COMPLEX ANALYSISInternational Series in Pure and Applied Mathematics 6. Springer and E. H. Spanier, Consulting Editors Aljora: Complex Analysis Bender and Orszag: Advanced Mathematical Methods for Scientists and Engineers Buck: Advanced Calculus Busacker ond Seaty: Finite Graphs aud Networks Cheney: Introduction to Approximation Theory Chester: Techniques in Partial Differential Equations Coddington and Levinson: Theory of Ordinary Differential Equations Conte and de Boor: Elementary Numerical Analysis: An Algorithmie Approach Dennemeyer: Introduction to Partial Differential Equations and Boundary Value Problems Detiman: Mathematical Methods in Physics and Engineering Golomb ond Shanks: Elements of Ordinary Differential Equations Hamming: Numerical Methods for Scientists and Engineers Hildebrand: Introduction to Numorieal Analysis Householder: ~The Numerical Treatment of a Single Nonlinear Equation Kelman, Falb, and. Arbib: Topics in Mathematical Systems Theory Ease: Vector and Tensor Analysis McCarty: Topology: An Introduction with Applications to Topological Groups Mork; Tuatroduction to Set Theory Moore: Tlements of Linear Algebra and Matrix Theory Moursund and Duria: Elementary Theory and Application of Numerical Analysis Pearl: Matrix Theory and Finite Mathematics Pipes and Hervil: Applied Mathematica for Engineers and Physicists Ralston and Rabinowitz: A First, Course in Numerical Analysis Ritger and Rose: Differential Equations with Applications Rudin: Principles of Mathematical Analysis Shapiro: Introduction to Abstract Algebra Simmons: Differential Equations with Applications and Historical Notes Simmons: Introduction to Topology and Modern Analysis Struble: Nonlinear Differential Equationsot COMPLEX ANALYSIS An Introduction to the Theory of Analytic Functions of One Complex Variable Third Edition Wass, bead OTCALRU Lars V. Ahlfors Professor of Mathematics, Emeritus Harvard University McGraw-Hill, lic. New York St. Louis San Francisco Auckland Bogoté Caracas Lisbon London Madrid Mexico City MitanAmeCisat) Td~ 0} COMPLEX ANALYSIS Copyright © 1978, 1966 by McGraw-Hill, Inc. All righta reserved. Copyright 1983 by McGraw-Hill, Inc. All rights reserved. Printed in the United Btates of America. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by sny means, electronic, mechanical, photocopying, recording, or otherwise, without the prior written psttiation of the publisher. 161718192021 BRBBRB 969876543210 ‘This book was set in Modern 8A by Monotype Composition Company, Inc. ‘The editors were Carol Napier and Stephea Wagley; ‘the production supervisor was Joe Campanella. Libeary of Congress Catzloging in Publication Data Ablfora, Lars Valerian, date Comples analysia. (international series in pure and applied mathematics)iedteah ia trier conS A.M heContents PTER 1 COMPLEX NUMBERS The Algebra of Complex Numbers 1.1 Arithmetic Operations L2 Square Roots 43 Justification 1.4 Conjugation, Absolute Value 5 Inequalities The Geometric Representation of Complex Numbers Geometric Addition and Multiplication The Binomial Equation Analytic Geometry ‘The Spherical Repreasatation ER 2 COMPLEX FUNCTIONS 828 SRexy sCONTENTS 2.8 Uniform Convergence 2.4 Power Series 2.5 Abel's Limit Theorem The Exponential and Trigonometric Functions 3.1 The Exponential 3.2 The Trigonometrie Functions 3.3 The Periodicity: 3.4 The Logarithm CHAPTER 3 ANALYTIC FUNCTIONS AS MAPPINGS i Elementary Point Set Topology 1.1 Bets and Elements 1.2 Metric Spaces 18 Connectedness 14 Compactness 1.5 Continuous Funotions 1.6 ‘Topological Spaces Conformality 2.1 Arcs and Closed Curves 22 Analytic Functions in Regions 23 Conformal Mapping 2.4 Length and Area Lineor Transformations 3.1 The Linear Group 3.2 The Crow Ratio 3.3 Symmetry 3.4 Oriented Circles 3.5 Families of Circles Elementary Conformal Mappings 4.1. The Use of Level Curves 4.2 A Survey of Elementary Mappings 43. Blementary Riemann Surfacea CHAPTER 4 COMPLEX INTEGRATION 1 Fundamental Theorems 11 Tine Integrals 1.2 Rectifieble Ares 5 eh ESR SEES BS GSSESS 8 S ssQ s 3 a Ree £48aa aCONTENTS x 2 Cauchy's Integral Formula 14 2.1 The Index of a Point with Respect to 2 Closed Curve 114 22° The Integral Formula 118 23 Higher Derivatives 120 3 Local Properties of Analytical Functions 4 3.1 Bemovable Singularities. Taylor's Theorem 124 3.2 Zeros and Poles 126 3.2 The Local Mapping 130 34 The Maximum Principle 133 4 The General Form of Cauchy's Theorem 137 4 Chains and Cycles 137 42 Simple Connectivity 138 43 Homology 141 44 The General Statement of Cauchy’s Theorem 141 4.5 Proof of Cauchy's Theorem 142 4.6 Locally Exact Differentials Wat 4.7 Multiply Connected Regions M46 5 The Calculus of Residues 48 5.1 The Residue Theorem 148 5.2 The Argument Principle 152 63 Evaluation of Definite Integrals 4 Harmonic Functions 182 6.1 Definition and Basic Properties 162 62 The Mean-value Property 165 6.3 Poisson’s Formula 166 64 Schwarz’s Theorem. 168 6.5 The Reflection Principle Wz : GHAPTER 5 SERIES AND PRODUCT DEVELOPMENTS 175 Dt ‘Power Series Expansions 178 1.1 Weierstrass’s Theorem 175 ; 12) The Taylor Series 179 13 The Laurent Series 184 187 187 Ww. 193 198= CONTENTS 3 Entire Functions 3.1 Jensen’s Formula 8.2 Hadamard’s Theorem 4 The Riemann Zete Function 41 The Product Development 4.2 Extension of {(8) to the Whole Plane 43 ‘The Functional Equation 44 ‘The Zeros of the Zeta Function 5 Normal Families 5.1 Equicontinuity 52 Normality and Compactness 5.3 Arsela’s Theorem 54 Families of Analytic Functions 5.5 The Classical Definition CHAPTER 6 CONFORMAL MAPPING. DIRICHLET’S PROBLEM 1 The Riemann Mapping Theorem 1.1 Statement and Proof 1.2 Boundary Behavior 1.3 Use of the Reflection Principle 14 Analytic Arcs 2 Conformal Mapping of Polygons 2.1 The Behavior at an Angle 2.2 The Schwarz-Christoffel Formula 2.3 Mapping on a Rectangle 24 The Triangle Functions of Schwars 3 A Closer Look at Harmonic Functions 3.1 Functions with the Mean-value Property 3.2 Harnack’s Principle 4 The Dirichlet Problem 4.1 Subharmonic Functions 4.2 Solution of Dirichlet’s Problem 5 Canonical Mappings of Multiply Connected Regions 5.1 Harmonic Mesaures ga 8 B12 213 214 216 218 N S I 8 3S EERE BREE E 8 SERS BEE ESE ERCONTENTS CHAPTER 7 ELLIPTIC FUNCTIONS 1 Simply Periodic Functions 1.1 Representation by Exponentials 1.2 The Fourier Development 1.8 Functions of Finite Order 2 Doubly Periodic Functions 2.1 The Period Module 2.2 Unimodular Transformations 2.3 The Canonical Basis 24 General Properties of Elliptic Functions 3 The Weierstrass Theory 3.1 The Weierstrass p-function 3.2 The Funetions {(2) and o(2) 3.3 The Differential Equation 3.4 The Modular Function Mz) 3.5 The Conformal Mapping by Mr) CHAPTER 8 GLOBAL ANALYTIC FUNCTIONS 1 Analytic Continuation 1.1 The Weierstrass Theory 12 Germs and Sheaves 18 Sections and Riemann Surfaces 1.4 Analytic Continuations along Ares 1.6 Homotopic Curves 1.6 The Monodromy Theorem 1.7 Branch Points 2 Algebraic Functions 2.1 The Resultant of Two Polynomials 2.2 Definition and Properties of Algebraic Functions 2.3 Behavior at the Critical Points 3. Picard’s Theorem = 3.1 Lacunary Values uf Linear Differential Equations 41 Ordinary Points 42 Regular Singular Points ya:,fe0 Solutions at Infinity 44 The Hypergeometric Differential Equation 45 Riemann’s Point of View 264 264 265 265 270 272 272 273 275 279Preface Complez Analysis has successfully maintained its place as the standard elementary text on functions of one complex variable. There is, never- theless, need for a new edition, partly because of changes in current mathe- matical terminology, partly because of differences in student, preparedness * and aims. There are no radical innovations in the new edition. The author still believes strongly in a geometric approach to the basics, and for this reason ‘the introductory chapters are virtually unchanged. In a few places, ‘throughout the book, it was desirable to clarify certain points that ex- “perience has shown to have been a source of possible misunderstanding or difficulties. Misprints and minor errors that have come to my attention ‘have been corrected. Otherwise, the main difierences between the second Vand third editions can be summarized as follows: E 1. Notations and terminology have been modernized, but it did not Lecom necessary to change the style in any significant way. ' &. In Chapter 2 a brief section on the change of length and area under formal mapping has been added. To same degree this infringes on the ‘ize self-contained exposition, for it forces the reader to fall back on lus for the definition and manipulation of double integrals. The vantage is minor. 