.

,
I

MATHEMATICAL ANALYSIS

I I

MATHEMATICAL ANALYSIS
SECOND EDmON

TOM M. APOSTOL Colijomitllnstitute of Technology

ADDISON· WESLEY PUBLISHING COMPANY Reading, Massachusetts Amsterdam· London· Manila • Singapore • Sydney. Tokyo

•;
,9ii;-~
11,00'"

To my parents

WORLD S1UDENT SERIES EDmON

FIFTH PR.IN1lNG 1981

A complete and unabridged reprint of the original American textbook, this World Student Series edition may be sold only in those countries to which it is oonldgned by Addison-Wesley or its authorized trade distrlbutOfS.1t may not be re-exported from the country to which it has been consigned, and it may not be sold in the United States of America or its possessions.
Consulting Editor: Lynn Loomis

191I Copyright © 1974 by Addison-Wesley Publishing Company, right 1974 by Addison-Wesley PublUhing Company, Inc.

Inc. Philippines copy-

AD rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical. photocopying, recording or otherwise. without the prior written permission of the publisher. Original edition prlnted in the United States of America. Published simultaneously in Canada. Librat'Y of Congress Catalog Card No. 72-11473.

,
.

-'.

/.~~I
'.~. "'j

PREFACE
A glance at the table of contents will reveal that this textbook treats topics in analysis at the "Advanced Calculus" level. The aim has been to provide a development of the subject which is honest. rigorous, up to date, and, at the same time, not too pedantic. The book provides a transition from elementary calculus to advanced courses in real and complex function theory, and it introduces the reader to some of the abstract thinking that pervades modem analysis. The second edition differs from the first in many respects. Point set topology is developed in the setting of general metric spaces as well as in Euclidean n-space, and two new chapters have been added on Lebesgue integration. The material on line integrals, vector analysis, and surface integrals has been deleted. The order of some chapters has been rearranged. many sections ha-ve been completely rewritten. and several new exercises have been added. The development of Lebesgue integration foDows the Riesz-Nagy approach which focuses dicectly on functions and their integrals and does not depend on measure theory. The treatment here is simplified. spread out, and somewhat rearranged for presentation at the undergraduate level. The first edition has been used in mathematics courses at a variety of levels, from first-year undergraduate to first-year graduate. both as a text and as supplementary reference. The second edition preserves this llexibility. For example, Chapters 1 throngh S, 12, and 13 provide a course in differential calculus of functions of one or more variables. Chapters 6 through 11, 14, and 15 provide a course in integration theory. Many other combinations are possible; individual instructors can choose topics to suit their needs by consulting the diagram on the next page, which displays the logical interdependence of the chapters. I would like to express my gratitude to the many people who have taken the trouble to write me about the first edition. Their comments and suggestions influenced the preparation of the second edition. Special thanks are due Dr. Charalambos Aliprantis who carefully read the entire manuscript and made numerous helpful suggestions. He also provided some of the new exercises. Finally, I would like to acknowledge my debt to the undergraduate students of CaItech whose enthusiasm for mathematics provided the original inc.entive for this work. pastJdena September 1973 T.M.A.

LOGICAl,

INTERDEPENDENCE

OF THE

CHAPTERS

THE REAL AND COMPLEX NUMBER SYSTEMS

CONTENTS

<lIapter

I The Real aJld Complex Number Systems 1.1 1.2 1.3 1.4 1.5 1.6 Introduction . The field axioms . The order axioms Geometric representation of real numbers Intervals . Integers The unique factorization theorem for integers Rational numbers Irrational numbers Upper bounds" maximum elemenr. least upper bound (supremum) . The completeness axiom Some properties of the supremum Properties of the integers deduced from the completeness axiom The Archimedean property or the real-number system . Rational numbers with finite decimal representlltion Finite decimal approximations to real numbers . Infinite decimal representation of real numbers Absolute values and the triangle inequality The Cauchy-Schwarz inequality Plus and minus infinity and the extended real number system R* Complex numbers Geometric representation of compleJ:: numbers The imaginaI)' unit . Absolute value of a complex number. Impossibility Qf ordering the complex numbers Complex exponentials Further properties of complex exponentials . The argument of a complex number . Integral powers and roots of complex numbers . Complex logarithms Complex powers Complex sines and cosines . Infinity and tbe extended complex plane c+ . Exercises .
is. 1

1 2

3 3
4 4 6

FUNCTIONS OF BOUNDED VARIATION AND REGTIl<'IABLE CURVES

1.7
1.8 1.9 1.l0
12

7 8
9 9 JO 10 11 11 J2
12

INFINITE SERIES AND INFINITE PRODUCTS

MULTtVARIABLE DIFFERENTIAL CALCUf,US IS

tIl 1.12

1.13
1.14 1.15 1.16

IMPLICIT FUNCTIONS
AND EXTREMUM PROBLEMS

1.17
1.18 1.19 1.20 1.21 1.22 1.23 1.24 1.25 1.26 1.27 1.28 1.29 1.30 1.31 1.32

13 14 15 17 18
18

19 19 20 20 21 22 23

CAUCHY'S THEOREM AND TIlE RESIDUE CALCULUS

16

Hi MULTIPLE LEBESGUE INTEGRALS

24
24 25

1.33

't·
:'~~

',\

II

eoat.ts Z Some BaRe: Notions of Set 'l1teOIry Introduction . Notations Ordered pairs Cartesian product of two sets Relations and functions Further terminology concerning functions One-to-one functions and inverses Composite functions. Sequences . Similar (equinumerous) sets Finite and infinite sets Countable and uncountable sets Uocouotability of the real-number system Set algebra Countable collections of countable sets Exercises .
0( Pomt

ea.taJIs 4.7 4.8 4.9 4.10 4.11 4.12 4.13 4.14 4.15 4.16 4.11 4.18 4.19 4.20 4.21 4.22 4.23 Limits of vector-valued fuoctions . Continuous functions Continuity of composite functions. Continuous complex-valued and vector-valued functions Examples of continuous functions Continuity and inverse ima.ges of open or dosed sets Functions continuous on compact sets Topological mappings (homeomorphisms) Bolzano's theorem Coacectedness Components of a metric space . Arcwise connectedness . Uniform continuity. Unifonn continuity and compact sets Fixed-point theorem for contractions Discontinuities of real-valued functions Monotonic functions Exercises •

Ii

Chapter

2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 2.10

32 32 33

33
34

z.u

35 36 31 31 38

77 18 19 80 80 81 82 84 84 86 81

2.12 2.13 2.14 2.15

38
39 39 40 42 43

88
90 91 92 92 94 95

CltQteF 3 EJemenU
3.1 3.2 3.3

Set Topology 47 47 49 SO 52 52
53 54

ClJapter

5 DeriYatJves 5,1 Introduction . 5.2 Definition of derivative . 5.3 Derivatives and continuity . 5.4 Algebra of derivatives 5.S TIle chain rule 5.6 One-sided derivatives and infinite derivatives 5.1 Functions with nonzero derivative 5.8 Zero derivatives and local extrema 5.9 Rolle's theorem . 5.10 The Mean-Value Theorem for dmivatives 5.11 Intermediate-value theorem for derivatives 5.12 Taylor's formula with remainder 5.13 Derivatives of vector-valued functions 5.14 Partial derivatives S.Is Differentiation of functions of a complex variable 5.16 TIle Cauchy-Riemann equati~ Exercises .

Introduction . Euclidean space R- . Open balls and open sets in R" 3.• The structure of open sets in Rl 3.5 Closed sets 3.6 Adherent points. Accumwation points . 3.1 Closed sets and a.dbenint points 3.8 The Bobaao-Weierstrass theorem 3.9 The Cantor intersection theorem . 3.10 The Lindelof covering tbeorem 3.11 The Heine-Borel covering theorem 3.12 Compactness in R" 3~13 Metric spaces 3.14 Point set topology in metric spaces s 3.15 C-0111pact ubsets of a metric space 3.16 Boundary of a set Execcises . Cllaptcr

104104105
106 106

107
108 109

56 56 58 59 60

llO

61
63 64
6S

110 111 113 114 115 116 118

121

4
4.1 4.2 4.3

LImIts and Continuity

4.4
4.5 4.6

Introduction . Convergent seqUClJCeS in a metric: space Cauchy sequences Complete metric spaces . Limit of a function . Limits of complex-valued functions

10
1t}

CIlapler

6 6.1 6.2 6.3 6.4 6.5

F!lllctions of Bounded Variatioo and Rectifiable Curves Introduction . Properties of monotonic functions Functions of bounded variation Total variation Additive property of total variation 127 127 128 129
130

72

74 74 76

6.6 6.7 6.8 6.9 6.10 6.1 J 6.12

TotaJ variation on [a. xl as a function of x Functions of bounded variation expressed as the diffe£enCe of increa$ing functions . Continuous functions of bounded variation Curves and paths :Rectifiable paths and arc length Additive and continuity properties of arc length Equivalence of paths. Change of parameter Exercises .
Jntegnll

131 132 131 133 134 135 136 137

CbapUlr

7 The Riemaan-Stlelijes 7.1 7.2 7.3 7.4 7.5 1.6 7.7 7.8 7.9

1.10 7.11

7.12 7.13 7.14 7.15 7. [6 7.17 7.18 7.19 7.20 7.21 7.22 7.23 7.24 7.2S 7.26 7.27

Introduction. Notation. The definition of the Riemann-Stieltjes integral. Linear properties Integration by parts . Change of variable in a Riernann-Stieltjes integral Reduction to a Riemann integral. Step functions as integrators Reduction of a Riemaon-Stieltjes integral to a finite sum Euler's summation formula MonotonicallyiDcreasingintegrators. Upper andlowec integrals . Additive and linearity propertie, of upper and lower integrals. Riemann's condition Comparison theorems Integrators of bounded variation Sufficient conditions for existence of Riemann-Stieltjes integrals Necessary conditions for existence of Riemann-Stieltjes integrals. Mean Value Theorems for Riemann-Stieltjes integrals. The integral as a function of the interval. Second fundamental theorem of integral calculus Change of variable in a Riemann integra] Second Mean-Value Theorem for Riemann integrals Riemann-Stieltjes integrals depending on a parameter. Dilferentiation under the integral sign Interchanging the order of integration Lebesgue's criterion for existence of Riemann integrals Complex-valued Riemann-Stieltjes integrals . Exercises . Infinite Series and JllfInite Produets Introduction. Convergent and divergent sequences of complex numbers . Limit superior and limit inferior of a real-valued sequence Monotonic sequences of real numbers Infinite series.

140 141 141 142 144 144 145 147 148 149" ISO 1S3 IS3 155 156 159 160 160 161 162 163 165 167 167 169 173 174

8.6 Inserting and removing parentheses 8.1 Alternating series 8.8 Absolute and conditional convergence 8.9 Real and imaginary pam of a complex series 8.10 Tests for convergence of series with positive terms 8.11 The geometric series 8.12 The integral test. 8.13 The big oh and little oh notation 8.14 The ratio test and the root test 8.15 Dirichlet's test and Abel's test . 8.16 Partial sums of the geometric series E zIt on the unit circle 8.17 Rearrangements of series 8.18 Riemann's theorem on oonditionaUy convergent series . 8.19 Subseries. 8.20 Double sequences 8.21 Double series 8.22 Rearrangement theorem for double series 8.23 A sufficient condition for equality of iterated series. 8.24 Multiplication of series . 8.25 CesAro summabiJity . 8.26 Infinite products . 8.27 Euler's product for the Riemann zeta function Exercises .

Izl

=t

187 188 189 189 190 190 191 192 193 193 195 196 191 191 199 200 201 202 203 205 206
2(J9

210

9.1

.66

9.2 9.3 9.4 9.5 9.6 9.1 9.8 9.9 9.10

9.11

Chapter

8 8.1 8.2 11.3 11.4 8.5

183 183 184 18S 18S

9.17: 9.13 9.14 9.15 9.16 9.11 9.18 9.19 9.20 9.21

Pointwise convergence of sequences of functions Examples of sequences of real-valued functions. Definition of uniform convergence Uniform convergence and continuity. The cauchy condition for uniform convergence Uniform convergence of infinite series of functions. A space-filling curve Uniform convergence and Riemann-Stieltjes integration Nonuniformly convergent sequences that can be integrated term by term Uniform convergence and differentiation Sufficient conditions for uniform convergence of a series Uniform convergence and double Sequences. Mean convergence Power series . Multiplication of power series. The substitution theorem Reciprocal of a power series Real power series The Taylor's series generated by a function Bernstein's theorem . The binomial series .

218 219 220 221 222 213 224 225 226 228 230 231 232 234 231 238 239 240 241 242 244

:':. -:
m
9.22 9.23 Abers limit tbc:on:m Tauber's theorem
,\

,Of

~."..

Con __ :1"

Exercises •
Chlpter"tO

244 246
247

I1.12 11.13 11.14 11.15 11.16

)1.11

The ..........

IDtegra]

10.1 10.2 10.3 10.4 105

10.6

10.7
10.8

10.9 10.10 10.11 10.12 10.1l 10.14 10.lS 10.16 10.17 10.18 10.19 10.20 10.21 10.22 10.23 10.24

Introduction . The integral of a step function Monotonic sequences of step fUJlCtions Upper functioBli and their integrals Riemann-integrable functions as examples of upper functions The class of Lebesgue-integrable functions on a general ioterval Basic properties of the Lebesgue integral. Lebesgue integration and sets of measure zero The Levi monotone convergence theomns The Lebesgue dominated converpna: theorem Applications of Lebesgue's dominated convergence theorem Lebesgue inlf:grals on unbounded intervals as limits of integrals on bounded intervals Improper Riemann intcgrali Measurable functions Continuity of functions defined by Lebesgue integrals Differentiation under the integral sign Interchanging the order of integration Measurable sets OD the real line The Lebesgue integral over arbitrary subsets of R Lebesgue integrals of complex-valued functions . Inner products and norms . The set L2(1) of square-integrable functions • The set L2(1) as a semimctric space A wnvcrgcncc theorem for series of functions in L2(1) Exercises .

252 253
2S4

256 2.59 260

261 264 265
270

11.18 11.19 11.20 11.21 11.22

Sufficient conditions for conversence of II Fourier series at a particular point Cesaro summability of Fourier series • Consequences of Fejer's theorem . The Weierstrass approximation theorem Other forms of Fourier series . The Fourier integral theorem The exponential form of the Fourier integral theorem Integral transforms . Convolutions The convolution theorem for Fourier transforms The Poisson summation formula Exercises
MAltivariable

319 119 321 322 322

323

325
326

327 329
332 335

Chaptet" 12
12.1

DifferentiaJ Calculus
3M

272 274 276

279
281

283 287
289 291 292 293

12.2 12.3 12.4 I2.S 12.6

12.7

294295

12.8 12.9 12.10 12.11 12.12

J2.13
12.14

)0.25 The Riesz-Fischcr theorem

295
297
298

Introduction . The directional derivative Directional derivatives and continuity The total derivative. The total derivative expressed in terms of partial derivatives An application to complex-valued functions. The matrix of a linear function The Jacobian matrix The chain rule Matrix form of the chain rule. The Mean-Value Theorem for differentiable functions A sufficient condition for differentiability A sufficient condition for equality of mixed partial derivatives Taylor's formula for functions from R" to Rl Exercises.

344
345 346 347 348 349

3S1 352 353 355 357 358 361 362

Cbapter 13 Implicit Ftmctioos and Extremum Problems a.apterU Il.l

11.3

Fourier Series and F ......

11.1 Introduction
11.4
115

11.6 11.7 11.8

ILIO
lUI

11.9

. Orthogonal systems of functions . TIle theorem on best approximation The Fourier series of a function relative to an orthonormal system . Properties or the Fourier coefficients . The Riesz-Fiscber theorem The convergence and representation problems for trigonometric series The Riemann-Lebesgue lemma The Dirichlet integrals . An integral representation for the partial sums of a Fourier series Riemann's localization theorem

306 306
307

309 309 311 312 313 314 317 318

13.1 Introduction . 13.2 Functions with nonzero Jacobian determinant 13.:1 The inverse function theorem 13.4 The implicit function theorem . 13.5 Extrema of real-valued functions of one variable 13.6 Extrema of real-valued functions of several variables 13.7 El\trcmum problems with side conditions Exercises .
Cbapter 14

367 368

372
373 375 376
380

384

Multiple Riemann Integrals 388 388

14.1 Introduction n 14.2 The measure of a bounded interval in R

llyii

The RiemaDn integral of a bounded function defuIed on a compact interval in R- . ... .. 14.4 Sets of measure zero and Lebesgue's criterion for existence of a multiple Riemann integral .. 14.5 Evaluation of a multiple integral by iterated integration 14.6 Jordan-measurable sets in R- . 14.7 Multiple integration over Jordan-measurable sets 14.8 Jordan content cx.J)I'es&ed 118 a Riemann integnd 14.9 Additive property of the Riemann integral . 14.10 Mean-Value Theorem for multiple integrals • Exercises .
14.3

389
391 391

396 3!i7 398

399 400 402

16.16 16.17 16.18 16.19 16.20 J6.21 16.22 16.23

The identity theorem for analytic functions . The maximum and minimum modulus of an analytic function The open mapping theorem Laurent expansions for functions analytic in an annulus

452 453 454

455
457 459
4(i()

Isolated singularities The residue of a function at an isolated

The Cauchy residue theorem .

singular

point

Chapter 15 Multiple J..ebesgae bte&mIs 15.1 Introduction. J 5.2 Step functiollll and their integrals . 15.3 Upper functioos and Lebesgue-intesmblc fUDctiOJlS 15.4 Measwablc functions and IJ:Jea$IJA.ble sets in R••... 15;5 Fubw's reduction theorem for the double integral ofa step function 1S.6 Some properties of sets of mc:asurc zero. •• 15.7 Fubini's reduction tbeoRIIl for double integrals 15.8 The TODClli-Hobson test for integrability 15.9 Coordinate transforma.t.ions. . 1s.t0 TIle transformation formula for multiple integnds. •..•. 15.11 Proor of the transformation formula for linear coordinate tramlfonna-

Counting zeros and poles in a region Evaluation of real-valued integrals by means of residues 16.24 16.25 Evaluation of Gauss's sum by residue calculus . 16.26 Application of the residue theorem to the inversion fonnula transforms 16.27 Conformal mappings
Exercises . IDdex of Special Symbols

461 462 464 for Laplace

468
410 412 481 485

405
406 406
407 409 411 413 415

Index .

CJ

416
421 421

tions

..

15.12 15.13

Proof of the transformation

compact cube Completion of

Exercises .
<hIpter 16 16.1 16.2 16.3 16.4
16.5

too

fOImuia (or the characteristic function of a .. proof of the transformation formula

423 429
430

CaudlTs

1beorem and die Residue

CakuIoI
434

Analytic functions .• Paths and curves in the complex plane

16.6 16.1 16.8

16.9

16.10 16.11 16.12 16.13

16.14

J6.1S

Contour integrals • . . •. The integral along a cin:ular path as a fum::tion of tile radilJ8 • atucby"s integral theorem for a circle Homotopic curves • Invarianee of contour integrals under homotopy GtDera1 form of atuchy's integral theorem • Cauchy's integral formula.. •••... The winding number of a circuit with mspcct to a point . .. The unboundedness of the set of points with winding number zero Analytic functions defined by contour integrals. Power-series expansiOIlll for analytic functions Cauchy's inequalities. LiouviUc's theorem . Isolation of the zeros of an analytic function

43.5 436 438

439

439 442 443

443

444 446 447

~ 450

451

o CHAPTER 1

THE REAL AND COMPLEX NUMBER SYSTEMS

1.1 IN'IRODUcnON

Mathematica1 analysis studies concepts related in some way to real numbers, so

we begin our study of analysis with a discussion of the real-number system.

".
"

Several methods are used to introduce real numbers. One method starts with the positive integers I, 2, 3. . .. as undefined concepts and uses them to build a larger system, the positive rational numbers (quotients of positive integers), their negatives. and zero. The rational numbers, in tum. are then used to construct the irratiorwi numbers. real numbers like.,fi and ~"""hichan: not rational. The rational and irrational. numbers together constitute the real-number system. Although these matters are an important part of the foundations of mathematics, they will not be described in detail here. As a matter of fact, in most phases of analysis it is only the properties of real numbers that concern us, rather than the methods used to construct them. Therefore, we lhall talce the real numbers themselves as undefined objects satisfying certain axioms from which further properties will be derived. Since the reader is probably familiar with most of the properties of real numbers discussed in the next few pages, the presentation will be rather brief. Its purpose is to review the important features and persuade the reader that, if it were necessary to do so, all the properties could be traced back to the axioms. More detailed treatments can be found in the references at the end of this chapter. For convenience we usc some elementary set notation and terminology. Let S denote a set (a collection of objects). The notation XES means that the object x is in the set S, and we write x ¢ S to indicate that x is not in S. A set S is said to be a subset of T, and we write S !;; T, if every object in S is . also in T. A set is called 1I01U!mpty if it contains at least one object. We assume there exists a nonempty set R of objects, caned real numbers. which satisfy the ten axioms listed below. The axioms fall in a natural way into three groups which we refer to as the field axioms. the order axioms. and the completeness axiom (also cailCd the least-upper-bownd axiom or the axiom of continuity). 1.1 THE FIElD AXIOMS

'.

Along with the set R of real numbers we assume the existence of two operations. called addition and multiplication, such that for every pair of real numbers x and y
1

I
;

!~

As. 1

11L 1.1

the sum x + y and the product xy are real numbers uniquely determined by x and y satisfying the following axioms. (In the axioms that appear below, x, y, z represent arbitrary real numbers unless something is said to the oontrary.) A:ri(!m 1. x

+ y = y + x, xy = yx A:dom2. x + (y + z) = (x + y) + Ax;om 1. x(y + z) = xy + xz

(commutative laws).
z; x(yz) = (xy)z iassociatioe laws).

x < y < z,

Thus we have 2 s: 3 since 2 < 3; and 2 s 2 since 2 2. The symbol ~ is similarly used. A real number x is called nonnegative if x ~ O. A pair of ;imultaneous inequalities such as x < y, y < Z is usually written more brir fiy as The following theorem, which is a simple consequence is often used in proofs in analysis. Theorem 1.1. Gioen real numbers a and b such that of the foregoing axioms,

(distributive law).

Axiom 4. Given any two real numbers x and y, there exists a real number z such that x + z = y. This z is denoted by y -'x; the number x - x is denoted by O. (11 can be proved that 0 is independe1tt of x.) We write =x for 0 - x and call -x the

Then a

.s

for every t h.

> O.

(1)

negative 0/ x,

Proof, If b < a, then inequality (1) is violated for e b

Therefore, Axiom

(a - b)/2 because
=
Q.

There exists at least one real number x #- O. If x and y are two real numbers with x #c 0, then there exists a real number z such that xz = y. This z is denoted by yix : the number xlx is denoted by 1 and can be shown. to be independent of x. We write X-I for l/x if x of. 0 and call x-1 the reciprocal of x.
A:ri(!m 5. From these axioms all the usual laws of arithmetic (-x) x, (X-t)-l x, -(x - y) = y x, x a more detailed explanation, see Reference 1.1.) 1.3 THE ORDER AXIOMS We also assume the existence of a relation < which establishes the real numbers and which satisfies the following axioms: A.xiom 6. Exactly
NOTE.

+z

+a- b

22.2

b<

+a

by Axiom 6 we must have a ,:s:; h. 10, the completeness axiom, will be described in Section 1.11.

can be derived; for example, y = x + (-y), etc. (For

1.4 GEOMETRIC

REPRESENTATION

OF REAL NUMBERS

an ordering among

one uf the relations x = y, x

< y,

>

y holds.

x > y means the same as y < x. 7. If x < y. then for every z we hove x

The real numbers are often represented geometrically as points on a line (called the real line or the real axis). A point is selected to represent 0 and another to represent I, as shown in Fig. 1.1. This choice determines the scale. Under an appropriate set of axioms for Euclidean geometry, each point on the real line corresponds to one and only one real number and, conversely, each real number is represented by one and only one point on the line. It is customary to refer to the point x rather than the point representing the real number x.

Axiom

+z <y +

z,

ex

Figure 1.1
II

Axiom 8. A.xiom 9.

If X >

0 and y and y

> O. >
Z,

then xy then x

>

O. The order relation has a simple geometric interpretation. If x < y, the point x lies to the left of the point y. as shown in Fig. 1.1. Positive numbers lie to the right of 0, and negative numbers to the left of O. If a < b, a point x satisfies the inequalities a < x < b if and only if x is between a and b. 1.5 INTERVALS The set of all points between a and b is called an interval. Sometimes it is important to distinguish between intervals which include their endpoints and intervals which do not.
NOTATION. jz,"

If x > y

> z,

NOTE. A real number x is called positive if x > 0, and negative if x < O. We denote by R + the set of all positive real numbers, and by R - the set of all negative real numbers.

From these axioms we can derive the usual rules for operating with inequalities. For example, if we have x < y, then xt < yz if z is positive, whereas xz > yz if z is negative. Also, if x > y and z > w where both y and w are positive, then xz > yw. (For a complete discussion of these rules see Reference L I.)
NOTE.

The symbolism

x :s:; y is used as an abbreviation "x < y or x

for the statement:

The notation {x: x satisfies all real numbers x which satisfy property

PJ
P.

will be used to designate

the set of

..
(a, b) = {x: a

Def.l.2

n.1.7
a prime if n > 1 and if the only positive divisors of n are 1 and n. If n > 1 and n is not prime, then n is called composite. The integer 1 is neither prime nor composite. This section derives some elementary results on factorization of integers, culminating in the unique factorization theorem, also called the fundamental theorem of arithmetic. The fundamental theorem states that (I) every integer n > 1 can be represented as a product of prime factors, and (2) this factorization can be done in only one way. apart from the order of the factors. It is easy to prove part (I). Tlu!onm 105. Every integer n > I is either a prime or a product of primes. Proof We use induction on n. The theorem holds trivially for n = 2. Assume it is true for every integer k with 1 < k < n. If n is not prime it has a positive divisor d with 1 < d < n. Hence n = cd, where 1 < e < n. Since both c and dare «n, each is a prime or a product of primes; hence n is a product of primes. Before proving part (2). uniqueness of the factorization, we introduce some further concepts, If dla and dlb we say d is a common divisor of Q and b. The next theorem shows that every pair of integers a and b has a common divisor which is a tinear combination of a and b. 1'Iu!or_ 1.6. Every pair of integers a and b has a common dipuor d of the form d=ax+by
this d.

lk./illitio" 1.2. Assume a < b. The open interval (a, b) is d4ined to be the set

<

<

b}.

The closed interval [a, b] is the set {x: Q.':-:;; x ;S; b}. The half-open intervals (a. b] and [a, b) are similarly tkjined, using the inequalities a < x ;S; b and a :$ x < b, respectively. Infinite intervals are tkfined as follows: (Q,+oo) (-GO, a)
=

{x:x>a}, {x:x

[a. +00) (-GO. a]

{x:x
= {x:x:5:

~ a}, a}.

< a},

The real tine R is. sometimes referred to as the open interval (-00, +00). A single point is. considered as a "degenerate" dosed interval.
NOTE.

The symbols + GO and - GO are used here purely for convenience in notation and are Dot to be considered as being real numbers. Later we shall extend the real-number system to include these two symbols, but until this is done, the reader should understand that all real numbers are "finite."
1.6 INTEGEllS

This section describes the integers, a special subset of R. Before we define the integers it is convenient to introduce first the notion of an inductioe set.

De./illititm 1.3.
two properties.-

A set of real numbers is called an inductise set

if it ho.J

the following

a) The number 1 is in the set. b) For every x in the set, the number x

where x and y are integers.

. .

Moreover, every common divisor of a and b divides

+ 1 is also in

the set.

For example, R is an inductive set. So is the set R.... Now we shall define the positive integers to be those real numbers which belong to every inductive set.
De}i1litilnll.4. A real number is called a positive integer inductive set. The set of positive integers is tknoted by Z +.

if

it belongs to enery

The set Z + is itself an inductive set. It contains the number I. the number 1 + 1 (denoted by 2), the number 2 + I (denoted by 3), and so on. Since Z+ is a subset of every inductive set, we refer to Z+ as the smalJest inductive set. This. property of Z+ is sometimes called the principle of inductwn. We assume the reader is familiar with proofs by induction which are based on this principle, (See Reference 1.1.) Examples of such proofs are given in the next section. The negatives of the positive integers are called the negative integers. The positive integers, together with the negative integers and 0 (zero), form a set Z which we call simply the set of integers.
1.7 THE UNIQUE FACTORIZATION THEOREM FOR IN1EGERS

Proof First assume that a ~ O. b :2-. 0 and use induction on n a + h. If n = 0 then a = b = 0, and we can take d = 0 with x = y = O. Assume, then, that the theorem has been proved for 0, 1,2, ...• n 1. By symmetry, we can assume a ~ h. If b = 0 take d = a, x = I. Y = O. If b ~ 1 we can apply the induction hypothesis to a - b and b, since their sum is a n - b ~ PI - 1. Hence there is a common divisor d of a - band b of the form d = (a - h)x + by. This d also divides (a - b) + b = a, so d is a common divisor of a and band we have d = ax + (y - x)b, a linear combination a and b. To complete the proof we need to' show that every common divisor divides d. Since a common divisor divides a and b, it also divides the linear combination ax + (y - x)h = d. This completes the proof if a ~ 0 and b ~ O. If one or both of a and h is negative, apply the result just proved to lal and Ibl.

of

If nand d are integers and if n = cd for some integer c, we say d is a divisor of n; or n is a multiple of d, and we write din (read: d divides n). An integer n is called

NOTE. If d is a common divisor of a and b of the form d = ax + by, then -d IS also a divisor of the same form, - d = a( - x) + b( - y). Of these two common divisors, the nonnegative one is called the greatest common divisor of a and b; and is denoted by gcd(a, b) or, simply by (a, b). If (0. b) = l then a and b are said to be relatively prime.

Tlu!orem 1.7(EIICHtI's Lemmq).

If albc and (a. b)

I, then

ale.

Ii

Real ...

Complex NIUIlbe£ Sy:rtems

Th.U

1b. 1.11

Proof, Since (a, b) = I we can write 1 But alacx and albcy, so ale.

ax

+ by.

Therefore c

acx

bey.
p

1.9 IRRATIONAL NUMBERS

Theorem 1.B. If a prime p divides ab, then pia or plb. More generally, divides a product a 1 ••• Ok. then p divides at least one of the fa(,;tors.

if a prime
;0;;

J"j, e,

Proof, Assume plab and that p does not divide a. If we prove that (p, a) 1, then Euclid's Lemma implies plb. Let d = (P. a). Then dip so d = 1 or d '" p. We cannot have d = p because dla but p does not divide 0. Hence d J. To prove the more general statement we use induction on k, the number of factors. Details are left to the reader.

Real numbers that are not rational are called irrational. For example, the numbers 'It" and e'" are irrational. Ordinarily it is not too easy to prove that some particular number is irrational. There is no simple proof, for example, of the irrationality of e". However, the irrationality of certain numbers such as../2 and";3 is not too difficult to establish and, in fact, we easily prove the following; Tlu!im!m 1.10. 1/!'I is a positive integer which is not a perfect square, then";;'
irrational.

is

TAeorellf 1.9(U";qfle/tldOTiUltio" theorem). Every integer n > I can be represented us If product of prime factors in only one way, apart from the order of the factors. Proof We use induction on n. The theorem is true for n "" 2. Assume, then, that it is true for all integers greater than I and less than n. If n is prime there is nothing more to prove. Therefore assume that n is composite and that n has two factorizations into prime factors, say
(2)

We wish to show that s = t and that each p equals some q, Since PI divides the product q!qz ... qf' it divides at least one factor. Relabel the q's if necessary so that pdq" Then PI = ql since both PI and qt are primes. In (2) we cancel PI on both sides to obtain n
PI

Proof. Suppose first that n contains no square factor> l. We assume that ..;;, is rational and obtain a contradiction. Let";; = alb, where a and b are integers having no factor in common. Then nb2 = 0,2 and, since the left side of this equation is a multiple of n, so too is a2• However. if cl is a multiple of n, a itself must be a multiple of n, since n has no square factors. > I. (This is easily seen by examining the factorization of a into its prime factors.) This means that a = en, where c is some integer. Then the equation nbz = 0,2 becomes nb2 = c2n2• or b2 nez. The same argument shows that b must also be a multiple of n. Thus a and b are both multiples of n, which contradicts the fact that they have no factor in common. This completes the proof if n has no square factor > I. If n has a square factor, we can write 11 = m2k. where k > I and k has no square factor> I. Then";; = mil.; and jf"/; were rational, the number ../k would also be rational, contradicting that which was just proved. A different type of argument is needed to prove that the number e is irrational. (We assume familiarity with the exponential e" from elementary calculus and its representation as an infinite serie~.) Tieorem 1.11. If e" number e is irrational. I

Since n is composite, I < nJpI < n; so by the induction hypothesis the two factorizations of nIp! are identical, apart from the order of the factors. Therefore the same is true in (2) and the proof is complete.
U RATIONAL NUMBERS

+ x + x2f2! +

xlJ3!

+ ... + x"fn! + ... , then the

Quotients of integers afb (where b "'" 0) are called rational numbers. For example, 1/2, -1/5, and 6 are rational numbers. The set of rational numbers, which we denote by Q. contains Z as a subset. The reader should note that all the field axioms and the order axioms are satisfied by Q. We assume that the reader is familiar with certain elementary properties of rational numbers. For example, if a and b are rational, their average (a + b)/2 is also rational"and lies between a and b. Therefore between any two rational numbers there are infinitely many rational numbers, which implies that if we are given a certain rational number we Cannot speak of the "next largest" rational number.

Proof We shall prove that e-I is irrational. The series for e-1 is an alternating series with terms which decrease steadily in absolute value. In such an alternating series the error made by stopping at the nth term has the algebraic sign of the first neglected term and is less in absolute value than the first neglected term. Hence. if Sit = Ll=o (_l)c/k!. we have the inequality

O < e -[ from which we obtain O«2~-1)!(e-l for any integer k ;::::I. Now (2k rational, then we could-choose k

Slk-l

1 <--,
(2k)!

-Slk-I)<2~:::;;~' f)!
SQ S2i-l

(3)

is always an integer. If e-1 were large that (2k - I)! e-1 would also be an

•
NOrE.

Dfl.l.1l

1'11. 1.14

integer. Because of (3) the dift"erence of these two integers would be a number between 0 and i. which is impossible. Thus e-1 cannot be rational, and hence e cannot be rational.

For sets like the one in Example 3. which are bounded above but have no maximum element. there is a concept which takes the place of the maximum ele-ment. It is called the least upper bound or supremum of the set and is defined as

For a proof that

'It:

is irrational. see Exercise 7.33.

follows: De./iIdtion 1.13. Let S be a set of real numbers bounded aboDe. A real number b ir called a least upper bound for $ if it has the following two properties; a) b is an upper hound for S. b) No number less than b is an upper boundfor S.
&•..,...,.. S If [0, 1] the maximum element I ill also a least upper bound for S. If S = [0,1) the number 1 is a least upper bound for S, even though S has no maximum element. It is an easy exercise to prove that a set cannot haw two different least upper bounds. Therefore. if there is a least upper bound for S. there is only one and we can speak of the least upper bound. It is common practice to refer to the least upper bound of a set by the more concise term supremum, abbreviated sup. We shall adopt this convention and write

The ancient Greeks were aware of the existence of irrational numbers as early as 500 D.C. However, a satisfactory theory of such numbers was not developed until late in the nineteenth century, at which time three different theories were introduced by Cantor. Dedekind. and Weierstrass. For an account of the theories of Dedekind and Cantor and their equivalence. see Reference 1.6. LtO UPPER. BOUNDS, MAXIMUM
(SVPREMlJM)

ELEMENT,

LEAST UPPER BOUND

Irrational numbers arise in algebra when we try to solve certain quadratic equations. For example, it is desirable to have a real number x such that = 2. From the nine axioms listed above we cannot prove that such an x exists in R. because these nine axioms are also satisfied by Q and we have shown that there is no rational number whose square is 2. The completeness: axiom allows us to introduce irrational numbers in the real-number system, and it gives the real-number system a property of continuity that is fundamental to many theorems in analysis. Before we describe the completeness axiom, it is convenient to introduce additional terminology and notation.

b
to indicate

supS

De/Ulitbnr 1.12. LeI S be "

that b is the supremum of S. If S has a maximum element. then sup S. The greatest lower bound, or i"./imum of S. denoted by inf S, is defined in an analogous fashion.

max: S

set of real1rUl11bers. If there is a real number b such that x :s: b for every x in S, the" b is called an upper bound for S and we say that S is bounded aboDe by b.

I.II

THE COMPI..ETENliliS AXIOM system involves the notion of supremum.

We say an upper bound because every number greater than b will also be an upper bound. H an upper bound h is also a member of S. then b is called the largest member or the maximum element of S. There can be at most one such h. If it exists. we write b = max S. A set with no upper bound is said to be unhouruled above. Definitions of the terms lower bound, bounded below. smalJest member (or minimum element) can be similarly Connulated. If S has a minimum clement we denote it by min S.
ED...,...,.

Our final axiom for the real number

A.xiom 10. Every nonempty set S of real monbers which is bounded above IuJs a supremum; that is, there is a real number b such that b = sup S.

As. a consequence
numbers

of this axiom it follows that every nonempty which is bounded below has an infimum. OF THE SUPREMUM

set of real

1.12 SOME PROPERTlFS This section discusses some useful in this text. There is the reader should formulate The fint property shows trariJy close to its supremum

I. TIle set R + "" (0, + 00) is unbounded above. It has no upper bounds and no mu· imum element. It is bounded below by 0 but has no minimum elooxot.

.z.

The closed. interval S = [0, 1] is bounded above by I and is bounded below by O. In fact, max S = 1 and min S ... O.

fundamental properties of the supremum that will be a corresponding set of properties of the infimum that for himself. that a set with a supremum contains numbers arbi.

3. The balf-open interval S [0, 1) is bounded above by 1 but it has no maximum element. Its minimum element is O.

T_rem 1.U(Approximatiollpropel'ty). Let S be a nonempty set o/real numbers with a supremum, say b = sup S, Then for every a < b there is some x in S such that a < x s; h.

to

Real aocI CompJex N.mer SysteJn$

Tb. US

Tb.UO

11

Proof. First of all, x !5: h for all x in S. If we had x !5: a for every x in S, then a would be an upper bound for S smaller than the least upper bound. Therefore x > a for at least one x in S. Tllunm 1.15 (Additir;e properly).
Given nonempty subsets A and B of R, let C denote the set

covered by a finite number of line segments of a given positive length, no matter how small.
Theorem 1.19. If x > 0 and integer n such that nx > y.

if

y is an arbitrary

real number, there is a positive

c=

{x

+y =

:x

A,

Y E B}.

Proof. Apply Theorem L 18 with x replaced by ylx.

1.1S RATIONAL NUMBERS

1/ each of A and B has a supremum,

then C has a supremum and

wrra

FINITE DECIMAL

REPRESENTATION

sup C

sup A

+ sup

B.

A real number of the form

Proof. Let a = sup A, b = sup B. If Z E C then z = x + y, where x E A, Y E B, so z = x + y :::;; + h. Hence a + b is an upper bound for C, so C has a a supremum, say c = sup C, and c :::;; + b. We show next that a + b ~ c. a Choose any 6 > O. By Theorem 1.14 there is an x in A and a y in B such that and Adding these inequalities we find a
Thus,
0

1=60+-+ a1

10

1()2

aa + ... +-

10" '
aj !5:

b-

s < y.

where a(} is a nonnegative integer and a1> ... is usually written more briefty as follows:

, a~

are Integers satisfying 0 ~

9,

+b£

2s

<x +

y :::;; . c

This is said to be ajinite decimal representation of r, For example,

+b<c+

2s for every

> 0 so, by Theorem ].1, a

+ b s; c.

-=-=0.5

1 2

5 10

'

50

-=-

lal

=002 .•

- = 7+- +-

29 4

2 10

5 = 7.25. 10~

The proof of the next theorem is left as an exercise for the reader. Theorem 1.16 (Compariwnproperly). that 8 :..; t for every s tn S and t in T.
and Given nonempty subsets Sand T of R such If T has a supremum then S has a supremum

sup S :::;; T. sup

1.13 PROPERTIES OF THE INlEGERS COMPLETENESS AXIOM Tlletlrem 1.17. The set Proof. DEDUCED FROM THE

Real numbers like these are necessarily rational and, in fact, they all have the form r = a/1ft, where Q is an integer. However. not all rational numbers can be expressed with finite decimal representations. For example. if 1- could be so expressed, then we would have i = a/10" or 30 = lO" for some integer a. But this is impossible since 3 does not divide any power of 10.
1.1' FINITE DECIMAL APPROXIMATIONS TO REAL NUMBERS

z+ of positive

integers 1,2,3, ... is unbounded above.

If Z+ were bounded above then Z+ would have a supremum, say a = sup Z+. By Theorem 1.14 we would have a - I < n for some n'in Z+. Then n + 1 > (.l for this n. Since n + I € Z + this contradicts the fact that a = sup Z +. TMoum I.IB. For every real x there is
0 positive

This section uses the completeness axiom to show that real numbers can be approximated to any desired degree of accuracy by rational numbers with finite decimal representations.
Tltevrem 1.10. Assume x ~ O. Then for every integer n ~ I there is a jinite

decimal r.

= 00' ala:.! ••• alt

such that

integer

II

such that n > x.

Proof: If this were not true, some x would be an upper bound for Z ", contradicting Theorem 1.17. 1.14 THE ARCHlMEDEAN PROPERTY OF THE REAL NUMBER SYSTEM

r~ s x <

r; +

1 10'"

The next theorem describes the Archimedean property of the real number system. Geometrically, it tells us that any line segment, no matter how long, can be

Proof Let S be the set of all nonnegative integers Sx. Then Sis nonempty, since 0 E S, and S is bounded above by x, Therefore S has a supremum, say ao = sup S. It is easily verified that ao E S, so aD is a nonnegative integer. We call ao the greatest integer in x, and we write llo = [x]. Clearly, we have

ao ~ x <

00

I.

u
Now let "1 = [lOx· - l~]. 5 lOx - lOuo = 100x the greatest integer in lOx - 10ao. 10, we have 0 ~ al :S 9 and lOao

n. 1.2.1
Since

111.1.11 number. the absolute value of x, denoted by [x], is defined as follows: if x ~ 0,

ao) <

o. In other
words,
Q1

:S lOx -

<

Q1

1.
A fundamental inequality

Ixl

={

-x,

x,

if x :s:; O.
values is given in the following:

is the largest integer satisfying the inequalities

concerning

absolute

tlo

+ "1 :s: x < tlo + a1 +

10

to

1.

More generally. having chosen a1> •••• largest integer satisfying the inequalities Qo

a .._1 with 0 :S a,

:s: 9,
+
t

T/u!orem 1.21. If a <::: 0, then we have the inequafity Ixl =a s: x:s:: a.

let a" be the

:s:

if. and only if,

10

til

+ ... + ..!3 :s:

10"

<

a
0

10

01

+ ., . +

aM

10" .

(4)

Then 0 :s:; a" :s:; r

I) and we have

Proof From the definition of lxi, we have the inequality -Ixl s: x ::s; ixl, since x = Ixl or x -Ixl. If we assume that Ixl :s: a; then we can write - a :s:; -Ixl :s:; x :s:; Ixl :s:; a and thus half of the theorem is proved. Conversely, let us assume -a :s:; x :s:; a. Then if x ~ 0, we have Ixl = x :s:; a, whereas if x < 0, we have [x] = -x :s a. In either case we have [x] :s:; a and the theorem is proved.
We can use this theorem to prove the triangle inequality.

< "" _ x -c r,

+ -, 1

to"
It is easy to verify that x is

where rIO = ao .alaz .. , a... This completes the proof. actually the supremum of the set of rational numbers 1.17 INFINI1E DECIMAL REPRESENTATIONS

Tlreorem 1.21. For arbitrary real x and y we have

r" rz•....

Ix + yl

:s: Ixi + Iyl

proves

(the triangle inequality).

OF REAL NUMBERS 1.20 can be used to

Proof. We have -Ixl.:s:: x :s:; -(lxl + Iy!) :s:; x + y :s: Ixl +

The integers Qo. at> a2• • •• obtained in the proof of Theorem define an infinite decimal representation of x. We write

Ix
x

+ yj ~ Ixl + Iyl·

Ixl and -lyl:S: y :s:; Iyl. Addition gives us Iyl. and from Theorem 1.21 we conclude that
the theorem. For example, if we take

This

x
00 = 0,

ao'"la~'"

The triangle inequality is often used in other forms. a - , and y = , - b in Theorem 1,22 we find la -

to mean that a,. is the largest integer satilifyi.t18 (4). For example, if x = 1 find we a, = 1. a2 2, a3 = 5. and a" 0 for all n ~ 4. Therefore we can write i = 0.125(XX}···

hi ~

la -

c]

[c -

hi.
Taking

Also, y=

from Theorem b, we obtain

1.22 we have

Ixl ~ Ix

+ yl - Iyl·

b,

If we interchange the inequality signs :S and < in (4), we obtain a slightly ditJerept definition of decimal expansions. The finite decimals T.. satisfy T. < x :S '" + 10-" although the digits ao, a1> Oz, ••• need not be the same as those in (4). For example, ifwe apply this second definition to x = i we find the infinite decimal
representation .

la + hi ~ lal - Ibl. la + hi ;;:: Ihl - lal

=

Interchanging hence

a and b we also find

-(lal - Ihl), and

itl + hi ~ Iial - Ihll·

0.124999'':''

The fact a real number might have two different decimal representations is merely a reflection of the fact that two different sets of real numbers can have the same supremum. 1.18 ABSOLUTE VALUES AND THE TRIANGLE INEQUALITY

mat

. By induction we can also prove the generalizations

and

Calculations with inequalities arise quite frequently in ana1ysis. They are of particular importance in dealing with the notion of absolute value. If x is any real

1.19 THE CAUCHY-SCHWARZ We shall now derive another

INEQUALITY which is often used in analysis.

inequality

14

Real lind Complu. Nombel' Systems

Th.l.13 and

Del. t.26

IS

Theorem 1.23 (CtlJlChy-Scitwurr; inequlllity). arbitrary real numbers, we haul"

If

aj,

•.••

Q"

bj,

•••

b"

are

b) If x > 0, then we have

x(+oo)

+00,
we haxe

X(-oo)

00.

c) If
Moreover, QlX + hie Proof

x<

0, then
-00, + (+00) + (-00)

if

some aj # 0 equality holds Of or each k 1,2, ... ,n.

if

and only

if

there is a real x such lha!

x(+oo)

x(-oo) = (+00)(+00) = (+00)(-00)

+00. +00,

d) (+00)
(-00)

A sum of squares can never be negative. Hence we have

= =

(-00)(-00) -00.

k~t

L (a,x + bli
II

e) If x E R, then. we have -00

<x <

+00.

~0

for every real x, with equality if and only if each term is zero. This inequality can be written in the form
Ax:.!

NOTATION. WedenoteRby 00, +00) andR* by [-00, +00]' ThepointsinR are called "finite" to distinguish them from the "infinite" points + 00 and - 00.

where
A

2Bx

C 2 0,

Loi,
l~l

IfA> If A

0, put x = - B/ A to obtain B z - A C 0, the proof is trivial.

s;

0, which is the desired inequality.

The principal reason for introducing the symbols + 00 and - 00 is one of convenience. For example, if we define + 00 to be the sup of a set of real numbers which is not bounded above, then every nonempty subset of R has a supremum in R*. The sup is finite if the set is bounded above and infinite if it is not bounded above. Similarly, we define - 00 to be the inf of any set of real numbers which is not bounded below. Then every nonempty subset of R has an inf in R*. For some of the later work concerned with limits. it is also convenient to introduce the following terminology. Definitioll 1.25. Every open interval (a, + co) is called a neighborhood of + 00 or a ball with center + 00. Every open interval ( - 00, a) is called a neighborhood of - 00 or a ball with center 00.
I.II COMPLEX NUMBERS

NOTE.

In vector notation the Cauchy-Schwarz inequality takes the form

(B'W
where a = (aJ> ... , a,,), b
(b l'
... ,

~ lIall'"Ifbll1.,
h.) are two n·dimensional vectors, .

a'b is their dot product, and

k=1

• L alA,

Iiall

= (a' a)1/2 is the length of B. AND THE EXTENDED REAL NUMBER

1.20 PLUS AND MINUS INFINITY

SYSTEM R* by the symbols

Next we extend the real number system by adjoining two "ideal points" denoted + 00 and - 00 ("plus infinity" and "minus infinity").

