°., ba(2—20)" has a positive radius of convergence Ro. We shall prove that their Cauchy product converges inPOWER SERIES 59 the disk |z — zo| < min{R1, Ro} to the product of the functions represented by the two original series. We can assume without loss of generality that 29 = 0. Suppose |z| < min{ Ry, Ry}. For W a positive integer we have ono y = YS ahsitt- > YO ajheit# On It follows that Ie dew’) (Sa) = ene’ | < Slats OSs RSW S+k>N < 3 Jaybyzit*| Fcmax{jk} Sw clude that as desired. Exercises 1. Deduce the law of exponents, ee = e+", from the power series represen- tation of e*.60 NOTES ON COMPLEX FUNCTION THEORY 2. Consider the correspondence that associates with cach power series 7°°.y an2” centered at 0 the infinite matrix (jx )% ja defined by — faj-rk forgzk ai Ag for j C is a piecewise C™ curve with (a) = 7(b), and let a, be a point in (a,b). Let 6) = a1 +(b—a), and define the curve 7 : [a1,b] —+ C by is a I, fl2dz = 0. (VL9) THE REVERSE OF A CURVE. If 7 : [a,b] — C is a curve, then the reverse of +y is the curve —y : [-b, a] —+ C defined by (—7)(t) = 7(-t). It is the curve one obtains from ¥ by reversing its direction. If 7 is piecewise C', then so is ~7, and [_, f(z\dz =— ff f(z)dz for any function f that is defined and continuous on 7. ‘The reader will easily verify this on the basis of the change of variables formula for the Riemann integral. (VL10) ESTIMATE OF THE INTEGRAL. Let 7 : [a,b] — C be a piecewise C! curve and f a continuous complez-valued function on y. Let M be the maximum of |f| on y, that is, M = mox(|f(7(#))|! @St 5 8}. ‘Then | f Hae] < ML(7). In fact, by (VI3) we have [[ sere iy 2 If fore IA , f[ ownin'eiae.COMPLEX INTEGRATION 67 The last integral is majorized by M f° |y'(t)|dt, which equals M L(7). (VI.11) INTEGRAL OF A LIMIT. Let 7 : [a,b] —> C be a piecewise C? curve. Let the sequence (fn)%2, of continuous complez-valued functions on converge uniformly on + to the function f. Then Jim, i fulz)dz = 7 flz)dz. In fact, let My denote the maximum of |f — fal on 7. By (VI10), | [tere - f teres Since lim My = 0 (by the hypothesis that f, —+ f uniformly on 7), the desired < M,L(7). = Lf Lie) = fale) dz n=00 conclusion follows. (V1.12) AN EXAMPLE. It is now possible to illustrate how complex meth- ods can be used to evaluate improper Riemann integrals that are inaccessible by the usual techniques of calculus. Many other examples of this kind will be given later, in Chapters VII and X, after more machinery has been developed. The function e~*” is by (V.16) represented in the whole plane by the power See tet) eee teens sorics Deo ar: Ot the power series patho de ofthe: Bente f is therefore the derivative of the function reproscnted by co (a1)hztntt n=0 GnFijnl # e~** around any closed curve in the plane is 0. . By (V1.6) we can conclude that the integral of ta + ib, oriented counterclockwise. By the observation in the last paragraph, fume ang Jia = We are going to hold 6 fixed and let a tend to oo. The preceding integral is the sum of the integrals of 6" over the four segumenis [—a, a}, {a,a + id}, [a + id, —a + id}, [-a+ib, —a}; let those integrals be denoted by Iy(a), I2(a), Is(a), Lu(a), respectively. Obviously, a sy lw) =f Mae. Parameterizing the segment [—a + ib,a + ib] by the function »(t) = ¢ + ib, where68 NOTES ON COMPLEX FUNCTION THEORY —a St 0, and letting R —+ oo. 2. Evaluate the integrals f° cost?dt and fy” sint?dt (the Fresnel integrals) by integrating e~*" around the triangle with vertices 0, R, Re“, where R > 0, and letting R — co.VII. SIMPLE VERSIONS OF CAUCHY’S THEOREM, AND CONSEQUENCES Cauchy’s theorem is the central theorem of complex function theory. It states, in one of its versions, that f\ f(z)dz = 0 whenever 7 is a closed curve and the function f is holomorphic in an open set containing +y and its interior. In a more sophisticated version, the single curve 7 is replaced by a system of curves A general version of Cauchy’s theorem is presented in Chapter IX; it requires considerable preparation. In the present chapter two special cases will be estab- lished, the case where 7 is a triangle, and, on the basis of that, the case where the function f is holomorphic in a convex open set containing . This will be enough to enable us to draw far-reaching conclusions concerning the behavior of holomorphic functions. Some applications to harmonic functions are given at the end of the chapter. (VIL. 1) CAUCHY’S