Download the full version of the ebook at

https://ebookfinal.com

Several Complex Variables with Connections to

Algebraic Geometry and Lie Groups 1st Edition
Joseph L. Taylor

https://ebookfinal.com/download/several-complex-
variables-with-connections-to-algebraic-geometry-
and-lie-groups-1st-edition-joseph-l-taylor/

Explore and download more ebook at https://ebookfinal.com

 Recommended digital products (PDF, EPUB, MOBI) that
you can download immediately if you are interested.

Complex Variables 1st Edition Joseph L. Taylor

https://ebookfinal.com/download/complex-variables-1st-edition-joseph-
l-taylor/

ebookfinal.com

The structure of complex Lie groups 1st Edition Dong Hoon

Lee

https://ebookfinal.com/download/the-structure-of-complex-lie-
groups-1st-edition-dong-hoon-lee/

ebookfinal.com

Functional Equations and Inequalities in Several Variables

1st Edition Stefan Czerwik

https://ebookfinal.com/download/functional-equations-and-inequalities-
in-several-variables-1st-edition-stefan-czerwik/

ebookfinal.com

Mathematical analysis An introduction to functions of

several variables 1st Edition Mariano Giaquinta

https://ebookfinal.com/download/mathematical-analysis-an-introduction-
to-functions-of-several-variables-1st-edition-mariano-giaquinta/

ebookfinal.com
Introduction to algebraic geometry 1st Edition Brendan
Hassett

https://ebookfinal.com/download/introduction-to-algebraic-
geometry-1st-edition-brendan-hassett/

ebookfinal.com

Geometry and Groups Carne T.K

https://ebookfinal.com/download/geometry-and-groups-carne-t-k/

ebookfinal.com

An Introduction to Complex Analysis and Geometry 1st

Edition John P. D'Angelo

https://ebookfinal.com/download/an-introduction-to-complex-analysis-
and-geometry-1st-edition-john-p-dangelo/

ebookfinal.com

Group Theory Birdtracks Lie s and Exceptional Groups

Predrag Cvitanovi■

https://ebookfinal.com/download/group-theory-birdtracks-lie-s-and-
exceptional-groups-predrag-cvitanovic/

ebookfinal.com

Abstract Algebra Applications to Galois Theory Algebraic

Geometry and Cryptography 1st Edition Celine Carstensen

https://ebookfinal.com/download/abstract-algebra-applications-to-
galois-theory-algebraic-geometry-and-cryptography-1st-edition-celine-
carstensen/
ebookfinal.com
Several Complex Variables with Connections to Algebraic
Geometry and Lie Groups 1st Edition Joseph L. Taylor
Digital Instant Download
Author(s): Joseph L. Taylor
ISBN(s): 9780821831786, 082183178X
Edition: 1
File Details: PDF, 6.50 MB
Year: 2002
Language: english
 Several Complex Variables
with Connections
to Algebraic Geometry
and 'Lie Groups

Joseph L. Taylor

Grad'uab `Studies
kin. MatMbiAtics
VoIume,4b,

M.ttherrlaticahSgci .e
Several Complex Variables
with Connections
to Algebraic Geometry
and Lie Groups
Several Complex Variables
with Connections
to Algebraic Geometry
and Lie Groups

Joseph L. Taylor

Graduate Studies
in Mathematics
Volume 46

American Mathematical Society

Providence, Rhode Island
 Editorial Board
Steven G. Krantz
David Saltman (Chair)
David Sattinger
Ronald Stern

2000 Mathematics Subject Classification. Primary 34-01, 14-01, 22-01, 43-01.

ABSTRACT. A graduate text with an integrated treatment of several complex variables and complex
algebraic geometry, with applications to the structure theory and representation theory of Lie
groups.

Library of Congress Cataloging-in-Publication Data

Taylor, Joseph L., 1941-
Several complex variables with connections to algebraic geometry and Lie groups / Joseph L.
Taylor.
p. cm. - (Graduate studies in mathematics, ISSN 1065-7339 ; v. 46)
Includes bibliographical references and index.
ISBN 0-8218-3178-X (alk. paper)
1. Functions of several complex variables. 2. Geometry, Algebraic. I. Title. II. Series.

QA331.7.T39 2002
515'.94-dc2l 2002018346

Copying and reprinting. Individual readers of this publication, and nonprofit libraries
acting for them, are permitted to make fair use of the material, such as to copy a chapter for use
in teaching or research. Permission is granted to quote brief passages from this publication in
reviews, provided the customary acknowledgment of the source is given.
Republication, systematic copying, or multiple reproduction of any material in this publication
is permitted only under license from the American Mathematical Society. Requests for such
permission should be addressed to the Acquisitions Department, American Mathematical Society,
P. 0. Box 6248, Providence, Rhode Island 02940-6248. Requests can also be made by e-mail to
[email protected].
© 2002 by the American Mathematical Society. All rights reserved.
The American Mathematical Society retains all rights
except those granted to the United States Government.
Printed in the United States of America.
The paper used in this book is acid-free and falls within the guidelines
established to ensure permanence and durability.
Visit the ARMS home page at URL: http : //www. ams. org/
1098765432 1 070605040302
Contents

Preface xiii

Chapter 1. Selected Problems in One Complex Variable 1

§1.1. Preliminaries 2

§1.2. A Simple Problem 2

§ 1.3. Partitions of Unity 4

§1.4. The Cauchy-Riemann Equations 7

§ 1.5. The Proof of Proposition 1.2.2 10

§1.6. The Mittag-Leffler and Weierstrass Theorems 12

§ 1.7. Conclusions and Comments 16

Exercises 18

Chapter 2. Holomorphic Functions of Several Variables 23

§2.1. Cauchy's Formula and Power Series Expansions 23

§2.2. Hartog's Theorem 26

§2.3. The Cauchy-Riemann Equations 29

§2.4. Convergence Theorems 29

§2.5. Domains of Holomorphy 31

Exercises 35

Chapter 3. Local Rings and Varieties 37

§3.1. Rings of Germs of Holomorphic Functions 38
§3.2. Hilbert's Basis Theorem 39

vii
 Contents

X3.3. The Weierstrass Theorems 40

X3.4. The Local Ring of Holomorphic Functions is Noetherian 44
§3.5. Varieties 45
§3.6. Irreducible Varieties 49
X3.7. Implicit and Inverse Mapping Theorems 50
X3.8. Holomorphic Functions on a Subvariety 55
Exercises 57

Chapter 4. The Nullstellensatz

§4.1. Reduction to the Case of Prime Ideals
X4.2. Survey of Results on Ring and Field Extensions
X4.3. Hilbert's Nullstellensatz
§4.4. Finite Branched Holomorphic Covers
§4.5. The Nullstellensatz
§4.G. Morphisms of Germs of Varieties
Exercises

Chapter 5. Dimension 95
§5.1. Topological Dimension 95
§5.2. Subvarieties of Codimension 1 97
§5.3. Krull Dimension 99
§5.4. Tangential Dimension 100
§5.5. Dimension and Regularity 103
§5.6. Dimension of Algebraic Varieties 104
§5.7. Algebraic vs. Holomorphic Dimension 108
Exercises 110

Chapter 6. Homological Algebra 113

§6.1. Abelian Categories 113
§G.2. Complexes 119
§6.3. Injective and Projective Resolutions 122
§6.4. Higher Derived Functors 126
§6.5. Ext 131
§6.6. The Category of Modules, Tor 133
§G.7. Hilbert's Syzygy Theorem 137
Exercises 142
Contents ix

Chapter 7. Sheaves and Sheaf Cohomology 145

§7.1. Sheaves 145
§7.2. Morphisms of Sheaves 150
§7.3. Operations on Sheaves 152
§7.4. Sheaf Cohomology 157
§7.5. Classes of Acyclic Sheaves 163
§7.G. Ringed Spaces 168
§7.7. De Rham Cohomology 172
§7.8. Cech Cohomology 174
§7.9. Line Bundles and Cech Cohomology 180
Exercises 182