It foliows, from the axioms governing the relation < that the square of a real number is never negative. Thus, for example, the elementary quadratic equation x2 = - I has no solution among the real numbers. New types of numbers, called complex numbers, have been introduced to provide solutions to such equations. It turns out that the introduction of complex numbers provides, at the same time, solutions to general algebraic equations of the form
Q(}

+ a1x + ... + a"x!' =

0,

/kfiwitWn 1.U. By the extended rea/number system R* we shall mean the set of real nllmbers R together with tWQ symbols + 00 and 00 which satisfy the followinq

properties .a) If x x
E

where the coefficients 00' Ot •••• , a. are arbitrary real numbers. (This fact is known as the FundamentaI_Theorem of A (qebra.) We shall now define complex. numbers and discuss them in further detail. De./illition 1.26. Bya complex number we shaJ/ mean an ordered pair of rea/numbers which we denote by (xl> x2). The first member, Xl' is called the real part of the
complex number; the second member, Xl. is called the imaginary part. Two complex numbers x = (x,; Xl) and y = (Yl, Y2) are called equal, and we write x = y, if,

R, then we have

+ (+ (0) = + 00,

x - (+ 00) xl( + 00)

= - 00, = x/( 00)

o.

x+ 00) x-(-oo)::::

00,

+00,

16

D.l.n

17 convenient to think oftbe real number system as being a special case of the complex number system. and we agree to identify the complex number (x, 0) and the real number x. Therefore, we write x = (x. 0). In particular, 0 = (0, 0) and 1 = (1,0). 1.2Z GEOMETRIC REPRESENTATION OF COMPLEX NUMBERS

and only if, Xl

Yl and Xl

Yz. We define the sum

+ Y and the product xy

by

the equations

NOTE.

The set of all complex numbers will be denoted by C.

Tluwum 1.17. The operations of addiliOf1 and multiplication just defined satisfy the commutative, associatioe, and distributive laws.
Proof We prove only the distributive law; proofs of the others are simpler. x = (x,; x2),! = (Yl> Y2), and z = (z,. 22). then we have x{y

If"

Just as real numbers are represented geometrically by points on a line, so complex numbers are represented by points in a plane. The complex number x = (XI> X2) can be thought of as the "point" with coordinates (x 1> x2). When this is done, the definition of addition amounts to addition by the parallelogram law. (See Fig. 1.2.)

z)

= =

(Xb

X2)(}'1

(XI}'I (XlYI

xy

+ XI.

+ ZI> Fz + Z2) + XIZI X2Yl - X2Z2, XI Y2 + XII2 + XlYI + XlII) - X2Y,., X1Yl + X2Y.) + (XlIt - XlZl, XII2 + X2I,) =

TltHum 1.18.
(X1o
Xl)

+ (0.0) =
0) =

(XI>

x2),

(x I> X2)(0, 0)
(XI> X2)

(0, 0),
-X2)

(Xl' X1)(I,

(Xl' Xl),

(-Xl>

= (0,0).
of

Figure 1.2

Proof The proofs here are immediate Theorems 1.29, 1.30, 1.32, and 1.33.

from the definition.

as are the proofs

TIIeonm 1.19. Given two complex numbers x = (x I' x2) and Y = (y h Y2). there exists a complex number z such that X + I = y. Infact, Z = (YI Xh Y2 X2)' This Z is denoted by y - x. The complex number (-x., -x2) is denoted by -x. TllloulIIlJO. /JejiIIitio1II.Jl.
X-l

For allY two complex numbers x and y, we Juwe

(-x)y
= x(-y)
(Xl>

= -(xy)

{-I, OXxy}.

[xl/(~

+ xi),

If x

x2)

-X2/(X;

+ xin.

(0,0) and y are complex "umbers, we define and yix = yx-1,

X

The idea of expressing complex numbers geometrically as points on a plane was formulated by Gauss in his dissertation in 1799 and, independently, by Argand in 1806. Gauss later coined the somewhat unfortunate phrase "complex number." Other geometric interpretations of complex numbers are possible. Instead of using points on a plane, we can use points on other surfaces. Riemann found the sphere particularly convenient for this purpose. Points ofthe sphere are projected from the North Pole onto the tangent plane at the South Pole and thus there corresponds to each point of the plane a definite point of the sphere, With the exception of the North Pole itself. each point of the sphere corresponds to exactly one point of the plane. This correspondence is called a stereographic projectwn. (See Fig. 1.3.)

TIteoUllt IJ1. If x and r are complex numbers with complex number z :ruch that xz y, namely, Z = yx-l•
Of special interest part is O. are operations with complex

#- (0,0), there exists a

whose imaginary

numbers

TIteoum 1.33.

(XI' 0)

+ (Yh

0)

(Xl

+ Y,. 0),
a},

(x" OXYIo 0) = (XtYI' 0),

(Xl'

O)/(y,> 0) =

(xt/r"

NOTE. It is evident from Theorem J.33 that we can perform arithmetic operations on complex numbers with zero imaginary part by performing the u~iial real-number operations on the real parts alone. Hence the complex numbers of the form (x, 0) have the same arithmetic properties as the real numbers. For this reason it is

Figure 1.3

i.
1.23 THE IMAGINARY UNIT

Del. 1.34

Del. 1.40

.,
[x

TlwDrem 1..39. If x and yare complex numbers. then we have

Xl) as a two-dimensional vector with components XI and X2' Adding two complex numbers by means of Definition 1.26 is then the same as adding two vectors component by component. The complex number I = (1,0) plays the same role as a unit vector in the horizontal direction. The analog of a unit vector in the vertical direction will now be introduced.

It is often convenient to think of the complex number (Xl'

+ yj

:::; Ixl

+ Iyl

(triangle inequality).

The proof is left as an exercise for the reader. 1.25 IMPOSSIBILiTY OF ORDERING THE COMPLEX NUMBERS

lkjbtititnll..3l.
inary Imit.

The complex number (a, I) is denoted by i and is called the imag-

1'IIeonm 1..35. Every complex number x = (Xl' x,) can be represented in the form

= XI

Proof.

+ ix2• XI = (Xl, 0), XI + iX2 = (X1o =-

ix,_ = (0, l)(x2'0) = (0. X2), 0) + (0. X2) = (XI> X2)"

The next theorem te1ls us that the complex number t provides us with a solution to the equation xl = -1. TllMnm 1.36. i Z Proof.
L

= (0, 1)(0. I) = (-1.0) = -I. VALUE OF A COMPLEX. NUMBER value to the complex number system.

1.24 ABSOLUIE

We now extend the concept of absolute

be the nonnegative real number Ixl gim!ll by

IhforitU»t 1.37.. If x =

(X .. X2), we define the modulus. or absolute value. of X to

I
,

As yet we have not defined a relation of the form x < y if x and y are arbitrary complex numbers, for the reason that it is impossible to give a definition of -c for complex numbers which will have all the properties in Axioms 6 through 8. To illustrate, suppose we were able to define an order relation < satisfying Axioms 6. 7, and 8. Then, since j ¢ 0, we mUlit have either i > 0 or i < 0, by Axiom 6. Let us assume i > O. Then taking, x = y = i in Axiom 8, we get ;2 > 0, or - I > O. Adding 1 to both sides (Axiom 7), we get 0 > L On the other hand, applying Axiom 8 to -1 > 0 we find I > O. Thus we have both 0 > 1 .and I > 0, which. by Axiom 6, is impossible. Hence the assumption i > 0 leads us to a contradiction. [Why was the inequality - 1 > 0 not already a contradiction '/] A similar argument shows that we cannot have j < O. Hence the complex numbers cannot be ordered in such a way that Axioms 6. 7, and 8 will be satisfied. 1.26 COMPLEX EXPONENTIALS The exponential e" (x real) was mentioned earlier. We now wish to define e" when z is a complex number in such a way that the principal properties of the real exponential function wiD be preserved. The main properties of e" for real are the law of exponents, trll!"l = e"1+;r1,and the equation e{) = I. We shall give a definition of e" for complex z which preserves these properties and reduces to the ordinary exponential when z is real. If we write z = x + iy (x, y real). then for the law of exponents to hold we want.l!"+ il = e"eiy• It remains, therefore. to define what we shall mean by eb•

Ixl 1honm1.38.
i) 1(0,0)1 iii)

Jxi +.G.
ii)

0, and Ixl > 0 i/x

+ O.

Ixyl = Ixllyl.
0)1= Ixd.

lx/yl

Ixi/lyl, if Y + O.

iv) I(xl>

e"

DefoUtiOll 1.40. If z = x = I!" (cos y + i sin y).

iy, we define e"

e"+b to be the complex number

Proof. Statements (i) and (iv) are immediate. To prove (ii), we write x = XI + ixz• Y = YI + i)'z. so that xy = X1Y1 - XIYZ + i(x1yz + xZYI)' Statement (ii)
follows from the relation Ixyl2 = Equation

This definition" agrees with the real exponential function when is, Y = 0). We prove next that the law of exponents still bolds.

is real (that

xry~ + x~~ + xiyZ + x~y:

= (xt

(iii) can be derived from (ii) by writing

+ x~XYf + yZ) = Ix111y12. it in the form Ixl = lyllxlJlI.

Geometrically, Ixl represents the length of the segment joining the origin to the point x. More generally, [x - yl is the distance between the points X and y. Using this geometric interpretation, the following theorem states that one side of a triangle is less than the sum of the other two sides.

• Several III'gUments can be gi\len to motivate the equation elF = cos y + j sin y. For example, let us write el~ = f(y) + ig(y) try to determine the real-valued functionsf and 9 so that the usual rules of operating with real exponentials will also apply to complex exponentials. Formal dlfferentiation yields e" = g'(y) - if'(y). if we assume that (et,y = te", Comparing the two expressions for .ly. we see that/and g must satisfy the equalions/(y) g'(y),f'(y) -g(y). Elimination of 9 yieldsf{y) -f'(y). Since we want eO I, we must havef(O) "" I and/,(O) O. It follows thatf(y) ... cos y and g(y) "" - f'(y) '" sin y. or course, this argument proves nothing, but it strongly suggests that the definition e'Y = COli Y + f-sin y is reasonable.

ana

'l'b. 1.41

Del. 1M

11

then we have

PrI)Qf,
of'

= e'"l(COS Y1 +

i sin Yl).

e'l

= e'"l{cos Y1

i sin Y2),

The two numbers r and I} uniquely determine z, Conversely, the positive number r is uniquely determined by z; in fact. r = [z]. However, z determines the angle 8 only up to multiples of 211:.There are infinitely many values of fJ which satisfy the equations x = [z ] cos 0, Y = [r] sin 0 but, of course, any two of them differ by some multiple of 211:.Each such (}is called an argument of z but one of these values is singled out and is called the principol argument of z, DejUritiQlf 1.46. Let z = x + iy be a nonzero complex number. number (} which satisfies the conditions

e'Jt!1.

e""c'[cos y, cos Y2

sin Y1 sin Yl sin YI eos Y2)]. are both reaL Also,

The unique real

+
Now e"'e" and

;(oos Yl sin Y2
Xl

+
.1£'2

= c' +"", since

= [z ]

cos 9,

[z] sin 6,

-n <

(l ~ +1t

and

is called the principal argument of z, denoted by 8

arg (z).

cos YI sin Y2 and hence

sin Y. cos Y2

sin (YI

+ h),

1.2.7 FURTHER PROPERTIES OF COMPLEX EXPONENTIALS In the following theorems, Tileorem 1.42.

I
."

The above discussion TMorem 1.47. Every

immediately

yields the following theorem: in the form useful in

reil, where r

complex [z] and (}

number z #- 0 can be represented arg (z) + 21tn, n being any integer.

NOTE.

This method of representing complex numbers is particularly connection with multiplication and division, since we have and

z,

Z,'

Z2

denote complex numbers.

e is never
eO

zero. be zero.

Proof, e'e:"

=
x

1. Hence e" cannot

Tlleorem 1.48. If
where

ZIZl

#- O. ,hen arg (zl.t:!.)

arg (Zl) (ZI) (ZI)

(%1)

arg (Z2) (Z2) (Z2)

S; S;

+ 21tn(ZJ'
+n,

Z2)'

Tileorem 1.43. If.1£' is real, (hen

lei" I

I.

Proof,

!eIX~2 =

C(lS2

+ sin2

J, and je;"1 > O.

if -71: < arg if -271: < arg if 1t < aeg

+ arg + arg

+ aeg (Z2)

S; -ll:.

2n.

Tluwnm 1.4". Ii'

Proof.

= I if, and

only if, z is an integral multiple of2ni.

then (2nn) I.

If z

= 2nin, where 11 is an integer,

e'

cos (2nn)

+ i sin

Conversely, suppose that e' = I. This means that e" C(lS Y = I and e'" sin y = O. Since e" #- 0, we must have sin Y = 0, y hr., where k is an integer. But cos- (kn) = (-] 1. Hence e" = (-1)"', since e'" cos (kJI) = l. Since e" > 0, k must be even. Therefore e" 1 and hence x O. This proves the theorem.

Proof, Write Z1 = Izde'41" Z2 = IZ2Ie"" where 61 = arg (ZI) and (;}l = arg (%2)' Then ZlZ2 = IZ1z11/(f,'HI... Since -1[ < 91 :s;; +n and -71: < 82 ~ +)'1:. we have -21t < 81 + 61 :s: 2n. Hence there is an integer n such that -n < fJ, + 61 + 2mr ~ n. This n is the same as the integer nCz., Z2) given in the theorem, and for this n we have arg (ZIZ2) = fll + 92 + 21111. This proves the theorem.
1.29 INTEGRAL POWERS AND ROOTS OF COMPLEX NUMBERS Defolitiml 1.49. Given a complex number z and an integer of z as follows ..
/1,

IMonm 1.45. Proof

c'

= e'" if, and only

if;

we define the nth power

Zl -

Zl ""

2nin (where n is an in/eger).

e'

e'" if, and only if, if'

-z, = I.

if n ~

o,
and n

1.28 THE ARGUMENT OF A COMPLEX NUMBER

if z #- 0
Theorem

> O.

If the point z
can

= (x, y) =

write x =r

iy is represented by polar coordinates rand 6, we cos 6 and y = r sin 6, so that z = r cos 6 + ir sin 8 = re'"

by mathematical

1.50, which states that the usual laws of exponents hold, can be proved induction. The proof is left as an exercise.

TiIetwem 1,s0.

Given two integers m and n, we have, lor z of; 0,

1
II

[)ef..l.55

13

and
Tkorem 1.5/. Ilz"'_ 0, and if n is a positive integer. then there are exactly distinct complex numbers zo, zl> ... , %._1 (eo/Jed the nth roots olz), $lien. that

TIu!orem 1.52. If z is a complex number #0, then rhf're exist complex numbers such that e" = 2. One such w is the complex number

'III

log Izi

+ i arg(z}, + 2mri,
z, But if

z:

z,

for each k

and any other such w must hove th« form

0, I, 2, ...•

II -

l. where n is an integer.

Furthermore,

these roots are given by the formulas where

log [z]
Proof. Since el"'1"1+''''''(2)

i arg (z)

fbI<
NOTE.

arg (z)
II

+ 2nk
11

(k = 0, 1.2 •...•

11 -

l~

log [s] + i arg (z) is a solution of the equation eW solution, then e" = e"" and hence w - WI = 2J17ri.

e",I<Ie'"rj:(%.

= Izle'· ...(%) = z, we see

WI

that 'III is any other

The n nth roots of z are equally spaced on the circle ofradius R center at the origin.
Proof.

DejiaitiOIt

!zll/.,

$lich that e"

1.53. Let z ::J:. 0 be a giLlen complex number. If w is a complex number = z, then w is called a logarithm of z, The partiCular !JOlueof w gjven
'III

by is caJ/ed the prindpallogarithm

The II complex; numbers Re~\ an nth root of z, since (Reo;~r = R"J"h

S k :5: 11 -

I, are distinct and each is

Jog [z]

+ iarg

(z)

of s, and for this w we write

'III

= Iz Ie'l.r I (If) + 20] = z.

EumpIes

We must now show that there are no other nth roots of z, Suppose w = Ae'" is a complex number such that IV' z. Then Iwlft = Irl. and hence A" = [z], A = Izl"-. Therefore. w" = z can be written el"" = eila'1l«)]. which implies
1IiX -

Log z.
=

1. Since 1. Since

Iii =
1- il

I and M8 (i) = 11;/2.Log (i)

1 and arg (-i)
=

iiefl. ... -[n/2.

Hi .

arg (z)

21£k

for some integer k,

= -11'/2. Log (-;0

=
11'.

3. Since i-It 5. Since

1 andarg(-I)
=

Log (-I)

==

Hence If = [arg (z) + 2l1'k]/n. But when k runs through all integral values, w takes only the distinct values zo •...• ZR-l' (See Fig. 1.4.)

... If x > 0, Log (x)

II + il

J-Z and arg (I

(ZIZ1)

log x. since

Ixl

= x and arg (x) ~ O. = 1f/4. Log' (1

+ i)

+ i)

'"

log"/i + i1t/4.
22)'

Tlleorem lSI. If

21Z2

# 0, then
= Log z,

Log

Log

%2

+' 2nin(zl>

whe~e n(Zl. Z2) is the integer defined in TheQrem Proof.

1.~8.

Log (zlzz)

= =

log log

IZlz21

IZll + log 1:21+ j [arg

+ t arg

(Z122)

(2t)

+ arg

(Z::)

2mr(zl' z~)].

Figure 1.4

•.3.

COMPLEX

POWERS

1.30 COMPLEX LOGARITHMS By Theorem 1.42. e" is never zero. It is natural to ask if there are other values that e" cannot assume. The next theorem shows that zero is the only exceptional value.

Using complex logarithms. we can now give a definition of complex powers of complex numbers.
Dejillition 1.55. If z '" 0 and ifw is any complex number, we define

111. 1.56

EumpJes 1. it = e1LofI =

etU'"/~) = e-'lt/2.
01"(1.)

2. (_1)1 = e'LotlI-1l =

= e-1t•

3. If /J is an integer. then z ·+! e"'+ ])LotI. = not conflict with Definition 1.49.

e"Lof'ros"

= :"z,

SO

Ddinition 1.55 docs

The next two theorems give rules for calcu1ating with complex powers: 7'INNHrIm 1.56. z .. , z ..,

, I
I

Def.l.Cil

b) III. c) so

e C, but z

# O. then z(CQ) = CQand z!O

= a»,

00 = (ooXaJ)

00.

DejUritiotJ 1.61. Every set in C olthi!form hood of 00, or a ball with center at 00.

{z: Iz1 > r ~ O} is called a neighbor-

zlP, ",.. ,

if z oF

O.

.1

I .
",

where n(xt, Pr(H)j.

X2)

is the iIItt!{Jertkfined in Theorem 1.48.

(%1%2)'"

ewlA>tI(x,.,)

e...' ......Zl+IAIZ>+Z

.. I«(.".>U.

1.31 COMPLEX SINES AND COSINJIS

lhfodtioll 1.58. Gwen a complex number z, we define oos

NOTE.
%

= iI' + e-b: _._-2'

sin z

tJz = --- - e-~

2i 1.40.

When z is real, these equations

I.

agree with Definition

TIttoran 1.59, If

+ tr: then we have

cos

The reader may wonder wby two symbols, +C(l and co, are adjoined to R but only one symbol, oo, is adjoined to C. The answer lies in the fact that there is an ordering relation -c among the real numbers, but no such relation occurs among the complex. numbers. In order that certain properties of real numbers involving the relation < hold without exception, we need two symbols, + 00 and - 00, as defined above. We have already mentioned that in R* every nonempty set has a sup, for example. In C it turns out to be more convenient to have just one ideal point. By way of illustration, let us recall the stereographic projection which establishes a oneto-one correspondence between the points of the complex plane and those points on the surface of the sphere distinct from the North Pole. The apparent exception at the North Pole can be removed by regarding it as the geometric representative of the ideal point 00. We then get a one-to-one correspondence between the extended. complex plane C* and the total surface of the sphere. It is geometrically evident that if the South Pole is placed on the origin of the complex plane, the exterior of a "large" circle in the plane will correspond. by stereographic projection, to a "small" spherical cap about the North Pole. This illustrates vividly why we have defined a neighborhood of CQ by an inequality of the form [z 1 > r.
=

cos z

cosh)' y

Pr(H)/
2cosz

sin z
=

= sin x cosh + e-iz

i sin x sinh y, j cos x sinh y.

EXERCISES Irrtegers 1.1 Prove tbat there is no largest prime. (A proof was known to Euclid.) 1.2 If n is a positive integer, prove the algebraic identity

= e-'(oos x + i sin x) + eY(cos x - i sin x) = cos x(e' + e-J') - t sin x(eY - e-,) = 2 cos x cosh Y 21 sin x sinh y. is similar.

eiz

The proof for sin Further

I.

properties

of sines and cosines are given in the exercises,

1.33 INFINITY AND TIlE EXTENDED COMPLEX PLANE C. Next we extend the complex number system by adjoining the symbol CQ. an ideal point denoted by

1.3 If 2" I is prime, prove that n is prime. A prime of the form 2" - I, where p is prime, is called a Mersenne prime. 1.4 If 2~ + 1 is prime, prove that n is a power of 2. A prime of the form 22m called a Fermat prime. Him. Use Exercise 1.2.
Xw 1

I is

IhjillititJII1.60. By the extended complex number system C· we slwll mean the complex plane C along with a zymbol 00 which satisfies the folloWing properties: a) If z
E

1.5 The Fibonacci "umbers I. I. 2. 3, 5, 8, 13, ... are defined by the recursion formula + = X" + X~-h with x] = Xl = 1. Prove that (xn' X~+l) = I and that x" = (a" - b-)/(a - b}, where a and b are the roots of the quadratic equation x2 - x-I = O. 1.6 Prove that every nonempty set of positive integers contains a smallest member. This is called the welf-orde,ing principle.

C, then we halJe Z

(J)

z-

<::IJ =

co, z/oo

O.

26 Ratioml and irrational IIlIDIben 1.7 Find the rational number whose decimal expansion is O.3~44444 ... 1.8 Prove that the decimal expansion of x will end in zeros (or in nines) if. and only if, x is a rational number whose denominator is of the form 2"5"', where m and n are nonnegative integers. 1.9 Prove that 1.22 Given x :> 0 and an integer k 2: 2. Let QI) denote the largest integer ~x and, assuming that alb a1, ... , 0._1 have been defined, let Q" denote the largest integer such that al 01 a Qo + - + -:; + ... + ....!! ,.; x. k k" k" a) Prove that 0 ~ Q, ~ k - 1 for each i = 1, 2, ... b) Let r,. = ao + alk-I + il2k-2 + ... + a,/(.-· and show that x is the sup of the set of rational numbers '1' '2.' •
NOTE.

"'/2 + .f3 is irrational.

:>

1.10 If a. b. c, d are rational and if x is irrational, prove that (ax irrational. When do exceptions occur? 1.11 Given any real x 1.11 If alb < c:/d with b > 0, d andc/d.
:>

h)j(cx

d) is usually

0, prove that tfiere is an irrational number between 0 and x. 0, prove that

«(I + c)/(h +

d) lies between alb

When k = 10 the integers a{h a,. a2 •... are the digits in a decimal representation of x. For general k they provide a representation in the scale of k .

1.13 Let a and b be positive integers. Prove that "./2 always lies between the two fractions alb and (a + 2b)/(a + b). Which fraction is closer to ·'/i.1 1.14 Prove that

Inequalities 1.23 Prove Lagrange's identity for real numbers:

.J n

-I

Jn +

1is irrational for every integer n

2:

I. Note that this identity implies the Cauchy-Schwarz 1.24 Prove that for arbitrary real
01.,

1.tS Given a real x and an integer N:> I. prove that there exist integers h and k with o < k: ~ N such that Ih - hi < lIN. Hint. Consider the N + 1 numbers tx - [tx] for t 0, 1,2, ... , N and show that some pair differs by at most lIN. l.t6 If x is irrational prove that there are infinitely many rational numbers hlk with k :> 0 such thai Ix - h/k1 < 1/k2• Hint. Assume there are only a finIte number h,/kl' ... , h.lk. and obtain a contradiction by applying Exercise 1.15 with N:> Ifo, where Ii is the smallest of the numbers Ix - hdkll. 1.11 Let x be a positive rational number of the form x

ls.k<.J~JI.

(Utbj -

ajblt)".

inequality.

b., c~ we have

1.25 Prove Minkowski's inequality:

l=1

• :E~, k!

(t
•••

(Ok

bt}2)

10 s (t; air! + (t b~r2

n-dimensional vectors. where
= (~

This
II ~

is the
(a1o
•••

where each Uk is a nonnegative integer with a.. ~ k - I for k 2: 2 and a" > O. Let [x] denote the greatest integer in x. Prove that UI "" [x], that ll" = [k! x] - k[(k - J)! x] for k = 2, ... , n, and that n is the smallest integer such that n! x is an integer. Conversely, show that every positive rational number x can be expressed in this form in one and only one way.
Uppe£

triangle inequality , Dn), b = (hl>""

[a + hll s ~all + Ijbll for

b.) and

lIall
1.26 If al
~

a~)
2:

112

02

;::

an and hi

:2:

b2

:2: •••

h", prove that

bound!>

1.18 Show that the sup and inf of a set are uniquely determined whenever they exist. 1.1' Find the sup and inf of each of the following sets of real numbers: a) All numbers of the form 2 -p + 3 - .. integer values. b) S e) S
=>

5 "", where p, q, and r take on all positive

{x: 3x2

lOx
a)(x

3 < OJ.
c)(x d) < OJ,

Complex nlUl1bers 1.27 Express the following complex numbers in the form a
a) (1 c)
j5

{x : (x

b)(x

where

<

b<c

<

d.

+
+

i)J,
;16,

b) (2
d)

1.20 Prove the comparison property for suprema (Theorem 1.16). 1.21 Let A and B be two sets of positive numbers bounded above, and let a = sup A, b "" sup B. Let C be the set of all products of the form xy, where x E A and Y E B. Prove that ab = sup C.

+ bi. + 3;)J(3 - 4i), W + n(l + i-8).

c)

1.28 In each case, determine all real x and y which satisfy the given relation. a) x

+ iy

Ix

iy[,

b) x

iy

(.y _ iy)',

l~O

:E i"

100

iy,

Real and Complex

NIIIIIber Systems

1.19 If z x + tr, x and i = x - if. Prove that:

a)
%1 Z

real, the complex conjugate of z is the complex number

b) Z1Z2
I,

1.36 Define the following "pseudo-ordcring"

if we have either

of the complex numbers: we say

ZI <:

%2,

d)

+ Z2 +z=

il

zz.

iliJ"

or
Which of Axioms 6, 1, 8, 9 are satisfied by this relation 1 1.37 Which of Axioms 6, 7, 8,9 are satisfied if the pseudo-ordering is defined as follows i We say (XI' Y1) < (x2, fz) if we have either

twice the real part of

e) (x - i)f; = twice the imaginary part of z. 1.39 Describe geometrically the set of complex numbers z which satisfies each of the following: conditions:

a) tzl ~
d) z

I, I,

b)
e)

Ixl
z-

< 1,

c) Izl ~
f) I

1,
= Iz12,

or
1.38 State and prove a theorem analogous to Theorem 1.48, expressing arg (zl/I2) in terms of arg (z 1) and ars (Z2)' 1..39 State and prove II. theorem analogous to Theorem 1.54, expressing Log (ZI/Z2) in terms of Log (Il) and Log (Z2)' 1.40 Prove that the nth roots of 1 (also called the nth roots of unity) are given by «, 1lf2. ••••• fi', where Of = ,,2.1/., and show that the roots # 1 satisfy the equation
t

i""

= i,

1.31 Given three complex numbers II' Z2. Z;3 such that IZII = Izzl = IZ31 = I and ZI + Zz + Z;l, = O. Show that these numbers are vertices of an equilateral triangle inscribed in the unit circle with center at the origin. 1.32: If
II

and b are complex numbers, prove that:

a) la -

W s (I + lal2)(1 + IW)·
lal + Ibl if, and
111 - bl
=

b} If 11 # 0, then 111+ hi =

only if, b/11 is real and nonnegative.

;(2

+ ... +

x"-1

o.
leas z I
..... UlJ.

1.33 If II and b are complex numbers. prove that

II - 001

1.41 a) Prove that Izil -c e« for all complex z l' O. b) Prove that there is no constant M > 0 such tbat

< M for all complex z.

if, and only if, vatid?

lal =

I or

Ibl

= 1. For which a and b is the inequality

la - bl < II -

libl

1.41 If w =

II

+ iv (u, v real), show that

ZW

= e"1oI1o:1-"..'I(.lel[

• ..,.I.I+

1.34 If a and c are real constants, b complex, show that the equation

1.43 a} Prove that Log (ZW) "" w Log I b) Prove that (z",)", ~ z-e2" 1.44
i) If (J and
Q

mill, where

11

is an integer. that
j

azi

+ bZ + liz + c

=0

(a #- 0, Z ... x

iy)

.....

where n is an integer.

represents a circle in the .o-plane, 1.35 Recall the definition of the inverse tangent: given a real number t, tan-1 (t) is the unique real number 0 which satisfies the two conditions tan 0 = t, If I
"'" X

are real numbers, - n < 8 ~ (cos 0

+ It. prove

+ i sin Of

-= cos (aU)

sin (00),

Show that, in general, the restriction -7t < ()~ +n i$ necessary in (i) by taking ()= -Jr, a = i. iii) If a is an integer, show thai the fonnula in (i) holds without any restriction on 0, In this case it is known as DeMoivre's theorem,
ji)

iy, show that

tan1 (~) ,

1.45 Use DeMoivre's theorem (Exercise 1.44) to derive the trigonometric identities sin 3(} ... 3 eas2 0 sin () - sin:! fJ, if x>
0,

a) arg (z)

cos 30
2:

cos! {}~ 3 cos (}sin:.!

e,
+
iy, we have

valid for real fJ. Are these valid when fJ is complex i

b) arg (z)

tan-

(~)

11",

if

x < 0, y

0,

1.46 Define tan z = (sin z)/(cos z) and show that for z ~ x

tanz=

c) arg (z)

tan-

(~)

Ir,

ifx<O,y<O,

sin 2x + i sinh 2y • cos 2x + cosh 2y

d) arg (z)

= ~ if x =
2

0, y > 0; arg [z]

=-

'J! if 2

x = 0, y < O.

z=x

1.47 Lei w be a given complex number. If w # ±], show that there exist two values of + iy satisfying the conditions cos z = wand - II: < X S + ft. Find these values when w = iand when w "" 2.

Ileal BDd C(IDIIIIn

NIIIIIber

Systems

31

1.41 Prove Lagrange's identity for complex numbers:

I~

I'l.tb.tr ... i110tll

i;;
(J -

].6 Hobson, E. W., The Theoryof Functions of a Real Variable and the Theory of Fourier's Series, Vol. I, 3rd ed, Cambr-idge University Press, 1927.

Ib,J2

l:<6tSD

Ia.Oj -

I'lJ~12.

1.7 Landau, E., FOlllfdiJtionsof Analysis, 2nd ed, Chelsea, New York, 1960. 1.8 Robinson, A.• N(}n-tllandard Analysis. North-HoUand. Amsterdam, 1966. 1.9 Thurston, H. A., The Number System. Blackie, London, 1956.

Use this to deduce a Cauchy-Schwarz inequality for complex numbers. ].49 a) By equating ilrulginary parts in DeMolvre's formula prove that sin nO b} If 0 <
(J =

sin~ 8 {(;)

cot"-1

(;)

COt'-3 8

+ (;)

cot"-5 8 -

+ ... } .

1.10 Wilder, R. L., Introduction to the Foundation;; of Mathematics, 2nd ed, Wiley, New York,I965.

< n/2, prove that sin (2m

1)0 = siolm+18P ...(co(l 8)

where P,. is the polynomial of degree m given by P"J.x) =

em 1+ I)

x'" -

em 3+ 1)
7ft _ 2m

x"'-l

em 5+ 1)

x"'-3 -

+ ... +
I)}

Use this to show that P", has zeros at them distinct points x; = cot2 {n:k/(2m for k = I, 2, ... , m. c) Show that the sum of the zeros of Pm is given by

i: cor' _
+
1

knl

= m(2m - 1) , 3

and that the sum of their- squares is given by

'" k L;cot4_R_1=1 NOTE.

m(2m -

1)(4m1

+
1l 2

10m -

9)

2m

45 j6andI::'
I

These identities can be used to prove thatL~l n-2 (See Exercises 8.46 and 8.47.)

n-4

...

1f4./90.

1.50 Prove that z· - I = formula

11;=1 (2
0-1

e2~ilf") for all complex z,

Use this to derive the

k~1

. SIU-

kt: n

n = -2"-1

for- n

2::

2.

SUGGESlED

REFERENCES

FOR FURTHER STUDY

1961.

1.1 Apostol, T. M., Calculus, Vol. I, 2nd ed. Xerox, Waltham,

1.2 Birkhoff, G .• and MacLane, S., A Survey of Modern. Algebra, 3rd ed. Macmillan, New York, 1965. 1.3 Cohen, L., and Ehrlich, G., The Structure of the Real-Number System. trand,.Princeton, 1%3. 1.4 Gleason, A., Fundamentals of Abstract Analysis. Addison-Wesley, Van Nos1966,

Reading,

1.5 Hardy, G. H., A Course of Pure Math"malics, 10th ed. Cambridge University Press, 1952.

Dd'.l.l

CHAPTER 2

SOME BASIC NOTIONS OF SET THEORY

11 INTRODUcnON

In discussing any branch of mathematics it is helpful to use the notation and terminology of set theory. This subject, which was developed by Doole and Cantor in the latter part of the 19th century. has had a profound influence on the development of mathematics in the 20th century. It has unified many seemingly disconnected ideas and has helped reduce many mathematical concepts to their logical foundations in an elegant and systematic way. We shan not attempt a systematic treatment of the theory of sets but shall confine ourselves to a discussion of some of the more basic concepts. The reader who wishes to explore the subject further can consult the references at the end of this chapter. A collection of objects viewed as a single entity will be referred to as a set. The objects in the collection will be called elements or members of the set. and they will be said to belong to or to be ccmtai"ed in the set. The set. in tum, will be said to contain or to be composed of its elements. For the most part we shall be interested in sets of mathematical objects; that is, sets of numbers, points, functions, curves, etc. However, since much of the theory of sets does not depend on the nature of the individual objects in the collection, we gain a great economy of thought by discussing sets whose elements may be objects of any kind. It is because of this quality of generality that the theory of sets has had such a strong effect in furthering tbe development of mathematics.
l.2 NOTADQNS

I
I 4

by 4, namely. {4, 8}, is-a subset of the set of even integen less than 10. In general, we say that a set A is a subset of B, and we write A !;;; B whenever every element of .A also belongs to B. The statement A. 6 B does not rule out the possibility that B E A. In fact, we have both A. £; Band B £: A if, and only if, A and B have the same elements. In this case we shall call the sets A. and B equal and we write A = B. If A and B are not equal, we write A ,p B. H A £: B but A. "" B. then we say that A is a proper subset of B. It is convenient to consider the possibility of a set which contains no elements whatever; this set is called the empty set and we agree to call it a subset of every set. The reader may find it helpful to picture a set as a box containing certain objects. its elements. The empty set is then an empty box. We denote tJiC empty set by the symbol 0.

2.J ORDERED PAIRS

Suppose we have a set consisting of two elements a and b; that is, the set {a. b}. By our definition of equality this set is the same as the set o}, since no question of order is involved. However, it is also necessary to consider sets of two elements in which order is important. For example. tn analytic geometry of the plane, the coordinates (x, y) of a point represent an ordered pair of numbers. The point (3, 4) is different from the point (4. 1), whereas the set {3. 4} is the same as the set {4, 3}. When we wish to consider a set of two elements a and b as being ordered, we shall enclose the elements in parentheses: (a, b). Then a is called the :firstelement and b the second. It is possible to give a purely set-theoretic definition of the concept of an ordered pair of objects (a, b). One such definition is tbe following:

rh.

lkforititM2.1.

(a, h)

{{a}, {a, b}}.

This definition states that (a, b) is a set containing two elements, {a} and {a, b}. Using this definition, we can prove the following theorem: Tlteore",2.2. (a, b)

(c, d) if, and only if, a = c and b = d.

Sets will usually be denoted by capital letters:

A, B, C, ...• X. Y, Z,

This theorem shows that Definition 2.1 is a "reasonable" definition of an ordered pair, in the sense that the object a has been distinguished from the object h. The proof of Theorem 2.2 will be an instructive exercise for the reader. (See Exercise 2.1.)

and elements by lower-case letters: 0, b, c•... , x, y, z, We write XES to mean "x is an element of S," or "x belongs to S." If x does not belong to S, we write x f S. We sometimes designate sets by displaying the elements in braces; for example, the set of positive even integers less than 10 is denoted by {2, 4, 6, 8}. If S is the collection of all x which satisfy a property P, we indicate this briefly by writing S = {x: x satisfies Pl. From a given set we can form new sets, called subsets of the given set. For example, the set consisting of all positive integers less than 10 which are divisible
32

2.4 CARTESIAN PRODUCT OF TWO SETS Ihjinitioll 13. Given two sets A and B. the set of all ordered pairs (a. b) such that o E .A and b e B is cQlled the cartesian product of A and B. and is denoted by A x B. &ample. If R denotes the set of aU real numbers, then R x R is the set of aU complex

numbers.

Th.2.6

tuncttons F and fi are

if

andG
every

=
numbers, Such a set of ordered.

1,

2.6 FURTHER TERMINOLOGY

CONCERNING

FUNCTIONS

expressions defines a certain set of ordered the set of all J') for which the

relation. The OO!teS·POII.dirll!: of the remtion, are shown in

2.1.

('"~nnh>,, number

then F .is called a one real system, then Fis called a

iii. £;>M~f;o,~

COlJ'1J'Jl,?X variable. is a subset of a cartesian A B, then F is called two variables. In this case we denote the function values

A fUllction of two real variables

is one whose domain

is a subset of

If S is a subset of we say that F is on S. such that ES caned t.he under F and T is any set which contains then F is also called a rnajt>tJil~(J This is often denoted F:51-+ T. of If The oorlcel~ of relation can be formulated need not be numbers hut may be is called a relation. IfS 1s a the set of all elements x that occur as first members of y) in S is caned the domain denoted The set of second members y is called. the ral1(jeof S, denoted The first shown in kind of relation know.n as a Dejinititm 2.1. A ,un", m", the member. That is, The definition of function that for every x in the dormdn of one y such that Y)E F. ltis to call y the value to write y= x ?md
If two functions F and G the inclusion relat.ion G <;;;; F; we say that G is a restriction of F or that F is an extension of G. In particular, jf S is a subset of and if G is defined the

of any kind.

Consider, for example, the function of a variable defined the eqaa= zZ. This function maps every sector S of the form 0 ::;;. rg a ::;; of the onto a sector described the ::;; ::;;arg

instead E F to indicate that the As an alternative to a function it coetams, to describe the domain of F, and for each x in the """en,",,,, to describe how the function value ill obtained. In this COlrme)cIIOIJ, we have the theorem whose is left as an exercise for the reader,

for all x in S,
then we call G the restriction

such that x

of F to 51. The function G consist" of those S. Its domain is S and its range is

l6

Def.1..7

Def.1.U

1.7 ONE-TO-ONE FUNCI'IONS AND INVERSES

77It'orem 1.10. If the function F il Orte-to-one on its domain, then

Oil

F is also afunction.

2.7. Let F he a function tkfored on S. We say F is one-to-one and only if. for every x and Y in S,
Jhjiltitiln; F(x)

S if.

Proof. To show that lisa function, we must show that if {x, J) E Fand (x, z) e F, then y = z: But (x, y) E i means that (y, x) E F; that is. x = F(y). Similarly,
ex, z) e f means that x = F(z). Thus F(y) = F(z) and, since we are assuming that F is one-to-one, this implies y = z: Hence, is a function.

F(y)

implies

x = y.

This is the same as saying that It function which is one-to-one on S assigns distinct function values to distinct members of S. Such functions are also called injective. They are important because. as we shall presently see. they possess inverses. However, before stating the definition of the inverse of a function, it is convenient to introduce a more general notion, that of the converse of a relation. lhjUJitimr 2.8. Given a rewtion S, t~ new relation

NOTE.

then the restriction

The same argument shows that if F is of F to S has an inverse.

one-to-one

on a subset S of 5j{F),

S defined

:l.B COMPOSITE FUNCTIONS

by

S=
is called the converse of S.

{(a, b) : (b, a) E S}

lkfotitio. 2.11. Givm two functions F and G such that !1/(F) S;; ~(O). we CWl form a n.ew junction, the camposite G " F of G and F, defined as follows: fOT every x in the domain of F. (G" F)(x) = G[F(x)].
Since ut(F) S;; !i(O), tbe element F(x) is in the domain of G. and therefore it makes sense to consider G[F(x)]. In general, it is not true that G" F = F., G. In fact. FoG may be meaningless unless the range of G is contained in the domain of F. However. the associative law,

Thus an ordered pair (a, b) belongs to if•.and only if, the pair (b, a), with elements interchanged, belongs to S. When S is a plane relatlan, this simply means that the graph of S is the reflection of the graph of S with respect to the line y x. In the relation defined by x < J. the converse relation is defined by y < x.
JhjUrJtkm 2.9. Suppose that the relation F is a function. Consider the conuerse relation i, which maJ or may not be a function. If F is also a function. then F is called the inverle of F and II tknoted by F-1•

always holds whenever each side of the equation has a meaning. be an interesting exercise for the reader, See Exercise 2.4.)

(Verification

will

Figure 2.3(a) illustrates an example ofa function Fforwhlch iis not a function. In Fig. 2.3(b) both F and its converse are functions. The next theorem tells us that a function which is one-to-one on its domain always has an inverse.

2.9 SEQUENCES

, . ,
I

Among the important integers.

examples

of fUDctions are those defined on subsets of the F

Ihjillition 2.12. By a finite sequence of" terms we shall under;stand a function whose domain is the set of numbers {I. 2, ... , n}.

The range of F is the set {F(l), F(2), F(3), ...• F(n)}. customarily written {FI' F2• F3 •••• ,FH}· The elements of the range are called terms of the sequence and, of course, they may be arbitrary objects of any kind. Definition 2.13.

By an infinite sequence we shall mean a function F whose dOfHllin is the set {I, 2, 3, } of all positive integers. The range of F. that is, the set {F(I), F(2), F(3), }, is also written {FlO FI, F3, ••• }. and the function value F.
is called the nth term of the sequence.

(al

(b)

For brevity, we &baH occasionally use the notation {F,,} to denote the infinite sequence whose nth term is F Let s = {s,,} be an infinite sequence, and let k be a function whose domain is the set of positive integers and whose range is a subset of the positive integers.
II •

Del.. 2.l4

I1L. 117.

Assume that k is "order-preserving," that is, assume that

k(m)

2.12 COUNI'ABLE AND UNCOUNTABLE SETS

< k(I'I),

ifm < II.

Then the composite function so k is defined for all integers n ~ I, and for every such n we have
(s
0

A set S is said to be countably infinite if it is equinumerous with the set of all positive integers; that is, if S'"" {I. 2. 3, ... }. In this case there is a function f which establishes a one-to-one correspondence between the positive integers and the elements of S; hence the set S can be displayed as follows: S = (f(I)./(2)./(3), ... }. Often we use subscripts and denote f(k) by al (or by a similar notation) and we write S = {o., 02' OJ •••• }. The important thing here is that the correspondence enables us to use the positive integers as "labels" for the clements of S. A countably infinite set is said to have cardinal number ~(l (read; aleph nought). 1h./htitioII2.15. A set S is ca//ed countable if it is either finite or countably infinite. A set which is not countable is called uncountable. The words denumerable and nomlenumerable are sometimes used in place of countable and "ncountable. 1'Iworem 2.16. Every subset of a countable set is countable. Proof. Let S be the given countable set and assume A ~ S. If A is finite, there i~ nothing to prove, so we can assume that A is infinite (which means S is also infinite). Let s = {s,,} be an infinite sequence of distinct terms such that S

k)(n) = s.ll;(,,)'

Sw::h a composite function is said to be a subsequence of s. Again. for brevity, we often use the notation {s.(.)} or {jt~}to denote the subsequence of {s.} whose nth term is st(It).
ED.,.... Let s
= {lIn}

andlet k be defined by ken) = 2". Then s» k

=;

{l/2"}.

2.tO SIMILAR (EQU1NUMEROUS)

SETS

Ikjiaitiorll.U. 1Wo sets A and Ii are called similar, or equinumerous, and we write A ~ B, if and only if there exist!J a one-tO-OlUl!unction F whose domain is the set A and whose range is the set B. We also say that F establishes a one-to-one correspondence between the sets A and B. Clearly, every set A is similar to itself (take Fto be the "identity" function for which F(x) = x for all x in A). Furthermore. if A ,.., B then B ,..,A, because if F is a one-to-one function which makes A similar to B. then F - 1 will make B similar to A. Also, if A -- B and if B ,.., C, then A ~ c. (The proof is left to the reader.)

{Sl>

3'1..' •. }.

1.11 FINlTE AND INFINITE SETS

A set S is called finite and is said to contain n elements if S - {I. 2, ... , n}. The integer n is called the cardinal number of S. It is an easy exercise to prove that if {I, 2, ... , n} - {l, 2, ... , m} then m = lJ. Therefore, the cardinal number of a finite set is well defined. The empty set is also considered finite. Its cardinal number is defined to be O. Sets which are not finite are called infinite sets. The chief difference between the two is that an infinite set must be similar to some proper subset of itself. whereas a finite set cannot be similar to any proper subset of itself. (See Exercise 2.13.) For example, the set Z+ of'all positive integers is simi1ar to the proper subset {2, 4, 8, 16, ... } consisting of powers of 2. The one-to-one function F which makes them similar is defined by F(x) = 2"' for each x in Z + •

Define a function on the positive integers as follows: Let k(l) be the smallest positive integer m such that s'" e A. Assuming that k(l), k(2), ... , ken - I) have been defined, let k(1'I) be the smallest positive integer m > ken - I) such that 8.. EA. Then k is order-preserving: m > n implies k(m) > k(I'l). Form the composite function II" k. The domain of II" k is the set of positive integers and the range of s e k is A. Furthermore. s k is oneto-one, since s[k(n)] = s[k(m)]. implies
0

which implies k(n) = k(m), and this implies n

m. This proves the theorem.

2.13 UNCOUNTABILITY OF THE REAL NUMBER SYSTEM

The next theorem shows that there are infinite sets which are not countable. Theorem 1.17. The set of all real numbers is uncountable.

'I1I.1..1

Del. l.U

41

Proof It suffices to show that the set of x satisfying 0 < x < 1 is uncountable. If the real numbers in this interval were countable, there would be a sequence

s = {sIll whose terms would constitute the whole interval. We shall &howthat this is impossible by constructing, in the interval, a real number which is Dot a term of this sequence. Write each s. as an infinite decimal:

The definition of union can be extended to any finite or infinite collection of sets:
DejinitioIl2.20. IJ F is an arbitrary collection of sets, then the union of all the sets in F is defined to be the set of those elements which belong to at least one of the sets in F, and is denoted by

when: each expansion where

U..,I

is O. I, ' .. .or 9. Consider the real number y which has the decimal

If F is a finite collection of sets, F

~=

"

{1. 2,

A",

U A = U A. =
1-1

..

{At, ••• , A,,}, we write

U'"

AI u A:a
{AI> Ai

u A•.
we write

ifll if It

oj:.

I,

If Fis a countable collection, F

<J)

•... },

I.

Then no term of the sequence {s.} can be equal to y, since y differs from 81 in the first decimal place, differs from 82 in the second decimal place, ... , from 6" in the nth decimal place. (A situation like s. = 0.1999 ... and y = 0.2000 ... cannot occur here because of the way the v. are chosen.) Since 0 < y < I, the

""I'

U A = U A" =
1:-1

A 1 U A.z v ...

theorem is proved.
product Z.. x Z + is countable.

Dejilfitioll 2.11. If F is an arbitrary collection of sets, the intersection of all sets in F is defined to be the set of those elements which belong to everyone of lhe sets in F, and is denoted by
The intersection of two sets A 1 and A 2 is denoted by At n A 2 and consists of those clements common to both sets. If A I and A 2 have no elements in common. then Al n Al is the empty set and At and Al are said to be disjoint. If F is a finite collection (as above), we write

ntorem 2.18. Let Z+ denote rhe set

of all positive integers.

Then the cartesilm

PrtJ()j. Define a functionJon

Z+ x Z+ as follows: if(m. n) e Z+ x Z+.

f(m, n) = 2"3·,

Then f is one-to-one on Z'" x Z+ and the range off is a subset of Z ", and if F is a countable collection, we write
2.14 SET ALGEBRA

Given tWo sets Al and Az, we define a new set, called the union of Al and AI, denoted by Al U A2, as follows:
/JejiJritio1l2J9. The UIIUm Al either to A I or to A 2 or to both.
U

Il.fiF

n A = n A.
ee

l=1

At

f"'1

A2

f"'1 ...

