. 5x - 6y + 28 = 0
47. ^τrab2 49. 8ττ2
(x - x n)- (y - y n)2
53. ------ —----------- 75-----
2
= 1, and x ≥ x n + a
a- b
55. 12√3 - 6 ln(2 + √ 3) ≈ 12.882 8
57. a. 24τr b. 24√37r
59. Maximum distance is 9.46 × 107 mi;
minimum distance is 9.14 × 107 mi.
CHAPTER 9 REVIEW

27. a. Parabola b. Hyperbola
c. Circle d. Circle
e. Line f. Parabola
g. Four-leaved rose h. Cardioid
28. Parabola 29. Circle
Supplementary Problems, Page 627
Horizontal tangent at (2, 3ττ∕2)

29. r 2 —3rcos θ = 5 31. r2 = θ
33. tan θ = m 35. r = 2α cos θ
39. x 2 = -2 0 y
(x - 6)2 y 2 _
41. 43. 4 5 1
16 48
4 (x ~ ⅛ ) 2 (y -3 )2
(x + 5)2 (y - 4)2
45. 47.
1 3 25 4
49. (0, 0), (2, 0)
51. (0, 0), ( ⅛ 5 '⅜ )
53. (2.24, 0.464) 55. (1, (1,
57. (0, 0), (a, 0), (a, π) 59. 8τr
65. ⅛(e4 - - 1)
o
69. 0.905
A -8 6 Appendices
73. a. y = x - a∖½ - 2 j b. y = 2a c. x = 2aπ 9 1 4(a - b)b

89.
1 9 16 a
7 5 .(1 9 ,2 ),(3 ,0 ) 77. 3ττα2 (x - 5)2 y2
93. = 1
0.1736 ^^ 24.8267
83. -1 85. ~ 9 ~
4√2 95. See American Mathematical Monthly, Vol. 83.
3—
87. a. 2 b. Circle area is TT; cardioid area is - γ .
CHAPTER 10
PROBLEM SET 10.1, Page 637 5. x 2 + y 2 + z 2 = 1

7. x 2 + ( y - 4)2 + (z + 5)2 = 9
9. (0, 1, -1); 2 11. (3, - 1 , 1); 1
13. B 15. E 17. A 19. G 21. I
23. Isosceles, but not right
15. ∣i - ⅛ 17. 5 = 6, r = 24 19. 5 = - 1 , t = - 5

5 5j
21. 1 7 i-1 8 j 23. 3 5 i- 3 5 j 25. x = 3, y = 2
27. x = - 1 + Vz7, y = 1 + V 7 or x = -1 - V 7 , y = 1 - V 7
2 9 .⅛ b + ) )
35. (3, 10) 37. (3, -5 ), (7, -3 ) 39. ∣

∣
v∣
∣= V COS 2 Θ + sin 2 0 = 1
43. a. Points on the circle with center (x0 , J ∙,G ) and radius 1
b. Points on or inside the circle with center (xfl, v0) and
radius 2
45. a = - 2 t , b = t, c = t, for any t
47. - 6 i + 3j
49. Horizontal component is 60 cos 40o or 45.96 ft/s; vertical
component is 60 sin 40o or 38.56 ft/s.
( ∣ - 5 √ 3 ji - (13 - ∣ √ 3 j j
51.
PROBLEM SET 10.2, Page 0 0 0

Ellipsoid; traces are ellipses.
43. Hyperboloid of one sheet; PROBLEM SET 10.4, Page 663

trace in the xy-plane is an
ellipse; traces in the xz-plane 1. k 3. - 2 i + 4j + 3k
and yz are hyperbolas. 5. - 2 i + 25j + 14k 7. - 2 i + 16j + I lk
9. -1 4 i - 4j - k 11 —
2
√3 15. ^ √ 3 3
13.
2
45. Elliptic paraboloid; trace in (i + 3 j - 2k) 16i + 7j + 9k)
17. v f e 1 9 ∙ Λ ∕⅛ ∑ (
the Λ'j,-plane is a single point V Joo
(0, 0, 0); trace in the yz-plane 21. √26 23. 2√59
is a parabola; trace in the xz-
plane is a parabola. 25. -V z3 2 7 .i√ 3
2
29. a. Does not exist b. Scalar
31. a. Scalar c. Vector
47. Elliptic cone; trace in the xy- 33. 5 35. 8 39. a. No b. No c. Yes
plane is a single point (0, 0, 0); 41. t = 0
trace in the yz-plane is a pair 43. (u × v) × w ≠ u × (v × w); cross product is not associa-
of lines; trace in the xz-plane tive
is a pair of lines.
45. w = 0 47. -4 0 √ 3 i
3. 2x - y + 3z - 5 = 0 5. 3x —2_y + 2z - 12 = 0
7 χ - 1= y + 1
= z + 2,
'∙ 3 -2 5
49.
x = 1 + It, y = -1 - 2t, z = - 2 + 5t
fl 1 U J 9. = = ½-γ½∙, x = 1 + t, y = -1 + 2t, z = 2 + t
51. \3’ 3 3 /
11. 2 l~ = = x = 1 + t, y = - 3 - 31, z = 6 - 5t
13. γγ = 4 η ∏ = x = 1 1Z ’ >= 4 ~ 3 1 Z = - 3 + 5t
3. - 2i ÷ 2j + k; 3
5. - 4 i - 4j - 4k; 4√3 15. x = 3, y = - 1 + t, z = 0
7. -1 6 9. 5 11. Yes 13. Yes 17. (0, - 6 , -3 ) , (8, 0 ,- 1 ) , (12, 3, 0)
15. √3 17. 14 19. - 4 i - 16j + 7k 19. (0, 4, 9), (8, 0, -3 ), (6, 1, 0)
21. V 32 ? 23. No 25. No 27. 11 21. Parallel 23. Parallel 25. (1, 2, 3)
29. 0 31. 1.23 or 710
27. -⅛=> -z ⅛ > - ⅛ ∙ , 1.24 or 71o, 2.08 or 119o , 2.52 or 1440
33. 1.98 or 114° 35. k; 1 √38 √38 √38
37. -⅛ (2 i - ' 3k ); ~7=7 39. ± j 29. 4=> - z τ=- ~ ⅛ 0.79 or 45o, 2.17 or 124o, 1.13 or 65o
’ V13 √2 5√2 5√2
4 l∙ ± ξ⅛ (i - k) 43. s(2i + 3j - 2k) where 5 <C 0 31. -4 = ’ 1.47 or 84°, 1.89 or 108o, 0.34 or 19o
√9T √91 √9T
45. x = - l , y = - l , z = 4 33. 2x + ⅛y —3z + 5 = 0 35. 2y —3z + 6 = 0
47. a = ∣ 49. 35o, 66o, 114° 37. z = 0 39. W 4 i + 2i k
÷ )
.- _2 _ „ 9 . _7
51. a. 4 b 3√7 c. 2 d∙ 2 41. ÷ v ⅛(2i ÷ 4i - 3k) 45. ± ∣( 2 i - 2j + k)
53. 2√6 55. 8 units X | N
47. II II 49. x + 2y + z - 1 = 0
57. 868 lb 59. 25,000 ft-lb o
t 0
l
63. 50 65. v = 0, and w can be any vector. 51. 3x + 5y + 2z —9 = 0 55. 1.18 or 68°
A -8 8 Appendices
χ _ y ~i _ z + i ,0 x _ y _ z + 5 45. - 4 =
ς7 46. ~ γ =
14 2 15 8 -5 19 √30 √26
61. -⅜=> -⅛ > - = = ; 67o, 140°, 121°
47. 48. 168.4 mi/hr
V61 V61 √6T √35
63. The lines intersect at (x0 , y g , z β) and are perpendicular to 49. 75V3 ft-lb
each other.
PROBLEM SET 10.6, Page 677 Supplementary Problems, Page 680

19 1. 8√2 + √14 3. 125°, 26°, 30°
1. 0 3, n 5. 7∙⅛
√10 - x - 1 T + 2
,. z - 3 .
3 1 -1 ’
9. √6 11, ⅛ ⅛ a 13 7
15. - F = x = 1 + 3t, y = - 2 + t, z = 3 - Z;
V 5α 2 + 1 √14 √35
Points may vary: ( -2 , —3, 4), (7, 0, 1).
17. φ 4∣a∣√5
19. 21. 7. λ ∖ , ; = ∣> y = 4; x = 1 + 2z, y = 4, z = t
V14 3 3
23. (x + 2)2 + (τ - 3)2 + (z - 7)2 = 316 Points may vary: (3, 4, 1), ( -1 , 4, -1 ).
4 x —2 _ y 3 _ z+ 1
25. V5,435 ≈ 73.72 0 χ - 3 = T~ = z + 1 11
8 -7 3 14 8 13
27. (3y + z + 3)2 + (3x - 4z - 3)2 + (x + 4y + 3)2 = 650
13. z = 0 15. 3x + 4y - z + 13 = 0
29. - ~ 31. ∣V6
√122 17. 5x - 2γ + 3z - 32 = 0 19. 3x + 2 y - l l z + 19 = 0
21. V14; - i - 3j + 4k; 18i - j - 7k
∖D l ~ P 2 l b .⅛
33. a. d = 23. √38j 6i - 5j + 5k; 7i - 5j
V .4 2 + B 2 + C 2 √6
25. 0; k 27. 6; -21 + 4j + 3k
29. 2; 6i + 3j - 4k
31. a. 14i + 7j + 20k
CHAPTER 10 REVIEW
c- 3 1 + 6j - 6k
34. a. 13i - 12j + 2k b. 1 33 ———> 1 / 3 . , 1 ■, 3 , ∖
35. ± ∖VΓ9 l √ 19 j √19 /
’ V 4 6 ’ √46^ √46
1 2 , ,. τ∙ . 12
c. -∏r(v 3ι —2j)l d. - 7 = 37. √74
13 vι 4
35. a. - 8 b. 14i + 6j + 2k 39. (0, 14, -10 ), (21, 0, 25), (6, 10, 0)
36. a. 40 41. (3, - 2 , 1) 43. 2x —z + 13 = 0
b. Not possible to cross multiply a scalar and a vector 45. ∣√ 497 47. l,250√3
c. 25i - 10j — 15k
49. (∣
∣
v[∣
w+ ∣
∣
w∣∣
v) ∙ (∣
∣
v∣
∣w- ∣
∣
w∣∣
v) = 0
d. Not possible because dot product is a scalar
51. 25°, 65°, 90° 53. x + y + z = a
37. f = j⅛ - = 38∙ 2x + 3z - 11 = 0
55. ∣√ 2 57.
13
r +
39. — y ^ - = z z3- = ^ - 40. 1 7 x + 1 9 y + 1 3 z - 2 5 = 0 59. 2x - 3y - 7z = 0 61. = -LΞL2, Z = 0
42. “ / = ’ ^ 7 = , ~7⅛ 2.13 or 122°, 0.64 or 37°, 1.30 or 74° 65. This set o f vectors is the line x = a + At, y - b + Bt,
V14 V14 V14 z = c + Ct
43. a. Skew lines b. (2, 2, 3) 81. See American Mathematical Monthly, Vol. 46.
44. a. 6 b. V z2 83. See American Mathematical Monthly, Vol. 90.
CHAPTER 11
PROBLEM SET 11.1, Page 690 5. F'(Z) = 2Zi - Z- 2 j + 2e 2 l k-

F''(Z) = 2i + 2z^3j + 4e 2 t k
1. t ≠ 0 7. F'(s) = cos si - sin sj + 2sk
F''(s) = - s i n si - cos sj + 2k
9 . ∕' ( x ) = - 9 x 2 - 2 x ∏ ,g J = v = ^ =
13. V(Z) = i + 2zj + 2k; V (l) = i + 2j + 2k;

∣∣V(1)∣∣ = 3 in the direction ∣ i + ∣ j + ∣⅛
A(Z) = 2 j; A ( l) = 2 j
15. V(Z) = -s i n Zi + cos Zj + 3k;
v ( f ) = - f i + f j + 3k;
∣∣v (4)∣∣ = in the direction + j^ jj +
A(Z) = -c o s Zi - sin zj; A θ ^ = - — i - j
17. V(Z) = e z i — e~ t j + 2e2Zk; V(ln 2) = 2i - 5j + 8k;
∣∣V(ln 2)∣∣ = i∏ the direction
77⅛⅛i ^^ + k; = e'i + e 'i + 4<?2Zk;
A(ln 2) = 2 i + + 16k
19. F'(0) = 2j; F '(l) = 2i + 2j + 5k; F ' ( - l ) = - 2 i + 2j + k

21. F'(0) = i; F'(2) - ⅛ i ÷ l∣ i + g k
23. F , ( j l ----- j + ilk: F'(π) = —I + nk
25. y i - ∣ e 3zj + 3zk + C
25. t ∖ + Z2 j + 2k 27. 2Zi + (1 — Z)j + sin Zk 27. (Z ln∣Z∣ - Z)i - ⅞(2 ln∣Z∣ - l)j + (3z ln∣Z∣ - 3Z)k + C
29. Z2i + Zj + V 9 - Z2 - Z4 k
29. ln∣Z∣ - ^ i - cos(l - Z)j + ∣ k + C
31. (7Z - 3)i - 10j + ^2Z2 - ∣ J k
33. i + 2j - 3k 35. 2√14

33. 3z - 2z2 35. (1 - Z)sin Z
37. - t 2 e t i + (Z2sin Z)j + {2te t + 5 sin Z)k 37. [6z + 2 cos(2 - z)]i + 9 sin 3zj - (3z~2 - 4e>2z)k
39. R(Z) = e t i + ( f - l ) j 41. ∣( 2 √ 2 - l)i + ∣ j
39. (4zez + Z - Z2 + 10 sin Z)i + (2z2e z + l)k
41. t2 e t — t3 e t — 2e t — —sin Z
Z
43. 3i + e 2j 45. 3i - ∣ j + 2k
47. i + <?k 49. 2i - 3j + <?k
1. Tf = 4.6 sec; R = 481 ft
51. all real Z 53. Z ≠ 0, Z ≠ - 1
3. Tf = 129.9 sec; R = 73,175 m
55. t ≠ 0 5. Tf = 1.9 sec; R = 41 ft 7. T f = 1.5 sec; R = 148 ft
9. V(Z) = V 5 u .; A(Z) = 0
11. V(Z) = (2 cos 2z)uf + (2 sin 2Z)uβ
1. F'(Z) = i + 2zj + (1 + 3z2)k A(Z) = ( - 4 sin 2z)ur + (8 cos 2∕)u β
3. F'(s) = (1 + In s)i + Q j j — e ’lln s + -∣Jk 13. V(Z) = - 1 0 sin(2z + l)u. + 10[l + cos(2z + l)u β
A(Z) = —20[2 cos(2z + 1) + 1]uj — 40 sin(2Z + l)u β
A -9 0 Appendices
15. 140 m/s 17. 21.7o or 68.3o 33. Maximum curvature (of 24) occurs at (0, 3), (0, -3 )
19. Maximum height is 35.64 ft; range is 225.93 ft; the
35. At x = 0, p = at Λ^= ± 1, p = 24
distance to the fence is 217.47 ft.
21. 3.45 sec 23. 5.5 sec 37. x = 2 + 21, y = - 1 - /, z = 1 + 21
25. 23.53 ft _ 2 V 9 ∕4 + 9 ∕2 + 1
K 41 R∙ “
27. V(∕) = (2 cos ∕)ur + + ∕)uβ (9 ∕2 612sin (1 + 4 ∕ 2 + 9 ∕ 4) 3z2 ( l + 4 .V ) v2

