Textbook Introduction To Complex Theory of Differential Equations Savin Ebook All Chapter PDF

Introduction to Complex Theory of
Differential Equations Savin

Visit to download the full and correct content document:
https://textbookfull.com/product/introduction-to-complex-theory-of-differential-equation
s-savin/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...
Introduction to Partial Differential Equations Peter J.

Olver
https://textbookfull.com/product/introduction-to-partial-
differential-equations-peter-j-olver/
Distribution Theory Applied to Differential Equations

1st Edition Adina Chirila
https://textbookfull.com/product/distribution-theory-applied-to-
differential-equations-1st-edition-adina-chirila/
A Modern Introduction to Differential Equations 3rd

Edition Henry J. Ricardo
https://textbookfull.com/product/a-modern-introduction-to-
differential-equations-3rd-edition-henry-j-ricardo/
Introduction to the Theory of Complex Systems Stefan

Thurner
https://textbookfull.com/product/introduction-to-the-theory-of-
complex-systems-stefan-thurner/
Ordinary differential equations : an introduction to
the fundamentals Second Edition. Edition Howell
https://textbookfull.com/product/ordinary-differential-equations-
an-introduction-to-the-fundamentals-second-edition-edition-
howell/
Numerical Solution of Differential Equations:

Introduction to Finite Difference and Finite Element
Methods 1st Edition Zhilin Li
https://textbookfull.com/product/numerical-solution-of-
differential-equations-introduction-to-finite-difference-and-
finite-element-methods-1st-edition-zhilin-li/
Theory of Differential Equations in Engineering and

Mechanics 1st Edition Kam Tim Chau
https://textbookfull.com/product/theory-of-differential-
equations-in-engineering-and-mechanics-1st-edition-kam-tim-chau/
Modelling with Ordinary Differential Equations Dreyer
https://textbookfull.com/product/modelling-with-ordinary-
differential-equations-dreyer/
Elementary differential equations Second Edition

Roberts
https://textbookfull.com/product/elementary-differential-
equations-second-edition-roberts/
Anton Savin • Boris Sternin
Introduction to Complex Theory

of Differential Equations
Anton Savin Boris Sternin
RUDN University RUDN University
Department of Applied Mathematics Department of Applied Mathematics
Moscow, Russia Moscow, Russia
and and
Leibniz Universität Hannover Leibniz Universität Hannover

Institut für Analysis Institut für Analysis
Hannover, Germany Hannover, Germany
ISSN 1660-8046 ISSN 1660-8054 (electronic)

Frontiers in Mathematics
ISBN 978-3-319-51743-8 ISBN 978-3-319-51744-5 (eBook)
DOI 10.1007/978-3-319-51744-5
Library of Congress Control Number: 2017934913
Mathematics Subject Classification (2010): 58-02, 58J32, 58Z05, 58J47, 35Gxx, 32W50, 86-XX, 78-XX
© Springer International Publishing AG 2017

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the
material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation,
broadcasting, reproduction on microfilms or in any other physical way, and transmission or information
storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now
known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this book are
believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors
give a warranty, express or implied, with respect to the material contained herein or for any errors or
omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in
published maps and institutional affiliations.
Printed on acid-free paper
This book is published under the trade name Birkhäuser, www.birkhauser-science.com

The registered company is Springer International Publishing AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Preface
The present book is devoted to the complex theory of differential equations or, more
precisely, to the theory of differential equations on complex-analytic manifolds.
The theory of differential equations on real manifolds is very well known. At the
same time, the complex theory of partial differential equations is still somewhat outside
the focus of interest of specialists on differential equations, which, as we believe, is com-
pletely unjust. Indeed, apart from the remarkable beauty of this theory, it can be used to
solve problems in mathematics and physics in purely real situations. An example of this
is given by the solution to the famous Poincaré problem on balayage inwards, and also
the study of an absolutely nontrivial “mother body” problem arising in geoprospecting, as
well as a number of other problems. It seems that these ideas were understood by the out-
standing French mathematician Jean Leray, who together with his coauthors and pupils
(L. Gårding, T. Kotake, and others) attempted to construct the theory of differential equa-
tions on complex manifolds. Unfortunately, their theory did not become as widely known,
as it definitely deserved, which could be explained by the nontraditionality of the subject
from the standpoint of the classical theory of partial differential equations and also by
the fact that Jean Leray did not have time to solve a number of natural problems of the
theory. To make the latter statement more precise, note that until recently in complex the-
ory there was no satisfactory analogue of a transform similar to the Fourier transform in
real theory, which enables solving equations with constant coefficients exactly. Moreover,
one of the most important problems in complex theory, namely, the problem of describ-
ing the asymptotics of solutions of Cauchy problems, was solved by Leray only in the
small, while applications required on studying singularities of the solution in the large,
i.e., if we speak about Cauchy problems, far from the original manifold. This and other
problems — e.g., the problem of antenna size optimization — were recently solved using
complex theory. In other words, it turned out that complex theory makes it possible to
solve successfully purely real problems, for instance, the balayage problem mentioned
above.
It is also worth mentioning that monographs dealing with complex theory are quite
difficult to understand and to some extent this probably hindered wider applications of
complex theory to many important problems in physics and engineering.
The authors of the present book tried to write it so that it could be read by and be
interesting to a wide range of mathematicians who may not be familiar with complicated
and advanced notions in complex analysis.
Here we will share a few words about the contents of this book. The first three
chapters are devoted to auxiliary material, which will be used in subsequent chapters: we
describe Leray residues, and ramified integrals and their asymptotics. Chapters 4-7 are
the central part of the book: here a new integral transform is introduced that allows us
to obtain not only an explicit formula for the exact solution of the Cauchy problem for
equations with constant coefficients, but also study singularities of these solutions. More
precisely, in Chapters 4 and 5 the transform in question is introduced and its properties
are discussed. Note that this transform was introduced by Sternin and Shatalov in [45,46].
In Chapters 6 and 7 we use the transform to obtain an explicit formula for the solution
v
vi Preface
of the Cauchy problem and describe the singularities of the solution. Note that solutions
of complex Cauchy problems always have singularities (possibly located at infinity), and
therefore the function classes, in which the Cauchy problem is considered, are the classes
of ramified analytic functions. Chapter 8 is devoted to studying Cauchy problems for
equations with variable coefficients using Leray’s uniformization method. The results ob-
tained here are valid in the small, even though presently there is an apparatus that permits
us to construct asymptotics in the large, i.e., far from the original manifold. Unfortunately,
this apparatus is far more complicated technically and its exposition is beyond the scope
of this book. The final Chapters 9 and 10 are devoted to applications: to the solution of
Poincaré’s balayage problem mentioned above as well as to an effective construction of
“mother bodies.”
At the end of each chapter we give bibliographic remarks on the references related
to that chapter. We should mention that during the preparation of this book we widely
used classical works by J. Leray and also the works by B.Yu. Sternin and V.E. Shatalov
and their coauthors on complex theory of differential equations. These are briefly the
contents of the book.
Acknowledgments. The results discussed in this book were delivered at a number of
scientific seminars [seminars of Prof. A.S. Mishchenko and others (MSU), Acad. A.T.
Fomenko (MSU), Prof. E. Schrohe (Leibniz University of Hannover, Germany)]; at inter-
national conferences in Bialowieza (Poland), Voronezh, Saint-Petersburg, Tambov (Rus-
sia) and were taught many times at various scientific centers: Independent University of
Moscow, RUDN University, etc. We are grateful to all the participants of the seminars
and lectures for attention and constructive criticism during the talks. The authors are also
grateful to Vladimir Nazaikinskii and Pavel Sipailo.
During the preparation of this book, the authors were supported in part by the Si-
mons foundation and the Ministry of Education and Science of the Russian Federation
(agreement no. 02.a03.21.0008).
Moscow–Hannover Anton Savin

2016 Boris Sternin
Contents
Preface v
1 Leray residues 1
1.1 A glimpse of one-dimensional residues . . . . . . . . . . . . . . . . . . . 1
1.2 First definition of Leray residues . . . . . . . . . . . . . . . . . . . . . . 2
1.3 Second definition of Leray residues . . . . . . . . . . . . . . . . . . . . 3
1.4 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.5 Leray exact sequences and residue theorem . . . . . . . . . . . . . . . . 5
1.6 Appendix. Some notions of (co)homology theory . . . . . . . . . . . . . 7
1.7 Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2 Ramified integrals 11
2.1 Why do integrals ramify? . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.2 General theory. Landau manifolds . . . . . . . . . . . . . . . . . . . . . 14
2.3 Integrals over relative cycles . . . . . . . . . . . . . . . . . . . . . . . . 19
2.4 Appendix. Differentiation of parametric integrals . . . . . . . . . . . . . 24
2.5 Appendix. Stratified sets . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.6 Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
3 Asymptotics of ramified integrals 31

3.1 Statement of the problem . . . . . . . . . . . . . . . . . . . . . . . . . . 31
3.2 Ramification of cycles around Landau manifolds (Picard–Lefschetz
theorem) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
3.3 Leray’s theorem on asymptotics of integrals . . . . . . . . . . . . . . . . 37
3.4 Computation of Leray’s asymptotics . . . . . . . . . . . . . . . . . . . . 39
3.5 Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
4 Ramified Fourier transform 43

4.1 Definition of the ramified Fourier transform . . . . . . . . . . . . . . . . 44
4.2 Construction of ramified homology classes . . . . . . . . . . . . . . . . . 45
4.3 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
4.4 Appendix. Projective spaces . . . . . . . . . . . . . . . . . . . . . . . . 50
4.5 Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
vii
viii Contents
5 Properties of the ramified Fourier transform 53

5.1 Action of the transform in function spaces . . . . . . . . . . . . . . . . . 53
5.2 Inverse transform . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
5.3 Commutation relations . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
5.4 Singularities of the transform . . . . . . . . . . . . . . . . . . . . . . . . 56
5.5 Properties of the ramified Fourier transform of nonhomogeneous functions 58
5.6 Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
6 The Cauchy problem for equations with constant coefficients 61

6.2 Solution of the Cauchy problem . . . . . . . . . . . . . . . . . . . . . . 63
6.3 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
6.4 A formula for the solution of the Cauchy problem . . . . . . . . . . . . . 67
6.5 Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
7 Singularities of the solution of the Cauchy problem 69

7.1 Preliminary description of singularities . . . . . . . . . . . . . . . . . . . 69
7.2 Geometric description of singularities . . . . . . . . . . . . . . . . . . . 70
7.3 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
7.4 Singularities of the solution if X is singular . . . . . . . . . . . . . . . . 74
7.5 Singularities of the solution if char H is singular . . . . . . . . . . . . . . 75
7.6 Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
7.7 Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
8 The Cauchy problem for equations with variable coefficients.