A2 is the set of those elements

which belo1l!J

If the sets in the collection have no elements in common. their intersection is the empty set. Our definitions of union and intersection apply. of course, even when F is not countable. Because of the way we have defined unions and intersections, the commutative and associative laws are automatically satisfied.
Dejildlitm 2.22. The complement to be the set of A relative to B, denoted by B - A, is defined

This is the same as sayiog that Al U A 1consists of those elements which belong to at least one ofthe sets AI' A:a. Since there is 00 question of order involved in this definition, the union Al u Ali is the same as A2 u A J; that is, set addition is commutative. The definition is also phrased in such a way that set addition is

{x:xeB,butxjA}.

associative :

Note that B - (D - A) = A whenever A s;;_ B. Also note that B - A = B if B n A is empty. The notions of union, intersection, and complement are illustrated in Fig. 2.4.

43

B-

Theorem 1.21. Let F and

a collection

have E:xamplel,

countable collection The set Q of aU rational numbers is a countable ret. numbers having A.. From this

U 11 then s: E but it not true that x to at least one in therefore x to no A In F. for every A in F% x E B A. But this x 1~ SCI that J1 - S s;;; T. the we obtain T s;;; J1 S, and this proves that J1 - S = To prove the second statement, use a similar argument.
1.15 COUNTABLE COLLECTIONS OF CnUNTABI,E SETS

let A. denote tile ret positive The set all positive rational equal to Q eeentable, since each A~ is countable.

, , ) denote the set of rational numbers and leI. A. be the set endpoint and whose right endpoint rational, Then A. IS countable and S =
I

Ak•

Definition 2•. 24. is a collection M""v,m. then .F is said

EXF..RCiSES :U Prove Theorem ':t2. appeal the definition 2.2 Let Sbe a let

d)

fa, S is said
be

A relation
am:~~_!.

is the set nfnn,~itiv" fore S-

Determine

reflexive, and which of these properties possessed numbers (x, such thai
.1'2
f,n"·· e s·""~

Theorem 1.26,

G=

collection

sets, let
F
0

y.

each case where

4ft formula

composite function G F call

(or formulas) for

Then G is a collection

44

Find F(x) if G(x) and G[F(x)] are given as follows: d) G(x) == x e) G{x)
3 ,

e) For all subsets A and B of S with B ~ A, we have = x' - 3x

= XZ
2

G[F(x)]

3x -

I.
2.11 If

I(A - B)

= /(A)

- [(B).

+ x + x2,

G[F(x»

+ S.

2.10 Prove that if A .. B and B - C, then A-C.

2.4 Given three functions F, G. H. what restrictions must be placed on their domains so that the following four composite functions can be defined? HoG, Assuming that H
a

fl,

2, ... , n} ~ {1, 2, ... , m}, prove that m =

If.

H» (Go F),

(HoG)·

F.

2.11 If S is an infinite set, prove that S contains a eountably infinite subset. Hint. Choose an element (II in S and consider S - (til). 2.13 Prove that every infinite set S contains a proper subset similar to S. 2.14 If A is a countable set and B an uncountable set, prove that B - A is similar to B. 2.15 A real number is called algebraic If it is a root of an algebraic equation fex-) = 0, where [(x) = «0 + alx + ... + a,.;i" is a polynomial with integer coefficients. Prove that the set of all polynomials with integer coefficients is countable and deduce that the set of algebraic numbers is also countable. 2.16 Let S be a finite set consisting of n elements and let Tbe the collection of all subsets of S. Show that T is a finite set and find the number of clements in T. 2.17 Let It denote the set of real numbers and let S denote the set of all real-valued functions whose domain is It. Sbow that S and It are not equinumerous. Hint. Assume S - R and let / be a one-to-one function such that /(R) S. If« E R. let D. ,.. [(a) be the real-valued function in S which corresponds to the real number a. Now define h by the equation h(x) = I + nx(x) if x E R, and show that h f S. 2.18 Let S be the collection of all sequences whose terms arc the integers 0 and 1. Show that S is uncountable. 2.1!' Show that the following sets are countable: a} the set 'Of circles in the complex plane having rational radii and centers with rational coordinates, b} any oollection

(G

F) and (H

G) e F can be defined, prove the associative law:

2.5 Prove the following set-theoretic identities for union and intersection: a) A
b) A
V (B

u C)

=>

(A u B) u C. (A

A rv (B n C)

(A n.B) n C.

(B V C)

n n

B) V (A

C).

c) (A

V B) "

(A u C) = A v (B n C). (C v A) "" (A (A n B)

d) (A V B) " (B V C) f) (A 2.6 Let C) f'I (B

n!1)
C.

V (A

C)

(B

C).

e) A n (B - C) ... (A n B) - (A n C).
C) g) (A - B) u B

A if, and only if, B ~ A.

I:

S.... T be a function,
f(A

If A and B are arbitrary subsets of S, prove that and

[(A f'I B) ~ /(A) n /(B).

V B) = [(A) V [(B)

Generalize to arbitrary unions and intersections. 2.7 Let [ : S .... Tbe a function. If Y ~ T, we denote by f-I(Y) the largest subset of S which / maps into Y. That is, .
f-l(y)
= {x: XES

and [(x) co Y}.

ot disjoint

intervals of positive length.

The set I-I (Y) is called the inverse inwge of Y under /. Prove the following for arbitrary subsets X of Sand Yof T.
a)X~rl[J(X)],

2.20 Let [be a real-valued (unction defined for every x in the interval 0 :;;; x s L Suppose there is a positive number M having the following property: for every choice of a finite number of points XI' xl •.••• .:rftin the interval 0 :5 x :5 I. the sum
I/(x!)

b)J[r'(Y)]~ v /-'(Y;z),
Y2) [-l(l"l) n[-I(Y2)'

Y,

c) [-I [Y, u Y2] = [-I(yl)

d) I-I(y, e) /-I(T

+ ... + I(x ..)i

:5 M.

n
-

Let S be the set of those x in 0 s

X-

s 1 fOT which/(x)

"# O. Prove that S is countable.

Y) = S - rl(y).

f) Generalize (c) and (d) 10 arbitrary unions and intersections.

2.8 Refer to Exercise 2.7. Prove that[[t-J(y)] only if, T = liS). 2.9 Let / : S
-+

Yforevery

subset Yof Tif, and

T be a function.

Prove that the following statements are equivalent. for all subsets A, B of S.

a) [is one-to-one on S. b} [(A n B) = /(A) n /(B)

2.21 Find the fallacy in the following "proof" that the set of all intervals of positive length is countable. Let {Xl> X2 •••• } denote the countable set of rational numbers ami let I be any interval of positive length. Then I contains infinitely many rational points x.., but among these tbere will be one with smallest index n. Define a function Fby means of the equation F(I) = n, if x. is the rational number with smallest index in the interval I. This function establishes a one-to-one correspondence between the set of all intervals and a subset of the positive integers. Hence the set of all intervals is countable.

c)

t:'

2.n

[/(A)]

for every subset A of S.

d) For all disjoint subsets A and B of S, the images f(A) and J (B) are disjoint.

Let S denote the collection of all subsets of a given set T. Let/: S .... R be a realy.uued function defined on S. The functionfis called oddttioe iff(A v B) [(A) + [(B) whenever A and B are disjoint subsets of T. If I is additive, prove that for any two subsets

A and B we have {(A. v B)

CHAPTER - A) and {(A v B) "'"f(A)

= 1(.11.) + I(B

+ feB)

- I(A

f\

B).

1.2.3 Refer to Exercise 2.22. Assume I is additive and assume also that the following
relations bold for two particular subsct$ A and B of T: I(A v B} = I(A.,) !(A " B)
=

+ RB,)

- {(A1/(B') 1(.14)

f(A)/(B),

+ feB)

[(n.
and com3.1 INTRODUCTION

ELEMENTS OF POINT SET TOPOLOGY

where A'

= T - A. B' = T' -.8. Prove that these relations detennine!(l1. pute the value of [(11.
REFERENCES FOR FURTHER FUllCtions. STUDY

SUGGESTED

2.1 Boas. R. P., A. Primer York, 1960.

01 Real

Carns Monograph No. 13. Wiley New ' Amsterdam, 1965.

2.2 FraenkeJ, A .• AbstrQct Set T1reory, 3rd ed. North-Holland.

2.3 Gleason, A., FlUIlhmenlah oj AbstrBt.:t A1Illl"sis. Addison-Wesley, Reading, 1966. 2.4 HaImos, P. R.., Naive Set 7Mory. Van Nostrand, New York, 1960. 2.5 Kamke, E., TIteory of Sets. F. Bagemihl. translator. Dover. New York, 19.50. 2.6 Kaplanst:y, I., Set Theory and Metric Spat:es. Allyn and Bacon, Boston, 1972. 2.7 Ro~an, B., and Kneebone, G. T., The Theory of Sets and Transfinite Numbers. ElscvICr, New Yruk, 1968.

A large part of the previous chapter dealt with "abstract" sets, that is, sets of arbitrary objects, In this chapter we specialize our sets to be sets of real numbers, sets of complex numbers, and more generally, sets in higher-dimensional spaces. In this area of study it is convenient and helpful 10 use geometric terminology, Thus, we speak about sets of points on the real line, sets of points in the plane. Or sets of points in some higher-dimensional space. Later in this book we will study functions defined on point sets, and it is desirable to become acquainted with certain fundamental types of point sets, such as open sets, closed sets, and compact sets, before beginning the study of functions, The study of these sets is called

point set topology.

3.2 EUCLIDEAN SPACE R" A point in two-dimensional space is an ordered pair of real numbers (XI' ;r~). Similarly, a point in three-dimensional space is an ordered triple of real numbers (x" X2, X3)' It is just as easy to consider an ordered '!-tuple of real numbers (XI> X2.· ..• x.) and to refer to this as a point in n-dimensional space. Dejinition 3,1. Let n > 0 be an imeger. An ordered set of n real numbers , x.) is called an n-dimensional point or a vector with n components. Points or tectors will usually be denoted by single bofd-/ace letters; for example,
(XI> X2, •••

or

The number Xi is called the klh coordinate of the point x or the kth component of the vector x, The set of all n-dimeasiona! points is called n-dimenstona! Euclidean space or simply n-space, and is denoted by R~. The reader may wonder whether there is any advantage in discussing spaces of dimension greater than three. Actually, the language of a-space makes many complicated situations much easier to comprehend. The reader is probably familiar enough with three-dimensional vector analysis to realize the advantage of writing the equations of motion of a system having three degrees of freedom as a single vector equation rather than as three scalar equations. There is a similar advantage if the system has n degrees of freedom.
47

EIeaads of Point Set Tepojogy

DeC. 3.2

Def.3.6

49

Another advantage in studying n-space for a general n is that we are able to deal in one stroke with many properties common to I-spaee, 2-space, 3-space, etc .• that is, properties independent of the dimensionality of the space. Higher-dimensional spaces arise quite naturally in such fields as relativity, and statistical and quantum mechanics. In fact, even infinite-dimensional spaces are quite common in quantum mechanics. . Algebraic operations on n-dimensionai points are defined as follows: Dejilritiml 1.2. a) Equality: b) Sum: c) Multiplication d) Difference: x y x y if, and only if, XI x

from (d) because

1Ix. + yl12

k=1

(x~

Yk)2

l~l

t (xi +

2xkY~

+ Y!)

NOTE.

Sometimes

the triangle inequality [x - zll

is written

in the form

Let x

(Xl" .. , X,.) andy

(Yh ...• Y .....•

X.

j>,,) be in R".

We define:

Ilx - yll

+ Ill' - zll.
y and y by y - z. We also have

X~

Y•.

This follows from (e) by replacing x by x -

+Y

= (XI

+ y"

... ,

+ y.).

IlIxll - tll'lll .:::.;lx - 111· I

Definitio" 3.4. The unit coordinate vector lit, in X8 is tne vector whose kth component is I and whose remaining components are zero. Thus,
01 =

by real nu.mbera (acalars): ax = (ax".. ,ax,.) (a real).

(1,0, .. , ,0),

U2

(0, I, 0, ... , 0), ... , n,

(0, 0, ... , 0, I).

Xl

e) Zero vector or origin;

f) Inner prfHhlCI or dot prfHhlct;

+ (-I)y.

If x

0= (0, .... 0).

x.' Ul'

(Xl>""

••• , X"

x.J then x = XlOl + ... + x"u.. and XI X'OI' x' u". The vectors u., ... , u" are also called basis vectors.

==

3.3 OPEN BALLS AND OPEN SETS IN Kif x'y = .~I

L xlY ..·
"
=

g) Norm or lenglh:

Let a be a given point in R" and let r be a given positive number. points x in R" such that

The set of

an

I[xll
The norm "x NOTE.

= (X'X)1/2

" L xi )111 .
1~1

IIx

all < r,

yll is called the d,stance between x and

y.

In the terminology

of linear algebra, X" is an example of a linear space.

is called an open n-bal! of radius r and center a. We denote this set by B{a) or by B(a; r). The ball 8(a; r) consists of all points whose distance from a is less than r. In R1 this is simply an open i-nterval with center at a. In X2 it is a circular disk, and in R3 it is a spherical solid with center at a and radius r.

Theorem 3.3. LeI x andy denote points in R". Then we haoe: a) [x]
;'<>:

0, Dnd [x]

0 if, and only if, "

O.

b) Ilax.
c) [x d)

lalll"ll for every real a. yll = II,. - x],

(Cauchy-Schwarz inequality).
)'[1 .:::.; Ilxll

3.5 DejUritiOIl of WI illlerior poillt. Let S be a subset o/R·, and assume that:ll E S. Then II is called an interior point of S if there is an open n-bal! with center at a. all of whose points belong to S.
In other words, every interior point :II of 8 can be surrounded by an n-ball 9(11) ~ S. The set of all interior points of S is called the interior of S and is denoted by jnt S. Any set containing a ball with center a is sometimes called a neighborhood of a.

IX'yl ~ IlxlillYll

e) IIx +

+ IIYII

(triangle-inequality).

Proof, Statements (a), (b) and (c) are immediate from the definition, and the Cauchy-Schwarz inequality was proved in Theorem 1.23. Statement (e) follows

3.6 Definition of points.

NOTE.

WI

open set. A set Sin R" is called open if all its points are interior

A set S is open if and only if S = int S. (See Exercise 3.9.)

lb.3.7

'lb. 3.11

ne

&nJctuR

or Opea Sets In R 1

51

EDmpJes. In R I the simplest type of nonempty open set is an open interval. The union of two or more open intervals is also open. A closed interval [a, h] is not an open set because the endpoints a and b are not interior points of the interval. Examples of open sets in the plane are: the interior of a disk; the cartesian product of two one-dimensional open intervals. The reader should be cautioned that an open interval in RI Is no longer an open set when it is considered as a. subset of the plane. In fact, no subset of R' (except the empty set) can be open in R2, because such a set cannot contain a 2"0011.

3.9 DejillitiDlIDfcomponentwerMl. Let S be an open subset of RI. An open internal I (which may be finite or infinite) is called a componem fnterwU of S if J s;; S and if there is no open internal J #- 1 such that I f:::: J <;;;; S. In other words, a component open interval contained in S. interval of S is not a proper subset of any other open set S belongs to one and only one

T/uwrl!m 3.10. Every point of a nonemp(v component interval of S.

In R" the empty set is open (Why?) as is the whole space R". Every open It-ball is an Open set in R~. The cartesian product
(OJ>

bl)

)( • ..

x (a", bit)

Proof. Assume XES. Then x is contained in some open interval I with I <;;;; S. There are many such intervals but the "largest" of these will be the desired component interval. We leave it to the reader to verify that this largest interval is Ix = (o(x), b(x», where a(x)

ofn one-dimensional open intervals (al> htl, ... , (a", b,,) is an open set in R8 called an n-dimensiona! open interval. We denote it by (a, b), where a = (al, ... , aJ and b = (blo ... , b,,). The next two theorems show how additional open sets in R" can be constructed from given open sets. Tluorem

inf {a : (a, x) S S},

b(x)

sup {b : (x, b) £ S}.

3.7.

The union of any coilection of open sets is an open

Set.

Proof. Let F be a collection of open sets and let S denote Assume XES. Then lit must belong to at least one of Since A is open, there exists an open n-ball B(x) f:::: A. and hence lit is an interior point of S. Since every point S is open.

their union, S = A~" A. the sets in F, say lit EA. But A f:::: S. so B(x) s,: S of S is an interior point,

Here a(x) might be - 00 and b(x) might be + 00. Clearly, there is no open interval J such that Ix 0;; J f:::: S, so Ix is a component interval of S containing x. If J x is another component interval of S containing x, then the union Ix u J" is an open interval contained in S and containing both Iex and J x' Hence, by the definition of component interval, it follows that I~ u J" = I~ and I~ u J" = J", so I", = J
I•

Tlu!orem 3.11 (RepreJelflatioll theQrem/or open lets ON tile rem liM). Every nonempty open set Sin Rt is the union of a countable collection of disjoint component intervals of S.

TheDrem 3.8. The intersection of a finite collection of open sets is open. Proof. Let S = A~ where each Ak is open. Assume XES. (If S is empty, there is nothing to prove.) Then lit E Ak for every k I, 2, ... , m, and hence there is an open n-ball B(x; l'k) s,: At. Let r be the smallest of the positive numbers rl. r2, •.. ,1",. Then lit E B(x; r) f:::: S. That is, x is an interior point, so S is open. Thus we see that from given open sets, new open sets can be formed by taking arbitrary unions or finite intersections. Arbitrary intersections, on the other band, will not always lead to open sets. For example, the intersection of all open intervals of the form (-lIn, I In), where n = 1,2, 3•.. , , is the set consisting of 0 alone. 3A THE STRUCTURE
1

Proof. If XES, let i" denote the component interval of S containing x. The union of all such intervals Ix is clearly S. If two of them. 1;,<and I" have a point in
common. then their union Ix U I; is an open interval contained in S and containing both (~and i,. Hence I" u I, I; and I,. u l, = so 1" = Iy• Therefore the intervals (. form a disjoint collection. It remains to show that they form a countable collection. For this purpose, let {XI' Xl' X3' ... } denote the countable set of rational numbers. In each component interval 1" there will be infinitely many x •• but among these there will be exactly one with smallest index 11. We then define a function F by means of the equation F((,,) = 11, if x. is the rational number in I~ with smallest index II. This function F is one-to-one since F(l,,) = F(l,) = 11 implies that I~ and Iy have x. in common and this implies Ix = I,. Therefore F establishes a one-to-one correspondence between the intervals (" and a subset of the positive integers. This completes the proof.

n:=1

r,

OF OPEN SETS IN R I

NOTE.

In R the union of a countable collection of disjoint open intervals is an open set and, remarkably enough, every nonempty open set in R_l can be obtained in this way. This section is devoted to a proof of this statement. First we introduce the concept of a component interval.

This representation of S is unique. In fact; if S is a union of disjoint open intervals, then these intervals must be the component intervals of S. This is an immediate consequence of Theorem 3.10. If S is an open interval, then the representation contains only one component interval, namely S itself. Therefore an open interval in RI cannot be expressed as

Def.3.12

DeI'.3.19

53

the union of two nonempty disjoint open sets. This property is also described by saying that an open interval is connected. The concept of connectedness for sets in RIO will be discussed further in Section 4.16. 3.5 CLOSED SETS J.12 Defillitioll 0/ II cleNd ret. R~ - S is open. Examples. A closed interval A set S in R" is called closed

In other words, x is an accumulation point of S if. and only if, x adheres to {x}. If xeS but x is not an accumulation point of S. then x is called an isolated point of S.

Examples

I, The

set of numbers

of the form Ifn, rr ~ 1,

2. 3, ...•

has 0 as an accumulation point.

if

its complement

2.. The set of rational numbers has every real number as an accumulation point. 3. Every point of the closed interval [a, b 1 is an accumulation point of the set of numbers in the open interval (a, b). Theorem 1.17_ If x is an accumulation point of S, then ellery n-ball B(x) contains infinitely many [Win's of S. Assume the contrary; that is, suppose an n-ball B(x) exists which contains only a finite number of points of S distinct from x, say .1' .2' ... ,am' If r denotes Proof the smallest of the positive numbers

la, b J in R

is a closed set. The cartesian product x [a., b.]

[01' b,] x ...

of n one-dimensional closed intervals is a closed set in R" called an n;;/imensiona/ closed interval [a, b],

The next theorem, a consequence of Theorems construct further closed sets from given ones.

3.7 and 3.8, shows how to

and the

Ilx
then B(x; ,/2) will be an n-ball about" from x. This is a contradiction. which contains

.",Ii.

TAwrem J.13. The union of a finite collection of closed sets is dosed, intersection of an arbitrary collection of closed sets is closed. A further relation theorem. Theorem 1.14. closed. between open and closed sets is described

no points of S distinct

by the following

If A is open and B is closed, then A

B is open and B - A is

Proof We simply note that A - B = A n (Rn - B), the intersection of .two open sets, and that B - A = B n (R" - A), the intersection of two dosed sets. 3.6 ADHERENT POINTS. ACCUMULATION POINTS points and accumulation

This theorem implies, in particular, that a set cannot have an aC(:umulation point unless it contains infinitely many points to begin with. The converse, however, is not true in general. For example, the set of integers {I, 2, 3, ... } is an infinite set with no accumulation points. In a later section we will show that infinite sets contained in some n-ball always have an accumulation point. This is an important result known as the Bolzano-Weierstrass theorem. 3.7 CLOSED SETS AND ADHERENT POINTS of an open set. The next theorem

Closed sets can also be described points.

in terms of adherent

A closed set was defined to be the complement describes closed sets in another way.

3.15 Dejinitio" of - adIIereflt poillt. Let S be a subse: of R", an.l x a point in R", x not necessarily in S, Then x is said to be adherent to S if euery n-ball B(x) contains at least one point of S.

Theorem 3.1B. A set S in R" is closed if, and only if, it contains all its adherent points. Proof Assume S is closed and let x be adherent to S. We wish to prove that lit € S. We assume x f S and obtain a contradiction. If x f S then x E R~ - S and, since R· - S is open, some n-ball B(x) lies in R" - S. Thus B(x) contains no points of S, contradicting the fact that x adheres to S. To prove the converse, we assume S contains all its adherent points and show that S is closed. Assume x € R" - S. Then x ¢ s, so X does not adhere to S. Hence some ball B(:x) does not intersect S, so B(x) S RIO - S. Therefore RIO S is open, and hence S is closed. 3.19 Defotitio" of CIeSllU. The set of all adherent points of a set S is called the closure of S and is denoted by S.

ExamJ*s
1. If XES, then x adheres to S for the trivial reason that every n-ball B(x) contains x, 2. If S is a subset of R which is bounded above, then sup S is adherent to S. Some points adhere to S because every ball B(x) contains points of S distinct from x. These are called accumulation points. 3.16 Definition of - IlccumuliIlioll poillt. If S S Rn and x E R", then x is called I1n accumulation point of S if every n-ball B(x) contains at least one point of S distinct from x,

54

EIemeots

or Point Set

Topofogy

Tb. 3.20

n..3.24

For any set we have S £; S since every point of S adheres to S. Theorem 3.18 shows that the opposite inclusion S £ S holds if and only if S is closed. Therefore we have:
Theorem 3.20. A set S is closed

----- -------j--

if and

only

if S = S.

I.U DejinitiOIi of derived set. The set of all accumulation points of a set S is called the derived set of S and is denoted by S'. Clearly, we have S = SuS' for any set S. Hence Theorem 3.20 implies that S is dosed if and only if S' £ S. In other words, we have:
Tlreorem 3.12.

---6:

Figure 3.1

A set S in R" is closed if, and only if, it contains all its accumulation

points.
3.8 THE BOLZANO-WEJERSTRASS 3.23 DejittitiDlI 0/ a bounded set. THEOREM

A set S in R" is said to be bounded if it lies entirely within an n-ball B(a; r) for some r > 0 and some. in R".

form two subintervals

11~1and I~~i, defined by the inequalities

Theorem 3.14 (&lZJllW-Werersfras).

man}'

If a bounded set S in R" contains infinitely points, then there is at least one point in R which is an accumulation point of S.
W

Next, we consider all possible cartesian products of Ihe form

(1) 11,.1:, x I(I}

2.t>

x ... x I (~)
" ...n'

(a)

Proof. To help fix the ideas we give the proof first for RI. Since S is bounded, it lies in some interval [-D, aJ. At least one of the subintervals [-a, 0] or [0, a] contains an infinite subset of S. Call one such subinterval [at> bI]. Bisect [al• b.] and obtain a subinterval [a2> bl] containing an infinite subset of S, and continue this process. In this way a countable collection of intervals is obtained. the nth interval [a., b,,] being of length h. - a~ = aI2·-1• Clearly, the sup of the left endpoints a. and the inf of the right endpoints b" must be equal, say to x, [Why are they equalf] The point x will be an accumulation point of S because, if ,. is any positive number, the interval [a.., b~] will be contained in B(x; r) as soon as n is large enough so that bw a. < r/2. The interval B(x; r) contains a point of S distinct from x and hence x is an accumulation point of S. This proves the theorem for Rt. (Observe that the accumulation' point x mayor may not belong to S.) Next we give a proof for R", n > I, by an extension of the ideas used in treating RI. (The reader may find it helpful to visualize the proof in'R2 by referring to Fig. 3.1.)

where each k i = I or 2.. There are exactly 2" such products and, of course, each such product is an n-dimensional interval. The union of these 2~ intervals is the original interval II' which contains S; and hence at least one of the 2" intervals, in (a) must contain infinitely many points of S. One of these we denote by 12• which can then be expressed as 11

IF)

11

2)

x ...

1!2).

where each 1~2) is one of the subintervals of Tfi of length Q. We now proceed with J2 as we did with II' bisecting each interval /pl and arriving at an n-dimensional interval 13 containing an infinite subset of S. If we continue the process, we obtain a countable collection of n-dimensional intervals It, 12, I'j, ... , where the mth interval 1m has the property that it contains an infinite subset of Sand can be expressed in the form

Since S is bounded, S lies in some n-baU B(O; a), a > 0, and therefore within the n-dimensional interval II defined by the inequalities
-Q ~

Writing we have
(k

Xl

Sa

(k

1,2, ...• /I).

x 1;1);

Here II denotes the cartesian product J.

=

1, 2•... , n).

Ifl)
,
Xk

x 1£1) x ...

that is, the set of points (Xl> ... one-dimensional interval -a :s;;

x,,), where x" E Itl) and where each IP' is a ~ a. Each interval JP) can be bisected to

For each fixed k, the sup of all left endpoints a1"'\ (m = I, 2, ... ), must therefore be equal to the inf of all right endpoints h~"i,(m = 1, 2, ... ), and their common value we denote by fl' We now assert that the point t = t2; •••• In) is an

o..

n. 3.28
accumulation point of S. To course, belongs to each of the m is such that aI2 .. -2 < rl2, contains infinitely many points an accumulation point of S. see this, take any n-ball .8(t; r), The point t, of intervals JI• Ji ••.• constructed above. and when this neighborhood will include J.... But since J. of S, SO will B(t; r), which proves tbat t is indeed 3. Let S = {(x, y) : x > O. y > o}. The collection F of all circular ~isks wi~ ~ters at (x, x) and with radius x. where x > O. is a covering of S. This covermg IS ,not countable. However, it contains a countable covering of S, namely, all those disks in which x is rational. (Sec.Exercise 3.18.) The Lindelbf covering theorem contains a countable subcolleetion the following preliminary 3.9 THE CANTOR INTERSECIlON THEOREM theorem we prove the Cantor As an application of the Balzano-Weierstrass intersection theorem. result: states that every open covering of a set S in R~ which also covers S. The proof makes use of

TIIuNtn 1.15. LeI {Q .. Q2' ... } be a countable collection of nonemply sets in Rsuch that:
(k ii) Each set

Tlwiwelfl 3.17 Let G {A It A.2" .} denote the countable collection of all nballs having rational radii and centers at points with rational coordinates. Assume x E R8 and let S be an open set in R· which contains x. Then at least one of the n-balls in G contains x and is contained in S. That is, we have x
E

Ak

!,;;

for some At in G. of Theorem 2.27, If x E R" and if S n-ball B(x; r} !,;; S., We s_haU~nd a is '~nea~" x ~n~, using this pomt. as which lies wlthm .8(x; r) and which x,,), that

1, 2, 3, ,'_. ).

Q. is closed and QI !S bounded.

I

Then the intersection

ni'~QJ: is closed and nonempty.

Proof. Let S = ():'=l Qk' Then S is closed because of Theorem 3.13. To show that S is nonempty, we exhibit a point x in S. We can assume ,bat each Q" contains infinitely many points; otherwise the proof is trivial, Now form a collection of distinct points A = {lei' :1:2. ..}. where Xi e Q.. Since A is an infinite set contained in the bounded set Ql' it bas an accumulation point. say lI:. We'shall show tbat xeS by verifying that x e Qi for each k, It will suffice to Show that 11:
is an accumulation point of each Qb since they are all closed sets. But every neighborhood of:x contains infinitely many points of A. and since all except (possibly) a finite number the points of A belong to Qk> this neighborhood also contains infinitely many points of QI' Therefore x is an accumulation point of Q. and the theorem is proved.

Proof. The collection G is countable because is an open set containing x, then there is an point y in S with rational coordinates. that center will then find a neighborhood In G
-

contains and k:

x. Write
be
1'1.

x
a rational Then

(Xt. Xl •••••

let Yl

number

such

Ir.. -

XlI

<

r/(4n)

for

each

1,2, ... ,

l1y - x]

S;

IYl

XII

+ .. , + Iy" -

x.1 <

4.'

or

Next, let q be a rational number such that rl4 < q < r/2, Then x E ~Y; q) and B(y; q) !;; B(:r;:; r) ~ S. But B(y; q) e G and hence the theorem IS proved. (See Fig. 3.2 for the situatiOfl in R 2,)

·3.1. TIlE LlNDELOF

COVERING TIIEORJ.!:M

In this section we introduce the concept of a covering of a set and prove the Linde/Of coveriTIg theorem. The usefulness of this concept wilt become apparent in some of the later work.
3.26 DejbliriM •• l:fIfJt/r;.g. A collection F 0/ sets is said to be a cOlJerillg of a given set S if S !;; A.-I" A. The. collectiem F is also said to cover S. If F is a collection of open sets, then F is called an open covering of S.

EDmples I. The collection of all intervals of the form lin < x < 2/n, (n = 2, 3,4, ' . ,), is an open covering of the interval 0 < x < 1. This is an example of a countable covering. 2. The real line gl is covered by the collection of al1open intervals (0. b). This covering is not countable. However, it contains a countable covering of Rl, namely. all intervals of the form (n, n + 2), where n runs through the integers.

CODnilfg

TII~o,em 1..21 (Lillilelif ctneriltg tlleorelll). Assume A !;; R· and let F be an open of A. Then there is a countable subco/leetion of F which also ('OVt'7'S A.

Proof. Let G

= {A" A2,.,,} denote the countab~ collection of all n-balls having rational centers and rational radii. This set G Will be used to help us extract a countable subcollection of F which covers A.

TIL 3.19

Th. 3.31

Co~inR'

59

Assume x E A. Then there is an open set Sin Fsuch that xeS. By Theorem 3.27 there is an n.ball Ai in G such that x E A.I;; !:;;; S. There are, of course, infinitely many such AI; corresponding to each S, but we choose only 6ne of these, for example, the one of smallest index, say m = m(x). Then we have x E A cS mC"'> - , The set of all n-balls A"'JI) obtained as x varies over all elements of A is a countable collection of open sets which covers A. To get a countable subcollectlon of F, which covers A. we simply correlate to each set At(,")one of the sets S of F which contained AI(,,)' This completes the proof.

3.11 COMPACfNESS IN a" We have just seen that if a set S in R" is closed and bounded, then any open covering of 8 can be reduced to a finite covering. It is natural to inquire whether there might be sets other than dosed and bounded sets which also have this property. Such sets will be called compact.

3.10 Ikftnitit", of IJ compact set; A set Sin R" is said to be compact if. and only if, every open covering of S contains 11 finite subcooer, that is, afinite subcollection which also covers S.
The Heine-Borel theorem states that every closed and bounded compact. Now we prove the converse result. set in R" is

3.11 THE HElNE-BOREL

COVERING

THEOREM

The Lindelof covering theorem states that from any open covering of an arbitrary set A in R~ we call extract a countable covering. The Heine-Borel theorem tells us that if, in addition, we know that A is closed and bounded, we can reduce the covering to a finite covering. 'The proof makes use of the Cantor intersection theorem. Theorem 1.19 (Hebre-Borel). Let F be an open covering of a closed and bounded set A in R". Then a finite subcollection of F also covers A.

Tlleorem 3.11. Let S he a subset of R", equivalent:

a) S is compact,

Then the fO{((Jwingthree statements are

b) S is closed and bounded.

c) Eoery infinite subset

of S has an accumulation point in S.

Proof. A countable sabeollection of F, say {II, 12, 3.28. Consider, for m ~ I, the finite union

•••

},

covers A, by Theorem

This is open, since it is the union of open sets. We shall show that for some value of m the union S", COvers A.

For this purpose we consider the complement RIO - S"" which is closed. Define a countable collection of sets {Q .. Q1' ••. } as follows: Q1 = A, and for m :» I, Q..
= A n (R" -

Proof As noted above, (b) implies (a). If we prove that (a) implies (b), that (b) implies (e) and that (c) implies (b), this will establish the equivalence of all three statements. Assume (a) holds. We shall prove first that S is bounded. Choose a point p in S. The collection of ,,·balls B{p; k), k = 1,2, ... , is an open covering of S. By compactness a finite subcollection also COvers S and hence S is bounded. Next we prove that S is closed. Suppose S is not closed. Then there is an accumulation point y of S such that y If; S. If x E S, let r" = Ilx - Y11/2. Each r.. is positive since y ¢ S and the collection {B(x; I J :XES} is an open covering of S. By compactness, a finite number of these neighborhoods cover S, say S
£;

SOl)'

U B(xl; l= I

It

rl)'

That is, Q.. consists of those points of A which lie outside of S"" If we can show that for some ",'\IU(; of m the set Q...is empty, then we will have shown that for thism no point or" A lies outside S ...; in other words, we will have shown that some Sm covers A. Observe the following properties of the sets Qm: Each set Q.. is closed since of the closed set A and the closed set R" - SOl' The sets Q.. ar~ decreasing, since the SIll are increasing; that is, Q.. + I £; Q... The Sets Qm' being subsets of A, are all bounded, Therefore, if no set Q.. is empty, we can apply the Cantor intersection theorem to conclude that the intersection 1 Qk is also not empty. This means that there is some point in A which is in all the sets Q... or, what is the same thing, outside all the sets 8M, But this is impossible, since , A £; Uk'= 1 Sl' Therefore some Q... must be empty, and this completes the proof.

Let r denote the smallest of the radii rl, rh . , . , r p' Then it is easy to prove that the ball B(y; r) has no points in common with any of the balls 8(X1; rk)' In fact, if x € B(y; r), then [x - yll < r ::; Tk, and by the triangle inequality we have lIy - xkll ::; 111- x] + IIx - Xlii, so

it is the intersection

tlx - Xlii <:: Ily

x.11 - [x

yll

2r.

[x - yll >

'k'

nk~

Hence x ¢ B(x.; r,.). Therefore B(y; r) n S is empty, contradicting the fact that y is an accumulation point of S. This contradiction shows that S is closed and hence (a) implies (b). Assume (b) holds. In this case the proof of (e) is immediate, because if Tis an infinite subset of S then T is bounded (since S is bounded), and hence by the Bolzano-Weierstrass theorem T has an accumulation point x, say. Now x is also

6}

an accumulation point of S and hence XES, since S is closed. Therefore (b) implies (c). Assume (e) holds. We shaJI prove (b). If S is unbounded, tben for every m > o there exists a point x... in Swith Ux...11 > m. The collection T"" {xJ. X2 .... } is an infinite subset of S and hence, by (c), T has an accumulation point y in S. But for m > I + IIyIl we have

We sometimes denote a metric space by (M, d) to emphasize that both the set M and the metric d playa role in the definition of a metric space. Examples 1. M = 11:"; d(I, Y) = Ilx YI. This is called tile Euclidean metric. Whenever we refer to Euclidean spate RIO, it will be understood that the metric is the Euclidean metric unless another metric is specifically IllCIltioned. 2. M = C. the complex plane; d(ZI' 2,,) = IZI ~ z~l. As a metric space, C is indistinguishable from Euclidean space R2 because it has the same points and the same metric, 3. Many nonempty set; d(x, y) = 0 if x = y, dtx, y} = I if x "" y. This is called the discrettl metric, and (M, d) is called a discrete metric space. 4. If (M. d) is a metrk space and if S is any nonempty subset of M, then (S, d) is also a metric space with the same metric or, more precisely, with the restriction of d to S x S WI metric. This is sometimes called tile relatioe metric: induced by don S, and S is called a metric sub$j1QCe f M. For example. the rational numbers Q with the o metric d(x, y) = Ix - yl form a metric subspace of R. R2; d(x, Y) = .J(XI - J'1)2-i- 4(Xl =-:;2)2, where x = (Xl> Xl) and y = The metric space (M, d) is not a metric subspace of Euclidean space R2 because the metric is different. M
()'1, )'2)'

1Ix... - yll ~ IIx .. II - IIyIl

>m-

~yU >

I,

contradicting the fact that y is an accumulation point of T. This proves that Sis bounded. To complete the proof we must show that S is closed. Let x be an accumulation point of S. Since every neighborhood of x contains infinitely many points of S, we can consider the neighborhoods B(x; Ilk), where k = 1,2 •...• and obtain a countable set of distinct points. say T = {Xl> Xl •••. }, contained in S, such that x.: EO' B(x; Ilk). The point x is also an accumulation point of T. Since T is an infinite subset of S, part (c) of the theorem tells us that T must have an accumulation point in S. The theorem will then be proved if we show that x is the only accumulation point of T. To do this, suppose that y ¢ x, Then by the triangle inequality we have

.s.

+ 11x.: - xII < IIY - xiii + Ilk, if~ E T. If to- is taken so large that Ifk < -lily - xii whenever k ;;:-:ko. the last inequality leads to -iIiY - xii < lr - xlii· This shows that Xl f. B(y; r) when k ~ ko• if r = llY - x], Hence y cannot be an accumulation point of T. This completes
lIy x] xLII

:s: lIy -

6.. M = {(Xl' XJ) : + I}, the unit circle in R"; d(x., y) smaller arc joining the two points I and y on the unit circle.

x~ xi "" xi

the length of tile

=

7. M = {(xl> X2, xl): + x~ + x~ = I}. the unit sphere in &3; d(x, y) of the smaller arc along the great circle joining the two points x and y. 8. M 9. M

the length

the proof that (c) implies (b). 3.13 METRIC SPACES

RR; d(x, y) = IXI RR; d(x, y)

)'11

+ ... +
)'11 •...•

Ix. -

)'.1.
y.I}.

max {lxI

Ix" -

The proofs of some of the theorems of this chapter depend only on a few properties of the distance between points and not on the fact that the points are in R". When these properties of distance are studied abstractly they lead to the concept of a
metric space.

3.14 POINT SET TOPOLOGY IN MEI1UC SPACES

J.12 IhjiItitiM of .. _rk IJIIII". If metric space is a nonempty set M of objects (CQ//ed points) tOfP!ther with a functiorr d from M x M to R (cal/ed the metric of the space) satisfying the follOWing four properties for all points x, y, z ill M:
1. d(x. x) = 3. d(x, y)

The basic notions of point set topology can be extended to an arbitrary metric space (M, d). If a E M. the ball B(a; r) with center a and radius r > 0 is defined to be the set of all x in M such that
d(x, a)

<

r,

o.
d(y, x),

2. d(x.y) > 0 ifx F y.

Sometimes we denote this ball by B1>I(a; r) to emphasize the fact that its points corne from M. If S is a metric subspace of M. the ball Bia; r) is the intersection of S with the ball BM(a~r).
~ In Euclidean space Rl the ball 8(0; I) is the open interval ( I. I). In the metric subspace S = [0, I] the ball .BJ.O~1) is the half-open interval [0, 1).
NOTE.

4. d(x. y) :i:;; d(x. z)

+ d(z, y).

Tbe nonnegative number d(x. y) is to be thought of as the distance from x to y. In these terms the intuitive meaning of properties 1 through 4 is dear. Property 4 is called the triangle inequality.

metric is not the Euclidean metric.

The geometric appearance of a ball in R~ need not be "spherical" if the (See Exercise 3.21,)

If S s; M, a point a in S is called an interior point of S if some ball BM(a; r) lies entirely in S. The interior, tnt S. is the set of interior peints of S. A set S is

FJeInaIIs

of Potut Set TOfIIlIccy

Th.l.33

Th.3.38

called open in M if all its points are interior points; it is called closed in M if M

is open in M.
.Examples.

The following theorems are valid in every metric space (M, d) and are proved exactly as they were for Euclidean space RW. In the proofs. the Euclidean distance IIx - yll need only be replaced by the metric d(x, y) . TlIeorem 3.35. a) The union of any collection of open sets is open, and the intersection of a finite collection of open sets is open. b) The union of a fin ite collection of closed sets is closed, and the intersection of any collection of closed sets is dosed. Tlleorl!m J.36. If A is open and B is closed, then A - B is open and B A is

1. Every ball BJg(o; r) in a metric space M is open in M,

2. In a discrete metric space M tJWry subset S is open.

In fact, if x e S, the ball R(x; ;) consists entirely of points of S (since it contains only x), so S is open. Therefore every subset of M is also closed t

l. In the metric subspace S ~ [0, 1] of Euclidean space Rl, evelY interval of the form [0, x) or (x, 1]. where 0 -c x <1, lS an open set in S. These sets are nol open in R '. Example 3 shows that if S is a metric subspace of M the open sets in S need not be open in M. The next theorem describes the relation between open sets in M and those in S. TIIeorem 3.33. Let (S, d) be a metric subspace of (M. d), and let X be a subset of S. Then X is open in S if, and only if,

closed. Theorem 337. For any subset S of M the following statements are equivalent,'
a) S is closed in M. b) S contains all its adherent points. c) S contains all its accumulation points.

X=A('\S
for some set A which is open in M. Proof. Assume A is open in M and let X = .A n S. If x E X, then x E A so BJI(x; r) <;; A for some r > O. Hence Bs(x; r) = B,H(x; t) ('\ S <;; A r, S = X so X is open in S. Conversely. assume X is open in S, We will show that X = A n S for some open set A in M. For every x in X there is a haJl Bs(x; r x) contained in X. Now Bs(x; 'x) = BJl(x; rJ n S, so if we let

d) S =S.
Example. Let M = Q. the set of rational numbers. with the Euclidean metric of RI. Let S consist of al! rational numbers in the open interval (0, b), where both a and b are irrational. Then S is a closed subset of Q.

»eX

U B",{x;

rJ,

then A is open in M and it is easy to verify that A ('\ S = X. T6eonm 3.34. Let (8, d) be a metric subspace of (M, d) and let Y be a subset of S. Then Y is closed in S if, aiu:/ only if. Y = B n S for some set B which is closed in M, Proof. If Y = B n S, where B is closed in M. then B= M - A where A is open in M so Y S II B = S 11 (M - A) = S - A; hence Y is closed in S. Conversely, if Y is closed in S, let X = S - Y. Then X is open in S so X = A ('\ S, where A is open in M and

Our proofs of the Bolzano-Weierstrass theorem, the Cantor intersection theorem, and the covering theorems of Lindelof and Heine-Bore! used not only the metric properties of Euclidean space R" but also special proper .ies of RR not generally valid in an arbitrary metric space (M, d). Further restrictions on M are required to extend these theorems to metric spaces. One of these extensions is outlined in Exercise 3.34. The next section describes compactness in an arbitrary metric space.

3.15 COMPACf SUBSETS OF A METRIC SPACE let (M, d) be a metric space and let S be a subset of M. A collection F of open subsets of M is said to be an open covering of S if S !;;;: A. A subset S of M is called compact if every open covering of S contains a finite subcover. S is called bounded if S ,; B(a; r) for some r > 0 and some a in M.

U~."

y
where B

= S=M-

S - (A n S)

S n (M
the proof.

A)

S ('\ B,

Tlreorem 3.38. Let S be a compact subset of a metric space M. Then:

i) S is closed and bounded. ii) Every infinite subset of S has an accumulation point in S.

A is closed in M. This completes

If S <;; M, a point x in M is called an adherent point of S if every ball B",(x; r) contains at least one point of S. If x adheres to S - {.t'} then x is called an accumulation point of S. The closure S of S is the set of all adherent points of S, and the derived set S' is the set of all accumulation points of S. Thus, S = SuS'.

Proof To prove (i) we refer to the proof of Theorem 3.31 and use that part of the argument which showed that (a) implies (b). The only change is that the Euclidean distance [x - rl is to be replaced throughout by the metric d(x, y).

Th.l.39 To prove (ii) we argue by contradiction. Let T be an infinite subset of Sand assume that no point of S is an accumulation point of T, Then for each point x in S there is a ball B(x) which contains no point of T (if x ; T) or exactly one point of T (x itself, if x E T). As x runs through S, the union of these balls B(x) is an open covering of S. Since S is compact, a finite subcolleetion covers S and hence also covers T. But this is a contradiction because T is an infinite set and each ball contains at most one point of T. EXERCISFS
Open and dosed sets in HI tIIIfl H2

3.1 Prove that an open interval in R I is an open set and tbat a closed interval is a closed set. 3.2 Determine all the ao::umuJa.tion points of the following the sets are open or closed (or neitber). a) All integers.
b) The interval (Q, b

sets in R 1 and decide whether

Nom. In Euclidean space R", each of properties (i) and (ii) is equivalent to compactness (Theorem 3,31). In a general metric space, property (ii) is equivalent to compactness (for a proof see Reference 3;4), but property (i) is not. Exercise 3.42 gives an example of a metric space M in which certain closed and bounded subsets are not compact.
lhormJ 3.3'. Let X be a closed subset of compact.
Q

J.
lIn, +
5-",

c) All numbers of the form d) All rational numbers.

(n

I, 2, 3.... ). ). ), ).

e) All numbers of the form 2 -. g) All numbers of the form

(m,1I = 1,2, (m,n = 1,2, (m,

It

compact metric space M.

Then X is

f) All numbers of the form (-I)" b) All numbers of the form

Proof. Let F be an open covering of X, say X s;; UA"F A. We will show finite number of Ihe sets A cover X. Since X is closed its complement M open, so F u {(M - X)} is an open covering of M. But M is compact, covering contains a finite subcover which we can assume includes M - X.

fore

that a - X is so this There-

+ (lIm)., (lIn) + Q/m), (- 1)"/[1 + (1/n»,

1,2,

(n = 1,2, ... ).

3.J The same as Exercise 3.2 for the foOowing sets in R2: a) All complex z such that Izi > 1.
b} All complex c)

M~

Al

U •••

u A" u (M - X).

z such that Izl ~ !. All complex numbers of the fonn (lIn) + (i/m).

(m, n

1.2. ... ).

dj All points (x, y) such that x2 ~ y2 < I. e) All POints (x, y) such that x > O.
f)

This subcover also covers X and. sinceM - X contains no points of X. we can delete the set M - X from the subcover and still cover X. Thus X s;; A I U' •• u A l' so X is compact.

All points

(x, y) such that x :!: O.

3.4 PrOVi: that eveIY nooempty open set S in RI contains both rational and irrational numbers. 35 Prove that the only sets in R 1 which are both open and dosed are the empty set and R.l itself. Is a similar statement true for R.2?
I.

3.16 BOUNDARY OF A SET lk./illitJoll 3.40. Let S be a subset of a metric space M, A point x in M is called a boundary point of S if every boll BAI(x; r) contains at least one point 'of S and at least one point of M - S. The set of all boundary points of S is called the boundary of S and is denoted by as.

3.6 Prove that every dosed set in Rl is the intersection of a countable collection of open sets. 3.7 Prove that a nonempty, bounded closed set S in R 1 is either a closed interval. or tbat S can be obtained from a closed interval by removing Ii rountable di~oint collection of open intervals whose endpoints belong to S. Open BDII closed sets in R" 3.8 Prove that open n-baJls and n-dimensional open intervals are open sets in R"'. 3.9 Prove thar the interior of a
set

The reader can easily verify that

iJS = This formula

I'l

M-

S.

shows that oS is closed in M.

in R." is open in R.".

EI:llIIIpIe In R·, the boundary ofa ball B(a; r) is the set of points 1 such that In R 1, the boundary of the set of rational numbers is all of R 1.'

Ix - _I

= r.

3.10 If S s;; R". prove that int S is the union of 311open subsets of R- which are contained in S. This is described by saying that int S is the largest open subset of S. 3.1t If S and T are subsets of RIt, prove that (iot S) n (int T) int (5" T), and (int S) u (int T) int (5 u T).