A(∕) = ( -2 sin t — 2714 - 18∕4sin ∕)ur 2 3
( 1 +W
45∙ κ = √ b
43∙ X 4) M
+ (181 + 12/ sin t + 12∕2 cos ∕)uβ R
29. aw 31. aw~ 33. 1.82 sec; 56.7 ft

14 = ■4 = ----- 2-----
■ τ V l + 412’ n V l + 412
1- T W √4 + 9 ∕2 , + V4 9p J
+
N W - i Γ
L V4- 3+1912 1. + 7 9 ∕2 j ]
V4 +
, t (,Zλ) =
3∙ T
cos 1 - sin 1 . . cos 1 + sin 1 . £ A = — L A — —
------ √ Σ ---- , + ------- √ Σ ------- j ’ τ V2 + 4 P ’ n V l + 212
N(ι) = ± [ s ιn ^ co sf i + s in f ~ co s * j] _ . _ 15 sin 1 cos t . . _ 4
τ V l + 15 sin 2/’ n V l + 15 sin2 /
r 4
c5 τv'∕ ∖ sin t . , cos t . . 1 . 9. A τ = 0; A N = 4
∙ w = ^ √ Γ 1 + √ Γ j + √5k
11. A τ = 2 v J^ ≈ 7.99; A N = 0.4
N(l) = ±[cos li + sin lj]
2 ∏ 2
7∙ τ ι '> - √ Γ1 ⅛ , + √ 2f1⅛ k
13. Pressure at highest point is lb/ft2;
N(,) = i [√f+⅛ i + √ r ⅛ k] smallest value for ω is 4 ft∕s.
15. a. 3,536.5 lb b. θ = 44.7o
17. ω ≈ 14 rev/min
19. A τ = ■. ↑1t - - , A nN = 2
∖∕⅛t + 5
JV 4 ∕,25+ 5c
21. A τ = ∖⅛e t ', A N = ∖' l e l 23. a. A v = — (60τr2 ω2)
b. ω = 4.83 rev/min
25. A = —gj; V = vθcos αi + ( - g ∕ + v0sin α)j
Γ 1
R = (v0cos α)∕ι + I —2^ + (v os ι n α )z ∙∣
v0sin a
/ = ---------for maximum height; at this height,
S
∣
∣V∣∣= Vvθcos2α + ( - g ∕ + v0sin α)2
A o .-...,1 A ., /2 cos α sin α - sin2 α
A r = 0, and A N = v0 ^ -----------------------------
29. About 36,000 km
CHAPTER 11 REVIEW

Length is 2-πVΓθ.
r⅜ rτ, ( '∏"I 1 ∙
^⅜
29. 11 ~-7I — /——ι !3——j,∙ wτ∣ 77 ∣ 3 ∙ . 1 ∙
N ^τr I — i——ι + ι——j
∖3∕ V10 V10 ∖3∕ V ιo V10
31. a. T(τr) = - ⅛ + 4 = k b. κ = ∣ c. L = V2ττ
’ V2 V2
25. F'(∕) = i+ zcosu

f2
s in Z j - (sin Ok 29. e t (t - l)i + ∣Z 2(2 In t - l)j + ∣Z 2k + C
F"(∕) = — ⅛ 2 sin r - 2t cos t - ∕ 2sin t. _ 31. V = i - j; A = 0; ⅛ = √2

+ (c o sZ )k j
(1 + z) t dt
26. (3 - 6 In 2)i + + (β In 2 - 33. V = (sin t + t cos Z)i + e - z (l — Z)j + k;
A = (2 cos t — t sin ∕)i + e~ t(t - 2)j;
27. (√ - Z)i “ (⅛ + 2jj + ( ∣Z2 + 3/Jk ⅛ = Vsin¾ + t sin 2t + Z2cos2Z + e^2z(l x
- ∕), 2 + 1
∕lf
35∙ τ - ⅛ i -⅛ *
28. V = i + 2j + e , (t + l)k
A = e t(t + 2)k
N = - U i+ -U j
⅛ = √5 + e 2t{l + 1)2
at v ’ √2 V2
2 / 3 3 / 4 ∖ 1 /4 ∖
29. T = , , i + , ,j j - . , k 37. T = I — sin H i - 5J + ( 5 c o s ∏k
V 4 ∕ 2 + 18 V 4 ∕ 2 + 18 √ 4 ∕ 2 + 18
N = (-co s Z)i + (sin Z)k
NJ _ — 3 . I ∙ , t .
2z
√ 2 ∕2 + 9 √2 ∕2 + 9 J V2 Z2 + 9 39. A τ = , 4/
7 + /p ≡i
τ √ 4 ∕ 2 + 4 + e2z
3
κ = (2t2 + 9)3z2
Aλ
∣
t1e 2t - 2tel t + 2el t + 4
2√2z 6 N 4 t 2 + 4 + e 2i
τ √2OV9, n √2Z T 9 T
_ 2^Vt2e 2t —2te2t + 2e 2t +^4
κ ~ (4 ∕2 + 4 + e2z) 3z2
30. a. ~ ~ ≈ 9.8 ft
o4
41. A τ = 0; AN = 43/2; κ = ∣
b. Time of flight is ≈ 1.6 sec; range is ≈ 67.7 ft.
43. κ = l 45. κ = -‰ ≈ 1.92
Supplementary Problems, Page 730 V2
1
1. 2k 3. -5 k 47. = 49. t ≠ 1
6
5. i - j 7. F'(Z) = e t(t + l)i + 2zj 2
F"(t) = e t (t + 2)i + 2j 51. F(∕) = (2t - sin t + l)i + ( ^t + cos t - 1 π
9. F'(Z) = - 2 r 2i - 2j + e^z(l - Z)k ∕4 ∣t 2 k
F"(∕) = 4z- 3 i + e~ t (t - 2)k 24τr 2
11. F'(Z) = (2Z + aeat )i + e~a , (l - at)j + aea t + 'k 53. Answers may vary.
F''(Z) = (2 + α2e"z)i + e~at(a2t —2α)j + a 2ea t + 'k
57. a. V = —e^z(cos t + sin t)i + e - z (cos t — sin ∕)j — e~‘ k;

speed = ∣
∣
V∣∣= V 3 e~ ‘;
A = (2e^⅛in ∕)i - (2e- z cos ∕)j + e ~, k
b. κ = ∣√ 2 e'
59. A τ = 0; AN = aw2 61.18,150 1b
x = t
y = 0
23. 0 25. (e - e 1)i + 6k

27. e t (t - l)i + ( ∣ cos 2t )j + j k + C
A -9 2 Appendices
65. Maximum velocity is 11.6 m ∕s. 19. Diverges 21. Diverges

67. Yes 23. Diverges 25. Converges
69. F'(x) = (sin x)i - (cos 2x)j + e^λ k
27. a. 5<T1 b. ∣
(' Λ ∖ ∕≠2 ∖ /,2 ∖
71. F(t) = ( ⅛ + 1 )i + ( ⅛ + 2 )j - (⅛ + 3 )k
29. F'(0 = 2i - 3e-3 z j + 4t3k
F''(t) = 9e-3 z j + 12t2k
73. F(t) = ∣(3 - cos 2r)i
(2 cos 2∕)i + (—2 sin 2z)j + 3k
+ ^[ez(sin t + cos /) - l]j - [3 ln(t + 1 )+ 3]k √13
N = ±[sin 2zi + cos 2t j]
75. See American Mathematical Monthly, Vol. 46. ∞ (—1∖kγ1k+2
77. See American Mathematical Monthly, Vol. 53. 33. a. ∕( x ) = x ⅛ - 2 = ∑ ------
⅛=o ∙
CUMULATIVE REVIEW, Page 736

5. y , = cosh 1x 3— , * 7. i x 2 (ln x - i ) + C
Vx2 - 1 6 V 2/
9. 2 sec x tan x - - ln∣sec x + tan x∣ + C
11. tan - l (si∏h x) + C

13. ∣ [ - i - tan~l ^ ^ + C 15. sin~1(x - 1) + C
37. y = 8x + 10; y = 2
17. 4(x - 5 ) + 14(y - 1) - 2(z - 2) = 0
or 2x + 7y - z - 15 = 0 39. v = 4.3726 mi/s ≈ 15,741 mi/hr
CHAPTER 12
1. a. 0 b. 0 c. 0 d. 2 e. 48 f. 2z3
g. Z4 + t5 h. t - Z2
3. a. 0 b. 1 — e2 c. e~ 2 + 1
d. 2x 3e 2x + 3x 2e2x + 2x
e. e 2 + 2y f. 4z(z2 —2)
9. Domain: x > y ; range: ∕( x , y) > 0
11. Domain: ∣ ≥ 0, x ≠ 0; range: ∕( x , y) ≥ 0
13. Domain: u sin y ≥ 0; range: ∕( x , j∕) ≥ 0

15. Domain: y ≠ 2; range: ∕( x , y) > 0
17. Domain: x 2 + y 2 < 9; range: ∕( x , y) > 0
31. An ellipsoid; the traces in

the planes parallel to the
coordinate planes are ellipses.
33. A hyperboloid of two sheets whose 57. a. Production more than doubles.
axis is the x-axis with vertices b. Production will increase more slowly
(±3, 0, 0). The traces in the xz- (less than doubles).
and xy-planes are hyperbolas. If
c. Production will double.
∣A^∣ > 3, the trace in x = k is an
ellipse. If ∣fc∣ < 3, the trace does 59. Q exactly doubles; Q exactly triples.
not exist. kT
61. P(T, Jz ) = - y ; the level curves are lines through
the origin.
35. A hyperbolic paraboloid. In the 2 2
xy-plane, a pair of lines is 63. P(x, y , z) = + ™ ~ ⅞∩ + 60x + 50y
intersecting at (0, 0). Traces in
planes parallel to xy-plane are
hyperbolas, which open in the
y-direction above and in the 3. 5 5. - 2 7. 1 9. 1
x-direction below. The traces in 11. 0 13. 1 15. 1 17. 50
planes parallel to the xz- and
yz-planes are parabolas. 19. 4 21. 10 23. Does not exist
25. 1 27. -1 29. 1
37. A hyperboloid of two sheets
whose axis is the z-axis with Paths shown in Problems 31 and 33 may vary.
vertices at (0, 0, ±1). The traces 31. Let x = 0, then lim f( x , y) = - 1 ;
(x,y)→ (0,0)
in the xz- and yz-planes are let y = 0, then lim f( x , y) = ±∞.
hyperbolas.If∣fc∣ > l,thetraces (.v,v)→ (0,0)
in z = k are ellipses. I f ∣⅛∣ < 1, Therefore, the limit does not exist.
there is no trace in z = k. 33. Let x = 0, then lim f i x , y) = 0;
let x = y, then lim
fix , y) = »•
2
(Λ∙,J )→(O,OJJ V
Z '
Therefore, the limit does not exist.

35. a. If x = 0, lim -⅞ = 0; if y = 0, lim -⅛ = 0;
(x,y)→ (0,0) yb (χ ,r)→ (0,0) x2
if zy = kx, lim ; = 0.
(x ,y )→ (0,0) 1 + k 6x 4
b. If y 3 = x, lim = ⅛∙
(x,j<)→ (o,o) 2x 2 2
Because ∣ ≠ 0, the function is not continuous.
37. a. 1.1
b. Along the x-axis, y = 0 and ∕( x , 0) → 1; along the
y-axis, x = 0 and /(0, y) → 2.
39. B = 0 43. If V x 2 + y 2 < δ, then (for x = 0)
∣x + y 2 - 0∣ ≤ ∣x∣ + y 2 < 2δ
Choose δ = ⅛.
45. If V (x - 1)2 + (y + 1)2 < δ, then (for x = 1)
∣ξ ~ z - 2∣ = ∣x - y - 2∣ = ∣(X - 1) - (y + 1)∣
≤ ∣x — 11+ ∣y + 11 < δ. Choose δ = ε.