Leray’s uniformization 81
8.1 Uniformization theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
8.2 Propagation of singularities . . . . . . . . . . . . . . . . . . . . . . . . . 86
8.3 Leray’s asymptotics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
8.4 Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
8.5 Asymptotics with respect to independent variables . . . . . . . . . . . . . 89
8.6 Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
9 Balayage inwards problem 95

9.2 Reduction to a complex Cauchy problem . . . . . . . . . . . . . . . . . . 97
9.3 Solution of the balayage problem . . . . . . . . . . . . . . . . . . . . . . 99
9.4 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
9.5 Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
10 Mother body problem 109

10.2 Singularities of the continuation of potentials and Schwarz functions . . . 113
10.3 How to make cuts? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
Contents ix
10.4 Construction of mother body (algorithm) . . . . . . . . . . . . . . . . . . 121

10.5 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
10.6 Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128
Hints for exercises 129
Bibliography 131
Index 137
Chapter 1
Leray residues
Residues of differential forms depending on several variables were first considered by

Poincaré (1887). Later, such residues appeared in the works of Picard (1901), de Rham
(1932—1936), and others. In this chapter we explain the theory of residues constructed
by Leray (1959).
1.1 A glimpse of one-dimensional residues

Let f (x) be a function of a complex variable x ∈ C that is holomorphic in a punctured
neighborhood of x0 ∈ C. Then the residue of the form f (x)dx at x0 denoted by
res f (x)dx ∈ C
x=x0
is defined in one of two equivalent ways:

1) the residue is expressed as the integral

1
res f (x)dx = f (x)dx (1.1)
x=x0 2πi γ
(here γ is a sufficiently small contour going around x0 one turn counterclockwise);

2) f (x) is expanded in the Laurent series
∞
f (x) = ∑ an (x − x0 )n ,
n=−∞
and we set
res f (x)dx = a−1 . (1.2)
x=x0
These definitions of the residue of forms are invariant (in contrast to the definition
of residues of functions). More precisely, a direct computation shows that
© Springer International Publishing AG 2017 1

A. Savin, B. Sternin, Introduction to Complex Theory of Differential Equations,
Frontiers in Mathematics, DOI 10.1007/978-3-319-51744-5_1
2 Chapter 1. Leray residues
both expressions (1.1) and (1.2) are invariant with respect to changes of variable!
Note also that the form f (x)dx is closed by Cauchy–Riemann equations.
Let us now translate (1.2) into the language of differential forms. To this end, we
substitute the Laurent expansion of f (x) in f (x)dx and represent this form as the sum
−2 ∞
an a−1 dx
f (x)dx = d ∑ (x − x0 ) n+1
+ + ∑ an (x − x0 )n dx (1.3)
n=−∞ n + 1 x − x0 n=0
of an exact form, a form with a first-order pole, and a holomorphic form. We can now
give the second definition of residue. Namely, it follows from (1.3) that the cohomology
class
[ f (x)dx] ∈ H 1 (C \ {x0 })
has the representative
dx
ϕ(x) + ψ(x)dx,
x − x0
where ϕ(x) and ψ(x) are regular functions at x = x0 . Then
res f (x)dx = ϕ(x0 ). (1.4)

x=x0
Exercise 1.1. Show that (1.4) is invariant with respect to changes of variable.
Let us now turn to the multidimensional theory of residues.
1.2 First definition of Leray residues

Let M be a complex manifold and X ⊂ M be a complex submanifold of codimension one,
i.e., X is locally defined by an equation
X = {x ∈ X | s(x) = 0},
where ds = 0 on X. The function s(x) is called a (local) defining function of X. Under these
assumptions, a small tubular neighborhood1 of X can be represented as the total space of
a fibration with fiber equal to a small disc in C, hence, slightly simplifying the situation,
Leray residues can be thought of as residues in the direction normal to X, while we do
not take residues along the tangent directions. We now proceed to the precise definitions.
Let ω be a closed differential form of degree k that is smooth in a punctured neigh-
borhood of X, and consider a small tubular neighborhood of X. Denote the boundary of
this tubular neighborhood by SX (see Fig. 1.1). This boundary fibers over X with the fiber
S1 . Denote the projection for this bundle by
π : SX −→ X.
1 Recall
that a tubular neighborhood of radius r of a submanifold is the set of points of the ambient manifold,
whose distance to the submanifold is less then r.
1.3. Second definition of Leray residues 3
Figure 1.1: A tubular neighborhood of a submanifold.
Definition 1.1. The Leray residue of the cohomology class [ω] ∈ H k (M \ X) of the form
ω is the cohomology class

1
Res[ω] = ω SX ∈ H k−1 (X), (1.5)
X 2πi S1
where
: Λk (SX) → Λk−1 (X)
S1
stands for the operator of integration along the fibers of the fibration SX → X (see the
Appendix to this chapter) oriented using the complex structure. Here and in what follows
Λk denotes the space of smooth k-forms on a manifold.
Clearly, in the one-dimensional case Definition 1.1 reduces to the classical definition
of residues using contour integration (1.1).
Let us also mention that the class (1.5) is well defined since integration along the
fiber (anti)commutes with the exterior differential.
1.3 Second definition of Leray residues

From now on we shall work with forms holomorphic in the sense of the following defini-
tion.
Definition 1.2. A form ω on a complex manifold with coordinates x1 , x2 , . . . , xn is holo-
morphic, if
ω = ∑ ωI (x)dxI , where dxI = dxi1 ∧ dxi2 ∧ · · · ∧ dxik , (1.6)

I
I = (i1 , . . . , in ) is a multi-index, and the functions ωI (x) are holomorphic. Note that such
forms have no antiholomorphic differentials dx j . If the coefficients in (1.6) have singular-
ities, then we say that the form is analytic.
Suppose that X ⊂ M is a submanifold with local defining function s(x). As in the

one-dimensional case, the definition of residue is based on decompositions of closed
forms as sums of exact forms, forms with first-order poles, and holomorphic forms. The
following proposition will enable us to formulate the second definition of Leray residues.
Proposition 1.1. Given a closed form ω in a punctured neighborhood of X, its cohomol-
ogy class [ω] ∈ H k (M \ X) has a representative equal to
ds
∧ ψreg + θreg (1.7)
s
(such forms are called forms with first-order poles), where ψreg and θreg are holomorphic
forms in a neighborhood of X and
1) the restriction ψreg |X is a closed form independent of the choice of s(x) and the
decomposition (1.7);
2) if the sum (1.7) is holomorphic outside X, then ψreg and θreg are holomorphic.
Definition 1.3. The Leray residue of a class [ω] ∈ H k (M \ X) is the class

Res[ω] = ψreg |X ∈ H k−1 (X), (1.8)
X
where ψreg is taken from a representative of this class in terms of a form with a first-order
pole; see (1.7).
Remark 1.1. For forms with first-order poles (1.7), Definition 1.3 gives an explicit for-
mula for Leray residues. However, the computation of residues for forms with more com-
plicated singularities can be quite cumbersome. Nonetheless, for forms with poles2 of
order k ≥ 2 the residue can be calculated using the formula (see Pham [36])
ds θ ψ 1 dψ
ω= ψ + k−1 = d − + k−1 +θ
sk s (k − 1)sk−1 s k−1
and induction over k. Here the first summand on the right-hand side is exact, while the
second summand has a pole of order k − 1.
Exercise 1.2. Prove that Definitions 1.3 and 1.1 are equivalent using the Stokes formula.
1.4 Examples
1.
x1 dx1 ∧ dx2 ∧ dx3 x1 d(x1 + x2 x3 ) ∧ dx2 ∧ dx3
Res = Res
x1 +x2 x3 =0 x1 + x2 x3 x1 +x2 x3 =0 x1 + x2 x3
= −x2 x3 dx2 ∧ dx3 .
2 A form ω has a pole of order k on a submanifold {s(x) = 0}, if sk ω is holomorphic in a neighborhood of
this submanifold.
1.5. Leray exact sequences and residue theorem 5
2.
x1 x2 dx1 ∧ dx2 (ζ − x2 )x2 dζ ∧ dx2

Res = Res
x1 +x2 =0 (x + x ) ζ2
1 2 2 ζ =0
dζ 2 (x2 )2 dζ ∧ dx2
= Res x ∧ dx2 − Res
ζ =0 ζ ζ =0 ζ2