Chapter 8. Coherent Algebraic Sheaves 185

§8.1. Abstract Varieties 186
§8.2. Localization 189
§8.3. Coherent and Quasi-coherent Algebraic Sheaves 194
§8.4. Theorems of Artin-Rees and Krull 197
§8.5. The Vanishing Theorem for Quasi-coherent Sheaves 199
§8.6. Cohomological Characterization of Affine Varieties 200
§8.7. Morphisms -Direct and Inverse Image 204
§8.8. An Open Mapping Theorem 207
Exercises 212

Chapter 9. Coherent Analytic Sheaves 215

§9.1. Coherence in the Analytic Case 215
§9.2. Oka's Theorem 217
§9.3. Ideal Sheaves 221
§9.4. Coherent Sheaves on Varieties 225
§9.5. Morphisms between Coherent Sheaves 226
§9.6. Direct and Inverse Image 229
Exercises 234

Chapter 10. Stein Spaces 237

§10.1. Dolbeault Cohomology 237
§10.2. Chains of Syzygies 243
§10.3. Functional Analysis Preliminaries 245
x Contents

§10.4. Cartan's Factorization Lemma 248

§10.5. Amalgamation of Syzygies 252
§10.6. Stein Spaces 257
Exercises 260

Chapter 11. Frechet Sheaves - Cartan's Theorems 263

§11.1. Topological Vector Spaces 264
§11.2. The Topology of H(X) 266
§11.3. Frechet Sheaves 2 74
§i1.4. Cartan's Theorems 277
§11.5. Applications of Cartan's Theorems 281
§11.6. Invertible Groups and Line Bundles 283
§11.7. Merornorphic Functions 284
§11.8. Holomorphic Functional Calculus 288
X11.9. Localization 298
§11. 10. Coherent Sheaves on Compact Varieties 300
§11.11. Schwartz's Theorem 302
Exercises 309

Chapter 12. Projective Varieties 313

§12.1. Complex Projective Space 313
§ 12.2. Projective Space as an Algebraic and a Holomorphic Variety 314
§ 12.3. The Sheaves O(k) and 7-[(k) 317
§12.4. Applications of the Sheaves O(k) 323
§12.5. Embeddings in Projective Space 325
Exercises 328

Chapter 13. Algebraic vs. Analytic - Serre's Theorems 331

§13.1. Faithfully Flat Ring Extensions 331
§ 13.2. Completion of Local Rings 334
§ 13.3. Local Rings of Algebraic vs. Holomorphic Functions 338
§13.4. The Algebraic to Holomorphic Functor 341
§ 13.5. Serre's Theorems 344
X13.6. Applications 351
Exercises 355
Contents xi

Chapter 14. Lie Groups and Their Representations 357

§14.1. Topological Groups 358
§14.2. Compact Topological Groups 363
§14.3. Lie Groups and Lie Algebras 376
§ 14.4. Lie Algebras 385
§14.5. Structure of Semisimple Lie Algebras 392
§ 14.6. Representations of xC2(C) 400
§14.7. Representations of Semisimple Lie Algebras 404
§14.8. Compact Semisimple Groups 409
Exercises 416

Chapter 15. Algebraic Groups 419

§15.1. Algebraic Groups and Their Representations 419
§15.2. Quotients and Group Actions 423
§ 15.3. Existence of the Quotient 427
§15.4. Jordan Decomposition 430
§15.5. Tori 433
§ 15.6. Solvable Algebraic Groups 437
X15.7. Semisimple Groups and Borel Subgroups 442
§15.8. Complex Semisimple Lie Groups 451
Exercises 456

Chapter 16. The Borel-Weil-Bott Theorem 459

§16.1. Vector Bundles and Induced Representations 460
§16.2. Equivariant Line Bundles on the Flag Variety 464
§16.3. The Casimir Operator 469
§ 16.4. The Borel-Weil Theorem 474
§16.5. The Borel-Weil-Bott Theorem 478
X16.6. Consequences for Real Semisimple Lie Groups 483
§ 16.7. Infinite Dimensional Representations 484
Exercises 493

Bibliography 497

Index 501
Preface

This text evolved from notes I developed for use in a course on several com-
plex variables at the University of Utah. The eclectic nature of the topics
presented in the text reflects the interests and motivation of the graduate
students who tended to enroll for this course. These students were almost all
planning to specialize in either algebraic geometry or representation theory
of semisimple Lie groups. The algebraic geometry students were primarily
interested in several complex variables because of its connections with al-
gebraic geometry, while the group representations students were primarily
interested in applications of complex analysis - both algebraic and analytic
- to group representations.
The course I designed to serve this mix of students involved a simulta-
neous development of basic complex algebraic geometry and basic several
complex variables, which emphasized and capitalized on the similarities in
technique of much of the foundational material in the two subjects. The
course began with an exposition of the algebraic properties of the local
rings of regular and holomorphic functions, first on C'z and then on vari-
eties. This was followed by a development of abstract sheaf theory and sheaf
cohomology and then by the introduction of coherent sheaves in both the
algebraic and analytic settings. The fundamental vanishing theorems for
both kinds of coherent sheaves were proved and then exploited. Typically
the course ended with a proof and applications of Serre's GAGA theorems,
which show the equivalence of the algebraic and analytic theories in the case
of projective varieties. The notes for this course were corrected and refined,
with the help of the students, each time the course was taught. This text is
the result of that process.

xiii
xiv Preface

There were instances where the course continued through the summer
as a reading course for students in group representations. One summer, the
objective was to prove the Borel-Weil-Bott theorem; another time, it was to
explore a complex analysis approach to the study of representations of real
semisimple Lie groups. Material from these summer courses was expanded
and then included in the text as the final three chapters.
The material on several complex variables in the text owes a great debt
to the text of Gunning and Rossi [GR], and the recent rewriting of that text
by Gunning [Gu] It was from Gunning and Rossi that I learned the subject,
.

and the approach to the material that is used in Gunning and Rossi is also
the approach used in this text. This means a thorough treatment of the local
theory using the tools of commutative algebra, an extensive development of
sheaf theory and the theory of coherent analytic sheaves, proofs of the main
vanishing theorem for such sheaves (Cartan's Theorem B) in full general-
ity, and a complete proof of the finite dimensionality of the cohomologies
of coherent sheaves on compact varieties (the Cartan-Serre theorem). This
does not mean that I have included treatments of all the topics covered in
Gunning and Rossi. There is no discussion of pseudo convexity, for example,
or global embeddings, or the proper mapping theorem, or envelopes of holo-
morphy. I have included, however, a more extensive list of applications of
the main results of the subject - particularly if one includes in this category
Serre's GAGA theorems and the material on complex semisimple Lie groups
and the proof of the Borel-Weil-Bott theorem.
Several complex variables is a very rich subject, which can be approached
from a variety of points of view. The serious student of several complex
variables should consult, not only Gunning's rewriting of Gunning and Rossi,
but also the many excellent texts which approach the subject from other
points of view. These include [D], [Fi], [GRe], [GRe2], [Ho], [K], and [N],
to name just a few.
Interwoven with the material on several complex variables in this text is
a simultaneous treatment of basic complex algebraic geometry. This includes
the structure theory of local rings of regular functions and germs of varieties,
dimension theory, the vanishing theorems for coherent and quasi-coherent
algebraic sheaves, structure of regular maps between varieties, and the main
theorems on the cohomology of coherent sheaves on projective spaces.
There are real advantages to this simultaneous development of algebraic
and analytic geometry. Results in the two subjects often have essentially
the same proofs; they both rely heavily on the same background material -
commutative algebra for the local theory and homological algebra and sheaf
theory for the global theory; and often a difficult proof in several complex
variables can be motivated and clarified by an understanding of the often
Preface xv

