Professional Documents
Culture Documents
INSTRUCTOR’S EDITION
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
DIFFERENTIAL EQUATIONS
Introduction to Differential Equations, see Section 4.5, p. 287
Growth and Decay, see Table 6.3, p. 420
INSTRUCTOR’S EDITION
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)
INSTRUCTOR’S EDITION
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:
Cover photo: Koji Horiuchi, courtesy Pei Cobb Freed and Partners
ISBN □ -1 3 -3 3 □ 3 ≡ 5 -X
Preface xiii
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
Chapter 4 Review, 319
GUEST ESSAY: “Kinematics of Jogging,” by Ralph Boas, 322
BOOK REPORT: To Infinity and Beyond, A Cultural History to the
Infinite, 324
Contents vii
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
Chapter 11 Review, 728
GUEST ESSAY: “The Stimulation of Science,” by Howard
Eves, 733
≡≡ BOOK REPORT: Hypatia's Heritage, 736
Cumulative Review, Chapters 7-11, 736
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 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
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.
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.
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.
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
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
1
2 Ch. 1 Preview of Calculus: Functions and Limits
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.
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’
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’
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:
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.
Even though Archimedes did not use the following notation, here is the essence of
what he did, using a method called “ exhaustion” :
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
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
Sec. 1.1 What Is Calculus? 7
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
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
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.
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.
Property Comment
a ≤ x [a, ∞) *" a
x ≤ b (-∞, ⅛] < b
a < x (a, ∞) *- a
X < b ( ∞, ⅛) ■*“ b
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.
12 Ch. 1 Preview of Calculus: Functions and Limits
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.
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. ∣≡≡≡
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.
midpoint midpoint
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
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.
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.
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
Sec. 1.2 Preliminaries 15
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.
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.
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.
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.
16 Ch. 1 Preview of Calculus: Functions and Limits
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
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.
■ 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
≈ 57.29577951°
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.
n — ιτ 5ττ
“ 3’ 3
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
Sec. 1.2 Preliminaries 19
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.
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∣ < b and b > 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)-
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.
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
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.
22 Ch. 1 Preview of Calculus: Functions and Limits
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
m . 66 -- 34 . 23
*Hooke, s law is useful for small displacements, but for larger displacements, it may not be a good
model.
Sec. 1.3 Lines in the Plane 23
£ = ∣w + 2
e = ∣(0) + 2 = 2
The original length was 2 cm. To find the weight that corresponds to € = 5, we
solve the equation
2= w
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
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
24 Ch. 1 Preview of Calculus: Functions and Limits
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 . ■
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:
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)
Sec. 1.3 Lines in the Plane 25
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.
26 Ch. 1 Preview of Calculus: Functions and Limits
■ 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.
■ 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 ).
*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.
28 Ch. 1 Preview of Calculus: Functions and Limits
To find ∕( f ) we use the first line of the formula because f ≈ 1.57 < 2 :
Finally, because 2 ≥ 2, we use the second line of the formula to find ∕(2 ).
∕( 2 ) = 3(2)2 + 1 = 13
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.
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 = (-∞, ∞)
30 Ch. 1 Preview of Calculus: Functions and Limits
∕( 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.
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.
Sec. 1.4 Functions and Their Graphs 31
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.
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:
32 Ch. 1 Preview of Calculus: Functions and Limits
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 —
a. ∕( x ) = (x 2 + 5x + 1)5 b. f( x ) = cos 3x
c. ∕( x ) = sin x 3 d. f ( x ) = V 5x 2 - 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.
Sec. 1.4 Functions and Their Graphs 33
■ 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.
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.
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.
34 Ch. 1 Preview of Calculus: Functions and Limits
TABLE 1.3 Directory of Curves
Identity Function Standard Quadratic Function Standard Cubic Function
y ~X y = x2 y = x3
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
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.
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:
Sec. 1.4 Functions and Their Graphs 39
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)
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.
42 Ch. 1 Preview of Calculus: Functions and Limits
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. ■■■
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
Sec. 1.5 The Limit of a Function 43
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:
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
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
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.
44 Ch. 1 Preview of Calculus: Functions and Limits
■ 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.
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 + .
v(() approaches 64 from the left. —> <→,(0 approaches 64 from the right.
Evaluate lim
x→0 x
Solution ∕( x ) = s ιn ^
x is an even function because
Sec. 1.5 The Limit of a Function 45
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.
Computational Window
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.
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.
x approaches 0 from the left; x → 0 —> <—x approaches 0 from the right; x → 0 +
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
Graph of y = sin
Figure 1.36 lim sin - diverges by oscillation.
Sec. 1.5 The Limit of a Function 47
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.
f(χ ) 15.33
∕( v ) 3
r 3 _ o
14. χ →
lim
ι.
g(x),
O where Sg(x)
∖ > =Χ 2—+ 2—Χ -—
+ 4
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.
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
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
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.
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.
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. ⊂=
Sec. 1.6 Properties of Limits 51
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.
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.
= 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
52 Ch. 1 Preview of Calculus: Functions and Limits
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. ■
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)
lim 6 = lim (x + 3) = 5
x→ 2 X ~ 2 χ→ 2
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.
x →4 (χ—-4)(√χ + 2)
lim —FJ— -
√X + 2
-v→ 4
i = 1
√4 + 2 4
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)
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
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
Sec. 1.6 Properties of Limits 55
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
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.
56 Ch. 1 Preview of Calculus: Functions and Limits
1 . 7 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.)
■ DEFINITION OF CONTINUITY
1. ∕( c ) is defined;
2. lim ∕( x ) exists;
3. lim ∕( x ) = ∕( c ) .
x →c
*∙ ∕W = *2 χ-x i 3
b. g(x) = λ + 2-λ ]~ 3 if χ ≠ ι and β if _
= χ 1
e. G(x) = 7x 3 + 3x 2 - 2
f. H(x) = 2 sin x - tanx
58 Ch. 1 Preview of Calculus: Functions and Limits
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.
Continuous at .v = c:
lim ∕(x) exists and is equal to fic ).
lim∕'(x) = L and
lim ∕(x) = R
Sec. 1.7 Continuity 59
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.”
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.
■ 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
Sec. 1.7 Continuity 61
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.
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.
(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
62 Ch. 1 Preview of Calculus: Functions and Limits
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
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.
Sec. 1.7 Continuity 63
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.
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.
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.
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.
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.
Sec. 1.7 Continuity 65
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
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.
66 Ch. 1 Preview of Calculus: Functions and Limits
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
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
This section can be omitted without disrupting the development of the remainder of the text.
68 Ch. 1 Preview of Calculus: Functions and Limits
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.
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” ).
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
Thus, the conditions of the definition of limit are satisfied, and we have
lim (2x + 1) = 5
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∣
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
Sec. 1.8 Introduction to the Theory of Limits 71
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.
- ∣χ 2 - χ - 2∣
x -4
1+ 1 _ 2
1 -4 -3
Jv- _1_ 1
and the smallest value of the fraction χ _ is
so that
—4 < ^γ + ! < 4 Because - S < 4
x - 4 3
and
* + 1 < 4
lx -4 ∣ 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
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
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
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.
Sec. 1.8 Introduction to the Theory of Limits 73
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.
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
C H A P T E R 1 R E V IE W
P R O F I C I E N C Y E X A M IN A T IO N
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
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
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{^i, 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
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:
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.
Computational Window
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.
80 Ch. 1 Preview of Calculus: Functions and Limits
GUEST ESSAY
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.
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
■ TANGENT LINES
■ 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.
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
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
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.
86 Ch. 2 Techniques of Differentiation with Selected Applications
PREVIEW
/.0 1
Xmin= ~10 Y n in = -10
Xnax=10 9∩ax=10
X s c l= l V s c l= l
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.
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.
88 Ch. 2 Techniques of Differentiation with Selected Applications
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.
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)
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
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
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
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
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.
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.
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.
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
■ 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
92 Ch. 2 Techniques of Differentiation with Selected Applications
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
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.”
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
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.
94 Ch. 2 Techniques of Differentiation with Selected Applications
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
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.
=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. ■
Here is a simple rule for finding the derivative of any power function.
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
96 Ch. 2 Techniques of Differentiation with Selected Applications
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
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
Solution
a. Applying the power rule with n = 8, we find that
^ ( x 8) = 8x 8^ 1 = 8x 7
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
Sec. 2.2 Techniques of Differentiation 97
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:
∕ ( 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
p(χ + ∆x)
∖_________ J
= ∖∕( x + ∆x) ∖g(x + ∆x)
__________________ √ _________sz________ √
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)
The last step in deriving the product rule is to take the limit as ∆x → 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
Sec. 2.2 Techniques of Differentiation 99
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:
F(x) = jx ^ 2 - ∣x + 5 + 1 + x ^ 1
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
Example 3 illustrates how the extended linearity rule can be used to differenti
ate a polynomial.
Find the standard form equation for the line tangent to the graph of
2
∕ ω = 2 x3x + 5
2 + x - 3
,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
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)(2x + 4) + (l)(x 2 + 4x - 7) = 2x 2 - 8 + x 2 + 4x - 7
= 3x2 + 4x - 15 = (3x - 5)(x + 3)
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.
■ HIGHER DERIVATIVES
d"y
Λth derivative y(") ∕<">(x)
dx"
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.
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
3. Find the derivatives of the functions given in Problems
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)
4. Find the derivatives of the functions given in Problems
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
Sec. 2.2 Techniques of Differentiation 103
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
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.
104 Ch. 2 Techniques of Differentiation with Selected Applications
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.
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.
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
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
Examples 1 and 2 show how the two special limits in Theorem 2.6 can be used
to compute trigonometric limits.
Find lim
.v→o 5x
*This string of inequalities was suggested by Leonard Gillman in the American Mathematical
Monthly, April 1991.
106 Ch. 2 Techniques of Differentiation with Selected Applications
Find lim ≡ ≠
1 →o sm 7x
= ¾ l)(τ)(l)= 7
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
Sec. 2.3 Derivatives of the Trigonometric Functions 107
≡
Computational Window
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
Differentiate ∕( x ) = x 2sinx.
Differentiate ∕z(∕) = - ^ - ∙
v ’ cos t
Solution Write V t as ∕ ιz2.
= ^ →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/
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
The six basic trigonometric functions sinx, cosx, tanx, cscx, secx, and cotx are
all differentiable wherever they are defined, and
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
Sec. 2.3 Derivatives of the Trigonometric Functions 109
the derivative of the tangent function and leave the rest as problems.
_/ tt a n r
_ dL sinx Trigonometric identity
dx dx cos x
D ifferentiate/(0 ) = 30 sec0.
Solution ∕'( 0 ) = ⅛ (3 0 s e c 0 )
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 ^ ~ ^
= - c s c 2x - 2(- c s c x cotx)
= 2 cscx co tx - csc 2x
110 Ch. 2 Techniques of Differentiation with Selected Applications
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
J , + √3 = 0
Computational Window
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
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
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:
y = A x cosx + B x sinx
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
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
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
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.
Let f(x) = x 2 - 4x + 7.
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 ^e r a don a(f) ~ ~ft = when these derivatives exist. The
speed of the object is ∣v(Z)∣.
decreasing We see th at a(t) < 0 on [0, 4.5) Velocity is decreasing-, that is,
decreasing increasing the object is decelerating.
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
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
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
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?
Sec. 2.4 Rates of Change: Rectilinear Motion 119
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.
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.
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
= 6x3(x2 + 2) t≡m
122 Ch. 2 Techniques of Differentiation with Selected Applications
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
= ∣ ( 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
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:
The chain rule is so important that you need to see several examples, with a
few variations.