(x2 )2 dx2
= x2 dx2 − Res d − = x2 dx2 .
ζ =0 ζ
Here we set ζ = x1 + x2 .
3. (Poincaré 1887). Given any 1 ≤ j ≤ n, one has
j
f (x)dx1 ∧ · · · ∧ dxn dx1 ∧ · · · ∧ dx ∧ · · · ∧ dxn
Res = (−1) j+1 f (x)
s(x)=0 s(x) ∂s
∂xj
provided that x ∈ M is such that at this point ∂ s/∂ x j = 0.
1.5 Leray exact sequences and residue theorem

From the standpoint of (co)homology theory, it is natural to consider the Leray residue as
one of the maps in the Leray exact sequence. Let us describe this sequence.
Leray sequence in cohomology. Consider the triangle
H ∗ (M) (1.9)
d
ω∗
y i∗
H ∗ (M \ X) / H ∗ (X)
δ ∗ =(2πi) Res
Here i∗ is induced by the embedding M \ X → M, while ω ∗ increases degrees of forms

by two and is defined using product with the Thom class of the normal bundle (see the
Appendix to this chapter):
1 ds
ω ∗ϕ = − d χ ∧ π ∗ ϕ,
2πi s
where π : U → X stands for the projection of the tubular neighborhood on X, while χ is
a cut-off function equal to one on X and zero outside U. Finally, δ ∗ decreases degrees of
forms by one and takes a class to its Leray residue on X.
One can show that the triangle (1.9) is commutative.
Leray sequence in homology. By duality (see the Appendix to this chapter) we also
have a commutative triangle of homology groups (with compact supports)
H∗ (M) (1.10)
9
i∗
ω
δ
$
H∗ (M \ X) o H∗ (X)
Here i∗ is induced by the embedding M \ X → M (see the Appendix to this chapter),

the mapping ω decreases the dimension of cycles by two and takes a cycle on M to its
intersection with X, while the mapping (called the Leray coboundary)
δ : H∗ (X) −→ H∗+1 (M \ X)
increases the dimension of cycles by one and takes a cycle on X to the pull-back of this
cycle on the circle bundle SX ⊂ M.
Let us give examples, where the Leray coboundary can be computed explicitly.
Example 1.1. Consider the embedding {0} ⊂ C of the origin. Then δ ({0}) is the homol-
ogy class of a circle around the origin.
Example 1.2. Consider the quadric

X = (x1 )2 + (x2 )2 = 1 ⊂ C2 .
Its real part X ∩ R2 is the circle defined by the same equation. The homology class of this
circle is denoted by
[X ∩ R2 ] ∈ H1 (X).
It is easy to see that its Leray coboundary
δ [X ∩ R2 ] ∈ H2 (C2 \ X)
is the homology class of the torus

x1 (1 + z), x2 (1 + z) ∈ C2 \ X |z| = ε, (x1 )2 + (x2 )2 = 1, x1 , x2 ∈ R ⊂ C2 .
Theorem 1.1 (Residue theorem). Given a cycle h on X and a closed form ω on M \ X,

we have
ω = 2πi Res ω. (1.11)
δh h X
Exercise 1.3. Prove residue formula (1.11).

1.6. Appendix. Some notions of (co)homology theory 7
Residues in relative theory. Suppose that we are also given a codimension one sub-
manifold S ⊂ M transversal to X. In this situation, we consider closed differential forms
on M that have no singularities on the complement of X, while their restriction to S is
zero. Recall that such forms define classes in relative de Rham cohomology H ∗ (M \ X, S).
Hereinafter, when we write a pair of spaces (A, B), we understand this pair as (A, A ∩ B).
In this relative theory the residue is a mapping
Res : H ∗ (M \ X, S) −→ H ∗−1 (X, S).

X
Its construction is similar to the above and we do not stop on it.
1.6 Appendix. Some notions of (co)homology theory

In this Appendix we recall some notions and results of (co)homology theory, which are
used in this book. All of them are contained in standard textbooks (e.g., see [5]). We
collected them here for the convenience of the reader.
Let M be a compact closed smooth manifold.
1. Relative cohomology. The de Rham cohomology of M is denoted by H k (M). It is

generated by closed differential forms of degree k on M modulo exact forms. If the man-
ifold is noncompact, then one considers cohomology with compact supports, denoted by
Hck (M), which is generated by compactly supported forms.
Further, given a submanifold X ⊂ M, one defines the relative cohomology H k (M, X)
generated by forms that have zero restriction to X. Moreover, cohomology groups for M,
X, and the pair (M, X) are connected by the long exact sequence of a pair in cohomology
δ i∗ δ
· · · → H k−1 (X) −→ H k (M, X) −→ H k (M) −→ H k (X) −→ · · · , (1.12)
where i∗ is induced by the embedding i : X ⊂ M, while δ is called the coboundary map-

ping.
2. Relative homology. By Hk (M) we denote the homology group generated by cycles of

dimension k in M, which are considered up to boundaries. If X ⊂ M is a submanifold, then
one has the relative homology group Hk (M, X) generated by chains, whose boundaries lie
on X. One has the following long exact sequence of a pair in homology
i∗ ∂ ∗ i
· · · → Hk (X) −→ Hk (M) −→ Hk (M, X) −→ Hk−1 (X) −→ ··· , (1.13)
where ∂ is the boundary mapping in homology (it takes each chain to its boundary).
3. Duality between homology and cohomology. Integration of forms over chains de-
fines the de Rham pairing
H k (M) × Hk (M) −→ C

[ω] , [c] −→ ω
c
on a closed manifold M and also the pairing (defined by the same formula)
H k (M, X) × Hk (M, X) −→ C
for a pair (M, X).
4. Integration over the fiber. Let π : E → B be a locally trivial fiber bundle with com-
pact smooth base B, fiber F, and the total space E. Suppose that the fibration is oriented,
i.e., there is an orientation in the fibers that depends continuously on the point of the base.
In this case one can define the mapping of integration over the fiber denoted by

: Λk (E) −→ Λk−n (B), n = dim F. (1.14)
F
Recall the definition of (1.14). To this end we choose local coordinates on E and B such
that the projection π has the form
(x1 , . . . , xk , y1 , . . . , yn ) −→ (x1 , . . . , xk ).
Then integration over the fiber is defined as

F
∑ aI (x, y)dy1 ∧ · · · ∧ dyn ∧ dxI = ∑ Y
aI (x, y)dy1 · · · dyn ,
I I
where dxI = dxi1 ∧ dxi2 ∧ · · · ∧ dxik , and I = (i1 , i2 , . . . , in ) is a multi-index. Here we sup-
pose that y1 , . . . , yn is a positively oriented coordinate system on the fiber.
Integration over the fiber has the following properties (they can be checked by a di-
rect computation):
1) it is well defined (i.e., independent of the choice of coordinates);
2) for all ω1 ∈ Λ(E) and ω2 ∈ Λ(B) we have

(ω1 ∧ π ∗ ω2 ) = ω1 ∧ ω2 ;
F F
(This property means that integration over the fiber is a homomorphism of Λ(B)-
modules.)
3) given ω ∈ Λ(E), we have

d ω = (−1)n dω.
F F
1.6. Appendix. Some notions of (co)homology theory 9
4) for all ω1 ∈ Λ(E), ω2 ∈ Λ(B) we have

(ω1 ∧ π ∗ ω2 ) = ω1 ∧ ω2 .
E B F
Here we suppose that B is oriented.

It follows from these properties that integration over the fiber induces a mapping in coho-
mology3
: H k (E) −→ H k−n (B).
F
This integration can be readily generalized to relative cohomology and cohomology
with compact supports.
5. Thom isomorphism and Thom class. Let E be an oriented vector bundle (with fiber
Rn ) over the base B.
Theorem 1.2 (Thom isomorphism). Integration over the fiber

: Hck (E) −→ H k−n (B) (1.15)
Rn
is an isomorphism for all k. The inverse mapping to (1.15) is called the Thom isomorphism
and is equal to the product with a special class
−1
TE = (1) ∈ Hcn (E), where 1 ∈ H 0 (B) is the generator,
Rn
called the Thom class. In other words, the group Hc∗ (E) is a free H ∗ (B)-module with one
generator equal to the Thom class.
Let us give an explicit formula for the Thom class of a one-dimensional complex
bundle E. Let us start with an example.
Example 1.3. Let C → pt be a one-dimensional complex line bundle over a point. Denote
the complex coordinate by x. Then we have Hc2 (C) C, and the Thom class is defined
by the 2-form
1 dx
d χ ,
2πi x
where χ ∈ C∞ (C) is a smooth function equal to zero at zero and equal to one at infinity.
A simple computation using Stokes’ formula shows that the integral of this 2-form is
equal to one. Hence, it can be taken as a representative of the Thom class.
Now let E be a trivial one-dimensional bundle E over an arbitrary base B. In this
case, the Thom class is defined by the same formula
1 ds
d χ ,
2πi s
3 It is also called the direct image mapping for the projection π.
where χ ∈ C∞ (E) is a smooth function equal to zero in a neighborhood of the zero section
of E and equal to one at infinity, while s ∈ C∞ (E) is an arbitrary function, whose restric-
tion to the fiber of E is proportional to the complex coordinate on the fiber.4 A simple
computation shows that this 2-form is closed and its integral over the fiber is equal to one
identically; hence this form is indeed a representative of the Thom class for E.
Explicit expressions for differential forms representing Thom classes of nontrivial
bundles can be found in [5], Chapters 1 and 2; see also [30].
1.7 Remarks
Leray residues were defined above as cohomology classes, and the natural question arises:
is it possible to define the residue in terms of holomorphic forms representing this coho-
mology class? It turns out, that the answer is no: even though the residue is defined as
a cohomology class, it might be the case, that there is no holomorphic representative in
this class! A counterexample was constructed by Leray (see his book [27]). Thus, the
Leray residue of a holomorphic form can be a smooth (nonholomorphic) form. How-
ever, a holomorphic representative exists if we deal with Stein manifolds (e.g., see [1] for
details).
More results and applications of Leray residues can be found in (Leray [27], Chap-
ters 1-6; Pham [36], Chapter 3; Sternin and Shatalov [53]; Shabat [40]; see also the survey
by Dolbeaut [7] and the references therein).
4 Such functions exist, because the bundle is trivial.