similar but technically simpler proof of the analogous result in algebraic

geometry.
Several complex variables and complex algebraic geometry are not just
similar; they are equivalent when done in the context of projective varieties.
This is the content of Serre's GAGA theorems. We give complete proofs of
these results in Chapter 13, after first studying the cohomology of coherent
sheaves on projective spaces in Chapter 12.
The text could easily have ended with Chapter 13. This is where the
course typically ends. The material in Chapters 14 through 16 is on quite
a different subject - Lie groups and their representations - albeit one that
involves the extensive use of several complex variables and algebraic geome-
try. Chapter 16 is devoted to a proof of the Borel-Weil-Bott theorem. This
is the theorem which pinpoints the relationship between finite dimensional
holomorphic representations of a complex semisimple Lie group G and the
cohomologies of G-equivariant holomorphic line bundles on a projective va-
riety, called the flag variety, constructed from G. Chapter 15 is a brief
treatment of the subject of complex algebraic groups. This is included in
order to provide proofs of some of the basic structure results for complex
semisimple Lie groups that are needed in the formulation and proof of the
Borel-Weil-Bott theorem. Chapter 14 is a survey of the background mate-
rial needed if one is to understand Chapters 15 and 16. It includes material
on topological groups and their representations, compact groups, Lie groups
and Lie algebras, and finite dimensional representations of semisimple Lie
algebras. These last three chapters are included primarily for the benefit
of the student of Lie theory and group representations. This material illus-
trates that both several complex variables and complex algebraic geometry
are essential tools in the modern study of group representations. The chapter
on algebraic groups (Chapter 15) provides particularly compelling examples
of the utility of algebraic geometry applied in the context of the structure
theory of Lie groups. The proof of the Borel-Weil-Bott theorem in Chapter
16 involves applications of a wide range of material from several complex
variables and algebraic geometry. In particular, it provides nice applications
of the sheaf theory of Chapter 7, the Cartan-Serre theorem from Chapter
11, the material on projective varieties in Chapter 12, Serre's theorems in
Chapter 13, and of course, the background material on algebraic groups and
general Lie theory from Chapters 14 and 15.
I have tried to make the text as self-contained as possible. However,
students who attempt to use it will need some background. This should
include knowledge of the material from typical first year graduate courses
in real and complex analysis, modern algebra, and topology. Also, students
who wishes to confront the material in Chapters 14 through 16 will be
xvi Preface

helped greatly if they have had a basic introduction to Lie theory. Though
the background material in Chapter 14 is reasonably self-contained, it is
intended as a survey, and so some of the more technical proofs have been
left out. For example, the basic theorems relating Lie algebras and Lie
groups are stated without proof, as is the existence of compact real forms
for complex semisimple groups and the classification of finite dimensional
representations of semisimple Lie algebras.
Each chapter ends with an exercise set. Many exercises involve filling in
details of proofs in the text or proving results that are needed elsewhere in
the text, while others supplement the text by exploring examples or addi-
tional material. Cross-references in the text to exercises indicate both the
chapter and the exercise number; that is, Exercise 2.5 refers to Exercise 5
of Chapter 2.
There are many individuals who contributed to the completion of this
text. Edward Dunne, Editor for the AMS book program, noticed an early
version of the course notes on my website and suggested that I consider turn-
ing them into a textbook. Without this suggestion and Ed's further advice
and encouragement, the text would not exist. Several of my colleagues pro-
vided valuable ideas and suggestions. I received encouragement and much
useful advice on issues in several complex variables from Hugo Rossi. Aaron
Bertram, Herb Clemens, Dragan Milicic, Paul Roberts, and Angelo Vistoli
gave me valuable advice on algebraic geometry and commutative algebra,
making up, in part, for my lack of expertise in these areas. Henryk Hecht,
Dragan Milicic, and Peter Trombi provided help on Lie theory and group
representations. Without Dragan's help and advice, the chapters on Lie
theory, algebraic groups, and the Borel-Weil-Bott theorem would not exist.
The proof of the Borel-Weil-Bott theorem presented in Chapter 16 is due
to Dragan, and he was the one who insisted that I approach structure the-
orems for semisimple Lie groups from the point of view of algebraic groups.
The students who took the course the three times it was offered while the
notes were being developed caught many errors and offered many useful
suggestions. One of these students, Laura Smithies, after leaving Utah with
a Ph.D. and taking a position at Kent State, volunteered to proofread the
entire manuscript. I gratefully accepted this offer, and the result was nu-
merous corrections and improvements. My sincere thanks goes out to all
of these individuals and to my wife, Ulla, who showed great patience and
understanding while this seemingly endless project was underway.

Joseph L. Taylor
 Chapter 1

Selected Problems in
One Complex Variable

The study of holomorphic functions of several complex variables involves the

use of powerful tools from many areas of modern mathematics, areas such
as commutative algebra, functional analysis, homological algebra, sheaf the-
ory, and algebraic topology. For this reason, a course in several complex
variables is a great opportunity to teach students how the seemingly sep-
arate fields of pure mathematics can be used in concert to solve difficult
problems and produce striking results. However, this fact also makes the
study of holomorphic functions in several variables much more difficult and
sophisticated than the study of other classes of functions - continuous func-
tions, differentiable functions, holomorphic functions of a single variable -
that students encounter in their early graduate work. When they begin to
realize this, students tend to ask questions such as: Why are we developing
all this machinery? Where is this headed? What is this good for? It is diffi-
cult to answer these questions until much of the language and machinery of
several complex variables has been developed. However, in this introductory
section we will attempt to give some indication of where we are headed and
why we are headed there, by discussing several problems from the theory
of a single complex variable that illustrate some of the issues that will be
central in the several variable theory.
While the main purpose of this chapter is to illustrate and motivate what
is to come, that is not its only purpose. Some of the results developed in
this chapter will be needed later. This is true, for example, of the results on
partitions of unity in section 1.3 and those on the inhomogeneous Cauchy-
Riemann equation in section 1.4.

1
2 1. Selected Problems in One Complex Variable

1.1 Preliminaries
The complex plane will be denoted by C, while complex n-space, the Carte-
sian product of n copies of C, will be denoted by Cn. The open disc of
radius r > 0 centered at a E C will be denoted 0 (a, r), while the closed disc
with this radius and center will be denoted 0 (a, r). If U is an open set in
C and f a complex valued function defined on U, then f is holomorphic if
its complex derivative
f (Z) f (W)
f '(Z) = w-+z
lim z - w

exists for each z E U. We will denote the space of all holomorphic functions
on U by 7-l(U).
We assume that the reader is familiar with the basic properties of holo-
morphic functions of a single variable as presented in standard texts (e.g.
[R]) in the subject: A holomorphic function on U has a convergent power
series expansion in a neighborhood of each point of U; a differentiable func-
tion is holomorphic if and only if it satisfies the Cauchy-Riemann equations;
the space H(U) is an algebra over the complex field under the operations of
pointwise addition, multiplication, and scalar multiplication; if a sequence
{ f7} in x(U) converges uniformly on each compact subset of U, then the
limit function is also holomorphic on U; holomorphic functions satisfy the
Cauchy integral theorem and formula, the identity theorem, and the maxi-
mum modulus theorem.
A function on an open set U in (Cn is holomorphic if it is holomorphic
in each variable separately (in the next chapter we shall prove that this is
equivalent to the existence of local multi-variable power series expansions
of the function). In the n variable case, we shall also denote the space of
holomorphic functions on U by R(U). The space of continuous functions
on a topological space U will be denoted by C(U), the space of n times
continuously differentiable functions on an open set U in a Euclidean space
by Cn (U), and the space of infinitely differentiable functions on U by C°°(U).
As usual, the Euclidean norm of a point z =(Z1, z2i ... , zn) E C" is defined
by Ilzll = (ziI2 + Izzl2 + ... + I