124 Ch. 2 Techniques of Differentiation with Selected Applications
Solution = 4 cos3 (3x + 1)2 ^ c o s ( 3 x + 1)2 Power rule for cosine function
Differentiate p(x) = . 1t a n J⅜ .
rv ’ (1 - 4x)5
Differentiate ∕z(x) = V s e c x 3.
Solution Write h(x) = (secx 3) ιz2.
Computational Window
The carbon monoxide level will be changing at the rate of 0.24 parts per
million. B M
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
Sec. 2.5 The Chain Rule 127
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 + -
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
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
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.
130 Ch. 2 Techniques of Differentiation with Selected Applications
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
√x∣*= - 1 (-1 ,3 )
E X A M P L E 2 Implicit differentiation
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)
Sec. 2.6 Implicit Differentiation 131
⅛ U W ) 1 + 2 ⅛ M X) ] ≈ - 3 ⅛ X + 2 ⅛ ^ )
.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≡
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
cos(x 2 + y) ∣
2x + = y 2(3) + (3x + l)^2y
132 Ch. 2 Techniques of Differentiation with Selected Applications
dy _ 3y2 - 2x cos (x 2 + y)
dx cos(x 2 + y) - 2y(3x + 1) B H
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
d*y= ± ( - χ ∖ = y ⅛ -⅛ -< ~ x ⅛ y = - JZ + X ⅜
dx 1 dx∖ y ) y2 y2
Quotient rule
Sec. 2.6 Implicit Differentiation 133
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
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
13 4 Ch. 2 Techniques of Differentiation with Selected Applications
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
(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)
Sec. 2.6 Implicit Differentiation 135
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
Figure 2.26 Problem 42
X 2 + y 2 = 6y - 10
-A
—( — ∣κv ∖ι k1-
-i / x
⅞*
a2
W
b2
= 1
-2--
a2 b2
13 6 Ch. 2 Techniques of Differentiation with Selected Applications
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.
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
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:
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
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.
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
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.
EXAM PLE 5 Related rates
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
140 Ch. 2 Techniques of Differentiation with Selected Applications
Let V denote the volume of water in the tank after t minutes. We know that
Jz = ⅛πr2h
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.
Sec. 2.7 Related Rates 141
A general relationship is
cot θ = ∣x
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
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
142 Ch. 2 Techniques of Differentiation with Selected Applications
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
Sec. 2.7 Related Rates 143
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?
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
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.
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:
∕'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,
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.
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
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).
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.
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 ™
■ ERROR PROPAGATION
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
150 Ch. 2 Techniques of Differentiation with Selected Applications
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 %.
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
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
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.
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.
152 Ch. 2 Techniques of Differentiation with Selected Applications
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:
e. The actual revenue (nearest cent) obtained from the sale of the 9th unit is
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?
Sec. 2.8 Differentials and Tangent Line Approximations 153
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
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 +
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
a. Estimating a root, r, of y = f∖ x)
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
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.
Computational Window
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_ -.— ........
Sec. 2.9 The Newton-Raphson Method for Approximating Roots 157
4. Repeat step 3 until ∣xπ+1 - x j < ε. The estimate x = x π then has the
required accuracy.
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
∣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.
Historical Note
root sought.
Figure 2.53 The Newton-Raphson method (part a) might fail (part b).
© 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?
160 Ch. 2 Techniques of Differentiation with Selected Applications
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
PROFICIENCY EXAMINATION
SUPPLEMENTARY PROBLEMS
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)∙
162 Ch. 2 Techniques of Differentiation with Selected Applications
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
Chapter 2 Review 163
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
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
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.
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
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.
Using the graph in Figure 3.3, locate the extreme values of the function f
defined on the 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
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
168 Ch. 3 Additional Applications of the Derivative
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].
a 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.
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).
Sec. 3.1 Extreme Values of a Continuous Function 169
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
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.
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):
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
Sec. 3.1 Extreme Values of a Continuous Function 171
EXAM PLE 5 Critical values where the derivative does not exist
■ 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.
EXAM PLE 7 Absolute extrema when the derivative does not exist
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.
= ⅛(¾+ ⅛) + 2 (λ f ) - y = ∣ + V 3 - f ≈ 1-309853256
■ OPTIMIZATION
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)
V'(b) = 0
Figure 3.10 Volume of a box
206 - 36 2 = 0
6(20 - 36) = 0
6 = 0 ,f
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
Sec. 3.1 Extreme Values of a Continuous Function 175
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]
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.
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
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.
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
The negative value is not in the open interval (1, 2), but the positive value
1 + √ 31 ≈ 1.522588121
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).
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. ■
b- a
Now we are ready to prove the mean value theorem.
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
180 Ch. 3 Additional Applications of the Derivative
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.
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. ■
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].
Sec. 3.2 The Mean Value Theorem 181
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
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
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)
3 .3 FIRST-DERIVATIVE TEST
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.
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.
a. Intervals where/is increasing and where it is decreasing b. The graph of/along with the sign graphs fo r/a n d /'
The function / i s increasing for x < -1 and for x > 3 ; / i s decreasing for
-1 < x < 3. —
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/'?
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.
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.
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.
Sec. 3.3 First-Derivative Test 187
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
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.
Computational Window
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
Sec. 3.3 First-Derivative Test 189
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
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 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
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
192 Ch. 3 Additional Applications of the Derivative
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.
■ 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.
■ 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.
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
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
Sec. 3.4 Concavity and the Second-Derivative Test 195
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.
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
1 96 Ch. 3 Additional Applications of the Derivative
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].)
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.
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
EXAM PLE 5 Using the first and second derivative to graph a function
Computational Window
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.
/ (-1 ) = 3 / (0 )= 0 ∕Q ) = λ^ = ≈ 3.7 7 9 7 6 3 15
198 Ch. 3 Additional Applications of the Derivative
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.
b. Completed sketch
Figure 3.37 Graph of T(x) = sin x + cosx on [0, 2π]
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.
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)
200 Ch. 3 Additional Applications of the Derivative
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.
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.
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.
Sec. 3.4 Concavity and the Second-Derivative Test 201
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?
202 Ch. 3 Additional Applications of the Derivative
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
Computational Window
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 ....
■ 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.
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'→ —∞
(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.
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 ,
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→+∞.
The following theorem will allow us to evaluate certain limits to infinity with
ease.
lim j ⅜l = 0
X'
A → +∞
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 /
206 Ch. 3 Additional Applications of the Derivative
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
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.
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
Sec. 3.5 Infinite Limits and Asymptotes 207
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→+∞
= +∞
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
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.
Λ → -Γ ∞ A Λ → -Γ ∞ ∖ A /
208 Ch. 3 Additional Applications of the Derivative
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.
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,
are infinite limits. The graphs of these functions are shown in Figure 3.46.
Sec. 3.5 Infinite Limits and Asymptotes 20 9
lim2~ x -~x2 , = -∞
lim x -~~2 = +∞ and χ→
Λ → 2÷
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∙
Solution
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
Sec. 3.5 Infinite Limits and Asymptotes 211
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
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
■ G EN E R A L CURVE SKETCHING
Find the first and second derivatives, along with the critical values for y ' and
y"∙
Derivatives Critical values
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.
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
Sec. 3.6 Summary of Curve Sketching 217
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.
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
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
■ 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.
220 Ch. 3 Additional Applications of the Derivative
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.
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
y = 4 -⅛ x
We can now write O as a function of x alone:
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 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 .
A = ⅛bh
In this case, the base b is 4 and, because sin θ = the height h is 4 sin θ. Thus,
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
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
Sec. 3.7 Optimization in the Physical Sciences and Engineering 223
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.
_____ x_____ _ 1
45√ χ 2 + 1,600 75
5x —3Vx2 + 1,600
225Vx 2 + 1,600
_ √ 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. ≡≡≡
Wire, length L
Figure 3.57 Forming a circle and a square from a wire with length L
= 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
O.O35OO6 197L2
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≡≡
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
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.
Sec. 3.7 Optimization in the Physical Sciences and Engineering 227
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
228 Ch. 3 Additional Applications of the Derivative
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.
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.
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.
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.
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
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.
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.
x = 20
Thus, the price that corresponds to the maximum profit is
jp(20) = 49 - 20 = 29 dollars H
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).
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.
⅛x + 4 = ∣x + 4 + 200x~l
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.
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.
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.
Sec. 3.8 Optimization in Business, Economics, and the Life Sciences 235
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.
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.
236 Ch. 3 Additional Applications of the Derivative
/(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
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?
238 Ch. 3 Additional Applications of the Derivative
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]
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.
∕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.
_ 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
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
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
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
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
Computational Window
Evaluate lim —— i
Λ-→O x
Solution This is a 0/0 indeterminate form, and we find that
Evaluate lim 1
x →o secx
Solution
0 If you apply ΓH6pitaΓs rule blindly, you obtain the W R O N G answer:
=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
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 ⅛ = 1
ll→0 1 MM
■ INDETERMINATE FORM ∞ ∕∞
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
Solution The limit is indeterminate of the form ∞∕∞. By applying l, HδpitaΓs rule,
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'
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.
244 Ch. 3 Additional Applications of the Derivative
Evaluate lim ( - -
χ → 0 '∖ X
∙J, ,).
v
SHIA./
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
Evaluate lim -x + s *n x .
%→+∞ x - cosx
Solution This limit has the indeterminate form ∞∕∞. If you try to apply
ΓH6pitaΓs rule, you find
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
3 . 1 0 ANTIDERIVATIVES
■ 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.
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
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
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
■ ANTIDIFFERENTIATION FORMULAS
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
x 3z2
= 5- Γ^ + 4(-cosx) + C
= -yX3z2 - 4 COSX + C
■ APPLICATIONS
= 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
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 ( - 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
EXAM PLE 7 Finding the demand function given the marginal revenue
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 _
Chapter 3 Review 251
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
PROFICIENCY EXAMINATION
P r a c tic e P r o b le m s m = 2o - τ f j
SUPPLEMENTARY PROBLEMS
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
x →+∞ 6x + 7 Vx
Λ∙→ +≈ X + 1 J
Chapter 3 Review 253
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
Computational Window
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
Chapter 3 Review 255
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.
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)
This project is to be clone in groups o f three or four students. Each group will submit
a single written report.
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.
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.
■ 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
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.
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
∆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 _
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.
260 Ch. 4 Integration
Computational Window
38.375
39.84375
40.58594
Sec. 4.1 Area As the Limit of a Sum; Summation Notation 261
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
262 Ch. 4 Integration
index k is called the index of summation (or running index). The terminology used
in connection with the summation notation is shown:
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 /
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
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
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( ZL±jΛ ( 2n + A f ι + 1V 2 + 1
1. n∖ n J∖ ) 1. ∖ n)∖ n
= lim ---------- τ---------- = lim --------- τ-------
n→+∞ 0 ∕t→+∞ O
264 Ch. 4 Integration
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
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
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.
266 Ch. 4 Integration
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.
Given
Given
∣t ∕< A =
k x k " ⅞ -ι = ∆χ k x k 2(4)
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
∣∣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
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
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:
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-
270 Ch. 4 Integration
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.
Our next example illustrates how to use the definition to find a definite
integral.
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
= lim -⅛ z
∑ ( - 2 H + 3∕<)
π →+≈ ∕γ fc=]
= 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.
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
[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
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.
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
= 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
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.
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
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.
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
Ja
f( x ) d x =
Ja
f( x ) d x +
35.