A-94 Appendices
PROBLEM SET 12.3, Page 762 17. 2z 2cos(2x - 3y) dx - 3z2cos(2x - 3y) dy
+ 2z sin(2x - 3y) dz
3. f χ = 3x 2 + 2xy + y 2; f y = x 2 + 2xy + 3y2 ; f χ χ = 6x + 2y; 19. ∕( x , y) = x y 3 + 3xy 2; ⅛ = y 3 + 3y2 ; ⅛ = 3xy 2 + 6xy;
f y χ = 2x + 2y
all are continuous, so / is differentiable.
5∙Λ= 2 ⅛ - - ^ ∕≈ - O⅛ - ~z
21. ∕( x , y) = x 3y 2 l ∕ . = 3x 2y 2 ∖f y = 2x 2y, all are continuous, so
/ is differentiable.
7 /j χ = ---- - ---- ∙, jf = ----- - ---- ∙, j fχ x = -------- - -----■
2
’ 2x + 3y y 2x+3y (2x + 3y) ’ 23. 37.04; actual is 37.085 445 65
25. 0; actual is 0
jy x (2x + 3y)2
27. 2.69; actual is 2.690 696 279
9. a. f χ = 2x cos x 2 cos y , f = -s in x 2 sin y
29. z = 9 31. c. 1(1 - x - y )
b. f χ = 2x cos y cos(x2 cos y); f = - x 2 sin y cos(x2 cos y)
11. f = r- 3% ∙ f = . 2yi 33. The maximum possible error is $1.14.
x
V 3 x + y ’ ' y ∖∕3x 2 + y 4
2 4 35. a. R(x, y) = 4,000x - 500x2 + 3,000y - 450y2
13. f = x (x + 2)ejt+y cos y , f = X 2C Λ + J '(COS y — sin y) b. Estimate the change in revenue to be $180.
y 37. 11% 39. 0.036

x V 1 —x 2y 2 ’ -*, V 1 —x 2y 2 41. Profit is increasing at $24/week.
17∙ fx = y2 + y z '->f y = lχ y + χ z + z i >f = χ y + iy z2 43. -0.0014 45. Not continuous at (0, 0)
z
19. f x = z i ∙,f = 2yz '-,f z = - z 2(x + y 2) 47. Increases by 1%
y
21. f = ------- ⅛------√, f = 2y f = 3z2

jχ 2
x + y + z i j y x + y + z ’ - x + y 2 + z3
2 3 PROBLEM SET 12.5, Page 779
-2 x . _ 6xy 3x 2 + 3y2z 5. a. 8e8z + 8<?2Z
23. zχ 25. z v = - ------- 3z =y -------------- 35-
9z ’ y 2z - 2z - y 2z - y b. (4 + y 2)(2e2t) + 2xy(3e3z) = 8e2z + 8e8'
1 _ ~ 2τ
27. zx = z
X 2x 3 β cos(xz) ’ y xcos(xz)
29. f χ = - ( x + I) 2;/ , = (y + 1)2
, - 3 sin 3( 3 ∕ 6 sin(3t) 3 sin 2(3t) r . .
b. -------τ2 √ - + -----⅛r√ = ----- τ-⅛~ + 6 sin 31
∙3, 11 ∙ - -5 33. jf xx = - jfyy = 6-v cos (3t) cos2(3t) cos2(3t)
35. f χ χ = - f y y = e λ sin y 37. a. - 4 b. 4 9. a. 11- = 2u sin 2v + 2u - 4v

∂u
39. fχy =
f y x ~ ~2xy 3cos(xy2) - 2y sin(xy2)
~ = 2w2sin v cos v - 4w + 8v
41. 2(sin z - x cos z) 43. a. -0 .6 7 σ (Γ — t')m~y 6 ^, ∂v
b. σm~ 0-67 b. ∣ 1 = 2x sin v + 2y sin(-2) = 2ιι sin 2 v - 2(z∕ + 2v)
c. -σ m " αδ7
∂F
sv k dv= _v — = 2xu cos v - 4y = (2M sin V)(M COS V) - 4(M - 2v)
45 a = b
∂T P ∂P P = 2 ι∕2sin v cos V —4(w - 2v)
∂P ∂V dT = _P k
∂V ∂T^ ∂P V P k 11. a. ~ = v2 + v; y - = 2 MV + u
∂u ∂v
47. a. 9 b. 14
b. = -v 2e uv2 + 1 ve"r = v2 + v
49. This does not violate the mixed partials theorem because ∂u x y
fx^fy’ fxy’ fyx a r e n o t a11 continuous at (0, 0). = -2uve uv2 + 1 ιιeub = 2uv + u
Sv x y
51. a. δ = -⅞
a2 ∣, ∂w _ ∂w Sx l ∂w ^y gw ∂z
∂s ∂x ∂s ∂y ∂s ∂z ∂s
PROBLEM SET 12.4, Page 772 ∂w _ ∂w Sx l ∂w Sy ∂w ∂z
∂t ∂x ∂t ∂y ∂t ∂z ∂t
1. 3x + y - vT θ z = 0 3. 2x + 4y —z —4 = 0
ιc ∂w ∂w ∂x l ∂wd y . ∂w ∂z
5. x - y + 4 z - τ r = 0 7. 10xy 3 dx + 15x 2y 2 dy
∂s ∂x ∂s ∂y 3s ∂z ∂s
9. y cos(xy) dx + x cos(xy) dy ∂w dw Sx l ∂w dy gw ∂z
ι ι
V 1 ∂t ∂x ∂t ∂y ∂t ∂z ∂t
11. —— dx + - dy 13. ye x dx + ex dy
x x ∂w ∂w 3x l ∂w dy gw ∂z
15. 9x 2 dx — 8y3 dy + 5 dz ∂u ∂x ∂u ∂y ∂u ∂z ∂u
17. cos(xyz)[-3yz - xze' t + 4xy] 35. ± , , ,=(αi + ⅛j)

V α 2 + b- ’
19. eχ 3 + yz ∖ + 2fy +
37. ± /, -, =¾=⅞(⅛2x∩i 2
0 + α 'yσj)
V⅛4x 2 + α4y 2 "
t cos(rt) _ 2s + t cos(rt)
21. = 39. 1 + √3 41. jξ⅛ 43.
∂r 2 —z 2- z 2 —z
∂w r cos(∕7) (x + y)(2sf) √17 15
=
∂t 2- z (2 - z)2 45. 2V3 in the direction ÷ j + k)
_ (2 - z)[r cos(rZ)] + 2st(x + y)
3 5
(2 - z)2 47. She turns in the direction τμ - yj∙
23. a. - 22 -2 = l z 2 b. 4 - 4
c. 49. 8.06 51. ^ = ( 2 i - 6j)
x y 2 xy xy3
2z 3
2
1 2z (x + z)
2z 2(y + z) y-∖ b∕a
25. a. b. ------- 4----- c.
sec2y cos x 2 2
7τ
(X
Λ O∕
27. a. b. 2z sin x 2- cos x 55. The spy walks along the parabola x = 6 —y j( y ^^ 1)2
4z 3 4z
from (1, 5) to (6, 1), a distance of 6.7 ft; he reaches the
4z3 sec2 y tan y - sec4y
c. hole in about 1 min 40 sec, which is plenty of time for a
4z3 cup o f tea before Chapter 13!
2 ∂2z _
29. ∂ Z2 _ 2 i⅛ . >2 J Z _
∂u dx 2 ’ ∂v2 ∂y2 PROBLEM SET 12.7, Page 802
31. Both the volume and the surface area are decreasing.
5. (2, 3) is a rel. min. 7. (0, 0) is a rel. max.
41. Let u = x 2 + y 2;
9. (0, 0) is a rel. min.
a. 2 f'(x 2 + y 2) + 4x 2∕" ( x 2 + y 2)
11. (√3, 0) is a rel. max. ( - V 3 , 0) is a saddle pt.
b. 2 ∕'( x 2 + y 2) + 4y2∕" ( x 2 + y 2)
13. (- 1 , 0) and (1, 0) are saddle pts; (0, —1) and (0, 1) are rel.
4xy f " (x 2 + y 2)
c. max.; (0, 0) is a rel. min.
, u 2 - ιιv — ux + vy 15. (2- ιz 3 , 2 - lz 3 ) is a rel. min.
47. ux = ------ 5--------------------------------------------
x + y - ux — uy - vx + vy
17. (—2, 0) and (0, 3) are saddle points; (—2, 3) is a rel. max.;
v2 - uv — uy — vx
vx ~ ~->
2--------- 2
1--------------------------------------------- (0, 9) is a rel. min.
x + y - u x - uy - v x + vy
/3 ∖
19. I 5 ’ 2 I is a saddle point.
PROBLEM SET 12.6, Page 791 22 / 1∖
f - z + n i + f i - χ ∖ ∙ 21. Maximum of 18 at (0, -3 ); minimum of —y at I 2, —j )
1. (2x - 2-y)i - 2xj 3. ∖ χ 2 y) ∖χ y 2∕ j
23. Maximum o f 1 at any point on the boundary;
5. e 3^v(i - wj) 7. cos(x + 2y)(i ÷ 2j) minimum of e~' at (—1, 0)
√2 25. Maximum o f 16 at (0, 2); minimum of 0 at (0, 0)
9. Cy+3z (i + γ j + 3x k) 11.
2 27. y = 1.02x + 0.80 29. y = -0 .5 8 x - 0.60
5√2 33. Maximum profit at (20, 20)
13. 15. 0
8 35. Min. at x = y = z = W θ 37. Min. at (2, 2, 8)
17. ±^7∣(1 - j + k ) ; x - y + z - 3 = 0 39. Price the first type at $3,000 and the second at $4,500.
41. y = 0.78x + 0.19; GPA 3.13 y-
19∙ + j- k ) ;x + y - z - f = 0 4--
• ∙
3- 9 ''∙
21- i ξ7 ∣( i —j + k ) ;x —y + z = 0 ,∙' ∙
2-- z ×⅛
23. ± /— (3i + 6j + k); 3x + 6y + z — 3 = 0
Answers to Problems 25-33 may vary.

25. a. 3i - 2j; VL3 b. 3i + 2j; √13 43. a. Y = In y , X = In x , K = In k
27. a. i + j; 27√2 b. i + j; 27√2 b. Linear fit is T = —1.6099 + 1.49996^4; so k = e κ
29. 2α2i + 2⅛2j + 2c¾ 2√α 4 + ⅛4 + c 4 = 0.1999 ≈ 0.2 and nt ≈ 1.5
31. i + 2j; 4 = 33. 10i + 4j + 10k; 6√6 as Kepler predicted.
49. a. relative minimum b. saddle point
A-96 Appendices
PROBLEM SET 12.8, Page 813 Supplementary Problems, Page 816

1 / 5 5 V 25 , √ 6 12∖ 44 1. The closed disc x 2 + y 2 ≤ 16
* ∙∙'V * , 4 ∕ 16 'i - j ∖5 5 ) 5 3. -1 ≤ x ≤ 1, -1 ≤ y ≤ 1
5 ./ (1 , 1 ) = / ( - 1 , - 1 ) = 2 5 ∙ ∕v = L ∕, = ~ 1 7. f χ = 2xy - x ^ 2y cos(y∕x)
7. /(0, 2) = /(0, -2 ) = - 4 f y = x 2 + x ~ 1 cos(y∕x)
9. f x = 6x 2y + 3y2 - x ~ 2y, f = 2x 3 + 6xy + x ^ 1
9. For x between 0 and 2ττ, ∕( - ⅛ ^ , L ⅛ j ≈ 1.8478.
∖ 0 0 /
H r( 2, _ 4 (Λ _ 8
j ∖7 7 7/ 7
13. f ( i . - i . lθ = y ; there is no maximum value.
10 , _ /0 ., 10 ∖
15. Maximum is lθV 3 at
V I ’ √3, √ 3 ∕
17. The distance is 0 because (0, 0, 0) is on the given plane,
regardless o f the values of A, B, and C.
19 ( V . - G
∖3 6 6/
21. Maximum is 324 for x = 3, y = 6, and z = 3.
23. Minimum temperature is y p ∙
25. 80 yd by 80 yd 27. r — 1 and h = 4

29. Spend $2,000 on development and $6,000 on promotion.
31. 10 acre-feet of water and 3 acre-feet of fertilizer
33. The largest box is 18 in. by 18 in. by 36 in.
35. Minimum of jy at L∣> - y ’
15. a. Level surface for c = 1: b. Level surface for c = 2:
37. Maximum o f ~~~ at f y -y>
39. a. Spend $3,000 on development and $5,000 on pro

motion.
b. The profit will increase about $130.61.
c. $4,000 on development and $6,000 on promotion
d. λ= 0
a b c - abc
41. x = —7= ’ y = ∙>anda z = —7=; Vf7 = — 7=
√3 ∙ √3 √3 3√3 17. ∣ 19. 0
CHAPTER 12 REVIEW 21. Along the line x = 0, the limit is - 1; along the line y = 0
the limit is 1; the limit does not exist.
Practice Problems, Page 815 23. Along the line x = 0, the limit is - 1; along the line y = 0
32. (1 - x ⅞ ∕ 3''2 33. 6τr2 the limit is 1; the limit does not exist.
34. a. V ∕( l , 2, -1 ) = i + 3k 25. y e t(t 1 - t 2) + (x + 2y)sec2t
b. - ~ = c. i + 3k; √Tθ
√5
35. Along the line x = 0, the limit is 0, but along the line
y = 0, the limit is 0.5.
36. f x = - ∖∕ x,f y = l∕y, f yy = - l ∕y 2 , f x y = 0
29. = - e x ~z - ⅛ = -e ^ - z
37. ( x + j> + z)2 38. -32(1 + √3) ≈ -8 7 .4 ∂x ∂y
39. Critical points are (0, 0) or any point on the circle j∣ ∂z _ 2xe2- - z 2e 2z - 2
x2 + y 2 = l. ^ ∂x - 2 z - 2(x - y)(2x - z 2)e 2 z '
40. Maximum of 12 at (1, ±V3); ∂z _ ________ 2xe 2z — z 2e 2z_______
minimum o f 3 at (—2, 0) ∂y 2∑ + 2(x —y)(2x — z 2)e2z
-ι-ι ∂Z _ z ∂z _ 2z 57. 2x + 2y —z + 1 = 0; x = 2t, y = 1 + 2t, z = 3 — t

' ∂x 3z + 1’ ∂y 3z + 1
59. Answers may vary; g(x, y, z) = x 2y + y 2z + x z 2
35 f = f = x
x ^V1 —x 2y 2 ’ y V 1 —χ 2y 2 2 2xt 2
61. .2 1 4- „ 2
e‘
f x x = f lι -‰ 3 '⅛ = ° ^ χ 2 y 2 )~ 3'2
63. - ^ = ( 3 + 4 In 2)
∖ λ y )
37. f χ = 2xe χ 2 + y 2 ∙ f y = 2ye χ 2 + y 2 ∙,

f x x = 2(2x2 + 'i}eχ 2 + y 2 - f y χ = 4xye χ2 +y2 65. a. -3 √ 6 b. —6i + j + 7k
67. x = 6, y = 3, z = 3 69. √6
39. f χ = 3x 2^ 2; f
y = ^ ^ - ',f xx = f>xe3yl^ f y χ =
71. a. y = 11.54x + 44.45 b. 102.15
41. f χ = 3x 2y 2V rz + 2x cos(x2 + y In z) 73. 24xy2z 3sin(x2 + y3 + z 4)
f = 2x 3y V z + (In z) cos(x2 + y In z)
7/35. - -3
f χ χ = 6xy 2V z + 2 cos(x2 + y In z) - 4x 2sin(x2 + y In z)
f = 6x 2y V z - sin(x2 + y In z)(2x In z) 77. ± ⅛ i + ∣j -
43. Tangent plane: 16x - 7y - 9 = 0;
Normal line: -1 6 (x - 1 ) + 7(y - 1 ) = 0 or λ ⅛-■= y sin 2(x + z)cos(x + y) + sin 2(x + y)cos(x + z)
79.
and z = 1 sin 3(x + z)
45. Relative minimum at (3, -1 ) 81. Relative minimum of 18 at (2, 3)
47. Relative maximum at I y 3 I; saddle points at (0, 1),(1, 1), 83. Minimum at x = —x2---- ⅛------ τ- >
(α + b 2 + c 2)
and (1,0)
49. Relative maximum at (0, 3); relative minimum at (2, 9); 'l (α2 + b 2 + c 2) an ^ z (α2 + b 2 + c 2)
saddle points at (0, 9), and (2, 3)
89. 0 = 2, r2 = A 91. a. 10τr b. 6ττ
51. = 4x - 12xyt - 6y2
93. 0.64 cm 3 has been removed.
53. ∣∣= e<1'2 - v2>[8xw2 + 4x - 12vx] 95. The sum is V fc
97. See The College Mathematics Journal, May 1991, ρ. 227.
∣ ∣ = eû2 ~v2 ∖ 12yu2 + 6y + 8ιιvy)
55. m = - 1 101. See American Mathematical Monthly, Vol. 45.
CHAPTER 13
PROBLEM SET 13.1, Page 828 y.