Chapter 2
Ramified integrals
We consider the notion of integrals depending on a parameter. Unexpectedly, this notion

turns out to be quite nontrivial in complex analysis: even in the situation, when the inte-
grand is a single-valued form, the integral is a ramified function as a rule. For this reason,
parametric integrals in complex analysis are also called ramified integrals. The aim of this
chapter is to describe this phenomenon in examples and then present the general theory
of such integrals.
2.1 Why do integrals ramify?

1. Integration over cycles. Consider the function f (t) defined by the parametric inte-
gral
dx
f (t) = , (2.1)
x −t
2
γ
where the contour of integration γ ⊂ C is a circle of radius 1/2 with center at x = 1 (see
Fig. 2.1).
Clearly, expression (2.1) determines f (t) as a holomorphic function in a neighbor-
hood of t = 1.1
The question then arises: how do we construct an analytic continuation of f (t) and
describe the singularities of this analytic continuation? Let us construct the analytic con-
tinuation of f (t), without explicitly computing the integral.2 Unfortunately, (2.1) does not
define f (t) for all t, since for some t zeroes of the denominator lie on the contour γ and
the integral diverges. The solution of this problem is obvious: we should simply move the
contour, or, more precisely, make the contour of integration depend on t. Indeed,
√ note that
(2.1) will not change, if√we take as the contour the circle with center at x = t and a small
radius such that x = − t lies outside this circle. Obviously, for t = 0 (i.e., when the roots
√
1 Indeed, the zeroes of the denominator are x = ± t, and therefore for t close to t = 1 the integrand has no
singularities on the contour of integration. Hence, the integral defines a holomorphic function of t by a standard
theorem of analysis on the differentiation of integrals, when the integrand√smoothly depends on the parameter.
2 This integral is easy to compute explicitly using the substitution x = tz.
© Springer International Publishing AG 2017 11

A. Savin, B. Sternin, Introduction to Complex Theory of Differential Equations,
Frontiers in Mathematics, DOI 10.1007/978-3-319-51744-5_2
12 Chapter 2. Ramified integrals
Figure 2.1: Contour of integration.
do not coincide) we can choose a continuous family of such contours. For instance, we
can take the family (see Fig. 2.2)
√ √
γt = x ∈ C |x − t| = 12 | t| .
√
Roughly speaking, our contour γt is associated with the root x = t. However, this root
has ramification, when the parameter goes around t = 0. Hence, our contour also ramifies
when we go around this point. Thus, (2.1) also has ramification when we go around the
point t =√0, while its Riemannian surface coincides with the Riemannian surface of the
function t. Note, finally, that the singularity at t = 0 corresponds to the simple fact that
for this value of t the contour of integration gets pinched, that is, it is impossible to deform
it continuously in such a way that it avoids the singularities of the integrand.
Let us give a homological interpretation of the constructions above. For each t the
integrand in (2.1) defines a de Rham cohomology class
dx √
∈ H 1 Cx \ {± t} . (2.2)
x −t
2
The constructed family of cycles γt defines a homology class

√
[γt ] ∈ H1 Cx \ {± t} . (2.3)
Note that [γt ] is a ramified function of t. Finally, the analytic continuation of (2.1) is
defined as the pairing of the classes (2.2) and (2.3). Since (2.3) is ramified, the integral
(2.1) is a ramified function as well.
2. Integration over relative cycles. Consider the function f (t) defined by the paramet-
ric integral
f (t) = dx, (2.4)
γt
2.1. Why do integrals ramify? 13
Figure 2.2: Deformation of the contour of integration.
√ √
where the chain γt ⊂ C is equal to the segment from − t to t. In this example the
integrand is nonsingular, but the chain has singularity.
Let us compute the ramification of (2.4) without computing the integral explicitly.
Obviously, the chain γt degenerates at t = 0 and has ramification of order two when we
go around the point t = 0. Thus, (2.4) has ramification of order two at t = 0.
Let us give a homological interpretation. The integrand in (2.4) defines a relative de
Rham cohomology class
√
[dx] ∈ H 1 Cx , {± t} . (2.5)
The chain γt defines a relative homology class
√
[γt ] ∈ H1 Cx , {± t} , (2.6)
which is a ramified function of t. Hence, the analytic continuation of (2.4) is given by the
pairing of classes (2.5) and (2.6).
Exercise 2.1. 1. Find singularities of the analytic continuation of the integral

dx
,
γ x2 + xt
where the contour γ goes around the pole x = −t.

2. Find singularities of the analytic continuation of the integral

dx
,
γ x3 + xt1 + t2
where the contour γ goes around a) only one zero of the denominator; b) two zeroes of
the denominator; c) all the zeroes of the denominator.
3. Find singularities of the analytic continuation of the integrals

dx dy, dx dy dz.
x2 +y2 ≤t x2 +y2 +z2 ≤t

dx dx
, .
γ x − t2
2
γ x2 − t 3

dx dy dx dy dz
, ,
γt x2 + y2 − t γt x2 + y2 + z2 − t
where γt is a cycle in the complement of the zero set of the denominator.
2.2 General theory. Landau manifolds

Let us describe the general theory of parametric integrals or, in other words, ramified
integrals. For simplicity we shall consider only integrals with single-valued integrands.
1. Statement of the problem. Consider the parametric integral

f (t) = ω(t) (2.7)
γ0
with parameter t ∈ T = Ctm , where

• {ω(t)} is a family of closed analytic k-forms on X = Cnx . We suppose that the coef-
ficients of the forms analytically depend on (x,t). Let S(t) ⊂ X be the set of singu-
larities of ω(t).

• [γ0 ] ∈ Hk X \ S(t0 ) is a homology class.
By the standard theorems of analysis (2.7) is well defined and holomorphic near t = t0 .3
Now the problem is to find analytic continuation of f (t) and describe its singular-
ities. The idea of constructing the analytic continuation is to extend the homology class
[γ0 ] to a class continuously depending on t. Such continuation gives a ramified function
as a rule. And for this reason this continuation is called the ramified homology class. As
soon as the ramified homology class is constructed, our problem is solved: indeed, the
analytic continuation of f (t) is given as an integral of ω(t) over this ramified homology
class.
3 Indeed, the integrand has no singularities on γ for t = t and, by continuity, near t = t . Therefore, the
0 0 0
integral admits differentiation under the integral sign.
2.2. General theory. Landau manifolds 15
2. Construction of ramified
homology classes. Landau manifolds. Let us extend the
class [γ0 ] ∈ Hk X \ S(t0 ) up to a family of homology classes

[γt ] ∈ Hk X \ S(t) (2.8)
continuously depending on the parameter. Roughly speaking, this extension is possible

until the moment t, when the homology of the complement X \ S(t) changes (i.e., the
topological type of X \ S(t) changes). Below we shall show that the set of those values of
the parameter can be effectively computed.
Consider the set
S = S(t) × {t} ⊂ X × T.
t
In general, this set has singularities. Let us assume for simplicity that this set is non-
singular, i.e., S is a regular submanifold. The general case is treated at the end of this
subsection.
Definition 2.1. The projection
π : (X × T, S) −→ T
(x,t) −→ t
of the pair (X × T, S) is locally trivial, if each t0 ∈ T has a neighborhood U and a contin-

uous family of homeomorphisms of pairs

gt : X, S(t) −→ X, S(t0 ) ,
defined for all t ∈ U.

Example 2.1. The projection onto the horizontal axis shown on Fig. 2.3 A, is locally
trivial, while the projection on Fig. 2.3 B is not locally trivial. Local triviality condition
is not satisfied in the second example at t = t0 , where the topology of the fiber changes:
S(t0 − ε) consists of two points, while S(t0 + ε) is empty.
It turns out that an arbitrary projection is locally trivial on the complement of a spe-
cial set, which we now define.
Definition 2.2. Consider the singularity set for the projection π|S : S → T , i.e., the set of
points (x,t), at which the rank of the differential dπ|(x,t) is less than dim T . The projection
of this singularity set to the base T is called the Landau manifold for the projection π and
denoted by L.
Example 2.2. The Landau manifold for the projection on Fig. 2.3 A is empty, while
for the projection on Fig. 2.3 B it consists of the point t0 . The Landau manifold for the
projection of a surface S on the screen (see Fig. 2.4) is the visible contour of this surface.
For this reason, Landau manifolds are also called visible contours.
Theorem 2.1 (Thom isotopy theorem). The projection π is locally trivial on the comple-
ment T \ L provided that X is compact.
Figure 2.3: Projections and their Landau manifolds.
Figure 2.4: Projection of a surface on the screen.
Remark 2.1. The compactness condition in Thom’s theorem is necessary, as Fig. 2.5
shows. Here we consider the projection R2y,t → Rt , while S = {ty = 1}. Obviously, t = 0
does not lie on the Landau manifold, while at this point the projection is not locally trivial,
since points of S(t) go to infinity as t → 0. In applications, the compactness condition
means that we have to pass from the original problem on a noncompact space Cn to its
2.2. General theory. Landau manifolds 17
Figure 2.5: The projection S → T is not locally trivial at t = 0.
compactification CPn and, in particular, the Landau manifold also has contributions from
points at infinity.
Now we can construct the ramified homology class (2.8). Consider the bundle over
T with the fiber at t ∈ T equal to the homology group H∗ (X \ S(t)). Since π is locally
trivial over the complement T \ L of the Landau manifold, it follows that this bundle is
also locally trivial. This and the fact that H∗ (X \ S(t)) has discrete topology implies that
[γ0 ] extends uniquely up to a continuous family of homology classes
[γt ] ∈ Hk (X \ S(t))
defined over a Riemannian surface T → T \ L.