Further properties of metric spaces are developed in the Exercises and also in
Chapter 4.

..t

Elements o( Point Set Topology

67 3.l3 Assume that S s Kft. A point x in K" is said to be a C()lfde~tion point of S if every n-bafl B(x) has the property that B(x) n S is not countable. Prove that if S is nc t countable, then there exists a point x in S such that x is a condensation point of S. 3.24 Assume that S S R- and assume that S is not countable. condensation points of S. Prove that; a) S - T is countable, c) T is a dosed set, Let T denote the set of

3.12 Let S' denote the derived set and S the closure of a set S in R". Prove that:
a) S' is dosed in R!'; that i$, (51'

~ S'. c) (S u D'S' u 7'. e) S is closed in R".

b) If S £ T, then S' s 7'. d) (5)'

f)
=

S'.

S is the intersection of all closed subsets of R" containing S. That is, S is the
smallest closed set containing S.

b) S" T is nol countable,

d) T contains no isolated points.

3.13 Let Sand T be subsets of R". Prove that S n T S S n

if S is open.
NOTE.

and that S n T £ S () T

Note that Exercise 3.23 is a special case of (b). 3.25 A set in R" is called perfect if 5 = S', that is, if S is a closed set which contains no isolated points. Prove that every uncountable closed set Fin Rit can be expressed in the form F = A V B, where A is perfect and B is countable (Cantor-.Bendixon theorem). Hint. Metric Use Exercise 3.24. SIIlCES

The statements in Exercises 3.9 through 3.13 are true in any metric space.

3.14 A set Sin RR is called convex if, for every pair of points x and yin S and every real (J satisfying 0 < 0 < J, we have (Ix + (1 O)y EO S. Interpret this statement geometrically (in K2 and K3) and prove that:
a) Every n-ball in Kif is convex,

b) Every n-dimensional open interval is convex. c) The interior of a convex set is convex.
d) The closure of a convex set is convex.

J.IS Let F be a collection of sets in R", and let S ~ A and T A. For each of the following statements, either give It proof or exhibit a counterexample. a) If x is an accumulation point of AinF.

U~"F

n~~F

3.26 In any metric space (M, d), prove that the empty set 0 and the whole space M are both open and dosed. 3.27 Consider the following two metries in R":
d1(x, y) = max
l~t;!;;JI

r. then x is an accumulation

point of each set

IXi - y,l,

d2(x, y) =

E IXI C:;;;;:I

YII·

b) If x is an accumulation point of S, then x is an accumulation point of at least one set A in F. 3.16 Prove that the set S of rational numbers in the interval (0, I) cannot be expressed as the intersection of a countable collection of open sets. Hint. Write S = {XI, X2' .•• }, assume S =1 Sk' where each Sk is open, and construct a sequence {Q.} of closed intervals such that Qu 1 £ Qn S S" and such that x" ¢ Q". Then use the Cantor intersection theorem to obtain a contradiction.

In each of the following metric spaces prove that the ball B(a; r) has the geometric appearance indicated: a) 10 (R", d1), a square with sides parallel to the coordinate axes. b) In (R2, d,), a square with diagonals parallel 10 Ihe axes.
c) A cube in (R3, d1). d) An octahedron

n:,~

in (R3, d2).

3.28 Let dJ and dz be the metrics of Exercise 3.27 and let [x - 111 denote the usual Euclidean metric, Prove the following inequalities for all x and y in R": dl(x, y) ~ [x -

3.17 If S S Rn, prove that the collection of isolated points of S is countable. 3.18 Prove that the set of open disks in the xy-plane with center at (x, x) and radius x > 0, x rational, is a countable covering of the ~ {(x, Y): x > 0,,, > .oj. 3.19 The collection F of open intervals of the form (lIn, 2In), where n = 2, 3, ... , is an open covering of the open interval (0, l). Prove (without using Theorem 3.31) that no finite subcollection of F covers (0, I).
3.%0 Give an example of a set S which is closed but not bounded and exhibit a countable open covering F such that no finite subset of F covers S.

yll ~

d2(x,

y)

and

d2(x, y)

~ ...1;llx - yll ~

nd.(x, Yl.

3.29 If (M, d) is a metric space, define

d'(x, y) =

1 + d(x, y)

d{x, y)

Prove that d' is also a metric for M. Note that 0 ~ d'(x. y) < I for all x, y in M. 3.30 Prove that every finite subset of a metric space is closed. 3.31 In a metric space (M, d) the closed ball of radius r > 0 about a point a in M is the set B(a; r ) = {x: d(x. a) s r }, a) Prove that Bt«; r) is a closed set. b) Give an example of a metric space in which B(a; r) is not the closure of the open
ball B(a; r).

3.21 Given a set Sin R" with the property that for every x in S tbere is an n·ball B(x) such that B(x) n S is countable. Prove that S is countable.

3.n

Prove that a collection of disjoint open sets in R" is necessarily countable. example of a collection of disjoint closed sets which is not countable.

Give an

3.32 In a metric space M. if subsets satisfy A !i 5 5 A. where A is the closure of A. then A is said to be t:k1lS# in $. For example, the set Q of rational n umbers is dense in R. If A is dense in $ and if S is dense in T, prove that A is dense in T. 3.33 Refer to Exercise 3.32. A metric space M is said to be separable if there is a countubk subset A which is dense in M. For example, R is separable because the set Q of rational numbers ill a countable I.k!nsesubset. Prove that every Euclidean space Rt is separable. 3.34 Refer to Exercise 3.33. Prove that the Lindelof covering theorem (Theorem 3.28) is valid in any separable metric space. 3.35 Refer to Exercise 3.32. If A is dense in 5 and if B is open in S, prove that B 5 A n B. Hilft. Exercise 3.1 3: 3.36 Refer 10 Exercise 3.32. U each of A and B is dense in 5 and if B is open in S, prove that A n B is dense in 5. 3.37 Given two metric spaces (S1> d1) and (S2' d2). a metric p for the Cartesian product S, x 52 can be constructed from dl and d'l in many ways. For example, if x (x 1> Xl) and y = (Yr. y,) are in 5. )( S'h Ji:t p(x. y) = dl(Xh ytl + d2(X2, Y2)' Prove that pis a metric for $1 x 52 and construct further examples.
COIIIfII'CI; su.bsets of:ll metric space

3.49 If int A

int B

8: and

if A is closed in M, then int (Il U B)

=

0.

3.50 Give an example in which inl A

J.5Z If A

int B

="

but int (A u B)

= M.

351 all "" A ..... - A and cA = Cl(M - A). M

n 11 = 0, then

0(1l u B) = oA V oB.

SUGGESTED

REFERENCES

FOR FURTHF.It STUDY Carus Monograph No. 13. Wiley, New Addison-Wesley, Reading. 1966. New

3.1 Boas, R. P_, A Primu of RIal Flmetiotu.

Yod:.l960.
3.2 GIeuon., A., FIIIIGb'tt!1ftolJ of ~ 3.4 Simmons, O. F.. l~ YOI'k.I963.
AlfIQ/y.rU.

3.3 ICapJansky, I .• Set 'T'heory IIIfIl Metric Sjut-n. Allyn and Bacon, Boston, 1m. to Tupo/IIgy_ Modem Ana/yAr. McGraw-HiD,

Prove each of the roHowing statements concerning an arbitrary metric space (M, d) and subsets S, T of M. 3.38 AIISUffiC SsT SO M. Then S is compact in (M. d) if, and only if, S is compact in the metric subspace (T, J). 3.39 If S is closed and T is compact. then S nTis compact. 3.40 The intersection of an arbitnuy col1cction of compact subsets of M is aJmpact. 3.41 The union of a finite number of compact subsets of M is compact, 3.42 Consider tbe metric !ipMle Q of rational numbers with the Eudidcan metric of' R. Let 5 consist of all rational numben in the open interval (a. b). wIlcR= a and b an:: irrational, Then S is a dosed and bounded subset of Q which is not compact.. MiIIcdIueouIIi 3..43 int A pnIIICI1ies of Ole iDkriGr ... M - M - A. M - .:4.
= tile ~

If A and B denote arbitrary subsets of a metric sp;a.ce M, prove that:

=

3.44 int (M - A) 3.4ti a) int

3.45 int (int A) = int A.

b) s ,,1' (int A), if Eis an infinite coIli!ctioo of s.uhlieiS of M. c) Give an example where equality does not bokI in (b). lA' a)

«()~~1 AI) int <()"'''F A)

n:-t
int

n...

(int Ai). where each Aj SO M.

U"'~F(int A)

SO

(U...er A).

b) Give an example of a finite collection Fin whdl «)U3.lity does not hold in (a). 3.41 a) int (ilA) = 0 if A is open or if A is- closed in M.
b) Give an example in which int (aA) "'" M.

n.o CHAPTER 4

71

x •.... p, If there is no .mch pinS,

We also say that {x.} converges to p mul we write x" .... p as n .... eo, or simpl, the sequence {x ..} is said to diverge. The definition of convergence x.
-+

LIMITS AND CONTINUITY

NOTE.

implies that d(x ... p) ....

if and only if

o.

The convergence

space Ri.
4.1 INIRODUCDON

of the sequence (d(x... p)} to 0 takes place in the Euclidean metric

The reader is aJready familiar with the limit concept as introduced in elementary calculus where, in fact, several kinds of limits are usually presented. For example, the limit of a sequence of real numbers {x.}. denoted symbolically by writing
lim .... "" x" = A.. means that for every number s

1. In Euclidean space ai, a sequence {x.l is called Uu:retlSing if x~ S; x.+t for all 11. If an increasing sequence is bounded above (that is, if xR :s; M for some M > 0 and all 11). them {x.} converges to the supremum of its range. sup {Xl> X2 •••• }. Similarly, {x.} is called det:reusing if x.+ I S; x. for all 11. Every decreasing sequence wbk:h is bounded below converges to the infimum of its range. For example. {lIn} converges toO. 2. If {a,.} and {b.} are real sequences converging to 0, then {a,. + b.} also converges: to O. If 0 ;0;; c. ,..;; ,. for all a and if {o.l conveJBeS to O. then {c..} also converges to O. a These e1ementary properties of sequences in RI can be used to simplif'y some of the

>

0 there is an integer N such that whenever n ;;:.: . N

This limit process conveys the intuitive idea that x. can be made arbitrarily close to A provided that n is sufficiently large. There is also the limit of Q junction, indicated by notation such as lim !(x) = A, 11/;"',
which means that for every 8 If(x) - AI

proofs

conceming

limits in a

general
=

metric space.

3, In the complex plane C. let z. 1 + 2i because d(z .. I

+ 11-2 +
+

(2 - Iln)i.

Then {z.} converges to

2i)l

= Iz" - (I

> 0 there
Ii:

is another 0

number (j

> 0 such

that

so d(z", 1 TMormr4.2.

2iW = _: + _1_ .... 0 as n .... 00, n "J

2i) .... O.

<

whenever

< Ix - pi <

a.

A sequence {x.} in a metric space (S, d) can converge to at most one We will prove that p

This conveys the idea that f(x) can be made arbitrarily close to A by taking x sufficiently close to p. Applications of calculus to geometrical and physical problems in 3-space and to functions of seveml variables make it necessary to extend these concepts to R·, It is just as easy to go one step further and introduce limits in the more genera1 setting of metric spaces. This achieves a simplification in the theory by stripping it of unnecessary restrictions and at the same time covers nearly all the important aspects needed in analysis. First we discuss limits of sequences of points in a metric space. then we diSCWIS limits of functions and the concept of continuity,
4.% CONVERGENT Dt!ji1titim14.I.
Q

point in S.
Proof Assume that x •.... p and x •.... q. triangle inequality we have

q.

By the

o s; d(p,
Since dept x.)
-+

q) ~ d(p, x

H)

+ d(x",

q).

0 and d(x •• q) .-;. 0 this implies that d(p, q}

0, so p

q;

If a sequence {x.} converges, the unique point to which it the limit of the sequence and is denoted by lim x .. or by lim._oo

converges
x".

is called

SEQUENCES

IN A METRIC

SPACE

Example. In Euclidean space Rl we have Iirn,."'<I) lin = O. The same sequence in the metric subspace T = (0, 1] does not converge because the only candidate for the limit is o and 0 j T. This example shows that the convergence or divergence a sequence depends on the Wlderlying space as weJI as on the metric.

or

if there is point
For etJery
8

A sequeru:e {x.} of points in a metric space (S, d) is said to converge p in S with the following properly:

11IetJrem 4.3. In a metric space (S, d), assume x~ .... p and let T = {Xl'
be the range of {xR}. Then:

X2.·

•• }

> 0 there

is an integer N such that

&

d(x., p) <

a) T is bounded.
b) P is an adherent point of T.

whenever n :2:. N.
70

73

hoof. a) Let N be tbe integer ~ndinl to II 1 in the definition of convergence. Then every z" with n :=: N lies in the ball SCp; I); so every point in T lies in the ball B(p; F), where .
F

4.7 ~ (1/11 CllllCIty SqllfUlCe. A sequence {x,,} in a metric space (S, d) is called a Cauchy sequence if it satisfies the following condition (called the Cauchy
condition): For every e > 0 there is an integer N such that d(x,., x...) <
E

max {dept Xl), ...•

d{p,

XN-1)}.

Therefore

T is bounded. b) Since every ba1l B(p; e) contains. a point of T. p is an adhcn:nt point of T. If T is ~te.
an aa:umuJation point of T.

whenever

11

2: Nand m ~ N.

~
p
IS

every ball B(P; e) contains infinitely many points of T. so

The next theorem provides a converse to part (b). J'IemuI 4.4. GWen0 metric IPfI" (S, d) ond a su.baet T s;;;;; S. II a point pinS is adherent point of T, ther. there MO sequence (xJ of points in T which amot!T(IeS top.
all

Theorem 4.6 states that every convergent sequence is a Cauchy sequence. The converse is not true in a general metric space. For example, the sequence {J/n} is a Cauchy sequence in the Euclidean subspace T = (0, I] ofR!, but this sequence does not converge in T. However, the converse of Theorem 4.6 is true in every Euclidean space a-I.. T1Ietwem

4.8.

In Euclidean space Rl every Cauchy sequence is cfJlWergent.

Proof. For every integer n ~ 1 there is a point x. in T with d(P. x.) s; lIn. Hence d(p, xJ -+ 0, so z" -+ p.

if, eDeI')'

4.5. In Il metric space (S, d) a .requnrce {x.} C01IVf!rgesUJ p if. ond only
~

{x"(.)}

CUItDeI'fJD

to p.

Proof: Assume z" -+ p and consider any subsequence {xt(.,}. For every .. > 0 there ~ an N_ such that n ~ N implies d(x.. p) < 8. Since {x....)} is a subsequence, there IS an mteger M such that k(n) i!: N for n 2: M. Hence n 2: M implies d(x!{.}, p) < £, which proves that xl.(JI) -+ p. The converse statement bolds trivially since {x.} is itself a subsequencc.
4.3

CAUCHY SEQUENCES

Proof Let {x,,} be a Cauchy sequence in R' and let T {Xl' X2> ••• } be the range of the sequence. If Tis finite, then all except a finite number of the terms {x,,} are equal and hence {x.} converges to this common value. Now suppose T is infinite. We use the BoJzano-Weierstrass theorem to show that T has an accumulation point p, and then we show that {x.} converges to p. First we need to know that T is bounded. This rollows from the Cauchy condition. In fact, when 1) = 1 there is an N such that 11 :<! N implies 11x.. - xjIIlI < L This means that all points x, with n :<! N lie inside a ball of radius 1 about as center, so T lies inside a ball of radius I + M about 0, when: M is the largest of the numbers ~xlll •... ,II"NII. Therefore, since T is a bounded infinite set it has an accumulation point p in Ri; (by the Bolzano-Weierstrass theorem). We show next that {ll.} converges to p. Given .t > 0 there is an N such that Ilx" x..II < 1;/2 whenever n :<!: Nand m ~ N. The ball B(,; e/2) contains a paint x.. with m 2! N. Hence if n 2: N we have

"N

If a sequence {x.} converses to a limit P. its terms must uJtimately become close to p and hence close to each other. This property is stated more formally in the next
theorem. so lim
metric Sptlce

IIx.. ~ pll s I1x" -

X ..

II + Ux... - pU < 2: +

[;

2=
I:

a,

x..

p. This completes

the proof.

1'iuIum 4.4. h.nune tItlIl {x..} converges in II > 0 there is all integer N $IIch that d(x". x..) <
8

(S. d). Thenlor every

Examples

whenerJer n ~ N (1lf(/ m ~ N.

1. Theorem 4.8 is oRen used for proving the convergence of a sequence when the limit is not known in advance, For example. consider the sequence in RI defined by
X

Proof. Let p = lim z". Given Il > 0, Jet N be such that d(x"" p) < 8/2 whenever " ~ N. Then d(x•• p) < 6/2 if m ~ N. If both II 2: N and m 2: N the triangle" inequality gives us
d(x ..

=1--+---+"'+--.
234

{_I)""-1

x,J

::s;: d(x .. p)

+ d(p, x,J < .! + ~ "" s:

2 2

If m > n

;:0:

N, we find (by taking successive terms in pairs) that Ix..

x,,1

If

+ ... +

I <1< I -m n - N'

74
so <: e as seea as N:> lIt. Therefore Ii CAirucny seqeenee itnd hena; it converges to some limit. It can be shown (see Exercise lUS) dud this limit is 2, a fact which is not immediately obviQus. 1. Given It real sequence {an} such that :s for all n ;e: L We can. prove that {a,,} converges. without knowing its limit Lei a" "" I~ ThenO:s :S so, by :S Heoceb" -> 1Ji, Also,lfm :> 17 wehave
I-

1S

Definition I,ll.

em accumulation
"" b,

and

T, the notation

is

to mean

for ellery

t :> ()

there is a () :> 0 such that whenever x

E

A,

~. p, and
as x tends to p, is indicate this
1:1,."

<:

is read "the limit " We sometimes

or

This 4.4 COMPLETE METRIC SPACES d) is called c()f11plete is caJied eommete Rl is

Definition 1.9. A metric space in S converges in S. A subset is COf.npl,(ne.

as x - p. The definition conveys the intuitive idea that dose to h dose to p, We that p be of A to make certain that there will be x in A sufnci.,ently dose to p, with x ~ p. 17 need not he in the domain and b need not he in the range

EDmpre 1. Every Euclidean space R~ is complete (Tlloorem 4.8). In rnu'1ienl"r complete, but the T {O, 1] is not complete.

The next theorem relates COlnp:lek:ne:ss with compactness.

T/worem 4,10. In any metric space I?pery compact subset T iii cmnPtete.
Figure 4.1

sequence in T and let A = Xz, .. } denote the range of then converges to one of the elements hence converges in T. If A is Tbeorem 3,38 tens us that A has an accumulation pm T since Till compact. We sbownext that x" - p. Given 8 :> 0, choose N so that Nand m ;;::; N < The bal! contains a m ;;;:N, Therefore if n > N the us

NOTE.

The definition can also be formulated in terms of baUs. Thus, and if, for every ball there is a bali such that empty and such that
whenever
XE

holds if,

A ill not

<

.2

+~=t
2

'

sequence in Thal> limit in T, so Tis rornptete.

4.5 LIMIT OF A FL'NCTION

When formulated this way, the definition IS when p or b are in the extended rea! number R* or In the extended number system C*. However, In what it to be understood that p and b are finite unless it is stated that can be infinite, The next. theorem relates limits of functions to limits

in this section we consider two metric spaces and where denote the metries, Let A be a subset of S and let f :A function from A to T

and
T he a

TMtJrtm 4.11. AS$ume p is an accumulation lim

z~p

0,

TIl. 4.14 Then we also Itave:

77

if, and only

if.

..... ""
Proof

lim I(x.) = b,
{p} which converges to p. 0 there is a /j

(3)

a) lim,,_,

[J(x)

g(x)] = a = ab, = alb

b.

b) lim" .. ,/(x)g(x)

for every sequence {xR} of poillls in A -

c) lim, ....f(x)/g(x) ,

if

b "" O.

If (2) holds. then for every

dy{/(x). b)

>

>

0 such that p) -c 6. (4j

<e

wbenever x

eA

and 0

< ds(x,

Now take any sequence {x.} in A - {p} whlch converges to p. For the {J in (4), there is an integer N such that 11 ~ N implies dJ..x", p) < o. Therefore (4) implies drlf(xJ. b) < e for n ~ N. and hence {!(xR)} converges to b. Therefore (2) implies (3). To prove the converse we assume that (3) holds and that (2) is false and arrive at a contradiction. ]f (2) is false. then for some Il > 0 and every;; > 0 there is a point x in A (where x may depend on /j) such that

Proof Weprove(b),leavingtheotherpartusexercises. GivenllwithO < £ < I, let e' be a second number satisfying 0 < e' < 1, which will be made to depend on Bin a way to be described later. There is a b > 0 such that if x E A and d(x, p) < (j, then jf{x) - al < 1'.' lo(x) - bl < £'. and Then 1!(x)1 = la + (f(x) - a)1 < lal + £' < lal + I. Writing/(x)g(x) - ab

= I(x)g(x)
obi < (lal

- bf(x)

I/(x)g(x) -

S If(x)llo(x)

0< dS<x,p) <

{J

but

drlf(x),

b) ~

8.

(5) sequence of

I}l:'

+ bf(x) - ab, we have - hI + Ibllf(x) - al + Ibl&' = i(lal + !hI +

1).

Taking II = lin, n = 1,2, ... , this means there is a corresponding points {x.} in A - {pI such that

If we choose Il' x E A and d(x. p)

< b, and

elnal + Ibl +

I), we see that I/(x)g(x) this proves (b).

001 < Il whenever

o < ds(x.,

p)

<

lin

but does not con-

4.7 LIMITS OF .VECTOR-VALUED FUNCf(ONS Again, let (S, d) be a metric space and let A be a subset S. Consider two vectorvalued functions f and gdefined on A. each with values in gi,

Oearly. this sequence {x.} converges to p but the sequence {f(xJ} verge to b. contradicting (3).
NOTE.

of

Theorems

4.12 and 4.2 together

show that a function

cannot

have two

f:A ..... \ R
Quotients of vector-valued functions arc not defined (if k > 2), but we can define the sum f + I, the product Ai (if l is rea1) and the inner product f·g by the respec-

different

limits as x .....p.

4.6 LIMITS OF COMPLEX·VALUED FUNCTIONS

tive formulas
(f

Let (S, d) be a metric space, let A be a subset of S. and consider two complexvalued functions I and 9 defined on A, I:A ..... , C
g:A .....C. The sum f + 9 is defined to be the function whose value at each point x of A is the complex number I(x) + g(x). The difference I - g, the product I' s, and the quotientflg are similarly defined. It is understood that the quotient is defined only at those points x for which g(x) "" O. The usual rules for calculating with limits are given in the next theorem. TMtJre", 4.13. Let I turd 9 be complex-valued functions defined on a subset A 01 tJ metric apace (S, d). Let p be an accumulation point 01 A, and assume that Jim/(x) ;<-, = a,
x ....v

+ g)(x)

= ((x)

+ g(x).

(If){x)

1I(x),

(f'g)(x)

f(x)·g(x) with limits of

for each x in A. We then have the foUowing vector-valued functions. T6ttor~", 4.14.

rules

for calculating

Let p be an aCCTImuiation point 01 A and assume that

lim

"'

..,rex) = a,

lim g(x)
Z"'I'

b.

Then we also have:

a) lim,,_

[f(x)

b) lim,,_p If(x) d) lim",_l'

+ g(.x)] = • + b, = 18 lor !,,"v scalar 1,

= .·b,

c) lim,,_p I(x)·g(x)

Hilt g{x)

h.

IIf(x) II

11.11·

18

79

and
The
(}

To prove

we write

tends to O as x To prove
NOTE.

-+

that

:s;

p, so

a' b, This proves

A, and let f': A
E

Let /j' ' ,/"be n real-valued be the vector-valued fu.nction defined

funceions defined lfx

ll"

of 51 (a of S which is not an accumulation at p will be continuous at p because for ~llj'fic'i"'ntl\l < li, x = p, and eS.

of

A,

every sequence

in S convergent ta p, the sequence

,,11

, in

we have

Jim The of this theorem is similar to that. of Theorem 4.12 and is len as an exercise for the reader. (The result can also be deduced from 4.12 but there is a cOlrnpHCEttle,n in the argument due to the fact that some terms of the sequence could be to The theorem is often described that for continuous functions the UU;;n,;JI!<;UlgC'U with the function Some care is needed converges when
Example If x" _, x and.pM y in a metric space (S. d), then (Exercise 4.1). The reader can that d is continuous on the metric space where p is the metric of Exercise 3.37 with Sl "'" $2 = S. ..... t», y} d
x S, p),

These
for each r.

show that

=a

if.

and

if,

4•. CONTINUOUS fiUNCTlONS 8 The definition of in elementary functions from one metric space to another. Dejilfitiolf 4.1.5. Let S to T calculus can be extended to

be metric spaces and

o such

said
that

continuous
p) <

/'S-tTbea
a poinr P in S

Continuity of a function/at continuous an A,

delrlf'rl,j"

subset A This definitionreflects dose continuous at p jf and

the intuitive idea that dose to are It can also be stated in terms of balls; A function if, for every t > 0, there is a S >0 such that

a p is called a on the behavior in the immediate concerns the whole domain its domain is a property.

4.9 <:ONTlNUlTY OF COMPOSITE FUNCTIONS ThefRetn 4.17. Let

is a s) in 1'. is an accumulation
f)~

under

must be contained

in the ball that

-U

be metrtc spacer, Let [ : S 4 T on S x in S.

lim

continuous at p and

continuous

then h is continuous at

Let

Given

;;:. 0, there is a it ;;:. 0 sllch that

<:r.

whenever

b) <:

For this S there is a

such that whenever

p) <: we find that

these two statements So II is continuous at p.

4.1(. CONTINUOUS

and whenever

If S is a subset 011 C on which the not continuous on S. Therefore a rational function where g continuous at those C at which the denominator does not vru:llsn, The familiar al-valued functions such as the exand ''''l!;GH •• liJt.t,-, ttU1Cl:lODiS, U ele:rnelllUlry fllIlC:tiollS 3 , the common
vahle of the

<: II',

C"OMPLEX.-VALUED AND VECI"OR·V ALUE.D Ft}NCfIONS

a metric space

commex-oatuea tuncuons conusuous at a p in are each continuous llt p. The

is an accumulation

of the OOlrrlP'IeX ""IJUU.f;W'"'' and tnaonomemc functions is a c01~rej'PO'!ld:mg real-valued fW1ctions and consequence of the ooIltirnlil:y Theorem. 4.. ,0. 2
4.12 CONTINUITY AND INVERSE IMAGES OF OPEN OR CLOSED SETS

is also continuous at p The result is trivial is an isolated of S. of S, we obtain the result from Theorem 4.1.3. There is, is
II:

The of Inverse of continuous functions.

cOlrrel>PlJmding

in the same way,

TlteorIt1l14,19. Let r wid II: be continuous at a p in a metric space and assume that f and g have Faiues inR". Then each continumts at p: the sum f + g, the every real A, the inner f·g, and the narm
TlutOl'l!m

4.21.

Let

then we haoe: is an isolated

we note that k """1,2, .... , n. 4,.U EXAMPLES OF CONTINUOUS FUNCfIONS
tr.ll!ow""",

is an eccumulsfor each transtatron of the definition of the sym[t should he observed

It is a trivial exercise to show that the functions are oontinuollsoll C: fUn:cti()n$, defined

function defined nomiuJ:

= c for every z in "" z for every z ill C.

12

Note that the statements in Theorem 4.22 can also be cxprc:sscd as follows:

Note also thatf-'(A

u B)

= /-I(A)

u /-I(B)

for all subsets A and B of T.

TIvort!llf 4.23. Let f :S - T be afimctionfrom one metric space (S, ds) to QlWther (T, dr). Then / is contilfUOll3 on S I/. and OIIly if, for every open set Y in T, the inrJe1'seimage /-1(y) is open in S.

Proof. Let Fbe an open covering off(X). so thatf(X) !;;; A. We will show that a finite number oftbe sets A coverf(X). Since/is continuous on the metric subspace (X, ds) we can apply Theorem 4.23 to conclude that each set ~ -l(A) ~s open in (X, ds). The sets I-I(A) form an open covering of X and, smce X IS compact, a finite number of them COVeT X. say X s;;; f-1(A I) u ••. v /-l(A ,,). Hence I(x) £ /U-1(A1) U'" uj-l(A,,)] = f(f-l(A1)] U'" v fU-!(A,)]
~ Al
U ···V

UAeF

A",

r I(Y).

Proof. Let/be continuous on S, let Y be open in T, and let p be any point of We will prove that p is an interior point of/-l(y}. Let y = /(p). Since Y is open we have B.,{y; e) £ Y for some! > O. Since I is continuous at 11. there is a.5 > 0 such that/(Bs(p; .5) £ BrtY; t}. Hence,
8M;~}
£;

sof(X}

is compact. As a corollary of Theorem 3.38, we see thatf(X) is closed and bounded.

De./iltititJ1t 4.26. A function f: S -+ Xl; is called bounded on S number M such that ~f(x)11~ M for all x in S.

if there is tl positive

/-l(f(Bs(p;

6»]

s;; /-I[8.,{y;

g)] £ /-I(y)'

so p is an interior point of/-I(y). Conversely. assume that f-l(Y) is open in S for every open subset Yin T_ Choose pinS and let y = /(P), We will prove thatfis continuous at p. For every 6 > 0, the ball Brty; &} is open in T, so rl(Brty; e) is open in S. Now, p ef-1(B.,.{y; e) so there is a ~ > 0 such tbat Bs(p; 8) £ 1-'(BrtY; Therefore,/(Bs(p; 0») £ 8T(y; 8) so f is continuous at p. ,

Since r is. bounded on S if and only if f(S) is a bounded subset of Rl, we have the following corollary of Theorem 4.25. Tlworem 4.27. Let f: S -+ Ri be a/unction from a metric space S to Euclidean space Rt. If r is continuous on a compact subset X of S, then f is bounded on X. This theorem has important implications for real-valued functions. If / is real-valued and bounded on X, then I(x) is a bounded subset of R, so it has a supremum, supf(X), and an infimum, inff(X). Moreover, inf I(X)

e».

TlIMnm 4.24. Let f :S -+ T be aJutretionfrom one metric space (S, ds') to another (T. dr). Then I is continuous on S if. and only if, for every closed set Y in T, the inverse image f-l( Y) is closed in S.

s fix)

:S sup /(X)

for every x in X. takes on the values supf(X)

Proof. If Y is closed in T, then T - Y is open in T and f-l(T

Now apply Theorem 4.23. Examples. The image of an opeo set under a continuous mapping is not nea:ssarily open. A simple counterexample is. a constant function which maps all of S onto a single point in R 1. Similarly, the image of a closed set under a continuous mapping need not be closed. For clWTIple, the reaI·valued function i(x) = arctan x maps Rl ooto the open Interval ( - n12. 1lf2). 4.13 FlJNCTIONS CONTINUOUS ON COMPACT SETS

The next theorem shows that a continuous/actually and inf I(X) if X is compact.

- Y)

S - /-1(y).

Theorem 4.18. Let [': S -+ R be a rem-volued function from a metric space S to Euclidean space R. Assume that f is continuous on a compact subset X 0/ S. Then there exist points p and q in X such that
/(p)
NOTE.

inf/(X)

and

I(q)

sup f(X).

Since f(p) s /(x) s I(q) for all x in X, the numbers f(p) and /(q) are called, respectively, the absolute or global minimum and maximum values of /on x. Proof Theorem 4.25 shows that /(X) is a closed and bounded subset of R. Let m = inf/(X). Then m is adherent to /(X) and, sincef(X) is closed, m ef(X). Therefore m = /(p) for somep In X. SimilarJy,/(q) = supf(X) for some qin X. Theorem 4.29. Let ] : S .... T be o/unctionjrom one metric space (S, dJ to another (T. dT). Assume that I is one-to-one on S, so thas the inverse function exists. I/ S is compact and if I is continuous on S,. then /-1 is continuous on f(S).

The next theorem shows that the continuous image of a compact set is compact. This is another global property of continuous functions.
Tlu!orem 4.25. Let f: S -+ T be afunclion/rom one metric space (S, diJ to another (T, dr)_ If [is continuous on a compact subset X 01 S, then the image leX) is a compact subset 01 T: in particular, /(X) is closed and bounded in T.

r:

Proof By Theorem 4.24 (applied to f-I) we need only show that for every dosed set X in S the image f(X) is closed in T. (Note that I(X) is the inverse image of

Def.4.30 Since X is dosed and S is compact. X is compact (by Theorem sof(X) is compact (by Theorem 4.25) and bencef(X) is closed (by Theorem This completes the proof.

x under f-1.)

3.39), 3.38).

Ti;u1'Dll4.31. Let / be defined on an interval S in R, Assume that f is continuous at a paint c in S and that I(c) "'"O. Then there is D I-ball B(c; 5) such that /(x) has the same sign as f(c) in B(c~ ~ ('I s.

Rumple. This example shows that compactness or S is an essential part of Theorem 4.29. Let S = [0, l) with the usuaJ metric of R 1 and consider the complex-valued function jdefined by j(:x) 1r21<1", for 0 :!> x -e 1. This is D. OJlC.t<Hl1lC continuous mapping of the balf-opcn interval [0, I) onto the unit circle Izi = 1 in the complex plane. However.j-l is not continuous at the pointj(O). For example, if X8 = 1 - lin. the sequence {j(x.)} converges to 1(0) but {x,,} docs not CODverp in S. 4.14 TOPOLOGICAL MAPPINGS (HOMEOMORPIDSMS)

Proof

Assume f(c)

>

O. For every

>

0 there is a ~ whenever x

>
E

0 such that B(c;.?) n S. Then we have

f(c) - s </(x)
Take the 8 corresponding:

<J(c)

+ I!

to s = /(c)/2

(this t: is positive). whenever x

f"\ E

!/(c) < f(x) < t/(e)

B(c; 6) n S,

so I(x) has the same sign, as ftc) in B(c; 6) except that we take e = -if(c).

S. The proof is similar if/(c)

< O.

f :S -+ T be Q Junction from one metric space (S, ds) to another (T, dT). Assume also that J is one-to-one on S, $0 that the inverse function /-1 exists. 1/1 is contilr1Jow on S and if is Continu0U3 on f(S}. thenf is called Q topaiogicm mapping or a homeonwrphi8m. and the metric spaces (8, ds) and (/(S), dT) are said to be hcmeomorphic.
lk./Utilio. 4.30. Let

r:'

TMorem 4.32 (Boi{lllllJ). Let f be real-valued and continuous on a compact interval b] in R, and suppose that /«(1) and /(b) have opposite sign.r; that is, assume f(a)f(b) < O. Then there is at least one poi", c in the open interval (0, b) such that f(c) = O.

[D,

Proof

For definiteness,

assumef(a)
=

>0

andf(b) and

< O. Let
J(x) ~ O}.

then so is 1-1. Theorem 4.23 shows that a homeomorphism maps open subsets of S onto open subsets of f(8). It also maps closed subsets of S onto closed subsets of f(S). A property of a set which remains invariant under every topological mapping is called a topologicm property. Thus the properties of being open, closed. Or compact are topological properties. An important example of a homeomorphism is an isometry. This is a function -+ T which is one-to-one on S and which preserves the metric; that is.

If f is a homeomorphism,

{x:xe[o,b]

t .s

Then A is nonempty since a E A, and A is bounded above by b. Let e = sup A. Then a < e < h. We will prove thatf(c) = O. Iff(c) "" 0, there is a l-ball B(c; iI) in whichfhas the same sign as/(c). If fCc) > O. there are points x > cat whichf(x) > O. contradicting the definition of c. If f(c) < 0, then e - iI/2 is an upper bound for A, again contradicting the definition of c. Therefore we must havef(cl = O. From Bolzano's theorem we can easily deduce the intermediate

dr{J(x).!(y))

d,(x, y)

value theorem

for continuous functicns.

TlMwrem 4.33. Assume / is real-valued and continuow 011 a compact interval S in R. Suppose there are two points IX < PinS such that f(a.) "" fOO. Then f takes er;ery value betweenl(rt.) andf(fl) in the interval (a. fJ). Proof Let k be a number betweenf(ll) and/(ft) and apply Bolzano's the function 9 defined on [a. P] by the equation g(x) = f(x) - k: theorem to

for all pojnts x and y in S. If there is an isometry from (S, ds) to (I(S), dT} the two metric spaces are called isometric. Topological mappings are particularly important in the theory of space curves. For example, a simple arc is tbe topological image of an interval, and a simple closed curve is the topological image of a circle. 4.15 BOLZANO'S THEOREM

This section is devoted to a famous theorem of Bolzano which concerns a global property of real-valued functions continuous on compact intervals [a, b] in R. If the graph off lies above the z-axis at Q and below the x-axis at b, Bolzano's theorem asserts that the graph must cross the axis somewhere in between. Our proof will be based on a local property of continuous functions. known as the sign-preserving properly.

The intermediate value theorem, together with Theorem 4.28, implies that the continuous image of a compact interval S under a real-valued function is another compact interval, namely,

[inf/(S), supf(S»). (If/is constant on S. this will be a degenerate interval.) The next section extends this property to the more general setting of metric spaces.

87

4.16 CONNECrEDmss This section describes the concept of connectedness and its relation to continuity. Since A and Bare nonempty.ftakes both values 0 and I, sofis not constant. Also.fis continuous on S because the inverse image of every open subset of {O. I} is open in S. Next we show that the continuous image of a connected

Dtjilfitioa 4.34. A metric space S is coiled tfisco,u,ecled if S = A V B wher .A. and B are disjoint nonempty open sets in S. We call S connected if it Lr not ~ connected.
NOlI!. A subset X of a metric space S is called connected metric subspace of S. it is a connected metric space.

set is connected.

if when regarded as a

Tllure", 4.37. Let I: S --+ M be a function from (J metric space S to another metric space M. Let X be a connected subset 0/ S. II f is continuous on X, then f{X) is a connected subset of M. Let g be a two-valued function on f(X). We willsbow that g is constant. Consider the composite function h defined on X by the equation h(x) = g(/(x»). Then h is continuous on X and can only take the values 0 and 1, so h is two-valued on X. Since X is connected, h is constant on X and this implies that g is constant onf(X). Thereforef(X) is connected.
Eumple. Since an interval X in R 1 is connected. every continuous image !(X) is connected. If f has real values, the image f(X) is another interval. JUhas values in R", the image f(X) is called a curve in R". Thus, every curve in a- is connected. As a corollary theorem. of Theorem

Rumples

1. The metric: space S = R - to} with the-usual Euclidean metric is discoonected. . it is ~ union of two disjoint DODempty open sets. the positive J\\'aJ numbers and. ncpti\'C n!aI numbers. 2. Ew;ry open interval. in R is 00dnCCted. ql.1CDl:e of Theorem 3.11. 3.

u::

Proof

This was

proved in

Sa:tion 3.4 as a consc-

of rational numbers., regarded as a metric subspace of Euclidean space R I, tS disconoected. In fact, Q = A V B. where A consists of aU rational numbers < and B of an rational numbers Similarly. every ball in Q is disconnected.

:n-.set Q
J2

>"/i

4. Every metric space S contains nonempty ~ the set {P} is 00Iltleded. late coaneetedness To function.

subsets. In fact for ~~-.. . S • ........ p m introduce the concept of a two-valued

4.37 we have the following extension of Bolzano's

n:

with continuity

we

aaiti to be two-uaJued on S if I(s) ~

4.35. A real-fl4lued fiore/ion

{O, I}.

I which

is continuous on a metric space S is

Theorem 4.38 (IlfIt:rmediate-wdue 'Monm for ruJ ctIlftimtoIu I~). Let / be real-valued and continuous on a connected subset S oIR*. Iff takes on two different talues in S, say a and b, then for each real c between a and b there exists Q point x il'l S such that I(x) = c. Proof. The image /(S) is a connected subset of Ri. Hence, I(s) is an interval containing a and b (see Exercise 4.38). If some value c between a and h were not inf(S). then/(S) would be disconnected.
4.17 COMPONENTS OF A METRIC SPACE

I.D other words, a two-val~ function is a continuous function whose only possible ,,~ are 0 and I. This can be regarded as a continuous function from S to the metric space T = {Of 1}, where T has ~ discrete metric. We recall that every subset of a discrete metric space T is both open and closed in T.
T~em 4.36. A metric space S is connected Iun.ctwn on S rs constant. Proof

if.

and only

if. every two-valued

This section shows that every metric space S can be expressed in a unique way as a union of connected "pieces" caned components. First we prove the following:

Ass~ S is connected and letlbe a two-valued function on S. We must ~how that 11s constant. Let A = 1-1({O}) and B = I-I({I}) be tbe inverse l~ages of th~ subsets {OJ and {I}. Since {OJ and {Il are open subsets of the discrete metne space {O, I}, both A and B are open in S. Hence. S = .A. V B where A and B are dis~oint open sets. But since S is connected, either A is emptr and B = S, or else B IS empt~ and A = S. In either case. I is constant on S. Conv.e?:ly. assume that S IS disconnected, so that S = A u B, where A and B are disJomt nonempty open subsets of S. We will exhibit a two-valued fiun..... ' ",.. on on S w bih IS not constant. c Let

connected.

TMonm 4.39. Let F be a collection 0/ connected subsets 01 a metric space S such that the intersection T = A is not empty. Then the Wliol'l U = A is

n,..".

UA""

If x E U, then x IE' A for some A in F. Since A is connected, and. since t E A,/(x) = I(t).

Proof Since T #- 0, there is some t in T. Let f be a two-valued function on U. We will show that/is constant on U by showing thatf(x) = f(t) for all x in U.
/ is constant on A

f(x) = {~

ifxe.A.. ifxeB.

Every point x in a metric space S belongs to at least one connected subset of S, namely {x}. By Theorem 4.39, the union of all the connected subsets which contain x is also connected, We call this union a component of S, and we denote it by U(x). Thus. U(x) is the maximal connected subset of S wbich contains x.

•
1'IIeon", 4.40. Every point of a metric SJKI" S belonga 10 a uniquely determined component of S. in otlrer wor•• tire components of S form a coIlectkm 0/ diajoinl sets whole union is S. Proof. Two distinct components cannot contain a point 4.39) their union would be a larger connected set con~g
4.18 ARCWlSE CONNECl'EllNESS
X'

Oef.·US The next theorem Theorem 4.41. relates arewise connectedness with connectedness.

89

Every arcwise connected set S in R8 is connected.

otherw.ise (by 1Deon:m

x.

Proof. Let 9 be two-valued on S. We will prove that 9 is constant on S. Choose a point a in S. If x € S, join a to x by an arc r lying in S. Since r is connected, g is constant on I' so g(x) = g(a). But since x is an arbitrary point of S, this shows that 9 is constant on S, so S is connected. We have already noted that there are connected sets which are not arcwise connected. However, the concepts are equivalent for open sets.

This section describes

a special property. called arcwi.se CONtectedness, whicla is possessed by some (but not aU) eonneeted sets in Euclidean space R".

~ 1.11: set S i3 caIJed arcwi.se connecled if for any two points a and binS there IS a COIItinuous junCtiOll f : [0, 1] - S such tAat 1(0) = •
NOTE.

!".-

Theonm 4.13. Every open connected set in R" is arewise connected.

Proof. Let S be an open connected set in R~ and assume XES. We will show that " can be joined to every point y in S by an arc lying in S. Let A denote that subset of S which can be so joined to x, and let B = S - A. Then S = A v H, where A and H are disjoint. We will show that A and B are both open in R". Assume that a E A and join a to x by an are, say r, lying in S. Since It E S and S is open, there is an n-ball B(a) £ S. Every y in B(a) can be joined to • by a line segment (in S) and thence to x by r. Thus yEA if Y E B(a). That is, B(a) £ A. and hence A is open. To see that B is also open, assume that b E H. Then there is an n-ball H(b) s; S, since S is open. But if a point y in H(b) could be joined to x by an arc, say I", lying in S, the point b itself could also be so joined by first joining b to y (by a line segment in B(b» and then using I". But since b ¢ A. no point of B(b) can be in A. That is. B(b) s;; .8, so .8 is open. Therefore we have a decomposition S A u H, where A and H are disjoint open sets in R.". MQreover, A is not empty since x E A. Since S is connected, it follows that B must be empty. so S A. Now A is clearly arcwise connected, because any two of its points can be suitably joined by first joining each of them to x, Therefore. Sis arcwise connected and the proof is complete.
NOTE.. A path f: [0, 1] -+ S is said to be polygonal if the image of [0, under f is the union of a finite number of line segments. The same argument used to prove Theorem 4.43 also shows that every open connected set in R" is polygonally connected. That is, every pair of points. in the set can be joined by a polygonal arc

and

f(l)

b.

o ::;:

Such a ~unctiOft is ~ ~ P!'th from a to b. If 1(0) ¢ f(1l. the image of [0, I] ~ f IS called au arc JOIIlID8 • and b. ThUll, Sis arcwise CODIlCCted if every pur of distinct points in S can be joined by an arc lying in S.· An:wise connected sets are ~. ~Jed pathwtre co1lllected. If f(t) = tb + (J - t)a for t ::;:, the curve Jom:tng • and b is called a line segment. I
SIa II s
1. Efti)'
COIl'\'Q R" is an:wi!Je COI1JXded. sia:JI% tbc Iioe qrneot joining two points of sudl a set lies m the set. In parIicuIac, every n-baJJ is arcwiilC connected.

:a~

1.. The

set

in Fig. 4.4 (a 1DIion

or two tangent

closed disb) is arcwise connected.

IJ

Fjgure4.4 3. The set~n Fig. 4.5 C?nsists.of.those points on the curve described by y = sin (llx), < x _ 1, along With the points on the horizontal segment -I .s x s O. This set IS connected but not arcwise connected (Exercise 4.46).

lying in the set.

TlleoreM 1.14. Every open set Sin R" can be expressed in one and (}Illy one way as a countable disjoint union of open connected sets.
Proof By Theorem 4.40, the components of S form a collection of disjoint sets whose union is S. Each component T of S is open, because if x € T then there is an n-baU H(x) contained in S. Since B{x) is connected, B{x) s T, so T is open. By the Lindelof theorem (Theorem 3.28), the components of S fonn a countable collection, and by Theorem 4.40 the decomposition into components is unique.

Fq:ure4.S

lkji.tJititnt 4.45. A set in R" Is called 0 region if it is the union of on open connected set with some, none, ot all its boundary points. If none of the boundary points are

Def.4M included, the reqion is called an open region. the region is called a closed region. Some authors complex plane.
NOT!':.

Th. 4.47

UDif_ O>ntiauity aod Compact Sets

'.II

1/ all

the boundary points are included,

An instructive formly continuous 4.2.0 UNIFORM

exercise is to show that the function on R I. CONTINtJlTY AND COMPACT SETS

in Example

2 is not uni-

use the term domain instead of open region, especially in the Uniform continuity on a set A implies continuity this.) The converse is also true if A is compact. on A. (The reader should verify

4.19 UNIFORM

CONTINUITY

Suppose/is defined on a metric space (S, ds), with values·in another metric space (T. dr), and assume thatfi~ continuous on a subset A of S. Then, given any point p in A and any e > 0, there is a c5 > 0 (depending on p and on e) such that, if x E A, then d"{f(x), f(p»)

Ti4!orem 4.17 (Heine), Let f :S ....T be a function from one metric space (S, ds) to another (T, dT). Let A be a compact subset oj Sand assume that f is continuous
an A. Proof Then

is uniformly continuous on A.

<

whenever dS<x, p)

< 0.

Let

> 0 be given. Then each point a in A has associated with it a ball

depending dr<f(x),f(a» on

BS<a; r),

with,

a,

such that whenever

In general we cannot expect that for a fixed e the same value of (i will serve equaUy well for every point p in A. This might happen, however. When it does, the function is called uniformly continuous on A. lJe./iltitinll 4.46. Let f : S -+ Theafunction/rom one metric space (S, ds) toano/her (T, dr)· Then/is said to be uniformly continuous on a subset A of S if the follOWing condition holds. For every e > 0 there exists and p E A then dr<f(x)./(p»)
Q~

<;

x E BS<a; r)

f"'l

A.

Consider the collection of balls Bsf"a; ,/2) each with radius r12. These cover A and, since A is compact, a finite number of them also cover A, say

>

0 (depending only on 6) such that whenever dS<x, p)

if

x €A In any ball of twice the radius, B(al; 'I)' we have (6)

<

<

0.