THEOREM FOR A TRIANGLE. Let T be a tri- angle in C, and let f be a holomorphic function in an open set containing T and its Then {,, f(z\dz =O. ite Then fr fl2}de = 9. This is sometimes called Goursat’s lemma, after E. Goursat, who, in the late 1800's, was the first to recognize that, in the definition of a holomorphic function, one can dispense with the requirement that the derivative be continuous. To prove the theorem we assume for definiteness that T has the counterclock- wise orientation, and we let IT = J, f(z)dz. By means of segments joining the midpoints of the sides of T, we construct four triangles Tp;, j = 1,2,3,4, each sim- ilar to but one-half the size of T (see sketch). Orienting each Ta; counterclockwise, we have I. fede = > [ fledae. Hence there is a j such that le Fl2)de| > ale Pick such a j, and lei T; denote 7; for that 7. From T; we construct four triangles T1;, 7 = 1,2,3,4, in the same way that the triangles Tp; were constructed from T. Orienting each T;; counterclockwise, we have =a I fled, nf ot72 NOTES ON COMPLEX FUNCTION THEORY _— Carving up the triangle T ly s0, since |fn, $(z)de| > alt there must be a j such that |, Herte| > lth Pick such a j, and let Tp denote 7;; for that j. Proceeding in this manner, we obtain a sequence (T',) ~~, of triangles, each Ty being one of the four triangles obtained from T,_1 by the process of connecting midpoints, such that | [ jteae! 2 [ieee for each n. We note that L(T,) = 2-L(T) for each n.SIMPLE VERSIONS OF CAUCHY’S THEOREM, AND CONSEQUENCES 73 Let Kn be the union of T, and its interior. The sets K’, are compact, they decrease as n increases, and their diameters tend to 0. Hence, their intersection consists of a single point, which we call z. The point z is either on T or in the interior of T, and thus it is in the open set where f is holomorphic. Hence, by R (11.2), we can write f(z) = f(0) + f"(20)(2 ~ 20) + R(z), where lim , FO ae ee By (VI) the integral of the function f(zo) + {"(zo)(z — 20) around Tr is 0, 50 I. f(z)dz = I, R(z)dz. Ve thus have \z| <4" vi R(z)dz}. Th : R(z) For each n, let €, denote the maximum of for z in T,\{zo}. Then — 2 lim en = 0, aud for z in Tn, nae |R(z)| S €n diameter(In) < én L(Tn). Hence R(z)dz| < ) -L(Tn) (VII.2) CAUCHY’S THEOREM FOR A CONVEX EON: Let G n subset of C and f a holomorphic function in @. Then f, fejde = 0 for every piecewise C? closed curve +y in G. We recall that a subset of C is called convex if, whenever it contains two points x and zg, it contains also the entire segment with endpoints z; and z,. To prove the theorem we fix a point zy in G and define the function g in G by ac) =f oMs. We shall show that g is holomorphic and that g! = f. The desired conclusion will then follow by (V1.6).74 NOTES ON COMPLEX FUNCTION THEORY Fix a point z; in G. If z is any other point of G, then [Hows [+0,=1) F()de + i F(C)de = 0, ] [=,70] by Cauchy’s theorem for a triangle (or trivially in case the points zo, 21, 7 happen to be collinear). The preceding equality can be written as 92) ~ g(a) = F(¢)d¢. [21,2] Consequently 9(2) — 9(z1) = f(a) = 2-2 ff Fen]as. Now fix a positive number e. By the continuity of f, there is a positive number 6 such that [70 a fla)| < € whenever lc —a| <6. Hence, if |z—2:| < 6, then the absolute value of the integrand in the preceding integral is bounded by ¢, and we have 7 1 fe g(2) = g(21) fla) z- 24 "1 J. o-sen]ac = fm H7)- 94) _ 7, 0 OR Ry zy and g'(z1) = f(21), as desired. (VIL.3) EXISTENCE OF A PRIMITIVE. if the function f is defined in the open subset G of C, then by a primitive of f one means a holomorphic function in G whose derivative equals f. The preceding proof shows that f has a primitive in case it is holomorphic and @ is convex. The proof in fact establishes something that seems stronger; If the continuous complez-valued function f in the open conver set G has integral 0 around every triangle contained in G, then f has a primitive in G. This is not in reality stronger, because, as we shall show very shortly, the derivative of a holomorphic function is itself holomorphic (which means a function f satisfying the hypotheses in the preceding statement is in fact holomorphic). The statement will nevertheless be useful later, in Section (VIL.10), where a converse of Cauchy’s theorem is established. (VII.4) MORE DEFINITE INTEGRALS. Cauchy’s theorem for a convex region, although very special compared to the general Cauchy theorem, is by itselfSIMPLE VERSIONS OF CAUCHY’S THEOREM, AND CONSEQUENCES 75 enough for many applications. In the following exercises it is used to evaluate improper integrals. Exercises 1. Let 0 <8 <1. Derive the equality ma 1-b+¢2 d eae by integrating the function (1 + z?)