3. In Chapter 4 there is a new and simpler proof of the general form of A thy’s theorem. It is due to A. F. Beardon, who has kindly permitted @ to reproduce it. It complements but does not replace the old proof, ich has been retained and improved. 4. A short section on the Riemann zeta function has been included. ™ah PREFACE ‘This always fascinates students, and the proof of the functional equation illustrates the use of residues in a less trivial situation than the mere computation of definite integrals. 5. Large parts of Chapter 8 have been completely rewritten. The main purpose was to introduce the reader to the terminology of germs and sheaves while emphasizing all the classical concepts. It goes without saying that nothing beyond the basic notions of sheaf theory would have been compatible with the elementary nature of the book. 6. The author has successfully resisted the temptation to include Riemann surfaces as one-dimensional complex manifolds. The book would lose much of its usefulness if it went beyond its purpose of being no more than an introduction to the basic methods and results of complex function theory in the plane. It is my pleasant duty to thank the many who have helped me by pointing out misprints, weakmesses, and errors in the second edition. IT am particularly grateful to my colleague Lynn Loomis, who kindly let. me share student reaction to a recent course based on my book. Lars V. AhiforsCOMPLEX ANALYSIS1 LUNIPLEA NUNMNBENS L THE ALGEBRA OF COMPLEX NUMBERS It is fundamental that real and complex numbers obey the same basic laws of arithmetic. We begin our study of complex func- tion theory by stressing and implementing this analogy. 1.1. Avithmetie Operations. From elementary algebra the reader is acquainted with the émaginary unit i with the property #@= —1. Ef the imaginary unit is combined with two real num- bers a, 8 by the processes of addition and multiplication, we obtain a compler number «+28. a and § are the real and imaginary pari of the complex number. If « = 0, the number is said to be purely imaginary; if 8 = 0, it is of course real. Zero is the only number which is at once real and purely imaginary. ‘Two complex numbers are equal if and only if they have the same real part and the same imaginary part. Addition and multiplication do not lead out from the system. of complex numbers. Assuming that the ordinary rules of srithmetic apply to complex numbers we find indeed @ fa + #8) + (y+ 48) = @ +7) +i1G+ 9 and @ (@ + iB) + 48) = (ay — 88) + iad + By). In the second identity we have made use of the relation ## = —1. It is less obvious that division ia also possible. We wish to2 COMPLEX ANALYSIS show that (a + i8)/(y + 78) is a complex number, provided that y + i8 #0. If the quotient is denoted by + + iy, we must have atis= (y + ie + iy. By (2) this condition can be written. a+i8 = (yr — ay) + ir + vy), and we obtain the two equations a= ym by 6 = 8+ yy. This syatom of simultancous linear equations has the unique solution 2 or BS Whe By — 2 y= oe for we know that y? + 8% is mot zero. We have thus the result @) a+ iB _ oy +B By — 2d yt +R +e Once the existence of the quotient has been proved, its value can be found in a simpler way. If numerator and denominator are multiplied with y — 48, we find at once a+ t8 _ (a + 18)(y — i8) _ (oy + 88) + (By — 0d), ye G+ DG ~ 8) te fi As a special case the reciprocal of 2 complex number + O is given by 1 a — if ati ote We note that #* has only four possible values: 1, i, —1, —i. They correspond to values of 2 which divided by 4 leave the remainders 0, 1, 2,3. EXERCISES 1 Find the values of +29), 5 Ata pa-—o. =TERCOMPLEX NUMBERS 3 2 fz = x + ty (x and y real), find the real and imaginary parts of aos: 3. Show that -1tiv3y_ £1 tivay_ (244) 1 and Gus 1 for all combinations of signs. 4.2. Square Roots. We shail now show that the square root of a complex number can be found explicitly. If the given number is a + 16, we are looking for a number x + ty such that @ + ty)? = a + if. This is equivalent to the system of equations @) gao—yoa Qary = B. From these equations we obtain +P) = vt Aetyt = at lence we must have ety VE TR ere the square root is positive or sero. Together with the first equa- on (4) we find B= Ha+ Vat + 8) ya ia + Val + PF). ve that these quantities are positive or zero regardless of the sign ‘The equations (5) yield, in general, two opposite values for z and two iy... But these values cannot be combined arbitrarily, for the second tion (4) is not a consequence of (5). We must therefore be careful eg pnd y so that their product bas the sign of 6. This leads to the Vat p= + (JetVEEE 8 ast ere) dthit § 4 0.' Forf = Othe valiiesare + a). Tt is most easily defined in terms of the set B+ of positive real numbers: a < ¢ if and only if 8 — a ¢ Rt. The set Rt is characterized by the following properties: (1) Ois not a posi- tive number; (2) if a ~ 0 either a or —a is positive; (3) the sum and the product of two positive numbers are positive. From these conditions one derives all the usual rules for manipulation of inequalitiea, In particular one finds that every square a” is either positive or zero; therefore 1 = 1* is # positive number. By virtue of the order relation the suma 1,1-+1,1+1+41,... are all different. Hence R contains the natural numbers, and since it is 6 field it must contain the subfield formed by all rational numbers. Finally, R satisfies the following completeness condition: every incress- ing and bounded sequence of real numbers has a limit. Let a1 < a: < my < ++ Sa, < ++, and assume the existence of a real number B such that a, < Bforalln. Then the completeness condition requires the existence of a number A = lim... ¢. with the following property: given any « > 0 there exists a natural number 7 such that A — e mo. Our discussion of the real-number system is incomplete inasmuch as we have not proved the existence and uniqueness (up to isomorphisms) of a system R with the postulated properties.t The student who is not thoroughly familiar with one of the constructive processes by which real numbers can be introduced should not fail to fill this gap by consulting any textbook in which a full axiomatic treatment of real numbers is given. The equation zt + 1 = 0 has no solution in R, for a® + 1 is always positive. Suppose now that a field F can be found which contains R as a subfield, and in which the equation z7 -+ 1 = Ocan be solved. Denote a solution by i Then st+1= @+(@—1), and the equation a* +1 = 0 has exactiy two roois in F,é and ~7. Let C be the subset of F consisting of all elements which can be expressed in the form « + ¢8 with real a and & This representation is unique, for a + 18 = a" + 46" plies a — a” = —i(6 — §'); hence (a — a’)? = (6 — 8’), and this is ible only if a = a’, 8 = fp’, The subset C is a subfield of F. In fact, except for trivial verifica- which the reader is asked to carry out, thia is exactly what was shown Sec.1.1. What is more, the structure of C is independent of F. For if F is another field containing R and a root ¢’ of the equation 2* + 1 = 0, 1 An tsomarphiam between two fields i¢ s one-to-one eorrespondence which pre- suing and products. The woid is used quite generally to indicate a corre- whith is ofie'ts one and preserves aI relations that are considered important6 COMPLEX ANALYSIS the corresponding subset C’ is formed by all elements « + 78. There is @ one-to-one correspondence between C and C’ which associates a + 78 and « + 76, and this correspondence is evidently a field isomorphism, Tt is thus demonstrated that C and C’ are isomorphic. We now define the field of complex numbers to be the subfield C of an arbitrarily given F. We have just seen that the choice of F makes no difference, but we have not yet showa that there exista a field F with the required properties. In order to give our definition a meaning it remains to exhibit a field F which contains R (or a subfield isomorphic with R) and in which the equation z? + 1 = 0 has a root. There are many ways in which such a field can be constructed. The simplest and most direct method is the following: Consider all expressions of the form & -+ 48 where a, § are real numbers while the signs + and # are pure symbols (+ does not indicate addition, and ¢ is not an element of a field). These expressions are elements of a field F in which addition and multiplication are defined by (1) and (2) (cbserve the two different mean- ings of the sign +), The elements of the particular form a + iO are seen to constitute a subfield isomorphic to R, and the element 0 + #1 satisfies the equation z*-+ 1 =0; we obtain in fact (+ 11)? = —(1 + 10). The field F has thus the required properties; moreover, it is identical with ‘the corresponding subfield C, for we can write a+ ip = (2 + 0) + BO + il). The existence of the complex-number field is now proved, and we can go back to the simpler notation a + 18 where the + indicates addition in C and ¢ is a root of the equation 2* + 1 = 0. EXERCISES (For students with a background in algebra) 4. Show that the system of all matrices of the special form. a 6 —6 af’ combined by matrix addition and matrix multiplication, is isomorphic to the field of complex numbers. 