1.2 A Simple Problem

Here we introduce a problem which is easy to state, but not so easy to solve.
Its solution illustrates, in a relatively simple setting, many of the issues we
will have to deal with later in this text.
Let U be an open set in C. Since R (U) is an algebra, it is natural to
try to find all of its maximal ideals. It is easy to see that each point A E U
1.2 A Simple Problem 3

determines a maximal ideal

M,\ = Kerf f --> f (A) : R(U) --> C1

and, in fact, every maximal ideal with C as quotient field has this form.
To see this, observe that if M is a maximal ideal with quotient field C,
0 : H(U) --> C is the quotient homomorphism modulo M, and A = O(z),
then z - A belongs to M. But, for each f E 7-l(U), it follows from the power
series expansion of f at A that f (z) - f (A) is divisible by z - A. That is,
f(z) (A)
z -fA
has a removable singularity at A and so it defines an h E 7-{(U)
such that f (z) - f (A) = (z - A)h(z). Thus, f - f (A) also belongs to M
and 0(f) - f (A) = O(f - f (A)) = 0. This means that 0 agrees with the
evaluation homomorphism f --+ f (A) and M = Ma.
Is every maximal ideal of ?-L(U) of the form Ma? No. Let {an} C U
be a sequence of points which has no limit point in U. Let I be the set of
functions in R (U) which vanish at all but finitely many points of {ate,}. Then
I is a proper ideal of ?-l(U). However, there is no single point at which all
the functions in I vanish. This follows from the fact that, given any discrete
set S of points of U, there is a function in R(U) which vanishes exactly on
S (a corollary of the Weierstrass theorem, which we will prove later in the
chapter). Thus, no maximal ideal that contains I can be of the form Ma.
The following is true, however:
1.2.1 Theorem. Each finitely generated maximal ideal of 7-EC(U) is of the
form MA for some A E U.
If a maximal ideal M of 71(U) is finitely generated, say by {gi,. , 90 1
. .

and if these generators all vanish at some point A E U, then all the functions
in M vanish at A and we have M C M,\. This, of course, implies that
M = MA, since M is a maximal ideal. Thus, Theorem 1.2.1 will be proved
if we can show that: if a finite set of functions {gi,.. , g,z } does not have a
.

common zero, then it does not generate a proper ideal. That is the content
of the following proposition.
1.2.2 Proposition. If a finite set of functions {gi,. .. , gn} C H (U) has no
common zero in U, then the equation
(1.2.1) f1gi++fngn=1
has a solution for li,.., In E 71(U) .

The proof of this result will occupy most of the chapter. Along the way,
we shall prove a number of important results from the theory of holomor-
phic functions of one complex variable that are often not covered in a first
4 1. Selected Problems in One Complex Variable

year graduate course on the subject. Although the ideal theory problem
posed here provides motivation for these results, they have much broader
applicability.

1.3 Partitions of Unity

Proving Proposition 1.2.2 is a typical example of what we will call a local
to global problem. That is, we know that the equation (1.2.1) has local
solutions in the sense that for each w E U there is a neighborhood V of w
and a solution to (1.2.1) consisting of functions f2 holomorphic on V. In
fact, for some j, gj does not vanish at w. Then Vj = Jz E U : gj(z) 0} is a
neighborhood of w on which equation (1.2.1) has a solution, given by setting
fi = g 1 and fi = 0 for i 34 j. Thus, we will have proved Proposition 1.2.2 if
we can show that: if equation (1.2.1) has a solution locally in a neighborhood
of each point of U, then it has a global solution.
We will encounter many of these local to global problems in the course
of our study. Proposition 1.2.2 is a special case of a more general result
concerning a system of linear equations
(1.3.1) GF = H,
where U is an open set in C'z, G is a given p x q matrix with entries from
H(U), H is a given p vector of functions from ?H(U), and a solution F is
sought which is a q vector of functions from 7-((U). Is it true that, if this
system of equations has a solution locally in a neighborhood of each point
of U, then it has a global solution on U? The answer is "yes", provided U is
what is called a domain of holomorphy. To prove this result requires much
of the machinery that we shall develop in this text. Every open set in C is a
domain of holomorphy and so the answer is always "yes" for functions of a
single variable. We won't prove that in this chapter, although we will prove
it in the special case of equation (1.2.1).
While the local to global problem posed by (1.3.1) is quite formidable
for holomorphic functions on an open set U in C'z, the same problem for the
classes of continuous or infinitely differentiable functions is actually quite
easy. This is due to the fact that the classes of continuous and infinitely dif-
ferentiable functions have a strong separation property - Urysohn's lemma.
Urysohn's lemma for continuous functions on a locally compact Hausdorff
space should be familiar to the reader. A similar result holds for C°° func-
tions on Euclidean space.
1.3.1 Lemma. If K C U C W , with K compact and U open, then there
exists f E C°° (Rn) such that 0 < f (x) < 1 for all x, f (x) = 1 for x E K,
and f (x) = 0 for x V U.
1.3 Partitions of Unity 5

Proof. Let a < b be positive numbers. The function A(t) which is exp(-t-1)
for positive t and 0 at all other points on the line is infinitely differentiable
(Exercise 1.1). Thus, the function %(t) = A (t - a) A (b - t) is also in C' (R).
Note that

,%(t) = exp((a - t)-1 + (t - b)-1) for a<t<b

and sli(t) = 0 for all other values of t. We use 0 to define a function
6(fb)1
t) = f
which has the properties that 0 < (t) < 1 for all real t, (t) = 1 for t < a,
and fi(t) = 0 for t > b.
We can now prove the lemma in the case where K is the closed ball
{x E IEgn lix - yII < a} of radius a centered at y and U is the open ball
{x E (fin I Ix - yj I < b} of radius b centered at y. In fact, the function
f (x) _ 0 (11 x - y11) satisfies the conditions of the lemma in this case.
To prove the lemma in general, we cover K with finitely many closed
balls Bi contained in U. Each such ball is contained in an open ball V C U
with the same center. We then choose functions fi, as above, for each pair
Bi C Vi. The function

f(x)=1-fJ(1-f2(x))

then has the required properties.

Given an open cover V of an open set U in II8n, a C°O partition of unity

subordinate to V is a collection {q} C C°°(U) such that: 0 < OZ (x) < 1
for each x E U and each i; Oi (x) = 0 except on a compact subset of some
member of V; on any compact subset K C U, all but finitely many Oi vanish
identically; and
=1

for every x c U.
1.3.2 Lemma. Given an open cover V of an open set U in Ian, there exists
a C°° partition of unity subordinate to V.
Proof. Let {B2} be an enumeration of the set of all open balls in Ian with
rational radii, centered at points with rational coordinates, and with the
property that the closure of the ball is contained in some member of V. Let
6 1. Selected Problems in One Complex Variable

Ai be the ball with the same center as Bi but with half the radius. Clearly,
UZAi = U.
By Lemma 1.3.1, there are functions f2 with 0 < f(x) < 1 for all x,
f(x) = 1 on A2, and f2(x) = 0 on the complement of B2. We define 01 = fi
and
Oj = (1 - fi)(l - f2)...(1 - fi-i)fi for i > 1.

Note that Oi = 0 on the complement of Bi. Also, an induction argument

shows that

1 - (01 + 02 + - - + On) = (I - fl)(1 - f2) ... (1 - fn)

-

and so
01+02+...+0,z=1 on A1UA2U...UAn.

Given a compact subset K of U, there is an n so that K C Al UA2 U . . U An.

It follows that, for i > n, Oi vanishes identically on K. Thus, the infinite
sum > Oi makes sense and is identically 1 on U.

In what follows, C°° (U)p will denote the space of vectors of length p
with entries from C°° (U). We consider C°O (U)p a module over the algebra
C°° (U) through the usual coordinate-wise operations of addition and scalar
multiplication.
1.3.3 Theorem. If U C 1i8n is an open set, G is a p x q matrix with entries
from C°°(U), and H E C°O(U)p, then the equation

GF=H
has a solution F E C°O(U)q, provided it has a solution in C°°(V)q for some
neighborhood V of each point of U.
Proof. We choose an open cover V of U by sets on which the equation has
a solution. We then choose a partition of unity {q5j} subordinate to this
cover. Thus, each Oi vanishes off a compact subset of a set Vi in V and there
is an F2 E C°° (V) q such that GF2 = H on Vi. If we define O FZ to be 0 in
the complement of Vi, then the result is a vector which is in C°° (U)q, as is
the sum F = E OFurthermore,
2

GF = G OjFj = OjGF2 = OZH = H.

2 2 2

Thus, F is a global solution to our equation.

1.4 The Cauchy-Riemann Equations 7

Of course, this argument won't work in the case of systems of equa-

tions involving holomorphic functions, because we do not have anything like
Urysohn's lemma or partitions of unity in the class of holomorphic func-
tions - even in the case of one variable. How do we know this? The identity
theorem says that a holomorphic function on a connected open set in C
is identically 0, if it is 0 on an open subset. In fact, the set on which a
non-constant holomorphic function of one variable on a connected open set
is 0 (or any other constant value) is always a discrete set. Thus, there cer-
tainly are no holomorphic functions with properties like those of the Oi in
the above argument.
So how do we proceed to prove a result like Proposition 1.2.2? We
reduce the problem to the one we just solved by using the inhomogeneous
Cauchy-Riemann equations to relate holomorphic functions to C°° functions.