~? around the rectangle with vertices +a, +a+ivo (a > 0) and taking the limit as a + co. What goes wrong if b > 1? 2. Let a be a posilive number. Evaluate the integrals 1, ~ ow galas i, apa i Bat by integrating the function (2? + a”)~? around the triangle with vertices 0, R, Re%# (R > 0) and taking the limit as R — co (VII.5) CAUCHY’S FORMULA FOR A CIRCLE. Let C be a counter- clockwise oriented circle, and let f be a holomorphic function defined in an open set containing C and its interior. Then for all points z in the interior of C. Thus, the values of f in the interior of C can be reconstructed from the values on C. A holomorphic function “hangs together.” Cauchy’s formula for a circle is a very special case of a more general formula, to be established i in Chapter X. The case of a circle is all we shall need to show tee hola Betad leenlle by Gaioeeea Gad tenn that holom sented locally by ps serics and thus are differentiable to all orders. To establish the formula we fix a point zp in the interior of C. For 0 < « < dist (zo,C), let C, be the circle with center zo and radius e, oriented counterclock- £O aca ace The preceding equality will be ae from Cauchy’s ees for a convex region. For each € we subdivide the region between C and C, into four subregions by means of two horizontal and two vertical segments connecting C with C:, as in the sketch. The boundary of each subregion is a closed curve made up of a subarc of C, a subarc of C,, and two of the segments, one horizontal and one vertical. wise. Our first task is to prove that Le FO) ae for every €.76 NOTES ON COMPLEX FUNCTION THEORY t + Let these boundary curves be denoted bn “ty 74 76) 74) and let them be given the £0 open halfplane with an open disk containing C' in which f is holomorphic). By Cauchy’s theorem for a convex region, therefore, ee eb aehieh Each 7; is @ convex open set in which the function is holomorphic (for example, in the intersection of a suitable fox =o, j= 1,234 nS El fe. [dL ee Since —2n we have the desired equality, I i. d= hi HO 4 ta"SIMPLE VERSIONS OF CAUCHY’S THEOREM, AND CONSEQUENCES 77 In Section (VI.7) it is shown that 1 dC = Qmi. [ 2. 6 — 20 This in conjunction with the preceding casa gives li fend in peta af. £9- SO) = F20) ae Imi Jo C— 2 — for each ¢ between 0 and dist (zo,C). Let M, denote the maximum of Bie 20 for ¢ on C;. The absolute value of the last integral is then bounded by ie L(C.), implying that x f Poa - 120) As € tends to 0 so does « M; (since 7 —+ |f"(z0)|), and the desired equality follows. S< eM. Dri Exercise Let f and C be as in the statement of Cauchy’s formula. What is the value of if fou when z is in the exterior of C? gue (VIL.6) MEAN VALUE PROPERTY. Let the function f be holomorphic in the disk i ~ zo| < R, and for 0 . g°—*(20) am” 0,1,2,.... But this is just the well-known Leibniz formuta for the n-th derivative of the prod- uct of two functions, which the reader has probably encountered in real analysis. The case n = 1 is just the product rule for differentiation, and a straightforward induction argument yields the general case. Exercises 1. If the function f is holomorphic in a neighborhood of the origin and % 4 a,2 is its Taylor series about the origin, what is the Taylor series about the origin of the function (1 — z)~? f(z)?SIMPLE VERSIONS OF CAUCIIY’S THEOREM, AND CONSEQUENCES 81 2. Prove that the Taylor scrics about the origin of the function [Log(1 — z)]” is Se tt, att tl aust Hk where Hp = 1+ 3 + 3 +++++— (the n-th partial sum of the harmonic series). n (VII.10) CONVERSE OF CAUCHY’S THEOREM. Let f be a contin- uous complea-valued function in the open subset G of C. If fp f(z)dz = 0 for every triangle T contained with its interior in G, then f is holomorphic. Because being holomorphic is a local property, it is enough to prove this for the case where G is a disk. In that case, as pointed out in Section (VII3), the hypotheses imply that f has a primitive, in other words, there is a holomorphic fanction g in G such that g! — f. Since holomorphic functions have holomorphic derivatives, the desired conclusion follows. Exercise Prove Morera’s theorem (a stronger converse to Cauchy’s theorem than the one above): If f is a continuous complex-valued function in the open subset G of C, and if fp f(z)dz = 0 for every rectangle R, with edges parallel to the coordinate axes, contained with its intcrior in G, then f is holomorphic. (VII.11) LIOUVILLE’S THEOREM. A function that is holomorphic in all of C is called an entire function. The theorem of Liouville states that the only bounded entire functions are the constant functions. For the proof, assume f is a bounded entire function, say |f(z)| 0 let Cp denote the circle |z — z| = R, with the counterclockwise orientation. By (VIL8)(i) we have fo) = sf Pow. mi Jo, (e — 20)? i M The integrand in the preceding integral is bounded in absolute value by -p7, giving the estimate IF'eo)l < This being true for all R > 0, we have f"(zo) = 0. Since the derivative of f vanishes identically, the function f must be constant, as desired.82 NOTES ON COMPLEX FUNCTION THEORY Exercises 1. Let f be an entire function satisfying A limsup a] < 00, 2 where n is a positive integer. Prove that f is a polynomial of degree at most n 2. Let f be a holomorphic map of the open unit disk into itself. Prove that If@| for all z in the disk. 3. Prove that an entire function with a positive real part is constant. (VII.12) FUNDAMENTAL THEOREM OF ALGEBRA. Every non- constant polynomial with complex coefficients can be factored over C into linear faciors. Purely algebraic reasoning reduces this to the statement that every nonconstant polynomial with complex coefficients has a complex root. In fact, suppose the preceding statement is known, and let p be a nonconstant polynomial with complex coefficients. Let zo be a root of p. The standard division algorithm of algebra then enables one to write p(z) = (z — z0)q(z), where q is a polynomial whose degree is one less than that of p. The same reasoning can now be applied to g, and so on, yielding the desired linear factorization in finitely many steps. To prove that nonconstant polynomials with complex coefficients have complex roots we argue by contradiction. Suppose p(z) = anz"+an—12"-) +-+-+a1z+ag is a polynomial of positive degree n. with complex coefficients. but without complex 1 roots. Let f = 270, Then f is an entire function. Since a, # 0 we can write, for 7: ‘The quantity in parentheses on the right will be no smaller than 5 when |2| is sufficiently large, say when jzj > @. HeuceSIMPLE VERSIONS OF CAUCHY'S THEOREM, AND CONSEQUENCES 83 On the other hand, since |f| is a continuous function, it is bounded on the compact set |z| < R. Thus f is a bounded entire function, and so it is constant by Liouville’s theorem. This is the desired contradiction, because the polynomial p was assumed to be nonconstant. The proof just given is purely an existence proof; it does not provide one with a method of locating the roots of a polynomial. The problem of finding accurate approximations to the roots of complex polynomials is a major theme in numerical analysis. (VII.13). ZEROS OF HOLOMORPHIC FUNCTIONS. Zeros of holo- morphic functions can be treated like zeros of polynomials. Suppose f is a holo- morphic function in the open subset G of C, and suppose 29 is a point of G such that f(zo) = 0. The point zp is called a zero of f of order (or multiplicity) m, where m is a positive integer, if f(")(z)) = 0 for n =0,...,m—1, while f(™(29) #0. In that case the Taylor series for f about 2) has the form 7°, an(z— 20)", where £O"(z0) 5 o m. ; Sy #0. We can write f(z) = (2 — 20)"9(z), where the function g, gz) = {© — 2)" f(z), 2 € G\f{zo} am; z= 2, is holomorphic in @. The function g, being nonzero at zp, is nonzero in some neighborhood of zy, implying that zy is at a positive distance from f~1(0)\{zo}, the set of all the other zeros of f. In other words, zeros of finite order of holomorphic functions are isolated. ‘The point 29 is called a simple zero of f if it is a zero of order 1. In that case the function g defined above takes the value f!