9 27 ιπ ι . A1 2 13. 1 sin 32 ≈ 0.1253
2 0
1. 32 3. 24 5. 24 7. 7 15. ∣ 17. 5 19. -j=- 21. 13
0
9. -1 11. 4 In 2 13. 1 15. 8
23. ∣ + 2 In 2 ≈ 2.0530 25. 2e - 4 ≈ 1.4366
17. gγ + ab 3 19. ⅞=- 21. 2
2 r4 r4 -x
t? f 4 f 4 ~y
27. a. xy dy dx = - y b. 1 1 x y dx
23. 3
2 "+2
25 — - -------- -----
1 Jo Jo
(n + l)(n + 2)
l 29. a. J J ey ~x dy dx =■e - 2
- π 31. 1,400 ∣ ∣ rf ∕( x , y) dA
f( x , y) dy dx
33. il~ b. f f ey ~x dx dy + f f ey ~x dx dy = e - 2
37. answers vary from 1 to 3.
JO Jy/2 Jl Jy/2
39. 31. a. J J dy dx = y

b. 2 f 1 f dxdy = l
, 32 5∙ 7. 3π - 2√3
3. 3 3 3
A -9 8 Appendices
f 2V3 f √ 1 6 ^ 8 π i 2√3
33. a. dx ay = -τ- H— z—
Jθ Λ 2∕6 ' 5 5
b , f p d , dx + f , f ' 1 7 ^ dy dx - ⅛ + ψ
9. 16ττ 11. 6τr 13. 4ττ 135.

1 -2
17. i(277 +
3√3) 19. ⅛ - 2 21. 77
23. 25. 77 In 2 27. 2ττα(l- In 2)

29. 0 31. τr(e9 - 1)
33. i(6τr In 3 + 4ττln 2 - 4τr) ≈ 4.2131
35. 8τr 37. ∣
rπ∕4 ra sin θ ι
39. 2 1 1 r dr dθ = g fl2 (7r - 2)
41. 256ττ 43' — 45

⅛3. —
47. ⅜
3 4 8
15π
49. 5311.
4 4

7
1. 4√2T 3. ∣( 1 7 3 z 2 - l ) 5.
4
7. 2√Γ7 + 1 ln(4 + √Γ7)
9. 8√65 + ln(√65 + 8) 11. f ( 5 3 β - 1)
13. 2 T7 15. 877 — 16

17. 15√2 + ∣ln (2 √ 2 + 3) 19. f(3 7 3 z 2 - 1)
D 2VA 2 + B 2 + C 2 25.
3V377α2
λ - 2A B C 4
27. 8 Γ '" / 2a 2 r d r d θ = 4τ7β2
Jo Jo V ^ 2 - r2
29. 2τrrh 31. 36τ7
55. The integral is between 1.17 and 2.31.
33. J J Ve 2x + 1 dx dy
Vτr∕2 _______________
n
37. ∣ J
rv 4 r 2sin r2 + 1 dr dθ
V (2 x + 5y)2 + (2y + 5x)2 + 1 dy dx
39. 4∣M∣V M2 + 1 41. √9w 4 + 1
43. y(6 3z2 - 8) ≈ 7.01 45. 4ττ⅛⅛

PROBLEM SET 13.5, Page 863 r2 π ra 1 1

33. 1 1 r2cos θ(rdrdθ) = ⅞ = ≠
68
3. 45 5. 45 7. 9.
9 π ra r{bla y Ja 2
35 ∙ 4 H ~ χ 2 ,2 7 , √ - _ _ mb2
y1 zdy1 dxv
11. g(5 - e 2) 13. 6ττ — 18 In 2
1 P1 , f 3α 3a 3α^∣ 39. 1 - ∣ e ~ 1 ≈ 0.0803
15. 8 17. 19. 0 21. - e 7
720 Λ 8 8 ’ 8)
1 4ττ 41. 0.0902 43. 0.4731
23. 25. 27. 4π
6 3
45. 3(e - 2) 47. 0
29. Jo Γ Jo f( x , y, z) dz dx dy 2 ⅛
49 53. j ^ ( Λ 2 + ∕ 2 ) 55. 4 π 2α⅛
5
y<∖- z ι r V 4 - y 2- z 2
∏ Jo
f(x , y> z ) dx dy dz
33. 32π 35. 3 2π 37. 15 39. ∣ π r 3 PROBLEM SET 13.7, Page 882
41. ^ττabc 3. a. (1.41, 0.79, 1.00) b. (1.73, 0.79, 0.96)

5. a. (2.45, - 0 .9 6 , 1.73) b. (3.00, - 0 .9 6 , 0.96)
c(l-x∕a-y∕a)
43. dz dx dy 7. a. (2.24, 1.11, 3.00) b. (3.74, 1.11, 0.64)
9. a. ( - 1 .5 0 , 2.60, -3 .0 0 ) b. (4.24, 2.09, 2.36)
c(↑ -x la - y la )
dz dy dx 11. a. (1.41, 1.41, 3.14) b. (3.72, 0.79, 0.57)
13. a. ( - 0 .4 2 , 0 .91, 3.00) b. (3.16, 2.00, 0.32)
c(l -x∕d) rb( 1- x l a —zlc)
dy dz dx 15. a. (1.50, 0 .87, -1 .0 0 ) b. (1.73, 0.52, -1 .0 0 )
Io 17. a. (0.50, 0 .00, 1.00) b. (0.00, 0 .52, 1.00)
a(∖ -z lc ) ('b(l-x∕a-z∕c)
dy dx dz 19. a. ( - 0 .0 6 , 0 .13, -0 .9 9 ) b. (0.14, 2 .00, -0 .9 9 )
21. z = r 2 cos 20 23. 9r 2 co s 2 0 - 4r 2 sin 2 0 + 36z 2 = 0
a(∖-y∕b-z∕c)
dx dy dz
a(∖-y lb -z lc )
dx dz dy
47. xyz 1w1 dx dy dz dw = 14

9

s ∣‰ ∖ 7 Λ 8∖ 9 f l , 108∖
3. zy /. ^ ι, 5 y y. 135y ∏ . (1, 1, 1)
13. (0 , ∙∣4 ) 15. ( 7 , ⅞)

∖ 5ττ∕ ∖ ij
1- f 4 In 2 - 2 In 2∖
7∙ ∖ ln 2 2 ’ 3 )
19. a √ 0 , ^ )
∖ 2T7√ ∖ 3 π 3ττ∕ 2
21∙ (y?+ ^⅛⅛ΞLT )= (4∙97I >O∙762> 23.

32
945
25. (θ , U ) 2 7 .( 1 .3 8 ,0 )
29. b. (0, 0, 0) and I z = ⅛ a 3b 3c 3(a2 + ⅛2)
311 ∙ 64
A -100 Appendices
33. τr 35. a. 1 b. 0 37. r
39. 49 In 7 - - y In 5 - 27 In 3 + 6 ≈ 11.3331
41. ∣(ln 5)(ln 13) ≈ 2.064
CHAPTER 13 REVIEW

23. ~ - 1 + e~v j , 2 24. 0 25. 36
26. ∣sin 1 27. 8τr
√(4 X - X 2)∕2 ∣
∙4X
Π
29. y ∣ "
∣ J x 2 +2y2
dz dy dx = 2πV 2
30. ∣(e2 + 8e^1 - 3)
55. ⅛
59. Because water pours in at the rate of 25 ft 3∕m in , it takes

8.84 minutes to fill the cave to nose level. So he drowns,
you say? Nonsense! Any spy worth his or her salt can hold
his or her breath for a little more than a minute. The spy
frees his hands, stands (to buy more time), hops to the
door, pulls up the lever (to stop the water), and opens the
door.
63∙ ⅝

App. E: Answers to Selected Problems A -1 0 1
49. abc 51. ⅛ 53. √ 2 T7

6 o
55. 4 T7 57. ∣ 59. y ∖ 2 7 - 16√2)

3
61. ⅜ ^ (1 - cos φ)
1Sττk ,nr l,275τr

63. a. — b. 105ττ c.
1 2 3z2
65. ⅛½(H + Λ ) 67. 8
11. i ( e - D 13. 10 15. ^y^(2 - √2)
69. u —2v = 0; 2v - u = 5; v + 2u = 0; v + 2u = 5
17. 31 1ιy9. isi77 ^
21. 12π
30 8
71. a. -2 x y b. c. 1
2xy
IT (^ 2 n + 2 ∖
23. 25. 0 27. 0 1 4(2 - V2)
n+ Γ ’ 73
540 3τr
29. 31. 11 b. (θ, 0, ⅛∖
m - 1 τra⅛
77. a ∙ 2
33. τr 35. 77 37. 877
5 2aVsin~2θ
39. f
Jθ
f
Jz
f
J z 2∕x
f ( x , y, z) dy dx dz
85. (0, 0, 0)
Π fia
pr dr dθ
41. ∣ ( e 8 - 8e) 43. J J f ( x < J-,) dy dx gγ f 1 ~ In 2. 1 - In 2. 1 — In 2∖

' ∖ 4 — 77 4 —77 4 —77/
45. √3 - ? 47. 77 + 3√3
CHAPTER 14
PROBLEM SET 14.1, Page 905 27. 2x + 2y + 2z; 0 29. x + y + z; - y ι - zj - xk

31. yz + 2x 2y 2z 2 + 3y2z 2;
(2yz3 - 2x 2y 2z)i + xyj + (2xy2z 2 - xz)k
33. 3; i + j + k 35. Harmonic
37. Harmonic 39. 6xj - 3yk
41. 3x + 2z 43. 6xyz2 + 2xy 3
45. 0; B is solenoidal.
51. Let ω = ω,i + ω2j +ω3k, where ω1, ω2, and
ωj are constants.
a. (ω2z - ω3y)i + (α>3x - ω1z)j + (ω1y — ω,x)k
b. 0; 2ω
53. II

3 .1 26
5. 1 7
3
9. ». 4 b∙ 4 11 11 . 3 11 33 . —
1 5
11. div F= 2; curlF = 0 13. 0; 0 15 17. 12 19. 12

15. 7 ;j-3 k 17. 2 - 2e“ 6; - 3 i + 3<τ6k
21. a. .
b. τ4 77 „
23. a. 25 .1 1
b.
19. cos x - sin y; 0 21. 0; 0 o 8 2
23. . ,1 5; 0 25. a + b; 0 25. 0 27. 9 29. - ∣ 31. 0

V x2 + y2
A-102 Appendices
9. 0 16 θv ^
33. 35. 2ττ 37. 16 39. ∣ 11. 13. f
41. 42 43. ∣ 45. 27,750 ft-lb 15. ^(3 9 1√ T 7 + 1) ≈ 84.4635 17. 36τr 19
oU 3
2
z ιi . 777∖"2 Z J . -5 2
23 2Z55. — 27. 77
4 8 3
29. ∣(2 √ 5 + ln(2 + √5) 31. -1√]^7 - 1 ln(4 + √17)
5. Conservative; x 2y 3 7. Not conservative o 2 8
9. Not conservative 11. a. —2τr b. —4 c. -2ττ
33. 25√6 35. ¾5 √5 - 1) 37 —
13. a. 0 b. 32 c. 32 15. x 2y ; 1 0 2
17. x 2 - x y + | y 3; - 19. e~x co sxy + sin 2x; 0.1081

21. 23. 8 25. ^ - 2 27. 8 1. 18ττ 3. ∣ 5.-8ττ 7. -1 8
29. 32 31. -2.492
1
9. 0 11. -2τr 13.-3ττ 15.
33 — = = xe x y ' ∂M = — = yve χ y- 2
∂y ∂z x e ’ ∂z ∂x e ’
17. 0 19. -16ττ 21. ∣ 23. 0
∂N ∂M
— = ~τ~ = ze'-'; conservative
∂x ∂y
3 5 .^ = ^ = 2 y z ; ^ = ^ = 2xz; PROBLEM SET 14.7, Page 955
∂y ∂z ■ ’ ∂z ∂x
∂N ∂M ^t ,- 64τr ∙, 128
— = = 2 xy,
z conservative 1 5. 3 7. 24τr
∂x ∂y 3 3
37. ^ = ^ = 3y 2- ,ψ - = ^ = y, o 13 243τr
∂y ∂z
x + x ∂ z ∂ x j y. 3 ι11
ι. 2 13. 0 15. 12τr
τ y = ~ - = 2xy
j + z; conservative 17. ∣ 19. ∣ττα4
∂x ∂y 21. 4τrα4
39. a. j/ = 23. b. 6 25. 0

r
b. K m M ∖^ = = ∖ , - ., 1, , ∕∣
LVα 2 + b 2 + c 2 V α 2 + b 2 + c-J CHAPTER 14 REVIEW
41. 0
45. a. f = x y 2 - x 2z + c(y, z)
23. xyz + C
b. c(y, z) = yz + c(z)
c. f = x y 2 —x 2z + yz 24. div F = 2xy - zeyz ∙, curl F = ye yz i - -j - x 2 k
25. - 2 26. 4ττ√2 27. 0 28. 0

PROBLEM SET 14.4, Page 931 zψ
29. ∕ = ⅛ ∣2 3 0 .0
1. 0 3. 3 5. 0 7. —6τr
9. 0 11. 16 13. 2ττ 15. 4ττ Supplementary Problems, Page 962
τrfl4 1. Conservative; ∕( x , y) = 2x - 3y
17. 15 19.
2 4 3. Conservative; ∕( x , y) = xy~ 3 + sin y
25. 0 27. 0 29. 160τr 31. 1 5. Not conservative 7. -7
33. To the spy’s surprise and relief, he finds that the Death
9. - ∣ y + sin 4 — sin 1 11. 5
Force is a conservative field. It does not matter which 2
path he takes, the work will be the same! Because he wants
13. div F = 3; curl F = 0 15. div F = - ; curl F = 0
to get out of the room as quickly as possible, he walks in a r
straight line from (0, 0) to (5, 4) and leaves. 17. 77 COS 772 19. - 6
PROBLEM SET 14.5, Page 939 21. V2 23. 5