2 Show that the complex-number system can be thought of as the figid of all polynomials with real coefficients modulo the irreducible polynemial z* + 1. 1.4. Conjugation, Absolute Value. A complex number can be denoted either by asingle letter a, representing an element of the field C, or in the form @ + 78 with real « and 8. Other standard notations are z=a2+iy,t = § +4, wv = u + fv, and when used in this connection itCOMPLEX NUMBERS T is tacitly understood that z, y, £4, uv are real numbers, The real and imaginary part of a complex number ¢ will also be denoted by Re a, Ima. In deriving the rules for complex addition and multiplication we used. only the fact that i# = —1. Since 7 has the same property, all rules must remain valid if 7 is everywhere replaced by -#. Direct verification shows that this is indeed so. The transformation which replaces a + 48 by @ — 78 is called complex conjugation, and « — 78 is the conjugate of a +46. The conjugate of a is denoted by 4. A number is real if and only if it is equal to its conjugate. The conjugation is an ivolutory transformation: this means that 4 = a. The formulas até a-@ Rea= z + Ima= og express the real and imaginary part in terms of the complex number and its conjugate. By systematic use of the notations 2 and 4 it is hence possible to dispense with the use of separate letters for the real and imaginary part, It is more convenient, though, to make free use of both notations, ‘The fundamental property of conjugation ia the one already referred to, namely, that a+b =a4+6 ob = 4-5. ‘The corresponding property for quetients is a consoquence: if ax = b, then G@ = 6, and hence (b/a) = 6/é. More generally, let R(a,b,c, . - 3 stand for any rational operation applied to the complex numbers 4, }, ¢, «+. Then EGE, = RABE, ...). As an application, consider the equation eon” + cet b= + ogee + oy = 0. Hi ¢ is 8 root of this equation, then £ is a root of the equation : fa beet tet HO particular, if the coefficients are real, ¢ and F are roots of the same equa- and we have the familiar theorem that the nonreal roots of an equa- ion with real coefficients occur in pairs of conjugate roots. F The product ad = at + fis always positive or zero. Its nonnega- tive square root imealled the modulus or absolute value of the complex num- 4; it indenoted by ja: - The terminalogy and notation are justified bya COMPLEX ANALYSIS the fact that the modulus of a real number coincides with its numerical value taken with the positive sign. We repeat the definition 0 = |al*, where |a| = 0, and observe that [é] = |aj. For the absolute value of a product we obtain |abl? = ab- ab = abab = aabb = fal*(b\2, and hence abl = [a] - (| since both are = 0. In words: The absolute value of a product is equal to the product of the absolute values of the factors. It is clear that this property extends to arbitrary finite products: large + + + aol = fol [art - + > lanl. ‘The quotient a/b, b 7 0, satisfies (a/b) = a, and hence we have alvo [ol - [@/0| = lal; or lel The formula for the absolute value of asum is not as simple. We find la +B] = @ + d)@ +B = 0a + (ob + ba) + 0b @ le-+ 6 = [ol + B+ 2 Re, The corresponding formula for the difference is @’) lo — BI? = |al* + [b|? — 2 Re ab, and by addition we obtain the identity (8) ja + Bit + [a — Blt = 2(ol* + (b)*). EXERCISES L Verify by calculation that the values of 2. 241 for z = 2+ dy and z = x — iy are conjugate. 2 Find the absolute values of B+ 49-1429, “MBADA+ AUTH md STMT SCOMPLEX NUMBERS ® 3 Prove that antl, 1-@ if either |a| = Lor |b] = 1. Whatexception must be madoif |a| = |b| = 1? 4. Find the conditions under which the equation az + bz +c = 0 in one complex unknown has exactly one solution, and compute that solution. &. Prove Lagrange’s identity in the complex form [Samp = $ te § ; Dt — Slabs — ab st 0 indicate that iis real and positive), If b > 0 this condition can be written in the form |b|*(a/b) & 0, and-it is hence eqilivalent to a/b Zz 0. In the general10 COMPLEX ANALYSIS case we proceed as follows: Suppose that equality holds in (11); then loal + los] + -* + lal = Mon + as) + as + - + + asl S lor + a2] + [asl + +--+ [eal S loa] + Jes] + + > + [ost Hence |a: + a2| = |a;| + fas], and if ¢ # 0 we conclude that a:/a: 2 0. But the numbering of the terms is arbitrary; thus the ratio of any two nonzero terms must be positive. Suppose conversely that this condition is fulfilled. Assuming that a; # 0 we obtain lar + a2 + >>> +a = Jal: =lol(V+24 +++ az wae da 144 --- +8 =hil(as Sia. +) = jail + fas] + - = > + lanl. To sum up: the sign of equality holds in (11) if and only if the ratio of any two nonzero terms is positive. By (10) we have also la} = \(@ — 8) + 4] S [a — 8] + |B] or lol ~ [ol Ja — 4. For the same reason |b| — |aj = ]a — |, and these inequalities can be combined to (12) le — b| 2 |lal — [lj Of course the same estimate can be applied to |a + d|. A special ease of (10) is the inequality 413) la + #8] S lo] + [Al which expresses that the absolute value of a complex number is at most equal to the sum of the absolute values of the real and imaginary part. Many other inequalities whose proof is less immediate are also of fre- mMuent use. Foremost is Cauchy’s snequality which states that god@abs + + + +b aabal? S (lanl? + + + - + lanal?)(fbal? + + + + + Tal?) x, in shorter notation, ay [Eaefts S lot 3 tat 17 is a convenient summation index and, used ag a subscript, cannot be confused ‘with the imaginary unit. Jt seems pointless to ban its ue.EE ME ER TTY mY COMPLEX NUMBERS n To prove it, let 4 denote on arbitrary complex number. We obtain by (7) * # * * as) Y lac— ab = ¥ jolt + [at Y if? - 2ReX Yad. 1 ial = im il This expression is = 0 for all k. We can choose for if the denominator should vanish there is nothing to prove. This choice is not arbitrary, but it is dictated by the desire to make the expression (15) as small as possible. Substituting in (15) we find, after simplifications, jun 7 y [bel? which is equivalent to (14). From (15) we conclude further that the sign of equality holds in (14) if and only if the a; are proportional to the §,. Cauchy’s inequality can also be proved by means of Lagrange’s identity (Sec. 1.4, Ex. 4). zo . EXERCISES 1, Prove that Fe if lal <1 and fb <1, 2 Prove Cauchy’s inequality by induction. % ltja] <1,4 2 Oforé=1,...,nandutaet--- +h =], [show that Pana + Avae + Ande] <1. 4 Show that there are complex numbers z satisfying a [ze seca] the 4s o| = Sle]w COMPLEX ANALYSIS if and only if [a] S |c|. If this condition is fulfilled, what are the smallest and largest values of |z|? 2 THE GEOMETRIC REPRESENTATION OF COMPLEX NUMBERS With respect to a given rectangular coordinate system in a plane, the complex number a = a + #8 can be represented by the point with coordi- nates (a,8). This representation is constantly used, and we shall often spenk of the point a asa synonym of the number a. The frst coordinate axis (z-axis) takes the name of real aats, and the second coordinate axis (y-axis) is called the imaginary avis. The plane itself is referred to as the complex plane. The geometric representation derives its usefulness from the vivid Mental pictures associated with a geometric language. We take the point of view, however, that all conclusions in analysis should be derived from. the properties of real numbers, and not from the axioms of geometry. For this reason we shail use geometry only for descriptive purposes, and not for valid proof, unless the language is so thinly veiled that the analytic interpretation isself-evident, This attitude relieves us from the exigencies of rigor in connection with geometric considerations. 