(zo) at 2. The reasoning above shows that if zo is a zero of f of any order, then the function that equals (z — zo)~1 f(z) in G\{zo} and equals f’(za) at zo is holomorphic. What about zeros of infinite order, that is, points at which f and all of its derivatives vanish? It turns out that these exist only in trivial cases. First, if G is connected, then f can have a zero of infinite order only if it is the zero function. ee n set, arid, by the discussion in the preceding paragraph, so do the points of G that are not zeros of f of infinite order. If G is connected, then one of these sets must be empty. If G is not connected, and if f has a zero of infinite order at the point 2) of G, one can at. least conclude that f vanishes identically in the connected component of G containing zo. In contrast to the behavior of holomorphic functions, it is possible for a real- valued C° function on R to have a zero of infinite order and yet. not. vanish iden- tically. A standard example is the function that equals 0 at the origin and equals e~1/** elsewhere. One can show that the origin is a zero of infinite order of this function.84 NOTES ON COMPLEX FUNCTION THEORY Most often in complex function theory one is concerned with holomorphic func- tions in connected open sets. The following statement summarizes what is estab- lished above about the zeros of such functions. Let f be a holomorphic function, not the zero function, in the connected open subset G of C. Then cach zero of f is of finite order, and f—1(0), the zero set of f, has no limit points in G. Since an uncountable subset. of C always contains limit points of itself, one can conclude in particular that the set f~1(0) is at most countable. The same is true of f-?(w), for any complex number w, provided f is not the constant function w. Exercises 1. Prove that there is no holomorphic function f in the open unit disk such that f(4) = 27" for n = 2,3,.... 2. Let f be a holomorphic function in the open subset G of C. Let the point 2» of G be a zero of f of order m. Prove that there is a branch of f# in some open disk centered at zo. (VII.14) THE IDENTITY THEOREM. Lei f and y be holomorphic func- tions in the connected open subset G of C. If f(z) = 9(z) for all z in a subset of G that has a limit point in G, then f = 9. Thus, for example, a holomorphic function in the open unit disk is uniquely determined by the values it assumes on the set {2-" : n = 1,2,3,...}. ‘To prove the theorem one notes that, if f and g agree on a subset of G that has a limit point in G, then the zero set. of the function f —g has a limit point in G, which implies by what is proved in the last section that f — g is the zero function. Exercises 1. Prove that there is no holomorphic function f in the open unit disk such that i(2)- De ee Ore a ae 2. Let G be a nonempty, connected, open subset of C which is symmetric with respect to the real axis. Let f be a holomorphic function in G such that f is real-valued on GR. Prove that f(Z) = J(z) for all z in G. (Suggestion: Use Exercise 2 of Section (IL8).) 3. Let G be as in the preceding exercise, and let f be a holomorphic function in G such that f is real-valued on a nonempty subinterval of GR. Prove that f is real-valued on all of GN R.SIMPLE VERSIONS OF CAUCHY’S THEOREM, AND CONSEQUENCES 85 (VII.15) WEIERSTRASS CONVERGENCE THEOREM. Let G be an open subset of C and (fx)g_, @ sequence of holomorphic functions in G that converges locally uniformly in G to the function f. Then f is holomorphic, and for cach positive integer n, the sequence (f{”)<, converges locally uniformly in G to fo. Once again we have a contrast between real function theory and complex func- tion theory. The local uniform convergence on R of a sequence of real-valued C™ functions docs not imply even the differentiability, let alone the indefinite differen- tiability, of the limit function. In fact, by the classical Weierstrass approximation theorem, any continuous real-valued function on R, even a nowhere differentiable one, is the local uniform limit of a. sequence of polynomials. ‘To prove the Weierstrass convergence theorem, fix a point 29 in G, and let r be a positive number such that the disk |z ~ zo| 0 into the open unit disk. Prove that < lelfe)P If@l < Re for all z in the half-plane. 