6
1. 8π 3. 0 5. 16ττ 7. 4√2 25. 2 T72 —277 27. 0
29. 4τr 31. 2 33. 2c', 3b 35. 0 However, the spy reneges on his deal and subdues a
5 2abc rather surprised Red Purity. The Green Purity asks,
“ How did you know?” The spy smiles smugly and says,
3 7 .4 39. ⅞ 4 1 .- 1 43. ~
4 2 0 “ The woman who set those traps to kill me knows some
math. You don’t and your sister does.” The Red Purity
45. ln ( jq 7 ^) 47. 3x 2y + 2xy 2 - 5x - 6y = 0 snarls, “ So she wins because Γm smarter. I’ll remember
that next time, and, believe me, there will be a next time.”
51. div F = 0, so F is solenoidal 55. 0
1 + y2 y2
57. a. f ( x , y) = ~ y
CUMULATIVE REVIEW , page 9 6 9
134
b. — rx-; C cannot cross the y-axιs. 5. f x = e TΛ∙(2 x - γ jy = y'x - J χ y = e √ ι - J
1 o y xe
59y 5 + ∣16
' 4 τ in 7 2
7. f = 2y sin x cos x - y sin xy; f = sin 2x - x sin xy;
61. 2πc f = sin 2x - x y cos x y - sin xy
63. 0 65. 0 67. - ∣In2 x 2 - 2xy - y 2 2
= x + 2xy - y _
2
= -4 x y
' jχ ( x - y) 2 j
’ y (x - y) 2 ’ j χ y (x - y) 2
1⅞ r
69. - 73. 75. 2! 49,152
ε 3 11. 13 — 15. e 2 - 3 17. π
13 15
79. The minimum value is maximized when N is chosen to be
19. 8 21. 0 23. x = y = z = VFθ
5. If both Puritys “ play” properly, the Green Purity
chooses N = 5, the Red Purity chooses x = - N ± 1 (that 2 5 .^ + 2 27. ∣[ 1 7 3z2 - 1]
is, —4 or —6), and the resulting head start is approximate
ly 30 min. 29. F is conservative; W = 0
A -104 Appendices
APPENDIX F Credits
CHAPTER 1 p. 212 Journal problem (No. 56) from Parabola, Vol. 20,
issue 1, 1984.
pp. 2, 3 Quotation from “The Calculus According to
p. 226 Journal problem (No. 7) from Parabola, Vol. 19,
Newton and Leibniz,” in The Historical Development o f the
issue 1, 1983, p. 22.
GZ/CM/MS (New York: Springer-Verlag, 1979).
p. 7 Global modeling illustration from Scientific p. 238 Problem (No. 30) adapted from “Flight of Birds in
American, March 1991. Relation to Energetics and Wind Directions,” by V. A.
Tucker and K. Schmidt-Koenig, The Auk, Vol. 88, 1971, pp.
p. 20 Journal problem (No. 75) by Murray Klamkin
97-107.
reprinted from The Mathematics Student Journal, Vol. 28,
1980, issue 3, p. 2. p. 253 Journal problem (No. 35) from the Mathematics
p. 26 Journal problem (No. 41) reprinted from Ontario Teacher, December 1990, p. 718.
Secondary School Mathematics Bulletin, Vol. 18, 1982, p. 255 Putnam examination problem (No. 68), 1938,
issue 2, p. 7. reprinted by permission from The Mathematical
p. 41 Journal problem (No. 66) reprinted from The Association of America.
Mathematics Student Journal, Vol. 28, 1980, issue 3, p. 2. p. 255 Putnam examination problem (No. 69) 1941,
p. 79 Putnam examination problem (No. 49), 1959, reprinted by permission from The Mathematical
reprinted by permission from The Mathematical Association of America.
Association of America. p. 255 Putnam examination problem (No. 70) 1986,
p. 79 Putnam examination problem (No. 50) 1960, reprinted by permission from The Mathematical
reprinted by permission from The Mathematical Association of America.
Association of America. p. 255 Putnam examination problem (No. 71) 1985,
p. 80 Guest essay, “Calculus Was Inevitable,” by John L. reprinted by permission from The Mathematical
Troutman. Association of America.
p. 256 Group research project from Elgin Johnston of
CHAPTER 2 “The Venerable Cooper,” Quantum, May 1990. This group
research project comes from research done at Iowa State
p. 104 Journal problem (No. 65) by Michael W. Ecker
University as part of a National Science Foundation grant.
from The AMATYC Review, Vol. 6, 1984, issue 1, p. 55.
p. 136 Journal problem (No. 50) by Bruce W. King from
The Pi Mu Epsilon Journal, Vol. 7, 1981, p. 346. CHAPTER 4
p. 161 Newspaper clipping from The Wall Street Journal,
June 6, 1986. p. 258 Asian integration from Mathematics, by David
p. 163 Putnam examination problem (No. 74) 1946, Bergamini, 1963, p. 108. Reprinted by permission of Time,
reprinted by permission from The Mathematical Incorporated.
Association of America. p. 260 Software output from Converge by John R.
p. 163 Putnam examination problem (No. 75) 1939, Mowbray, JEMware, 567 South King Street, Suite 178,
reprinted by permission from The Mathematical Honolulu, HI 96813-3076 (808-523-4811).
Association of America. p. 286 Journal problem (No. 63) by Murray Klamkin from
p. 163 Putnam examination problem (No. 76) 1946, the College Mathematics Journal, September 1989, p. 343.
reprinted by permission from The Mathematical p. 289 Photograph of scattering experiment from
Association of America. Scientific American, May 1991.
p. 164 Group research project from Diane Schwartz from p. 307 Quotation from “A Short Account of the History of
Ithaca College, Ithaca, N.Y. Mathematics,” as quoted in Mathematical Circles Adieu by
Howard Eves (Boston: Prindle, Weber & Schmidt, Inc.,
CHAPTER 3 1977).
p. 192 Journal problem (No. 68) from Mathematics p. 308 Graph from The Wall Street Journal, February 26,
Magazine, Vol. 55, 1982, p. 300. 1986.
p. 202 Problem (No. 42) from a paper “The Mechanics of p. 322 Putnam examination problem (No. 73), 1970,
Bird Migration,” by C. J. Pennycuick, published by Ibis III, reprinted by permission from The Mathematical
1969, pp. 525-556. Association of America.
p. 212 Journal problem (No. 55) by Michael G. Murphy p. 322 Putnam examination problem (No. 74), 1951,
from the College Mathematics Journal, Vol. 22, May 1991, reprinted by permission from The Mathematical
p. 221. Association of America.
App. F: Credits A -105
p. 322 Putnam examination problem (No. 75), 1958, CHAPTER 8

reprinted by permission from The Mathematical Associa
tion of America. p. 567 Putnam examination problem (No. 97), 1984,
p. 322 Guest essay, “Kinematics of Jogging,” by Ralph
tion of America.
Boas.
p. 567 Putnam examination problem (No. 98), 1977,
CHAPTER 5 tion of America.
p. 568 Group research project from David Pengelley,
pp. 374-375 Quotation about Bernoulli from Leonhard Marcus Cohen, Edward D. Gaughan, Arthur Knoebel, and
Euler in Elements of Algebra. Douglas S. Kurtoro, Department of Mathematical Sciences,
p. 381 Putnam examination problem (No. 93), 1939, of New Mexico State University, Las Cruces, NM 88004.
tion of America.
p. 381 Putnam examination problem (No. 94), 1940, CHAPTER 9
reprinted by permission from The Mathematical Associa p. 628 Putnam examination problem (No. 94), 1974,
tion of America. reprinted by permission from The Mathematical Associa
p. 381 Putnam examination problem (No. 95), 1951, tion of America.
p. 382 Group research project from Peter D. Taylor of tion of America.
Queen’s University, Kingston, Ontario. p. 629 Group research project from Steve Hilbert, Ithaca
College, Ithaca, New York.
CHAPTER 6
CHAPTER 10
p. 434 Group research project from MAA Notes, Vol. 17, tion of America.
1991, “Priming the Calculus Pump: Innovations and p. 682 Putnam examination problem (No. 83), 1983,
Resources,” by Marcus S. Cohen, Edward D. Gaughan, R. reprinted by permission from The Mathematical Associa
Arthur Knoebel, Douglas S. Kurtz, and David J. Pengelley. tion of America.
p. 682 Group research project from MAA Notes, Vol. 17,
1991, “Priming the Calculus Pump: Innovations and
CHAPTER 7 Resources,” by Marcus S. Cohen, Edward D. Gaughan, R.
p. 477 Journal problem (No. 52) by Peter Lindstrom from Arthur Knoebel, Douglas S. Kurtz, and David J. Pengelley.
the College Mathematics Journal, Vol. 24, No. 4, September
1993, p. 343. CHAPTER 11
reprinted by permission from The Mathematical Associa p. 686 Photograph of double helix of DNA, courtesy of
tion of America. Scientific American, October 1985, p. 60, “DNA” by Gary
p. 487 Putnam examination problem (No. 73), 1968, Felsenfield.
reprinted by permission from The Mathematical Associa p. 705 Photograph of an earth satellite, courtesy AT & T
tion of America. Photo Center.
p. 487 Putnam examination problem (No. 74), 1985, p. 727 Photograph of a Ferris wheel, courtesy Eugene
reprinted by permission from The Mathematical Associa Gordon.
tion of America. p. 732 Putnam examination problem (No. 75), 1939,
p. 488 Group research project adapted from a computer reprinted by permission from The Mathematical Associa
project used at the U.S. Coast Guard Academy. tion of America.
A -1 0 6 Appendices
p. 732 Putnam examination problem (No. 76), 1947, p. 819 Putnam examination problem (No. 101) 1938,
reprinted by permission from The Mathematical Associa reprinted by permission from The Mathematical Associa
tion of America. tion of America.
p. 732 Putnam examination problem (No. 77), 1946, p. 819 Group research project is from MAA Notes, Vol. 17,
reprinted by permission from The Mathematical Associa 1991, “Priming the Calculus Pump: Innovations and
tion of America. Resources,” by Marcus S. Cohen, Edward D. Gaughan, R.
p. 733 Quotation by Johannes Kepler from The World of Arthur Knoebel, Douglas S. Kurtz, and David J. Pengelley.
Mathematics, Vol. I, James R. Newman, (New York: Simon
& Schuster, 1952), p. 220.
p. 733 Guest essay, “The Simulation of Science,” by CHAPTER 13
Howard Eves. This article is reprinted by permission from
Great Moments in Mathematics Before 1650, published by p. 895 Putnam examination problem (No. 89) 1942, re
The Mathematical Association of America, 1983, p. 194. printed by permission from The Mathematical Association
of America.
p. 895 Putnam examination problem (No. 90) 1957, re
CHAPTER 12
printed by permission from The Mathematical Association
p. 738 Computer-generated graph of Mount St. Helens, of America.
courtesy of Bill Lennox, Humboldt State University. p. 895 Putnam examination problem (No. 91) 1958, re
p. 741 Three-dimensional curves, Stanford Graphics im printed by permission from The Mathematical Association
age courtesy of Visual Numerics, Houston, TX. of America.
p. 741 Typographical map, courtesy of U.S. Geological p. 896 Photograph of an Apollo spacecraft, courtesy
Survey. NASA.
p. 742 Weather forcast map, courtesy of Accu-Weather, p. 896 Group research project is from MAA Notes, Vol. 17,
Inc. 1991, “Priming the Calculus Pump: Innovations and
p. 818 Problem No. 97 adapted from “Hidden Boundaries Resources,” by Marcus S. Cohen, Edward D. Gaughan, R.
in Constrained Max-Min Problems,” by Herbert R. Bailey Arthur Knoebel, Douglas S. Kurtz, and David J. Pengelley.
from The College Mathematics Journal, May, 1991, p. 227.
p. 819 Photograph of soap bubbles from Scientific Ameri
can, “The Geometry of Soap Films and Soap Bubblers,” by CHAPTER 14
Frederick Almgren, Jr., and Jean Taylor.
p. 819 Putnam examination problem (No. 99) 1967, re p. 964 Photograph of an earth satellite, courtesy NASA.
printed by permission from The Mathematical Association p. 965 Putnam examination problem (No. 82), 1948,
of America. reprinted by permission from The Mathematical Associa
p. 819 Putnam examination problem (No. 100) 1946, tion of America.
reprinted by permission from The Mathematical Associa p. 966 Guest essay, “Continuous vs. Discrete Mathemat
tion of America. ics,” by William Lucas.
NOTE: Portraits of Newton, Leibniz, Euler, Lagrange, Laplace, Gauss, Cauchy, Hamilton, Fermat, Weirstrass, and Rieman are
based upon portraits from the David Eugene Smith Collection of the Rare Book and Manuscript Library, Columbia University.
IN D E X
A ⅛I ⅛⅛8C X ⅛ ⅛ ⅛ 3 s ⅛ ⅛ S f c ⅛ 5 ⅛ f i ¾ ⅛ 3 ⅞ ⅛ F 7 5 ⅛ i ^ S ^ ⅛ ⅛ s s ^ f^ " ’
parameterization, 712-15 Carbon-14 dating, 417-18