2.1, Geometric Addition and Multiplication. The addition of com- plex numbers can be visualized as vector addition. To this end we let a complex number be represented not only by a point, but also by a vector pointing from the origin to the point, The number, the point, and the vector wil! all be denoted by the same letter a, As usual we identify all vectors which can be obtained from each other by parallel displacements. Place a second vector b so that its initial point coincides with the end Point ofa. Then a + bis represented by the vector from the initial point of a to the end point of 6. To construct the difference b — a we draw both vectors a and } from the same initial point; then b — ¢ points from the end point of a to the end point of b. Observe that a + @ anda — b are the diagonals in a parallelogram with the sides a and b (Fig. 1-1). Anadditional advantage of the vector representation is that the length of the vector a is equal to [a|. Hence the distance between the points a and b is ja — d|. With this interpretation the triangle mequality a+ d| S |a| + [b| and the identity fa + dP + |a — B[* = 2(lal* + 6) become familiar geometric theorems. ‘The point ¢ and its conjugate é le symmetrically with respect. to the teal axis. The symmetric point of a with respect to the imaginary axis isCOMPLEX MUMBEAS = 13 Fig.1-. Vector addition. —4d. The four points a, —4, —a, d are the verticea of a rectangle which is symmetric with respect to both axes. In order to derive a geometric interpretation of the product of two complex numbers weintroduce polar coordinates. If the polar coordinates of the point (a6) are (r,¢), we know that @=rcog Beraing, Hence we can write a = « + 18 = r(cos@ + isin y). In this trigo- nometric form of a complex number r is always 2 0 and equal to the modulus |a|. The polar angle ¢ is called the argument or amplitude of the complex number, and we denote it by arg a. Consider two complex numbers a; = r:{cos ¢; + sin ¢:) and a, = 2(c08 gp: + ¢8in ge). Their product can be written in the from 02 = T1r2|(cob p) cO8 gs — Bin ¢; Bin yy) 4 é(sin y, COS yx +- COS —, Bin G2). By means of the addition theorems of the cosine and the sine this expres- sion can be simplified to (6) 103 = rirs{cos (e1 + 2) + én (eo + ¢3)}- We recognize that the product has the modulus riz and the argument gi +2. The latter result is new, and we express it through the equation an arg (a.a3) = arg a, + arg as. It in clear that this formula can be extended to arbitrary products, and we can therefore state; fi The argument of a product is equal te the sum of the arguments of the factors. This is fundamental, The rule that we have just formulated gives a deep and unexpected justification of the geometric representation of com- Plex numbers. We must be fully aware, however, that the manner in which we have arrived at-the formula (17):violates our principles. In the1“ GOMPLEX ANALYSIS first place the equation (17) is betweon angles rather than hetween num- bers, and secondly its proof rested on the use of trigonometry. Thus it remains to define the argument in analytic terms and to prove (17) by purely analytic means. For the moment we postpone this proof and shall be content to discuss the consequences of (17) from a less critical standpoint. We remark first that the argument of 0 is not defined, and hence (17) has a meaning only if a; and a are # 0. Secondly, the polar angle is determined only up to multiples of 360°. For this reason, if we want to interpret (17) numerically, we must agree that multiples of 360° shall not count. By means of (17) a simple geometric construction of the product ai@3 can be obtained. It follows indeed that the triangle with the vertices O, 1, @: ia similar to the triangle whose vertices are @, @:, aa2, The points 0, 1, a1, and ¢ being given, this similarity determines the point aia. (Fig. 1-2).In the case of division (17) is replaced by (18) arg 2 = arg ay — arg a. The geometric construction is the same, except that the similar triangles ave now 0, 1, a: and 0, a:/a;, a2. Remark: A perfectly acceptable way to define angles and arguments would be to apply the familiar methods of caleulus which permit us to expreas the length of a circular arc as a definite integral. This leads to a correct definition of the trigonometric functions, and to s computational proof of the addition theorems. ~ The reason we do not follow this path is that complex analysis, as a,a, a2 [a ~COMPLEX NUMBERS 15 opposed to real analysis, offers a much more direct approach. The clue lies in a direct connection between the exponential function and the trigonometric functions, to be derived in Chap. 2, Sec. 5. Until we reach this point the reader is asked to subdue his quest for complete rigor. EXERCISES i. Find the symmetric points of ¢ with respect to the lines which bisect the angles between the coordinate axes. 2 Prove that the points ai, a2, a; are vertices of an equilateral triangle if and only if ai -+ a? + a? = aaa + Gotz + Arta. 3. Suppose that « and } are two vertices of a square. Find the two other vertices in all possible cases. 4. Find the center and the radius of the circle which circumscribes the triangle with vertices a1, a1, a2. Express the result in sy mmetrie form. 2.2. The Binomial Equation. From the preceding results we derive that the powers of a = r(cos ¢ + isin ¢) are given by (19) a* = r*(cos mp + isin ng). | This formula is trivially valid for n = 0, and since a = (cos p — isin g) = rfoos (—¢) + isin (—¢)] it holds also when n is a negative integer. For r = 1 we obtain de Moivre’s formula (20) (cos g £ ¢ sin g)* = cosny + isinny which provides an extremely simple way to express cos np and sin ng is = terms of cos 9 and sin ¢. To find the nth root of a complex number @ we have to solve the equation amen ee 21) Poa16 COMPLEX ANALYSIS This equation is certainly fulfilled if »» =r and 2¢=y Hence we obtain the root t= /r(couf + «sin 2) where ~/r denotes the positive nth rool of the positive number r. But this is not the only solution. In fact, (22) is also fulfilled if n@ differs from ¢ by 8 multiple of the full angle. If angles are expressed in radians the full angle is 2x, and we find that (22) is satisfied if and only if 4 Em, om fakes, where k isany integer. However, only the valuesk = 0,1,...,% —1 give different values of 2. Hence the complete solution of the equation @1) is given by sf [om (2 +42) + tain (2+ 42)), k-O1,...jn—1. There are n nth roots of any complet number =O. They have the same modulus, and their argumenis are equaily spaced. Geometrically, the nth roots are the vertices of a regular polygon with n sides. The case a = 1 is particularly important. The roots of the equation 2 = | are called nth roots of unity, and if we set = cos 2% or (28) @ = cos + dain all the roots can be expressed by 1, a, w7,... , eo". It is also quite evident that if 0. An easy argument shows that this distinction is independent of the parametric representation. EXERCISES 3, When does az + 0 + ¢ ~ 0 represent a line? 2. Write the equation of an ellipse, hyperbola, parabola in complex form. 3. Prove that the diagonals of a parallelogram bisect each other and that the diagonals of « rhombus are orthogonal. “ - & Prove analytically that the midpoints of parallel chords to a gircle ‘He on 4 diameter perpendicular to the chords. a & Show that ell citsles that’ pass through « and 1/4 intersect the ‘Wivele fel" 1 at-zight angles“18 COMPLEX ANALYSIS 2.4, The Spherical Representation. For many purposes it is useful to extend the system C of complex numbers by introduetion of a aymbal to represent infinity. Its connection with the finite numbers is estab- lished by setting ¢@ + © = © +¢@= © forall finite c, and b-o=w-b= oe for all 6 #0, including b= . It is impossible, however, to define © + © and @- » without violating the laws of arithmetic. By special convention we shall nevertheless write 6/0 = © fora # Oandb/o =0 for b # ©, In the plane there is no room for a point corresponding to © , but we can of course introduce an “ideal” point which we call the point at infinaty. The points in the plane together with the point at infinity form the ertended complex plane. We agree that every straight line shall pase through the point at infinity. By contrast, no half plane shall contain the ideal