Definition 2.3. A ramifying homology class for the projection
π : (X × T, S) −→ T
of the pair (X × T, S) is a locally constant function

γ(t) ∈ Hk X \ S(t)
defined over a Riemannian surface over T \ L, where L stands for the Landau manifold
for π.
3. Singularity sets of parametric integrals.

Theorem 2.2. Let ω(t) be a form with analytic coefficients and γ(t) a ramified homol-
ogy class defined on a Riemannian surface T . Then the integral (2.7) is a holomorphic
function on T , while its singularities lie on the Landau manifold L.
Exercise 2.2. Prove Theorem 2.2.
4. Example. Let us find the singularities and the Riemannian surface of the integral

dx
f (t) = 2 −t
. (2.9)
γ(t) x
The set S = {x2 = t} ⊂ C2x,t is nonsingular. In this case the Landau manifold is just a one-
point space
L = {t = 0} ⊂ Ct ,
which is just the set of singularities of the projection of S to the t-axis. A ramified cycle
can be defined as √ 1√ √
γt = x |x − t| = | t| ⊂ Cx \ {± t}
2
√
and has Riemannian surface equal to that for the function t. √
Hence, the Riemannian surface for (2.9) coincides with that for the function t.
5. Stratified singularity sets. In general, the set S is stratified, i.e., it is a finite union

S= Aj
j
of smooth manifolds A j (called strata) such that the boundary A j \A j is a union of strata of
lower dimensions and the Whitney conditions are satisfied by any pair of adjacent strata
(for more details, see the Appendix to this chapter).
The results described above translate to the situation of stratified S with the follow-
ing natural changes in the definition of local triviality and Landau manifolds:
• in Definition 2.1 we need to suppose that homeomorphisms gt are stratified, i.e.,
they map strata to strata;
• in Definition 2.2 the Landau manifold for S should be defined as the union of Landau
manifolds of the strata A j .
Taking into account these changes, Thom’s Theorem 2.1 and Theorem 2.2 on the
singularities of parametric integrals remain valid.
6. Example. Consider the integral

dx
f (t) = ,
γ x2 − t 3
where the contour γ encircles the root x = t 3/2 .

We see that S = {x2 = t 3 } ⊂ C2x,t is a semicubical parabola with the singular point
x = t = 0. Consider the stratification (see Fig. 2.6)

S = x = t 3/2 = 0 ∪ {x = t = 0} ≡ A1 ∪ A2 .
2.3. Integrals over relative cycles 19
Figure 2.6: Stratification of a semicubical parabola.
A computation shows that the Landau manifold is equal to
L = {t = 0} ⊂ Ct .
As a ramified cycle we can take

1
γt = x |x − t 3/2 | = |t 3/2 | ⊂ Cx \ {±t 3/2 }.
2
Hence, the integral has ramification of the type t 1/2 .
Exercise 2.3. Given a ∈ C, find the Riemannian surface of the integral

dx
f (t) = ,
γt (x2 − t)(x − a)
√
where γt is a closed contour, which encircles the points x = a and x = t. Give two
solutions of this problem : 1) first, use analytic continuation of the integral by means of
contour deformations; 2) second, compute the integral explicitly using residues. Compare
the two answers.
2.3 Integrals over relative cycles

1. Integrals of holomorphic forms with zeroes. Given a closed holomorphic form
ω(t) of degree m with coefficients holomorphic in x and t, we suppose that this form
vanishes on a hypersurface S(t)
ω(t)|S(t) = 0. (2.10)
Such a form defines a relative cohomology class
[ω(t)] ∈ H m (X, S(t)),

which can be integrated over a relative homology class
[γ0 ] ∈ Hm (X, S(t0 )).
Let us give analytic continuation of the integral

f (t) = ω(t) (2.11)
γ0
and describe its singularity set. To this end, consider the projection
π : (X × T, S) −→ T (2.12)
of the pair (X × T, S). By Thom’s theorem this projection is locally trivial over T \ L,
where L stands for the Landau manifold. Hence, [γ0 ] extends to a ramified homology
class denoted by
[γt ] ∈ Hm X, S(t) ,
where t runs over a Riemannian surface T over T \ L.
Thus, the analytic continuation of (2.11) is defined by de Rham pairing of ω(t) and
γt , and is a holomorphic function on the Riemannian surface T , while the singularities of
the integral lie on the Landau manifold.
Example 2.3. Consider the integral

f (t) = dx ∧ dy, (2.13)
γt

where γt = (x, y) | x2 + y2 ≤ t . The analytic continuation of this relative cycle is equal
to √ √
γt = (u t, v t) ∈ C2 | u2 + v2 ≤ 1, u, v ∈ R .
In this case, the Landau manifold is L = {t = 0}. Indeed, we have a paraboloid S =
{x2 + y2 = t} (see Fig. 2.7). Clearly, the singularity of the projection of this surface to the
axis Rt is precisely the vertex x = y = t = 0. A direct computation shows that neither the
family of cycles

[γt ] ∈ H2 C2 , S(t) , where S(t) = (x, y) | x2 + y2 = t ,
nor the integral (2.13) have ramification at t = 0.

Example 2.4. Consider the circle S1 = {x2 + y2 = 1} ⊂ R2 and the family of straight
lines Lt = {y = t}. Define f (t) as the area of the circular segment cut off by Lt (see
Fig. 2.8). Let us show that f (t) is analytic in t and find the Riemannian surface of its
analytic continuation. To this end, we write f (t) as the integral

f (t) = dx ∧ dy,
Dt
2.3. Integrals over relative cycles 21
Figure 2.7: Surface S = {x2 + y2 = t}.
Figure 2.8: Circular segment.

√
Figure 2.9: Riemannian surface of 1 − t 2.
where Dt stands for the circular segment. The relative cycle in this integral defines the
class
[Dt ] ∈ H2 (C2 , Σ ∪ Lt ),
where Σ = {x2 + y2 = 1} ⊂ C2 .
Exercise 2.4. Show that [Dt ] has ramification of order two around the points
√ t = ±1,
while the Riemannian surface of f (t) is the same as that for the function 1 − t 2 , see
Fig. 2.9.
2. Integrals of holomorphic forms with zeroes and singularities. Let ω(t) be a closed
holomorphic form of degree m, which is equal to zero on a submanifold denoted by S(t)
and has singularities on a set denoted by Y (t). Such a form defines a relative cohomology
class
[ω(t)] ∈ H m X \Y (t), S(t) .
Suppose we are also given a homology class

γ0 ∈ Hm X \Y (t0 ), S(t0 ) .
Let us obtain analytic continuation for the integral

f (t) = ω(t) (2.14)
γ0
and describe its singularity set.