A __________________ in polar coordinates, 604 (Prob. 65) Cardioid, 577
Absolute convergence, 535 Area (in the plane): Cartesian coordinate system, 20, 639
Absolute extrema: by double integration, 832 Catenary (hanging cable), 478
o f∕(x ), 166 by Green’s theorem, 929 Cauchy, Augustin-Louis (1598-1647), 68
o f ∕( x , y), 793 using horizontal strips, 313-14 Cauchy mean-value theorem, A-27
procedure for finding, 171, 799 as an integral, 271 Cauchy-Riemann condition, 818 (Prob. 87)
Absolute maxima and minima, 166 as the limit of a sum, 258-61 Central force field, 707
Absolute value, 10 in polar coordinates, 588 Centrifugal and centripetal force, 726
Acceleration, 114, 698 of a triangle, 661 Centroid:
with constant speed, 725 between two curves, 310 of a planar laminar, 411
normal component of, 722 of an unbounded region, 469 of a plane region, 410-11
in polar form, 706 using vertical strips, 312 of a solid region, 867
summary, 729 Area problem, 5 Chain rule:
tangential component of, 722 Aryabhata (476-550?), 99 justification of, 126
Addition (resolution) of vectors: Asymptote, 202 for partial derivatives, 774, 777
parallelogram law, 632 horizontal, 203, 209 single variable, 121
triangle law, 632 oblique (slant), 203, 209 Change of logarithm base, 341
Agnesi, Maria (1718-1799), 158 vertical, 203, 209 Change of variables:
Alternating harmonic series, 532 Attitude numbers, 670 in double integral, 885
Alternating series: Average: in single integral, 280-81
convergence test, 530 cost, 232-33 in a triple integral, 888
definition of, 529 rate of change, 112 Cheops (Great Pyramid), 395 (Prob. 57)
error estimate, 533 value of a function, 297 Circle:
Ampere’s law, 948 (Prob. 27) equation of, 15-16
Analytic geometry, 15
Angle:
B ___________________ osculating, 718
Circulation of a field, 959
Ballistics, 701-5
between two curves, 136 (Probs. 48-49) Classification of quadric surfaces, 644
Base for logarithm, 338
between two lines, 26 (Prob. 40) Closed surface, 933
Believer/doubter formal, 69
between vectors, 651 Cobb-Douglas production function, 747
Bernoulli equation, 426 (Prob. 41)
Angle of inclination, 21 (Prob. 59), 809
Bernoulli, Johann (1667-1748), 239, 324
Antiderivative: Coefficients of a power series, 544
(Prob. 6), 374, 559
notation, 247 Comparison:
Bessel function, 564 (Probs. 76-77)
of a scalar function, 246 property of integrals, 274
Bhaskaracharya, 205
of a vector function, 699 test for series, 516, 518
Binomial series, 560
Antidifferentiation: Completing the square, 16
Binormal vector, 721 (Prob. 57)
general rules, 248 Composition of functions:
Bisection method, 64 definition, 30-31
by substitution, 280-81 Boas, R. P., Jr. (1912-1992) (essay),
Applications of integration (summary), limit rule, 60
322-24
427 Compound interest, 329-30
Boundary, 934
Applied optimization problems, 219-38 Concave (down, up), 193
Boundary point, 934
Approximation: Conditionally convergent series, 535
Bounded function, 27
by bisection method, 64 Conjugate axis of a hyperbola, 616
Bounded monotone convergence theorem,
linear, 147 Conservative vector field, 916
496 cross-partials test for, 917
with rectangles, 303 Bounded region, 934
Simpson rule, 305 scalar potential for, 916
Boyle’s law, 142 (Prob. 19) Constant:
to the sum of a series, 533, 556-60 Brahe, Tycho (1546-1601), 734
tangent line, 147 derivative rule, 95
Brahmagupta, 4, 501 of integration, 247
tangent plane, 766 Braking rate, 418
trapezoidal rule, 303 limit rule, 49
Archimedes (2877—212 B.C.), 5. 81, 159 Constrained optimization, 805-6
(Prob. 34), 410 c ___________________ Consumer’s surplus, 316
Arc length: Calculus: Continuity, 57, 751
function, 713 differential, 5 of common functions, 59
as an integral, 397-98 integral, 6 of compositions, 60
A-107
A-108 Index
Continuity (cont’d): tangent line to, 84 Dilution problems, 422

implied by differentiability, 90, 768 unit normal vector to, 710-11 Diminishing returns, 196
of a function of two variables, 751 unit tangent vector to, 719-20 Direction angles, 669
on an interval, 60 Curve sketching procedure, 187, 196 Direction cosines, 655 (Prob. 49), 669
one-sided, 60 Cycloid, 598 Directional derivative, 781-82
of vector functions, 690 Cylinder, 643 Directory of common graphs, 35
Continuity equation, 953 Cylindrical coordinates: Dirichlet, Gustav (1805-1859), 422
Continuously compounded interest, 330 definition, 876 Discontinuity, 57-58
Convergence: integration with, 877-78 Displacement, 114
bounded, monotone, 496 Cylindrical shells, method of, 391 Distance:
of an improper integral, 473-74, 844 as an integral, 271
of an infinite series, 501 between point and line 26 (Prob. 39), 675
interval of a power series, 541 D ___________________ between point and plane, 676
of a power sequence, 498 Dam, fluid force on, 409 between two lines, 678 (Prob. 32)
radius of, 541 Dating (carbon), 417—18 in the plane, 14
of a sequence, 492 Decay equation, 416 in space, 714
Convergence tests for series: Decay, radioactive, 416 on a number line, 10
absolute convergence test, 534 Decrease without bound, 204 traveled by a particle, 272
alternating series test, 530 Decreasing function, 183 Distance formula, 641
conditional convergence, 535 Definite integral, 268 Divergence:
direct comparison test, 516 opposite of, 269 of an improper integral, 473, 844
divergence test, 510 at a point, 269 of a series, 501
integral test, 511 properties of, 274 Divergence test, 510
limit comparison test, 518 Definite integration: Divergence of a vector field, 900, 904
p-series test, 514 definition, 267-68 Divergence theorem, 949
ratio test, 523 by parts, 451 Domain convention, 29
root test, 525 by substitution, 284 Domain of a function, 26, 29
zero-infinity comparison test, 520 Dependent variable, 27 Dot (scalar) product, 651
Coordinate planes, 639 Derivative: use in computing work, 654
Coordinates: chain rule for, 120 geometric interpretation of, 651-52
Cartesian (rectangular), 20, 639 of a constant, 95 properties of, 651
cylindrical, 876-77 directional, 781-82 Double integral, 822
polar, 570 of an exponential, 346-47 exchanging order in, 834
spherical, 879-80 of a function f(x ), 86 general properties, 836
Copernicus, Nicholas (1473-1543), 734 higher order, 101-2 in polar coordinates, 839
Cosine function see Sine and cosine of hyperbolic functions, 480 over type 1 and type II regions, 830-31
functions of inverse trigonometric functions, 367 Drug elimination, 353, 506
Cost function, 231 of linear combinations, 97
Critical point: of the natural logarithm, 345
o f ∕( x , y), 795 notation, 91 E ___________________
first-order for f(x ), 168 partial, 756 e (natural exponential base), 329
second-order for ∕( x j, 194 of a polynomial, 100 as a limit, 330, 500 (Prob. 53)
Critical value, 168 as a rate of change, 112 as the sum of a series, 552
Cross-partials test, 917 second, 101-2 Eccentricity of conic sections, 619
Cross product of vectors, 657 of trigonometric functions, 108 Ellipse:
use in computing area and volume, of a vector function, 692, 695 centroid, 431 (Prob. 34)
661-63 Descartes, Rene (1596-1650), 20 directrix of, 619
determinant formula for, 657 Determinant: eccentricity, 619
geometric interpretation, 659 for cross product, 657 equation (standard), 614
magnitude of, 660 for triple scalar product, 662 foci of, 612
properties of, 658 Difference quotient, 28, 86 major and minor axes, 612
Cross-sectional method (volume), 385 Differentiable function: in polar coordinates, 619-21
Curl of a vector field: is continuous, 90, 768 reflection property, 622
definition, 901 of one variable, 87 vertex, 613
physical interpretation, 903 of two variables, 769 Epsilon-delta proofs, 71 ff.
Current in an electrical circuit: Differential: Error:
steady-state, 424 dx, dy, 147-48 absolute, 14
transient, 424 total, of f( x , y^), 768 for alternating series, 533
Curvature: Differential equation: propagation, 149
formulas for, 717, 719 exact, 958 relative, 150
radius of, 718 first-order linear, 421 in Simpson’s rule, 306
of a smooth curve, 715 general solution, 287 in Taylor’s theorem, 550-53
summary of formulas for, 715 logistic, 418 in the trapezoidal rule, 306
Curve: nth order, 953 Escape velocity, 291-93
closed, 920 separable, 288, 970 Euler, Leonhard (1707-1783), 374
concavity of, 193 solution, 287 Even and odd functions, 30
Jordan, 925 Differentiation: Eventually monotone sequence, 497
length of, 397, 712 basic rules for, 97 Exact differential equation, 958
normal line to, 89 implicit, 130 Exchanging order of integration, 834
parametric, 594, 687 logarithmic, 348-49 Exponential function:
piecewise smooth, 694 partial, 756 base b, 326
smooth, 694 of power series, 95 graph of, 326
symmetry of polar, 578 of vector functions, 692 natural, 330
Index A -1 0 9
Exponential growth, decay, 342, 418 Gauss, Karl Friedrich (1777-1855), 949 transverse axis, 616
Extended mean value theorem, A27 Gauss’s law, 956 (Prob. 28) vertices, 616
Extent, 213 General solution, 287 Hyperbolic functions:
Extrema: Generalized mean-value theorem, A27 definitions, 478
absolute, 166, 793 Generalized derivative formulas, 123 derivatives involving, 480
relative, 168, 794 Geometric and physical interpretations: graphs of, 479
Extreme value theorem, 166, 798 of the cross product, 659 identities involving, 479
of the curl, 903 integrals involving, 481
of the divergence, 903-4 inverses of, 481-82
F _____________________ of the dot product, 651-52
Falling body problem, 116 of Lagrange multiplier method, 812, 807
of MVT for derivatives, 179
Faraday’s law, 948 (Prob. 26)
Fermat, Pierre de (1601-1665), 116, 483 of MVT for integrals, 295-96 I _____________________
Fibonacci sequence, 499 (Prob. 48) Geometric series: Implicit differentiation (procedure for),
First derivative test, 186, 200 definition, 504 130
First-order linear differential equation, and repeating decimals, 505 Implicitly defined function, 129
421 sum of, 504 Improper single integrals, 473-74
Fluid flow, 290-91 Gradient vector: Improper double integrals, 843
Flux density, 900, 944 definition, 784 Incompressible, 953
in directional derivative, 784 Increase without bound, 204
Flux integral, 944, 959
Force: as a normal vector, 787 Increasing:
exerted by a spring, 405 as an optimal direction, 786 function, 183
fluid, 407-8 properties, 785 sequence, 496
Frobenius, method of, 1003 Graph: Increasing, decreasing function theorem,
Fubini’s theorem, 825, 831, 857 of an equation, 15 184
geometric justification, 827 of a function, 33, 741 Increment:
Function: of a parametric curve, 594-97 o f∕(x ), 147
in polar coordinates, 574-83 off( x , tj, 766
algebraic, 38
bounded, 27 Graphing strategy, 216 Independence of path, 916, 920
Gravitational field, 899 Independent variable, 27
classification of, 38
composition of, 30-31 Green, George ( 1793-1841), 925 Indeterminate forms, 239
Green’s identities, 930, 955, 956 (Prob. 24) Inequality:
continuous, 57, 751
Green’s theorem, 925 absolute value, 13
cost, 231
decreasing, 183 Growth models (summary), 460 triangle, 10
definition, 26 Gyration (radius of), 415 (Prob. 47) Infinite limit of integration, 469
derivative of, 86 Infinite limits, 208
Inflection point, 193
differentiable, 87, 769
Inhibited growth, 418
domain of, 26, 739
elementary, 27
H ___________________ Initial value problem, 422
even, 30 Half-life, 417 Instantaneous rate of change, 113
Hamilton, William Rowan (1805-1865), Instantaneous velocity, 42
exponential, 39
967 Integrable function, 268
gamma, 487 (Prob. 62)
Harmonic function, 904 Integral:
graph of, 33, 741
Harmonic series, 512 approximation to, 302-7
hyperbolic, 478
Heat: basic properties, 248, 274
implicitly defined, 129
equation (derivation), 954 change of variable, 280-81, 284
increasing, 183
flow, 294 (Prob. 47), 789 constant multiple property, 248
inverse hyperbolic, 481-82
Helix, 686 of continuous functions, 270
inverse trigonometric, 364-65
Higher-order derivatives: in cylindrical coordinates, 877-78
logarithm, 39
o f∕(x ), 101-2, 132 definite, 268
odd, 30
partial, 760-61 double, 822-23
one-to-one, 27
of vector functions, 695 of hyperbolic functions, 481
piecewise-defined, 28
power, 38, 95 Hooke’s law, 141 (Prob. 10), 405 improper, 473-74, 843
Horizontal asymptote, 203, 209 indefinite, 247
price, 231
Horizontal integrating strip, 313-14 involving ex , 359
profit, 231
Horizontal: involving inverse trigonometric functions,
range of, 26
line, 21-22 369
rational, 38
tangent, 168, 590 involving fn x, 360
real-valued, 27 iterated, 824, 857
Horizontal line test, 336
revenue, 231
Houdini, Harry (1874-1926), 434 as a limit of Riemann sums, 268
smooth, 694
Hydrostatic (fluid) force, 407-8 line, 906-7
of two variables, 739
Hypatia (A.D . 370-415), 605 linearity property, 248
transcendental, 38
Hyperbola: multiple, 822
trigonometric, 39
asymptotes, 617 in polar coordinates, 588
Fundamental theorem: center, 617 in spherical coordinates, 881-82
of calculus (first), 227
central rectangle of, 617 of trigonometric functions, A-46 to A-47
of calculus (second), 299
conjugate axis, 616 triple, 855-57
on line integrals, 918 directrix of, 619 of a vector function, 699-700
eccentricity, 619 Integral tables, 440-45, see Appendix D
equation, 617 Integral test, 511
G ____________________ foci, 616 Integrand, 269
Galilei, Galileo (1564-1642), 733 polar equation, 619 Integrating factor, 421-22
Gamma function, 487 (Prob. 62) reflection property, 623, 624 (Prob. 61) Integration (see Integral)
A-110 Index
Intercepts, 33 of a spring, 405 Logistic equation, 418