J a
∕(x)g(x) dx =
'-Ja
∕( x ) dx
-^ ∙-J a
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
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.
Finally, we take the limit as ∣∣P∣∣ → 0, and because F(b) —F(a) is a constant, we
have
Γ1
Evaluate 4x dx using the definition of the definite integral and also using
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'
= hm 1 2 (- 4 n 2 + 3 n 2 + 3n∖
n→ +∞ n ∖ Z /
= -6
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
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
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
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
Integration Formulas
Constant rule: ∣ 0 du = C
Power rule: u n du - , + C; ∏Ψ 1
, 77+ 11
cos u du = sin u + C
∣ sec2 u du = tan u + C
1 CSC 2 M du = -cotw + 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.
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
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.
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,
Sec. 4.3 The Fundamental Theorem of Calculus; Integration by Substitution 281
∕ ω = ⅛[G(<oι = G '(> )g = s ( u ) ⅛
as required.
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
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
9u8 du
Find J V 3 x + 7 dx.
J V 3 x + 7 dx = j V u = j ∣ MIZ2 du
0 After you have made your substitution and simplified, there should be
no leftover x-values in the integrand. 0
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
Computational Window
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).
x dx
Find
V x 2 + 1’
/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
If you wish to check your work with integration problems, you can check by
using differentiation:
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?
284 Ch. 4 Integration
u = 3t ÷ 1
du = 3 dt
V + C l = ∣ J u l/2 du
Notice that the definite integral eliminates the need for finding C. Furthermore,
the following theorem eliminates the need for returning to the original variable.
Because F is an antiderivative
o ff ■
Sec. 4.3 The Fundamental Theorem of Calculus; Integration by Substitution 2 85
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
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
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
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 √ ∕
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
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
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.
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
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
‘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.
Sec. 4.4 Introduction to Differential Equations 291
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
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
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
-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^.
Sec. 4.4 Introduction to Differential Equations 293
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.
294 Ch. 4 Integration
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
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.
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
296 Ch. 4 Integration
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.
Find a value of c guaranteed by the mean value theorem for integrals for
∕( x ) = sinx on [0, τr].
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.
Sec. 4.5 The Mean Value Theorem for Integrals; Average Value 297
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
∆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'
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.
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
298 Ch. 4 Integration
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.
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
The fundamental theorem of calculus tells us that the definite integral can be
evaluated as a difference:
= ∕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
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 -
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
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
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
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>
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
■ APPROXIMATION BY RECTANGLES
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
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
304 Ch. 4 Integration
Our first example uses the trapezoidal rule to estimate the value of an integral
that we can compute exactly by using the fundamental theorem.
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
---------
■ 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-
Sec. 4.6 Numerical Integration: The Trapezoidal Rule and Simpson's Rule 305
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
By adding the area of the approximating parabolic strips and combining terms, we
obtain the sum S n of n parabolic regions:
+ + 3 [∕( ⅞ - 2) + 4∕( ∙V ι) + ∕( x χ
n)]∆
*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.
306 Ch. 4 Integration
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. ■
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.
Sec. 4.6 Numerical Integration: The Trapezoidal Rule and Simpson's Rule 307
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.
308 Ch. 4 Integration
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.
J 2
"2
V l ÷ sinx dx with n = 4
x^ 1 dx
θ
J o
x cosx dx with n = 6
J o
cos2 x dx accurate to three decimal places
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.
310 Ch. 4 Integration
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
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
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
312 Ch. 4 Integration
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.
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
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
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
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.
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.
Sec. 4.7 Area Between Two Curves 315
In the following example, the net earnings of a machine are calculated as a definite
integral and interpreted as the area between two curves.
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
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.
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
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
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
Chapter 4 Review
PROFICIENCY EXAMINATION
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.
SUPPLEMENTARY PROBLEMS
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
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.
322 Ch. 4 Integration
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
324 Ch. 4 Integration
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
■ EXPONENTIAL FUNCTIONS
Furthermore, if b ≠ 0, then
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.
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
x = lim rn
n→+∞
Then, bx is defined by
bx = lim br"
n→+∞
The bracketing theorem follows from these definitions.
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.
328 Ch. 5 Exponential, Logarithmic, and Inverse Trigonometric Functions
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 < b < 1 is shown in Figure 5.3b. In particular,
notice that in this case, the reciprocal c = ∖ ∕b satisfies c > 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.
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. ■
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
Sec. 5.1 Exponential Functions: The Number e 329
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.
■ NATURAL BASE e
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
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
Sec. 5.1 Exponential Functions: The Number e 331
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'
Figure 5.4 The growth of an account over a one-year period with different compounding
frequencies
332 Ch. 5 Exponential, Logarithmic, and Inverse Trigonometric Functions
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
$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
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)
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.
334 Ch. 5 Exponential, Logarithmic, and Inverse Trigonometric Functions
■ 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
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
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
for all x. M M
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.
■ 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.
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.
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 ).
Suppose/is strictly monotonic and differentiable for all x in its domain. Then
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
338 Ch. 5 Exponential, Logarithmic, and Inverse Trigonometric Functions
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:
To check this result, note that because f(x ) = V x , we have f 1(x) = x 3. Thus,
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
Sec. 5.2 Inverse Functions; Logarithms 339
so (1, 0) and (b, 1) lie on the logarithmic curve. Theorem 5.6 lists several general
properties of logarithms.
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.
Solution You can evaluate this on your calculator, or you can use Theorem 5.6
□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
5x = log 10 195 - 3
log 10 195 - 3
5
≈ —0.141 993 077 7 By calculator
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.
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
Sec. 5.2 Inverse Functions; Logarithms 341
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.
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.
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)
342 Ch. 5 Exponential, Logarithmic, and Inverse Trigonometric Functions
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.
Computational Window
≈ 0.0168236
Sec. 5.2 Inverse Functions; Logarithms 343
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
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
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).
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
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
⅞ = —X τ (lθ * + 2) 1Ox + 2
dx 5x 2 + 2x + Γ ,
5x 2 + 2x + 1
Figure 5.10 Graph of
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.
346 Ch. 5 Exponential, Logarithmic, and Inverse Trigonometric Functions
T H E O R E M 5.10 Derivative of In ∣
w∣
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. ■
Differentiate ∕( x ) = sin x + V in x.
Solution Write ∕( x ) = sinx + (In x ) ιz2.
f (x) = cosx + ∖ (In x) ιz2 Q = cosx +
Differentiate ∕( x ) = (x2 + ln V x) 3.
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
∣ _ 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 ■
Solution e χ l + x (2x + 1)
≡ i^ = (in *)*-
The second part of this theorem follows from the chain rule.
348 Ch. 5 Exponential, Logarithmic, and Inverse Trigonometric Functions
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 ( 10¾χ ) = ( 1 ⅛
■ LOGARITHMIC DIFFERENTIATION
product or quotient functions and power functions where variables appear in both
the base and the exponent.
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
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)
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
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
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 →+ ∞ .
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
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
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
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.
1 2,2
Figure 5.12 Preliminary sketches for the graph of ∕( x ) = ~ y = e^x
χ 2ττ
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
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
354 Ch. 5 Exponential, Logarithmic, and Inverse Trigonometric Functions
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.
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.
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 < p < 50 and
Figure 5.15 The graph of the con for p > 50:
sumer expenditure function
£(p) = 5,OO(⅛<ro o ⅛ If 0 < p < 50, then E '(p) = 5,000f-° 0⅛(l - 0.02∕t) > 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.
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.
= 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.
356 Ch. 5 Exponential, Logarithmic, and Inverse Trigonometric Functions
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?
358 Ch. 5 Exponential, Logarithmic, and Inverse Trigonometric Functions
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?
■ INTEGRATION OF e x AND in x
J ex dx = ex + C
Proof: This follows directly from the differentiation formula
⅛(<--) - <"
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 ■■■
± d x = ln∣x∣ + C
j
Proof: This proof follows directly from Theorem 5.10. Because
we have I dx = ln∣x∣ + C _
{n+ 1
ln∣x∣ + C n = -1
360 Ch. 5 Exponential, Logarithmic, and Inverse Trigonometric Functions
F ta d ∫ ⅛ ⅛
Computational Window
Find sinx dx
3 + 2 cosx
Solution Let u = 3 + 2 cosx, SO that du = - 2 sinx dx.
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
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.
The formula for the integral of the cotangent may be derived in a similar fashion
(see Problem 62). ■
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.
PROBLEM SET 5 .5
θ Find the given integral in Problems 1-32.
35. I
pin 2 ,
∕( < V - e~x )dx
f 2
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.
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.
364 Ch. 5 Exponential, Logarithmic, and Inverse Trigonometric Functions
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.
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
The definition of inverse functions yields four formulas, which we call the
inversion formulas for sine and tangent.
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
cos l x = y - sin 1x
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
∕ ,ω = ⅛ ta n -1 vχ)
b. Let ιι = (1 - Z);
^ '( Z ) = ⅛ s in - 1 (l -Z)]
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.
θ = (α + θ) —a = cot 1
~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.
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
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
dx
Find
V9 - 4 V
Sec. 5.6 The Inverse Trigonometric Functions 371
_ f | du
J √ 3 2 - ιι2
= 1 s ∙i n” 1 ∕1z√
β∖I + C From the formula with a = 3
= sin” 1 + C
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
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.
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 )
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
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
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
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
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.
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.
x dt
Historical Note THEOREM 5.20 Properties of a logarithm as defined by In x = . t
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.
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).
376 Ch. 5 Exponential, Logarithmic, and Inverse Trigonometric Functions
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.
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
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.
378 Ch. 5 Exponential, Logarithmic, and Inverse Trigonometric Functions
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
PROFICIENCY EXAMINATION
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
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
SUPPLEMENTARY PROBLEMS
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
380 Ch. 5 Exponential, Logarithmic, and Inverse Trigonometric Functions
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
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∣ < P that
63. I 64. 1
x + x(ln x) 2 x - x(ln x ) 2
■
4 - sec lx = J = for ∣x∣ > 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)'
Chapter 5 Review 381
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.
96. P U T N A M E X A M IN A T IO N P RO BLEM
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
382 Ch. 5 Exponential, Logarithmic, and Inverse Trigonometric Functions
"Quality Control”
This project is to be done in groups o f three or four students. Each group will submit
a single written report.
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.
*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.
383
384 Ch. 6 Additional Applications of the Integral
Cube Side 5 V = s3
Z
cone Circular base of Y = j π r 2h
radius r
Cylinder Height h
A
Area of base B V = Bh
Λ
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 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
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
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
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
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
V= A(x) dx = ττ[∕(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.
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 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.
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
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
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
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
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.
Length (Height)
Distance to axis
Distance to axis
Sec. 6.1 Volume: Disks, Washers, and Shells 393
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
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
X 0 1 2 3 4 5
(x - a)2 + y 2 = b 2
F=∣(.1, + ∖ < ξ ξ + ,l 2)
Figure 6.14 Volume o f a section of a sphere
= √(Δ⅞)2 + ( ∆ γ j 2
U ⅞ ) 2 + (∆yfcj 2
(Δ⅞)2
∑ ∖ i + [ ∕'( 4 ) ] 2 δ ⅞
fc=ι
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.
398 Ch. 6 Additional Applications of the Integral
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
Find the arc length of the curve y = x 3'2 on the interval [0, 4].
Solution Let / ( / = x 3Z2; therefore,/'/) = ∣ x ιz2.
⊂3 Computational Window ▲
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
Sec. 6.2 Arc Length and Surface Area 399
Computational Window
--------------------- - 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.
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
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.