To emphasize the difference between continuity on A and uniform continuity on A we consider the following examples of real-valued functions.
Examples

d~f(x), !(at) < ;

Let 5 be the smallest of the numbers rtl2, ... , r.../2. We shall show that this b works in the definition of uniform continuity. For this purpose, consider two points of A, say x and p with ds(x, p) < b. By the above discussion there is some ball BS<a~;rJ2) containing x, so

1. Let lex) = llx for x ;:. 0 and take A ~ (0, 1 J. This function is continuous on A but not uniformly continuous on A. To prove this, let It = 10, and suppose we could find a 0, 0 < J < 1, to satisfy the condition of the definition. Taking x = 0, p = 0/11. we obtain Ix - pi < J and

d.(f(x), f(aJ) By the triangle inequality we have

<

2:'
+ '" 2
r, -i

I/(x) -/(p)1

= ~I -

olio

! = !Q ;;. 10.
I(p) I >

Hence, for these two points we would always have I/(x) the definition of uniform continuity. 2. Let I(x) = Xl if x E R 1 and take A = (0, 1] as above. continuous on A. To prove this. observe that
I/(x) - l(p)1 ~

to, contradicting
Hence, p

dsCp, a~)
€

dsCp, x)

ds(x, at) < ii

+ 'k '2
< £/2.
6

rk•
Using the

This function is uniformly 21x - pI.

Bsf"aA; r~) ('\ S, so we also have d~/(p),f(al))

once more we find

Ix" - p21

I(x -

pXx

+ p)1 <

triangle inequality

I/(x) - f(P)1 < U. Hence. if £ is given, we need only take = 1l/2 to guarantee that If(x) - 1(P)1 < e for every pair x, p with Ix - pI < J. This shows that/is uniformly continuous on A.

If Ix

pi < 0, then

d,,{!(x).!{p»)
This completes

d.(f(x).!(a,,))

dr<f(asJ,f(p»

<

+2=

11

s,

the proof.

9Z
4.21 FIXElJ..POINT THEOREM FOR CONTRACTIONS

n......

Del. 4.49

93

a function from a metric space (8, d) into itself. A point p in 8 is called ajixed point of/if /(p) = p. The function/is called a contraction of 8 if there is a positive number IX < I (called a cantraction constant), such that Let / : d(/(x), /(y» ~ rJ.d(x, y)

s .....be S

Let / be defined 011 an interval (0, b). Assume C E [0, b). If J(x) -> A as x -+ c through values greater than e, we say that A is the righthand limit off at c and we indicate this by writing lim f(x)

A.

for all x, y in S.

(1)

Clearly, a contraction of any metric space is uniformly continuous on S.

TlutorelII 4.48 (Fixl!ll-poilft tMOl'I!lII). A contraction/ 0/ a complete metric space 8 has a unique fixed point p, Proof. If p and p' are two fixed points, (7) implies d(p, p') .$ IXd(p, p'), so so d(p, p') 0 and p = p'. Hence/has at most one fixed point. To prove it has one, take any point x in S and consider the sequence of iterates:

The righthand limit A is also denoted by /(c+). In the s, <> terminology this means that for every .& > 0 there is a {) > 0 such that

1/(;-=)~ /(c+)1 < It

whenever c < x < c

/j

< h.

x, lex). /(/(x)),
Po x,
p,,+ I /(P,.),

...
n = 0, 1,2, ...

Note that f need not be defined at the point c itself. If/is defined at c and if f(c+) = /(c), we say that/is continuousfrom the right at c. .. . Lefthand limits and continuity from the left at e are similarly defined If c E (0, b]. . If a < e < b, thenfis continuous at c if, and only If,
/(c) the following conditions

That is, define a sequence {p.} inductively as follows:

/(c+)

J(c-}.

We say c is a discontinuity of / if / is not continuous at c. In this case one of is satisfied: a) Either /(c+) or f(c-) does not exist. b) Both/(c+) and f(c-) exist but have different values.
c) 8othf(c+)

We will prove that {pft} converges to a fixed point off First we show that {p.} is a Cauchy sequence. From (7) we have
d(Pw+l'P.)

d{J(p.).!(Pft-'».$

I1.d(P."Pn_I)'

andf{c-)

existandf(c+)

=/Cc-)

¥-/(c).

so, by induction, we find

d(p 11+ 1, p,.) ~ a." d(p
1,

Po) =

crt,.,
a!"

where c = d(pl> po)· Using the triangle inequality we find, for m > n,
• -1

In case (c), the point c is called a removable discontinuity, since the discontinuity could be removed by redefiningf at c to have the value /(c +) .= /(c.-): In cases (a) and (bl, we call c an irremovable discontinuity because the discontinuity cannot be removed by redefining/at e.

d(p"" PII)

L d(PUh
~~.

~-1

Pt) ::::; e

t-8

rt' L a: c--I-IX

c < --ci'.
I-a:

~JUritWlf
both exist

at some

4.49. Let / be definNl on a closed internal [a. b]. 1/ f(c+) ami /(c-) interior point c, then: is called the lefthandjump o/fat
is called the jump off at c. three numbers is different from 0, then c is called a jump dis-

Since rt' -+ 0 as n -> co, this inequality shows that {p.} is a Cauchy sequence. But 8 is complete so there is a point P in S such that p. -+ p. By continuity of J.
f(p) = /(lim

a) /(c) - /(c-)
b) /(c+) c) /(c+)

c,
0.1

- !(c) is called the righthand jump of/ - /(c-)

c,

"-'liD

p,,\ J

lim f(p,,)
.-+eJ;,lo

"-,110

lim P..+ 1

p,

so P is a fixed point off This completes the proof.

If anyone
continuity

0/ these
off.

Many important existence theorems in analysis are easy consequences 'Of the fixed point theorem. Examples are given in Exercises 7.36 and 7.37. Reference 4.4 gives appflcations to numerical analysis.
4.22 DISCON11Nl.lmES OF REAL-VALUED FUNCIlONS

For the endpoints a and b, only one-sided jumps are considered, tbe righthand jump at 0, f(o +) /(a). and the lefthand jump at bJ(h) - f(b- ).
Eumpies 1. The function [defined by /(x) = x/lxl if x #; 0,[(0) = A, bas a jump discontinuity at 0, regardless ofthe value of A. Here [{O+ ) t andf(O-) I. (See Fig. 4.6.) 2. The functi(ln / defined by /(x) = i if x of- 0, /(0) continuity at O. In this case [(0+ ) = /(0-) = 1.
=

The rest of this chapter is devoted to special properties of real-valued functions defined on subintervals of R.

0, has a removable jump dis-

LimiI5 and Contillllity

Def.4.50

Th. 4.53

Monotonic Functions

We shall prove that functions which are monotonic on compact intervals always have finite right- and lefthand limits. Hence their discontinuities (if any) must be jump discontinuities. Theorem 4.51. If[is increasinq on [a. b], then/(c+) c in (a, b) and we haoe [(c-) sf(c) S/(c+).
A t the endpoints we have
FJgUre4.7

and/(c-)

both exist for each

f(a) S /(a+)

and

/(b-)

s [{b).

l. The function I defined by /(x) = l/x if x of;. 0, [(0) = A, hal; an irremovable discontinuity at O. In this case neither /(0+) nor /(0-) exists. (See Fig. 4.7.)
4. Thefunclionfdefined by lex) =, sin (l/x) if x '# 0./(0) '" A, has an irremovable discontinuity at 0 since neither 1(0+) nor /(0-) exists. (See Fig, 4.8.)

Proof. Let A = {I(x): a < x < c}. Since I is increasing. this set is bounded above by f(c). Let 0: = sup A. Then a; S f(c) and we shall prove that /(c-) exists and equals a. To do this we must show that for every I; > 0 there is' a (j > 0 such that c-6<x<c implies
I/(x) a;1< e.

5. The function f defined by f(x) ... x sin (J/x) if x ,t. 0, /(0) = I, has a removable jump discontinuity at 0, sinre[(O+) /(0-) O. (See Fig. 4.9.)

But since Cl = sup A, there is an element/(xt) of A such that a - t < f(xd $; 0:. Since / is increasing, for every x in (Xl' c) we also have It - e < [(x) $; 0:, and hence I/(x) ~ al < e. Therefore the number {j = c - XI has the required property. (The proof that [(c+) exists and is ';;;!f(c) is similar, and only trivial modifications are needed for the endpoints.) There is, of course. a corresponding theorem for decreasing functions which the reader can formulate for himself, Theorem 4.52. Let [be strictly increasing on a set S in R. Then strictly increasing on/(S).

r: exists

and is

FiKIIRl 4.8

Figure 4.9

4.11 MONOTONIC Definition 4.50. Let

FUNCTIONS

Since I is strictly increasing it is one-to-one on S, so exists. To see is strictly increasing, let j-, < Y2 be two points in/(S) and let Xt = f-I(y,), X2 = [-I(Y2)' We cannot have XI ';;;! Xl' for then we would also have y, ~ Y2' The only alternative is
Proof

tbat[-t

r:

! said to be increasing is
in S,

I be

a real-valued function defined on a subset S 01 R. Then (or nondecreasinqs on S if /()r every pair of points x and y implies

and this means that

f -1 is strictly increasing.

x<y

I{x) $;J(y).

Theorem 4.52, together with Theorem 4.29, now gives us: Theonm 4.53. Let I be strictly increasing and continuous On a compact interval [a. bJ. Then 1-1 is continuous and strictly increasing on the interval (/(a),f(b)].
NOTE.

II x :: y impli~s.r.(x) < J~ then]' is said to be strictly increasing on S. (Decreasing y), functions are similarly defined.) A function is called monotonic on S if it is increasing on S or decreasing on S.

. If [is an increasing function, then -f IS a decreasing function. Because of this simple fact, in many situations involving monotonic functions it sufficesto consider only the case of increasing functions.

Theorem 4.53 tells us that a continuous, strictly increasing function is a topological mapping. Conversely, every topological mapping of an interval [a, b] onto an interval [c, dJ must be a strictly monotonic function. The verification of this fact will be an instructive exercise for the reader (Exercise 4.62).

EXERCISES Limits

Now consider the functions/defined x2 _ y2 a) [(:r,y) = --Xl

on RZ as follows: if (x, y) '" (0,0).1(0,0) "" o.

or sequences
->

y2 (Xy)l (x _

4.1 Prove each a) z~ 0

b) If z~ d) If a~ =
1,

or the following
if

statements about sequences in C. if

1=1

« 1; {zoO} diverges

1=1

> 1.

b)

lex,

.1') ..

(xy)2'+
X

y)2

if (x, y) "# (0, 0), /(O, 0) If X sin (l/y) 0,/(0, y) ~ }'.

= 0,

0 and if {c,,} is bounded, then {c.z,,} -> O.

c) f(x, y) ~ !_ sin (xy) d) f(x, y) = { 0

Qtt+l _

""

c) z~/lI! ..... 0

for every complex z, that a•......QI (

.../;? + 2 - n, then Ow ...... O. 4.2 If au (an I + Q.)/2 for all 11 ~ I, show a.+! = Ha" - a. + Il.

(x

.1') sin (l/x)

if x '" 0 and y "# 0,

if x = 0 or y "" O.

'2a1,)/3. Hint.

4.3 If 0 -c Xl <: 1 and if X0+1 = 1 - ~~ for all n ;?: 1, prove that {x.J is a decreasing sequence with Iirnit O. Prove also that x~+tlx" .....1-. 4.4 Two sequences of positive integers {o"l and {b.} are defined recursively by taking til = b, = 1 and equating rational and irrationaJ parts in the equation a. Prove that Q~ and that 2b,./a" .....

e) f(x, y)

Sinx = tan x
cos3 x

u.;
'\:2

+
=

b•../2 =. (Q,,_ 1

+ b"_l

.../2)2

for n ~ 2;

I for II ?: 2. Deduce that Q,./h•...... through values -c ,/1..

.../2

In each of the pr«:eding examples, determine whether the following limits exist and evaluate those limits that do exist: lim f(x, y). lim [lim [ix, y)] ; lim pim/(x, y)] ;
x-rO r-()
y-+O K-O

sin y Ian .I'

if Ian x '" tan

.1',

if Ian x ~ tan y.

through values

>.../i,

(r,71-(O,O)

4.12 If x

[0, I) prove that the following limit exists, lim [lim cos~· (m! 1fX)] ,
m-l>!j;()

4.S A real sequence {x,,} satisfies 7x,,+1 = x! + 6 for n ?: 1. If Xl = t, prove that the sequence increases and find its limit. What happens if Xl = Of if Xl = i1

tt ...... ro

4.6 If

con verges.

la.l

<: 2 and

la..+2

an+ll

:$

ifa;+, - a~1 for

all

n~

1,

prove

and that its value is 0 or I, according to whether x is irrational or rational. COIItinuity or real-valued. fUDCtions 4.13 Let f be continuous on [0, b] and let f{x) '" 0 when x is rational. Prove that f(x) =: 0 for every x in [a, b J. 4.14 Let f be continuous: at the point a (Ob 0l' ... , a.. in RO.. Keep 02' 03' ... , o. ) fixed and define a new function 9 of one real variable by the equation g(x)

that

{aN}

4.7 In a metric space (S, d), assume that x •......x and Y. _,. y, Prove that d(l(", Y.) ..... dix, y). 4.8 Prove that in a compact metric space (8, d), e'llety sequence in S has a subsequence which converges in S. This property also implies that S is compact but you arenot required to prove this. (For a proof see either Reference 4.2 or 4.3.) 4.9 Let A be a subset of a metric space S. If A is complete, prove that A is dosed. that the converse also holds if S is complete. Limits of functions
NOTE.

Prove

= f(x,

02·' .••

Q~).

Prove that 9 is continuous at the point x al' (This is ~metimes stated as follows: A continuous function of 11 variables is cominuous in each oariable separately.) 4.15 Show by an example that the converse of the statement in Exercise 4.14 is not true in general. 4.16 Let/, g, and h be defined on [0, I] as follows: f(x) g(x) h(x) 0, whenever x is irrational;
/(x) = 1 and g(x) = x,

In Exercises 4.10 through 4.28, all functions are real-valued.

(0, b)

4.10 Let [be defined on an open interval statements a) lim

and assume x

r;:

(a, b). Consider the two

Prove that (a) always implies (b), and give an example in which (b) holds but (a) does not. 4.11 Let f be defined
Oil

"-0

If(x +

h) - f(x)t

'" 0;

b) lim tf(x

.-0

h) - I(x

- illt

o.

whenever x is rational;

h(x) ~ I/n, if x is the rational number min (in lowest terms); h(O} ... 1.

R z. If lim f(x, y) L

Prove that / is not continuous anywhere in [0, 1]. that 9 is continuous only at x that h is continuous only at the irrational points in [0, I J.

0, and

and if the one-dimensional limits lim.._./(x, y) and lim?_./(x, y) both exist, prove that
... _111

lim [lim/(x, .1')] = lim [lim/ex, y)J "" L. .v_v ,x_d'

)t-t'b

4.17 For each x in (0, I}, let I(x) = x if x is rational, and let f(x) = I - x if x is irrational. Prove that: a) f(f{x» = x for all x in [0, 1]. b) /(x) + /(l - x) = 1 for all x in [0, 1).

c) I is continuous only at the point x

d)

I assumes every e) I(x + yJ - f(x)

value between

° and

1.
I.

a) Prove that/cannot

be continuous on [0, I}. has the above property.

I(y) is rational for aU x and y in

[0, I].

b) Construct a function Iwhich

I is continuous.

4.18 Let fbe defined on R. and assume that there exists at least one point Xo in R at which

SUPPOsealso that, for every x and y in R, I satisfies the equation I{x

y) = f(x)

+ fly).
=

c) Prove that any function with this property has infinitely many discontinuities on [0, 1). 4.28 In each case, give an example of a function J. continuous on S and such that I(S) = T, or else explain why there can be no such I;
a) S = (0, 1),

Prove that there exists a constant a such that f(x)

ax for all x.

T T T
= =

(0, I]. the set of rational numbers. {O, II.

x

4.19 Letlbeoontinuous on [a. bland defineg as follows:g{a) = f(a)and. for a < x ~ V, let g(x) be the maximum value of fin the subinterval [a, x]. Show that g is continuous' on

b) S = (0, I),
c) S = R',

T = (0, 1) v (l, 2).

(a, b].

d) S = [0, I] v [2. 3].

.. ,I,. bem real-valued functions defined on a set Sin R". Assume that each h is continuous at the point a s. Define a new function I as follows: For each x in S, /(x) is the largest of the m numbers Ii(x), ... ,/,..(x). Discuss the continuity of I at a.
4.20 LetJi,

or

e) S "" {O, t] x (0, I], f) S "" [0, 1] x [0, 1], g) S = (0, 1) x (0, I),

T= R2. T = (0, 1) T= R2. . T is a function from one metric (0, I).

4.21 Let /: S --> R be continuous on an open set S in R", assume that , E S, and assume thatl(P) > O. Prove that there is an If-ball B(P; r) such that.f(x) > 0 for every x in the ball. 4.22 Let f be defined and continuous on a dosed set S in R. Let
A {x:xeS

and

/(x)

= OJ.

Continuity in PJetric spaces In Exercises 4.29 through 4.33, we assume that space (5, ds) to another (T, dT).
4.29 /-'l(int B) S int/-I(B) 4.30 Prove thatlis

r: S

-->

Prove that A is. a closed subset of R. 4.23 Given a function

Prove that lis continuous on S if, and only if, for every subset B of T.

I:

R ..... R, define two sets A and Bin R2 as follows:

A = {(x. y); y

< /(x)},

B = {(x, y):.v

> /(x)}.

continuous on S if, and only if, for every subset A of S.

Prove that / is continuous on R if, and only if, both A and B are open subsets of R 2. 4.24 Let

/be

defined and bounded on a compact interval S in R. If T s S, the number

Q,(T) = sup {/(x) - I(y): x e T, yET}

4.31 Prove that .f is continuous on S if, and only if, f is ~ontinuous on every compact subset of S. Hint. If X" ..... pin S, the set {p, XI> X2'··· } IS compact.
4.32 A function.f:

is called the oscillation (or spa") off on T. If XES, be tbe number

m_,(x):o
...... 0+

the oscillation off at x is defined to

S --> Tis called a ctosed mapping on S if the imagel(A) is.dosed in T for every closed subset A of S. Prove that/is continuous and dosed. on S If, and only

lim Q f(B(x ;h) ,...,S).

=

Prove thaI this limit always exists and that w_,(x)

0 if, and only if. I is continuous at x.

4.25 Let I be continuous on a compact interval (a, b). Suppose that / has a local maximum at XI and a local maximum at Xl' Show that there must be a third point between XI and X2 where/has a local minimum,
NOTe. To say that I has a local maximum at .1:"1 means that there is a l-ball B(Xl) such that I(x) -s f(x I) for all x in B(x I) rv [u, b J. local minimum is similarly defined.

if, /(it) = M for every subset A of S. . 4.33 Give an example of a continuous f and a Cauchy sequence {x~} in some metnc space S for which {f(x.)} is not a Cauchy sequence in T. 4.34 Prove that the interval (- 1, I) in R I is hOH_Ieomorphie to R I. This shows that neither boundedness nor completeness is a topological property.

I(

4.35 Section 9.7 contains an example of a function t. continuous on [0, I]. with [0, I]) = [0, 1] x [0, I]. Prove that no such f can be one-to-one on [0, I ]. CoonectedDeSS
4.3(; Prove that a metric space S is disconnected.if, and only if, there is a nonempty subset

4.26 Let /be a real-valued function, continuous on [0. I), with the following property: For every real r. either there is no x in [0, I] for which [(x) = y or there is exactly one such x. Prove that/is strictly monotonic on {O, I]. 4.27 Let I be a function defmed on [0, I] with the following property: For every real number y, either there is no x in [0, I] for whichf(x) '" y or there are exactly two values of x in [0, 1) for whichf(x) = y,

A of SA",

S which is both open and closed. III S.

4.37 P~ove tlla; a metric space S is connected if, and o~\Y if. the only subsets of S which are both open and closed in S are the empty set and S Itself. .. 4.38 Prove that the only connected subsets of R are (a) the em.pty (b) sets consssung of a single point, and (e) intervals (open, closed, half-open, or infinite),

s:et,

100
4.39 Let X be a connected subset of a metric; space S. Let Y be a subset of S such that X <;; y <;; X. where X is the closure of X. Prove tha.t Y is also connected. In particular,

101

this shows that X is connected.

4.40 If x is a point in a metric space S, lei U{x) be the component of S containing x. Prove that U(x) is closed in S. 4.41 Let S be an open subset of R. By Theorem 3.11, S is the union of a countable disjoint collection of open intervals in R. Prove that each of these open intervals is a component of the metric subspace S. Explain why this does not contradict Exercise 4.40. 4.41 Given a compact set S in R"' with the following property: For every pair of points a and h in S and for every II > 0 there exists II. finite set of points {so. Xt. •.• , x,,} in S with Xo = a and xN = b such that

4.55 Let f :S -+ T be a function from a metric space S to another metric space T. Assume I is uniformly continuous on a subset A of S and that T is com~ete. prove that there is a unique extension of /to A which is uniformly rontinuous on A. 4.56 In a metric space (S, d). let A be a nonempty subset of S. Define a function

fA : S

.....R by the equation

I,b}
for each x in S. The number
b) Prove that

= inf {d(x.y):

yEA}

I.. {x)
XES

is called the distance from x to A.

and

a) Prove that IA is uniformly continuous on S.

= {x;

I..(x)

= O}.

for k
Prove or disprove; S is connected.

= I, 2•...

II.

4.57 In a metric space (S, d). let A and B be disjoint dosed subsets of S. Prove that there exist dildoint open subsets U and Y of S such that A <;; U and B s V. Hint. Let g(x) = h,{x) - IJ.x), in the notation of Exercise 4.56, and consider 9- I{ - 00. 0) and
g~](O,

+ 00).

4.43 Prove that a metric space S is connected if, and only if, every nonempty proper subset of S has a nonempty boundary.
4.44 Prove that every convex subset of R" is connected.

DiscoIIliI'lllities

4.S8 Locate and classify the discontinuities of the functions! defined on R [ by the following equations: a) I(x) = (sin x)lx
b) I(x) = ell"

4A5 Given a function f: R't .... R"' which is one-to-one and continuous on R". IT A is open aDd disconoectl:d in R-, prove that f(A) is open and disconnected in f(R"). 4.46 LetA = {(x.y):O -c x s 1, y = sin I/x}. B = {(x,y):y = 0, -I s x sO}, and let S = A v B. Prove that S is connected but not arcwise ooonected. (See Fig. 4.5. Section 4.18,) 4.47 Let F = {Fl' Fl •... } be a countable colle<;tion of connected compact sets in R" such that FH 1 s Fk for each Ie 2:; 1. Prove that the intcnCction t FA:is connected and closed.

if x ". O./(O) if x
'f: 0.[(0)

c) f(x)

= ellx

+ sin

(l/x)

if x ". 0, /(0)

= o. = o. = o.

n.i..

d) f(x) = I/O ell"} if x ". O,{(O} = O. • 4.59 Locate the points in R:.l at wbidt each of the functions in Exercise 4.11 is not con-

tinuous. Moootooie flllPctioos 4.60 Let/be defined in the open interval (0, b) and assume that for each interior point x of (a. b) there exists a I-ball B(x) in which / is increasing. Prove that! is an increasing function throughout (0, b). 4.61 Let/be continuous on a compact interval {a. b] and assume that/does not hav~ a local maximum or II. local minimum at any interior point. (See the NOTE following Exercise 4.25.) Prove that I must be monotonic on [a. b]. 4.61 IfI is one-to-one and continuous on [a, b]. prove that! must be strictly monotonic on Ia. b}. That is, prove that every topological mapping of [a. b] onto an interval [e, d] must be strictly monotonic. 4.63 Let I be an increasing function defined on [a. bland the interior such that a -c XI < X2 < ... < x~ < b.
a} Show that

4.48 Let Sbe an open connected sci in R". Let Tbea component ofR« R- - T is connected. 4.4' Let (S, ti) be a connected metric space which is not bounded. a in S and every r :> 0, the set {z : di«, a) = r} is aonemply.
Uniform aJDtinuity

S. Prove that

Prove that for e"""'Y

4.50 Prove that a function which is uniformly 4.51 If lex) =

Xl

continuous on S is also continuous on S.

00

for x in R, prove thai lis Dot uniformly continuous

R.

4.52 Assume that/is

uniformly continuous on a bounded set Sin R". Prove that/must

be bounded on S. 4.53 Let f be a function defined on a set S in R- and assume thar f(S) <;; R", Let g be defined on r(S) with value in Rl, and let .. denote the composite function defined by h(x) g[f(x)] ih E S. Ifris unifol"mly oonlmuous on Sand ifg is uniformlycontmuous on f{S). show that b is unirormly continuous 00 S. 454 Assume I : s -+ T is uniformly continuous on S, w~re Sand T are metric spaces. If {x.} is any Cauchy sequence in S. prove that {/{x,,)} is a Cauchy sequence in T. (Compare with Exercise 4.33.)

let x I,

...

Xw

be n points in

D=l
I has

(f(Xt+)

- I(Xl-)]

;;;f(b-)

- I(a+}.

b) Deduce from part (a) that the set of discontinuities of f is countable. c) Prove that that points of continuity in every open subinterval of [a, b]. defined and strictly increasing on a set S in R. such
4.64 Give an example of a functionf,

/-1 is not

continuous on!(S}.

103
10l

4.65 Let/be strictly inCtt.!lsing 00 a subset S of R. Assume that the image/(S) has one of the following properties: (a)f(S) is open; (b)/(S) is connected; (c)/(S) is dosed. Prove that/must be continuous 00 S. Metric spaces and hed JI(lOlts 4.66 Let B(S) denote the set of all real-valued functions which are defined and bounded on a nonempty set S. Ir /e B(S).let

4.'72 Assume that / satisfies

the condition

in Exercise 4.71.

If x

IE S,

let Po = x,

PH I = /(p,J, and Cit ~ d(p~, p~+ I) for II 2: O.

a) Prove that {e.) is a decreasing sequence, and let c = lim "". b) Assume there is a subsequence {PM"») which converges to a point q in S. Prove that

d(q,/(q»

d(!(q),/[J(q»)).

Deduce that q is a fixed point of / and that P •.... q.

II/II
The number

= sup 1/(x)J.

Ifl is called

the "sup norm" of f=

suGGESTED
g~ defines a metric don B(S).

REFERENCES

FOR FURTHER

STUDY

a) Prove that the formula del, g)

Irf ~

4.1 Boas,. R. P., A Prime,. of Real FflflCliortS. Carns Monograph No. 13. Wiley, New York,I960. 4.2 Gleason, A .• FlUItiamentais 0/ Abstract Analysis. Addison-Wesley. Reading, 1966,

b) Prove that the 'metric space (B(S). d) is complete. Hint. If {f.} is a Cauchy sequence in B(S), show that {f.(x)} is a Cauchy sequence of real numbers for each x in S. 4.67 Refer to Exercise 4.66 and let C(S) denote the subset or B(S) consisting of all functions continuous and bounded on S, where now S is a metric space. a) Prove that C(S} is a closed subset of B{S). b) Prove that the metric subspace C(S) is complete. 4.68 Refer to the proof of the fixed-point theorem (Theorem 4.48) for notation. a) Prove that dip, p,,) -s d(x.J(x)}:t'/(1 - a), This inequality. whk:h is useful in numerical work. provides an estimate for the distance from P.. to the fixed point p. An example is given in (b). b) Take/(x) = t(x + 2jx). S = [I. + co). Prove that/is a contraction of S with contraction constant a = t and fixed point p =..12. Form the sequence {pft} starting with x = Po = 1 and show that Ip. - ../21 .s; 2-", 4.69 Show by counterexamples that the fixed-point theorem for contractions need not hold if either (a) the underlying metric space is not complete, or (b) the contraction constant II ;:::1. 4.70 Let f :S .... S be a function from a complete metric space (S, d) into itself. Assume there is a real sequence {a~} which converges to 0 such that dU"(x), P'(y» :$ «,p(x, y) for all 11 ~ 1 and all x, y in S, where/" is the nth iterate of /; that is,
P(x) = /(.1'). f"+1(x) = /(/"(.1'»

4.3 Simmons. G. F., Introduction to Topology and Modtrm AlUllysis. McGraw-Hili, New York, 1963. 4.4 Todd, J" Survey 0/ Numerkal A1Illlys;s, McGraw·HiIl, New York, 1962.

forn2:L

Prove that/has suitable m.

a uaiqae

fu!:oo

point.

Hint, Apply the fixed-point theorem tor' for a

4.71 Let j": S .... S be a function from a metric space(S, d) into irsell such that whenever x ~ y. d(J(x),

fey»~

< di», y)

a) Prove that [has at most one fixed point, and give an example of such an /with no fixed peim, b) If S is compact, prove that f has exactly one fixed point. g(.1') d(x. /(.1'» attains its minimum on S, Hint. Show tbat

c) Give an example with S compact in which [is not a contraction.

Th.5.3

105

CHAPTER

derivative off Similarly, the nth derivative of J, denoted by P,.}, is defined to be the first derivative of for n = 2, 3, . . .. (By our definition, we do not consider p.) unless r:» is defined on an open intervaL) Other notations with

r:».

which the reader may be familiar are

DERIVATIVES
S.1 INTRODUCfION This chapter treats the derivative, the central concept of differential calculus. Two different types of problem-the physical problem of finding the instantaneous ~elocity of a moving particle, and the geometrical problem of finding the tangent h,ne to a curve at a given point-both lead quite naturally to the notion of derivatrve ', Here. w~ shall not be concerned with applications to mechanics and geometry, but Instead wIll confine our study to general properties of derivatives, This chapter deals primarily with derivatives of functions of one real variable specifically, real-valued functions defined on intervals in R. It also discusses bri~fly .deriv~tives of vect~:}f-~a1ued functions of 0n.e real variable, and partial derivatives, srnee these tOpICS Involve no new ideas. Much of this material should be familiar to the reader from elementary calculus. A more detailed treatment of derivative theory for functions of several variables involves significant changes and is dealt with in Chapter 12. The last part of the chapter discusses derivatives of complex-valued functions of a complex variable. S.2 DEFINITION OF DERIVATIVE If f is defined on an open interval (a, b), then for two distinct points x and c in (a, b) we can form the difference quotient

/'(e)

= D/(e)

== ~f (c) = ~.~
dx dx

[where y = J(x)], written

"me

or similar notations. The function I itself is sometimes which produces f' from I is called differentiation.

PO).

The process

5.3 DERIVATIVES to theorems

AND CONTINUITY on derivatives

The next theorem makes it possible to reduce some of the theorems on continuity.

Tlworem 5.2. If I is defined on (a, b) and differentiable at a point c in (a, b), then there is a function 1* (depending on! and on e) which is continuous at c and which satisjit'f the equation

I(x) - J(c)

(x - e)f*(x),

(I)

for all x in (0, b), with j*(c) = /'(c). Conversely, tinuous at e, which satisfies (I), then f is differentiable Proof If/,(c) exists, let/*

if

there is a function f*, conat c and /'(e) = f*(c}.

be defined on (0, b) as follows:

xif x :;I; e, f*(c)

/*(x)
Thenj* is continuous

f(x} -:-J(c1
C

= /,(c).

at c and (I) holds for all x in (a, b). c

Conversely, if (I) holds for somef* continuous at e, then by dividing by x and Jetting x ...... we see that/,(c) c exists and equalsf*(c).

f(x)

- fCc) x c
as x .... c.
CE

As an immediate consequence
Tlworem 5.3. Iff Proof Let x
-+ C

of (I) we obtain: at c, then f is continuous at c.

is differentiable in (I).

We keep c fixed and study the behavior of this quotient Then

Definition 5.1. Let f be defined on an open interval (a, b), and assume that

f is said

(a, b).

NOTE.

IQ

be differentiable

01

c whenever the limit

x--+('

. f(x)j 1m X-

Equation (I) bas a geometric interpretation which helps us gain insight into its meaning. SinceJ* is continuous at cJ*(x) is nearly equal toft(c) = /'«(') if x is neal' c. Replacing/*(x) by /'(1:) in (1) we obtain the equation

I(x)

f(c)

+ !'(c)(x

- c),

exists.

The limit, denoted by,f'(c),

is called the derivative off at c.

This limit process defines a new function I', whose domain consists of those points in (a, b) at which I is differentiable. The function /' is called the first 104

which should be approximately correct when x c is small. In other words, if lis differentiable at c, then/is approximately a linear function near c. (See Fig, 5.1). Differential calculus continually exploits this geometric property of functions.

106

Deri'ratives

Th.S.4

Def.5.6

One-Sided Derivati_

aDd bdinite Derimtive5

107

Theorem 5.5 (CNUII ruk). Let f be defined on an open interval S, let 9 be defined on I(S), and consider the composite function 9 f defined on S by the equation
0

(g JXx)
0

g[j(x)).

Figure 5.1

Assume there is a point c in 5 such that f(c) is on interior point of 1(5). II f is di.fferenliable at c and if g is d{trerentiable at fCc) then go f is differentiable at c and we have (g fnc) = gTf(c)]J'(c).
Q

Proof

Using Theorem 5.2 we can write

f(x) - ftc) (x c)f*(x)
=

5.4 ALGEBRA OF DERIVAJ'lVFS

for all x in S,

The next theorem describes the usual formulas for differentiating the sum, difference, product and quotient of two functions.

where f* is continuous at c and I*(c) g(y) - g[f(c)]

/'(c). Similarly,

Lv -

f(c)]g*(y),

_r + gJ
a)
b) c) Proof

Theorem 5.4. Assume f and 9 are defined on (0, b) and differentiable - g, and]> gare alsodijferentiobJe at c. This is also trueolllg The derivatives at c are given by the following formulas."

at c. Then if g(c) #- O.

for all y in some open subinterval T of/(S) containingf(e). at/(e) and g*[j{c)] g'[f(c)]. Choosing x in S so that y = f(x)
g[j(x)] - g[f(e)]
= E

Here g* is continuous

U±

g)'(c)

= /'(e) ±
=
=

T, we then have
=

g'(c),

[J(x) - f(c)]g*[f(x}]

(x - c)f*{x)g*(f(x)].

(2)

(f'g),(c) U/g)'(c)

f(c)g'(e)
g(c)f'(c)

+ J'(c)g(c),
- f(e)g'(~) g(e)2 •

By the continuity theorem [or composite functions,

provided gee) =f:.

o.

g*lf(x)]

..... g*(f(c)]

== 9Tf(c)]

as x

-jo

c.

Therefore, if we divide by x - c in (2) and let x .....c, we obtain

g(x) = gee)

We shall prove (b). Using Theorem 5.2 we write

f(x) = I(c)

(x -

c)f*(x),

+ +

lim g[f(x)]
(x - c)g*(x).

Thus,
I(x)g(x) - I(c)g(c)

as required.
5.6 ONE-SIDED

>;-"

X -

g(J(c)]
C

= g'[I(c)]/,{{'),

== (x - c)(f(c)g*(x)
-+

+ f*(x)g(C)]

(x - C)2f* (x)g* (x).

DERIVATIVES

AND INFINITE

DERIVATIVES

Dividing by x - c and letting x are similar.

c we obtain (b). Proofs of the other statements

From the definition we see at once that iff is constant on (a, b) then f' = 0 on (a, b). Also, if f{x) x, then f'(x) I for all x. Repeated application of Theorem 5.4 tells us that jf f(x) "" x" (n a positive integer), then f'(x) "" nx"-1 for all x. Applying Theorem 5.4 again, we see that every polynomial has a derivative everywhere in R and every rational function has a derivative wherever it is defined. 5.5 THE CHAIN RULE

Up to this point, the statement that I has a derivative at c has meant that c was interior to an interval in which f was defined and that the limit definingj'(c) was finite. It is convenient to extend the scope of our ideas somewhat in order to discuss derivatives at endpoints of intervals. It is also desirable to introduce infinite derivatives, so that the usual geometric interpretation of a derivative as the slope of a tangent line will still be valid in case the tangent line happens to be vertical. In such a case we cannot prove that I is contin uous at c. Therefore, we explicitly require it to be so. Definiti6R 5.6. Let I be defined on a closed interval S and assume that f is continuous at the point c in S. Then I is said to have a righthand derioauoe at c if the righthand limit lim f(x)
x-,'+

A much deeper result is the so-called chain rule for differentiating composite functions.

- ftc)
XC

lOCI

lb. 5.7

n.5.9

lot

exists as a finite value. or if the limit is +00 or - 00. This limit will be deIWted by f~(c). Lefthand deril.llJtives, denoted by f:"'(e), are similariy defined. In addition, if c is an !nterior point of S, then we say that / has the derivative f'(c) = + OJ if both the rlgh/~ and /efthand derivatives at care + 00. (The derivative f'(c) = _ 00 is Similarly defined.)

If/,(e)

= + 0:).

there is a l-ball B(c) in which

f(x) - f(c)

>1

x-c

whenever x #- c.

It is clear thatfhas a derivative (finite or infinite) at an interior point c if and only if,f~(e) = f ~(c), in which casef~(c) = f'_(c) "= f'(e). •

In this ball the quotient is again positive and the conclusion follows as before. A result similar to Theorem 5.7 holds, of course, jf f'(e) at some interior point c of (0, b).

< 0 or if /'(c)

=-

0:)

5.8 ZERO DERIVATIVES

AND LOCAL EXTREMA

IHjiJrititnI S.8. Let f be a real-fi4lued /unction defined on a subset S 0/ a metric space M, and assume a e S. Then. f is said to have a local maximum at a if there is a ball 8(a) such that f(x) If f(x)
NOTE.

~ /(D)

for aU

in 8(a) n S.

~ f(a) lor alJ x iii B(a) n S, then / is said to hove a local minimum at a.

A local maximum at a is the absolute maximum off on the subset B(a) n S. Iffhas an absolute maximum at a, then a is also a local maximum. However,/ can have local maxima at several points in S without having an absolute maximum on the whole set S.
YIgUU

S.Z

The next theorem shows a connection between zero derivatives and local extrema (maxima or minima) at interior points.
TItsoum j.!J. Let f be defined on an open interval (a. b) and assume that / has a local maximu.m or a local minimum at all interior point e of ia, h). If I has a deriuatite (finite or infinite) at e, then f'(e) must be O.

Figure 5.2 illustrates SOme of these concepts. At the point Xl we havef~(xl) = At Xl the lefthand derivative is 0 and the righthand derivative is 1. Also, f'(x3) ~ 00, f~(X4) = -1, .f~(X4) = +1, f'(x6) = +00, and /'-(x7) =2. There IS no derivative (one-sided or otherwise) at xs. since f is not continuous there.
00.

5.1 FUNCfIONS
~ In (~. b) 111 which

WITH NONZERO

DERIVATIVE

If f'(c) is positive or +00, then f cannot have a local extremum at c because of Theorem 5.1. Similady,f'(c) cannot be negative or -00. However, .because there is a derivative at c. the only other possibility is/,(c) O.
Proof.

T":orem 5.7. Let f be defined on an open interval (a, b) and assume that for SOme we have f'(e) > 0 or f'(c) = + co. Then there is a I-ball B(c) s (a, b) f(x)

> ftc)

if x > c,
f(x)

and

/(x)

< fee)

if

< c.

Proof

If r(e) is finite and positive we can write f(c) (x c)f*(x),

o_f contm:ous functl~ns there is a I-ball

sign as!
(c), and this means thatf(x)

where!* iscontinu~us at c andf*(c)

r(c) > O. By the sign preserving property B(e) r;;;; (a, b) in whichf*(x) has the same f(c) has the same sign as x-c.

The converse of Theorem 5.9 is not true. In general, knowing thatf'(c) = 0 is not enough to determine whether I has an extremum at c. In fact, it may have neither, as can be verified by the example f(x) = x3 and e = O. In this case, 1'(0) = 0 but/is increasing in every neighborhood ofO. Furthermore, it should be emphasized that f can have a local extremum at c without f'(c) being zero. The example I(x) = Ixl has a minimum at x = 0 but, of course, there is no derivative at O. Theorem 5.9 assumes that I has a derivative (finite or infinite) at c. The theorem also assumes that c is an interior point of (a, b). In the examplef(x) = x, where a !>: x !>: b,ftakes on its maximum and minimum at the endpoints butf'(x) is never zero in [a, b].

110

Th.5.10

Cor. 5.15

111

S.9 ROLLE'S

THEOREM

It is geometrically evident that a sufficiently "smooth" curve which crosses the x-axis at both endpoints of an interval [a, b] must have a "turning point" somewhere between a and b. The precise statement of this ract is known as Rolle's

NOTE. The reader should interpret Theorem 5.12 geometrically by referring to the curve in the xy-plane described by the parametric equations x = g(I), Y = I(t}, a:S;; t :s;; h.

theorem.

There is also an extension which does not require continuity at the endpoints. lhtJulfl 5.13. Let I and 9 be two lunctions, each having a derivative (finite or infinitf?)at each point 01 (a, h). At the endpoints assume that the limits I(a +), g(a+). I(b-) and g(b-) exist as finite values. Assume further that there is no interior point x at which both /'(x) and gt(x) are infinite. Then for some interior
point c we have

Tlteorem 5.10 (Rolle). Assume f IuJs (1 derivative (finite or infinite) at each point 01 an open interval (0, b), and assume that I is continuous at both endpoint» a olld b. 111(a) = f(b) there is at least one interior point c at whichf'(c) =

o.

Proof We assume f' is never 0 in (a. b) and obtain a contradiction. Since I is continuous on a compact set, it attains its maximum M and its minimum m somewhere in [a, bJ. Neither extreme value is attained at an interior point (otherwise I' would vanish there) so both are attained at the endpoints. Since I{a) = I(b), then m = M. and hencelis constant on [a, b]. This contradicts the assumption tharj" is never 0 on (a, b). Therefore /,(c) = 0 for some c in (a. b).
S.10 THE MEAN-VALUE THEOREM FOR DERIVATIVES

j'(c)[g(b-) Proof

- g(a+)] = g'(c)(f(b-)

f(!t)].

Define new functions F and G on [a, b] as follows: F(x) F(a)

=

= I(x)
G(a)

and

G(x) = g(x}

if x

IE

(a, b);

1(a+).

= g(a+). F(b) = f(b-),

G(b)

= g(b-).

TAeorem 5.11 (Mellllf-Yalue TfIeorem). Assume that I has a derivative (finite or infinite) at each point of an open interval (a, h). and assume also that I is continuous at both endpoints a and b. Then there is a point c in (a. b) such that

Then F and G are continuous on [a, b] and we can apply Theorem 5.12 to F and G to obtain the desired conclusion. The next result is an immediate consequence of the Mean-Value Theorem. Theorem 5.14. Assume I has a derivative (finite or infinite) at each point of an open interval (0. b) and 1M' I is continuous at the endpoints a and b. a) If f" takes only positive values (finite or infinite) in (a, b), then f is strictly Increasing on [a, h].
b) e)

I(b) - f(a)

= /'(c)(b

a).

Geometrically, this states that a sufficiently smooth curve joining two points We will deduce Theorem 5.11 from a more general version which involves two functions f and 9 in a symmetric fashion.
A and .B has a tangent line with the same slope as the chord AB. Tltevrem 5.12 (GeneraIivd Mea". Value Theorem), Let I and 9 be two functions, each having a derivative (finite or infinite) at each point 01 an open interval (a, b) and each continuous at theendpoints a and h. Assume also that there is no interior point x at which bOlhP(x) and g'(x) are infinite. Then lor some interior pcint c we have

II/' takes only negative values (finite or infinite) in (a, b), then decrt'tlJi1l!l on [0, b]. II/'
is zero f!Vf!rywhf're in (a, b) then

is strictly

I is

constant on [a, b].

Proof Choose x < y and apply the Mean-Value Theorem to the subinterval [x, y] of [a. bJ to obtain I(y) - I(x)
= /,(c){y

- x)

where c € (x, y).

/,(c}[g(b) - g(a)]
NOTE.

= g'(c)[J(b)

- 1(0)].

All the statements of the theorem follow at once from this equation.
By applying Theorem 5.14 (c) to the difference f Corolltuy 5.15. in (a, b), then f

When g(x)

x, this gives Theorem 5.1 I.

9 we obtain:

Proof. Let hex) = f(x)[g(b) - g{a)] - g(x)[J(b) - 1(0.)]. Then h'(x) is finite if both!'(x) and g'(x) are finite, and h'(x) is infinite if exactly one of/,(x) or g'(x) is infinite. (The hypothesis excludes the case of both and g'(x) being infinite.) Also, " is continuous at the endpoints, and h(a) = h(b) = I(a)g(b) - g(a)f(b}. By Rolle's theorem we have h'(c) = 0 for some interior point, and this proves the assertion.

II I

and 9 are continuous on [0, b] and haoe equal finite derivatir,'es 9 is constant on [a, b]. TIJEOREM FOR DERIVATIVES

rex)

5.11 lNTERMEDIATE-VALUE

In Theorem 4.33 we proved that a function I which is continuous on a compact interval [a, b] assumes every value between its maximum and its "minimum on

III

Th.S-It;

Th. !UO

Taylor's Formula willi Remainder

113

the interval. In particular,J assumes every value between f(a} and feb}. A similar result will now be proved for functions which are derivatives. v'orem S.16 (l"'e''''dmt~ tlt.uremfor tkr~N). Assume that f is defined on a compact interval [a, b] and that f bas a derilJiZtive(finite or infinite) (11 each interior point, Assume also that / has finite one-sided derivatives I~(a) and f'_(b) at the endJWints. with f+(a) .p. /'-(b}. Then, if c is a real number between /+(a) and f'-(b), there exists at least one interior point x !luch that /'(x) = c. Proof. Define a new function 9 as follows'

discontinuity (by Theorem 4.5 J). Hence f' omits some value between /'(Cl) and /'(fJ), contradicting the intermediate-value theorem.
5.12 TAYLOR'S FORMULA WITH REMAINDER

As noted earlier, iff is differentiable at c, thenj" is approximately a linear function near c. That is, the equation f(x)
=

/(c)

+ f'(cXx

- c),

g(x) =/(x) - f~Q~ x-a

if x

f; tl,

g(a)

= f';'(a).

Then 9 is continuous on the closed interval [a, h]. By the intermediate-value theorem for continuous functions, 9 takes on every value between /';'(a) and [feb) - l(o)]/(b - a) in the interior (a, b). By the Mean-Value Theorem, we have g(x) = J'(k) for some k in (u, x) whenever x E (a, b). Therefore /' takes on every value between 1+(0) and [/(b) - l(a)]/(b - a) in the interior (a, b). A similar argument applied to the function h, defined by
h(x)

is approximately correct when x c is small. Taylor's theorem tells us that. more generally,fcan be approximated by a polynomial of degree n I ir/has a derivative of order n. Moreover, Taylor's theorem gives a useful expression for the error made by this approximation. TllHrem 5.19 (Taylor). Let f be a !unction having finite nth der;lJiZtivep.) everywhere in an open interval (a, b) and assume that j<,,-l) is continuous on the dosed interval [a, b]. Assume that e E [a, b]. Then,for every x in [0. b]. x .p. C, there exists a point x 1 interior to the interval joining x and c such that
f(x)

= f(x)

- feb)

x~b

jf x '# b,

h(b)

f:(b),

= f(c) +

.-1 f(I:)( }
.=1

L __e (x ~ k!

e)"

+~

f(O), n!

(x - C)R.

shots that ]" takes on every value between [/(b) - f(a)]/(b a) andf~(b) in the interior (a, b). Combining these results. we see that/, takes on every value between I~(a) andf:(b) in the interior (a, b), and this proves the theorem. Theorem 5.16 is still valid if one or both of the one-sided derivatives I~(a), f~(b), is infinite. The proof in this case can be given by considering the auxiliary function g defined by the equation g(x) = /(x) - ex, jf x E [a, b]. Details are left to the reader.
NOTE.

Taylor's theorem will be obtained as a consequence of a more general result that is a direct extension of the generalized Mean-Value Theorem. Theorem 5.20. Let / and 9 be two functions having finite nth derivalilJes l and g('" in an open interoal (0,11) and continuous (n - l)st derivatives in the closed interval [a, b]. Assume that C E [a, b]. Then, for every x in [a, b], x #; C, there exists 0 point Xl interior to the interval joining.;,K and c such that

r'

The intermediate-value theorem shows that a derivative cannot change sign in an interval without taking the value O. Therefore, we have the following strengthening of Theorem 5. 14(a) and (b). Tlreorem 5.17. Assume f has a derivative (finite or infinite) on (a, b) and is continuous at the endpoints a and h. If /'(x) #: 0 lor all x in (fl, b) then f is strictly monotonic on [a, b J. The intermediate-value theorem also shows that monotonic derivatives are necessarily continuous.

/(x) -

.-If(l) )]
t~O

L~ k!

(x - e)l

g(·)(XI)

= P")(xt)

[0-1

g(x) -

l=O

(l}() L g~k! (x

- c)~ .

NOTE.

o~

For the special case in which g(x) = (x - e)", we have g(i)(C) = 0 for k s; n - I and g(lO'(X) = n!. This theorem then reduces to Taylor's theorem.
C

Proof.