5. Prove that, except for the identity function, a holomorphic map of the open unit disk into itself has at most one fixed point in the disk. (Vii.i8) EXISTENCE OF HARMONIC CONJUGATES. Zeit u be a real-valued harmonic function in the convex open subset G of C. Then there is a holomorphic function g in G such that u = Reg. The function g is unique to within addition of an imaginary constant. In other words, u has a harmonic conjugate in G, unique to within addition of a constant. The conclusion is true under a much weaker assumption on G than convexity, as will be shown later on the basis of the general Cauchy theorem. To prove the proposition we consider the function f = =— Ou which is holomorphic in G (by Exercise 2 of Section (II.16)). As shown in the proof of Cauchy’s theorem for a convex region, the function f has a primitive in G; that nection g in G euch that g! — f. Fix a p. G. By is, ther 2 holomorp! adding a constant to g, we may assume g(79) We hi: fe have a go = atta and also =a oon i ae "ay" a Hence ee a and $Y = a The first partial derivatives of the function u—u thus vanish throughout G, implying that u — uy is constant. Since u(zo) = wi(zo), the constant is 0, and «= val As for uniqueness, if two holomorphic functions in G have the same real part, then their difference has derivative 0, by the Cauchy-Riemann equations. In that case the functions differ by a constant.SIMPLE VERSIONS OF CAUCHY’S THEOREM, AND CONSEQUENCES 89 Exercises 1. Prove that a real-valued harmonic function u in the open unit disk has a series representation u(re"®) = ao + > r"(ancosnO + bn sinné), n=l where the coefficients ap and bp are real numbers. 2. Prove that a positive harmonic function in C is constant. 3. Let u be a nonconstant real-valucd harmonic function in C. Prove that the set u71(c) is unbounded for every real number c. (VIL.19) INDEFINITE DIFFERENTIABILITY OF HARMONIC FUNCTIONS. Harmonic functions are of class C*. Because differentiability is a local property, it will suffice to prove this for a harmonic function w in an open disk D. We may assume without loss of generality ae ea eee erage ere eT a BY that u is real-valued. Then, by (VIL18), we can te u — Reg with g holomor- phic in D. The desired conclusion now follows by the indefinite differentiability of holomorphic functions. (VII.20) MEAN VALUE PROPERTY FOR HARMONIC FUNC- TIONS. Let u be a harmonic function in the open subset G of C, and let 29 be a point of G. Then = uz) = xf u(zo t+reit\dt, 0 0 such that90 NOTES ON COMPLEX FUNCTION THEORY the disk |z—%0| 1 by a l-z re different series, the series — —z oo a series in negative powers of z. For a slightly more complicated example, consider the function eaness mae : 1 1 : . . which can be written as ——; ~ ———. The function — is represented in the disk |z| < 2 by a power series in z, and we have just seen that the function is represented in the region jz| > i by a series in negative powers of z. Hence, by using a series in both positive and negative powers of z, we can represent the function ———> in the annulus 1 < |z| < 2. (z= 1(z-2) Al Exercise Let a and 6 be complex numbers such that 0 < |al < |b]. Find a series in (2 = a)(z— positive and negative powers of z that represents the function annulus |a| < |z| < [6]. (VIII.2) LAURENT SERIES. A Laurent series centered at the point zo of C is, by definition, a series of the form 7° __, an(z — zo)", where the coefficients a, are complex constants. The series is said to converge at the point z if both of the series 1°29 an(z — zo)" and \>%, a_n(z — zy)~" converge. In that case lim, a,(z — 20)” exists and is defined to be the sum of the series. Let Ri = limsup|a_n|!/", 1 R, = ———_-. ? Tim sup Jan|*/* nes92 NOTES ON COMPLEX FUNCTION THEORY By the Cauchy-Hadamard theorem (V.12), the series 37°) an(z— 20)" con- verges absolutely and locally uniformly in the disk |z — zo| < Re, and the series © Ja—n(z—20)~ converges absolutely and locally uniformly in the region |z — 20| > Ri. Hence, if Ry < Ro, then the Laurent series [%_.. dn(z — 20)” converges absolutely and locally uniformly in the annulus Ri < zo| < Ro, called the annulus of convergence of the series. Assume Ri < Re, and let f be the function to which the Laurent series converges in the annulus Ri < |z—2zo|< Ro, in other words, let f(z) = an(z — 20)". Then f is holomorphic, by the Weierstrass convergence . For Ry R. We have aaa if ¢-2 ~ (€-2)— (2-2) iz 1 — 2) |1— G2 =z eae nai the series converging locally uniformly (with respect to ¢) in the region |¢ — zo| < Jz — zo|, and hence uniformly on 7. The series — YS HONG - 20)" Me - 20)" n= 9(¢) thus converges uniformly to ago We can therefore integrate the series term-by-term over +7 to obtain in other words, the Laurent expansion f(z) = S771_,, an(z — 20)", with an f no) OCMC, , holds in the region |z — zo| > R. (VIII.4) CAUCHY’S THEOREM FOR TWO CONCENTRIC CIRCLES. Let the function f be holomorphic in the annulus Ry < |z — 2| < Ko For Ri