1.4 The Cauchy-Riemann Equations

In this section we will make use of the notation of differential forms in the
plane. Thus, if f is a C°° function on an open set in the plane, then

df = dx + ay ay.

An expression of the form g dx + h dy is called a differential 1-form. Equation

(1.4.1) defines an operator d from C°° functions to C°O differential 1-forms.
A C°° differential 2-form is an expression of the form f dx A dy with f a
C°° function. The exterior product u A v of two 1-forms, u = ul dx + u2 dy
and v = vl dx + v2 dy, is a 2-form defined by

u n v = (u1v2 - u2v1)dx A dy.

This product is distributive and anti-commutative (u n v = -v n u). Given

a C' 1-form u = g dx +h dy we set

du = (ate - ag I dx A dy.
y/
This defines an operator from 1-forms to 2-forms. Note that d(df) = 0, for
any C°° function f.
The coefficient functions of 1-forms and 2-forms are allowed to be com-
plex valued. In particular there are 1-forms

dz = dx + i dy and d z = dx - i dy.
8 1. Selected Problems in One Complex Variable

For a complex C°° function f , we can write df in terms of dz and dz as

follows:
df = z dz +
09Z dz
where the complex differential operators 8/az and o9/8z are defined by

az 1/2 \ax y
and

8z 1/2(Bx+Z y
Using the above notation, the homogeneous Cauchy-Riemann equations
can be written as the single complex differential equation
a
0.
8z f (z) =
A complex differentiable function f on an open set U is holomorphic on U
if and only if it satisfies this equation.
It is also important to consider the inhomogeneous Cauchy-Riemann
equation
az9(z) = f(z).
We shall show that this has a solution g E C°O(U) for every open set U and
every f E C°O(U). The first step is to prove the generalized Cauchy integral
formula:
1.4.1 Theorem. Let U be an open subset of C bounded by a simple closed
rectifiable curve 'y. If f is a C°° function on a neighborhood of U and z E U,
then
d( af d( A d(
f(z)=
27i ,y
f -z 2,77r a z
U

Proof. Note that, as forms in for fixed z,

d f (() d( a f af (() d( A d(
d(Ad(-- -
(-Z ZJ 19( (-Z
Thus, if ryT is the boundary of the disc 0(z, r) and if r is chosen small enough
that this disc is contained in U, then Stokes' theorem implies that
d( A d( d( d(
if
U,.
o9f
Z fv
f17
Z JVYr
f (() Z
1.4 The Cauchy-Riemann Equations 9

where U, = U - 0 (z, r). Now ((- z) -' is a function of ( on any bounded

region of the plane, as is easily seen by integrating its absolute value using
polar coordinates centered at z. Thus, by the Lebesgue dominated conver-
gence theorem,

of[0f(()d(Ad(
o.
at = Jf as
U
_

Also,
27r
fd(
lim
r C z r o f f (z + reZt) idt = 2i f (z).
o

The result follows.

1.4.2 Proposition. If f E C°° (U), for an open set U C C containing a

compact set K, then there exists an open set V, with K C V C U, and a
g E C°° (V ), such that ag/az = f in V.
Proof. We modify f so that it is actually C°° on all of C, with compact
support in U, by multiplying it by a C°° function which is 1 in a neighborhood
V of K and has compact support in U, and then extending the resulting
function to be 0 on the complement of U. With f so modified, the integral

is defined for all z c C and defines a function g c C' (C). We calculate the
derivative ag/8z of g using the change of variables (-* ( + z:
a a d( A d(
ff
1
9W f (( + Z)_
09z- 2-7ri (9z- (
a f (( + z) d( n d( /' a f (( + z) d( n d(
1

1
ff az
af(() dC A d(
1

if
if
z, JJ aS (-z = J(=).

wh ere the last two identities follow from reversing the change of variables
and then using the generalized Cauchy integral theorem on U (recall that
f vanishes on the complement of a compact subset of U and so the line
integral in Theorem 1.4.1 vanishes). Thus, ag/az = f on all of C. Of
course, we modified f on the complement of V and so this equation holds
for our original f only on V, but this is what was to be shown.
10 1. Selected Problems in One Complex Variable

1.4.3 Theorem. If U is any open set in the plane and f E C°O(U), then
there exists g c C°O(U) such that ag/az = f .

Proof. Let {K} be a sequence of compact sets with the following proper-
ties: (a) Kn is contained in the interior of K,+1 for each n; (b) the union
of the interiors of the Kn's is U; and (c) each bounded component of the
complement of Kn meets the complement of U. This can be done (Exercise
1.2 or [R], 13.3) and, by Runge's theorem, it ensures that each function
holomorphic in a neighborhood of Kn can be uniformly approximated on
Kn by rational functions which have poles only in the complement of U
([R], 13.6).
We will prove by induction that there is a sequence {gn}, with gn a
C°° function satisfying (99n/9z_ = f on a neighborhood of Kn, and with
I - gn_ 1(z) I < 2-n for all z E Kn_ 1 if n > 1. We may choose g1
satisfying these conditions by Proposition 1.4.2. Suppose a sequence {g}
satisfying these conditions has been chosen for n < m. We apply Proposition
1.4.2 to choose a function h which is C°° on a neighborhood of K,.,2+1 and
satisfies o9h/az = f on this neighborhood. On a neighborhood of Km we then
have a (h - gm) /az = 0, and so h - g,n is holomorphic on this neighborhood.
By Runge's theorem, we may choose a rational function r, with poles in the
complement of U, such that jh(z) - gm(z) - r(z) l < 2-'n for z E Km. If
we set gm+1 = h - r, then ag72+1 /az = f on a neighborhood of K,n+1 and
Igm+1(z) - gm(z) l < 2-'n on Km. By induction, a sequence {g} with the
required properties exists.
Clearly, the sequence {gm} converges uniformly on each Kn to a function
g defined on U. Furthermore, gn is holomorphic on a neighborhood
of Kn for each m > n. Thus, for each fixed n, gn} is a sequence of
holomorphic functions on a neighborhood of Kn which is uniformly conver-
gent on Kn. It follows that the limit function g - gn is holomorphic on the
interior of Kn. Hence, g is C°° on the interior of Kn. Since this is true of
each n, g is C°° on all of U. Clearly, ag/az = f.

1.5 The Proof of Proposition 1.2.2.

We think of the n-tuple of functions {g1, g2 , ... , gn } as defining, for each
open set V C U, an 7i (V)-module homomorphism

00 : H M' -- H (V)
by
00(fli f2i ... ,fn) = 91,f1 + 92.f2 + ... + 9n,fn = FGt)
i
1.5 The Proof of Proposition 1.2.2. 11

where G and F are the vectors G =(91, 92, ... , g,) and F = (fl, f2,.. , M.
The hypothesis of Proposition 1.2.2 is that each point of U has a neighbor-
hood V for which this map is surjective. We want to conclude that it is
surjective when V = U. To do this, we need to know something about the
kernel of the map 0o.
We define another 7-l(V)-module homomorphism 01 : H(V)'n --+ 7-l(V)n,
n(n - 1)
where m = 2 Here we represent 7-L(V)' as the space of n x n
skew-symmetric matrices with entries from x(V) and, for such a matrix
A = {a}, define 01(A) = GA, where G is the vector (gi, g , ,gam,). Then
. .
.