400 Ch. 6 Additional Applications of the Integral
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.
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].
Let it = 1 + 9x 4 ; du = 36 x i dx.
If x = 1, then ιι = 10.
If x = 0, then u = 1.
EXAM PLE 5 Derive the formula for the surface area of a sphere
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
given in Problems 1-12.
'^ 1. fix ) = 3x + 2 on [ - 1 , 2] 5. f ix ) = ∣ x 3z2 + 1 on [0, 4]
9. f ix ) = ∣ x 4 + ∣x ^ 2 on [1, 2]
402 Ch. 6 Additional Applications of the Integral
y = ⅜x3 + (4x)^l
Give your answer correct to the nearest hundredth. between x = 1 and x = 27.
Sec. 6.2 Arc Length and Surface Area 403
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 .
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?
404 Ch. 6 Additional Applications of the Integral
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
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
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.
‘ 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.
406 Ch. 6 Additional Applications of the Integral
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
*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.
Sec. 6.3 Physical Applications: Work, Liquid Force, and Centroids 407
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
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.
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.
408 Ch. 6 Additional Applications of the Integral
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.
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
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
Sec. 6.3 Physical Applications: Work, Liquid Force, and Centroids 409
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)
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
The fluid force formula applies equally well if the plate is submerged in a fluid
whose density varies with the depth h.
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
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
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°
= 532,480 lb
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
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
and
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.
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.
Sec. 6.3 Physical Applications: Work, Liquid Force, and Centroids 413
Axis of revolution
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
414 Ch. 6 Additional Applications of the Integral
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?
Sec. 6.3 Physical Applications: Work, Liquid Force, and Centroids 415
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
+ 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.
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
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
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≡
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
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 ≡
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 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
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
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
± ln Q
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.
420 Ch. 6 Additional Applications of the Integral
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
Sec. 6.4 Growth, Decay, and First-Order Linear Differential Equations 421
The following theorem shows how the general solution of a first-order linear
differential equation y ' + P(x)y = Q(x) may be obtained.
+ P(x)j = <2(x)
is given by
y = 7 ⅛ [/ 2( χ )z (λ∙) dx + C
We now begin the proof by starting with the given differential equation.
∕U ) + Kχ)P{×)y = Jr(χ)Q(χ)
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
5 0 + l l b ⅛a l
Sec. 6.4 Growth, Decay, and First-Order Linear Differential Equations 42 3
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
- p o ⅛ P ( 5θ÷')≈ + q
= 2 (5 0 + '> + ( 5 δ ⅛
^(°) = 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
77
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.
Sec. 6.4 Growth, Decay, and First-Order Linear Differential Equations 425
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
426 Ch. 6 Additional Applications of the Integral
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
Figure 6.38 Problem 42
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
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
y ,Λ ⅞ >
2
cb
Shells: v-axis Σ ¾ ∕( x * ) ∆ x i Volume = 2τr 1 x∕(x) dx
fc=l Ja
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
PROFICIENCY EXAMINATION
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].
SUPPLEMENTARY PROBLEMS
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
430 Ch. 6 Additional Applications of the Integral
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
Chapter 6 Review 431
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.
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?
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
72. P U T N A M E X A M IN A T IO N P RO BLEM
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
434 Ch. 6 Additional Applications of the Integral
"Houdini’s Escape”
This project is to be done in groups of three or four students. Each group will submit
a single written report.
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
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.
■ 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
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.
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'' ’
.-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
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 τ ⅛ = f ≠ ⅛ ⅞ = ∖ ^ f Γ = -ln∣M∣ + C = -ln(e-'- + 1) + C
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.
_ ! 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
440 Ch. 7 Methods of Integration
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 ,
Computational Window
(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:
Table of Integrals
INTEGRALS INVOLVING UU + b
— ~7-7 = - ln[αw + b∖
au + b a 1
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.
= x(ln x)4 —4^x(ln x)3 —3 J (In x) 3-1 dx^ Formula 501 again
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
444 Ch. 7 Methods of Integration
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.
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).
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
Sec. 7.1 Review of Substitution and Integration by Table 445
We now show how one of the formulas in a table of integrals might be derived
by using substitution.
Formula 403
=taιΓax
2a
+ - ln∣cosαx∣
a ' 1
+ C
■ TRIGONOMETRIC SUBSTITUTIONS
Find J V 4 - x 2 dx.
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
446 Ch. 7 Methods of Integration
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)
= 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
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
Sec. 7.1 Review of Substitution and Integration by Table 447
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
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
sin 2 x dx = ∣ x - ∣ sin 2x
cos2 x dx = ⅛x + ∣ sin 2x
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
∙ ∙
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,
ιιv = J u dv + J v du
If we rewrite this last equation, we obtain the formula summarized in the
following box.
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.
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
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 .
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
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 √
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.
x e l x dx.
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
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.
Sec. 7.2 Integration by Parts 453
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
454 Ch. 7 Methods of Integration
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
■ DECOMPOSITION
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
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
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
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’
—3x*3 — V
Decompose 2 2.
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.
= l∏∣x - 2∣ + ξ ⅛i + 2 & ~ 2 )2 + C
Sec. 7.3 The Method of Partial Fractions 459
Find f -Γ 4 + 2A-; - 4x 2 + A - 3 d χ
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
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
Find 5A 2 + 2 1A + 4
(A + 1)2(A - 3)
Find [ x 22 j^ ⅛
x ~ 23 dx.
J (x + 4)(x + 3)
{
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.
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
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
The German mathematician Karl Weierstrass (1815-1897; see the historical note
on page 59) noticed that the substitution
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
Weierstrass Substitution For -ττ < x < IT, let u = ta∏y so that
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 ’
_ 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
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
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
Computational Window
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 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)
464 Ch. 7 Methods of Integration
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
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).
466 Ch. 7 Methods of Integration
Indicate a procedure to set up each integral. It is not necessary to carry out the
integration.
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
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+ e^i 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
468 Ch. 7 Methods of Integration
J tan m x sec"x dx
45. f
I X√ X 2 + 1 J
2
46. I∣ x V x + 1 dx
7 .5 IMPROPER INTEGRALS
rN
Improper Integrals (First Type) Let a be a fixed num ber and assume ∕( x ) dx exists for all N ≥ a. Then if
dx
v2∙
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.
Evaluate
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.
+
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
Thus, the improper integral diverges for p ≤ 1 and converges for p > 1. ≡ ≡
√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
Evaluate xe 2x dx.
'o
1
4
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.)
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
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→ -∞
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
474 Ch. 7 Methods of Integration
∕< 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
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.
Sec. 7.5 Improper Integrals 475
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
= 3 h m [ ( i - l ) 173- ( - l ) ] = 3
That is, the improper integral converges and has the value 3.
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 = —∞
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
Sec. 7.5 Improper Integrals 477
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.
478 Ch. 7 Methods of Integration
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
■ 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 )
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.
The list of properties in the following theorem shows how the basic hyperbolic
functions are similar to the trigonometric functions.
Proof: We will verify the first identity and leave the others for the reader.
_ e 2x + 2 + e~2x _ e 2x —2 + e~2x _
4 4
coth x = -—⅛— = e x + e - x
tan h x ex - e x
sech x = — V— = γ ,~ -77
cosh x ex + e
Rules for differentiating the hyperbolic functions are listed in Theorem 7.2.
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
Proof: The proof of each of these formulas follows directly from the corre
sponding derivative formula. ■
f 7√4 1
∫ cosh x sinhx dx =
3 w3 du = + C = ∣ cosh4 x + C
Let u = coshx;
dιι = sinh x dx.
Let u = x 2;
du = 2x dx.
= ∣ 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).
482 Ch. 7 Methods of Integration
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) ■
[ . /------- 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
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
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
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
Chapter 7 Review
PROFICIENCY EXAMINATION
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
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
30. Find the volume of the solid generated when the region
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
SUPPLEMENTARY PROBLEMS
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 )
486 Ch. 7 Methods of Integration
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
J τ
e^ , tx ~' dt < 0.005
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 .
488 Ch. 7 Methods of Integration
"Buoy Design”
This project is to be done in groups of three or four students. Each group will submit
a single written report.
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
You should make the following assumptions:
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
Your paper is not limited to the following questions but should include these
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
■ CONTENTS PREVIEW
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
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
Find the 1st, 2nd, and 15th terms of the sequence {a }, where the general term
is
^ = 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। ∙ ∙ ∙
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
« = — 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
492 Ch. 8 Infinite Series
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.
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
∫ 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. {(-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 +
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
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.
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
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
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.
Proof: Let ε > 0 be given. Because lim ∕'(x) = L, there exists a number
N > 0 such that
Sec. 8.1 Sequences and Their Limits 495
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).
Figure 8.9 Graphical comparison of lim a and lim ∕( x ) , where f(ri) = a for n = 1, 2, . . .
11
M —»20 r → ≡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.
Show that the following sequences converge, and find their limits.
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
⅛ bn f
0 ≤ - ⅛n ≤ - ∙⅛ ≤ - because n ∙ n∖ < πn
n n πn n
Decreasing Decreasing if α1 > α9 > α3 > ∙∙∙ > a jt-1 > ak > ∙∙∙
Nonincreasing Nonincreasing if α1 ≥ a 2 ≥ ≥ ak _ { ≥ ak ≥
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.
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.
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
498 Ch. 8 Infinite Series
e2 is between 7 and 8). We see that the sequence is bounded from below
because
0 < U 114 for all n > 1
V/7
Therefore, the given sequence is bounded from below and is eventually
decreasing, so it must converge. ■■■■
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. ■
Sec. 8.1 Sequences and Their Limits 499
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
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
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
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
≈ ∣
Show that the senes ∑ converges.
A-= ι 4
502 Ch. 8 Infinite Series
S' = l ⅛ l = i
*-, 2 2 4 4
S' =
d - 1 + 1+ 1= 2
3 2 4 8 8
$n ~ 2 + 4 + ■" + 2"
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 /
∑ ( - l) * = -1 + 1 - 1 + 1 - 1 + 1 - ∙ ∙ ∙
⅛ =1
X 1
Show that the series ∑ -∩---- τ converges and find its sum.
k=∖ k~ + k
n+ 1
Sec. 8.2 Introduction to Infinite Series: Geometric Series 503
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
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. ■
∞ 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 - '
We can now use Examples 1 and 3 to conclude that the series converges and
that the sum is
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.
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
504 Ch. 8 Infinite Series
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
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
Sec. 8.2 Introduction to Infinite Series: Geometric Series 505
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. ■
y √ - B5 t - a = 25 - X
ι⅛ ' 1- r 1 _ (_1) 10
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,
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.
— 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
506 Ch. 8 Infinite Series
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.
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
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 →∞
10e~ιz5
1 - e^ ιz5
≈ 45.166 556
We see that about 45 units remain in the patient’s body.
/ 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.
508 Ch. 8 Infinite Series
, ” 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
■ DIVERGENCE TEST
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.
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∞
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.
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→∞
Sec. 8.3 The Integral Test: p-Series 511
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.
∑ 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
512 Ch. 8 Infinite Series
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 %Λ^ia, 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
Now, suppose the improper integral ∕( x ) dx converges and has the value /;
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
∞ Λ-
Test the series ∑ for convergence.
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
= 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.
514 Ch. 8 Infinite Series
∞ 1
The p-series y converges if p > 1 and diverges if 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 → ∞.
E X A M P L E 4 p-series test
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
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.
516 Ch. 8 Infinite Series
8 . 4 COMPARISON TESTS
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.”