point. It is desirable to introduce a geometric model in which all points of the extended plane have a concrete representative. To this end wo con- sider the unit sphere S whose equation in three-dimensional space is at +f 42% —= 1. With overy point on S, except (0,0,1), we can assooi- ate a complex number (24) go titi and this correspondence is one to one. Indeed, from (24) we obtain a+d iti “G-a) T-as and hence es (25) ne i =z i. Further computation yields e+2 1-7 74102 I a e vm 3° FED The correspondence can be completed by letting the point, at infinity correspond to (0,0,1), and we can thus regard the sphere as a repre- gentatinn at the avtandad nlene ar of the evtanded numhar evetom) | |6OWeCOMPLEX NUMBERS 138 hemisphere x; > 0 to its outside |zj > 1, In function theory the sphere &S is referred to as the Riemann sphere, If the complex plane is identified with the (x1,z:)-plane with the mr and sraxis corresponding to the real and imaginary axis, respectively, the transformation (24} takes on a simple geometric meaning. Writing z= + ty we can verify that Q7) ey: —1 = fizete — 1, and this means that the points (2,y,0) (z,2:,01), and (0,0,1) are in a straight line. Hence the correspondence is a central projection from the center (0,0,1) as shown in Fig. 1-3. It is called a stereographic projection. The context will make it clear whether the stereographic projection is regarded as a mapping from S to the extended complex plane, or vice versa, In the spherical representation there is no simple interpretation of addition and multiplication, Its advantage lies in the fact that the point at infinity is no longer distinguished. It is geometrically evident that the stereographic projection trans- forms every straight line in the z-plane into a circle on 8 which passes through the pole (0,0,1), and the converse is also true. More generally, any circle on the sphere corresponds to a circle or straight line in thez-plane. To prove this we observe that a circle on the sphere lies in a plane @at1 + ante + aati = ap, where we can assume that of + a? + af = 1 and 0 5 as <1. In terms of z and 2 this equation takes the form anle +2) — asile — 2) + anllel? — 1) = anlle? +0) or {ay — a3)(2" + 97) — Gaye — Za + a + a2 = 0. For os ¥ a: this is the equation of a circle, and for ac = as it represents 8 straight Tine. Conversely, the equation of any circle or straight line AN: =x20 COMPLEX ANALYSIS ean be written in this form. The correspondence is consequently one to one. It is easy to calculate the distance d(z,2’) between the stereographic projections of z and z, If the points on the sphere are denoted by (21,2221), (GLtea4), we have first (a: — 2i)® + (as — 28 + Crs — 2h)? = 2 — Bleazt + cox + aur). ¥rom (35) and (36) we obtain after a short computation styey tet + 275 = G+ OE +7) — @ Ae - 7) 4+ Ce? - DUE? ~ 1) @+ eh + FP — Oth) + 2) — 22 — a+ 0 + eP) As a result we find that (28) alee) = Se al VO+EPG + kh) For 7 = the corresponding formuls is 2 4,2) = Spy EXERCISES 1. Show thas z and 2’ correspond to diametrically opposite points on the Riemann sphere if and only: if 22 = -1, 2 A cube has its vertices on the sphere S and its edges parallel to the coordinate axes, Find the stereographic projections of the vertices. 3. Same problem for a regular tetrahedron in general position. 4, Let Z, 2’ denote the stereographie projections of z, 2’, and let N be the north pole. Show that the trianglea NZZ’ and Nzz’ are similar, and use this to derive (28). 5. Find the radius of the spherical image of the circle in the plane whose center is a and redius RB.2 COMPLEX FUNCTIONS 1. INTRODUCTION To THE CONCEPT oF ANALYTIC FUNCTION The theory of functions of 2, complex variable aims at extending calculus to the complex domain. Both differentiation and inte- gration acquire new depth and significance; et the same time the range of applicability becomes radically restricted. Indeed, only the analytic or holomorphic functions can be freely differentiated. and integrated. They are the only true “functions” in the sense of the French “Théorie des fonctions” or the German “Funktionentheorie.” Nevertheless, we shall use the term “function” in its modern meaning. Therefore, when stepping up to complex numbers we have to consider four different: kinds of functions: real functions of a real variable, real functions of a complex variable, complex functions of a real variable, and complex functions of a complex variable. Asa practical matter we agree that the letters 2 and w shall always denote complex variables; thus, to indicate a complex function of a complex variable we use the notation w = f(z).t ‘The notation y = f(2) will be used in a neutral manner with the understanding that + and y can be either real or complex. When we want to indicate that a variable is definitely restricted to real values, we shall usually denote it by f. By these agreements we + Modern students sso well aware aware that f standa for the function and jz) ~for » value'of the function. However, analysts are traditionally minded and continue to apeak of: ‘the fanetion fle)? ao:a COMPLEX ANALYSIS do not wish to cancel the earlier convention whereby a notation z = x + ty automatically implies that z and y are real. . It is essential that the law by which a function is defined be formulated in clear and unambiguous terms. In other words, all functions must be well defined and consequently, until further notice, single-valwed.f Tt is not necessary that a function be defined for all values of the independent variable. For the moment we shall deliberately under- emphasize the role of point set theory, Therefore we make merely an informal agreement that every function be defined on an open set, by which we mean that if f(a) is defined, then f(z) is defined for all x suffi- ciently close toa. The formal treatment of point set topology is deferred until the next chapter. LI. Limits and Continuity. The following basie definition will be adopted: Definition 1. The function f(z) is said to have the limit A as x tends to a, a) lim f(z) = A, if and only tf the following is true: For every ¢ > 0 there exists a number § > 0 with the property that \f@) — Al < « for all values of x such that |x — al < 5 and x ¥ a. ‘4 This definition makes decisive use of the absolute value. Since the notidti of absolute value has a meaning for complex as well as for real numbers, we can use the same definition regardless of whether the variable a‘and the function f(z) are real or complex. + Att an alternative simpler notation we sometimes write: f(z) — A for a, 4 There are some familiar variants of the definition which correspond tthe case where a or A is infinite. In the real case we can distinguish bebween the limits + © and — ©, but in the complex case thore is only oré-infinite limit. We trust the reader to formulate correct definitions to’vover all the possibilities. ' «The well-known results concerning the limit of a sum, a product, and 8 quotient continue to hold in the complex case. Indeed, the proofs depend only on the properties of the absolute value expressed by abl = fal - |b] and = a + Bi S [al + [BI + We shall sometimes use the pleonastic term single-valued function to underline ‘that the function has only one value for each value of the variable.COMPLEX FUNCTIONS 23 Condition (1) is evidently equivalent to @ lim Flay = From (1) and (2) we obtain lim Re f(z) = (3) lim Im f(z) = Im A. Conversely, (1) is a consequence of (3), The function f(z} is said to be condinuous af a if and only if lim f(z) = f(a). A continuous function, without further qualification, ma is one which is continuous at all points where it is defined. The sum f(z) + g(x) and the product f(z)g(z) of two continuous fune- tions are continuous; the quotient f(z}/q(x) is defined and continuous at aif and only if g(a) # 0. If f(z) is continuous, so are Re f(a), Im f(z), and |/(z)|- ‘The derivative of a function is defined asa particular limit and cau be considered regardless of whether the variables are real or complex, The formal definition.is 4 f'@) = tim £2 -f@, ato z~-a “The usual rules for forming the derivative of a sum, a product, or a : quotient are all valid. The derivative of a composite function is deter- + Inined by the chain rule. There is nevertheless a fundamental difference between the cases of a ‘ real and a complex independent variable. To illustrate our point, let Hz) be a real function of a compler variable whose derivative exists at z=a. Then J’(e) is on-one side real, for it ia the limit of the quotients fla +h) —f@ & S&F teats to zero through real values, On the other aide it is also the ‘Ymit of the quotients Kat ss -{@ pend ag such purely imaginary. Therefore f(a) must be zero, Thus a teal fanetion of a complex variable either has the derivative zero, or else the dorivative does not exist. 