00 o 01(A) = GAGt = 0, since A is skew-symmetric and GAGt, being a

scalar function, is symmetric. Thus, 01 has image contained in the kernel
of 00. In fact, locally, 01 maps onto the kernel of 00. To see this, let V be
a neighborhood on which 00 is surjective. We choose F = (fl, f2,. . , fn) .

in 7-l(V)n, so that FGt = Oo(F) = 1 on V, and let H be any vector in the

kernel of 0o. Then A = FtH - HtF is a skew-symmetric n x n matrix, with
the property that
01(A) = GFt H - GHt F = H.
Thus, we have a sequence of 7-1(V)-module homomorphisms
-H V tm > 7-1(V) n > 7-1(V) 0
for which the composition of any two succeeding maps is 0 and which is
locally exact - that is, each point of U has a neighborhood V on which the
kernel of each map in the sequence is the image of the preceding map.
Note that the maps Oo and 01 are defined for tuples of C°° functions as
well and have the same exactness properties. In fact, we have a commutative
diagram

0 0 0

x(U)m 1
x(U)n ° n(u) 0

coo (U) m -- C°° (U) n ° -- C°° (U) ) 0

a a
8z 8z 8x

C°°(U)m
0 C°° (U) n 0> C°° (U) 0

I t t
0 0 0,
12 1. Selected Problems in One Complex Variable

where i is the inclusion of holomorphic vectors into Coo vectors and a/az acts
on C°° vectors by acting on each component. The fact that a/az commutes
with 00 and 01 follows from the fact that these maps are defined by matrices
with holomorphic entries - entries which are killed by a/az (Exercise 1.6).
The rows and columns of the above diagram are complexes (the com-
position of successive maps is 0), the columns are exact (the kernel of each
map is the image of the preceding map) by Theorem 1.4.3, and the second
and third rows are exact by Theorem 1.3.3.
Now the proof of Proposition 1.2.2 is a simple diagram chase. We choose
F E C' (U)' so that q5o (F) = 1. Then Oo (aF/(9z) = a1/az = 0. Since the
bottom row of the diagram is exact, there exists A E C°°(U)72 such that
q1(A) = aF/az. Since the columns are exact, there exists B E C°°(U)m
such that aB/az = A. Then F = F - 01 (B) is holomorphic on U, because

F (F - 01(B)) = -F - 01(A) = 0.
a Z_ 09 Z_ 09 Z_

Furthermore, Op (F) _ Op (F) - Op(01(B)) = 1. This completes the proof of

Proposition 1.2.2.

1.6 The Mittag-Leffler and Weierstrass Theorems

The solvability of the inhomogeneous Cauchy-Riemann equation was the key
to the above proof of Proposition 1.2.2. It has other important applications
as well, among them proofs of the Mittag-Leffler and Weierstrass theorems.
We will conclude this chapter with a development of these results. They
both follow easily from the following proposition, which is a solution to a
single variable version of a problem which was posed in several variables by
Cousin. Thus, we will call it the single variable Cousin problem.
1.6.1 Proposition. Let {U} be an indexed collection of open sets in C and
{gjj} an indexed set of holomorphic functions satisfying:
(i) gzj c H(Ui fl Uj) for each pair (i, j);
(ii) gzj = -gji for each pair (i, j); and
(iii) g2j + gjk + gki = 0 for each triple (i, j, k).
Then there exists an indexed set {h2} of functions, with hi c H(Ui) for each
i and gzj = hj - hi for each pair (i, j).
Proof. Let fopj be a C°O partition of unity subordinate to the cover {U}.
Then, for each p, there is a kp so that Op has compact support in Ukp; on
any given compact subset of U = UiU2, all but finitely many of the Op vanish
identically; and E op = 1. Note that OpgkPi may be defined to be a function
1.6 The Mittag-Leffler and Weierstrass Theorems 13

in C°° (Ui) by defining it to be 0 on Ui - (ui n Uk). For each i, we set

Opgkpi on Ui.
p

Then, on Ui n u,

fi - A = Op (gkpj - 9kpz) = q5pgij = gi3

p p

This gives us a C°° solution to the Cousin problem.

Note that afi/az = 3f2/3 z on Ui n Uj, since fj - fi = gi.7 is holomorphic.
It follows that there is a function u E C°° (U) such that u = 3f/0 z on
Ui for each i. If v c C°°(U) is chosen so that av/az = u, then hi =
fi - v is holomorphic on Ui and hj - hi = The set {h} is the required
gig.

holomorphic solution to the Cousin problem.

An indexed set of holomorphic functions {gjj } satisfying conditions (i),

(ii), (iii) of the above proposition is called a set of Cousin data for the cover
{U2}.
The Mittag-Leffler theorem is a simple application of the above result. It
concerns the existence of meromorphic functions with given principal parts.
A meromorphic function on an open set U C C is a function which is defined
and holomorphic on U, except at a discrete set of points where it has poles.
If A is a point at which a meromorphic function f has a pole of order k, then
(z - A) k f (z) has a removable singularity at A. Thus, f (z) = g(z)(z - A)-k,
where g(z) is holomorphic in a neighborhood of A and, hence, has a power
series expansion which converges in some disc centered at A. In this disc,
we can use the power series expansion of g to write f in the form

f(z) = a-lc(z-A)
-k+...+a_i(z-A)-i+ao+al(z-A) +a2(z-A)2+...

The rational function a_k(z - A) -k + + a_1(z - A)-' is then called the

principal part of f at A.
1.6.2 Mittag-Leffler Theorem. Let U be an open set in C, let S be a
discrete subset of U, and, for each point A E S, let PA be a polynomial in
(z - A)-1 with no constant term. Then there is a meromorphic function f
on U which has a pole with principal part PA at each A E S and which has
no other poles.
Proof. We choose an open cover {U} of U with the property that each Ui
contains at most one point of S. We assign a meromorphic function fi on
14 1. Selected Problems in One Complex Variable

Ui to each index i in the following way: If Ui contains a point wi E S, then

fi = Pwi ; otherwise, f i = 0. We then define Cousin data for the cover {U2}
by setting
gij = f3- fi on UinU3
and noting that gij is holomorphic on Ui n Uj. It follows easily that the
collection {gjj} satisfies (i), (ii), and (iii) of Proposition 1.6.1. Hence, by
that proposition, there exists a collection {h}, with hi holomorphic on Ui
and with h j - hi = gij on Ui n Uj . Then,
f-h=f-h j on UinUj
for each pair i, j, and so these functions fit together to define a meromorphic
function f on U. Because adding a holomorphic function to fi doesn't affect
its poles and principal parts, f has poles exactly at the points of S and, at
each such pole A, the principal part of f is PA.

With a little more work, we can prove the Weierstrass theorem con-
cerning the existence of holomorphic functions with prescribed zeroes and
poles.
1.6.3 Weierstrass Theorem. Given an open set U C C, a discrete set
S C U, and a non-zero integer ka assigned to each point A of S, there exists
a meromorphic function f which has a zero of order ka at each A E S with
kA > 0 and a pole of order -ka at each A E S with ka < 0 and which has no
other zeroes or poles.
Proof. We may choose a sequence {U} of open sets with compact closures
such that Un C Un+1, the union of the Un's is U, and each bounded compo-
nent of the complement of Un contains a point of the complement of U (see
Exercise 1.2 or [R], 13.3). It is easy to see that this can be done in such a
way that no points of S lie on the boundary of any set Un. We set Kn = Un
and Sn = S n Kn = S n Un. We will inductively construct a sequence {fn}
with the following properties:
(1) fn is a rational function on C;
(2) fn has zeroes and poles of the required orders at the points of Sn and
no other zeroes or poles on Kn; and
(3) the zeroes and poles of ffn+i occur in pairs, with each pair consist-
ing of a zero and a pole of the same order occurring at points which
lie in the same component of the complement of Kn.

We set f, (z) _ fj (z - A) kA Then f, is a rational function with the

.