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.
the divergent p-series ∑ τ∣p (divergent because p = ∣ < 1). The direct
*=ι κ ~
comparison test tells us the given series must diverge.
Computational Window
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
∞ 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
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.
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.
Sec. 8.4 Comparison Tests 519
∞ 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. ■■
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
7/c + 100
k li e k,5 (7k + 100)
lim e - 70
k→∞ k k(e k,5 - 70)
e k' 5
520 Ch. 8 Infinite Series
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.
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.
Proof: The first part of the proof is outlined in Problem 62; the second part
follows similarly. ■
∞ 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
Because Σl∕Zd, converges (∕ι-series with p > 1), the series converges by the
zero-infinity comparison test.
Sec. 8.4 Comparison Tests 521
∞ ∞ |
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
5. Σ 1.5A 6. ∑ 2 w2 7Q
2 ^∙ λV 1 30 V 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.
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
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^∙
∞ ∩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
∞ L-∕c
Test the series ∑ for convergence.
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.
Sec. 8.5 The Ratio Test and the Root Test 525
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.
Proof: The proof of this theorem is similar to that of the ratio test.
∞ 1
Test the series 7j—ττ⅛ for convergence.
∞ 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
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=>
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.
Sec. 8.5 The Ratio Test and the Root Test 527
TABLE 8.1
Guidelines for Determining Convergence of the Series Xu, for ak > 0
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
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.
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∕^ik — 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. ■■
∞ ∕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.
PROBLEM SET 8 .5
θ 1 . 1 What Does This Say? Compare and contrast the ratio 6 ∙ f∞ ⅛
τk y k 8'
∞ ')k
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?
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:
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.
1. lim a.κ = 0
k→∞
2. {ak} is a decreasing sequence; that is, ¾ +1 < ak for all k.
∑ , ( - 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)
Moreover, because
(^ 2
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
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.
_ , . ≈ ( - l) i+ 1 In k c
Test the series ∑ ------- -------- for convergence.
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
≈ ∕- ] U + l
Test the series ∑ -----_ . , for convergence.
⅛=∣ tan k
lim a, = lim 1 l= 2 ≠ 0
∕<→≈ * k→∞ .tan -1 k, 77 77
2
∞ (^~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.
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. ≡≡≡
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
Sec. 8.6 Alternating Series; Absolute and Conditional Convergence 533
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.
534 Ch. 8 Infinite Series
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
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. ⊂~~)
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.
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
Sec. 8.6 Alternating Series; Absolute and Conditional Convergence 535
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
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,
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
536 Ch. 8 Infinite Series
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
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
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
∞ ')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.
If L = 1, the test fails.
8 .7 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.
54 0 Ch. 8 Infinite Series
∞ 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
X
Show that the power series ∑ ^∙χ k converges only when x = 0.
*= ι
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
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
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 < u < R) and diverges for ∣M∣ > 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.
542 Ch. 8 Infinite Series
∞ 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
Find the radius of convergence and the interval of convergence for the power
series
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.
Then:
Proof: The proof is similar to the proof of Theorem 8.23 and is omitted. ■
,,— 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)
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).
544 Ch. 8 Infinite Series
∕( 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
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.
Sec. 8.7 Power Series 545
∕'(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.
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
(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 =
Sec. 8.7 Power Series 547
∑ 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
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
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 ) .
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
■ TAYLOR’S TH EO REM
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?
β k ∖ x) = k∖ak so that ak = -
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)
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.
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 + ∙ ∙ ∙
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 +
E X A M P L E 1 Maclaurin series
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
e ≈ l + l + ⅛+ ⅜+ ⅛ + i⅛ + Λ 5(l)
= 2.716+ Λ 5(1)
Sec. 8.8 Taylor and Maclaurin Series 553
for some z i between 0 and 1. Because 0 < z5 < 1, we have e" 5 < e and
∕'( x ) = ⅛ ∕' ( i ) = ι
∕''( χ ) = ⅞r ∕" ( 0 = - ι
∕ ,4 V ) = ⅛ ∕ ( 4>(1) = - 6
A
_ ≈ ( - 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 ∣
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'
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.
The next example illustrates a function whose Maclaurin series does not
represent that function on any interval.
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
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.
-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 | 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
Sec. 8.8 Taylor and Maclaurin Series 55 7
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
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.
558 Ch. 8 Infinite Series
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
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 , 1 _ 2x 2 , 2 3x 4 _ 2 5x 6 ,
2 2 2 !4 ! 6!
= 1 —x 2 + ∣ x 4 - ⅛x 6 + ∙ ∙ ∙ for all x
Sec. 8.8 Taylor and Maclaurin Series 55 9
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,
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
560 Ch. 8 Infinite Series
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.
p ( p - l ) - ( p → + 1)
+ k∖ x +
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 + ( —∞ , ∞ )
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]
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. ■
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
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
Sec. 8.8 Taylor and Maclaurin Series 563
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.
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 =
65. Find the Maclaurin series for
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 ^-
66. Find the Maclaurin series for
50. Use the Maclaurin series for ex and e~x to find the
Maclaurin series for ∕( x ) Λ- + j i nx
=
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/'.
Chapter 8 Review 565
C H A P TER 8 R E V IEW
PROFICIENCY EXAMINATION
SUPPLEMENTARY PROBLEMS
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)
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 ,
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
56 8 Ch. 8 Infinite Series
This project is to be done in groups of three or four students. Each group will submit
a single written report.
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?
Your paper is not limited to the following questions but should include these
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.
*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
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 ) .
Rotate 0 = - y
6
Sec. 9.1 The Polar Coordinate System 571
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.
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
The relationship between polar and rectangular coordinates can be found by using
the definition of trigonometric functions (see Figure 9.3).
y = - 3 sin¾r = - 3 ^ - ^ = -=⅛p-
∕3 V z2 3V2∖
The rectangular coordinates are (~⅜~, )∙
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 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.
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.
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 )
574 Ch. 9 Polar Coordinates and the Conic Sections
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
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
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).
Sec. 9.1 The Polar Coordinate System 575
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
576 Ch. 9 Polar Coordinates and the Conic Sections
© 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,)?
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.
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
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
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
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 general form for a cardioid in standard position is given in the following
box.
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.
When sketching a polar graph, it is often useful to determine whether the graph
has been rotated or if it has any symmetry.
578 Ch. 9 Polar Coordinates and the Conic Sections
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.
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
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
■ LIMACONS
Graph r = 3 + 2 cos θ.
Solution Let /(0) = 3 + 2 cos θ.
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
Figure 9.10 Graph of
r = 3 + 2 cos θ r - b ± a cos θ or r - b ± a sin θ
580 Ch. 9 Polar Coordinates and the Conic Sections
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.
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
■ ROSE CURVES
There are several polar-form curves known as rose curves, which consist of several
loops, called leaves or petals.
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.
r = 5 sin 40
= 5 cos - 40) Cofunctions o f complementary angles
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. ≡≡≡
58 2 Ch. 9 Polar Coordinates and the Conic Sections
■ LEMNISCATES
The last general type of polar-form curve we will consider is called a lemniscate.
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.
r 2 = a 2 cos 2θ
r = a sin θ standard form
π
2 rotation,
one (circular) petal
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
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 ∙
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
586 Ch. 9 Polar Coordinates and the Conic Sections
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.
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
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.
r = ∣-c o s ⅛ = j - ( 4 ) = 2
Sec. 9.3 Area and Tangent Lines in Polar Coordinates 587
0 = 2 - cos θ or cos θ = ⅜
-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
1 ⅛j - cos(¾r)
- 3 _ 1
- 2 2
Satisfied
Be sure to check for points of intersection that you may have missed,
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.
A = ⅛r2 θ
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.
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
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.
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
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
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)
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
590 Ch. 9 Polar Coordinates and the Conic Sections
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
■ 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.
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
dy
dy = dθ
dx dx
dθ
Sec. 9.3 Area and Tangent Lines in Polar Coordinates 591
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
= -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. — ∣
592 Ch. 9 Polar Coordinates and the Conic Sections
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
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.
594 Ch. 9 Polar Coordinates and the Conic Sections
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))∙
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. ■■■
t x y
- 3 0 -1 (Starting point)
- 2 -5 -j
- 1 -8 -∣
0 - 9 0
1 -8 ⅜
2 -5 j (Endingpoint)
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.
- = 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=□
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
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.
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 ∖
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.
Sec. 9.4 Parametric Representation of Curves 599
Computational Window
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.
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
Proof: This formula follows directly from the chain rule, ~J~t n
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□
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.
Sec. 9.4 Parametric Representation of Curves 601
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
Computational Window ~ π T H E O R E M 9.5 Formula for the area under a curve defined parametrically
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
Sec. 9.4 Parametric Representation of Curves 603
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
604 Ch. 9 Polar Coordinates and the Conic Sections
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.
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.
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.
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
x 2 = -8 y
■ 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
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
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,
608 Ch. 9 Polar Coordinates and the Conic Sections
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
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)
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.
Sec. 9.5 Conic Sections: the Parabola 609
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
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
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.
610 Ch. 9 Polar Coordinates and the Conic Sections
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.
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.
*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
612 Ch. 9 Polar Coordinates and the Conic Sections
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.
■ 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
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
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.
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
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
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.
616 Ch. 9 Polar Coordinates and the Conic Sections
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.
where b 2 = c 2 —a 2 or c2 = a 2 + b 2
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.
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∣ → ∞.
618 Ch. 9 Polar Coordinates and the Conic Sections
Find the set of points such that, for any point, the difference of its distances
from (6, 2) and (6, - 5 ) is always 3.
Sec. 9.6 Conic Sections: The Ellipse and the Hyperbola 619
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
(T ÷ ¾ 2 (x -6 )2 (-y - 6)2
9 10 io
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
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.
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
620 Ch. 9 Polar Coordinates and the Conic Sections
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 9.60 Forms for the equation of an ellipse (ε < 1) in standard polar form
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
Figure 9.62 Forms for the equation o f a hyperbola (ε > 1) in standard polar form
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 = ∣
.
622 Ch. 9 Polar Coordinates and the Conic Sections
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
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.
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 )
6 24 Ch. 9 Polar Coordinates and the Conic Sections
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,
^i<) 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
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
Sec. 9.6 Conic Sections: The Ellipse and the Hyperbola 625
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~
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
shown in Figure 9.67.
Chapter 9 Review
PROFICIENCY EXAMINATION
SUPPLEMENTARY PROBLEMS
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?
628 Ch. 9 Polar Coordinates and the Conic Sections
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 < u < "Vz2 and v > 0.
Chapter 9 Review 629
This project is to be done in groups of three or four students. Each group will submit
a single written report.
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.
Your paper is not limited to the following questions but should include these
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
■ INTRODUCTION TO VECTORS
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
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.
Sec. 10.1 Vectors in the Plane 633
a. u + v b. c. 3u - ∣
v
c. 3 u - ∣ v = 3 ( 2 ,- 3 > - ∣ ( - l , 7)
= (β + —9 — Add vectors.
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
= ∣A B + ∣B C + P Q - ⅛ C
= ∣A B + PQ
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.
Sec. 10.1 Vectors in the Plane 635
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 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
u = AL = (2, -3 ) = _ j_ _ = ∕j =
I∣∣ √∏
v∣ √ IT ’ ' W TTV T3∕ —
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 .
636 Ch. 10 Vectors in the Plane and in Space
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
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. ≡≡≡
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
- 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
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.
638 Ch. 10 Vectors in the Plane and in Space
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
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 10.11 A typical classroom; assume the dimensions are 25 by 30 ft with an 8-ft
ceiling
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).