2.2a COMPLEX ANALYSIS The case of a complex function of areal variable can be reduced to the real case. If we write 2(/) = x(¢) + ty() we find indeed ZH 270 +i’, and. the existence of ¢‘(é) is equivalent to the simultaneous existence of w'() and 7/(). The complex notation has nevertheless certain formal advantages which it would be unwise to give up. In contrast, the existence of the derivative of a complex function of 9 complex variable has far-reaching consequences for the structural proper- ties of the function. The investigation of these consequences is the cen- tral theme in complex-function theory. 1.2, Analytic Functions, The class of analytic functions is formed by the complex functions of a complex variable which possess a derivative wherever the function is defined. The term holomorphic function is used with identical meaning. For the purpose of this preliminary investiga- tion the reader may think primarily of functions which are defined in the whole plane. The sum and the product of two analytic functions are again analytic. The same is true of the quotient f(z)/g(z) of two analytic functions, pro- vided that g(z) does not vanish. In the general case it is necessary to exclude the points at which g(2) = 0. Strictly speaking, this very typi- eal case will thus not be included in our considerations, but it will be clear that the results remain valid exeept for obvious modifications. The definition of the derivative can be rewritten in the form on LEA) -~ SO [@=inS—. As a first consequence f(z) is necessarily continuous. Indeed, from Fe +h) — $@) = h- UY@ +b) — f@))/h we obtain lim (ff + 8) —f@) = 0° F'@ = 0. If we write f(z) = u(z) + tv(z) it follows, moreover, that u(z) and #(2) are both continuous. The limit of the difference quotient must be the same regardless of the way in which A approaches zero. If we choose real values for h, then the imaginary part y is kept constant, and the derivative becomes a partial derivative with respect tor. We have thus ay xy mu, fO-357 at tagGOMPLEX FUNCTIONS a Similarly, if we substitute purely imaginary values tk for h, we obtain . fle + ik} — fe) oF au f@ = kin ety +e It follows that f(c) noust satisfy the partial differential equation of of ® 3a ~~ ay which resolves into the real equations a oe a ayy oz ‘These are the Cauchy-Riemann differential equations which must be satisfied by the real and imaginary part of any analytic funetion.t We remark that the existence of the four partial derivatives in (6) is implied by the existence of f’(2). Using (6) we can write down four formally different expressions for f’(z); the simplest is @) py = ee, For the quantity {f'(e)[? we have, for instance, The last expression shows that |f/(z)|* is the Jacobian of u and # with respect to # and y, We shall prove later that the derivative of an analytic function is itself analytic. By this fact u and » will have continuous partial deriva- tives of all orders, and in particular the mixed derivatives will be equal. Using this information we obtain from (6) ae, ote da=e5+2,=0 = oe Ay = a +28 =o. A function u which satisfies Laplace’s equation Au = 0 is said to be harmonic, The real and imaginary part of an analytic function are thus harmonic. If two harmonic functions u and v satisfy the Cauchy- Riemann equations (6), then v is said to he the conjugate harmonic func- t Augustin Cauchy (1780-1857) and. Bernhard Riemann (1826-1869) are regarded. an the founders of couplex-funotion, theary- Blemsmle ore emphasised the geo” metric agpreta in gonirent.to.the purely analytic approngh of Cauchy. :as COMPLEX ANALYSIS tion of 4. Actually, » is determined only up to an additive constant, so that the use of the definite article, although traditional, is not quite accu- rate. In the same sense, u is the conjugate harmonic function of —v. This is not the place to discuss the weakest. conditions of regularity which can be imposed on harmonic functions. We wish to prove, how- ever, that the function u + @ determined by a pair of conjugate har- monic functions is always analytic, and for this purpose we make the explicit assumption that u and » have continuous first-order partial derivatives. It is proved in calculus, under exactly these regularity con- ditions, that we can write ule + hay +1) — weg) = Fh + ek +e ole + hy +k) — o(e,y) = eh +Eb+ ta, where the remainders «), ¢: tend to zero more rapidly than h + k in the sense that e:/{h + tk) 0 and e2:/(h + ik) 20 fork +ik—>0. With the notation f(z) = u(c,y) + ivGz,y) we obtain by virtue of the rela- tions (6) fle ++i) — J) = H+) O+ a te ties and hence im Leb ath) -—f@ _ du, ov we” RR ae fae We conclude that f(z) is analytic. -s If u(z,y) and 0(z,y) have continuous firsi-order partial derivatives which satisfy the Couchy-Riemann. differential equations, then f(z) = u(z) + t(z) te analytic with continuous derivative f’(z), and conversely. The conjugate of a harmonic function can be found by integration, and in simple cases the computation can be made explicit. For instance, «= 2? — y? is harmonic and du/dz = 22, du/dy = —2y. The conju- gate function must therefore satisfy ov oy BT 2y, wy = 2. From the first equation » = 2cy + oy), where oy} is a function of y alone. Substitution in the second equation yields y/(y) = 0, Hence ¢(y) is 4 constant, and the most general conjugate function of 2? — y? is 2zy + ¢ where cis constant. Observe that x? — y+ 2izy = 2. The analytic function with the real part ct — y? is hence z® + tc.COMPLEX FUNCTIONS 2 There is an interesting formal procedure which throws considerable light on the nature of analytic functions. We present this procedure with an explicit warning to the reader that it is purely formal and does not possess any power of proof, Consider a complex function f(z,y) of tworeal variables. Introducing the complex variable z = z + ty and its conjugate Z = x — iy, wé can write z = (+2), y = —2i(2—2). With this change of variable we ean consider f(x,y) a8 a function of 2 and 2 which we will treat as inde- pendent variables (forgetting that they are in fact conjugate to each other). If the rules of calculus were ve we would obtain af (2- in 4 an %) az 2 “5 as oy, These expressions have no convenient definition aa limite, but we can nevertheless introduce them as symbolic derivatives with respect to 2 and Z. By comparison with (5) we find that analytic functions are char- acterized by the condition df/az = 0, We are thus tempted to say that an analytic function is independent of 2, and a function of 2 alone. ‘This formal reasoning supports the point of view that analytic func- tions are true functions of a complex variable as opposed to functions which are more adequately described as complex functions of two real variables. By similar formal srguments we can derive 4 very simple method which allows us to compute, without use of integration, the analytic function j(z) whose real part is a given harmonic function u(z,y). We remark first that the conjugate function f(z) has the derivative zero with respect to 2 and may, therefore, be considered as a function of 2; we denote this function by f@). With this notation we can write down the identity . ulay) = Hi@ + iy) + fle — wl. It is reasonable to expect that this is a formal identity, and then it holds even when z and y are complex. If we substitute x = 2/2, y = 2/2i, we obtain u(e/2, 2/28) = 3@) + FO). Since f(z) is only determined up to a purely imaginary constant, we may as well assume that f(0) is real, which implies f(0) = u(0,0). The fune~ tion f(z) can thus be computed by means of the formula J) = 2u(e/2, 2/24) — u(0,0). A purely imaginary constant can be added. at will. Tn thin farm the method ia definitely Hmited ta functions uftw which28 COMPLEX ANALYSIS. are rational in x and y, for the function must have a meaning for com- plex values of the argument. Suffice it to say that the method can be extended to the general case and that a complete justification can be given. EXERCISES 1 If g(w) and f(z) are analytic functions, show that g(f(z)) is also analytic. 