AESI
required zeroes and poles on K1. Assume that f2 with the above properties
1.6 The Mittag-Lefi3er and Weierstrass Theorems 15

have been chosen for i < n. We will then show how to construct fn+1. We
choose a rational function v on C so that v has the required zeroes and poles
at points of Si n+and no other zeroes or poles on Kn+1. Now the function
f'v is rational and has its poles and zeroes in the complement of Kn. Thus,
this function has the form

fn-'(z)v(z) = 11(z - ai)mi,

with each ai in the complement of Kn. Now each ai is in the same component
of the complement of Kn as some point bi in the complement of Kn+1. If
we set
fn+ I (z) = v (z) fl (z - bi) -Mi
i
then f+i has the required zeroes and poles on Kn+1, and the zeroes and
poles of f'f+i occur in pairs (ai, bi) as in (3) above. Thus, by induction,
a sequence with properties (1), (2), and (3) exists.
Now if h is any rational function and 'y is a simple closed curve in C
which doesn't meet a zero or pole of h, then

(27ri)' hh'(()
dS

counts the number of zeroes minus the number of poles (counted according
to multiplicity) in the bounded component of the complement of ry ([R],
10.43). It follows that if V is an open set with the property that the poles
and zeroes of h occur in pairs, each consisting of a zero and a pole with the
same multiplicity, located in the same component of the complement of V,
then any such integral must vanish if ry lies in V. This, in turn, implies that
there is a well-defined holomorphic logarithm of h defined on V - that is, a
function g E 7-l(V) such that h(z) = exp(g(z)) on V. In fact, on any given
component of V, such a function is given by

h(()
9(z) = log h(zo) + 1^1Z h/(() d(

where zo is a fixed point of the component and -yz is any smooth curve in V
joining zo to z. If we apply this in the case where h = f'fm+' and V = Un,
we conclude that there is a function gn, holomorphic on Un, so that

eXp(9n) - .fn lfn+l

on Un.
16 1. Selected Problems in One Complex Variable

We now define Cousin data {gjj} for the cover {U} by requiring that
gig = -gji for all i, j and, for i < j, setting

9ij =9i+gi+1+...+gj-1 on Ui = UinUj.

It follows easily that exp (g2 j) = f1f3 on Ui n Uj for each pair of indices
(i,j). Given a triple of indices i < j < k, we have

9ij +9jk+Ski=gi+...+gi-1+9 +...+gk-1-gi-...-9k-1 =0

on UUnU, nUk = U2.

By Proposition 1.6.1, there exist hi E ?L (Ui) such that gij = hj - hi on
UU n U3. Then

fi exP(-hj) = fz eXP(9zj) exp(-hj) = fi exp(-h2) on Ui fl Uj,

and so there is a holomorphic function f, defined on U, with f = f2 exp (- hi )
on U. Since fi has zeroes and poles of the required orders at the required
points in U2, the same thing is true of f. Thus, f has the required zeroes
and poles of the required orders on all of U.

This is not the most efficient proof of the Weierstrass theorem but it is
a proof which illustrates the methods we have been developing.
Several nice applications of the Mittag-Leffler and Weierstrass theorems
appear in the exercises.

1.7 Conclusions and Comments

We have seen how the local to global problem posed by Proposition 1.2.2 can
be solved by using the solvability of the inhomogeneous Cauchy-Riemann
equation to pass from C°° solutions of a linear equation to holomorphic
solutions. This will be a major theme later in the several variable theory.
There, the local to global problem is formulated as the problem of showing
that certain sheaf cohomology groups vanish. The first theorem of this type,
and the one on which all later ones are based, is Dolbeault's lemma which
asserts the solvability of certain systems of differential equations which are
several variable analogues of the inhomogeneous Cauchy-Riemann equation.
The Cousin problem of section 1.6 makes sense for holomorphic functions
of several variables. In later chapters, we will formulate it as the problem of
showing the vanishing of Cech cohomology groups for sheaves of holomorphic
functions. The Mittag-Leffier and Weierstrass theorems also have several
variable analogues and they are related to Cech cohomology in the same way
1.7 Conclusions and Comments 17

the single variable versions were related to the Cousin problem in section
1.6. The strongest theorem we will prove on the vanishing of cohomology of
sheaves related to holomorphic functions is Cartan's Theorem B. Much of
the machinery developed in the text is aimed at proving this result.
The special nature of the zero set of a holomorphic function played a big
role in the preceding pages, as did detailed knowledge of the local structure
of holomorphic and meromorphic functions. Specifically, a non-trivial holo-
morphic function f has a discrete zero set; in a neighborhood V of each point
A in the domain of f there is a factorization f (z) = (z_A)kg(z), where g is a
,

non-vanishing holomorphic function on V and k is a non-negative integer; a

meromorphic function locally has a similar factorization with k an arbitrary
integer. A detailed knowledge of the zero sets of holomorphic functions and
the local structure of holomorphic functions will be just as essential in the
several variable theory. However, it will not be so easily obtained. Much
of the material in the next four chapters is devoted to these subjects. The
zero sets of holomorphic functions are no longer discrete in several variables.
They are, however, sets with very special structure (holomorphic varieties)
and will be studied in detail. The relationship between holomorphic vari-
eties and ideals in the local algebra of holomorphic functions will be a major
story, culminating in Chapter 4 in the Nullstellensatz, which is one of the
main theorems of the subject.
As we develop the machinery of several complex variables, we will also
give a parallel introductory development of complex algebraic geometry. The
general strategies in the two subjects, as treated in this text, are the same:
Develop detailed knowledge of the local structure and then use techniques
of sheaf cohomology to pass from local information to global information.
We feel it is quite instructive to see the two subjects developed in parallel.
We can point out the way in which the subjects differ as well as ways in
which they are similar. This also lays the foundation for proofs of Serre's
GAGA theorems in Chapter 13. These essentially show that the two subjects
are equivalent if one is working on a projective variety. Serre's proofs use
both the major results of the theory of several complex variables (including
Cartan's Theorem B) and the basics of complex algebraic geometry.
Although the approach we have adopted for this text is based heavily
on commutative algebra, we will also use a substantial amount of analysis.
There are several arguments in Chapter 10 which use properties of Banach
spaces. In Chapter 11 we will need information about more general topo-
logical vector spaces. The approximation argument using Runge's theorem
in the proof of Theorem 1.4.3 will recur in Chapter 10 and in Chapter 11
in a more abstract context, where we will need to use properties of topo-
logical vector spaces (specifically, Frechet spaces). We will use a certain
18 1. Selected Problems in One Complex Variable

amount of Hilbert space theory in Chapter 14. There will be arguments

which involve integrating or differentiating Banach space valued functions.
The definitions and basic properties of these calculus concepts, in the setting
of Banach space valued functions, are not in any substantial way different
from the versions for real or complex valued functions and so we will use
them without much comment. We will, in particular, have occasion to use
holomorphic functions which are Banach space valued in Chapter 11. The
elementary facts we will need concerning such functions in plane regions are
developed in Exercises 1.16 - 1.20. Most of the analysis background that
will be needed in the coming pages is contained in Walter Rudin's first year
graduate text [R], although there will be an occasional reference to more
advanced texts in functional analysis, such as [R2]

Exercises
1. Show that the function which is exp(-t-1) for t > 0 and 0 for t < 0 is
C°O on the real line.
2. Show that if U is an open set in C, then there exists a sequence {K} of
compact subsets of U such that
(i) Kn C int(Kn+1) for each n;
(ii) UnKn = U;
(iii) each bounded component of the complement of Kn contains a point
of the complement of U.
3. Verify that the function

hh(()
9(z) = log h(zo) + L
(() d(,

w hich appears in the proof of Theorem 1.6.3, is indeed a logarithm for h.

That is, verify that expg(z) = h(z) on V.
4. Show that if U is an open subset of C, every finitely generated ideal of
H(U) is a principal ideal.
5. If K is a compact subset of C, let H(K) denote the algebra of functions
holomorphic in a neighborhood of K, where two functions are identified
in R(K) if they agree on some neighborhood of K. Prove that each
maximal ideal of R(K) is of the form Ma = { f E 'H(K) : f (A) = 0} for
some point A E K.
6. Let U be an open set in C. Prove that each C°°(U)-module homomor-
phism 0: C°O(U)' --+ C°°(U)n is given by an n x m matrix with entries
 Exercises 19

in C°O(U). Prove that 0 commutes with 8/az if and only if this matrix
has holomorphic entries.
7. Show that every meromorphic function on an open set U C C is of the
form f(z)/g(z), where f and g are holomorphic functions on U.
8. Let U be an open subset of C. Prove there is a function f E H(U) which
cannot be analytically continued to any larger open set.
9. Use the Mittag-Leffler and Weierstrass theorems to prove the interpola-
tion theorem: If {\} is a sequence of points in the open set U, with no
limit point in U, and if {c} a sequence of complex numbers, then there
is a holomorphic function f E 'H(U) such that f (Ai) = az for each i.
10. Prove the following strong version of the interpolation theorem: Given
a discrete sequence {a2} in an open set U C C, a sequence of positive
integers {n}, and a sequence of functions {f}, where fi is holomorphic
in a neighborhood of ai for each i, show that there exists an f E 7-l(U)
such that f (z) agrees with fi to order nZ (that is, f - fi has a zero of
order at least nZ) for each i.
11. Let K be a compact connected subset of C and consider the ring 7-l(K),
defined in Exercise 1.5. Show that this ring is a principle ideal domain
(an integral domain such that every ideal is a principle ideal).
12. Let U be an open subset of C such that C - U has only finitely many
connected components. Let {)} be a set consisting of one point from
each of these components. Prove that each non-vanishing function f in
7-l(U) has the form

f (z) = rl(z - Aj )nj exp g (z),

7

for some g c 7-l(U).