640 Ch. 10 Vectors in the Plane and in Space
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.
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.
Sec. 10.2 Quadric Surfaces and Graphing in Three Dimensions 641
Solution
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
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).
642 Ch. 10 Vectors in the Plane and in Space
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
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.
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
644 Ch. 10 Vectors in the Plane and in Space
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.
ASSIGNMENT:
nGraph1, y = ------12=---- 2y
, ' 1 + Λ' + y
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.
Sec. 10.2 Quadric Surfaces and Graphing in Three Dimensions 645
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.
_ = y∑ _ x÷
b2 a2
x 2 + y 2 + z 2 = r2
646 Ch. 10 Vectors in the Plane and in Space
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.
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
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).
■ VECTORS IN T 3
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
∣
∣
AB∣
∣= √(61 - α1) 2 + (⅛2 - a2 )2 + (b3 - a3)2
Magnitude of a Vector
Solution ∣
∣∣= √22 + (-3 ) 2 + 52 = √38
v∣ and
I R I = √ (-2 - 1)2 + [3 - ( - l ) ] 2 + [8 - (-4 )] 2 = 13
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
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,.
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.
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. ■
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.
two vectors ∣
∣
v∣∣
∣∣
w∣∣
652 Ch. 10 Vectors in the Plane and in Space
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∣∣
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.
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
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.
= ^ (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
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.
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
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 = ^i - 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?
656 Ch. 10 Vectors in the Plane and in Space
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
■ 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
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
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
i j k j k
v × w= α ι ⅛2 ⅛3 = -(w × v)
z,ι ⅛2 b3 αι α2 f l 3
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
’Properly, it is the distributive property of vectors for cross product over addition, which, for
convenience, we shorten to vector distributivity.
Sec. 10.4 The Cross Product 659
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.
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)
= a^a2b3 - a χa3b2 - a3 a2b3 + a2a3b i + a χa3b2 - a^a3bi
=0
EXAM PLE 2 A vector orthogonal to two given vectors
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
-k
j k
0 0=0
0 0
Next, we will find the magnitude of the cross product of two vectors.
=H H √1-[M⅞]2
= ∖∣
(al 2 + a2 2 + α3 2)(⅛12 + ⅛2 2 + b3 2) - (ai bi + a2b2 + a3b3)2
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∣∣
Sec. 10.4 The Cross Product 661
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
A = ⅛∣
∣
PQ × PR ∣
∣
= ∣ V ( - 6 ) 2 + 12 + ( - l ) 2
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
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
662 Ch. 10 Vectors in the Plane and in Space
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∣
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.
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). ■
■ 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).
• 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)
664 Ch. 10 Vectors in the Plane and in Space
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
■ 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.
666 Ch. 10 Vectors in the Plane and 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
Sec. 10.5 Lines and Planes in Space 667
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:
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 )
Figure 10.37 Graph of
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.
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
668 Ch. 10 Vectors in the Plane and in Space
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
4 (4 ) + l = - l
⅞ = l+ 2 ( 4 ) = 0
To = - ι + ( 4 ) = 4
Graph of L 1 and L 2
⅞ = 2 + 4(4 ) = 0
■ 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.
670 Ch. 10 Vectors in the Plane and in Space
■ PLANES IN + 3
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
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.
Sec. 10.5 Lines and Planes in Space 671
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. ■
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). ■■■
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
= -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
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
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
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.
676 Ch. 10 Vectors in the Plane and in Space
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
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.
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
678 Ch. 10 Vectors in the Plane and in Space
-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
Chapter 10 Review
PROFICIENCY EXAMINATION
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.
SUPPLEMENTARY PROBLEMS
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
Chapter 10 Review 681
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.
682 Ch. 10 Vectors in the Plane and in Space
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
Figure 10.51 Problem 79
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.
Chapter 10 Review 683
This project is to be done in groups o f three or four students. Each group will submit
a single written report.
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.
■ CONTENTS PREVIEW
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
■ VECTOR-VALUED FUNCTIONS
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 11.2 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. ≡≡
1953 by James Watson and Francis Crick. Figure 11.4 Examples o f helixes
Sec. 11.1 Introduction to Vector Functions 687
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)
= Z3 + 1 - 5 sinZ
= 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.
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
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
Verify that lim [F(Z) × G(Z)] = [lim F(Z)j × [lim G(Z)] for the vector functions
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
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)
690 Ch. 11 Vector Calculus
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.
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 ,
53. G(,) - i i za
56. G(∕) = ∣
∣
^∣
j' where u = ∕i + Vtj
692 Ch. 11 Vector Calculus
■ VECTOR DERIVATIVES
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
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
■ 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)
694 Ch. 11 Vector Calculus
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
following definition.
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 .
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
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.
Sec. 11.2 Differentiation and Integration of Vector Functions 695
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.
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).
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
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.
696 Ch. 11 Vector Calculus
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 ').
b. ⅛[F(Z)∙G(Z)] = f ∙G(Z) + F ( Z ) ∙⅛
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.
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.
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.
698 Ch. 11 Vector Calculus
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
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
■ 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
= 57r2i + 3τrj - 2k
e t dt i + t2 dt j + cos2z dt 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)
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
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
702 Ch. 11 Vector Calculus
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
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
t = — -—
vθcos a
Sec. 11.3 Modeling Ballistics and Planetary Motion 703
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:
= -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
704 Ch. 11 Vector Calculus
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
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
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
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
The range R is
v2 /11 ∩∖2
R = -^-sin2α = k ^ y -(s in 80°) ≈ 372.38043
<5
Sec. 11.3 Modeling Ballistics and Planetary Motion 705
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 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
dr c^u 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.
706 Ch. 11 Vector Calculus
ur + dlθ + 2 drdθ∖ a
dt dt2 \.r dt2 dt dt∖ e
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
mA = -G ^ u r
A
λA = ≡ ≡Q^ u11
r2 r
Sec. 11.3 Modeling Ballistics and Planetary Motion 707
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. ■
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.
708 Ch. 11 Vector Calculus
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?
where
is constant.
p (t∖, = m r d^θ + 2m — — 1 2 dθ
W- dt- dt dt 2 dt
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.
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 .
∣
∣
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):
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
Computational Window
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
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,
Sec. 11.4 Unit Tangent and Normal Vectors; Curvature 713
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).
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
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
∣
∣
V(0 ∣
∣= ∣
∣
R , (O∣
I= ⅛ for a ≤ t ≤ b
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
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
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.
= (CURVATURE)(PRINCIPAL U N IT NORMAL) = K N
■ CURVATURE
Thus, curvature of the graph of R(Z) may be computed by finding T and then
substituting into the formula
«=∣∣f ∣∣ - ∣ f |v(-sinψ)≈+(∞s ⅛ - ∣ f ∣
This curvature is interpreted geometrically in Figure 11.19b.
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 ≡ ≡
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).
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
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
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.
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).
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]
720 Ch. 11 Vector Calculus
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
_ 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
■ 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.
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.
Computation of Normal The normal component AN can be found using the formula
Component A∣
AN = √I∣∣
2 - A2
Γ
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
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.
■ 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
Sec. 11.5 Tangential and Normal Components of Acceleration 725
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,
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
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.
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
726 Ch. 11 Vector Calculus
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
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?
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
∣
∣
V∣∣= ⅛ This is called the speed.
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
Torsion
-Ilf 1
PROFICIENCY EXAMINATION
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?
SUPPLEMENTARY PROBLEMS
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
Chapter 11 Review 731
Find the curvature at the given point on each o f the polar curves
given in Problems 43-48.
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?
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
⅞τ=
1
~32i
dt
initial conditions:
F = 50j and = 5i + 5k at t = 0
75. P U T N A M E X A M IN A T IO N P RO BLEM
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
Chapter 11 Review 733
GUEST ESSAY
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.
734 Ch. 11 Vector Calculus
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
Chapter 11 Review 735
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
equal intervals of time.
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
The best ones submitted will appear in the next edition
(along with credit to the problem poser).
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
Chapter 11 Review 737
738
Sec. 12.1 Functions of Several Variables 739
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
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.
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
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
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.
Sec. 12.1 Functions of Several Variables 741
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
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.
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.
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
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.
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
11. ∕( x , y) = y f -
10. f(ιι, v) = V MV
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 ~
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.
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
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
■ PROPERTIES OF LIMITS
The various rules for manipulating limits of a function of one variable all have
counterparts for limits of two variables.
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)
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.
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.
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
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
⅞ g ≡ ⅜ ∣S g ⅞
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
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
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
{ 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)
1 2 .3 PARTIAL DERIVATIVES
■ PARTIAL DIFFERENTIATION
∕'(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
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
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
Find a. ⅛ b. ⅛-.
∂x ∂y
The following example uses implicit differentiation to find the slope of the
tangent line to a level curve at a particular point.
75 8 Ch. 12 Partial Differentiation
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
■ GEOMETRIC INTERPRETATION
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).
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
Thus,
/.(1, 3) = ,1 + √1 + 3 = I
λ 2√1 + 3 4
βββ
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.
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
76 0 Ch. 12 Partial Differentiation
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.
= ∂ z _ ∂ (∂z∖ _ f ∩
d2f 2
= f
∂y2 ∂y2 θy∖ θy∕ ^ y y yy
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.
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 . ≡≡
Sec. 12.3 Partial Derivatives 761
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. ■
∙^∣7~, ")2T
The heat equation = c2 describes the heat in an insulated rod. The
constant c depends on the material in the rod.
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
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)
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)
Sec. 12.3 Partial Derivatives 763
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 ', ^
■ 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)
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
■ INCREMENTAL APPROXIMATIONS
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
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
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.
Δ V ≈ ^τrRH) ΔR + (∣τrΛ2) Δ H
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
„ . . 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
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
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
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
■ DIFFERENTIABILITY
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∕
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
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.
even if/ exists at (0, 0). Therefore,/is not differentiable at (0, 0).
∕√ * , = + 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)
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
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?
774 Ch. 12 Partial Differentiation
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
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.
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.
∆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
∕ 1 ∖2
z = x 2 + y 2 = (γ j + (Z2)2 = Γ 2 + Z4 for Z ≠ 0
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
Solution The volume of the cylinder is V = τrr2h, and we are given = 3 and
= - 5 . We find that
dz ∂z dx ∂z dy ~
d i = ∂ ⅛ d t+ ∂ } d t c ht a m ' u le
Sec. 12.5 Chain Rules 777
=g <°)+“[∂x2 + - ( 0 ) + b ∖ -^- a +⅛
∂x∂y ∂y Layax ∂y2
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.
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^
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
Sec. 12.5 Chain Rules 779
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
780 Ch. 12 Partial Differentiation
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
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.
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-------------------
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 ------------------ *-------------------
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 ) ( - ^ )
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)∙
■ THE GRADIENT
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
Then,
V∕(x, y) = 2xyi + (x2 + 3y2)j
The following theorem shows how the directional derivative can be expressed
in terms of the gradient.
W0 ( l , -3 ) = - ⅜ i - ⅛ j
=—
∣
∣∣= —
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.
Quotient rule s. ≠ o
∖ g) g2
Power rule V(P) = n f (n- 1) v /
^ ∕= 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 . —
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
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).
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.
78 8 Ch. 12 Partial Differentiation
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 ) ^
= ^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.
Plane z = 1
Figure 12.25 The surface f(x , y) = x2 - y2 and the level curve for which C = I
Sec. 12.6 Directional Derivatives and the Gradient 789
= - ( 6 i + 8j) = 3 _ 4
V (-6 )2 + (-8 )2 5 5j —
Normal line
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.