To this end, we consider the projection
π : (X × T, S ∪Y ) → T (2.15)
of the pair (X × T, S ∪ Y ). Let L be the Landau manifold for this projection. Then by
Thom theorem this projection is locally trivial over T \L. Hence, [γ0 ] extends to a ramified
homology class
[γt ] ∈ Hm X \Y (t), S(t) ,
Another random document with
no related content on Scribd:
artist, the decorator, the paper mills’ agent, and, last of all, the printer and the
binder. This was not the way the old-time printers had planned their books.
With all their mechanical limitations, they had followed architectural lines kept
consistent and harmonious because controlled by a single mind, while the
finished volume of the eighteen-nineties was a composite production of many
minds, with no architectural plan. No wonder that the volumes manufactured,
even in the most famous Presses, failed to compare with those produced in
Venice by Jenson and Aldus four centuries earlier!
When I succeeded John Wilson as head of the University Press in 1895, I
determined to carry out the resolution I had formed four years earlier, while
sitting in on the Eugene Field conference, of following the example of the
early master-printers so far as this could be done amidst modern conditions.
Some of my publisher friends were partially convinced by my contention that
if the printer properly fulfilled his function he must know how to express his
clients’ mental conception of the physical attributes of prospective volumes in
terms of type, paper, presswork, and binding better than they could do it
themselves. The Kelmscott publications, which appeared at this time, were of
great value in emphasizing my contention, for William Morris placed printing
back among the fine arts after it had lapsed into a trade.
I had no idea, when I presented my plan, of persuading my friends to
produce typographical monuments. No demand has ever existed for volumes
of this type adequate to the excessive cost involved by the perfection of
materials, the accuracy of editorial detail, the supreme excellence of
typography and presswork, and the glory of the binding. Sweynheim and
Pannartz, Gutenberg’s successors, were ruined by their experiments in Greek;
the Aldine Press in Venice was saved only by the intervention of Jean Grolier;
Henri Étienne was ruined by his famous Thesaurus, and Christophe Plantin
would have been bankrupted by his Polyglot Bible had he not retrieved his
fortunes by later and meaner publications. Nor was I unmindful of similar
examples that might have been cited from more modern efforts, made by
ambitious publishers and printers.
What I wanted to do was to build low-cost volumes upon the same
principles as de luxe editions, eliminating the expensive materials but retaining
the harmony and consistency that come from designing the book from an
architectural standpoint. It adds little to the expense to select a type that
properly expresses the thought which the author wishes to convey; or to have
the presses touch the letters into the paper in such a way as to become a part
of it, without that heavy impression which makes the reverse side appear like
an example of Braille; or to find a paper (even made by machine!) soft to the
feel and grateful to the eye, on which the page is placed with well-considered
margins; or to use illustrations or decorations, if warranted at all, in such a way
as to assist the imagination of the reader rather than to divert him from the
text; to plan a title page which, like the door to a house, invites the reader to
open it and proceed, its type lines carefully balanced with the blank; or to bind
(even in cloth!) with trig squares and with design or lettering in keeping with
the printing inside.
By degrees the publishers began to realize that this could be done, and
when once established, the idea of treating the making of books as a
manufacturing problem instead of as a series of contracts with different
concerns, no one of which knew what the others were doing, found favor. The
authors also preferred it, for their literary children now went forth to the
world in more becoming dress. Thus serving in the capacity of book architect
and typographical advisor, instead of merely as a contrasting printer, these
years have been lived in a veritable Kingdom of Books, in company with
interesting people,—authors and artists as well as publishers,—in a delightfully
intimate way because I have been permitted to be a part of the great
adventure.
During these years I have seen dramatic changes. Wages were somewhat
advanced between 1891 and the outbreak of the World War, but even at this
latter date the cost of manufacturing books was less than half of what it is now.
This is the great problem which publishers have to face today. When the cost
of everything doubled after the World War, the public accepted the necessity
of paying twice the price for a theater ticket as a matter of course; but when
the retail price of books was advanced in proportion to the cost of
manufacture, there was a great outcry among buyers that authors, publishers,
and booksellers were opportunists, demanding an unwarranted profit. As a
matter of fact, the novel which used to sell at $1.35 per copy should now sell
at $2.50 if the increased costs were properly apportioned. The publisher today
is forced to decline many promising first novels because the small margin of
profit demands a comparatively large first edition.
Unless a publisher can sell 5,000 copies as a minimum it is impossible for
him to make any profit upon a novel. Taking this as a basis, and a novel as
containing 320 pages, suppose we see how the $2.00 retail price distributes
itself. The cost of manufacture, including the typesetting, electrotype plates,
cover design, jacket, brass dies, presswork, paper, and binding, amounts to 42
cents per copy (in England, about 37 cents). The publisher’s cost of running
his office, which he calls “overhead,” is 36 cents per copy. The minimum
royalty received by an author is 10 per cent. of the retail price, which would
give him 20 cents. This makes a total cost of 98 cents a copy, without
advertising. But a book must be advertised.
Every fifty dollars spent in advertising on a five thousand edition adds a
cent to the publisher’s cost. The free copies distributed for press reviews
represent no trifling item. A thousand dollars is not a large amount to be spent
for advertising, and this means 20 cents a copy on a 5000 edition, making a
total cost of $1.18 per copy and reducing the publisher’s profit to 2 cents, since
he sells a two-dollar book to the retail bookseller for $1.20. The bookseller
figures that his cost of doing business is one-third the amount of his sales, or,
on a two-dollar book, 67 cents. This then shows a net profit to the retail
bookseller of 13 cents, to the publisher of 2 cents, and to the author of 20
cents a copy.
Beyond this, there is an additional expense to both bookseller and
publisher which the buyer of books is likely to overlook. It is impossible to
know just when the demand for a book will cease, and this means that the
publisher and the bookseller are frequently left with copies on hand which
have to be disposed of at a price below cost. This is an expense that has to be
included in the book business just as much as in handling fruit, flowers, or
other perishable goods.
When a publisher is able to figure on a large demand for the first edition,
he can cut down the cost of manufacture materially; but, on the other hand,
this is at least partially offset by the fact that authors whose books warrant
large first editions demand considerably more than 10 per cent. royalty, and
the advertising item on a big seller runs into large figures.
I wish I might say that I had seen a dramatic change in the methods
employed in the retail bookstores! There still exists, with a few notable
exceptions, the same lack of realization that familiarity with the goods one has
to sell is as necessary in merchandizing books as with any other commodity.
Salesmen in many otherwise well-organized retail bookstores are still painfully
ignorant of their proper functions and indifferent to the legitimate
requirements of their prospective customers.
Some years ago, when one of my novels was having its run, I happened to
be in New York at a time when a friend was sailing for Europe. He had
announced his intention of purchasing a copy of my book to read on the
steamer, and I asked him to permit me to send it to him with the author’s
compliments. Lest any reader be astonished to learn that an author ever buys a
copy of his own book, let me record the fact that except for the twelve which
form a part of his contract with the publisher, he pays cash for every copy he
gives away. Mark Twain dedicated the first edition of The Jumping Frog to “John
Smith.” In the second edition he omitted the dedication, explaining that in
dedicating the volume as he did, he had felt sure that at least all the John
Smiths would buy books. To his consternation he found that they all expected
complimentary copies, and he was hoist by his own petard!
With the idea of carrying out my promise to my friend, I stepped into one
of the largest bookstores in New York, and approached a clerk, asking him for
the book by title. My pride was somewhat hurt to find that even the name was
entirely unfamiliar to him. He ran over various volumes upon the counter, and
then turned to me, saying, “We don’t carry that book, but we have several
others here which I am sure you would like better.”
“Undoubtedly you have,” I agreed with him; “but that is beside the point. I
am the author of the book I asked for, and I wish to secure a copy to give to a
friend. I am surprised that a store like this does not carry it.”
Leaning nonchalantly on a large, circular pile of books near him, the clerk
took upon himself the education of the author.
“It would require a store much larger than this to carry every book that is
published, wouldn’t it?” he asked cheerfully. “Of course each author naturally
thinks his book should have the place of honor on the bookstalls, but we have
to be governed by the demand.”
It was humiliating to learn the real reason why this house failed to carry
my book. I had to say something to explain my presumption even in assuming
that I might find it there, so in my confusion I stammered,
“But I understood from the publishers that the book was selling very
well.”
“Oh, yes,” the clerk replied indulgently; “they have to say that to their
authors to keep them satisfied!”
With the matter thus definitely settled, nothing remained but to make my
escape as gracefully as circumstances would permit. As I started to leave, the
clerk resumed his standing position, and my eye happened to rest on the pile
of perhaps two hundred books upon which he had been half-reclining. The
jacket was strikingly familiar. Turning to the clerk I said severely,
“Would you mind glancing at that pile of books from which you have just
risen?”
“Oh!” he exclaimed, smiling and handing me a copy, “that is the very book
we were looking for, isn’t it?”
It seemed my opportunity to become the educator, and I seized it.
“Young man,” I said, “if you would discontinue the practice of letting my
books support you, and sell a few copies so that they might support me, it
would be a whole lot better for both of us.”
“Ha, ha!” he laughed, graciously pleased with my sally; “that’s a good line,
isn’t it? I really must read your book!”
The old-time publisher is passing, and the author is largely to blame. I have
seen the close association—in many cases the profound friendship—between
author and publisher broken by the commercialism fostered by some literary
agents and completed by competitive bids made by one publishing house to
beguile a popular author away from another. There was a time when a writer
was proud to be classified as a “Macmillan,” or a “Harper” author. He felt
himself a part of the publisher’s organization, and had no hesitation in taking
his literary problems to the editorial advisor of the house whose imprint
appeared upon the title pages of his volumes. A celebrated Boston authoress
once found herself absolutely at a standstill on a partially completed novel. She
confided her dilemma to her publisher, who immediately sent one of his
editorial staff to the rescue. They spent two weeks working together over the
manuscript, solved the problems, and the novel, when published, was the most
successful of the season.
Several publishers have acknowledged to me that in offering unusually
high royalties to authors they have no expectation of breaking even, but that to
have a popular title upon their list increases the sales of their entire line. The
publisher from whom the popular writer is filched has usually done his share
in helping him attain his popularity. The royalty he pays is a fair division of the
profits. He cannot, in justice to his other authors, pay him a further premium.
Ethics, perhaps, has no place in business, but the relation between author
and publisher seems to me to be beyond a business covenant. A publisher may
deliberately add an author to his list at a loss in order to accomplish a specific
purpose, but this practice cannot be continued indefinitely. A far-sighted
author will consider the matter seriously before he becomes an opportunist.
In England this questionable practice has been of much slower growth.