Interest, 330-31 of a vector, 634, 649 Log-power quotient series, 520
Intermediate value property, 63 Level curve, 741-42 Lower bound of a sequence, 496
Intersection of polar graphs, 585-87 Level surface, 741-42 Lucas, William F. (essay), 966-68
Interval notation, 11 ΓH6pital, G. F. A. de (1661-1704), 239
Interval of convergence, 540-41 ΓHδpitaΓs rule
Interval of integration, 269 0/0 and ∞∕∞ forms, 240 M___________________
Intervals of increase and decrease, 184 other forms, 243-44, 355 Maclaurin, Colin (1698-1746), 556
Inventory model, 234 used with sequences, 495 Maclaurin series, 551
Inverse function: Limaςon, 579 Magnitude of a vector, 634, 649
continuity of, 337 Limit: Marginal:
definition, 334 divergent, 45-47 analysis, 150, 232
derivative, 337 divergent to infinity, 46 cost, 150, 232
differentiability of, 337 divergent by oscillation, 46 revenue, 132, 150, 250
existence of, 336 exponential, 328, 352 Mass:
graphs of, 336-37 formal definition (∕ε), 68, 750 center of, 866-67
horizontal line test for, 336 by graphing, 43 of a planar lamina, 865
Inverse hyperbolic functions: geometric interpretation, 42 of a solid, 867
derivatives, 482 indeterminate form, 239 Mathematical modeling, 6-8
logarithm forms for, 482 infinite, 208 Maximum:
Inverse square law (Newton’s law of gravita of integration, 269 absolute, 166
tion), 119 (Prob. 40), 292 (footnote), intuitive formulation, 4, 42 procedure for finding, 171, 799
295 (Prob. 51) one-sided, 43 relative, 168
Inverse trigonometric functions, 364-65 of a piecewise-defined function, 54 Maximum profit, 232
derivatives of, 367 properties, 49 Maxwell, James Clerk (1831-1879), 904
graphs of, 364-65 of a rational function, 51 Maxwell’s equations, 946, 960
properties of, 366 of a sequence, 492 Mean-value theorem:
Irreducible quadratic polynomial, 455 of a sum, 258 Cauchy’s generalized, A27
Isotherms, 289 squeeze rule for, 50 for derivatives, 177
Iterated integral, 856 tabular computation of, 44 geometric interpretation, 179
Iteration, 824, 857 of trigonometric functions, 44, 51, 52, 104 for integrals, 295
of a vector function, 688-89 Midpoint formula, 15
J ___________________ Limit comparison test, 518
Limit point, 753
Minimum average cost, 232
Minimum of a function:
Jacobian, 885 Limit to infinity, 204 absolute, 166
Jacobi, Carl J. (1804-1851), 886 Limits of integration, 269 procedure for finding, 171, 799
Joint probability density function, 871 Linearly dependent and independent, 638 relative, 168
Jordan curve, 925 Lines in the plane: Mixed partial derivatives, 760
forms for the equation of, 22 Mobius strip, 933-34
κ ___________________ normal, 89 Moment:
Kelvin, Lord William Thompson parallel, 23 about an axis, 865
(1824-1907), 925, 927 perpendicular, 23 of a planar lamina, 868
Kepler, Johann (1571-1630), 256, 734 secant, 5 of a solid region, 865-66
Kepler’s laws of planetary motion, 705, 735 slope of, 21 Moment of inertia, 415 (Probs. 45-46), 869
Kinetic energy, 430 (Probs. 26-27) tangent, 84 Monotone function, 183
Lines in space, 665 Motion:
parallel, 667 of an object in space, 697
L___________________ parametric equations for, 665 in a resisting medium, 974
Lagrange form of the remainder, 551 skew, 667 rectilinear, 114
Lagrange identity, 682 (Prob. 75) symmetric equations for, 666 simple harmonic, 128 (Prob. 50)
Lagrange, Joseph Louis (1736-1813), 812 vector equation for, 665 in a vacuum, 702
Lagrange multiplier method: Line integral: Mt. St. Helens, 738
geometric interpretation, 812 with respect to arc length, 913 Multiplier effect, 506
procedure, 806 to compute work, 911-12 Muzzle velocity, 701
with two constraints, 810 definition, 906-7
Lamina, 410-11, 865-69 independent of path, 916
Laplace, Pierre Simon (1749-1827), 306 properties, 907 N ___________________
Laplace’s equation, 904 of vector fields, 908-9 Napier, John (1550-1617), 337
Laplacian, 904 Linear approximation, 147 Native Americans (Navajo), 58
Least squares approximation, 800-801 Linear combination: Natural exponential function, 330
Left-hand limit, 43 derivative of, 97 derivative of, 377
Legendre: of i, j, and k, 648 as inverse of fn x, 375-76
equation, 1002 (Prob. 32) Linearity: notation (exp), 330
polynomial, 1002 (Prob. 32) of differentiation, 97, 100 power series for, 552-53, 561
Leibniz, Gottfried Wilhelm (1646-1716), of integration, 248 properties of, 328
81 Logarithm function, 338 series for, 548-49
Leibniz notation, 92 change of base, 341 Natural (exponential) growth, 418
Leibniz’ rule, 300 common and natural, 340 Natural length of a spring, 405
Lemniscate, 582 defined as an integral, 373 Natural logarithm function:
Length: geometric interpretation, 374-75 as an integral, 373
of arc, 397-98 properties of, 340 as inverse of e, 338
of a space curve, 713 Logarithmic differentiation, 348-49 power series for, 547, 553
Index A -lll
Negative polar angle, 570-71 reflection property, 610 plotting points in, 570
Neigborhood, 746 vertex, 605 primary prepresentations in, 571
Net earnings, 314-15 Parabolic mirror, 611 (Prob. 55) conversion to rectangular coordinates, 572
Newton, Sir Isaac (1642-1727), 80 Parabolic trajectories, 703 curves, 573, 583 (summary)
Newton’s law of cooling, 333 (Prob. 36), Paraboloid, 743 intersections, 586
358 (Prob. 48) Parallel: partition of a sector, 588
Newton’s laws of motion: lines, 23 rectangle (cell), 588, 839
inverse-square law of gravitation, 119 planes, 670 rotation, 578
(Prob. 40), 292n., 295 (Prob. 51) vectors, 667 symmetry, 578
second law, 702 Parallelepiped, 661-62 unit vectors ur and vβ , 705
Newton-Raphson method: Parallelogram: Polar (radial) axis, 570
error in, 156 area, 661 Polynomial:
procedure, 155-56 law (vector addition), 632 continuity of, 59
when method fails, 158 Parametric form of: degree of, 38
Noether, Emmy (1882-1935), 442 arc length, 601 derivative of, 100
Nondecreasing sequence, 496 area, 602 graph of, 188 (Example 4)
Nonincreasing sequence, 496 a curve, 687 integral of, 248
Nonnegative term series, 511 a cycloid, 598 Maclaurin and Taylor, 548-49
Nonorientable surface, 933-34 a derivative, 599 in two variables, 740
Norm: a line, 665 Population models:
of a Riemann sum, 267 a surface, 938 exponential growth, 418
of a vector, 634 Parameterization of a curve, 594, 687 logistic growth, 418
Normal: Partial derivatives, 756 Position:
line, 89, 788 chain rules for, 774, 777 function, 114
principal unit N, 711 equality of mixed, 761 vector, 698
vector to a surface, 787 geometric interpretation, 758 Positive (polar) angle, 570-71
Normal component of acceleration, higher-order, 760 Potential energy, 430 (Prob. 26-27)
722 of implicit functions, 757 Potential function, scalar, 916-17
Number line, 9 rate interpretation, 759 Power rule:
Numerical integration, 302 Partial fractions decomposition, 455-58, for derivatives, 95
with Maclaurin series, 551-55 563 (Probs. 54-60) for integrals, 248
by rectangles, 303 Partial integration, 824 Power series:
by Simpson’s rule, 305 Partial sums of series, 501 absolute convergence set, 540
by trapezoidal rule, 303-4 Partition (for Riemann sums): binomial, 560
of an interval, 267 and differential equations, 1000-1
of a planar region, 822-23 in (x - c), 543
polar, 588 Maclaurin series, 548-49, 551
0 ___________________ regular, 268 numerical computations with, 552-54
Oblique asymptote, 203, 209 Parts (integration by), 448 radius of convergence, 541
One-sided limits, 43 Pendulum, 128 (Prob. 54), 977 representation only at x = 0, 555
One-to-one function, 27 (Prob. 46) substitution in, 555-56
Open interval, 11 Percentage rate of change, 119 (Probs. table of common series, 561
Optical properties: 42-47) Taylor series, 551
Fermat’s principle, 225 Perpendicular: term-by-term differentiation in, 545
reflection, 610 lines, 23 term-by-term integration in, 545
Snell’s law, 226 vectors, 653 uniqueness principle, 550
Optimization: Pi W Present value of future income, 330
of discrete functions, 230 computation of, 560 Pressure (fluid), 407
o f∕(x ), 173-74, 353, 369 series for, 560 Principal unit normal vector, 710
o f ∕( x , y), 793-98 Piecewise smooth: Prismoidal formula, 309 (Prob. 28)
general procedure, 220 curve, 694 Probability density function (joint), 871-72
Orbit: surface, 933 Product rule:
of earth, 624, 705-8 Plane(s), 641 for derivatives, 97
geosynchronous, 728 attitude numbers for, 670 for the gradient, 785
Order of integration, 824 coordinate, 639 for limits, 49, 750
Order properties, 9 equation for, 670 Product:
Orientable surface, 933 how to graph, 642 cross (vector), 657
Orthogonal trajectory, 289 parallel, 670 dot (scalar), 651
Osculating circle, 718 in space, 670 rule (for derivatives), 97
Overdamping, 970 Planetary motion, 708/. triple scalar, 661-62
Point of discontinuity, 57-58 Profit function, 231
Point of inflection, 193 Projectile:
Point-normal form for a plane, 670 range, 704
P ___________________ Point-slope form for a line, 22 time of flight, 704
Pappus’s volume theorem, 412 Poiseuille’s laws, 40 (Prob. 60), 153 (Prob. trajectory, 702
Pascal’s principle, 407 47), 235 Projection:
Parabola, 605 Pole (origin), 570 scalar, 653
axis, 605 Polar angle, 570 vector, 653
directrix, 605 Polar coordinates, 570 Propagated error, 150
equation, 606, 607 arc length in, 604 (Prob. 65) p-series test, 514
focus, 605 area in, 588 Pyramid of Cheops, volume of, 395 (Prob.
polar form, 609 of conic sections, 609, 621 57)
A -1 1 2 Index
Root law for limits, 49 power series, 539

Q----------------------------- Root location theorem, 63 radius of convergence, 541
Quadric surfaces, 643, 644 Root test for series, 525 ratio test, 523
Quotient rule: Rose curves, 580 rearrangement of terms, 537, 538 (Prob.
for derivatives, 97 53)
for the gradient, 785
for limits, 49, 750 s _____________ root test, 525
sequence of partial sums of, 501
Saddle point, 796 sum of, 501
R ___________________ Satellite orbits, 728 Taylor, 551
Radial: Scalar, 630 telescoping, 502
axis, 570 Scalar multiple of a vector, 631 terms of, 501
coordinate (polar form), 570-71 Scalar potential function, 916-17 Shells, method of cylindrical, 391
distance, 570 Scalar triple product, 661-62 Sign of the first derivative, 186, 200
unit vector u, 705 Secant line, 5 Sign of the second derivative, 199, 200
Radian measure, 17 Second derivative, 101 Simple harmonic motion, 128 (Prob. 50),
Radioactive decay, 416 Second derivative test, 199, 200 989 (Prob. 39)
carbon dating, 417-18 Second fundamental theorem of calculus, Simply connected region, 917, 925
half-life, 417 299 Simpson, Thomas (1710-1761), 305
Radius of convergence, 541 Second-order partial derivatives, 760 Simpson’s rule, 305
Radius of curvature, 718 Second partials test, 796 error in, 306
Range of a function, 26 Seki Kδwa (1642-1708), 258 Sine and cosine functions:
Range of a projectile, 704 Separable differential equation, 288, 972 continuity of, 59
Rate of change, 112, 759 Sequence, 490 derivatives of, 107
Ratio test, 523 bounded monotone, 496 power series for, 552, 561
Rational function: convergence, 492 special limits involving, 104
continuity of, 59 divergent, 492 Sink and source, 953
of sine and cosine, 462 eventually monotone, 497 Sketching graphs (strategy), 216
Rational trigonometric integrals, 462-63 Fibonacci, 499 (Prob. 48) Skew lines, 667
Rearrangement of series, 537, 538 (Prob. 53) general term, 491 Slope of a line, 21
Rectangular (Cartesian) coordinate system, graph of, 492 as a derivative, 84, 250
639-40 limit, 491-92 in polar form, 590
Rectilinear motion, 114 limit rules, 492 Slope-intercept form for a line, 22
Recursively defined sequence, 499 (Prob. monotone, 496 Smooth:
48) nondecreasing and nonincreasing, 496 curve, 694
Reduction formula, 442-43 notation, 491 function, 694
Reference triangle, 366 nth term, 491 surface, 933
Refinement of partition, 267-68 of partial sums, 501 Snell’s law of refraction, 226
Reflection of a graph, 37 power, 498 Snowplow problem, 433 (Prob. 70)
Reflection property: recursively defined, 499 (Prob. 48) Solid of revolution, 387
of ellipse, 622 squeeze law, 495 Solution (differential equation), 287
of hyperbola, 623, 624 (Prob. 61) terms of, 491 Somerville, Mary Fairfax (1780-1872), 751
of inverse functions, 336 Series: Space coordinates:
of parabola, 610 absolute convergence test, 534 cylindrical, 876-77
Refraction of light, 226 absolutely convergent, 535 rectangular, 639-40
Regular partition, 268 alternating harmonic, 532 spherical, 879-80
Related rates, 136—37 alternating p-series, 532 Space curve, 685
Relative rate, 418 alternating series test, 530 Speed, 1 14, 698, 714
Relative extrema, 168 binomial, 560 Sphere, 642
derivative tests for, 186, 199, 200 direct comparison test for, 516 equation for, 642
Remainder term (Lagrange form), 551 conditionally convergent, 535 surface area of, 400 (Ex. 5)
Repeating decimal, 505 convergent, 501 volume of, 393 (Probs. 15-16)
Residual concentration of drug, 506 definition, 501 Spherical coordinates:
Revenue function, 231 divergence test, 510 triple integrals in, 879-80
Reversing order of integration, 834 divergent, 501 Spiral, 576
Revolution: domination of, 516 Springs, 405
solid of, 387-88 error estimate for alternating series, 533 Spring constant, 405
surface of, 399-400 expanded form, 502 Spy problem, 78, 120, 227, 294, 380, 426,
Riemann, G. F. B. (1826-1866), 266 for (pi), 560 464, 486, 509, 604, 681, 728, 790, 884,
Riemann sum: general properties of, 503 932
formation of, 268, 822, 856 generalized ratio test, 536 Squeeze theorem, 50, 73
limit of, 267 geometric, 504 Standard form:
norm of, 267 harmonic, 512 for a circle, 15-16
Right endpoint approximation, 258 integral test, 511 for a line, 22
Right-hand limit, 43 interval of convergence, 541 for a plane, 670
Right-handed coordinate system, 639 limit comparison test, 518 Standard region, 926
Rise (and run), 20 linearity of, 503 Standard unit vectors:
RL-circuit, 424 log-power quotient series, 520 i, j in R 2 , 635
RLC-circuit, 998 Maclaurin, 551 i, j, k in R 3, 648
Rocket problem, 291-93 nth partial sum, 501 ur, uβ in polar form, 705
Rolle, Michel (1652-1719), 178 nonnegative term, 511 Steepest descent, ascent, 786
Rolle’s theorem, 178 p-serics, 514 Stirling’s formula, 487 (Prob. 64)
Index A -113
Stokes, George (1819—1903), 941 Termwise differentiation and integration of dependent, 27