2 Verify Cauchy-Riemann’s equations for the functions 2* and 2*. 3% Find the most general harmonic polynomial of the form az* + brty + czy? + dy? Determine the conjugate harmonic function and the corresponding analytic function by integration and by the formal method. 4. Show that an analytic function cannot have a constant absolute value without reducing to a constant. 5. Prove zigorously that the functions f(e) and J(@) are simultaneously analytic. 6. Prove that the functions u(z) and u(2) are simultaneously harmonic. 7. Show that a harmonic function satisfies the formal differential equation ate 2 a3 1.3. Polynomials. Every constant is an analytic function with the derivative 0. The simplest nonconstant analytic function is z whose derivative is 1. Since the sum and product of two analytic functions are again analytic, it follows that every polynomial 7% P@) =ao+az+ ++: fa, is an analytic function. ts derivative is P'(@) =a + 2a2+ °° -° + na. The notation (7) shall imply that @, x0, and the polynomial is then said to be of degree x. The constant 0, considered as a polynomial, is in many respects exceptional and will be excluded from our considerations.{ For x > 0 the equation P(z) = 0 has at least one root. This is the so-called fundamental theorem of algebra which we shall prove later. If P(as) = 0, itis shown in elementary algebra that P(@) = (2 — a:)Pilz) where P,(z) is a polynomial of degree n — 1, Repetition of this process finally leads to a complete factorization. (8) P(@) = an(z — on)(@ — as) - + > (@ — an)COMPLEX FUNCTIONS a where the a, az, ... , a, ate not necessarily distinct. From the fac- torization we conclude that P(z) does not vanish for any value of z different from a1, 02, . , . , an. Moreover, the factorization is uniquely determined except for the order of the factors. If exactly & of the a; coincide, their common value is called a zero of P(g) of the order h. We fiud that the sum of the orders of the zeros of a polynomial ig equal to its degree. More simply, if each zero is counted ag many times as its order indicates, a polynomial of degree n has exactly m geros, The order of a zero a can also be determined by consideration of the successive derivatives of P(z) for z= a Suppose that a is a zero of order k. Then we can write P(z) = (2 — a)*Pi(z) with Pile) ¥ 0. Sue cessive derivation yields P(a) = P’(a) = +> = P® (a) =0 while P® (a) 0. Tn other words, the order of a zero equals the order of the first nonvanishing derivative. A zero of order | is called 4 simple zero and ig characterized by the conditions P(a) = 0, P’(e) # 0. As an application we shall prove the following theorem, known as Lucas's theorem: Theorem 1. [f aff zeros of a polynomial P(z) le in a half plane, then all zeros of the derivative P’(¢) He in the same half plane. From (8) we obtain P@®_ 1 . 1 ® P@ "reat tite Suppose that the half plane A is defined as the part of the plane where Im (@ — @)/b < 0 (gee Chap. 1, Sec. 2.3). If ay is in H and 2 is not, we have then -In > 0, n-o & But the imaginary parts of reciprocal numbers have opposite sign. Therefore, under the same assumption, Im b(z — «)-' <0. If this is true for all & we conclude from (9) that bP’) _ ¥ b I - Yin <9, 2 7 Pe) ~ z— ay and consequently P’(z) 0. In a sharper formulation the theorem tells us that the smallest convex polygon that contains the zeros of P(e) siso contains the zeros of P’tz).0 COMPLEX ANALYSIS 1.4, Rational Functions, We turn to the case of a rational function ao) RG) = ag given as the quotient of two polynomials. We assume, and this is essen- tial, that P(z) and Q(z) have no common factors and hence no common zeros, 2(z) will be given the value © at the seros of Q(z). It must therefore be considered as a function with values in the extended plane, and as such it is continuous. The zeros of Q(z) are called poles of R(z), and the order of & pole is.by definition equal to the order of the corre- sponding sero of Q(z). The derivative 15) = FO2@) — YEP ay Fe= ——eF exists only when Q(z) ~ 0. However, as a tational function defined by the right-hand member of (11), #’(z) has the same poles as R(z), the order of each pole being inereased by one. In case Q(z) has multiple zeros, it should be noticed that the expression (11) does not appear in reduced form. Greater unity is achieved if we let the variable 2 as well as the values R(@) range over the extended plane. We may define R(>) ag the limit of R(z) asz— ~, but this definition would not determine the order of a zero or pole at 2%. It is thorefore preferable to consider the function R(1/z), which we can rewrite as a rational function Ri(z), and set R(%) = Ra(0). Jf R.(0) = 0 or «, the order of the zero or pole at © is defined as the order of the zero or pole of Rx(2) at the origin. With the notation _Motag + st bast RG) = Foe bo bbe we obtain Re a en eh ae + an bom bu Foe where the power 2"-* belongs either to the numerator or to the denomi- nator. Accordingly, if m > 7 R(@) has a zero of order m —n at ©, if m 0 there exists an mo such that ja, — A| < efor 2 no. A sequence with a finite limit is said to be convergent, and any sequence which does not con- verge is divergent. If lima.. Gn — ©, the sequence may be said to diverge to infinity, Only in rare cages can the convergence be proved by exhibiting the limit, 80 it is extremely important to make use of a method that permits proof of the existence of a limit even when it cannot be determined explicitly. The test that serves this purpose bears the name of Cauchy. A sequence will be called fundamental, or 2 Cauchy sequence, if it satisfies the follow- ing condition: given any < > 0 there exists an a9 auch that |a, — an| < © whenever » 2 no and m 2 ns. The test reads: A sequence is convergent tf and only tf it is a Cauchy sequence. The necessity is immediate. If a,— A we can find ma such that la, — Al < e/2 for n= m. For mm Z no it follows by the triangle inequality that |an — anal S (a, ~ Al + lan — Al 0 there exists an 7 such that A,, < A + ¢, and it follows that o, SA,, A—e If A = +c there are arbitrarily large o., and A = — © if and only if a, tends to — ©, In all cases there cannot be more than one number A with these properties. The limes inferior can be defined in the same manner with inequalities reversed. It is quite clear that the limes inferior and limes superior will bbe equal if and only if the sequence converges to a finite limit or diverges to + orto —#, The notations are frequently simplified to fim and lim. The reader should prove the following relations: Him ag + lim Ba & lim (as + A.) S lim a + Tim #, hm a, + Em 8, fim (a, + 8.) $ lim a + Im 8,. Now we return to the sufficieney of Cauchy’s condition. From len — oa] <2we obtain la,| < a, + ¢ for 2 2 mo, and it follows that A= lim a, and a = lim a, are both finite. If a * A choose 4-4) ces and determine a corresponding m. By definition of a and A there exista an a, A—s with mn 2m. It follows that A—a=(A — an) + (am — on) + (an — a} < 3, contrary to the choiceCOMPLEX FUNCTIONS a 2.2. Series. A very simple application of Cauchy’s condition permits us to deduce the convergence of one sequence from that of another. If it is true that |b, — bs| S ja, — 4,| for all pairs of subscripts, the sequence {5s} may be termed a ceniraction of the sequence {a,} (this is not a standard term), Under this condition, if (a,} is a Cauchy sequence, so is: {tn}. Hence convergence of {a,} implies convergence of {b,}. An infinite series is a formal infinite sum (15) mtarte ss bantec:. Associated with this series is the sequence of its partial sums & = at ast tae The series is said to converge if and only if the corresponding sequence is convergent, and if this is the case the limit of the sequence is the sum of the series. Applied to 2 series Cauchy's convergence test yields the following condition: The series (15) converges if and only if to every «¢ > 0 there exists an no such that |, +- .4, + + > - + aay] <¢ forall n 2 noand p20. For p = 0 we find in particular that jon] << Hence the gen- eral torm of a convergent series tends to zero. This condition is necessary, but of course not sufficient. If a finite number of the terms of the series (15) are omitted, the new series converges or diverges together with (15). In the case of conver- gence, let R, be the sum of the series which begins with the term das. ‘Then the sum of the whole series is 8 = s + Ra The series (15) can be compared with the series (18) jaa] + Jaa] +--+ > tad ts: formed by the absolute values of the terms. ‘I'he sequence of partial gums of (15) is 9 contraction of the sequence corresponding to (16), for [an Gata o> * $F aeeel S foe] + lames] + ++ + laneol. There- fore, convergence of (16) implies that the original series (15) is convergent. A series with the property that the series formed by the absolute values of the terms converges is said to be absolutely convergent. 