790 Ch. 12 Partial Differentiation
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
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.
Sec. 12.6 Directional Derivatives and the Gradient 791
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
79 2 Ch. 12 Partial Differentiation
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
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.
We begin by considering relative extrema, and then we turn to the more general
question of finding absolute extrema.
■ RELATIVE EXTREMA
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).
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
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
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).
796 Ch. 12 Partial Differentiation
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
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.
Graph of
∕(x , y) = 8x3 - 24xy + y 3
79 8 Ch. 12 Partial Differentiation
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). ∙≡ ≡
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).
Proof: The proof of this theorem is beyond the scope of this text and is usually
given in an advanced calculus course. ■
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.
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.
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
= 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=
Computational Window
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:
GPA 3.1 2.8 2.5 2.8 3.4 1.9 2.1 3.2 2.9 2.3
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 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?
804 Ch. 12 Partial Differentiation
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
b. z = -2
z = -1
z= 1
z=2
1 2 .8 LAGRANGE MULTIPLIERS
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. ■
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.
/(0, 1) = 1 - 02 - l 2 = 0
The last step is to evaluate T at the critical points and the endpoints:
< 1 ’ 1) = 3-6;
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.
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
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). ≡≡≡
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.
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
x + 2y + z = 10 x2 + y2 - z = 0
x + 2y - J = 10 x 2 + y2 + ∣ = 0
x + 2y = 10 + I x2 + / = - ∣
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.
Sec. 12.8 Lagrange Multipliers 813
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
814 Ch. 12 Partial Differentiation
CHAPTER 12 REVIEW
PROFICIENCY EXAMINATION
SUPPLEMENTARY PROBLEMS
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
5. ∕( x , y) = χ + 6. f( x , y) = x 3e 3>'z2x 3 2
30. x + 2xz - y z - z = 1 3
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).
818 Ch. 12 Partial Differentiation
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
Chapter 12 Review 819
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 ~
820 Ch. 12 Partial Differentiation
"Desertification”
This project is to be done in groups o f three or four students. Each group will submit
a single written report.
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
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
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
∕- ∑ ∕( ⅞ X ) ∆ 4 < ε
fc=l
■ 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
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
■ 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.
Sec. 13.1 Double Integration Over Rectangular Regions 825
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
√ Γ rb
or
. ↑
Limits of y (variable inside brackets) Limits of y (outside brackets)
Let us see how Fubini’s theorem can be used to evaluate double integrals. We
begin by taking another look at Example 1.
vertices (0, 0), (3, 0), (3, 2), and (0, 2).
826 Ch. 13 Multiple Integration
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.
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
Sec. 13.1 Double Integration Over Rectangular Regions 827
Limits of x
Limits of y
'
x cos xy dA cos u du dx
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
[ ∫ ∕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
828 Ch. 13 Multiple Integration
∠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.
Sec. 13.1 Double Integration Over Rectangular Regions 829
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 ∙ ∣- ∣∙
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.
■ NONRECTANGULAR REGIONS
In the last section, you saw how the double integral f(x , y) dA could be
J RJ
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.
g2W
T Y P E I: (vertical strip): ∕( x , y) dy dx
x-fixed, y varies (form dy dx)
Π '∣W
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.
J J (x + y) dA
τ
using an iterated integral with:
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
832 Ch. 13 Multiple Integration
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
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.
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
Find the area of the region D between y = cos x and y = sin x over the interval
0 ≤ x ≤ f using
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
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
shown in Figure 13.10.
Let u = 1 —y 2 to integrate.
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.
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.
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)
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
836 Ch. 13 Multiple Integration
Double integrals have many of the same properties as single integrals. Three of
these properties are contained in the following theorem.
f f [af(x, y)
D
PROBLEM SET 1 3 .2
O 1. ■ What Does This Say? Describe the process for find ∙2 r sin x "τr∕2 rsinx
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).
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
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
∏ 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
838 Ch. 13 Multiple Integration
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
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 = -
■ 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
∑ ι , < < p ∆ ∕l 1
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
Finally, we compute the given double integral in polar form by taking the limit
∕( r , 0)r dr dθ
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.
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
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
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
Let u = a 2 - r 2∙,
du = —2r dr.
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
Evaluate J J dA, where D is the region in the first quadrant that lies inside
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 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,
844 Ch. 13 Multiple Integration
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.
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
∣J
Q
b. Show that
1 3 . 4 SU R F A C E A R E A
■ 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 .
= 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
∆=1⅛ Σ √ [ ∕K > ‰ ) ] 2 + [ ∕ χ , ^ ) ] 2 + 1 ∆ ^
n → 0° m=∖
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
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
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
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.
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
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
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
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.
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
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
n
D
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)
854 Ch. 13 Multiple Integration
Problem 26
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
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
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
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
∫J> R
, y, z) dV - f(x , y, z) dx 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.
858 Ch. 13 Multiple Integration
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
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.
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
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.
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
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
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
As you would expect, this is the same as the result from projecting on the xy-
plane. ≡≡
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
shown in Figure 13.33.
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
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
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;
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 .
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,
so that
m
R
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
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
following definition.
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
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
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
M j κ - ∫ ∫ ∫ y p(χ, y , z) d v
R 1
Distance to the xz-plane
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.,
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 ∣∙ '
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
■ 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.
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
870 Ch. 13 Multiple Integration
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
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
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) dy dx
Γ1 fx 3(l - x ) 2 x (l - x ) 4 ] , _ χ
∣0 L 2 + 12 J a x 90
∕‰ 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.
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
872 Ch. 13 Multiple Integration
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
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
Sec. 13.6 Mass, Moments, and Probability Density Functions 873
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
874 Ch. 13 Multiple Integration
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
Computational Window
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
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.)
■ CYLINDRICAL COORDINATES
We have the following conversion formulas, which follow directly from the
rectangular-polar conversions.
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.
(√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 ≡≡
Sec. 13.7 Cylindrical and Spherical Coordinates 877
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 ( ,
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
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
Sec. 13.7 Cylindrical and Spherical Coordinates 879
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.
■ 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.
p= a Because p ≥ 0
b. z = x + 3y2
2
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
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
'--∣
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
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
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.
884 Ch. 13 Multiple Integration
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
∂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.
886 Ch. 13 Multiple Integration
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.
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
r 2 υ2
Find the area of the region E bounded by the ellipse + p = 1.
We will transform the elliptical region E to one that is circular. Because E can
be represented by
∂(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
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
∂(x, y, z)
∫ ∫ ∫ ∕⅛ r , z)dx dy dz = ∂(u, v, n,) du dv dw
R∙
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
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
890 Ch. 13 Multiple Integration
∂(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
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
C H A P T E R 1 3 R E V IE W
PROFICIENCY EXAMINATION
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
892 Ch. 13 Multiple Integration
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).
SUPPLEMENTARY PROBLEMS
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
∙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
∙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
∏ R ,
dx dy, where R is the rectangle a ≤ x ≤ b,
∂2f
44. Find numbers a, b, and c so that
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.
894 Ch. 13 Multiple Integration
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
√-∕ ∖ ,/ < 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?
∕(-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.
896 Ch. 13 Multiple Integration
"Space-Capsule Design”
This project is to be done in groups o f three or four students. Each group will submit
a single written report.
’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
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
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
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
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.
900 Ch. 14 Vector Analysis
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.
a. F(x, y) = x 2j ,i + xy 3j b. G(x, y, z) = x i + y 3z 2j + x z 3k
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
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
_ ∂ 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.
902 Ch. 14 Vector Analysis
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.
= (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
EXAM PLE 4 A constant vector field has divergence and curl zero
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
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
904 Ch. 14 Vector Analysis
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.
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)
906 Ch. 14 Vector Analysis
1 4 .2 LINE INTEGRALS
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
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. ■
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
908 Ch. 14 Vector Analysis
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
z7v f“ C~ r ^ι □
[ ( t ) 2 + (^ 2 ) ] -∕t dt = I 2t2 -2 td t = 4 I t3 dt = [ t 4Jθ = 16
t*,2 y
= (-t)e~ t dt Because -I
Ji c
= ⅛(e~4 - e~ , ) β
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 dR =
Jo
(6Z3 + 8Z2 - Z4) dt = ⅛4 + t^3- ⅛5υjo . 119
30
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 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
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.
J c,
F ∙ JR =
Jo
(sin2 / dt + cos2/ dt) =
Jo
dt = 2
F ∙ JR = 0= 0
'c, Jo
Sec. 14.2 Line Integrals 911
F ∙ dR = 0
Thus,
I F ■
dR = f F ∙ dR + ( F ∙ dR + f F ∙ dR = f + 0 + 0 = ?
Jc Jc, Jc, Jc, - λ
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
Solution Let C denote the path of the object. Then the total work W done by F on
Because I y I + J’2 = 1
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:
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.
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).
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.
916 Ch. 14 Vector Analysis
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
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
Verify that the vector field F = 2xyi + x 2j is conservative, with scalar potential
∕ = x 2y.
Sec. 14.3 Independence of Path 91 7
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.
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
= ex sin y - y x + c(y)
Where c(y) is a function of y alone— a
“constant” as far as ^-integration is concerned
918 Ch. 14 Vector Analysis
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.
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.
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)
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
= (e - ⅞ ) - 0 = e - ⅛ _
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.
920 Ch. 14 Vector Analysis
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.
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
Thus, F is conservative.
xy 2 dx + x 2y dy = 0
Jc
=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
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
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
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.)
=- 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. ■
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
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
= [ ∣(4xy - 2xy) dA = 2 ∣ xy dy dx = 2 ∣ dx
J DJ
= |o (x ~ χ5) d x = ⅛ 2 - ⅛χ 6 ]0 = |
928 Ch. 14 Vector Analysis
A line integral can be used to compute an area of a region in the plane by applying
the following theorem.
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.
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
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
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
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
930 Ch. 14 Vector Analysis
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.
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
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)
(3y dx - 2x dy)
21. Show that 29. Evaluate I ds, where z(x, p) = 2x 2 + 3y2 and C is the
Jc ∂n y j
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
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
1 4 .5 SURFACE INTEGRALS
■ 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.
934 Ch. 14 Vector Analysis
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
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
Sec. 14.5 Surface Integrals 935
-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.
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
936 Ch. 14 Vector Analysis
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.)
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
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.
Finally,
=0+4=4
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.
F ∙ N = ^ j ( x y + l - χ - y + χ + y) = - f y χ y + 1 )
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
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
Sec. 14.5 Surface Integrals 939
■U× R JI du dv
5 D
■2 r l -2
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
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
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
1 4 .6 STOKES'S THEOREM
■ STOKES’S THEOREM
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.
Notice that the surface 5, in the statement of Stokes’s theorem is not closed.
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
(1 - s ) s d s = ( ⅞ - f ) ∣ o = ∣
F ∙ √R =
∣
E1 Jθ
F ∙ <√R = 0
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
N = ^ ( i + j + k)
Because z = 1 - x - y, we find
= 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
∫ ∫ (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
by Stokes’s theorem.
Sec. 14.6 Stokes’s Theorem 94 3
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
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
944 Ch. 14 Vector Analysis
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
= 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
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.
F ∙r ∕R = F ∙ N) dS = 0 dS = 0
'c JS J
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.
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.
946 Ch. 14 Vector Analysis
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
H dR = I
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.
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
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.
Because z = 0 on Sj
950 Ch. 14 Vector Analysis
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)
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
= 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
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.
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
= 3 f x( 1 - x) 2 - ∣x( 1 - x) 2 dx = A
Jo °
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
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.