Stokes’s theorem, 941 a series, 544-46 independent, 27
physical interpretation of, 944 Test for concavity, 194 limits of integration, 298-300
Stopping distance, 249-50 Test for orthogonality of vectors, 653 Variation of parameters, 995-97
Subinterval representative, 267 Test for relative extrema: Vascular branching, 235
Substitution: fo r∕(x ), 186, 199 Vector:
integration by, 280-81, 284 f o r∕(x , y), 794 acceleration, 698
in power series, 556-58 Tests for convergence of series: addition, 632
Sum: absolute convergence, 535 angle between two, 651
of a geometric series, 504 alternating series, 530 angular unit, 705
of an infinite series, 501 direct comparison test, 516 Cartesian unit: i, j, and k, 635, 648
of vectors, 632 divergence test, 510 component form, 632
Summation notation, 261 integral test, 511 components, 653
general rules, 262 limit comparison test, 518 continuity, 690
index, 262 ratio test, 523 cross (vector) product, 657-58
limits, 262 root test, 525 derivative, 692
special formulas, 263 Theorems (Appendix A, from A-l to A-26) difference of two, 632
Surface, 641 Torque vector, 663 direction angles of, 669
closed, 933 Torricelli’s law, 290«. direction cosines of, 655 (Prob. 49),
normal line to, 788 Torsion, 721 (Prob. 57) 669
orientable and nonorientable, 933 Torus: direction vector, 635
parametric, 938 surface area of, 403 (Prob. 35), 855 dot (scalar) product, 650
piecewise smooth, 933 (Prob. 45) equality, 631
quadric, 643, 644 volume of, 396 (Prob. 66), 412 gradient, 784
of revolution, 399-400 Total differential, 767 initial point of, 631
tangent plane to, 788 Trace of a surface, 645 integral, 699
Surface area: Trajectory of a projectile, 698 length (magnitude), 634, 649
fundamental vector product, 847-48 Transcendental function, 38 limits, 688-89
of a parametric surface, 851-52 Transformation, 36 linear combination, 633
of surface z = f(x , y), 848 Transient current, 424 multiplication by scalar, 631
Surface integral, 834-835 Translation of a graph, 36 notation, 631-32
Symmetric equations of a line, 666 Transverse axis of a hyperbola, 616 negative of, 631
Symmetry: Trapezoidal rule, 303 operations, 633
in polar coordinates, 578 error in, 306 parallel, 650
in rectangular coordinates, 213 Triangle inequality, 10 parallelogram rule, 632
Symmetry of polar curves, 578 Triangle law (vector addition), 632 perpendicular (orthogonal), 652
Trigonometric: position, 698
τ _____________ derivatives, 108
equations, 17
principal unit normal, 710-11, 715
product (cross), 657
Tables of integrals (Appendix D. from A-42 graphs, 35 projection, 653
to A-48) limits, 52, 104 radial unit: u, 705
Tangent function: power series for, 561 scalar (dot) product, 651
continuity of, 59 substitution, 445-46 scalar multiple of, 631
derivative of, 108 Triple integration: stationary, 698
Tangent line: in cylindrical coordinates, 876-77 subtraction of, 632
to a graph, 84 in rectangular coordinates, 855-57 sum of, 632
horizontal, 168 in spherical coordinates, 879-80 terminal point of, 631
to a parametric curve, 600 Triple scalar product, 661-62 in three dimensions, 648
in polar coordinates, 590-91 Trochoid, 597-98 triple scalar product, 662
slope of, 84-85 Troutman, John (essay), 80-81 unit, 635, 649
vertical, 90 Two-intercept form for a line, 22 unit normal, 693
Tangent line approximation, 147 Type I and type II regions, 830-31 unit tangent, 693
Tangent plane: vector (cross) product, 657-58
approximation, 766
to surface F(x, y, z) = 0, 788 u _____________ velocity, 698
zero, 631
to surface z = f(x. y), 765 Unbounded interval, 11, 469 Vector (cross) product, 657
Tangent problem, 4 Underdamping, 987 properties, 658
Tangent vector (unit), 693 Undetermined coefficients, method of, 991, determinant formula for, 657
Tangential component of acceleration, 994 magnitude of, 660
722 Unit normal vector, 710-11, 715 orthogonality property, 652
Taylor, Brook (1685-1731), 549 Unit tangent vector, 710-11, 715 Vector derivative:
Taylor polynomial, 550-51 Upper bound for sequences, 496 definition, 692
Taylor series, 551 rules for, 695
with remainder, 550
theorem, 551 V_____________ Vector equation of a line, 665
Vector field, 898
Techniques of integration (summary), Value: acceleration, 698
465-67 average, 297 circulation, 959
Telescoping sum, 502 maximum and minimum, 166 curl of, 901
Terms: Variable, 27 conservative, 916
of a sequence, 490 change in double integral, 885 divergence of, 900
of a series, 501 change in triple integral, 888 flux of, 959
A-114 Index
Vector field (cont’d):

gradient, 784
Vertical asymptote, 203, 209
Vertical integrating strips, 312
w_____________
velocity, 698 Vertical line, 21, 22 Wallis’s formulas, 454 (Probs. 51-52)
Vector forms of Green’s theorem, 929 Vertical line test, 36 Weierstrass, Karl (1815-1897), 59, 68
Vector integration, 699 Vertical tangent, 901 Whispering room, 622
Vector functions: Vibrating springs, 987-88 Work:
continuity of, 690 Volume: as a dot product, 654
difference quotient of, 692 by cross sections, 385 to fill (drain) a tank, 406
differentiation of, 692, 695 by disk method, 388 as a line integral, 911-12
domain of, 685 in cylindrical coordinates, against a spring, 405
integration of, 699-700 877-78 units, 404
limits of, 688 by cylindrical shells, 391 done by a variable force, 405
operations with, 687 by double integration, 823, 832
range of, 685
Vector limits, 688
of an ellipsoid, 395 (Prob. 65)
of frustum of cone, 396
X_____________
x-intercept, 33
Velocity: (Prob. 68)
and acceleration, 114 of a parallelepiped, 662
average, 42 of a pyramid, 387
escape, 291-93 of a solid of revolution, 388
field, 898 of a sphere, 395 (Prob. 63), 881 ^-intercept, 33
initial, 702 of a tetrahedron, 395 (Prob. 64)
in polar form, 706
instantaneous, 42
of torus, 396 (Prob. 66), 412
by triple integration, 858 z_____________
summary, 729 under z = f(x , jj, 823, 859 Zeno’s paradox, 3, 508 (Prob. 57)
vector, 698 by washer method, 389 Zero vector, 631
DIFFERENTIATION FORMULAS
PROCEDURAL RULES
Constant multiple { c fY = c f'
Sum rule ( f + g ) ' = f ' + g'
Difference rule < j-g y = r -g ,
Linearity rule ( a f+ b g y = a f'+ b g '
Product rule { fg y = fg , + f g
/Z j = s f ' - f g '
Quotient rule ∖g ) g2
dy _ dy du
Chain rule
dx du dx
B A SIC F O R M U L A S
d ,, ∖du
Extended power rule — un = nu n 1 -
dx dx
d_ du d ■ du
Trigonometric rules cos u = -sin u — — sin u = cos u —
dx dx dx dx
d_ ∙-j du d . ∩ 2 du
2
tan u = sec u — — cot u = -e s c u—
dx dx dx dx
d_ du d . du
sec u = sec u tan u — — CSCu = —esc ucot u —
dx dx dx dx
d_ -1 du d ■-1, 1 du
Inverse trigonometric rules — sin u =
dx √ 1 - w2 dx dx ∙x∕ l - w2 dx
d __ l .. 1 du d -1 du
— cott -1ι u =
dx 1 + w2 dx dx 1 + u2 dx
d _i 1 du d ι -1 du
— CSC - 1 u =
dx ∣w∣ ∖∕u 2 - 1 dx dx ∣M∣√ W2 - 1 dx
d_ , 1 du log// du 1 du
Logarithmic rules
dx u dx i 10⅛l"l= u dx
=
u ln b dx
^~bu = bi ll n b ~
Exponential rules
dx dx dx dx
d ∙ ∣ du d ■, , du
Hyperbolic rules cosh1 u = sinh u — — sinh u = cosh u —
dx dx dx dx
d_ 2 Z coth u = -csc h 2 u
tanh u = sech u
dx dx dx dx
± csch u = - csch u coth u
sech u = -sechu tanh u
dx dx dx dx
d • . -1 1 du 4~ cosh 1 u= ±1 du
Inverse hyperbolic rules oHlll U r—-----
dx ∖∕ u2 + 1 dx dx ∖ ∕ H2 - 1 dx
± 1 du
tanh- 1 u = — -—2 — , ∣w∣ v^ coth - 1 u= ∣M∣>1
dx 1 - u dx dx 1 - w2 dx
d +1 du —1 du
oC/Ull I Λ !-------------- — csch - 1 u=
dx u ∖∕ 1 - w2 dx dx ∣M∣AV∕ 1 —M2 dx
INTEGRATION FORMULAS
P R O C E D U R A L R U LES Constant multiple c f ∖u) du = c I f(u ) du
Sum rule [f(μ ) + g ( ιi) ] d u = lf( u ) d u + I g(u) du
Difference rule [/(«) - g(w)] d u = f(u ) du -
Linearity rule [ a f(u ) + h g(u)]du —a I f(u) du b I g(u) du
J 0 du = C
B A SIC F O R M U L A S Constant rule
Power rules + C ; n ≠ -1
u 1 du - In ∣w∣ + C
Exponential rules el1 du = e u + C
a 11 du = --------1- C ; a > 0, a ≠ 1
In a
Logarithmic rule In u du = M In u —u + C ; w >0
Trigonometric rules sin u du = -c o s u + C cos u du = sin u + C
tan u du = - In ∣cos M∣ + C cot u du - In ∣sin M∣ + C
sec u du = In ∣sec u + tan u∖+ C esc u du = —In ∣csc u + cot u∣ + C
sec2 u du = tan u + C csc 2 u du = -c o t u + C
sec u tan u du = sec u + C esc u cot u du = —esc u + C
J cos2 u dιι = ∣ u + ∣ sin 2u + C

Cosine squared formula
Sine squared formula j sin2 u d u = ∣ u - ∣ sin 2u + C
J cosh u du = sinh u + C
Hyperbolic rules sinh u du = cosh u + C
I tanh u du = In (cosh u) + C coth u du = In (sinh w) + C
, u _
Inverse rules = cosh- 11 - + C
a
u
C tanh- 1 - + C if 1
J a a
u
I coth- 1 - + C if 1
a a
- - sech 1 1- I + C
a Ia I
= -csc h 1 ∣M∣ + C
,,
~≈ CALCULUS
GERALD L BRADLEY & KARL J. SM ITH
To enhance learning by the use of computational technology, we offer

the following manuals, which give specific, key-stroke guidance to prob
lems in the text and also offer supplementary problems.
Calculus Explorations Using Mathematica

0-13-328618-5
Calculus Explorations Using Maple

0-13-328576-6
Calculus Explorations Using MATLAB®

0-13-328600-2
Calculus Explorations Using Texas Instruments Graphics Calculators

0-13-328592-8
Calculus Explorations Using Hewlett-Packard Graphics Calculators

0-13-328584-7
PRENTICE HALL
ENGLEWOOD CLIFFS, NJ 0 7 6 3 2
9 780133303254

Calculus (Bradley &amp; Smith) - Instructor&#39;s Edition

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Calculus (Bradley &amp; Smith) - Instructor&#39;s Edition

Uploaded by

Copyright:

Available Formats

~r

iERALD L. BRADLEY & KARL J. SMITH

Properties of Absolute Value, see Table 1.1, p. 10

LIMITS, DIFFERENTIATION, AND GRAPHING

ANALYTIC GEOMETRY, VECTORS, AND THREE DIMENSIONS

FUNCTIONS OF TWO VARIABLES

Bradley, Gerald L., 1940-

Acquisition Editor: George Lobell

© 1995 by Prentice-Hall, Inc.

Prentice-Hall International (UK) Limited, London

About the Authors xii

Preview of Calculus: Functions and

2 Techniques of Differentiation with Selected

□ Additional Applications of the Derivative 165

5 Exponential, Logarithmic, and Inverse

OAdditional Applications of the Integral 383

7 Methods of Integration 436

8 Infinite Series 489

9 Polar Coordinates and Parametric Forms 569

10 Vectors in the Plane and in Space 630

12 Partial Differentiation 838

13 Multiple Integration 821

14 Vector Analysis 897

Gerald L. Bradley Karl J. Smith

Calculus teaching is undergoing great changes, most of which will be of lasting

Besides developing some minimal skills in algebraic manipulation and problem

Cooperative Learning (Group Research Projects) There is a great deal of interest

M athem atical Communication Students should be encouraged to communicate

Computational Windows Reform is driven partly by the need to maximize the

Technology Laboratory M anuals For those students with access to a computer or

S ign ifican t D igitsWith the availability of calculators and computers, we were

Problem s We believe that students learn mathematics by doing mathematics.

Journal Problem s In an effort to show that m athematicians work problems too,

competitions. The Putnam Examination is a national annual examination given

Student Mathematics Handbook We begin Calculus with a minimum of review.

Differential equations We also include differential equations early (in Section

B A Student Survival Manual by Ken Seydel offers a running commentary of

The writing and publishing of a calculus book is a tremendous undertaking. We

publication of this book. Our sincere appreciation to development editor Alan

Reviewers o f this edition:

1.1 WHAT IS CALCULUS?

■ THE LIMIT: ZENO'S PARADOX

4. Instantaneous changes in po­

5. Average of an infinite collec­ Figure 1.3 Achilles and the tortoise

can be described by writing a general term: —qyγ where n = 1, 2, 3, 4, . . . .

It is reasonable to guess that the sequence of fractions is approaching the

■ THE DERIVATIVE: THE TANGENT PROBLEM

lim ∕(Q + A) ~ /(a)

■ THE INTEGRAL: THE AREA PROBLEM

Figure 1.6 Approximating the area of a circle

Let .-43 be the area of the inscribed equilateral triangle;

Figure 1.7 Area under a curve

Figure 1.8 Approximating the area using circumscribed rectangles

From Scientific American, March

IN THIS SECTION Distance on a number line, absolute value, distance in

Order Properties For all real numbers a, b, and c:

■ A BSO LUTE VALUE

Absolute Value The absolute value of a real number a, denoted by ∣α∣, is

0 ∣α∣ is N O T the number a a if a < 0

TA B LE 1.1 Properties of Absolute Value

ι . a∖ ≥ 0 1. Absolute value is nonnegative.

A convenient notation for representing intervals on a number line is called

Name of Interval Inequality Notation Interval Notation Graph

Closed interval a ≤ x ≤ b [α> b] a b

Calculus (Bradley & Smith) - Instructor's Edition

Calculus (Bradley & Smith) - Instructor's Edition

4. Instantaneous changes in po

5. Average of an infinite collec Figure 1.3 Achilles and the tortoise

Figure 1.20 Graph of lines paral y - 7 = -¾(χ - 4)