The House of Murray, in London, is one of those still conducted on the old-
time basis. John Murray IV, the present head of the business, has no interest in
any author who comes to him for any reason other than a desire to have the
Murray imprint upon his book. It is more than a business. The publishing
offices at 50a, Albemarle Street adjoin and open out of the Murray home. In
the library is still shown the fireplace where John Murray III burned Byron’s
Memoirs, after purchasing them at an enormous price, because he deemed that
their publication would do injury to the reputation of the writer and of the
House itself.
John Murray II was one of the publishers of Scott’s Marmion. In those days
it was customary for publishers to share their contracts. Constable had
purchased from Scott for £1,000 the copyright of Marmion without having
seen a single line, and the honorarium was paid the author before the poem was
completed or the manuscript delivered. Constable, however, promptly
disposed of a one-fourth interest to Mr. Miller of Albemarle Street, and
another one fourth to John Murray, then of Fleet Street.
By 1829 Scott had succeeded in getting into his own hands nearly all his
copyrights, one of the outstanding items being this one-quarter interest in
Marmion held by Mr. Murray. Longmans and Constable had tried in vain to
purchase it. When, however, Scott himself approached Murray through
Lockhart, the following letter from Mr. Murray was the result:
So highly do I estimate the honour of being even in so small a degree the publisher
of the author of the poem that no pecuniary consideration whatever can induce me to
part with it. But there is a consideration of another kind that would make it painful
to me if I were to retain it a moment longer. I mean the knowledge of its being required
by the author, into whose hands it was spontaneously resigned at the same instant that
I read the request.
There has always been a vast difference in authors in the attitude they
assume toward the transformation of their manuscripts into printed books.
Most of them leave every detail to their publishers, but a few take a deep and
intelligent personal interest. Bernard Shaw is to be included in the latter group.
A leading Boston publisher once telephoned me that an unknown English
author had submitted a manuscript for publication, but that it was too
socialistic in its nature to be acceptable. Then the publisher added that the
author had asked, in case this house did not care to publish the volume, that
arrangements be made to have the book printed in this country in order to
secure American copyright.
“We don’t care to have anything to do with it,” was the statement; “but I
thought perhaps you might like to manufacture the book.”
“Who is the author?” I inquired.
“It’s a man named Shaw.”
“What is the rest of his name?”
“Wait a minute and I’ll find out.”
Leaving the telephone for a moment, the publisher returned and said,
“His name is G. Bernard Shaw. Did you ever hear of him?”
“Yes,” I replied; “I met him last summer in London through Cobden-
Sanderson, and I should be glad to undertake the manufacture of the book for
Mr. Shaw.”
“All right,” came the answer. “Have your boy call for the manuscript.”
This manuscript was Man and Superman.
From that day and for many years, Shaw and I carried on a desultory
correspondence, his letters proving most original and diverting. On one
occasion he took me severely to task for having used two sizes of type upon a
title page. He wrote four pages to prove what poor taste and workmanship this
represented, and then ended the letter with these words, “But, after all, any
other printer would have used sixteen instead of two, so I bless you for your
restraint!”
We had another lengthy discussion on the use of apostrophes in printing.
“I have made no attempt to deal with the apostrophes you introduce,” he
wrote; “but my own usage is carefully considered and the inconsistencies are
only apparent. For instance, Ive, youve, lets, thats, are quite unmistakable, but Ill,
hell, shell, for I’ll, he’ll, she’ll, are impossible without a phonetic alphabet to
distinguish between long and short e. In such cases I retain the apostrophe, in
all others I discard it. Now you may ask me why I discard it. Solely because it
spoils the printing. If you print a Bible you can make a handsome job of it
because there are no apostrophes or inverted commas to break up the
letterpress with holes and dots. Until people are forced to have some
consideration for a book as something to look at as well as something to read,
we shall never get rid of these senseless disfigurements that have destroyed all
the old sense of beauty in printing.”
“Ninety-nine per cent. of the secret of good printing,” Shaw continued, “is
not to have patches of white or trickling rivers of it trailing down a page, like
rain-drops on a window. Horrible! White is the enemy of the printer. Black,
rich, fat, even black, without gray patches, is, or should be, his pride. Leads and
quads and displays of different kinds of type should be reserved for insurance
prospectuses and advertisements of lost dogs.…”
His enthusiasm for William Morris’ leaf ornaments is not shared by all
booklovers. Glance at any of the Kelmscott volumes, and you will find these
glorified oak leaves scattered over the type page in absolutely unrelated
fashion,—a greater blemish, to some eyes, than occasional variation in spacing.
Shaw writes:
If you look at one of the books printed by William Morris, the greatest printer of
the XIX century, and one of the greatest printers of all the centuries, you will see that
he occasionally puts in a little leaf ornament, or something of the kind. The idiots in
America who tried to imitate Morris, not understanding this, peppered such things all
over their “art” books, and generally managed to stick in an extra large quad before
each to show how little they understood about the business. Morris doesn’t do this in
his own books. He rewrites the sentence so as to make it justify, without bringing one
gap underneath another in the line above. But in printing other people’s books, which
he had no right to alter, he sometimes found it impossible to avoid this. Then, sooner
than spoil the rich, even color of his block of letterpress by a big white hole, he filled it
up with a leaf.
Do not dismiss this as not being “business.” I assure you, I have a book which
Morris gave me, a single copy, by selling which I could cover the entire cost of printing
my books, and its value is due solely to its having been manufactured in the way I
advocate; there’s absolutely no other secret about it; and there is no reason why you
should not make yourself famous through all the ages by turning out editions of
standard works on these lines whilst other printers are exhausting themselves in dirty
felt end papers, sham Kelmscott capitals, leaf ornaments in quad sauce, and then
wondering why nobody in Europe will pay twopence for them, whilst Kelmscott books
and Doves Press books of Morris’ friends, Emery Walker and Cobden-Sanderson,
fetch fancy prices before the ink is thoroughly dry.… After this I shall have to get you
to print all my future books, so please have this treatise printed in letters of gold and
preserved for future reference
CHAPTER III
Friends through Type

III
FRIENDS THROUGH TYPE
In 1903 I again visited Italy to continue my study of the art of printing in

the old monasteries and libraries, sailing on the S. S. Canopic from Boston to
Naples. Among the passengers on board I met Horace Fletcher, returning to
his home in Venice. At that time his volume Menticulture was having a
tremendous run. I had enjoyed reading the book, and in its author I discovered
a unique and charming personality; in fact, I have never met so perfect an
expression of practical optimism. His humor was infectious, his philosophy
appealing, his quiet persistency irresistible.
To many people the name of Horace Fletcher has become associated with
the Gladstonian doctrine of excessive chewing, but this falls far short of the
whole truth. His scheme was the broadest imaginable, and thorough
mastication was only the hub into which the other spokes of the wheel of his
philosophy of life were to be fitted. The scheme was nothing less than a
cultivation of progressive human efficiency. Believing that absolute health is
the real basis of human happiness and advancement, and that health depends
upon an intelligent treatment of food in the mouth together with knowledge
of how best to furnish the fuel that is actually required to run the human
engine, Horace Fletcher sought for and found perfect guides among the
natural human instincts and physiologic facilities, and demonstrated that his
theories were facts.
During the years that followed I served as his typographic mentor. He was
eager to try weird and ingenious experiments to bring out the various points of
his theories through unique typographical arrangement (see opp. page). It
required all my skill and diplomacy to convince him that type possessed rigid
limitations, and that to gain his emphasis he must adopt less complicated
methods. From this association we became the closest of friends, and
presuming upon this relation I used to banter him upon being so casual. His
copy was never ready when the compositors needed it; he was always late in
returning his proofs. The manufacture of a Fletcher book was a hectic
experience, yet no one ever seemed to take exceptions. This was characteristic
of the man. He moved and acted upon suddenly formed impulses, never
planning ahead yet always securing exactly what he wanted, and those
inconvenienced the most always seemed to enjoy it.
A Page of Horace Fletcher Manuscript
“I believe,” he used to say, “in hitching one’s wagon to a star, but I always
keep my bag packed and close at hand ready to change stars at a moment’s
notice. It is only by doing this that you can give things a chance to happen to
you.”
Among the volumes Fletcher had with him on board ship was one he had
purchased in Italy, printed in a type I did not recognize but which greatly
attracted me by its beauty. The book bore the imprint: Parma: Co’tipi Bodoniani.
Some weeks later, in a small, second-hand bookstore in Florence, I happened
upon a volume printed in the same type, which I purchased and took at once
to my friend, Doctor Guido Biagi, at the Laurenziana Library.
“The work of Giambattista Bodoni is not familiar to you?” he inquired in
surprise. “It is he who revived in Italy the glory of the Aldi. He and Firmin
Didot in Paris were the fathers of modern type design at the beginning of the
nineteenth century.”
“Is this type still in use?” I inquired.
“No,” Biagi answered. “When Bodoni died there was no one worthy to
continue its use, so his matrices and punches are kept intact, exactly as he left
them. They are on exhibition in the library at Parma, just as the old Plantin
relics are preserved in the museum at Antwerp.”
GIAMBATTISTA BODONI, 1740–1813
From Engraving at the Bibliothèque Nationale, Paris
I immediately took steps through our Ambassador at Rome to gain

permission from the Italian Government to recut this face for use in America.
After considerable difficulty and delay this permission was granted, with a
proviso that I should not allow any of the type made from my proposed
matrices to get into the hands of Italian printers, as this would detract from
the prestige of the city of Parma. It was a condition to which I was quite
willing to subscribe! Within a year I have received a prospectus from a revived
Bodoni Press at Montagnola di Lugano, Switzerland, announcing that the
exclusive use of the original types of Giambattista Bodoni has been given them
by the Italian Government. This would seem to indicate that the early
governmental objections have disappeared.
While searching around to secure the fullest set of patterns, I stumbled
upon the fact that Bodoni and Didot had based their types upon the same
model, and that Didot had made use of his font particularly in the wonderful
editions published in Paris at the very beginning of the nineteenth century. I
then hurried to Paris to see whether these matrices were in existence. There,
after a search through the foundries, I discovered the original punches, long
discarded, in the foundry of Peignot, to whom I gave an order to cast the
different sizes of type, which I had shipped to America.
This was the first type based on this model ever to come into this country.
The Bodoni face has since been recut by typefounders as well as for the
typesetting machines, and is today one of the most popular faces in common
use. Personally I prefer the Bodoni letter to that of Didot (see opp. page). The
Frenchman succumbed to the elegance of his period, and by lightening the
thin lines robbed the design of the virility that Bodoni retained. I am not in
sympathy with the excessive height of the ascending letters, which frequently
extend beyond the capitals; but when one considers how radical a departure
from precedent this type was, he must admire the skill and courage of the
designers. William Morris cared little for it,—“The sweltering hideousness of
the Bodoni letter,” he exclaimed; “the most illegible type that was ever cut,
with its preposterous thicks and thins”; while Theodore L. De Vinne, in his
Practice of Typography, writes:
The beauty of the Bodoni letters consists in their regularity, in their clearness, and
in their conformity to the taste of the race, nation, and age in which the work was first
written, and finally in the grace of the characters, independent of time or place.
When authorities differ to such a wide extent, the student of type design
must draw his own conclusions!
The Bodoni Letter (bottom) compared with the Didot Letter (top)
Fletcher’s idea of an appointment was something to be kept if or when