For simplicity, assume that new functions F and G as follows: F{t) G(t)

< b and that

ft-t i~1

>

c. Keep x fixed and define

= f(t) +
=

I' is continuous on (a,

Proof

Theorem 5.18.

Assume r exists and is monotonic on an open interval (a, b). Then

b).

L: _t k!

ril( )

(x ~

tl,

9(t)

0-1
k~1

We assume j" has a discontinuity at some point c in (a, b) and arrive at a contradiction. Choose a closed subinterval [I)', P] of (a, b) which contains c in its interior. Since f' is monotonic on [a, If] the discontinuity at c must be a jump

L:-gk!' _

(l)(t)

(x - I)",

for each t in [c, x]. Then F and G are continuous on the closed interval [c, x] and have finite derivatives in the open interval (c, x), Therefore, Theorem 5.12 is

1:14

PartiIIl Derimi:,.

115

applicable

and we can write G(e)]

=

F'(xl)[G(x)

G'(x,)[F(x)

- F(c)] ,

where

XI E

(c, z).

The Mean-Value Theorem, as stated in Theorem 5.11, does not hold for vectorvalued functions. For example, if I(t) = (cos t, sin t) for all real t, then f(21t) - 1(0) O.

This reduces to the equation F'(xJ)[g(x) G(e)]

G'(x1)[f(x)

- F(c)],

(a)

since G(x) = g(x) and F(x) = f(x). If. now, we compute the derivative of the sum defining F(t), keeping in mind that each term of the sum is a product, we find that all terms cancel but one, and we are left with F'(t) SimiJarly. we obtain

but 1'(1) is never zero. In fact, Ilf'(t)U = 1 for all t, A modified version the Mean-Value Theorem for vector-valued functions is given in Chapter 12 (Theorem 12.8). 5.14 PARTIAL DERIVATIVES Let S be an open set in Euclidean space R", and let f: S -+ R be a real-valued function defined on S. If:l[ = (x, •... , xJ and c = (CI, , ..• cJ are two points of S having corresponding coordinates equal except for the kth, that is, if Xi = c. fOT i '" k and if xt '" c1> then we can consider the limit

or

(x - 1)0-1 P~)(t). (n - I)!

t)"-I

G'(t)
If we put t =
Xl

= (x -

g(.}(t). of the theorem.

(n - 1)! and substitute into (a). we obtain the formula

I· /(:11:) 1m
%~......

fCc)

X ... -

c"

. respect to the kth

5.13 DERJV ATIVES OF VECTOR-V ALUJID FUNCfIONS Let I: (a, b) _,. R" be a vector-valued function defined on an open interval (a, b) in R. Then f = (/1' ... .f.J where each eomponent j; is a real-valued function defined on (a, b). We say that f is differentiable at a point c in (a, b) if each component!. is differentiable at c and we define f/(e) = U;(c}. ' , . , f~(c)). In other words, the derivative f'(c) is obtained by differentiating each component of r at c. In view of this definition, it is not surprising to find that many of the theorems on differentiation are also valid for vector-valued functions. For example, if f and g are vector-valued functions differentiable at c and if A is a real-valued function differentiable at c, then the sum r + g, the produce U, and the dot product f· g are differentiable at c and we have (f

When this limit exists. it is called the partilll deril10tive of fwith coordinate and is denoted by DJ(c),

or by a similar expression. We shall adhere to the notation Dd(c}. This process produces n further functions Dt!, DzJ, ... , DJ defined at those points in S where the corresponding limits exist. . Partial differentiation is not really a new concept. We are merely trea tmg /(x!, .. , x,,) as a function of one variable at a time, holding the others fixed. That is, if we introduce a function g defined by
g(Xt)

= /(Ch

•.. , ('l-I> Xl' C.t+I, ...

c.),

g)'(c)

=
=

r'(c)

g'(c},

(ll')'(c) (f·g),(c)

}.'(c)f(c) f'(c)'g(e}

+ J.(c)f'(c),
+
f(c)·g'(c).

The proofs follow easily by considering components. There is also a chain rule for differentiating composite functions which is proved in the same way. If f is vectorvalued and if 11 is real-valued, then the composite function g given by g(x) "'" f[ u(x)] is vector-valued. The chain rule states that g'(c)

f'(u(c)Ju'{('), of u(c) and if u'(c) and

then the partial derivative DJ(c) is exactly the same as the ordinary derivative g'(Cl)' This is usually described by saying that we differentiate/with respect to the kth variable, holding the others fixed. In generalizing a concept from RI to R", we seek to preserve the important properties in the one-dimensional case. For example, in the one-dimensional case, the existence of the derivative at c implies continuity at c. Therefore it seems desirable to have a concept of derivative for functions of several variables which will imply continuity. Partial derivatives do not do this. A function of n variables can have partial derivatives at a point with respect to each of the variables and yet not be continuous at the point. We illustrate with the following example of a function of two variables:

if the domain of f contains exist.

a neighborhood

fT u(e)J

both

f(x, y)

= {~, + Y.

if x

= 0 or

= O.

otherwise.

ner.5.11 The partial derivatives DJ(O, 0) and Dd(O, 0) both exist. In fact, 1,

117

_ 1m x D 1J(O • 0) - li1m It». 0) - f(O. 0) - I' - ..-0 x- 0 x~{)x

become a field] which includes C. Since division is possible in C, ho~ever. we can form the fundamental difference quotient [/(z) - I(c)]/(z - c) which was used to define the derivative in R. and it now becomes clear how the derivative should be defined in C. DejinitWfJ 5.11. LeI I be a complex-valued [unction defin~d on Q~ ~pen set S in C, and assume c E S. Then/is said to be differentiable at c !f the limit lim
r-ec Pe;_;(;...-z

and, similarly, D·d(O. 0) = I. On the other hand. it is clear that this function is not continuous at (0,0). The existence of the partial derivatives with respect to each variable separately implies continuity in each variable separately; but, as we have just seen, this does not necessarily imply continuity in all the variables simultaneously. The difficulty with partial derivatives is that by their very definition we are forced to consider only one variable at a time. Partial derivatives give us the rate of change of a function in the direction of each coordinate axis, There is a more general concept of derivative which does not restrict our eonsiderations to the special directions of the coordinate axes. This will be studied in detail in Chapter 12. The purpose of this section is merely to introduce the notation for partial derivatives, since we shall use them occasionally before we reach Chapter 12. If f has partial derivatives Dd •... , DJ on an open set S, then we can also consider their partial derivatives. These are called second-order partial derivatives. We write Dr.d for the partial derivative of Dtf with respect to the rth variable. Thus, D•.d Higher-order partial derivatives D,(Dd). Other notations are

)'----'--1..;..,(
Z -

cc.-)

f'(c)

exists. By means of this limit process, a new complex-valued ~unc.tioll f~is _?etined at those points z of S where f'(z) exists. Higher-order derivatives f ,f , ... are, of course, similarly defined. . The following statements can now be proved for complex-valued functions defined on an open set S by exactly the same proofs used in the real case:

a) f is dijJerfntiable at c if. and only if, there is a function f"", continuous that
f(z) - J(c)

01

c, such

(z

c)[*(z),

for all z in S, with f*(c)

NOTE.

= f'(c).

are similarly defined.

If we let g(z) = r(z)

- r(c) the equation

in (a) can be put in the form - c). Taylor formula for

;PI D.f=-r,. "a' Xi IJX.

S.IS DIFFERENTIATION OF FUNCTIONS OF A COMPLEX VARIABLE

J(z)
where g(z)
--jo

f(c)

+ !'(c)(z

- c)

+ g(z)(z

0 as z

--jo

c.

This is called ajirs/-(}rder

f.

In this section we shall discuss briefly derivatives of complex-valued functions defined on subsets of the complex plane. Such functions are, of course, vectorvalued functions whose domain and range are subsets of R 2• An the conside ..atiens of Chapter 4 concerning limits and continuity of vector-valued functions apply, in particular, to functions of a complex variable. There is, however. one essential difference between the set of complex numbers C and the set of n-dimensional vectors R" (when n > 2) that plays an important role here. In the complex number system we have the four algebraic operations of addition. subtraction, multiplication, and division, and these operations satisfy most of the "usual" laws of algebra that hold for the real number system. In particular, they satisfy the first five axioms for real numbers listed in Chapter I. (Axioms 6 through 10 involve the ordering relation -c , which cannot exist among the complex numbers.) Any algebraic system which satisfies Axioms I through 5 is called a field. (For a thorough discussion of fields, see Reference 1.4.) Multiplication and division, it turns out, cannot be introduced in RIt (for n > 2) in such a way that R" will

b) If I is differentiable at c. then f is continuous at c. c) Iitwofunctionsfand 9 have derivatives at c, th~n their sum, difference, praduc:, and quotient also have derivatives at c and are gIVen by tire usualformulas (as In Theorem 5.4). In the case of JIg, we must assume gee) #- O. d) The chain rule is valid; thai is
(g
0

10 say.

11:0"

haue ]f'(c),

f)'(c)

= g'[/(e)

if the domain of g ('ontains a neighborhood of ftc) and if /,(e) and g'(f(c)] both exist.
When/(z) = z, we findf'(z) = I for all z. i~ C: Using (c) r:epeatedly, we find thatl'(z) = nz"-I when/(z) = z" (n is a positive integer), This also holds when For example, if it were possible to define multiplication in R:J so as to mak; R: a field 3 including C. we could argueas follows: For every x III R the veclor~ I, x; x ,x ,:,,",ouId be linearly dependent (see Reference SJ. p, SS8). Hence for each x In R • a relation of the form a + a x + a Xl + aJ,x3 = 4) would hold, where all> Qt. Ql. 03 are real hers pol;nomial of degree three with real coefficients is a product of a iynomial and a quadratic polynomial with real coefficients. The only roots such l~um PO. mear le be polynomials can have are either real numbers or romp x num rs.

i'U! ~

ttl

111. 5.12

Th. S.U

Caudly-R.IemamI EquatiOllS

119

n is a negative integer, provided z '" O. Therefore, we may compute derivatives of complex polynomials and complex rational functions by the same techniques used in elementary differential calculus.
5 .r 6 THE CAUCHY-RIEMANN EQUATIONS

wheref* is continuous z '"' x

at

andf*(c)

= /,(c).
ib, and

Write /*(z)

iy,

A(z)

iB(z).

Iffis a complex-valued value in tbe form

function of a complex variable. we can write each function f(z) = u(z}

where A(z) and 8(z} are real. Note that A(z) _,. A(e) and B(z) -> 8(c) as z _,. c. By considering only those z in S with y = band taking real and imaginary parts of (5), we find u(x, b) - Il(a, b) = (x - a)A(x

+ ib),
and

vex, b) -

!J(a, b) = (x -

a)B(x

+ lb).

~~.
. Similarly. whenf(z)

+ ;V(Z),
.

Dividing by x - a and letting x .....a we find Dlu(c)

where II ~d v are real-valued functions of a complex variable. We can, of course, also consider II and v to be real-valued functions of two real variables and tben f(z)
= u(x. y)

A(e)

D11)(c)

= =

B(e).

iv(x, y),

if z

+ iy.

Since!'(c)

= A(c) + i8(e), this proves (3). Similarly, by considering only those z in S with x

a we find

In either case, we write f = ~ + ivand we refer to II and 11 as tbe real and imagrnary parts off. For example, 10 the case of the complex exponential function J, defined by f(z) the real and imaginary

and which proves (4). Applications of the Cauchy-Riemann equations are given in the next theorem.

~cosy

i~siny.

parts are given by

u(x, y)

Z2

= e" cos )', = (x + ;y)2,

v(x, y) = ~ sin y. we find v(x, y) = lxy.

Theorem 5.11. Let f = u + io be a function with a derivative everywhere in an open disk D centered at (a, b). If anyone 01 u, D, or If I is constanrt on D, then [ is constant on D. Also.f is constant iff'(z) = o jar all z in D. Proof Suppose u is constant on D. The Cauchy-Riemann equations show that D2t:i = D,» = 0 on D. Applying the one-dimensional Mean-Value Theorem twice
places a we find, for some y' between band J"
(y -

In the next theorem we shall see that the existence of the derivative/, rather severe restriction on the real and imaginary parts u and v.

v(x,

y) -

v(x, b) =

b)D2v(x,

y') =

0,

5.22. LeI J = u + iv he defined on an open set S in C. IJ /,(e) exists for so~e C In S, then the partial derivatives D1u(c), D2u(c), Dlv(e) and D2v(e) also exist and we have /'(c) D11I(c) + i D1v(c), (3) and j'(c) = D2v(e) - i D2u(e). (4) This implies, in particular, that and
These last two equations are known They are usually seen in the form
NOTE. {JIJ

Tu",em_

and, for some x'' between a and x, v(x, b) l'(a, b) = (x - a}Dlv(x', b)

O.

Therefore v(x, y) v(a, b} for all (x, y) in D. so v is constant on D. A similar argument shows that if v is constant then u is constant. Now suppose If I is constant on D. Then 1/12 u2 + 1,1 is. constant on D. Taking partial derivatives. we find

as the Cauchy-Riemann

equations.

By the Cauchy-Riemann

equations
{'DIU

the second equation

uDIt'

can be written as

= O.
DID

iJu

Combining this with the first to eliminate

u2
11

iJx

oy
defined on S such that

+ (;2 =
Ilf

Proof

Since/'(c)

exists there is a functionJ* fez) - fCc)

is constant,

0, then u t' = 0, so so f is constant.

=0

= O. If uf

we find (112 + v1)Dlu = 0, If + [,2 oF 0 then DIU = 0; hence

(z - c)f*(z),

(5)

tHere

denotes the function whose value at z is 1/(2)1.

lZU Finally, jf f' = 0 on D. both partial derivatives DID and D1v are zero on D. Again. as in the first part of the proof, we find/is COnstant on D. Theorem 5.22 tells WI that a necessary condition for the functionj' = u + io to have a derivative at c is that the four partials DIu, D1.u, D1o, DIP, exist at c and satisfy the Cauchy-Riemann equations. This condition, however, is not sufficient. as we see by considering the following example. Enmple. Let
II

121

EXERCISES
ReaI.valued functions

In the following exercises assume, where necessary, .a knowledge. of ~he for~uJas for differentiating the elementary trigonometric, exponential, and logartthmic functions. 5.1 A function f is said to satisfy a Lipschitz condition of order a at c if there exists a positive number M (which may depend on c) and a I-ball B(c) such that
If(x} - /(c)1 < Mix - cl" whenever x E B(c), x "# c. a) Show that a function which satisfies a Lipschitz condition of order at c if It: > 0, and has a derivative at c if a ;. 1.

and v be defined as follows;

if (x, y)
if (x, y)

'" (0, 0),

11(0, 0) 0(0. 0)

0, O.

It:

is continuous

(0, 0),

It is easily seen that DIu(O, 0) = D1v(O, O} I and that D1II(0, 0) = Dl!l{O,O) = ~1, that the Cauchy-Riemann equations hold at (0, 0). Nevertheless, the function I '" u + if) cannot have a derivative at z = O. In fact, for x = 0, the difference quotient becomes
80

b) Give an example of a function satisfying a Lipschitz condition of order I at c for which /,(c) does not exist. S.Z In eaCh of the following cases, determine the intervals in which the function I is increasing or decreasing and find the maxima and minima (if any) in the set where each I is defined. a) /(x) .. x3 + ax + b, x E R. b) [(x)
c) f(x)
ee

/(z)

- 1(0)

z-o
whereas for x ~ Y. it becomes
/(z) - 1(0)

-y

JjI

.+ iy = 1 + i,
I+I 2

log (x2
);213(X -

=
=

9), 1)4,

Ix1 >
0~
X

3.
::;;

l.

d)

lex)

(sin x)/x if x "# O,f(O) = I,

0 ::;; x ::; n/2.

xi

z-O
and hencel'(O) cannot exist.

x+ix

5.3 Find a

polynomial/of
/(Xl) ." a..
i2l>

IOWQt possible degree such thai

[(Xl) = aJ,

/'(xl)

bl>

j'(X'l)

= b",

where In Chapter 12 we shall prove that the Cauchy-Riemann equations do suffice to establish existence of the derivative of J = u + if) at c if the partial derivatives of u and v are continuous in some neighborhood of e. To illustrate how this result is used in practice, we shall obtain the derivative of the exponential function, Let J(z) = e" = u + io. Then

x,

-#

x2

and

02'

b.. b2 are given real numbers.

5.4 Define

I as

folloW'l: f(x)

e-1/X'

if x "# 0,/(0)

O. Show that

a) f is continuous for all x, b} p.) is continuous for all x, and that /<0>(0) = 0, (n = 1,2, ... ).
g(x) = x sin (llx),

uix, y)
and hence

if" cos y,

v(x, y)

e" sin y,

S.5 Definej,g, and II asfollows:[(O) hex) = x2 sin (l/x).

g(O) = h(O) "" Oand,ih:F

O./(x)

sin (I/x),

Show that /,(0) does not exist.

g'(O) does not exist. h'(O)

a)

/'(x)

-1/x2

cos

(I{x), if x "# 0; Ilx cos (1/:.:), if x "# 0;

b) g'(x)

sin (1Ix)

c) 1I'(x) == 2x sin (If:.:) - cos (1Ix), if x ¥ 0; Since these partial derivatives are continuous everywhere in il2 and satisfy the Cauchy-Riemann equations. the derivative f'(z) exists for all z, To compute it we lise Theorem 5.22 to obtain

= 0; 1imx~o h'(x) does not exist.

5.6 Derive Leibnitz's formula for the nth derivative of the product II of two functions landg;
If·l(X) =
1:=0

f'(z)
Thus, the exponential

= if" cos y

ie" sin y

= fez).
(as in the real case). 57 Let

t (n) f
k

kl(X)u'"-1<J(x),

where

k! (II -

k)!

function

is its own derivative

and g"(x) ~r all

nd g be two functions defined and having finite third-order derivatives I-ex) x in R. If I(x)g(x) = 1 for all x, show that the relations in (a). (b), (c),

112

113

and (d) hold at those points where the denominators are not zero: a) I (x)lf(x)
b) I'(x)/f'(x)

Mean-Value 1'1IeoI'em 5.10 Given a function f defined and having a finite derivative in (a, b) and such that lim.. _I>_ I{x) + 00. Prove that limx-+b_ P(x) either fails to exist or is infinite.

+ g'(x)fg(x)
- 2/,(x)//(x}

o.
- g~{x)/g'(x) = 3r{x)

o.

c) /"'(x)
/'(x)

_ 31'(x)g"(x) /(x)g'(x)

d)

rex) _ ~ (r(X»)2
f'(x) 2 /'(x)

... g"'(x) _ ~ (g"(~~)<l.

g'(x) 2 g'(x)

lex)

_ g"'_(x) = gT(X)

o.

s.n

Show that the formula in the Mean-Value Theorem can be written as follows:

where 0 < 0 < 1. Determine

a) /(x) = x2, c) I(x)

(J

as a function of x and h when

b) f(x) d)
(J

= x\

The expression which appears on the left side of (d) is called the Schwarzian ikrilltltWe of/at x.
NOTE.

e", fixed, and find lim,,_Q

lex)

log x ,

.• > O.

e) Show that/and

9 have the same Schwarzian derivative if

Keep x#.O

in each case.

g(x) = (a/(x)

+ b l!{cf(x) +

dl,
+

where ad - be "# O.
(be

Hint.

part (d).

If c "# 0, write

(€if + b)/(ef +

d) = (a/c)

ad)/[C(cf

d)]. and apply

5.11 Take lex) = 3x4 - 2x3 - x2 + 1 and g(x) = 4x3 - 3x2 - 2x in Theorem 5.20. Show that f'(x)/g'(x} is never equal to the quotient [(0) - I{O)]/[g(l) - u(O)] if o < x .s I. How do you reeoncije this with the equation
- /(a) _ f'(x,) g(b) -

5.8 Letfl'/2, 91' Uj;be four functions having derivatives in (0,0). Define Fby means of the determinant
F(x) =

yea) -

g'{Xl)'

obtainable from Theorem 5.20 when

It ..

1'?
11

! II(x)

lUI (z]
'/~(X)
91(X)

Il(X)!, Ul(X)!

if x

(a, b).

S.13 In each of the following special cases of Theorem 5.20, take and show that XI = (# + h)/2. a) I(x)
=

= I, c = a, x ,.. b,

a) Show that F'(x) exists for each x in (a, b) and that

F'(x) ...

sin x,

g(x) = cos x;

b) f(x)

e",
X,

li(X)j 4- ) /:(x)
gix)
Ui(X)

92(X)

i2(X)I.

Can you find a general class of such pairs of functions land U for which (0 + b)/2 and such that both examples (a) and (b) are in this class'!

will always be

b) State and prove a more general result for nth order determinants. 5.9 Given n functioIlllh ••..• /", each haviDg nth order derivatives in (a, b). A function W, called the Wronskian off, •...• I", is defined as follows: For each x in (0, b), W(••) is the value of the determinant of order n whose element in the kth row and mth column is r,:-o(x), where k ~ 1,2, .•.• IT and m "" I, 2, ...• n. [The expressionr.,°)(x) is written for/ .. (x).] a) Show that W'(x) can be obtained by replacing the last row of the determinant defining W(x) by the nih derivatives!t")(x), ,r.·}{x). b) Assuming
cdl(x) + x in (a, b).

5.14 Given a function I defined and having a finite derivative /' in the .half-open interval < x S I and such that IP(x) I < 1. Define Q. = f(lfn) for n = I, 2, 3, ...• and show that iim"_,,, o. exists. Hint. Cauchy condition.

5.15 Assume that/has a finite derivative at each point of the open interval (8, 0). Assume also that limr_cP(x) exists and is finite for some interior point c. Prove that the value of this limit must be I'(c). 5.16 Let /be continuous on (a, b) with a finite derivative P everywhere in (a, b). except possibly at c. If lirnx->.. /'(x) exists and has the value A. show that p(e) must also exist and have the value A. 5.17 if tion 5.18

the existence of n constants c" , c•• not all zero, such that ... + c../,,(x) ~ 0 for every x in (a, b}, show that W(..-} = 0 for each

Let I be continuous on [0, I], /(0) = 0, j'(x) finite for each x in (0, 1). Prove thai is an increasing function on (0, 1), then so too is the function g defined by the equag(x) = /(x)/x.

Assume f has a finite derivative in (a, h) aDd is continuous on [a, b] with f(a) =
= O. Prove that for every real 1 there is some c in (a, b) such that

NOTE.

A set of functions satisfying such a relation is said to be a linearly dependent set oo~~ . c) The vanishing of the Wronskian throughout (a, b) is necessary, but not sufficient, for linear dependence of It..... f". Show that in the case of two functions, if the Wronskian vanishes throughout (0, b) and if one of the functions does not vanish in (a, b), then they form a linearly dependent set in (a, b).

f(fJ)

/'(d

= If(e).

Hint. Apply Rolle's theorem to g(x)/(x) for a suitable g depending on 1.

5.1' Assume/is continuous on (a, b] and has a finite second derivativ !' in the open interval (a, b). Assume that the line segment joining the points A = (a,/(a)) and B = (b,f(b» intersects the graph of lin a third point P different from A and B. Prove that!'(c) = 0 for some c in (a, b).

114 5.20 If /bM a finite third derivative I" in [a. b] and if

/(4) = /,(a)

us
5.28 Prove the foUowing theorem: Let / and g be two functions having /inite nth derisatioes in (a, b). For some interior point e in (0, b), alsume that /(c) = /,(c) ••• = P~-I)(C) "" 0, and thar g(c) = g'(c) = ... = g(.-H(C) = 0, but that g(ltl(x) is neoer zero in (a, b). Show thar lim f(x} = r'(c) .
x-+~

= /(b) =

/,(b) = 0,

prove thatj'(c)

= 0 for some c in (a, b).

5.11 Assume I is nonnegative and has a finite third derivative I" in the open interval (0, I). IfI(x) = 0 for at least two values of x in (0, I), prove that /"(c) = 0 for some e in (0. I). S.l2 Assume /has a) If/(x) b) If /'(x)
1;:) If rex)
-+

g(x)

g"')(c)

a finite derivative in some interval (a, 1 and/,(x) 1 as x 0 as x

-+ c

+ 00). + 00. + 00.

NOTE.

1(·) and

g("}

are Rot assumed to be continuous at c. Hint. Let

F(x) = /(x) (x - C)"-l/fJI-U(c) (II -

as x

-+

+ 00,

prove that c = O. ..... I as x

-+ 0

+ 00. + 00,

prove that/(x)/x prove that/(x)/x

I)!

as x

define G similarly, and apply Theorem 5.20 to tile functiOllll Fand G. 5.29 Show that the formula in Taylor's theorem can also be written as foUows: j(x) = ~~-(x ~ k!
K-l Flt.c)

5.l3 Let II be a fixed positive number. Show lhal there is no function / satisfying the foUowing three conditions: /,(x) exists for x ;;.:O./,(0) = 0, J'(x) ;;.:h for x > O. S.l4 If h > 0 and if rex) ellists (and is finite) for every x in (Il - h. a continuous on [0 - h;« + h), show that we have: a) f(a b) /(a

+ n). and

if / is

- cl

(x

c){x _ XI)"-l

(n _ I)!

rex!),

h) ~ I(a -

h)

= r(a +
h}

(lh)

+ r(a

_ fJh),

0 < IJ < I; 0 < A. < 1.

where Xl is interior to theintervaljojning x and c. Let 1 - () = (x XI)/(X - c). Show that 0 < 8 < I and deduce the following form of rhe remainder term (due to Caucny):
(1 (n -

h) -

2f~a)

+ f(a

= f'(a

..1.h) - r(a _ ..1.h),

(J).-l(X - c)"r)[Bx I)!

(I

8)c).

Hint. Take G(t)

= get)

= t in the proof of Theorem 5.20.

c) If r(a) exists, show that

/"(a) "" lim/(O

+ It)

2j(a)

+ I(a -

Vector-valued
h).

fimctions

.~O

hZ

5.30 If a vector-valued [uuction f is differentiable at c, prove that

d) Give an example where the limit of the quotient in (c) exists but where /"(0) does not exist. S.l! Let / have a finite derivative in (a. b) and assume that c eo(a, b). Consider the following condition: For every e > 0 there exists a l-ball 1I(c; 0), whose radius" depends only on t and not on c, such that if x e B(c; §), and x ¢ c, then
j(x)

r'(c)

= lim! [f(c
~.... h o

h) -

f(c)].

Conversely. if this limit exists, prove that f is differentiable at c. 5.31 A vector-valued function f is differentiable at each point of (0, b) and has constant norm I~I!. rove that f(t)· I'(t) = 0 on (0, b). P 532 A vector-valued function f is never zero and has a derivative f" which exists and is continuous on R. If there is a. real function A such that f'(t) "" l(t)(/) for all t, prove that there is it. positive real function u and a. constant vector c such that f(t) = u(t)c for all t. Partial deri1'lltives 5.33 Consider the function / defined on HZ by the following formulas:

x-c

- /(c) _

{(C)i <

1:.

Show that/'

is continuous on (a, b) if this condition holds throughout (a, b).

lex)

5.26 Assume f has a finite derivative in (a, b) and is continuous on [a, b], with a ::;; :$ b for aU x in [a, bland If'(x) I s ex < 1 for all x in (a, b). Prove that I has it unique fixed point in [a, b].
9 having finite derivatives in (0,1),

5.17 Give an example of a pair of functions/and such that

tt». y)

xy
X Z

"-1;1 g(x)
but such that limx_o!'(x)/g'(x)

limf(x)

0,

if (x,

y)

¢ (0,0)

/(0,0)

o.

does not exist. choosing 9 so that g'{x) is never zero.

Prove that the partial derivatives Dt/(x. y) and Dd(x., y) exist for every (x, y) in R2 and evaluate these derivatives explicitly in terms of x and y, Also, show that f is not continuous at (0,0).

126 5.34 Let/be defined on R2 as follows: /(x. y) = y ~ x

~-r

CHAPTER

+ J'

if (x

Y) ¢ (0, 0),

1(0, 0) = O.

Compute the first- and second-order partial derivatives off at the origin, when they exist. C4Jmp1n:-valued fllllCtiCDl 5.35 Let S be an open set in C and let S· be the set of complex conjugates z, where Z E S. If/is defined on S, defineg on S· as follows: g(Z) = 1{Zj. the complex conjugate of /(z). If / is differentiable at c prove that 9 is differentiable at c and that g'(c) = !'(c).
5..36

FUNCTIONS OF BOUNDED VARIATION AND RECTIFIABLE CURVES

6.1 INTRODUcnON Some of the basic properties of monotonic functions were derived in Chapter 4. This brief chapter discusses functions of bounded variation, a class of functions closely related to monotonic functions. We shall find that these functions are intimately connected with curves having finite arc length (rectifiable curves). They also playa role in the theory of Riemann-Stieltjes integration which is developed in the next chapter. 6.2 PROPERTIES OF MONOTONIC FUNCfIONS Theorem 6.1. Let f be an increasing function defined be II + I points such that
011

i) In each of the following eu.mples write / for u(x. y) and v(x, y); a) j{z) = sin z, c) /(z) = [z],
e) /(z) = arg z (z"# 0). g) /(z) = b) fez)

u
=

+ Iv and
cos z,
i, Log z
il

find explicit formulas

d) fez)
f) fez)

=
=

(z"# 0),

(These

h) fez) =.t (<< complex, z '" 0). functions are to be defined as indicated in Chapler L) ii) Show that the Caucby-Riemann equations for tile following values of z : AU z in (a). (b), (g); no z in (c), (d), (e); all z except real z :;;: in (f), (h). 0 (In part (b), the CallChy-R.icmann equations hold for all z if 1% is a nonnegative integer, and they hold for all z "# 0 if II is a negative integer.) the derivative/,(z) in (a), (b), (f), (g), (h), assuming
it exists.

e"',

u and v satisfy

(a, b] and let xo. x I, ...

.=

x"

a=
Then we have the inequality
n-I k~l

Xo

<

XI

<

X2

< '"

< x"

b.

iii} Compute 5.37 Write/

u + ivand assume that/has a derivative at each point of an open disk: D centered at (0, 0). If IlUZ + bv2 is Constant on D for some real a and b, not both 0, prove that / is constant on D.

L: [!(x +)
Q

- f(xc)]

s f(b)

f(a).

SUGGESTED

REFERENCES

FOR FURTHER

STUDY

5.1 Apostol, T. M., ell/elilus, Vol. 1, 2nd ed, Xerox, Waltham, J967.

Proof, AS5umethatYl€(xltXul)' Fori ~ksn l,wehavef(x1+)sf(yJ and J(Yt-)) s f(x_-). so that f(xf. +) - f(xlr.-) :-:;; - f(Ytf(Yl) If we add these inequalities, the sum on the right telescopes to ftV,,-l) - f(yo). Since f(Y"-I) - f(yo) :s; feb) - f(a), this completes the proof.

a·

5.2 Chaundy, T. W., The Differrntial

CalclIllI:J. Clarendon Press, Oxford, (9)5.

The difference f(x. +) f(xk-) is. of course, the jump of fat Xt.. The foregoing theorem tells us that for every finite collection of points xt in (a, b), the sum of the jumps at these points is always bounded by f(b) - f(a). This result can be used to prove the following theorem. Theorem 6.2. Iff countable. is monotonic on [a, b]. then the set of discontinuities off is

Proof Assume thatfis increasing and Jet S .. be the set of points in (11. b) at which the jump off exceeds l [m, m > O. If x I < Xl < ... < x_ -I are in S.., Theorem 6.1 tells us that
-._-

feb) - f(a).
127

128

Dd'.U

TOOIIVariatJon

129

This means tbat Sill must be a finite set, But the set of discontinuities of/jn (0, b) is a subset of the union U:=l S.. and hence is countable. (Iffis decreasing, the argument can be applied to -J.)
tj.J FUNCTIONS

Proof

Assume that x e (a, b). Using the special partition ii(x) - f(a)l
S;

{a, x, b}, we find

If(b) - l(x)1
S;

;$

M.
holds

OF BOUNDIID VARIATION

This implies If(x) - !(a)1 if X = 0 or x = b. Examples

M, I/(x)1

1/(0)\

M. The same inequality

Ihfotititm satisfying

6.3. 1/ [D, b] is D compact interval, a set oj points the inequalities

. J'," a = Xo

{Xlh

xl> ' .• , x~},

X,,-1

< Xl ••• <

<

X~ =

b,

I. It is easy to construct a continuous function which is not of bounded variation. For example,letf(x) = x cos {></(2x)}ifx # 0.1(0) = O. Then f'is continuous on [O.l}. but if we consider the partition into 2n subintervals P
=

is called a partition of [a. b]. The iktelT!Dl [Xl-I> Xl] is called the kth sublntel'VQI of P and we write Ax. = x. - Xl_I' so that l:l~l hl = b - Q. The collection of all pvssible partitiOllS of [Q, b] will he denoted by 9'[0, b].
Defod;tio1l6.1. Let f be defined on [a, b]. If P = {xo, XI' •••• x~} ts a partition of [0, b], write Ali = f(x ..) I(Xi-I), for k = 1.2, ... , n. If there exists a

to, 2n2n-l ....,!. !, l}, _!_. _I_ 32

we have

an easy calculation shows that

positive number M such that

" _,- + - + 6:1. lb.!. I = 2n2ll2n22n2 -t

2~

+ --

+ ... + - + 22

lIt

=1

+ - + ... + - .
2 n

for all partitions of [a, b], then f is said to be of bounded ,'arialioll on [Q. b].

This is not bounded for al1 n, since the series l.::~l(lIn) diverges. In this example the derivative]" exists in (0. I) but/' is not bounded ontO, 1), However,f'is bounded on any compact interval not containing tbe origin and hence [will be of bounded variation on such an interval. 2. An example similar to the fint is given by f(x) x" cos O/z) if x ". 0, [(0) O. This [is of bounded variation on [0, I}. since I' is bounded on [0, 1]. In fact, ['(0) = 0 and, for x ". O,f'(x) sin (I/z) + 2x cos (Ilx), so that If'(x) 1 :::;; 3 for aU x in (O, 1]. 3. Boundcdness of l' is not necessary for [to be of bounded variation. For example, let f(x) = xll3. This function is monotonic (md hence of bounded variation} on every finite interval. However. [,(x) ...... 00 as x _,. O. +

Examples of functions of bounded variation are provided by the next two

theorems. Tlret1rem

ss.
W

Iff

is monotonic on [a, b], thenfis

of bounded variation on [a. b]. of [a, b] we have

Proof. Letfbe increasing. and hence

ft

Then for every partition

ft

Aft ;;,._ 0

L IM.I = .=1 AfA; = l=l [J(x.) 1: 1: t=l

Tieorem 6.6. Iff interior. say 1J'(x) Proof.

- !(Xi-I)]

= !(b)

I(a).

is

s A for

continuous on [a. b] and iff' exists and is bounded in the all x in (a, b). thenf is of bounded variation on [a. Theorem, we have

bJ.

6.4 TOTAL VARIATION Dejinitkm

~=l

Applying the Mean-Value

IAJ,.I

6.8. Let Fbe ofboufldedvariation corresponding to the partition

on [a, b], and felL (P)denote the sum P = {Xo, Xl' •• ,. x.} 0/ [€I. b]. The

AI". = f(xt)
This implies

/(xt-1)

= f'(tt)(x"

- Xi:-l),

number

VJa,b)

= sup

{E(P):

Pe9"[a,b]},

is called the total variation off on the interval [a, b].

NOTE.

When there is no danger of misunderstanding, we will write VI instead of

Theorem 6.7. Iff is 0/ hounded variation on [a, b], say titions 0/[0, b], then f is bounded on [a. b]. In/act, If(x)1

L IAJ,.I

Vr(a, b).
;$

M for a/l par-

s 1/(a)1 +

for all

in [a, b].

Sincefis of bounded. variation on [0, b], the number VI is finite. Also, VI ;;,._ , 0 since each sum L (P) ;;,._ Moreover, V,(o, b) = 0 if, and only if, I is constant O. on [a, b].

Th.6.9

n. 6.JZ

Toad VariatiGrl

ClII.

I.. xl as _ Flmction

of x

131

$0

Tlreoum 6.9. Assume thatf and 9 are each 0/ bounded variation on [0, bJ. are their sum, difference, and product. Also, we have tJnd

Theil

This shows that each sum 1: (PI) and 1: (Pl.) is bounded by V,(o, b) and this means that/is of bounded variation on [a. c] and on [c, b]. From (I) we also ob.aia the inequality because of Theorem 1.1 S. To obtain the reverse inequality, let P = {xo, XU"'' x,,} e @[a, b] and let Po = P u {e} be the (possibly new) partition obtained by adjoining the point c. If c e [Xi _ .. X,,]. then we have
I/(Xl) - f(Xt-I)1 :-:;;If(x') - f(c)1

where A Proof

sup {lg(x)1 : x

[0, b]}.

sup fI/(x)

I:x

[0, b]}.

Let h(x) = /(x)g(x).

FOT

every partition P of [a. b]. we have

IM.I

= =

l/(xJg(X,l:)

f(Xk-l)g(X._I)~

I[/(x.)g(x.)

- /(xt-1)g(Xl)]
- /(xl-1)g(xt_I)]1
;S;

If(c) - f(xa.-I)!,

[/(Xk-l)g(XI)

AIA/kl +

BIA9ll.

This implies that his of bounded variation and that VII ;S; AVj for the sum and difference are simpler and will be omitted.
NOTE.

+ BY,..

The proofs

and hence L (P) ;S; L (Po). Now the points of Po in [a, c] determine a partition PI of [a. c] and those in [e, b] determine a partition Pz of [c, b]. The corresponding sums for all these partitions are connected by the relation

Quotients were net included in the foregoing theorem because the reciprocal of a function of bounded variation need not be of hounded variation. For example, if/(x) -+ 0 as x -+ Xo. then If/will not be bounded on any interval containing XI) and (by Theorem 6.7) I/f cannot be of bounded variation on such an interval. To extend Theorem 6.9 to quotients, it suffices to exclude functions whose values become arbitrarily close to zero.
Theorun 6JO. Let/be 0/ bounded variation on [a, h] and assume that f is bounded away from zero; that is, suppose that there exists a posit;ve number m such that < m :-:;; /(x)!for all x in [a, b]. Then g = II!is also of bounded variation on I [D, h], tJnd V, ;S; VJiml. PrQOf.

r:(P) .:::;; (Po) = r:

L: (PI) + L

(Pz)

:-:;; Via. c)

+ VAc, b).
1:(Pl. Since this

Therefore, VAU. c) + V,,<c,b) is an upper bound for every sum cannot be smaller than the least upper bound. we must have
Via, b) .:::;; V,,(o, c)

V,,(c, b),

and this completes the proof.

6.6 TOTAL VARIADON ON (II, xl AS A FUNCTION OF x

Now we keep the function f and the left endpoint of the interval fixed and study the total variation as a function of the right endpoint. The additive property implies important consequences for this function. TAet1rem 6.12. LeU he 0/ bounded variatitin on [a, h1 Lei V be defined on [a, h] asfollows: V(x) = Vjo, x) if a < x :::;; V(a) = O. Then: b,

6S ADDITIVE PROPERTY OF TOTAL VARIATION

i) V is an increasing function on [0, b].

ii) V-lis an increasing function on [0, b].

Proof. If a < x < Y :5: b, we can write Vf(l.l, y) = Vf(a, x) + ~f(x, y). This implies V(y) - V(x) = Vlx, y) ~ O. Hence V{x) s V(y), and (i) holds. To prove (ii), let D(x) = V(x) - f(x) if x e [0, b]. Then, if 0 :-:;; < y ;S; b, x we have

In the last two theorems the interval [a, b] was kept fixed and Vf(a, b) was considered as a function off If we keep f fixed and study the total variation as a function of the interval [a, b], we can prove. the following additive property.
Theorem 6.11. Let/be of bounded variation on [a, b]. and asStime that c Then f is 0/ bounded rariation on (a, c] and on [c, b] and we have V
EO

(a, b).

Jo, b)

= Via, c)

D(y) - D(x)

V(y) - V(x) - (f(y) - f(x)] = Vf(x. y) - (f(y)

- f(x)).

Vic, b).

But from the definition of Vjx, y) it follows that we have

Proof. We first prove that/is of bounded variation on [a, c] and on [c, b]. Let PI be a partition or [a, c] and let be a partition or [e, b]. Then Po = PI U is a partition of [a, bJ. If 1: (P) denotes the sum L IAli:! corresponding to the

r,

r,
····1 :.j.

fey) - f(x) :::;Vf(x, y). This means that DCy) - D(x) ~ 0, and (ii) holds.
NOTE. For some functions j, the total variation Via. x) can be expressed as an integral. (See Exercise 7.20.)

partition

(of the appropriate interval), we can write

(1)

111 6.7 FUNCTIONS OF BOUNDED VARIATION EXPRESSED DIFFERENCE OF INCREASING FUNcrIONS The following simple and elegant characterization tion is a consequence of Theorem 6.12. of functions AS THE of bounded

TIt. 6.13

Tb. 4i.l5

Adding more points to P can only increase the sum that 0 < Xl - Xu < O. This means that variaand the foregoing

l: IAftl and

hence we can assume

IAft!
inequality

I/(Xl) - f(c)1 < ;,

1'IIeorem 6.13. Let f be defined on [a, b]. Then f is afbounded variation on [a, b] if. and only if,f can be expressed as the difference of two increasing /unctions. Proof. Iff is of bounded variation on [a. b]. we can write f = V - D, where Vis the function of Theorem 6.12 and D = V - [. Both V and D are increasing functions on [a, bj. The converse follows at once from Theorems 6.5 and 6.9. The representation of a function of bounded variation as a difference of two increasing functions is by no means unique. Iff = fl - f2' where f. and 12 are increasing, we also have f = (i. + g) + g), where 9 is an arbitrary increasing function, and we get a new representation off If 9 is strictly increasing. thesame will be true off1 + 9 and/z + g. Therefore, Theorem 6.13 also holds if "increasing" is replaced by "strictly increasing."

now becomes

VIc, b) - ! < ~ +
2 2
since {Xl' But
X2"" ,

l~2

" L: IAftl 5 ~ + VJx" 2

b]. We
therefore

b),
have

.t,,} is a partition

of [Xl'

V~c, b) - Vf(xl, b) < e.

o s Vex,)

Vee) =

V,(a,

Xl) -

V~a. c)

u;,

= VI(e, Xl) Hence we have shown that

VJc, b) - V/(Xh b) < s,

0<

Xl -

C<

implies

6.8 CONTINUOUS Tlu!orem 6.14. Let

FUNCTIONS

OF BOUNDED

VARlt\T10N

of bounded variation on [a, b]. If x E (a. b]. let V(x) = O. Then every point of continuity off is also a point of continuity of V. The converse is also true.

I be

This proves that V(c+) = VCr). A similar argument yields V(c-) = V(c). The theorem is therefore proved for all interior points of [a, b]. (Trivial modifications are needed for the endpoints.) Combining Theorem

VAa, x) and put V(a)

6.14 with 6.13, we can state

Proof. Since V is monotonic, the right- and lefthand limits V(x+) and V(x-) exist for each point x in (a, b). Because of Theorem 6.13, the same is true of

TIIeorem 6.15. Let f be continuous on [a, h]. Then f is 01 bounded variation on [a, b] if, and only if,I can be expressed as the dif{l'rence ()f 111'0 increasing continuous

functions.
NOTE.

f(x+)
If

andf(x-).
Q

<

< Y s b, then os

we have [by definition I/(y) - f(x)J

of VAx, y)]

The theorem

also holds if "increasing"

is replaced by "strictly

increasing."

V(y) -

Vex).

Letting y ~ x, we find

Of course, discontinuities (if any) of a function of bounded variation must be jump discontinuities because of Theorem 6.13. Moreover, Theorem 6.2 tells us that they form a countable set. 6.9 CURVES AND PATHS

o « I/(x+) - l(x)1 s V(x+) - V(x).

Similarly, 0 s If (X) - f(x-)I s V(x) - V(x-). These: inequalities imply that a point of continuity of V is also a point of continuity off. To prove the converse, letfbe continuous at the point c in (a, b). Then, given t > 0, there exists a b > 0 such that 0 < Ix - ci < 0 implies !/(x) f(c) I < 2/2. For this same s, there also exists a partition P of [c, b], say
Xo

e, x.

such that

b,

[a, b] -+ RIO be a vector-valued function, continuous on a compact interval in R. As t runs through [D, b], the function values f(t) trace out a set of points in R" called the graph of f or the curve described by f. A curve is a compact and connected subset of R· since it is the continuous image of a compact interval. The function (itself is called a path. It is often helpful to imagine a curve as being traced out by a moving particle. The interval [a, b] is thought of as a time interval and the vector I(t) specifies the position of the particle at time t. In this interpretation, the function f itself is called a motion.
Let f:

[0, b)

Dd'.6.16

Th. 6.t8

135

Different paths can trace out the same curve, valued functions

For example.

the two complexI,

defined by the equation Al...a, b) = sup {AI...P): P 1/ the set of numbers A/,.P) is unbounded,
It is an easy matter to characterize
E

each trace out the unit circle xl + I, but the points are visited in opposite directions. The same circle is traced out five times by the function h(t) = e10.",

o !:> t !:>

!P[a, b]}.

r is ClJllednonrectifiable.

all rectifiable curves.

O!:>t~1.
6.10 RECTIFIABLE PATIIS AND .u,C LENGTH

Theorem 6.17. Consider a path f: [a, b] -+ RIt with comp(ments , (flO'" .t~). Then f is rectifiable if, and only if. each component fa is of bounded variation 011 [a, b]. If r is rectifiable, we have the inequaiities
V.(a, b) :s; Al...a, b) ~ V1(a, b) where Vl(a, b) denotes the
10101

Next we introduce the concept of arc ICIl8th of a curve. The idea is to approximate the curve by inscribed polygons. a technique learned from ancient geometers. Our intuition tells us that the length of any inscribed polygon should not exceed that of the eurve (since a straight line is the shortest path between two points). so the length of a curve should be an upper bound to the lengths of all inscribed polygons. Therefore, it seems natural to define the length of a curve to be the least upper bound of the lengths of all possible inscribed polygons. For most curves that arise in practice, this gives a useful definition of arc length. However, as we will see presently, there are curves for which there is no upper bound to the lengths of tbe inscribed polygons. Therefore, it becomes necessary to classify curves into two categories: those which bave a length, and those which do not. The former are called rectifiable, the latter lWnrectifiable. We now tum to a formal description ofthesc ideas. Let f : [a. b] -+ R" be a path in R". For any partition of [0, b] given by

+ ... +

v,,(a, b),

(k

1,2•...• n).

(2)

variation 0/ ft on [a. b l
of [a, b] we have

Proof

If P = {to. I), •• " t",} is a partition

J~I

i:

INti) - k(ti-1)!

AI...P) :;:;

I-I

i: t Ih{tt) - /j.,tj-1)1,
J~l

(3)

for each k, All assertions

of the theorem

follow easily from (3).

Examples 1. As noted earlier. the function given by f(x) x cos {n/(2x)} for x "" o. f{O) = 0, is oontinuous but not of bounded variation on [0, I]. Therefore its graph is a nonrectifiable curve. :t. It can be shown (Exercise 7.21) that if I' is continuous on [tl, b J. then f is rectifiable and its arc length can be expressed as an integral. AI-a, b)""

{to, t1,

••••

t..},

1"'(t)11 dt.
OF ARC LENGTH

the points '(to), C(t1)•••• , r(tm) are the vertices of an inscribed polygon. (An example is shown in Fig. 6.1.) The length of this polygon is denoted by MP) and is defined to be the sum A,(P) =

l~l

I:IIf(tt)

..

6.11 ADDITIVE AND CONTINUITY. PROPERTIES Let r = (fl' ... J.. is of bounded

- r(lt-l)lI·

DejWtio" 6.16. If th« set o/numbers Atf..P)is bounded/or all partitions P of [a, b],
then the path f is sajd to be rectifiable and its arc length, denoted by

,/.) be a rectifiable path defined on [a. bJ. Then each component variation on every subinterval [x. y] of [a, b]. In this section we keep f fixed and study the are length A,(x, y) as: a function of the. interval [x, y]. First we prove an additive property.

Ma, b). ;3

Thet1l'em 6JI. Proof

If c e (0, b) we have

Ma, b) =
[a, c]

A,(a, c)

+ At._c, b).
PI of

Adjoining the point c to a partition P of [a, b], we get a partition and a partition P 1. of [c, b] such that A/.P) ~ A/.P1)

+ A/.P2)

A1(a, c)
To obtain

+ At,(c.

b). let

This implies Al,.a. b) :s; A/..a. c) PI and Pl be arbitrary partitions

+ A,.(c. b).
P
=

af[a. c] and [c. b], respectively. P, u P2t

the reverse inequality. Then

1b. 6.19

137

is a

partition

of [a, b] for which we have

At<.P1)

+ AIPz) =

AlP)

s; A,(a, b).
is the sum A,{a. c) If x
E

Since the supremum of all sums A/..Pj) (see Theorem 1.15). the theorem follows.