13. Let g(z) be a meromorphic function on C which is holomorphic except at
a discrete set of points f Aj I C C, where it has poles of order one. Suppose
the residue of f at Aj is an integer nj for each j. Let U = cC - {)}. Prove

()dois
that if ryz is a smooth simple curve in U joining a fixed point zo E U to
a variable point z E U, then the function

f (z) = exp (17Z g(

a well-defined holomorphic function on U and has a zero of order nj

at Aj for each j. Show that the Weierstrass theorem for the domain C
follows from this result and the Mittag-Leffler theorem.
14. Give a simpler proof of the existence of solutions of the inhomogeneous
Cauchy-Riemann equation (Theorem 1.4.3) in the case where U = C, one
that does not involve Runge's theorem.
 20 1. Selected Problems in One Complex Variable

15. Here is another type of inhomogeneous Cauchy-Riemann problem: Let U

be an open set in C and f a function which is C°O on U and has compact
support in U. By Theorem 1.4.3, f = a9 for some C°O function on U,
az
but g may not have compact support. Prove that a solution g exists with
compact support in U if and only if I f(z)h(z)dz n dz = 0 for every
U
h E 7-l(U).
16. Let X be a complex Banach space (see [R]) and U an open subset of
C. A function f : U -> X is said to be holomorphic in U if its complex
derivative
f (Z) -- lim f (Z)z - wf (W)
/
W--+Z

exists at each point of U, in the sense that the indicated limit exists in
the norm of X. Contour integrals of continuous X-valued functions in
plane regions are defined in the same way as for complex valued functions,
with convergence of the integral taking place in the norm of X. Prove the
Cauchy theorem for Banach space valued functions; that is, prove that
if f : U --> X is holomorphic, r is a piecewise continuously differentiable
simple closed curve in U, and the bounded component of the complement
of r is contained in U, then

[f(d(=0 and f(z)= 27ri f f (0 d(

for each z in the bounded component of the complement of F.

17. If X is a Banach space and f is an X-valued holomorphic function in
the open disc {z E C : IzI < R}, then prove that f has a power series
expansion

f(z)=xo+xiz+--z 2
where
f (k)
f (Z)
1 (0)
xk
-- 27ri f zk+i dz =
k!

Prove this power series converges uniformly in the norm of X in every

disc {z E C : IzI < r} with r < R. Hint: Use the result of the previous
exercise.
18. Prove that if X is a Banach space, f is a holomorphic X-valued function
in a neighborhood of the closed disc {z E C : Izi < R}, and if I< M
on this disc, then
f (k) (0) MM

fork=1,2,...
 Exercises 21

19. Let X be a Banach space and let f be an X-valued holomorphic function

in some neighborhood of 0 in C. Prove that if R is the largest positive
number r such that f can be extended to be holomorphic on the disc
f z c C : jzj < r}, then HM SUP IIxk I 1l1k = R-1, where the elements xk are
the power series coefficients of f .
20. Prove Liouville's theorem for Banach space valued functions: If X is a
Banach space and f is a holomorphic X-valued function on C such that
If (z) I is bounded, then f is constant.
I
 Chapter 2

Holomorphic
Functions of Several
Variables

There are a number of possible ways to define what it means for a complex
valued function on an open set in C' to be holomorphic. One could simply
insist that the function be holomorphic in each variable separately. Or one
could insist the function be continuous (as a function of several variables)
in addition to being holomorphic in each variable separately. The a priori
strongest condition would be to insist that a holomorphic function have a
convergent expansion as a multi-variable power series in a neighborhood
of each point of the domain. The main object of this chapter is to show
that these possible definitions are all equivalent. We will also develop some
of the elementary properties of holomorphic functions of several variables
and introduce a concept that will play a central role in later chapters -
the concept of domain of holomorphy. All open sets in C are domains of
holomorphy. The fact that this is not true in several variables is one of the
things that makes the several variable theory strikingly different from the
single variable theory.

2.1 Cauchy's Formula and Power Series Expansions

For holomorphic functions of one complex variable, the existence of local
power series expansions is usually deduced from Cauchy's integral formula.
The same technique works in several variables.

23
Random documents with unrelated
content Scribd suggests to you:
9. Sklaven mit einem toten Löwen, Relief, assyrisch. 7.
Jahrhundert vor Chr.
10. Pagode, Südindien, wahrscheinlich 11. bis 12.
Jahrhundert
11. Kutub-Minar bei Delhi. 11. bis 14.
Jahrhundert
12. Palast der Winde, Indien, Jaipur. Anfang des 18. Jahrhunderts
13. Marmorfenster der Sidi-Said-Moschee, Indien, Hindostan
14. Nebenhalle eines Felsentempels, China,
Itschang. Nach einer Aufnahme von
Regierungsbaumeister Ernst Boerschmann
15. First eines Tempels, China, Kanton. Nach einer Aufnahme von
Regierungsbaumeister Ernst Boerschmann
16. Paukentürmchen in einem Familientempel,
China, Tschangschafu. Nach einer Aufnahme von
Regierungsbaumeister Ernst Boerschmann
17. Wu-I-Hsien, Regensturm, China, um 1400
18. Unkoku Togan, Landschaft (Teil eines
Setzschirms), Japan. 16. Jahrhundert
19. Holzstatue eines Priesters, Korea
20. Kopf einer Jünglingsstatue,
griechisch. Erste Hälfte des 6.
Jahrhunderts vor Chr.
21. Theseus. Detail einer
Gruppe „Theseus und
Antiope“, griechisch. Zweite
Hälfte des 6. Jahrhunderts vor
Chr.
22. Figuren aus dem Ostgiebel des Parthenon, griechisch. 5. Jahrhundert
vor Chr.

23. Sarkophagfiguren, etruskisch. 6. Jahrhundert vor Chr.

24. Die kapitolinische Wölfin, etruskisch. 5. Jahrhundert vor Chr. (Die
Kinderfiguren sind aus der Zeit der Renaissance)
25. Sarkophagfigur, etruskisch. 5. Jahrhundert vor Chr. (?)
26. Hades, etruskische Wandmalerei. Grotte dell’
Orco bei Corneto. 4. Jahrhundert vor Chr. Aus „Der
Schöne Mensch im Altertum“ von Heinr. Bulle (G.
Hirths Verlag)
27. Die Basilika des Konstantin, Rom. 4. Jahrhundert
28. Altes Gemäuer, Rom. 3. bis 4. Jahrhundert
29. S. Stefano Rotondo, Rom. 5. Jahrhundert
Welcome to our website – the ideal destination for book lovers and
knowledge seekers. With a mission to inspire endlessly, we offer a
vast collection of books, ranging from classic literary works to
specialized publications, self-development books, and children's
literature. Each book is a new journey of discovery, expanding
knowledge and enriching the soul of the reade

Our website is not just a platform for buying books, but a bridge
connecting readers to the timeless values of culture and wisdom. With
an elegant, user-friendly interface and an intelligent search system,
we are committed to providing a quick and convenient shopping
experience. Additionally, our special promotions and home delivery
services ensure that you save time and fully enjoy the joy of reading.

Let us accompany you on the journey of exploring knowledge and

personal growth!

ebookfinal.com