convenient, yet he never seemed to offend any one. He did nothing he did not
wish to do, and his methods of extricating himself from unwelcome
responsibilities always amused rather than annoyed. “If you don’t want to do a
thing very badly,” he confided to me on one such occasion, “do it very badly.”
HORACE FLETCHER IN 1915
On board the Canopic Fletcher was surrounded by an admiring and

interested group. General Leonard Wood was on his way to study colonial
government abroad before taking up his first administration as Governor of
the Philippines. On his staff was General Hugh Lennox Scott, who later
succeeded General Wood as Chief of Staff of the United States Army. The
conversations and discussions in the smokeroom each evening after dinner
were illuminating and fascinating. General Wood had but recently completed
his work as Governor of Cuba, and he talked freely of his experiences there,
while General Scott was full of reminiscences of his extraordinary adventures
with the Indians. He later played an important part in bringing peace to the
Philippines.
It was at one of these four-cornered sessions in the smokeroom that we
first learned of Fletcher’s ambition to revolutionize the world in its methods of
eating. That he would actually accomplish this no one of us believed, but the
fact remains. The smokeroom steward was serving the coffee, inquiring of
each one how many lumps of sugar he required. Fletcher, to our amazement,
called for five! It was a grand-stand play in a way, but he secured his audience
as completely as do the tambourines and the singing of the Salvation Army.
“Why are you surprised?” he demanded with seeming innocence. “I am
simply taking a coffee liqueur, in which there is less sugar now than there is in
your chartreuse or benedictine. But I am mixing it with the saliva, which is
more than you are doing. The sugar, as you take it, becomes acid in the
stomach and retards digestion; by my method, it is changed into grape sugar,
which is easily assimilated.”
“To insalivate one’s liquor,” he explained to us, “gives one the most
exquisite pleasure imaginable, but it is a terrific test of quality. It brings out the
richness of flavor, which is lost when one gulps the wine down. Did you ever
notice the way a tea-taster sips his tea?”
As he talked he exposed the ignorance of the entire group on physiological
matters to an embarrassing extent, clinching his remarks by asking General
Wood the question,
“Would you engage as chauffeur for your automobile a man who knew as
little about his motor as you know about your own human engine?”
No one ever loved a practical joke better than Horace Fletcher. I was a
guest at a dinner he once gave at the Graduates’ Club in New Haven. Among
the others present were President Hadley of Yale, John Hays Hammond,
Walter Camp, and Professor Lounsbury. There was considerable curiosity and
some speculation concerning what would constitute a Fletcher dinner. At the
proper time we were shown into a private room, where the table was set with
the severest simplicity. Instead of china, white crockery was used, and the chief
table decorations were three large crockery pitchers filled with ice water. At
each plate was a crockery saucer, containing a shredded-wheat biscuit. It was
amusing to glance around and note the expressions of dismay upon the faces
of the guests. Their worst apprehensions were being confirmed! Just as we
were well seated, the headwaiter came to the door and announced that by
mistake we had been shown into the wrong room, whereupon Fletcher, with
an inimitable twinkle in his eye, led the way into another private dining room,
where we sat down to one of the most sumptuous repasts I have ever enjoyed.
Today, twenty years after his campaign, it is almost forgotten that the
American breakfast was at that time a heavy meal. Horace Fletcher
revolutionized the practice of eating, and interjected the word fletcherize into
the English language. As a disciple of Fletcher Sir Thomas Barlow, physician-
in-chief to King Edward VII, persuaded royalty to set the style by cutting
down the formal dinner from three hours to an hour and a half, with a
corresponding relief to the digestive apparatus of the guests. In Belgium,
during the World War, working with Herbert Hoover, Fletcher taught the
impoverished people how to sustain themselves upon meager rations. Among
his admirers and devoted friends were such profound thinkers as William
James who, in response to a letter from him, wrote, “Your excessive reaction
to the stimulus of my grateful approval makes you remind me of those rich
soils which, when you tickle them with a straw, smile with a harvest”; and
Henry James, who closes a letter: “Come and bring with you plenary
absolution to the thankless subject who yet dares light the lamp of gratitude to
you at each day’s end of his life.”
My acquaintance with Henry James came through my close association

with the late Sir Sidney Lee, the Shakesperian authority, and Horace Fletcher.
“Don’t be surprised if he is brusque or uncivil,” Sir Sidney whispered to
me just before I met him at dinner; “one can never tell how he is going to act.”
As a matter of fact, I found Henry James a most genial and enjoyable
dinner companion, and never, during the few later occasions when I had the
pleasure of being with him, did he display those characteristics of ill humor
and brusqueness which have been attributed to him. It may not be generally
known that all his life—until he met Horace Fletcher—he suffered torments
from chronic indigestion, or that it was in Fletcherism that he found his first
relief. In a typically involved Jamesian letter to his brother William he writes
(February, 1909):
It is impossible save in a long talk to make you understand how the blessed
Fletcherism—so extra blessed—lulled me, charmed me, beguiled me, from the first
into the convenience of not having to drag myself out into eternal walking. One must

Textbook Introduction To Complex Theory of Differential Equations Savin Ebook All Chapter PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Textbook Introduction To Complex Theory of Differential Equations Savin Ebook All Chapter PDF

Uploaded by

Copyright:

Available Formats

Introduction to Complex Theory of

Differential Equations Savin

Introduction to Partial Differential Equations Peter J.

Distribution Theory Applied to Differential Equations

A Modern Introduction to Differential Equations 3rd

Introduction to the Theory of Complex Systems Stefan

Numerical Solution of Differential Equations:

Theory of Differential Equations in Engineering and

Modelling with Ordinary Differential Equations Dreyer

Elementary differential equations Second Edition

Introduction to Complex Theory

Leibniz Universität Hannover Leibniz Universität Hannover

ISSN 1660-8046 ISSN 1660-8054 (electronic)

Library of Congress Control Number: 2017934913

© Springer International Publishing AG 2017

Printed on acid-free paper

This book is published under the trade name Birkhäuser, www.birkhauser-science.com

Moscow–Hannover Anton Savin

3 Asymptotics of ramiﬁed integrals 31

4 Ramiﬁed Fourier transform 43

5 Properties of the ramiﬁed Fourier transform 53

6 The Cauchy problem for equations with constant coefﬁcients 61

7 Singularities of the solution of the Cauchy problem 69

8 The Cauchy problem for equations with variable coefﬁcients.

9 Balayage inwards problem 95

10 Mother body problem 109

10.4 Construction of mother body (algorithm) . . . . . . . . . . . . . . . . . . 121

Hints for exercises 129

Residues of differential forms depending on several variables were ﬁrst considered by

1.1 A glimpse of one-dimensional residues

is deﬁned in one of two equivalent ways:

(here γ is a sufﬁciently small contour going around x0 one turn counterclockwise);

© Springer International Publishing AG 2017 1

res f (x)dx = ϕ(x0 ). (1.4)

1.2 First deﬁnition of Leray residues

Figure 1.1: A tubular neighborhood of a submanifold.

1.3 Second deﬁnition of Leray residues

ω = ∑ ωI (x)dxI , where dxI = dxi1 ∧ dxi2 ∧ · · · ∧ dxik , (1.6)

Suppose that X ⊂ M is a submanifold with local deﬁning function s(x). As in the

x1 x2 dx1 ∧ dx2 (ζ − x2 )x2 dζ ∧ dx2

provided that x ∈ M is such that at this point ∂ s/∂ x j = 0.

1.5 Leray exact sequences and residue theorem

Leray sequence in cohomology. Consider the triangle

Here i∗ is induced by the embedding M \ X → M, while ω ∗ increases degrees of forms

Here i∗ is induced by the embedding M \ X → M (see the Appendix to this chapter),

Example 1.2. Consider the quadric

It is easy to see that its Leray coboundary

is the homology class of the torus

Theorem 1.1 (Residue theorem). Given a cycle h on X and a closed form ω on M \ X,

Exercise 1.3. Prove residue formula (1.11).

Res : H ∗ (M \ X, S) −→ H ∗−1 (X, S).

Its construction is similar to the above and we do not stop on it.

1.6 Appendix. Some notions of (co)homology theory

1. Relative cohomology. The de Rham cohomology of M is denoted by H k (M). It is

where i∗ is induced by the embedding i : X ⊂ M, while δ is called the coboundary map-

2. Relative homology. By Hk (M) we denote the homology group generated by cycles of

for a pair (M, X).

Then integration over the ﬁber is deﬁned as

4) for all ω1 ∈ Λ(E), ω2 ∈ Λ(B) we have

Here we suppose that B is oriented.

4 Such functions exist, because the bundle is trivial.

We consider the notion of integrals depending on a parameter. Unexpectedly, this notion

2.1 Why do integrals ramify?