+ A,{P2)

+ At(c. b)
(a, b], let

the compact set [a, b], Theorem 4.29 tells us that (-I exists and is continuous on its graph. Define U(/) = f-1[g(l)] if t e [e, d]. Then u is continuous on [c. d] and get) = f[u(t)]. The reader can verify that u is strictly monotonic, and hence f and g are equivalent paths. EXERCISES
FlIIICtious of boooded variation

TMorem 6.19. Consider a rectifiable path f defined on [a, b]. s(x) = Alp, x) and let .I'(a) = O. Then we have:

i) The function s $0 defined is increasing and continuous on [a, b]. Ii] If there is no subinteroal of [a, b] on which f is constant, then s is strictly increasing on [0, b].

6.1 Determine which of the following functions are of bounded variation on [0, 1].

Proof Has; x < y S; b, Theorem 6.1 g implies .~(y) - sex) = Al..x, y) ~ O. This proves that s is increasing on [a, b]. Furthermore, we have s(y) - s(x) > 0 unless AJ..x. y) = O. But, by inequality (2), A/..x, .v) = 0 implies V.(x, y) = 0 for each k and this, in turn, implies that f is constant on [x, yJ. Hence (ii] holds. To prove that s is continuous, we use inequality (2) again to write

if x l' 0,/(0) = O. if x #- 0./(0) = O. 6.2 A function /, defined on [a, b], is said to satisfy a uniform Lipschitz condition of order « :> 0 on [a, b 1 if there eMsts a constant M :> 0 such that I/(x) - f(Y)l < Mix - YI'"for all x and JI in [a, b 1. (Compare with Exercise 5.1.)
a) /(x)
x2 sin (llx)

b) I(x) =

.J~ sin (l/x)

a) If/is

IX "'" 1 impliesfis

sllch a function, show that Of :> 1 implies/isconstant of bounded variation [u, b].

on [a, b]. whereas

o :S s(y)

- sex)

= Ab.

y) :::;

1~1

:E Vt(x, y).
Similarly.

It

b) Give an example of a function/satisfying II unlfonn Lipschitz condition of order IX < 1 on [a. b] such that f is not of bounded variation on [a, b]. c) Give an example of a function / which is of bounded variation on [a, b J but which satisfies no uniform Lipschitz condition on [a, b]. 6.3 Show that a polynomial / is of bounded variation on every compact interval [a, b J Describe a method for finding the total variation off on [0, b] if the zeros of the derivative f' are known. 6.4 A nonempty set S of real-valued functions defined on an interval [a, b] is called a linear space of functions ifit has the following two properties: a)

Hwe let y __". we find each term y;.(x, y) _. 0 and hence s(x) = s(x+). x, s(x) = s(x-) and the proof is complete. 6.12 EQUIVALENCE OF PATHS. CHANGE OF PARAMETER

This section describes a class of paths. having the same graph. Let f. [a, b] .... R" be a path in R". Let u: [c, d] .... [a, b] be a real-valued function, continuous and strictly monotonic on [t, d] with range [0, b]. Then the composite function g = fOil given by g(t) = f(Il(I») for
C ;$,

Iff E ~.

then cf

S for every real number c.

s:

d,

b) !ffE S andge

5, then/+

s « S.

is a path having the same graph as f. Two paths f and g so related are called equivalent. They are said to provide different parametric representations of the same curve. The function u is said to define a change of parameter. Let C denote the common graph of two equivalent paths f and g. If u is strictly increasing, we say that f and g trace out C in the same direction. If u is strictly decreasing, we say that rand g trace out C in opposite directions. In the first case, u is said to be orientation-preserving; in the second case, orientationreversing. Tlleorf!m 6.20, LeI f : [a, b] ......RIO and g : [c, d] ..... RK be tWQ paths in R", each of which is one-tv-one on its domain. Then f and g are equivalent ih and only if, they have the same graph.

Theorem 6.9 shows loot the set Y of a.I1 functions of bounded variation on [a, b) is a linear space. If S is any linear space which contains all monotonic functions on [a, b}, prove that V 0;; S. Thill can be described by saying that the functions of bounded variation form the smallest linear space containing aU monotonic functions. . 6.S Let / be a real-valued function defined on [0, 11 such that /(0) > 0, f(x) if;. x for all x, andf(x) ~ fey) whenever x ~ y. Let A = {x :/(x) > z}, Prove that sup A e A and that/<l) > 1. 6.6 If / is defined everywhere in Rt, then / is said to be of bounded variation on (- 00, + 0',) if/is of bounded variation on every finite interval and ifthere exists a positive number M such that Y _,(a, b) < M for all compact intervals {a, b]. The total variation of / on (- 0',), + 0',) is then defined to be the sup of aU numbers Y_,(u, b), - 00 < u < b < + 00, and is denoted by V_,(- ro, + 00). Similar definitions apply to half-open infinite intervals [a. + 00) and ( - co, b]. a) Stale and prove theorems for the infinite interval (- 00, + 00) analogous to Theorems 6.7, 6.9. 6.10, 6.11. and 6.12.

Proof

Equivalent paths necessarily have the same grapb. To prove the converse, assume that f and g have the same graph. Since f is one-to-one and continuous on

13.
b) Show that Theomn 6.S is true for (- 00, + (0) if "monotonic" is replaced by "bounded and monotonic." State and prove a similar modification of Theorem 6.13. 6.7 Assume that f is of bounded variation on [a, b) and let p ~ {xo. x" ... , x,,} e (11 [a, b]. As usual, write Aft. The numbers a) Show that h describes a rectifiable curve c) Show that the set of points S

139

r.

b) Explain. by means of a sketch, the geometric relationship between/. g, and h.

{(x,y):a

S x:s

b, f(x)::;
CI:lIVe

y:s:

g(x)}

= [(xJ

- f(xt-1),

1,2 •.•••

n. Define

is a region in Rl whose boundary is the

r.
0] as roJlows: if a S t S b,

A(P) = {k!

Air. > OJ.

B(P) = (k: 6ft

< O}.

d) Let H be the complex-valued function defined on [a,2b

H(t) = 1 -

ti [g(t)

- f(l)]. t) f(2b t)],

pr.a, b) = sup { L: .lli:

and
n.t<a. b)
blAIn

P E (11[0, b] }

H(t)

=t

+ ti [g(2b

if b

sts

2b ":'

IJ.

Show that H describes a rectifiable curve

E /¥[a,

sup {

hB(P)

L IAlil: P

l}

ro which
.s

is the boundary of the region

f(x)},

So = {(x, y). a S x S b, [(x)

g(x)

2y :s g(x)

are called, respectively, the positive and negative variations off on [a, b]. For each x in (a, b]. let V{x) ee V,,(a. x), p(x) = pr<a. x), n(x} n.t<a. z), and let V(a) = p(a) = ma) = O. Show that we have: a) V(x)
b) 0 pCx)

e) Show that So has the x-axis as a line of symmetry. Syll'Utll!'l'ization of S with respect to the x-axis.}
f) Show that the length of Absolutely

(1he region So is called the

ru does. not

exceed the length of

r.

+ 1I(x).
Vex) and 0 S n(x) S V(x).

c) p and n are increasing on [a, b J. d) [(x) f(a)

s: p(x) s

eootinuous funetions

+ p(x)

A real-valued function [defined on {a, b J is said to be absolutely continuous on [a, b] if for every I! > 0 there is a B > 0 such that
". V(x} - /(x)

II(x). Part (d) gives an alternative proof of Theorem 6.13.

f(o), 2n(x)

e) 2P{x) = V(x)

+ lex) -

+ [(a).

f) Every point of continuity off is abo a point of continuity of p and of n.

Ic~l

L I/(b
R

k)

f(a~)j

<

Ii

(i.8 Let f and

g be

complex-valued functions defined as follows: = e11t1f if t

E

f(t)

[0, 1],

g{t) =

t'2~lt

if t

for every n disjOint open subintervals (Uk' bk) of [a, b], n = 1,2, ... , the sum of whose lengths 1l~1 (b,. - ail is Jess than ,j. AbsoJutely continuous functions occur in the Lebesgue theory of integration and differentiation. The following exercises give some of their elementary properties. 6.11 Prove that every absolutely continuous bounded variation on [a, b 1. function on [a, b 1 is continuous and of

[0,2].

a) Prove that f and g have the same graph but are not equivalent according to the definition in Section 6.12. ,b) Prove that the length of 9 is twice that of /. 6.9 Let f be a rectifiable path of Iength L defined on [0, b]. and assume that f is not constant on any subinterval of [a, b]. Let $ denote the arc-length function given by s(x} "" At(a. x) if a < x ::; b, s(a} O. a) Prove that
[J

NOTE. There exist functions which are coatlnuous and of bounded variation but not absolutely continuous.

6.12 Prove that f is absolutely continuous if it satisfies a uniform Lipschitz condition of order 1 on [a, b}. (See Exercise 6,2.) 6.13 Iff and g are absolutely continuous OD [a. b J. prove tba.t each of the following is also: III. cf(cconstant),f + e.I: g; a1so/tg if gill bounded away from zero.

exists and is continuous on {a, L].

b) Define g(t) = f[s-I(I)] if I E [0, L] and show that g is equivalent to f. Since f(/) = g[s(t)], the function g is said to provide a representation of the graph of f with arc length as parameter. 6.10 Let f and 9 be two real-valued continuous functions of bounded variation defined on [a, b 1. with 0 < f(x} < g(x) for each x in (0, b), f(a) = g(02),[(b) = n(b). Let h be the complex-valued function defined on the interval [a, 2b - oj as follows:
h(t} = t

SUGGESTED

REFERENCES

FOR. FURlHER

STUDY

6.1 Apostol, T. M., Ca/culll1J, Vol I, 2nd ed. Xerox, Waltham, 1967. 6.2 Natanson, 1. P., 17u:ory of Function» 0/ Q Real Boron, translator. Ungar, New York. 1961.
Variable,

Vol. I. rev. ed, Leo F.

if(t}, t

if 0 - t),

:S

_$

b,

h(t) = 2b -

+ ig(2b

ifb:St::;2b-02.

Del. 7.1

141

CHAPTER 7

THE RIEMANN-STIELTJES

lNlEGRAL

become special cases of this more general process. Problems in physics which involve mass distributions that are partly discrete and partly continuous can also be treated by using Stieltjes integrals. In the mathematical theory of probability this integral is a very useful tool that makes possible the simultaneous treatment of continuous and discrete random variables. In Chapter 10 we discuss another generalization of the Riemann integral known as the Lebesgue integral.
7.1. NOTATION

7.1 JNTRODUcnON

Calculus deals principally with two geometric problems: "finding the tangent line to a curve, and finding the area of a region under a curve. The first is studied by a limit process known as differt!1l.li4tkm; the second by another limit processintegration--to which we tum now. The reader will recall from elementary calculus that to find the area of the region under the graph of a positive function f defined on [a, b], we subdivide the interval [a, b] into a finite number of subintervals, say n, the kth subinterval having length LU.., and we consider sums of the fonn :E:;1 /(tkJ Ax... where I., is some point in the kth subinterval. Such a sum is an approximation to the area by means of rectangles. If / is sufficiently well behaved in [a. b}-continuous, for example-then there is some hope that these sums will tend to a limit as we let n -. 00, making the successive subdivisions finer and finer. This, roughly speaking, is what is involved in Riemann's definition of the definite integral J:f(x) dx. (A precise definition is given below.) The two concepts, derivative and integral, arise in entirely different ways and it is a remarkable fact indeed that the two are intimately connected. Hwe consider the definite integral of a continuous function f as a function of its upper limit, say we write F(x)

For brevity we make certain stipulations concerning notation and terminology to be used in this chapter. We shall be working with a compact interval (0, b] and. unless otherwise stated, all functions denoted by J, g, a, P. ete., will be assumed to be real-valued functions defined and bounded on [a, b]. Complex-valued functions are dealt with in Section 7.27, and extensions to unbounded functions and infinite intervals will be discussed in Chapter 10. As in Chapter 6, a partition P of [0, b] is a finite set of points, say
P such that a = Xo < Xl < X._I < X~= b. A partition P' of [a, b] is said to be finer than P (or a refinement of P) if P s;;; P', which we also write P' ;2 P. The symbol A.~ denotes the difference A.a. = ll(xJ - a(xl-1). so that

<.. ...

= lxo. Xl>

•••

x,,}.

1~1

" L: All" = a.(b)

""7 1%(0).

The set of all possible partitions of [a, b] is denoted by £'[0, b]. The norm of a partition P is the length of the largest subinterval of P and is denoted by IIPII. Note that
P';2 P

J." /(I) e.

implies

IIP'II ~ IIPII.

then F has a derivative and F'(x) = f(x). This important result shows that differentiation and integration are, in a sense. inverse operations. In this chapter we study the process of integration in some detail, Actually we consider a more general concept than that of Riemann: this is the RiemannStieltjes integral, which involves two functions f and «. The symbol for such an integral is J:f(x) da:(x), or something similar, and the usual Riemann integral occurs as the special case in which C!(x)= x. When a has a continuous derivative, the definition is such that the Stieltjes integral J:f(x) d«(x) becomes the Riemann integral J!f(x) a'(x) fix. However, the Stieltjes integral still makes sense when 1% is not differentiable or even when 1% is discontinuous. In ract, it is in dealing with discontinuous exthat tbe importance of the Stieltjes integral becomes apparent. By a suitable choice of a discontinuous a, any finite or infinite sum can be expressed as a Stieltjes integral, and summation and ordinary Riemann integration then 140

That is, refinement of a partition decreases its norm, but the converse does not necessarily hold.
7.3 THE DEfINITION OF THE RIEMANN-STIELTJES INTEGRAL
[6,

DejinititJlJ 7.1. Let P = {Xo. Xl' ... ,x ..} be a partition of point in the subinterval [Xk- t s Xl]' A sum of the form S(P.j. a) =

b] and let

11 be

l~ I

" :E j(IJ

AOCl

is called a Riemann-Stieftjfs sum off witll respect to ce. We say f is Riemannintegrable with respect to ex on [a, b], and we write "fE R(a) on [a, b]" if there exists a number A hailing the following property: For every I: > 0, there exists a partition p. of [0. b] such that for every partitwn P finer than r, and for every choice of the points tt in [Xt-l, xJ. we have IS(P.J, ex) - AJ < s,

141

n.7.1

Def,'U

141

When such a number A exists, it is uniquely determined and is denoted by r~f da.or by j!f(x) d«(x). We also say that the Riemann-Stiehjes integral da. existtJ. The functions f and IX are referred to as the integrand and the integmtor. respectively. In the special case when a(x) = x, we write S(P,/) instead of S(P,f, IX}, and Ie R instead ofle R(IX). The integral is then called a Riemann integral and is denoted by dx or by J:f(x) dx. The numerical value of J:f(x) da.(x) depends only onJ.«, ii, and b, and does Dot depend on the symbol x. The letter x is a "dummy variable" and may be replaced by any other convenient symbol.

$!I

TlluuIII7.4. Assume that e e (a, b). If two of the three integrals in (I) exist, then the third also exists and we have (I) Proof. If P is a partition
P' = P of
Ii

J!f

[a, h] [D, c]

such that c e P,let

and

P"

f"l

[c, b],
The Ric-

This is one of several accepted definitions of the Riemann-Stiel1jes An alternative (but not equivalent) definition is stated in Exercise 7.3.
NOTE.

integral.

denote the corresponding partitions of [a, c] and [c, b]. respectively. mann-Stieltjes sums for these partitions are connected by the equation

S(p./. IX)

S(P',f, IX)

+ S(P",f.
£

a).

7," LINEAR PROPER'I1ES It is an easy matter to prove that the integral operates in a linear fashion on both the integrand and the integrator. This is the context of the next two theorems. 17tetwt!1fI7.2. If Ie R(a.) and if 9 e R(a) 011 [0, b]. then [a, b (for any two COlUUlllttJ Cl and C2) and we have

(cd + e2g)

da.

Cl

f
II

f dCl +

Cz

cd + C29 e R(<<) on
9 da..

P: of [a, c] such that IS(P'. t.a.) and a partition jS(P"'f, Cl) -

Assume that J~f dll and

J~fch exist.
f

Then, given

> 0, there is a partition

dal

<~

whenever P'

is finer

than

p;.

P: of [c, b] such that

J-

<

wbenever

r: is finer than

P~.

Proof. Let h

= cd + clg.
l-l

Given a partition
All}

P of [a, b], we can write All,

so, h. IX)

h(t.t)

c1

tKl

E f(tJ
!X).

+ Cz

l=J

g(t,J

06.1%1;

Then P; = P; u P: is a partition of [a, b] such tbat P finer than p. implies P' ;i2 P; and P" ;;2 P~. Hence, if P is finer than p., we can combine the foregoing results to obtain the inequality

= cIS(P,J, a)

+ c2S(P, y,

IS(p,f, IX) -

f dlX -

s:

I dlXl < 11.

Given E; > O. choose P; so that P ~ P; implies IS(P, J, a) - J:f dill choose P: so that P 2 P: implies IS(P, g, a) g drxl < s, If P, = P~ v.p;, then, for P finer than P., we have

S(P, h, If) - c1

J:

< s, we

and take

I da. -

'2

This proves that I da. exists and equals f dCl + fda.. The reader can easily verify that a similar argument proves the theorem in the remaining cases. Using mathematical induction. we can prove a similar result for a decomposition of [a, b] into a finite number of subintervals.

J:

J;

J!

dl%l ~ lei I! + lc~IB,

and this proves the theorem. Theorem 7.3. life

(for any two constants c] and fd(Cli:t.

R(a) andfe
C2)

RCP) on [a, b], then/ER(c1a. and we have

1:1

+ cd!)

f~fda
C2fI) on [0, b]

NOTE.

The preceding type of argument cannot be used to prove that the integral exists whenever J:l da. exists. The conclusion is correct, however. For integrators a of bounded variation, this fact will later be proved in Theorem 7.25.

Ida

c1

fdlJ.

DefilliliOi' 7.5. If a < b, we define We also define J:f ds: = O.

The equation in Theorem

J:I d« = - J!f

da: whenever

$:1da.

exists.

The proof is similar to that of Theorem 7.2 and is left as an exercise. A result somewhat analogous to the previous two theorems tells us that the integral is also additive with respect to the interval of integration.

s:

7.4 can now be written as follows:

f d«

1'"fda

I d«

= O.

144 7.5 INTEGRA nON BY PARTS

Th.7.6

Th.7.7

Reduction to a Ri_

lDtegraJ

145

b = g(d).

Let hand fJ be the composite functions defined as follows:

hex)

A remarkable connection exists between the integrand and the integrator in a Riemann-Stieltjes integral. The existence of f drr. implies the existence of a df, and the converse is also true. Moreover, a very simple relation holds between the two integrals.

J!

S:

= [[g(x)],

{J(x) = a[g(x)],

if
That is,

XES.

Then h

R(jJ) on S tnd we have

S!f

da.

=
c

S~h dp.

TlIeol'em 7.6. Iff

R(a:) an [a, b]. then « e R(j) on [a, b] cmd we h4ve

!(x) d«(x)

f.

fI<d)

J(t) dlX(t)

fd

f[g(x)]

d{fX[g(X)]}.

g{,,)

a(%) dl(x)

= f(b)o:(b) - J(I1.)«(I1.). law for the integral, is P, of [a. b]

NOTE.

This equation. which provides a kind of reciprocity known as the formula for integration by parts.

Proof.

such that

Let IS> 0 be given. Since J:fdlX exists, there is a partition for every P' finer than P .. we have

Isept.!, a) Consider an arbitrary Riemann-Stieltjes

Proof For definiteness, assume that 9 is strictly increasing on S. (This implies , c < d.) Then 9 is one-to-one and has a strictly increasing, continuous inverse g-l defined on [a. b]. Therefore, for every partition P = {Yo, ... ,Ya} of [c, d], there corresponds one and only one partition P' = {xo, ... , x,,} of [a, b] with xA = g(y.). In fact. we can write
P'

g(P)

and

J dal < E.
sum for the integral

(2)

J:

IX

dJ, say

SCPo a./)

" = :E

~=I

a(tJ 611t. =

k=l

" :E a.(tJJ(xJ -

J<s

" L a:(tJJ(Xk-l),
1

Furthermore, a refinement of P produces a corresponding refinement of P', and the converse also holds. If I: > 0 is given. there is a partition P~ of [a, b] such that P' finer than P; implies IS(P',J, a:) - J:fdal < 8. Let P; = g-I(P;) be the corresponding partition of [c. dJ, and let P = {Yo •...• Y.} be a partition of [c. dJ finer than p•. Form a Riemann-Stieltjes sum S(P, h, fJ) = where
Xi

where P is finer than p a- Writing A

= f(b)lX(b)
"

- f(ap(a),

we have the identity

A
Subtracting A-

:E!(x~xJ
t= 1

II

- L!(Xi-l)a(Xi-1).
k= 1

k=1

" :E

h(uJ AjJ",

the last two displayed

S{P, a:,!)

LJ(xJ[a(xk)
1-1,

equations,

we find

Ul e [Yl-I> Yt] and AfJ~ = (J(y1t,) {1(Yl-I)' If we put tlo; g(Uk) and g(y.), then P' = {xo, ... , x.} is a partition of [a, b] finer than P;. Moreover,

- (X(li)]

~
LJ(Xl-l)[a:(tt)
l=l

we then have

- a(xt_1)].

S(P, h, (J)

The two sums on the right can be combined into a single sum of the form S(P' ,t. IX). where P' is that partition of [a, b] obtained by taking the points x. and lit. together. Then pi is finer than P and hence finer than p •. Therefore the inequality (2) is valid and this means that we have

" = EJ[g(Ut)]{a[g(y~)]
1=1

- a[g(Yl-l)]}
= S(P',J, IX),

=
since f" € proved.
NOTE.
(Xk-I,

k~l

" L J(lt){O:(Xl)

- a:(xt-1)}

IA whenever P is filler than p.. and equals A da.

S(P, IX,/)

xJ.

Therefore,!

S(P, h, fJ) -

J: f dal

<e

and the theorem

is

dill

<

E, that

s:r

But this is exactly the statement

J: a: df

exists

This theorem applies, in particular, to Riemann integrals, that is, when x, Another theorem of this type, in which g is not required to be monotonic, will later be proved for Riemann integrals. (See Theorem 7.36.)

I1(X)

7.6 CHANGE OF VARIABLE IN A RIEMANN-STIELTJES

INTEGRAL

7.7 REDUCTION

TO A RIEMANN

INTEGRAL

Theorem 7.7. Let f « R(a) on [a, b] and le: g be a strictly monotonic continuous [unction defined on an. interval S having endpoints c and d. Assume that a gee).

The next theorem tells us that we are permitted to replace the symbol dlX(x) by a:'{x) dx in the integral S:f(x) da(x) whenever IX has a continuous derivative «'.

146

11t. 7.8

111.7.9

147 AS INTEGRATORS

Theorem 7.8. Assume f e R(a) an [a, b) and assume that IX has a continUQUS derivative rT: on [a, b]. Then the Riemann integral f(x)a'(x) dx exists and we iuJve

J:

7.8 STEP FUNCtIONS

If

at

is constant throughout [a, b]. the integral

J!f d(J. exists and

has the value O.

f(x) da(x) = f.f{x):xt(X)

dx.

Proof. Let g(x)

= f(x)a'(x)
S(P, g)

since each sum S(P,/. a) = O. However, if a is constant except for a jump discontinuity at one point, the integral da: need not exist and, if it does exist, its value need not be zero. The situation is described more fully in the following

J:l

and consider a Riemann sum

theorem: Axi• Theorem 7.9. Given a < c a(e), a:(b)are arbitrary;

and a(x) a.(b)

1-1

" L g(/l)

Axk

" = L /(tJa'(tJ
t=1

< b. Define a on [a, b] as follows: The values a(a), if a :s; x < .C, if
c

The same partition P and the same choice of the Riemann-Stieitjes sum
S(P,J. a) =

til;

can be used to form the

a(x) = «(a)

l-l

L f(tt)

"

<

x:s; h.

Aocll'

Applying the Mean-Value Theorem, we can write

Let 1 defined on [a, b] in such a way that at least one of the functions / or IX is be continuous/ram the left at c and at least one is continuousfrom the right at c. Then / e R(a} on [a, b] and we have

and hence S(P, J, a) - S(p. g) =

dct = f(c)[ot(c+) - a{c- )].

k=!

" L /(tl)[a'(vl)

- a:'(tJ] ':uk'

NOTE. The result also holds if c = a, provided that we write aCe) for a(c-), and it holds for c = b if we write a(e) for a(c+). We win prove later (Theorem 7.29) that the integral does not exist if both f and 0: are discontinuous from the right or

Since f is bounded, we have I/(x)1 5: M for all x in [a, b], where M > O. Continuity of a.' on [a, b] implies uniform continuity on [a, bJ. Hence, if I> > 0 is given, there exists a ~ > 0 (depending only on e) such that

from the left at c. Proof. If c E P, every term in the sum S(P, /, a) is zero except the two terms arising from the subinterval separated by c, say
S(P,/. tx) !(tt-l)[CC(C) a(c-)]

os

[x -

yl <

implies

la'(x) - a'(Y)1

<

2M(b - a)

+ /(fk)[a(c+)

- «(c)].

If we take a partition P; with norm IIP;n < t;, then for any finer partition P we will have la'(v.} ~ ,,'(tt)I < 1l/[2M(b ~ a)] in the preceding equation. For such P we therefore have
IS(P,J, a) - S(P, g)!

where

tl;-1 S;

c s;

ti•

This equation can also be written as follows:

= [/(tll;-l)

- /(c)] [o:(c)- «(c-)]

+ U(tJ ~ /(c)][a.(c+ ) Hence we have I/(liJ - /(c)lla(c+)

a(c)],

where A = S(P,j. 1%) - I(e)[a(c+)

- a(e-)J.

< 2"

IAI s 1/(tk-1)

/(c)IIa:(c) - «(c-)I f(c)1 < e and

- a(c)l.

If/is continuous at c, for every e > 0 there is a 15

> 0 such that liP]] < 15implies

s,

On the other hand, since IE R(a) on [a, bJ, there exists a partition P; such that p finer than P:implies

I/(t"-I)

1/(tJ - l(c)1 <

- a(e)l.

IS(p,/,

Il)

In this case, we obtain the inequality

/dal < ;.
= P~ u P;,

IAI :s:: ela(e) -

a(c-)I

+ I>]a.(c+)

Combining the last two inequalities, we see that when P is finer than P, we will have IS(P, g) - J!I docl < s, and this proves the theorem.
NOT!'!.

A stronger result not requiring continuity of a' is proved in Theorem 7.35.

But this inequality holds whether or not/is continuous at c. For example, if/is discontinuous both from the right and from the left at c, then «(c) = I%(c-) and o:(e) = a(c+) and we get A = O. On the other hand, if fis continuous from the left and discontinuous from the right at c, we must have aCe) = a:(c+) and we,get

148

net'.

7.10

1b.1.13

149

IAI ~ eIOl(c) - OI(C- )1. Similarly. if f is continuous from the right and discontinuous from the left at c, we have «(c) = I%(c-) and IAI s tIIX(c+) - lX(c)l. Hence the last displayed inequality holds in every case. This proves the theorem. Theorem 7.9 tens us that the value of a Riemann-Stie1tjes integral can be altered by changing the value of f at a siogle point. The following example shows that the existence of the integral can also be affected by such a change. Let
Example. «(x) = 0,

discontinuous from the right have

01'

from the left at each. Xi' Then

$:1 da exists

and we

f(x) dct(x)

=~

!(X,clUl'
of integrals of the type

Proof. By Theorem 1.4, considered in Theorem 7.9.

r:f da can be written as a sum

ih if
=

i' 0, «(0) = -1, l~x~+1.

I(x) = 1, In this case Theorem 7.9 implies

/(x) = I if x ;f;. 0, we can easily see that

O. But if we re-define I so that /(0) = 2 and exist. In fact, when P is a par. tition which includes 0 as a point of subdivision, we find
1

J:

I d«

One of the simplest step functions is the greatest-integer junction. Its value at x is the greatest integer which is less than or equal to x and is denoted by [x]. Thus, [x] is the unique integer satisfying the inequalities [x] S X < [x] + L Tlreorem 7.12. Every finite sum can be written as a Riemann-Stieltjes fact, given 0 sum L:~J Q1. define/on [0, n] as follows: f(x) Then
=

f: 1 I drJ will not

integral. I"

S(P.J. II)

= J(r.J[«(x~)
ee

- «(0)] + 1(1.._.)[01(0) - OI(X~_l)]

a~

ifk - 1 <

s; k

(k

1.2, .... n), 1(0)

= o.

f(tl)

JOt-I),

where x,,_ 2 ~ t 0 ~ tk ~ x.. The value of this sum is 0, 1, or - 1, depending on the choice of t" and 'k-l' Hence, l~l / th does not exist in this case. However. in a Riemann integral J! J(x) dx, the values of 1 can be changed at a finite number of points without affecting either the existence or the value of the integral. To prove this, it suffices to consider the case where I(x} 0 for aU x in [a. b] except for one point. say x c. But for such a function it is obvious that IS(P,.n1 :5 If(c)1~PII. Since IIP~can be made arbitrarily small, it rollows that i:J(x) dx = O.

'Ii_

t i1
a" :;
s; x.

f(k)

f:

f(x) d[x],

where [x] is the greatest integer

Proof

The greatest-integer function is. a step function, continuous from the right and having jump I at each integer. The function/ is continuous from the left at 1,2 •...• n. Now apply Theorem 7.1 L 7.to EULER'S SUMMATION FORMULA

7.9 REDUCTION

OF A RlEMANN-STIELTJES

INTEGRAL TO A FINITE SUM

The integrator in Theorem 7.9 is a special case of an important class of functions known as step functions. These are functions which are constant throughout an . interval except for a finite number of jump discontinuities.

Definition 7.10 (Step IlIlICtion). Afunclion ;f there is Q par! ilion

II defined

on [0. b] is catted a step function

We shall illustrate the use of Riemann-Stieltjes integrals by deriving a remarkable formula known as Euler's summation formula, which relates the integral of a function over an interval [a, h] with the sum of the function values at the integers in [a, b]. It can sometimes be used to approximate integrals by sums or, conversely, to estimate the values of certain sums by means. of integrals.

such that II is constant on each open subinterval (Xl-" Xl)' The number ,,(Xl +) rx(x,,-) is called the jump at Xl; if I < k < I'. The jump at XI is :x(x1 +) - :X(x1). and the jump at X. is cc(x,,) - o:(x. - ).
Step functions provide the connecting and finite sums: link between Riemann-Stieltjes integrals

Tlwornn 7Jj (Ellier's sll1fUlUllion/orfllJlla). [a, h]. then we have

Q~b

fen)

f(x) dx

If/has

a continuous derioatioe

f' on

j'(x)«(x»

dx

+ f(o)(a»

- 1(b)(b»,

where «x»

x - [x).

T6eorem 7.I1 (Rehctioll of (J RiemaM-Stieitjes inteflrtll to ajitdte sum}, Let Cl be a step/unction defined on [a, b] with jump 7k al xt, where Xl' ... , are as described in Definition 7.10. Let f he defined on [a, b] in such a ~l'ay tluu not both f and ('(are

x.

];;1(n) ==
NOTE.

When

a and

f(x) dx

b are integers, this becomes

f'(x) (~ - [x] -.

dx

+ f(o) ; feb) .

L.,,~~~ means

the sum from n

[a]

1 to"

[b].

n,7.HI

151

we have

Since the

function has unit

at the

ask: What the value of upper sums? This leads us to consider the inf of all upper sums, a number called the upper The defined be the sup of all lower sums, for functions) both these will be m'F"'''''~ wiU be different and it becomes an will ensure that the disCJJSSthis type of
b] and tet

If

combine thi£
INTEGRATORS, UPPER AND

sup inf The numbers

for that.

,x
'XE

7.11 MONOTONICALLY INCREASINf; LOWER L"'lTEGRALS

The

and

IX) =

k~ i

" l:
with respect

appear in the Riemanna vital role ill the fl"lJ>'II~"_ we shalluse the abbreviation "a? 011

.;;::-' and we if ? em from which it follows that the lower sums do not exceed.the upper sums, Furthermore, if I. E j) then
We

when a r, we have tile

«)

the upper and lower sums to the '·<'11 '''''! "" sums. do not which are used in the material that
l!l""C"~UI"

.,,,,,,,,!.\,,,,,,i

function,

The next theorem shows for (1., refinement of the increases the lower sums and decreases the upper sums, Theorem 7.15. Assume that
IX?

on

Then:

than P, we have and For any two

we have

151

Dd'.7.16

Th. 7.19

153

Proof. It suffices to prove (i) when P' contains exactly one more point than P, say the point c. If c is in the ith subinterval of P, we can write U(P',f. (X) =

Iom1

:E Mk_(f)

Then for every partition P of [0, 1], we have Mt(f) = 1 and m~(f) = 0, since every subinterval contains both rational and irrational numbers. Therefore, U(P, j) = 1 and L(P, 0 for all P. It follows that we have, for {a, b] [0, 1],

.da,l; + M'[o:(c)

«(Xi-I)]

M"[oc(xJ - o:(e)].

l""

and

f dx

o,
=

where ME and M" denote the sup of/in and

[X'-I'

e] and [e; x;). But, since

M"

s M!f),

Observe that the same result holds if [(x) = 0 when x is rational, andf(x) irrational. 1.12 ADDITIVE AND LINEARITY LOWER INTEGRALS Upper and lower integrals ample. we have PROPERTIES OF UPPER AND

1 when x is

we have U(P',/. «) :::;:U(P.j, ct). (The inequality for lower sums is proved in a similar fashion.) To prove (ji), let P = PI U Pl' Then we have L(Pt.!.
NOTE.

a.)

s L(P,f,

0:)

U(P. f. a.) S U(P2.f.

a.).

share many of the properties

of the integral.

For ex-

It follows from this theorem that we also have (for increasing a)

m[o:(b) a(a)] :::;: (PI,f. L «) S U(P2,f,
0:) ~

M[(X(h) - a(a)], if a < c < b, and the same equation holds for lower integrals. However. certain equations which hold for integrals must be replaced by inequalities when they are stated for upper and lower integrals. For example, we have

where M and m denote the sup and inf of/on Definition 7.16. Assume that «» on [a, b]' respect to 0: is defined as follows:

[a, b].
The upper Stieltjes integral

all

with

s: f
f

da = inf {U(P,J,

«): p EO 9'[a, b]}.

and
E

U + g}
+ g)

da

s: f

d«

J:

g da.

The lower Stiettjes integral is similarly defined;

da.

sup {L(P,f,

a): P

.'P[a, b)}.

f (/

d« ~

S: f

da

g d«.

NOTE.

We sometimes write J(f. a.) and fU 0:) for the upper and lower integrals. In the special case where t.J:(x} = x, the upper and lower sums are denoted by U(P,f} and L(P.f} and are called upper and lower Riemann sums. The corresponding integrals. denoted by J:f(x) and by f(x) dx, are called upper and

These remarks can be easily verified by the reader.

(See Exercise 7.11.)

ax

J:

lower Riemann

integrals,

They were first introduced

Of)"

by

J.

7.13 RIEMANN'S

CONDITION

G. Darboux «).

(1875).

1'1Ietwt!m 7.17. Assume that

Proof. If 8

on [a. bJ. Thenl(f. a.)

s 1(f.

>0

is given, there exists a partition U(P1.f. <X)

PI such that

By Theorem 7.15. it follows that L(P, f. a). Hence, l(f. Of) ~ l(f, [(f. a) S J(f. «). EDmple. define/on

tu: a) + s is an upper
Of)

< J(f. a)

+ e,

If we are to expect equality of the upper and lower integrals, then we must also expect the upper sums to become arbitrarily close to the lower sums. Hence it seems reasonable to seek those funetions j for which the difference U(P,f, IX) L(P,J, a) can be made arbitrarily small.

a, and, since

1l

bound to all lower sums is arbitrary, this implies

Definition 7.18. We say that f satisfies Riemann's eondttton with respect to rx on [a, b] if,Jor every Il > 0, there exists a partition P; such that P finer than P; implies

os
Theort!m 7.19. Assume thai
equivalent: i)

U(P,J, a)

LJ..p.f,

e) < s,

It is easy to give an example in which l(/' «) < [0, J} as follows: [(x) = 1, if x is rational, f(x)
=

us. «).

Let «(x) = x and

«»

on [a, b]. 'Then the following three statements are

0,

if

.r is irrational.

f e R(a)

on [0, b).

151

n.1.)'
7.14 COMPARlSON DIEOKEMS

iii) lif, 1%)

ii) Isatisfies Riemann's condition with respect to a on [a, b]. = J(f, a).

Proof. We will prove that part (i) implies (ii), part (0) implies (iii), and part (ill) implies 0). Assume that (i) holds. H a(b) = «(a). then (ti) holds trivially, so we can assume that «(a) < a(b), Given Il > 0, choose P, so that for any finer P and all choices of and t~in [Xi-I> x,J, we have

TIteouIII7.1O. Assume that «» on (a. b]. lIfe R(<<) and g e R{rz) on [a. b) and flAx) s g{x)/or all x in [a, b]. then we hatJe

't

r
I;:r

f(x)

da:(x);o;;

r
•

g(x) da:(x). Riemann-Stieltjes sums satisfy

Proof. For every partition

S(P,/. a) =

P, the corresponding

Itf(tJ
1=1

Aa& ~

AI

< !!_ 3

and we find since and Xi]}, it follows that

• E/(tJ ~
1

;0;;

:where A

= J!I

.t-I

E g(tJ.!\al
follows

S(P,

e. a),
g(x) :<!: 0

da. Combining

these inequalities,

jt U(t.) ~ I(I.J) &:tl

i=l

a'" on [a, b].

From

tbis the

theorem

< ~ s.
3
[Xt-h

In particular, this theorem {J)' on [a. b].

implies that

r. g{x) da(x) 2 0
1!(x)1
da(x).

easily.

whenever

Since M.. (f) - m.(n = sup fl(x) - Ax'): x. x' in for every h > 0 we can choose t" and t~so that

TMot'e1ll 7.21.

[a, b] and we have the i1tequality

Auume that

fl./'

on [a, b].

f{tJ -

l(tfJ > M/c(f) -

Int.(/) - h.

Making a choice corresponding to h = te/[a(b) - I%(a)]. we can write U(P,I, (I) L(P,/. a)

If

I{x) d!Z{X)1
of Definition

5;

Ifl e R(IX) 011 [a, b]. then

III e R(f1.) 011

=
<

•-1

L [Min

"

Proof. Using the notation

1.14, we can write

- mk(f)] ~
f(tl)] At'll

Hence, (i) implies (ii). Next, assume that (ii) holds. He> 0 is given, there exists a partition P, such that P finer than p. implies U(P, J, IX) < L(P.J, a) + e. Hence, for such P we have

" L: [J(tJ .~I

1i~1

L Aa4

..

M1(f) - m.(f)

= sup {f(x) - f(y) : x, y in [XI:_1,xJ} .

S;

< e.

we

Since the inequality have

1I/(x)1 - II(Y)11

I/(x) - f(y)1 always holds, it follows that

- ml;{f).

Milll) - ma(l/l) ::;; M..(f)

on k, we obtain

MUltiplying

by AIr,!; and summing

U(p.

J(f. It)

S;

U(P,f, cr.) < L(P./, IX)

+s

III. fl.)

L(P.I/I.

fl.) ::;; U{p.f. a.) -

L(p.f. «).

S;

I(/' a.}

+ e.

That is, 1(f, a) ,s; l(f, a) + Il for every t > O. Therefore, I(f, a.) S; l(f. 12:). But, by Theorem 7.17. we also have the opposite inequality. Hence (il) implies (iii). Finally. assume that [(f, a) = l(f, tx) and let A denote their common value. We will prove tbat cia: exists and equals A. Given,fl > 0, choose P; so that U(P,j, Il) < l(f, a) + e for all P finer than P;. Also choose P~ such that

for every partition P of (0. b]. By applying Riemann's condition, we find that III € R(I2:)on [a, bJ. The inequality in the theorem follows by taking g = III in

Theorem 7.20.
NOTE.

J!I

The converse of Theorem

a/'

7.21 is not true.

(See Exercise 7.12.)

1ieorem 7.22. Assume that

on [a, b]. Ille

L(P,J, IX) for all P finer than

> l(f, al - s

[a, b].
Proof. Using the notation
of Definition

R(a.) on [a, b]. thenp

€ R(<<)

on

P;.
I:

If P,

= p~ v P:, we can write

1.14, we have and

l(f. a) -

< L(P,/,

ct:)

S;

S(P,f, (.()::;;U(P,f, IX) < Itf, a)

+t

Ml(f~)
Hence

[Ml(I/IW

m..(p) = [mA;(lfI)]2.
- mil/l))

for every P finer than P,. But, since l(f, a) = 1(f, a) = A, this means that IS(P, f, (.() A I < Ii: whenever P is finer than This proves that f tkt exists and equals A, and the proof of the theorem is now complete.

we

can write

r;

J:

M,,(f"} - mi(f~) = [MiG/I)

+ mt<:lfl)][MA;(lflJ

::;;2M[M,,(I/I) - m.,(lfl)].

Th 7.2.3

n.7.25

157

where M is an upper bound for the conelusien follows. X_rem

Ilion [a, b].

on [a, b]. II

By applying

Riemann's

condition,

To prove the first inequaJity,

we note that
-

Alk n

IAIX.I;;.: 0 and hence

7.21.

Assume that
E

IX?

the productf-g

R(a:) on [a.

bJ.

IE

R(a) turd g € R(a) on

[a. b].

then

~~l

L: [Mol(/)

mt(f)](AVt

IAalD

S;

2M

1;=

:E (AY.
I

- IAeltl)

Proof, We use Theorem 7.22 along with the identity

2/(x)g(x)
=

= 2M (V(b)

k~1

i: iAllkl)
< A},

<

t.

[I(x)

+ g(xW

[{(X)]2

- [g(x)r·

To prove the second inequality, let

A(P)

{k: AIX .. 2: O},

E

R(P)

{k: ~

7.1S INTEGRATORS

OF BOUNDED

VARIATION

In Theorem 6.13 we found that every function IX of bounded variation on [a, bJ can be expressed as the difference of two increasing functions. Jfr;x = 01:1 az is such a decomposition and if Ie R(al) andfE R(lXz) on [a, bJ. it follows by linearity thatfe R(a.) on [a, b]. However. the converse is not always true. Hfe R(a:) on [a. b], it is quite possible to choose increasing functions al and (X2 such that Cl ()!1 01::2, but such that neither integral drx" da.2 exists. The difficulty, of course, is due to the ncnuniqeeness of the decomposition a a} - 1%2' However, we can prove that Qi:ere is at least one decomposition for which the converse is true, namely, when a'l is the total variation of IX and 0:2 = a1 - a. (Recall Definition 6.8.)

and let h

= !J:./ Y(b).

Hk

A(P), choose tl; and

t; so that

f(t,;) - I(t~) > M,,(/) - m.(f) - II; but, if k e B(P). choose Then
It and Ii so that

f(tD - 1(/,,) > Mk{f) - m,,(/) - h.

J:I

J:f

It.= I

L [M.(f) -

..

m.U)] IAIl.1 <

k,4(P)

:E

[f(t.)

f(t.OJ fAa..1
- 1(t.)]

+ IEB(I',)[J(tt) L
=
i~l

I~I + h

.-1

L IAall

TMonm 7.24. Assume that a is o/boundedvariiltion on [a, b]. Let V(x) denote the total variation of IX on [D, x] if a < x s; b, and let Yea) = O. Let I he defined and bounded 011 [a, bJ. lIfE R(a) all [a, bJ, then/E R(V) on [a, b]. Proof, If V(b) 0, then V is contant and the result is trivial. Suppose therefore, that V(b) > O. Suppose also that rf(x)1 s; M if x EO [a, b]. Since V is increasing, we need only verify that/satisfies Riemann's condition with respect to Von [a, b]. Given e > O. choose P~ so thit for any finer P and all choices of points til; and
t~ in

L [J(t

l)

f(t.t)J Ar.rl
t

+
8

11

.~l

L IAall

<
It follows thatfe
NOTE.

8 + hV(b) = -- +

442

= -.

R(V) on [a. b].

[Xl ~ 1. Xt] we

I~

have

[f(tJ

- f(t0]

&x.j < ~

and

V(b) <

t~1

1: lAIlll

+ ....£.. •
4M

This theorem (together with Theorem 6.12) enables us to reduce the theory of Riemann-Stieltjes integration for integrators of bounded variation to the case of increasing integrators. Riemann's condition then becomes available and it turns out to be a particularly useful tool in this work. As a first application we shall obtain a result which is closely related to Theorem 7.4. Th«I'Dn 7.25. Let «. he 01 bounded variation on [a, b) and assume that [a, b]. Then I E R(r.r) on every subinterval [c, dJ 0/ [a, b].

For P finer than p. we will establish

the two inequalities

t;; [MJ<U)
"
and
which, by addition,

f e R(Il)

on

- mk(f)](AJIi. -

lAllA!) < ;,

yield U(P, J. V) - L(P,

J.

V) < e.

Proof. Let V(x) denote total variation of a on [D, x], with Veal = O. Then a = V - (V - IX). where both V and V - 0% are increasing on [a, b (Theorem 6.12). By Theorem 7.24,/e R(V), and hencefe R(V - «) on [a, b]. Therefore, if the theorem is true for increasing integrators, it follows that f E R( V) on [c, d] aud/e R(V - cr) on [c, sofe R(r.r) on [c. dJ.

too

dJ,

Th.736

'lb. 7.211

15'1

Hence. it suffices to prove the theorem when ('f.~ on (a, b]. By Theorem 7.4 it suffices to prove that each integral J~fda and J:fdCl: exists, Assume that a < c < b. If P is a partition of [a, x], let A(P, x) denote the difference A(P. x) = U(P,f, a) - L(P,/. IX), of the upper and lower sums associated with the interval [0, x], Since f e R(ex) on [a, b], Riemann's condition holds. Hence, if e > 0 is given, there exists a partition P, of [a, b] such that A(P, b) < I; if P is finer than P•• We can assume that C E r; The points of r, in [a. c] form a partition r; of [0, c]. If P' is a partition of [a, e] finer than P:. then P = P' v p. is a partition of [0, b] composed of the points of P' along with those points of p. in [c, b]. Now the sum defining A(P', c) contains only part of the terms in the sum defining A(P, b). Since each term is ~O and since P is finer than p .. we have
A{P', c) 5; A(P, b)

Therefore, if M,
IS(p,j, G) -

5;

Mil

t. S:.

r
=

sup {lg(x)1 ; x
f·g dal

I~f:-,
=

[a, b]}, we have

{I(t.) - f(t)}g(t)

1m.) - f(t)1 drz(t) ::::;M.

t l~-l

da.(t)1

[MAU) - mi(f)] drt{t)

M,{U(P,f,

a.) - L(P.!, a:)}.

Since f E R«('(). for every I': > 0 there is a partition P, such that P finer than P, implies U(P./. fl.) L(P,/. a.) < 1;. This proves that je R(G) on [a. b] and that J:f' 9 da. = J:f dG. A similar argument shows that 9 e R(F) on [a, b] and that J:f'g da = dF.

S:g

NOTE.

Theorem 7.26 is also valid if

(l

is of bounded variation on [0, b].

<

e. 7.16 SUFFICIENT CONDmONS INTEGRALS FOR EXIS'lENCE OF RIEMANN-STIELTJES

That is, P' finer than P; implies A(P', c) < e. Hence, f satisfies Riemann's condition on [a, e] and f~f da. exists. The same argument, of course, shows that f:f da. exists, and by Theorem 1.4 it follows that J~fda. exists. The next theorem is an application of Theorems 7.23, 7.21, and 7.25.
Tleonm Define 7.26. Assume fe R(o:) and 9 e R(o:) on [a, h], where (Xl"'

In most of the previous theorems we have assumed that certain integrals existed and then studied their properties. It is quite natural to ask: When does the integra] exist? Two useful sufficient conditions will be obtained.
1'Ieorem 7.27. Iff is continuous on [a, b] and

F(x) and

on [a, b].

then f e R(a) on [a. b].

NarE.

if

II is

of bounded variation on [a, b],

f(t) da.(t)

By Theorem 7.6, a second sufficient condition can be obtained by interchangingfand a in the hypothesis.

G(x)
Then

i""

Proof.

g(t)drx(t)
E

ifxe[a,b].

IE

R(G), 9

R(F), and the product f·g

f(x)g(x)

da.(x)

=
=

Proof

[a, b] we have
S(P,j,

The integral

s:r 9 drx exists by Theorem 7.23.

_t.f(tJ

r r

R(t>:)on [a. b], and we have

n» dG(x)
g(x) dF(x).