Evaluate J j F ∙ N dS, where F = xyi - z 2k and S' is the surface of the upper
(y - 2z) dy dz = (I - 2z) dz = -⅛
Io Jo Jo v z
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.
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 = - ∣
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 —
954 Ch. 14 Vector Analysis
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.
- 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
=
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.
956 Ch. 14 Vector Analysis
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
Chapter 14 Review 957
CHAPTER 14 REVIEW
■ 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
958 Ch. 14 Vector Analysis
= 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:
∣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,
Chapter 14 Review 959
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.
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)
fi x , y, z) dx
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
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
Stokes’s Theorem
(Section 14.6; Green’s theorem in three dimensions)
F ∙ dR = dS
'c
s
Π ' ∙ -
Γ N K
Chapter 14 Review 961
ELECTROMAGNETISM
Let E be the electric intensity field and H be the magnetic intensity field:
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
PROFICIENCY EXAMINATION
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.
SUPPLEMENTARY PROBLEMS
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.
Chapter 14 Review 963
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
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
Chapter 14 Review 965
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.
GUEST ESSAY
Historical Note
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
The best ones submitted will appear in the next edition
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
Chapter 14 Review 969
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
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 = ~ ~
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
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 )
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
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
( ∕o u)'(x) =f'(u)u'(x) or
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.
lim ⅛ = 0
χ→+∞ Λ
Furthermore, if x n is defined when x < 0, then
lim -2⅜
x→-∞ Xn
∕(x ) ∕'( x )
hm = lim ~ ÷ ∕
x→c g(Λ) x →c g∖ χ)
provided the limit on the right side exists (or is infinite).
lu = C1 f du = c J7(«) du
Constant multiple
(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:
Then
∣f(x) dx = g(u) ~^,dx = G(u) + C
for G an antiderivative of g.
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
(/ ) W - ∕ v - 1(*)]
T x^ =
lim = -∞ lim = 0
x→o+ x n x→+∞ x n
lim = +∞ lim x ne~kx = 0
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
a. E(x + y) = E(x)E(y)
b. E(x - y ) = E ( y j
c. [∕f(x)p = E(px)
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.
+ yP(x) = Q(x)
is given by
y = 7 ⅛ j QW dx+ c
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
⅛(sinhM) = c o s h a g ⅛ (c°sh “ ) = s in h M S
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
CHAPTER 8
X(c¾ + db^ = c ^∑
tak + d Xb k
Σ ark = . a
k=o 1- r
lχ - ¾l < ⅛÷1
where S n is the nth partial sum of the 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.
App. A: Theorems by Chapter A -17
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
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 9.5 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,
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:
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
PN S A_ v , w
M H
v∙w= 0
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
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 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
∙5, = l∣R'(0∣l dt = y∣
[x'(t)]2 + [ ∕( t ) ] 2 + [√(Z)]2 dt
Ja Ja
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 )
Quotient rule ,≠ 0
∖6 / g
Power rule V (/”) = n f n ~vV f
A-22 Appendices
∏' f(x, y) dx dy =
Ja Jc
f(x, τ) dy dx
D D
D D;
∣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
plane under the change of variable x —g(u, v), y = h(u, v) where g and h are
continuously differentiable on D*. Then
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
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.
= g Throughout D
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{^u) + ⅞ ] δ m = ∕ W ) + M ^ ∕IX∙*)]
G ^ + ^ J ΔM
lim Substitution
∆x
= 2≡0 [^ ΔM>+ S Δ^
Because g is continuous at t = 0
d f du
du dx
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
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
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
l. ∕W ,. ∕'( t )
w→a + g (W ) w →a
i g (t)
*This is a special case of ΓHδpitaΓs rule. The other cases can be found in most advanced calculus
textbooks.
A -2 8 Appendices
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.
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
This is true for all k > N. Now we can complete the proof by using the direct
App. B: Selected Proofs A -2 9
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. ■
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 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(c) _ F∖ z)
G(c) G'(z)
-β n+1∖ z)
n∖ (x - z)n
- ( x - z)n
n∖
=β n+1 ∖ z)
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)]
A'(u, v)
Figure B.2
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 ^I* + 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)
Let 5 be an oriented surface with unit normal vector N, and assume that S is
bounded by a closed, piecewise-smooth curve C whose orientation is compatible
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
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
( ∂h ∂z . ⅛ ∂z df ∂z . ∂h ∂z , ∂g ∂ f∖
∖ ay ∂x ^1^ ∂z ∂x ∂z ∂y ^r^ ∂x ∂y ∂x ∂y /
∫ ∫ [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
App. B: Selected Proofs A -35
—
∂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
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
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
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
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.
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.
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.
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
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
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
n +। ι1,
2. Power rule U (IV l / ≠ -1
u n du = In ∖u; n = -1
4. Logarithmic rule M∣
ln∣ du = u In z∕ - u
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 lz^a 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
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∖
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— —
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|
206. ∣
∖ 3√ π ^ = - ^ 7 2)5z2÷ ^ 2^ 2)3z2
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
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 - '^a
, 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
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
349. fI u
,, b
<,iinll 2 a
mu. √
au,, _ « ! _ u sin4β2 au _ cos 2au
350. f sin3(2Wdu = + ≤ ⅛
<2 3<2
, . . , 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 + ''
456.
∏ sin - 1 - ) du = M( sin - 1 ^ J - 2w + 2 V a 2 - w2 sin - 1 ^
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 ^au∖
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
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
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
45. a. b. - √ 2 c. - √ 3
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. - 9 3. - 8 s. 4 7. 2 9.
V3
9
53. y 55. Does not exist 57. 4 59. 8 61. 3 63. - 3x^2
App. E: Answers to Selected Problems A -5 1
discontinuous at x = / + nττ
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
CHAPTER 2
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
Y 1 B s in <2πcos X)
PROBLEM SET 2.3, Page 110
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
PROBLEM SET 2.2, Page 102
, 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
=
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 = ∣
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
Practice Problems, Page 160
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.
App. E: Answers to Selected Problems A -5 7
CHAPTER 3
PROBLEM SET 3.1, Page 175 PROBLEM SET 3.3, Page 190
5. Maximum, 1; minimum, - 1
0
7. Maximum, 0; minimum, —2
9. Maximum, 1; minimum, 0
3. 1 5. 1-5275
7. ∖ ∣.2 5 ≈ 1.0772 9. ∣
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
∕(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
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
23. a. 31. a. x = 0
b. Increasing on ^0, ~
decreasing on 0
c. Critical point: (0, 0), relative minimum
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)
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;
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)
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)
3. 8 ft × 8 ft
35. Relative maximum at x = 0;
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.
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
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
PROBLEM SET 3.10, Page 251
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
19. jx 5 + yx 3 + 25x + C
CHAPTER 3 REVIEW
Practice Problems, Page 252
14. 2 15. - ∣ 16. 0 17. 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.
CHAPTER 4
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
l l .F ( x ) = - i ( l - 3 x ) - 3 - ⅛ F ' ( x ) = (r ⅛ 3
23. ∣ 2 5 .- 1 0
37. 18.7 0C 3 9 . / ( 2 )= ∣
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
App. E: Answers to Selected Problems A -6 7
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
CHAPTER 5
—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%
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 )
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.
1 -χ 2
9. - ∣ e + C
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
1 3 7 2∕ 7 ≈ 0.2475
43. 45.
4 In 2
-⅛(e 12 - i ) ≈ 10.850
47.
15∖ e)
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
Practice Problems, Page 579
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
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
CHAPTER 6
-2
PROBLEM SET 6.1, Page 393
1. 9 3. i 5. 2 7. 36√3 J
o
[(1 + V 4 - j; 2)2 - 1] dy
-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
App. E: Answers to Selected Problems A -7 3
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
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
Λl = ∣m(v 0 - gt)2
CHAPTER 7
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 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 =
CHAPTER 8
PROBLEM SET 8.1, Page 499 11. 1, 3, 13, 183, 33673 13. 5 15. - 7
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
b. 0.083
PROBLEM SET 8.4, Page 521
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
App. E: Answers to Selected Problems A -7 9
1. R = 1; ( - 1 , 1) 3. R = 1; ( - 1 , 1)
'—∞ ox
Γ ∣^ 7 „ _ 4√ 13,
5. R = ∣m 7 ∙ K ~ 3’ ∑ T
“ (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
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
9. (l,77), ( -1 ,0 ) ; ( -1 ,0 )
∣ 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
9 .( 2 ,^ .( 2 ,^ ) ll.( 0 ,0 ) ,( 2 ,0 ) ,( 2 ,τ τ )
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
43. 5x - 6y + 28 = 0
(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
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 )
65. ⅛(e4 - - 1)
o
69. 0.905
A -8 6 Appendices
CHAPTER 10
2 9 .⅛ b + ) )
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
PROBLEM SET 10.3, Page 655
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.
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.
App. E: Answers to Selected Problems A -8 9
CHAPTER 11
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;
A(ln 2) = 2 i + + 16k
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
41. t2 e t — t3 e t — 2e t — —sin Z
Z
43. 3i + e 2j 45. 3i - ∣ j + 2k
PROBLEM SET 11.3, Page 708
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
PROBLEM SET 11.2, Page 700
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∙ “
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 ^ -----------------------------
CHAPTER 11 REVIEW
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
App. E: Answers to Selected Problems A -91
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
x = t
y = 0
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
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
PROBLEM SET 12.2, Page 754
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 '
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)∣
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.
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
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 ∙
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
App. E: Answers to Selected Problems A -9 7
f x x = f lι -‰ 3 '⅛ = ° ^ χ 2 y 2 )~ 3'2
63. - ^ = ( 3 + 4 In 2)
∖ λ y )
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.
∣ ∣ = e^u2 ~v2 ∖ 12yu2 + 6y + 8ιιvy)
99. See American Mathematical Monthly, Vol. 74.
55. m = - 1 101. See American Mathematical Monthly, Vol. 45.
CHAPTER 13
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
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
17. i(277 +
3√3) 19. ⅛ - 2 21. 77
rπ∕4 ra sin θ ι
39. 2 1 1 r dr dθ = g fl2 (7r - 2)
15π
49. 5311.
4 4
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
PROBLEM SET 13.3, Page 844
Vτr∕2 _______________
n
37. ∣ J
rv 4 r 2sin r2 + 1 dr dθ
∏ 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
a(∖-y lb -z lc )
dx dz dy
19. a √ 0 , ^ )
∖ 2T7√ ∖ 3 π 3ττ∕ 2
311 ∙ 64
A -100 Appendices
39. 49 In 7 - - y In 5 - 27 In 3 + 6 ≈ 11.3331
CHAPTER 13 REVIEW
√(4 X - X 2)∕2 ∣
∙4X
Π
29. y ∣ "
∣ J x 2 +2y2
dz dy dx = 2πV 2
55. ⅛
61. ⅜ ^ (1 - cos φ)
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
CHAPTER 14
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
PROBLEM SET 14.3, Page 922
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
τ y = ~ - = 2xy
j + z; conservative 17. ∣ 19. ∣ττα4
∂x ∂y 21. 4τrα4
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. 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^ " ’
A-107
A-108 Index
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
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
PROCEDURAL RULES
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
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
a 11 du = --------1- C ; a > 0, a ≠ 1
In a
Logarithmic rule In u du = M In u —u + C ; w >0
J cosh u du = sinh u + C
Hyperbolic rules sinh u du = cosh u + 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
PRENTICE HALL
ENGLEWOOD CLIFFS, NJ 0 7 6 3 2
9 780133303254