2.3. Uniform Convergence. Consider a sequence of functions f,(z), all defined on the aime set FE. If the sequence of values {n(e)] con- verges for every x that belongs to Z, then the limit f(z) is again a function on By ‘definition, if ¢ > 0 and = belongs to H there exists anno such a a Leo36 COMPLEX ANALYSIS For instance, it is true that . 1 im 1+ 3) r=2 for all x, but in order to have |(1 + 1/n)a — 2] = |a|/n < ¢ form = mo it is necessary that no > |z|/e. Such an mp exists for every fixed z, but the requirement cannot be met simultaneously for all x. We say in this situation that the sequence converges pointwise, but not uniformly. In positive formulation: The sequence {f,(r)] converges uniformly to f(x) on the set E af to every e > O there exists an ny auch that [f-(z) — f(x)| < © for ali n = ne and all in E. The most important consequence of uniform convergence is the following: The limit function of a uniformly convergent sequence of continuous functions ia itself continuous. Suppose that the functions f,(z) are continuous and tend uniformly to f(z) on the set FE. For any « > 0 we are able to find an n such that Ife) — f(@)| < 0/3 forallzin EZ. Letzbe apointin #. Because f,(x) is continuous at ay we can find § > O such that {fn(c) — fa(te)| < ¢/3 for all z in E with |x — 2¢| < 8 Under tho samo condition on z it follows that Wf@) — fed) = @) — FO) + Li) — fo + Ifo) — fledl <4 and we have proved that (z) is continuous at xp. In the theory of analytic functions we shall find uniform convergence much more important than pointwise convergence. However, in most cases it will be found that the convergence is uniform only on a Part of the set on which the functions are originally defined. Cauchy’s necessary and sufficient condition has a counterpart for uniform convergence. We assert: The sequence {f.(x)}} converges uniformly on E if and only if to every & > O there exisis an no such that |Ja(z) — falz)| < ¢ for allman = no and all cin EB. The necessity is again trivial. For the sufficiency we remark that the limit function f(2) exists by the ordinary form of Cauchy’s test. In the inequality |f.(z) — J.(z)| < ¢ we can keep n fixed and let m tend to «©, It follows that jj(z) — f.(z)| S eforn 2 mand allzin H. Hence the convergence is uniform. For practical use the following test is the most. applicable: If = sequence of functions {f,(z)} is a contraction of a convergent sequance of constants (a,}, then the sequence {f,(x)} is uniformly convergent. The hypothesis means that |f.(c) — f(z)| S lan — | on H, and the con-COMPLEX FUNCTIONS a7 clusion follows immediately by Cauchy’s condition, In the case of series this criterion, in a somewhat weaker form, becomes particularly simple. We say that a series with variable terms Flt) + fra) +o +h@+4+-°- has the series with positive terms Gta es bates for a majorant if it is true that |f.(2)| S Ma, for some constant Mf and for all sufficiently large n; conversely, the first series is a minorant of the second. In these circumstances we have Viale) + fotue) ++ + + ferele)| S Ma, + Gaz + +++ + ants) Therefore, if the majorant converges, the minorant converges uniformly. This condition is frequently referred to as the Wederstrass M test, It has the slight weakness that it applies only to séries which are also absolutely convergent. The general principle of contraction is more complicated, but has a wider range of applicability. EXERCISES 1, Prove that a convergent sequence is bounded. 2 If lim z, = A, prove that me tim 2 @ pesto + be) = de wre h 4 Show that the sum of an absolutely convergent series does not change if the terms are rearranged. 4. Diseuss completely the convergence and uniform convergence of the dequence {r2"}?. 5. Discuss the uniform convergence of the series < z 2 n(l + nz} for real values of 2. al CV=umtut+: ++, Vento: - are convergent series, prove that UV = aay 4 (eave 4 aa) + (cewa + eve + tats) +s + provided that at least one of the series is absolutely convergent. (It is easy if both series are absolutely convergent. ‘Try to arrange the proof so economically that the absohute convergence of the second series is not needed.)38 COMPLEX ANALYSIS 2.4. Power Series. A power series is of the form a7) Qo + az + agz? + - fag tees where the coefficients ¢, and the.variable z are complex. A little more generaily we may consider series ¥ an(e — a)" a=0 which are power series with respect to the center zo, but the difference isso slight that we need not do so in a formal manner. As an almost trivial example we consider the geometric series Leetotos tage: whose partial sums can be written in the form i— 2 —Zz lteter: pots Since 2 — 0 for {z| <1 and |z"| 2 1 for |e] = 1 we conelude that the geometric series converges to 1/(1 — z) for |z| < 1, diverges for |z| = 1 Tt turns out that the behavior of the geometric series is typical. Indeed, we shall find that every power series converges inside a circle and diverges outside the same circle, except that it may happen that the series converges only for z = 0, or that it converges for all values of 2. More precisely, we shall prove the following theorem due to Abel: Theorem 2. For every power series (17) there exists a number B, 0S RS ©, called the radius of convergence, with the following properties: (3) The series converges absolutely for every 2 with |e] B the terms of the series are unbounded, and the sertes is consequently divergent. Gii) In je] < B the sum of the series is an analytic function. The derivative can be obtained by termwise differentiation, and the derived sertes has the same radius of convergence. The circle |z| = # is called the circle of convergence; nothing is claimed about the convergence on the circle. We shall show that the assertions in the theorem are true if 2 is chosen according to the formulaCOMPLEX FUNCTIONS a This is known as Hadamard’s formula for the radius of convergenee. If |2| < & we can find p so that [2]

1/R, and by the definition of iimes superior there exists an no such that |a,|"* < 1/p, lan] < Ife" forn = no. It follows that (a,2*| < (\z]/p)* for large n, so that the power series (17) has a convergent geometric series as 4 majorant, and is consequently convergent. To prove the uniform convergence for lz] 3 ep < RB we choose a p’ with p < a! < Rand find |c,2"| € (o/p')* for nm 2m. Sinee the majorant is convergent and has constant terms we conclude by Weierstrass's M test that the power series is uniformly convergent. If |e] > R we choose p a0 that R

1/p, len] > 1/e%. Thus lenz*| > (jal/e)* for infinitely many n, and the terms are unbounded. The derived series ) raz") has the same radius of convergence, T because (/n— 1. Proof: Set 0, and by use of the binomial theorem 2 = (1 + 3,)* > 14+4n(n — 1)8. This gives 82 < 2/n, and hence 6,30. For [2| < & we shall write fa) = Foner = 448) + Role where al) =a tae t+ tanwt Ba) = 3 ot, and also = AG) = Yroget = lim £09, We have to show that f’(z) = S,(z). Consider the identity (1g) LYLE — pegy = (HO = AED — 3609) + (lea) = Hed) Pals) — Bales) + (BGS Bt), where we assume that 2 = z snd |2|, [so

O such that 0 < |z — zo| < 3 implies | $u(2) — 8,(20) zaas TS fn(ee) ‘When all these inequalities are combined it follows by (19) that when 0 < |z — zal < 5. We have proved that f'(zo) exists and equals Siler). Since the reasoning can be repeated we have in reality proved much more: A power series with positive radius of convergence has derivatives of all orders, and they are given explicitly by I@ =ay+az+az?*+--- F@) = a + 202 + 3ay?+ °°: 5°@) = 2a + Gaye + Wag? + - JO) = kta, 4+een et OE tt Lae In particular, if we look at the last line we see that a, = f"(0)/k!, and the power series becomes LOn,.. "1g9) $9 2504 FO2+ May... + ‘This is the familiar Taylor-Maclaurin development, but we have proved it only under the assumption that f(z) has a power series development. We do know that the development is uniquely determined, if it exists, but the main part ie still missing, namely that every analytic function has a [Peeler asmlenenont