It suffices to prove the theorem when 0: l' with a(a) < a(b). Continuity of f on [a, b] implies uniform continuity, so that if 1: > 0 is given, we can find Jj > 0 (depending only on 8) such that Ix - yl < ij implies If(x) - f{y)1 < etA, where A = 2[a:(b) - a(a)]. If P, is a partition with norm lIP.1I < a, then for P finer than p. we must have Mt(f) - fflk(f) 5; £'/A, since Mk(J) - m.(J) = sup (f(x) - f()-') ; x, y in [x._!, XI]}' inequality by.IJ.t'l.and summing, we find
U(P,f,

Multiplying the

a) - L(P.f,

ex)

For every partition P of

i: A
£, k=1

A2:k

-<£ 2 '

and we see that Riemann's condition holds. Hence.j e R(~) on [a, bJ. For the special case in which a(x) corollary:
Theorem 7.28. Each of the following Riemann integral x, Theorems 7.27 and 7,6 give the following conditions is sufficient for the existence of the

G)

=
6

i"''
~-J

get) dt>:(t) =

and

f f(x)g(x)
Q

drx(x)

e~ b; 1,,-.
~

At J~k-l ex"
1

f(t~)g(t)

da(t),

f(l)g(l)

drx(t).
,,,:

J: fix)

dx: b) f is of bounded variation on [0, b].

a) fis continuous on [a, h].

lb. 7.19

Th.7.32

Integnalas a Fwdioo of the IlItenal

161

7.17 NECESSARY CONDmONS INTEGRALS

FOR EXISI'ENCE

OF RIEMANN-S'I1ELTJES

m :>; c :>; M such that

When !:I: is of bounded variation on [a, oJ, continuity of/is sufficient for the existence of J:/ da, Continuity of / throughout [a. h] is by no means necessary, however. For example. in Theorem 7.9 we found that when a: is a step function, then/can be defined quite arbitrarily in [a. bJ provided only thatfis continuous at the discontinuities of 11. The next theorem tells us that common discontinuities from the right or from the left must be avoided if the integritl is to exist. TMOum 7.19. Assume that al" on (0, bJ and let a < c < h. Assume further that both 11 and / are discontinuous from the right at x = e; that is, assume that there exists em a > 0 such that for every Ii > 0 there are values 0/ x and y in the interval (e, c + 5) lor which
I/(x) - /(c)! ;:::: £ / are discontinUOus from the left at c. Proof. lA:t P be a partition of [a, b] containing c as a point of subdivision and form the differ:mce U(P,f, a) L(P.J. a:) = and I(I(Y) -

I(x) drz(x) = c

da(x) = c[a(b) - a:(a)]. c

In particular, if/is

continuous on [a,

bJ, then

= I(xo)/or

some

Xo

in [a,

bJ.

Proof If a(a) = a(b), the theorem holds trivially, both sides being O. Hence we .can assume that tx(a) < a(b). Since all upper and lower sums satisfy m[a:(b) - 1):(0)] S L(P,f, «) :>; U(P,/, a) :>; M[ll(b) - o:(a)],

da must lie between the same bounds, Therefore, the quotient (J!fda)/(J! da.) lies between m and M. When/is continuous on [0, bJ, the intermediate value theorem yields c = f{xo) for some Xo in [0, bJ.

the integral

J:/

a(e)1 ;::::II.

A second theorem of this type can be obtained from the first by using integra-

Then the integral J:/(x) th(x) cannot exist. The integral also/ai13 to exist

if a: and

tion by parts.
Tluorsm 7.31 (&cvml Muut-Yalue T1teorem for RiellJlllla-Stiehje.s integrflls). Assume that a; is continuous and that f /' on [a. b]. Then there exists a paint x(J
I !

in [a, b] such that

1:=1

L [M.(f)
-

- mlU)]

,6l.1l·

s:

f{x) da(x)

= I(a)

fO

da:(x)

+ f(b)

fo

da(x).

If the itb subinterval has c as its left endpoint, then

U(PJ. «) L(P,J, a} ;;>; [M;(f)

proof

By Theorem 7.6, we have

mil)][«(xJ - a(c)],

since each term of the sum is ;:::: If c is a common discontinuity from the right. O. we can assume that the point Xi is chosen so that a(XI) - «(c) 2!:. e, Furthermore. the hypothesis of the theorem implies M IJJ
U(P,

S:

f(x)

da:(x) = I(b)a:(b) - f(a)a:{a) -

a(x) df(x).

Applying Theorem 7.30 to the integra] on the right, we find

mlJJ:l!: e. Hence,

J. a:)

- L(P,

J. «)

:l!:e2, where Xo
E

f(x)

drx(x) = f(a)[a(xo)

rx{a)J

+ J(b)[a(b)

- l.I(xo)],

and Riemann's condition cannot be satisfied. (If e is a common discontinwty from the left, the argument is similar.)
7.18 MEAN-VALUE THEOREMS FOR RIEMANN-STlELTJES INTEGRALS

[a, b]. which is the statement we set out to prove.

AS A FUNCTION OF THE INTERVAL

7 .r 9 THE INTEGRAL

Although integrals occur in a wide variety of problems, there are relatively few eases in which the explicit vaJue of the integral can be obtained. However, it often suffices to have an estimate for the integral rather than its exact value. The M eon Value Theorems of this section are especially useful in making such estimates.
1'IIurem 7.JO(Fir. Metlll- V4Iae Tbttoremfar Rk~ie/tjtl, itIIegrllb). Assume that I%/' and let f « R{a) on [a, hJ. Let M and m denote, respectively, the sup and inf of the set {f(x): x e [a, bJ}. Then there exists a reall1umher e satisffing

Iff E R(a) on [a, b] and jfa: is of bounded variation, (hen (by Theorem 7.25) the integral J:/ da exists for each x in [a, b] and can be studied as a function of x. Some properties of this function will now be obtained.

[0, b].

Theorem 7.12. Let IX IN: 0/ bounded variation on [a, b] and assume that / Define F by the equation F(x) =

e R(a;) on

1""f d«,

if x e [a, b]'

· 161

Th.7.33

a.a..: of Variable
(a, b) and has the value

1()J

Then we have.'
i) F is of bounded variation on [a, Ii) Every point of continuity of
It

b].
Q

g'(x) point of continuity of F.

= f(x)

for every x in (a, b).

and g(b-) exist and satisfy g(b-). g(b) - g(a).

is also

iii) If «» on [a, b], the derivative F'(x) exists at each point x in (a, b) where a;'(x) exists and where f is continuous. For such x, we have F'{x) Proof
that F(J') - F(x) =
= f(x}«'{x).

At the endpoints assume that g(a+)

g(a) - g(o+)

= g(b) -

Then we have

It suffices to assume tbat a.? on [0, b]. If x ¢ y, Theorem 7.30 implies

f
~

f(x)

dx

g'(x) dx

J d« =

Proof
c[a.(y) - a(x)].

For every partition of [a, b) we can write

g(a) =

where m ::; e ::s;; M (in the notation of Theorem 7.30). Statements (i) and (ii) follow at once from this equation; To prove (iii), we divide by y - x and observe that c ~ fix) as y ~ x. . When Theorem 7.32 is used in conjunction with Theorem 7.26, we obtain the following theorem which converts a Riemann integral of a product f" 9 into a Riemann-Stieltjes integral J:f riG with a continuous integrator of bounded variation. Tkorem 7.33. lffe F(x)

gCb)

l~ 1

L: [g(x.)

- g(Xl-1)]

1;= 1

L: 0'(/,,)

..

Axl

L:f(t,)
k= 1

Ax...

where tic is a point in (Xt-l' xi) determined by the Mean-Value Theorem of differential calculus. But, for II. given B > 0, the partition can be taken so fine that 19(b) - g(a) -

/(x) dxj

l~f(tJ

4xk -

n» dxl < s,

R and 9 e Ron [0, b], let

J(t) dt,

G(x)

and this proves the theorem.

The second fundamental theorem can be combined with Theorem 1.33 to give the following strengthening of Theorem 7.&,
get) dt

ifx e [a, b].

Also,

fe

Then F and G tire continuous fimcti()ns of bounded f1QJ'iation on [a, b]. R(G) and 9 E R(F) on [a, b], and we haoe

f(x)g(x)

ax

f(x)

dG(x)

Tlietlrem 7J5. Assume fER 011 [a, b]. Let a.be Q junctUm which is cominuous on [a, b] and whose derivative a.' is Riemann integrable ,oil [a, b]. Then the foll,owing integrals exist tmd are equal:

g(x) dF(x}.

s:

/(x) drz(x) =

s:

f(x}:l'{x)

dx.

Proof Parts (i) and (ii) of Theorem 7.32 show that F and G are continuous functions of bounded variation on [a, b]. The existence of the integrals and the two formulas for J:f(x)g(x) dx follow by taking a(x) x in Theorem 7.26.
NOTE.

Proof. By the second fundamental theorem we have, for each x in [a, b],
«(x) - t.(a) =

I%'(t) dt.

fundamental

When a:(x) = x, part (iii) of Theorem 7.32 is sometimes called the first theorem of inteqral calculus. It states that F'(x) = fix) at each point of continuity off A companion result, called the second fundamental theorem, is given in the next section.
UO SECOND FUNDAMENTAL TIIEOREM OF INTEGRAL CALCULUS

Taking 9
NOTE.

a:' in Theorem 7.33 we obtain Theorem 1.35.

A related result is described in Exercise 7.34.

7.11 OJANGE OF VAR.IABLE IN A RIEMANN INtEGRAL

The next theorem tells how to integrate a derivative.

on [a, b]. Let 9 be a junction

The formula $:/ drz == J~h d{J of Theorem 1.7 for changing the variable in an integral assumes the form

Tlu!orem 7.34 (S«OM folldtmtelltl.d 'h#Ore/ll of illtegr. ~). Assume tiuu fER defined on [a. b] such that the derivative g' exists in

0(1' f(x)

dx =

fd
<

f[g(t)Jg'{t)

dt,

f(~)

164

TIt.. 7.36

Th.7.37

Secoo4 Mean-Value Theorem. for ruemaon Integrals

165

when «(x) = x and when 9 is a strictly monotonic function with a eontinuous derivative g'. It is valid if J e R on [a. b]: When J is continuous, we can use Theorem 7.32 to remove the restriction that 9 be monotonic. In fact, we have the following theorem: 7.36 (C~ of lUfridle ill " R1emt:ura intq,./). Assume that 9 has a continuous derivative g' on QII interval [c, d]. Let f be continuous on g([e, and define F by the equation

Iheou",

dJ)

In particular, the result is valid if g(c) = g(d). This makes the theorem especially useful in the applications. (See Fig. 7.2 for a permissible g.) Actually, there is a more general version of Theorem 7.36 which does not require continuity of J or of g', but the proof is considerably more difficult. Assume that II E'R on [c, d] and, if x e [c, d], let g(x) = h(t) dt. where a is a fixed point in [c, d]. Then if Ron g([c, d]) the integral S~![g(t)] h(¥) dt exists and we have

r«

S;

F(x)

JI'I<:)

r'"

f{t) dt

if x

g([ c, d]).
dt exists QIId has the value

9(~)

f(x) dx

Id J[g(t)]h(t)
c

dt.

IIle)

Then, fOT each x in [c, d] the integral S~f[g(t)]g'(t) F[g(x)]. In particular; we have

l'ld)

l(x) dx

11l[g(t)]g'(t)
c

dt.
on [c, d]

This appears to be the most general theorem On change of variable in a Riemann integral. (For a proof, see the article by H. Kestelman. Mathematical Gazette, 45 (1961), pp, 17-23.) Theorem 7.36 is the special case in which" is continuous on [c, d] andJis continuous on g([c, d]). 7.22 SECOND MEAN·VALUE THEOREM FOR RIEMANN INTEGRAlS Theorem 7.37. Let 9 be continuous and assume that two real numbers salis/ring the inequalities

,(c)

Proof Since both g' and the composite function fog are continuous the integral in question exists. Define G on [c, d] as follows: G(x)

f'

f»

on [a, b]. Let A and B be

J[g(t)]g'(t)
By Theorem

dt. 7.32, we have

A :sJCo+)
Then there exists a point

and

B ;;:f(b-).

We are to show that G(x) = F[g(x)].

G'(x) = f[g{x)]g'(x), and, by the chain rule, the derivative of F[g(x)] is also f[g(x}]g'(x), sinceF'(x) = j(x). Hence, G(x) - F[g(x)] is constant. But, when x c, we get G{c) = 0 and F[g(c)] = 0, so this constant must be O. Hence, G(x) = g(x)] for all x in [c, d]. In particular, when x = d, we get G(d) = F[g(d)] and this is the last equation in the theorem.

i)

r
I"
til

Xo

in [a, b] such that B

f(x)g(x) dx

1"" g(x) dx +
r g(x) dx, J;.:o
b

f:

g(x) dx,

Fr

In particular,

if /(x)

;;: 0 for all x in [tl. b]. we have

ii)

f(x)g(x)

dx = B

where

XI) €

[a,

bJ.

NOTE. Some texts prove the preceding theorem under the added hypothesis that g' is never zero on [c, d], which, of course, implies monotonicity of g. The above proof shows that this is not needed. It should be noted that g is continuous on [c, d]. so g([c. dJ) is an interval which contains the interval joining g{c) and g(d).

NOTE.

Part (ii) is known as Bonnet's theorem.

Proof

If oc(x)

J: get)

dr, then oc' = g, Theorem

f(x)g(;x) dx

= f(a)

I'·

7.31 is applicable,

g(x) dx

+ feb)

f:

and we get

g(x) dx.

Figure'.:!

This proves (i) whenever A = f(a) and B = f(b). Now if A and B are any two real numbers satisfying A :s f(a+) and B ~ J(b-). we can redefine/at the endpoints a and b to have the "Values/(a) = A andJ(b) = D. ThemodifiedJis still increasing on [a. b] and, as we have remarked before, changing the value off at a finite number of points does not affect the value of a Riemann integral. (Of course. the point Xo in (1) will depend on the choice of A and D.) By taking A = O. part (ii) follows from part (i),

166

'l'L7.l1

n.7.41 7.24 DIFFERENTIATION UNDER THE INTEGRAL SIGN

167

7.23 ItIEMANN-STIELTJES INTEGItALS DEPENDING ON A PARAMETER

Theorem 7.38.

Let f be continuous at each point (x. y) 01 a rectangle

I_rem

7.40. Let Q = {(x, y) : a

011

:s: x :s: b. c :s: Y :s: d}.

bounded variation

[a, b] and,for

Assume that a: is of each fixed y in [c. d], assume that the integral

f(x, y) : a ~

X~

h. c ~ y ~ d}..

.As$U11lethat IX is of bounded oariati(}1I tm [a. b) und let F be the fimction [c. d) by the equation F(y) Then F is continuous on (c, =

F(y) "'"

defined on

lex, y)

da.(x),

f(x, y) drx(x).

If the partial derivative D,./ is continuous on Q, the derivative F'(y) exists for each y in (e, d) and is given. by
exists.
F'(y)
NOTE.

d].

In other words.

if Jlo e
lim !(x.
'-+10

[c.

d]. we have

D:J(x, y) dll(x).

lim
'-10

f.b lex,
"

y)

dct{x)

=
=

r
J..

e"

y) d«(x)

In particular, when 9
F(y)

R on [a, b] and a:(x) = and

rt»

g(x)/(x. y) dx

J: get) dt. we get

g(x) Dzf(x, y) dx.

I(x, Yo) drx(x).

Proof. If Yo E (c, d) and y .. Yo. we have

F(y) - F(yo) y - Yo ~

Proof Assume that «I' on [a, b). Since Q is a compact set, I is uniformly continuous on Q. Hence. glven 8 > 0, there exists' a ;; > 0 (depending only on e) such that for every pair of points z = (x, y) and z' = (z', y') in Q with Iz - z' J < lJ. we have lAx, y) Ax. )"')1 < B. If Lv- )"'1 < lJ, we have
IF(y) F(y')1 s

5." I(x,
"

y) - I(x. Yo) da(x) Y - Yo

f.b
•

D:J(x.

Y) drx(x).

where j is between y and Yo. Since Dzf is continuous clusion by arguing as in the proof of Theorem 7.38. 7.25 INTERCHANGING

on Q. we obtain the con-

I/(x,

y) -

lex, y')1 d!X(x) s 6[tx(b) - !X(a)].

THE ORDER OF INTEGRATION

This establishes the continuity of F on [c, d].

Of course, when a(x) = x, this becomes a continuity theorem for Riemann integrals involving a parameter. However, we can derive a much more useful result for Riemann integrals than that obtained by simply setting o:(x) = x if we employ Theorem 7.26.

Tlleo"", 7.41. Let Q = {(x, y) : a :s: x S b. c ~ y s d}. Assume that oX is 01 bcunded variation on [a. b], P is 01 bounded variation on [c, d], and f is continuous on Q. If (x, y) € Q, define
F(y) Then
FE

I(x, y) da.(x),

G(x)

lex, y)

dP(Y).

.[a, b].

TMonm 7.39.

If I is

then the function

conf;nu01lS on the rectangle [a, b] x [c, d]. and F defined by lire equation .F(y)

if g e R

on

R(P) lm [c.

d]. G E R(a)

y(x)f(x.

)I) dx;

In other words. we may interchange the order 01 integration as follows:

is continuous on (c, d]. Tha: is, lim

1-tol0

if Yo

[c, d). we harJe

Yo) dx.

f.6 g(x)f(x, » dx = f.b g(x)/(x,

• ..

r [f

on [a, b], and we have

F(y) dP(y)

G(x) dix(x) .

!(x, y) dP(Y)] da.(x) =

f U:

f(x,

y) da(X)] dP(y)·

Now apply Theorem 7.38.

Proof. If G(x)

= J~g(t)

dt, Theorem 7.26 shows that F(y)

J!f(x, y) dG(x).

Similarly.

Proof. By Theorem 7.38, Fis continuous on [c, d] and hence F E R(P) on [c. d]. G E R(a) on [a, b]. To prove the equality of the two integrals, it suffices to consider the case in which a » on [a, b] and P » on [e, d].

168 By uniform continuity, given e > 0 there is aD> points z = (x, y) and z' = (x'. y') in Q. with Iz -

7.41

Def.7.43

16

z'l < &,

0 such that for every pair of we have

Tllurem 7A2. Let f be continuous on th« rectangle [a, b] x [c, d]. If 9 [a, b] and if heR on [c, d], then we have

lAx,

y) - f(x'.

i)1 < e.
[0. b] and [c, d]

Let us now subdivide Q into n1 equal rectangles by subdividing each into n equal parts. where n is chosen so that and Writing (d - c)

r U:

Ron

g(x)h(y)f(x,

y) dY] dx =

f U:

g(x)h(y)f(x,

_v)

dX]

dy. 7.26

Proof. let <lex) =

and 7.41.

J~g(u) du and

let P(y)

= H h(v)

dv, and apply Theorems

-n-

< .../2"

7.26 LEBFSGUE'S

CRlTERlON FOR EXISTENCE

OF RIEMANN INTEGRAlS

x.

=a

+ -'----'n n. we have

k(b -a)

and

for k == 0, 1, 2, ...•

f \Jcr"
b/
II

I(x, y) d{J(y» d~x) = ~

l~O J

't f"'' +1 (f'Jl'JJ+I

0 "'" (Xb

f(x,

y) dfJ(y}) dlX(x}.

We apply Theorem

7.30 twice on the right.

The double sum becomes

.-1 .-1
i::;JiliO JAO

L L f(x;.

yj)[p(Yj+l)
'

- P(YJ)][<r(xu I) - G1(xJ].

Every continuous function is Riemann integrable. However, continuity is certainly not necessary, for we have seen thatfE R whenfis of bounded variation on [a, b]. In particular.j" can be a monotonic function with a countable set of discontinuities and yet the integral J!f(i) dx will exist. Actually, there are Riemann-integrable functions whose discontinuities form a noncountable set. (See Exercise 7.32.) Therefore. it is. natural to ask "how many" discontinuities a function can have and still be Riemann integrable. The definitive theorem on this question was discovered by Lebesgue and is proved in this section. The idea behind Lebesgue's theorem is revealed by examining Riemann's condition to see the kind of restriction it puts on the set of discontinuities off. The difference between the upper and lower Riemann sums is given by

where (x;, yj) is in the rectangle vertices. Similarly. we find

Qt,j having

Y)

and

(X.l:+1.

Yj+l) as opposite

"~1

" L [Mi(f)

mi(f)}

Ax",

f (1" If

f(x,

y)

d~X»)
.. -I

dP(y)
,,-1

= where (x;, yj)

E Qk,j'

k=O

L L f(x;, yj)[,6(Yj+
j=O

I) - P(YJ)][G1(XU1)
E

- G1(xJ].

But If(x~, Y;) - f(x';., yj)1 <

G(x) da(x) -

and hence

F(y) dP(y)1

.. -1

<s
=

JEO

L [fJ(YJ+I)

,,-1

P(y)]

l~O

L [IX{X.Hl) -

and, roughly speaking, / will be integrable if, and only if, this sum can be made arbitrarily small. Split this sum into two parts, say SI + S2' where 51 comes from subintervals containing only points of continuity of J, and 82 contains the remaining terms. In S1> each difference Mt{f) - m.(f) is small because of continuity and hence a large number of such terms can OCCurand still keep SI small. In 8:1.' however, the differencesM.(f) - mi.(/) need not be small; but because they are bounded (say by M), we have IS21 :;;; M so that 82 will be small if the sum of the lengths of the subintervals corresponding to 8z is small. Hence we may expect that the set of discontinuities of an integrable function can be covered by tervals whose total length is small, This is the central idea in Lebesgue's theorem. To formulate it more precisely we introduce sets of measure aero,

Lax"

IX{xJ]

~[P(d) - p(c)][iX(b) - a(a)].

Defotititm 7.43. A set S of real numbers is mid to halJemeasure zero if, for every ~> 0, there is a counlab1tt covering of S by open intervals, the sum 0/ w/wse lengths i~ less than e. If the intervals are denoted by (ak'

Since e is arbitrary.

this implies equality of the two iategrals. with Theorem 7.26 gives the following result for Rie-

btl,

the definition

requires that (3)

Theorem 7.41 together mann integrals.

U (a
t

i•

bJ

and

t'7G

Th. 7M

1Ia. 7.48

171

If the collection of intervals is finite, the index k in (3) runs over a finite set. If the collection is countably infinite, then k goes from I to co, and the sum of the lengths is the sum of an infinite series given by lim
N""oo

only on 8) such that for every closed subinterval T whenever the length of T is less than h. Proof. For each x in [a, b] there exists a l-ball B"

<;;;

[a, b], we have

nAT) < e:

1:-1

L (bl.

B(x; 6..) such that e.

aJ.

Besides the definition, we need one more result about sets of measure zero. TMOTt!lft 7.44.

Let F be a countable collection of sets in R, say

F = {Fl' Fl.' ... }.

each of which has measure zero. Then their rmion S also has a measure zero.
=
~~I

U r;

..,

The set of all half size balls B(x; ~>J2) forms an open covering of [a, b]. By compactness, a finite number (say k) of these COVel' [a. b]. let their radii be 8,,2 •...• ~k{2 and let Jj be the smallest of these k numbers. When the interval T has length <0, then T is partly covered by at least one of these balls, say by B(x,,; 6,J2}. However, the ball B(x,,; h,,) completely covers T (since ij, 2: 20). Moreover. in B(x,.; lJ,,) n [a, b] the oscillation of lis less than e. This implies that !lAT) < £ and the theorem is proved .

Theorem 7.47. Let set J. asfollows:

f be defined and bounded on [a. b]. For each e > 0 define the
I. {x: x
E

Proof. Given 8 > 0, there is a countable covering of Fl by open intervals. the sum of whose lengths is less than 8/2t. The union of all these coverings is itself a countable covering of S by open intervals and the sum of the lengths of all the
intervals is less than .

[0,

bJ•.

Wf(X) 2: e},

Then J. is (J closed set. Proof. Let x be an accumulation point of Je. If x ¢ I.. we have w_,(x) Hence there is a l-ball B(x) such that
n_,(B(x) ('\ [a, b])

< e.

&=12'"

'" L _!_

= e.

<

e.

EumpIes. Since a set consisting of just one point has measure zero, it follows that ewry countable subset of R has measure zero. In particular, the set of rational numbers has measure zero. However, there are lIMountable sets which bave measure zero, (See Exercise 7,32.) Next we introduce
DefinitiOll

Thus

DO

points of B(x) can belong to J.., contradicting

point of I.. Therefore.

accumulation

the statement that x IS an e J. and J. is closed.

the concept of oscillation.

TMorem 7.48 (Lebesglle's uiteritJ1f for Rielfltfllll-iRtt!f/rability). Let f be defined and bounded on [a, b] and let D denote theset aldiscontinuities olfin [a. b]. Then feR on [a, b] if, and only if, D has measure zero.

7.45.

Let f be defined and bounded on an interval S. (lAT)

number

II T

S;

S, the

Proof (Necessity). First we assume that D does not have measure zero and show
thatfis not integrable. We can write D as a countable union of sets

sup {fix) - fly) : x € T, YET}, where D.

is called the osc;llation of Ion T. The oscillation off at x is defined to be the number mAx)
NOTE.

lim OAB(x; h) n 8). ,,~o+

This limit always exists, since OiB(x; h) n S) is a decreasing function of Jr. In fact. T, <;;; Tz implies nATI) s n.r<T2). Also, w1(x) = 0 if. and only if, lis continuous at x (Exercise 4.24),

{x: wjx)' ~ ;}.

, ,
,

If x e D, then wAx) > 0, so D is the union of the sets D, for r = l, 2, ... Now if D does not have measure zero, then some set Dr does not (by Theorem

[a, b], then O_,(T) <

The next theorem tells us that if wf(x) < II at each point of a compact I; for all sufficiently small subintervals T.

interval

7.44), Therefore, there is some £ > 0 such that every countable collection of open intervals covering D. has a sum of lengths ~e. For any partition P of [a, b] we

Tlrreor~1II7.46. Let I be defined and bounded on [~ b], and let e > 0 be given. Assume that wAx) < Bfor erJeryx in [0, b). Then there exists Q 6 > 0 (depending

have U(P,f) - L{P.f)

l;1

" L [Mt(f)

- m.(f)) 4xt

Sl

+ S2 ~

S"

111

1'h.1M

Th.7..50

173 on

where SI contains those terms coming from subintervals containing points of D in their interior, and 82 contains the remaining terms. The open intervals from SI cover Dr except possibly for a finite subset of D .. which has measure 0, so the sum of their lengths is at least e. Moreover, in these intervals we have I 6 M.(f) - m..(f) ~ and hence SI ;;:0; - •

that a bounded functionfon a compact interval [a, b) is Riemann integrable [a, b] if, and only if./is continuous almost everywhere on [a, b]. The following statements

are immediate consequences T/lurellt 7.19. a) If b) If/E

(some of which were proved earlier in the chapter) of Lebesgue's theorem. bounded variation on [a, b], then fER
E

I is of

on [D, b].

This means that

e U(P.J) - 4.P,J) ~ • r
P, so Riemann's condition cannot be satisfied. Therefore f is integrable. In other words, iff E R. then D has measure zero. (Sufficiency). Now we assume that D has measure zero and show that the Riemann condition is satisfied. Again we write D = I' Dr. where Dr is the set of points x at which wt<x) ~ I/r. Since Dr :!;;;: D. each D. has measure 0, so Dr can be covered by open intervals, the sum of whose lengths is < l/r. Since D"is compact (Theorem 7.47), a finite number of these intervals cover Dr' The union of these intervals is an open set whlch we denote by A•. The complement Br = [a, b] - A, is the union of a finite number of closed subintervals of [a. Let J be a typical subinterval of B,. If x E I. then wt<x) < llr so, by Theorem 7.46, there is a 6 > 0 (depending only on r) such that I can be further subdivided into a finite number of subintervals T of length <6 in which O.,(T) < l/r. The endpoints of all these subintervals determine a partition P, of [a, b]' If P is finer than P; we can write for

[a, b].
c) lffe

III E R and f2

R on [a. h). thenfe ROil [e, d]for E R on [D, b]. Also, f·g

every subinterval [c,

R on [0,

bJ

d] c [a, h], whenever g e R on

9 is bounded away

not

every partition

R andg e R on [a, b], then_fjo e R

011

[a,

U~

d) Iff and 9 are bounded functio1l8 having the same t/is(xmtinuities on [a. bJ. then fe R on [a, b] if, and only if. 9 e R on [0, b]. e) Let 9 e R on [a, b] and amune that m s g(x) ;5; M for all x in [a. b]. Iffis continuous on [m, M], the comp08;tefunction h defined by hex) = f[g(x)] is Riemann-integrable on [D, b].
NOTE.

from O.

b) whenever

oJ.

Statement (e) need not hold (See Exercise 7.29.)

if

we assume

only that

fe R

on

Em, M].

7.27 COMPLEX·V ALum

RIEMANN-sTIELTJF.S

INTEGRALS

U(P,j)

- [{PJ)

t=1

.. :E [M,(J)

- m.(f)] Ax. = SI

S2.
of

where S, contains those terms coming from subintervals containing points Dr> and S2 contains the remaining terms. In the kth term of S2 we have and hence Si Since A. covers all the intervals contributing

b-a < --.

r

to SI. we have

S 1:$
where M and m are the sup and inf U(P.f)

M
---

r m

off

on [D, b]. Therefore

Riemann-Stieltjes integrals of the form da.. in which / and ex: are complexvalued functions defined and bounded on an interval [a, b are of fundamental importance in the theory of functions of a complex variable. They can be introduced by exactly the same definition we have used in the real case. In fact. Definition 7.1 is meaningful when f and IX are complex-valued. The sums of the products f(tJ[ct(xJ - ct(Xt-I)] which are used to form Riemann-Stieltjes sums need only be in_terpreted as sums of products of complex numbers. Since complex numbers satisfy the commutative. associative, and distributive laws which bold for real numbers, it is not surprising that complex-valued integrals share many of the properties of real-valued integrals. In particular. Theorems 7.2, 7.3, 7.4. 7.6, and 1.1 (as well as their proofs) are all valid (word for word) whenfaod IX are complex-valued functions. (In Theorems 7.2 and 7.3, the constants Cl and C1 may now be complex numbers.) In addition, we have the following theorem which, in effect. reduces the tbeory of complex Stieltjes integrals to the real case.

S:f

J.

_ 4P,f)

<

M-

+
r

b-

a.
holds,
SOfE

Tllt!tJffm 7.50. Let f = fl defined on an interval [a.

on [a, b).

Since this holds for every r ~ I, we see that Riemann's

condition

NOfE. A property is said to hold a/most everywhere on a subset S of Rl if it holds everywhere on S except for a set of measure O. Thus, Lebesgue's theorem states

r a: r
bJ.
I
dll =

if2 and

C!l

+
j

i«2

be complex-valued

Then we have
-

11 dal

Iz

d(2)

(f 12

d~1

functions

II dal)

whenever all four integrals on the right exist.

174 The proof of Theorem 7.50 is immediate from the definition and is left to the reader. The use of this theorem permits us to extend most of the important properties of real integrals to the complex case. For example, the connection between differentiation and integration established in Theorem 7.32 remains valid for complex integrals if we simply define such notions as continuity, differentiability and bounded variation by components, as with vector-valued functions. Thus, we say that the complex-valued function tX Q!1 + ia.2 is of bounded variation on [6, b] if each component (Xl and !Xl is of bounded variation on [0, b]. Similarly. the derivative (X'(t) is defined by the equation a'(t) = ai(/) + iiX2(t) whenever the derivatives a:i(t) and ajJt) exist. (One-sided derivatives are defined in the same way.) With this understanding, Theorems 7.32 and 7.34 (the fundamental theorems of integral calculus) both remain valid when f and II are complex-valaed, The proofs follow from the real case by using Theorem 7.50 in a straightforward
and hence U(P.f) < 1+ e. Similarly, L(P'f) Hence IS(P,f) -

175

II <

if

~pn < min (J. 0').

A(x) =1I::!Ii:.x

>1

iqPll

-e 1/

for some 0'.

7.5 Let {a..} be a sequence of real numbers.

For x
=

2!:

O. define

::E a. ::E a..

.=1

r-l

where [x] is the greatest integer in x and empty sams are interpreted as zero. Let {have a continuous derivative in the interval I :s; x :s; a. Use Stieltjes integrals to derive the. following formula:

E a,.!(n) =
~~.

- [A(X)/'(X)
1

dx

A(a)/(a).

manner.

We shall return to complex-valued integrals in Chapter 16, when we study functions of a complex variable in more detail.

7.6 Use Euler's summation formula, or integration by parts in a Stieltjes intc:gra1, to derive the fonowing identities: I rt-1 b)

+ $ Jl
I

t'

[xl
x'+l

dx dx

if, '" 1.

EXERCISES Itienwm-SdeItjes integrals

7.1 Prove that

1-1

• :E -k1 = log n -

i"x-

--2~

{x]

1.

I! dlr(x)

Il(b)

Il(a).directly from

Definition 7.1.
is monotonic on [a, b],

7.Z If fE R(a:) on [a, b] and if J!I tk ... 0 for every {which prove that III must be constant on [a. b 1-

7.7 Assume/, is continuous on (1, tion by parts to prove that

2nJ

and use Euler's summation formula or integra.

7.3 The following definition of a R.iemann-StieltjOi integral is often used in the literature: We say lis integrable with respect t.o Ill: irthere exises a real number A having the property that for every Il > 0, there exists a {} > 0 such that for every partition P of [a. b] with nonn IIPII < 0 and for every choice of t1 in [Xl-I> XlI. we have IS(P,/. a) - AI < fl.

6
211

(-lll(k)

J.

t:

/'(x)([x}

- 2[x/2]) dx,

a) Show that jf j! , d", exists according to this definition, then it also existsaccordiqg to Definition 7.1 and the two integrals are equal. 0 for a ~ x < t:'/{x) ~ a(x) ." I for c < x ~ b.J(c) = O. a:(t) == 1. Show that tk exists according to Definition 7.1 but does not exist by this second definition.
"" a:(x)
ee

7.8 Let fll(X) = x - [x] - ! if x ~ integer. and let 1I'1(X) = 0 if x integer. Also, let "1(X) Jij '1(1) dt. If I" is continuous on [1. nJ prove that Euler's summation formula implies that

b) Let/(x)

$:{

i-Ill

:E '(Ie)

"i
=

'(x)

dx -

i"

fP2(x)/"(x)

dx

+ 1(1) + fen) 2

7.9 Take I(x) = log x in.Exercise 7.8 and prove that log n! 7.10 If x
;":!

7.4 IffE R accordiog to Definition 7.1. prove that S:I(x) dx also exists according to the definition of Exercise 7.3. [Contrast with Exercise 7.3(b).] Hint. Let I ""I!/(x) dx, M = sup {1/(x)1 : x E [a, b J}. Given e > 0, choose p. SO tbat U(P., f) < I + 8/2 (notation of Section 7.11). Let N be the number of subdivision points in p. and let ii = 8/(2MN). If IPII < 0, write U(P,f) =

(n

1) log

If -

If

+ (-

J.

'2!t)
t

lit.

I, let 1l(x) denote the number of primes :;;x. that is,

It{x) =

L MA(f)

Ax, = Sl

+ S2'

.E I,
JI",X

where S 1 is the sum of terms arising from those subintervals of P containing no points of p. and Sl is tbe sum of the remaining terms. ;Then SI -s U(P.,f) <J

when: the sum is extended over all primes p

:s
X

x. The prime number theorem states that

lim 1I(x) log x = 1.

,;;c-.(6)

z/2

and

81. ~

NMllPl1

< NMii = £/2,

116
This is usuaJly proved by studying a related function 8 given by

171 7.14 Assume IE R(gr) on [a, b 1, where « is of bounded variation On [a, b]. Let Vex) denote the total variation of /l on [II. x] for each x in (a, b J. and let Veal = O. Show that

8(x) ~

where again the sum is extended over all primes p s x. Both functions If and 9 are step functions with jumps at the primes. This exercise shows bow the Riemann-Stieltjes integra] can be used to relate these two functions.

.""

:E logp,

Ifldals

fill

dV

s MV(b),

where M is an upper bound for

a) If x :<!: 2, prove that n(x) and 9(x) can be expressed as the following RiemannSticl\jcs

becomes

Ilion [0, b].

In particular, when «(x) = x, the inequality

integrals:

8(x) =
NOTE.

Jm

(;It

log' dx(t),

B(x) =

i"
I

1
I

3/2108

d3(t).

If ul
I(x)

:50

M(b - a),

1be lower limit can be replaced by any number in the open interval (1. 2),
b) If x

~ 2, use integfation by

parts to show that

7.15 Let {.z,,} be a sequence of functions of bounded variation on [a, b]. Suppose ~ exists a function a defined on [a. 6] such that the total variation of gr - «" on [a, b] tends to 0 as n -+ 00. Assume also that Cl(a) = a,.(a) = 0 for each n = 1.2, ... If lis continuous on [a, b], prove that

8(x) = I(x) log x -

12

f'" 7Z(t) dI,

:~~ f
/(x)

dx..{x)

[/(X)

i'Al(x).

x(x)

8(x)
log X

+ [lI:

J2

9(t) dt. t logz t

7.16 If Ie R(<<),/l t R(<<), 9 e R(<<), and 92 e R(<<) on [a, b], prove that

These equations can be used to prove that the prime number theorem is equivalent to the relation lim" .. ", 8(x)/x = I. 7.11 If tV on [a. b J, prove that we have

2..

[il> I/(X} .. ICY)

g(y)

g(X)1

d:z(y)l

a)
b)

f '"

rr r. sri r
=

= (a < e <

(1" th(X»)(f
/(X)2

dz(x)

g(x/

th:(X») -

(1"

I(x)g(x) dx(X)t

I· +

I..

b).

When a)'" on [a, b], deduce the Cauchy-Sebwarz

inequality

if +

g) •

c) [

if +

g) tb.:<!:

1:
i

d«

9 ..

I th. + [

9 th:.

(Compare with Exercise 1.23.)

7.1l Give an example of a bounded function I and an increasing funaion c defined on [a. b] such that IIIEO R{«) but for which I! I. does not exist 7.13 Let 0; be a continuous function of bounded variation on [a, b). Assume g e R(o;)
on [a, b) and define P(x) =

i[ [f u:» -

7.17 Assume that

I E R(gr), 9

E R(fl). and

f· 9 E R(gr) on [II. b].

da(x)

Show that

I(x»)(g(y) - g(x»

th:(y)]

I!g(t)

th(t) if x e [a. b]. Show that:

a) If I?' on [II, b). there exists a point Xo in [a, b] such that

l>

I dfJ

= /(a)

..

r
..

(<«b) - a(a» [ f(x)g(x) dl(x) -

9•

+ I(b)

i~
;;'0-

U:

/(x)

til(X») (1" g(x) til(X») .

f(x)g(x) da(x)

9 ...

If gr)1 on {a, b], deduce the inequality

b) If. in addition, lis continuous on [a, b]. we also have f(x)g(x) dx(x) = I(a)
9•

+ 1(6)

J~

(J:

/(x)

d«(X»)(f

g(x)

iJx(X») S (Cl(b) - «(0»

g ••.

when both land 9 are increasing (or both an: decreasing) on [a, b]. Show that the reverse inequality holds if I increases and 9 decreases on [a, b],

178
~bmgraIs 7.18 Assume fER
OD

I"
[a, b]. Use Exercise 7.4 to prove that the Umit lim b .... '" "

c) Usc (b) to prove tho folJowiq tIJcm:an of L. F..,: The number of ehanp in sign of 1m [0. a] is not Jess than the mmtber of c:hanaes in sisn itt the ordered set
1(0).

ai:J(a
:t~1

k b - "')

f(t)

dt'l
•

t/(l)

41.

• •• •

1"/(1) lit.

exists and has the value

I!/(x)

dx. Deduce that

7.24 Let/be a positiveQOlltinuousfunctioD of Ion [a, b}. Show that lim

... co

in [a, b]. Let Mdenot:e the maximum value

(f/(x'!

dx)l/ll .. M.
if x is rational, if x is irrational.

7.1' Define

7..25 A function! of two real wri.a.bIes ia defined. for each point (x. y) in the unit square :s; x s: 1. 0 s )l :s; I as foRows:

a) Show that g'(x) + /,(x) "" 0 for all x and deduce that g(x) b) Use (a) to prove that

+ f(x)

= 71/4.

a) Compute

lA/(x, y) the and lA/(x, y) dx In terms

b) Show that I(x, y) dy CJIists of x and t for 0 :s; x s: 1. 0

JJ

7.20 Assume 9
integral

I:

R on [a. b] and define I(x) = J! g(l) dt if x Ig(t)1 tit gives the total variation of I on [4, x].
E

[a. b]. Prove that the

e) Let F(x)

fA !(X, Y) dy.

.s

rOl' each

of y.

filled x and compute

t s 1.

ro I(x, y) dy in tcnns
2-·, then/(x) "" 2-·,

Show that

fA F(x)

fix exists and find its value.

7.2.1 Let f = (11)' . .fJ be a w:ctor-valued function with a continuous derivative f' on [a, h). Prove that the curve described by f has length A,(a" b) =

7311 Let/be defined on [0, I] as foIlows:/(O) = 0; iU-·-1 for 11 = O. 1,2. ... a) Give two
I'C8!IOM why

<X

S;

b) Let F(x) = 111"(1)11 tit. where: A(x) ...

ro I(t) dt.
.... 21

IA /(x) dx exists.
F(x)

Show that for 0 -c x

s:

1 we have

xA(x) - tA(x?

73:1. If/(Hll is continuous on [a. z], define l.(x) = n! • (x - 1)"/0.+1)(1) dt. I a) Show that 1~_I(X)- ll(x) = .fkJ(Q~~ - a>", k = 1, 2, ...•
II.

r[-qJrI

and when: (y] is the: greatest integer in y.

7:r1 Assume I has a derivative whid1 is monotonic III > 0 for all x in [a, 0]. Prow that

decr'easinlI:

and satisfies 1'{;Ie) ~

If
Hint. Multiply and divide
7.28 Given a dcacasing &eqUCllCCof Define &function/on [0. I] in terms I(x) = 0; if x is the rational mJ,. oscillation w_,.(x) at each x in [0. I]

cos/(x)

thel s ~.

the intep"aDd by {'(x) and usc Theorem 7.37(ii).

b) Usc (a) to express the remainder in Taylor's formula (Theorem 5.19) as an integral. 7.23 Letlbe continuous on [0, aJ. If x
E

[0, a]. define/o(x)

n

= I(x) and let

em

I. +1(x)

... -1

n!

IX
0

mil numbers {G{n}} such that G(II) ~ 0 as ,. .... eo, of (G(II)} as foIlows:/(O) = 1; if x is iIrational, tbc:n lowest teIms). thmf(m/n) = G(n). Compute the and show that Ie R on [0. I].

(x - t)"f(t) dt,

0, I, 2, ...

a) Show that the 11thderivative of I. exists and equals! b) Prove the following theorem of M. Fekete: The number of changes in sign of 1 in [0, 0] is not Jess than the number of changes in sign in the ordered set of

7.l9 Let/be defined as in EKrcise 7.28 with G(n) = lIn. Let g(x) == 1 if 0 -c x S; I, g(O) O. Show that the composite function" dc6ned by h(x) = g[f(x)] is not Riemannintegrable on [0, 1], although both I € R and g e R on [0. I],
7.30 Use Lebesgue's theorem to prove Theorem 7.41). 7.31 Use Lebesgue's theorem to pJ'O'iC that jf / e R and g m > 0 for all x in [D, b], then the function h dcOned by
/(x) = I(x~)
E

numbers

R on [a, b] and if I(x) 2:.

l(a),/l(o), ... ,/,,(0). Hinl.

Use mathematical induction.

is Riemann-integrable

on [a, h].

..,
7.31 Let 1 = [0. I} and let AI = 1 1) be that subset of 1 obtained by removing those points whlch He in the open middle thinl of 1; that is. Al = [0. t) v 1J. Let A;I. be that subset of Al obtained by removing the open middle third of [0, tJ and of K. 1). Continue this process and define .43 .4... . .. The set C = n:'l A" is called the • Cantor set. Prove that:

lU

<t.

e.

Existenee tbeoRms for Integral_ dift'creidiaI .,....tionB The following exercises illustrate bow the fixed-point theorem for centractions (Theorem 4.48) is used to prow emte:nce theorems (or solutions of amain intcgIal and differential equations. We denote by C [a, b] the metric space of all real continuous fuoctioos on [a, b] with the rnetJ:i<:

a) C is a (:(Jrnpac:t set having measure zero. b) x e C if, and only if, JC = l:..";1 0.,3-", when each a,. is either 0 or 2. c) C is uncountable. d) Let {(x) = 1 if x
E C,/(x)

= 0 if x ; C. Prove that / e R on [0, 1).

and recall that C [a. b] is a complete metric space (Exercise 4.67). Let /(x)

7.33 This exercise outliDcs a proof (due to lvan Niveo) that x"(l - x'fln!. Prove that: a) 0 < {(x) < l/nl Now assume that if 0 < x < 1. . and{oo(1) is an integer. b) Each ktb dcrivative{Ct)(O)

r is irrational.

7.36 Given a funetien g in C [a, b J. and a functioo K continuous 00 the rectangle Q = [a, b] x [a, b). consider the function. T defined on C[a, b J by the equation T(V)(x) = o(x)
wbeIe

+ 1[

X(x, t)(l(t) dt.

r "" where alb,

a and b are positive integers, and let I/' ~ (-If/(ul(x) ,r.-u.

1is a given constant,

a) Prove that T maps C [a, bJ into itself.

F(x) =

..

Prove that: c) F(!) and F(I) are integers. d) ,rti'/(x) sin roc

b) If IK(x,y)1 s Mon Q. where M::> 0, and if P.I < M-l(b -:- a).-I. pro~ tbat T is a contraction of C [a, bJ and bencc has a fixed point , which IS a solution of the integral equation (l(x) = g(x) + 1 K(x, t)f'{t) dJ.

I:

= !.. {F(x)
dx

e) P(l) +

F(O) = 7fti'

f n»

aiD roc - xF{x) cos roc}. sin xx dx.

7.37 Assume 1 is continuous on a rectangle Q = [a - h. a where It > O. k > O.

It) x [b - k, b

k],

a) Let " be a. function, continuous on [a - It, a + It], such that (x, (l(x») e Q for all x in [a - h. a + II]. If 0 -c c ~ II., prove that Ip satisfies the differential equation y' y) on (a e, a + c) and the initial condition fl'{a) = b if, and only if. II satisfies the integral equation

n»;

f) Use (a) in (e)

to deduce that 0 -c F(l) + F(O) < 1 if n is sufficiently large. This contradicts (c) and shows that (and hence n) is irrational.

,(x) = b

f(t. ,,(t») dJ

on

(12 - c. a

c).

7.34 Given a real-valucd function a, continuous 00 the interval [a, b] and having a finite bounded derivative II: 00 (a, b). Let/be defined and bounded on [a, b] and asswne that both integrals

!(x) dx(x)

and

I(x) rr'(x) dx
/I.'

b) A&sume that. I/(x, )1)1s Moo Q. where M> O. and let c "" min {Ir. kiM}. Let S denote the metric subspace of C [a - e, a + e J consisting of all II such that If'{x} - bl ;s Me on [0 C, Q + c). Prove that S is a closed subspace of C [a - c, a + e] and hence that S is itself a complete metric space.
c) Prove that the function Tdefined on S by the equation T(9')(x) b

exist. Prove that these intcsrals are equal. (It is not assumed that

is eontiauous.)

7.35 Prove the following theorem" which implies that a function with a positive integral must itself be positive on someint.erval. Assume that/"" Ron [a, b] and that 0 :5; /(x) s M on la, b], where M ::> O. Let 1 = I:{(x) dx, let h = YI(M + b - a), and assume that 1 ::> O. Then the set T = {x :/(x) ~ h} contains a finite number of intervals, the sum of whose lengths is at least It. Hint. Let P be a _partition of [a, b) such that every Riemann sum S(P.f) = D_l1(tJ &xl satisfies S(P./) ::> 1/2.. Split S(P,/) into two parts, S(P,!) = k + :Et..a. where A = {k: [Xl_I' Xi]
LuAx,,::>
!; T}.

[(t.

,,(t»

dJ

maps S into itself. d) Now assume that 1satisfies a Lipschitz condition of the form

I/(x, y) - I(x,

%)1 ~

AIY -

zl

and

{k: k ¢ A}.

If k e A. use the inequality {{tJ s M; if k e B. choose It so that/(tJ h.

<

n.

Deduce that

for overy pair of points (x. y) and (x. z) in Q. where A > o. Prove that T is a contnldion of S if" < IIA. Deduce that for II < 1/.4 the differetltial equation y' = /(x.)') has exactly one solution )' = tp(x) on (a - c. Q + c) such that f(a) = b.