Full download ebook at ebookgate.

com

Automorphic forms and L Functions for the

group GL n R 1st Edition Dorian Goldfeld

https://ebookgate.com/product/automorphic-forms-
and-l-functions-for-the-group-gl-n-r-1st-edition-
dorian-goldfeld/

Download more ebook from https://ebookgate.com

More products digital (pdf, epub, mobi) instant
download maybe you interests ...

Automorphic Representations and L Functions for the

General Linear Group Volume 2 1st Edition Dorian
Goldfeld

https://ebookgate.com/product/automorphic-representations-and-l-
functions-for-the-general-linear-group-volume-2-1st-edition-
dorian-goldfeld/

Automorphic Representations and L Functions for the

General Linear Group Volume 1 1st Edition Dorian
Goldfeld

https://ebookgate.com/product/automorphic-representations-and-l-
functions-for-the-general-linear-group-volume-1-1st-edition-
dorian-goldfeld/

The GL Diet For Dummies 1st Edition Nigel Denby

https://ebookgate.com/product/the-gl-diet-for-dummies-1st-
edition-nigel-denby/

Complex Functions Examples c 3 Elementary Analytic

Functions and Harmonic Functions 1st edition Edition
Mejlbro L.

https://ebookgate.com/product/complex-functions-
examples-c-3-elementary-analytic-functions-and-harmonic-
functions-1st-edition-edition-mejlbro-l/
Telomerase Composition Functions and Clinical
Implications Composition Functions and Clinical
Implications 1st Edition Aiden N. Gagnon

https://ebookgate.com/product/telomerase-composition-functions-
and-clinical-implications-composition-functions-and-clinical-
implications-1st-edition-aiden-n-gagnon/

Operation PLUM The Ill fated 27th Bombardment Group and

the Fight for the Western Pacific 1st Edition Adrian R.
Martin

https://ebookgate.com/product/operation-plum-the-ill-fated-27th-
bombardment-group-and-the-fight-for-the-western-pacific-1st-
edition-adrian-r-martin/

The Ministry of Truth 1st Edition Dorian Lynskey

https://ebookgate.com/product/the-ministry-of-truth-1st-edition-
dorian-lynskey/

Contributions of Self Psychology to Group Psychotherapy

Selected Papers The New International Library of Group
Analysis 1st Edition Walter N. Stone

https://ebookgate.com/product/contributions-of-self-psychology-
to-group-psychotherapy-selected-papers-the-new-international-
library-of-group-analysis-1st-edition-walter-n-stone/

Advances in Group Processes 1st Edition Shane R. Thye

https://ebookgate.com/product/advances-in-group-processes-1st-
edition-shane-r-thye/
 CAMBRIDGE STUDIES IN ADVANCED MATHEMATICS
Editorial Board
B. Bollobas, W. Fulton, A. Katok, F. Kirwan, P. Sarnak, B. Simon, B. Totaro

AUTOMORPHIC FORMS AND L-FUNCTIONS FOR

THE GROUP GL (n, R)

L-functions associated with automorphic forms encode all classical num-

ber theoretic information. They are akin to elementary particles in physics.
This book provides an entirely self-contained introduction to the theory of
L-functions in a style accessible to graduate students with a basic knowledge
of classical analysis, complex variable theory, and algebra. Also within the
volume are many new results not yet found in the literature. The exposition
provides complete detailed proofs of results in an easy-to-read format using
many examples and without the need to know and remember many complex
definitions. The main themes of the book are first worked out for GL(2,R) and
GL(3,R), and then for the general case of GL(n,R). In an appendix to the book,
a set of Mathematica® functions is presented, designed to allow the reader to
explore the theory from a computational point of view.
CAMBRIDGE STUDIES IN ADVANCED MATHEMATICS
Editorial Board:
B. Bollobas, W. Fulton, A. Katok, F. Kirwan, P. Sarnak, B. Simon, B. Totaro
Already published
30 D.J. Benson Representations and cohomology I
31 D.J. Benson Representations and cohomology II
32 C. Allday & V. Puppe Cohomological methods in transformation groups
33 C. Soule et al. Lectures on Arakelov geometry
34 A. Ambrosetti & G. Prodi A primer of nonlinear analysis
35 J. Palis & F. Takens Hyperbolicity, stability and chaos at homoclinic bifurcations
37 Y. Meyer Wavelets and operators I
38 C. Weibel An introduction to homological algebra
39 W. Bruns & J. Herzog Cohen-Macaulay rings
40 V. Snaith Explicit Brauer induction
41 G. Laumon Cohomology of Drinfeld modular varieties I
42 E.B. Davies Spectral theory and differential operators
43 J. Diestel, H. Jarchow, & A. Tonge Absolutely summing operators
44 P. Mattila Geometry of sets and measures in Euclidean spaces
45 R. Pinsky Positive harmonic functions and diffusion
46 G. Tenenbaum Introduction to analytic and probabilistic number theory
47 C. Peskine An algebraic introduction to complex projective geometry
48 Y. Meyer & R. Coifman Wavelets
49 R. Stanley Enumerative combinatorics I
50 I. Porteous Clifford algebras and the classical groups
51 M. Audin Spinning tops
52 V. Jurdjevic Geometric control theory
53 H. Volklein Groups as Galois groups
54 J. Le Potier Lectures on vector bundles
55 D. Bump Automorphic forms and representations
56 G. Laumon Cohomology of Drinfeld modular varieties II
57 D.M. Clark & B.A. Davey Natural dualities for the working algebraist
58 J. McCleary A user’s guide to spectral sequences II
59 P. Taylor Practical foundations of mathematics
60 M.P. Brodmann & R.Y. Sharp Local cohomology
61 J.D. Dixon et al. Analytic pro-P groups
62 R. Stanley Enumerative combinatorics II
63 R.M. Dudley Uniform central limit theorems
64 J. Jost & X. Li-Jost Calculus of variations
65 A.J. Berrick & M.E. Keating An introduction to rings and modules
66 S. Morosawa Holomorphic dynamics
67 A.J. Berrick & M.E. Keating Categories and modules with K-theory in view
68 K. Sato Levy processes and infinitely divisible distributions
69 H. Hida Modular forms and Galois cohomology
70 R. Iorio & V. Iorio Fourier analysis and partial differential equations
71 R. Blei Analysis in integer and fractional dimensions
72 F. Borceaux & G. Janelidze Galois theories
73 B. Bollobas Random graphs
74 R.M. Dudley Real analysis and probability
75 T. Sheil-Small Complex polynomials
76 C. Voisin Hodge theory and complex algebraic geometry I
77 C. Voisin Hodge theory and complex algebraic geometry II
78 V. Paulsen Completely bounded maps and operator algebras
79 F. Gesztesy & H. Holden Soliton Equations and their Algebro-Geometric Solutions Volume 1
81 Shigeru Mukai An Introduction to Invariants and Moduli
82 G. Tourlakis Lectures in logic and set theory I
83 G. Tourlakis Lectures in logic and set theory II
84 R.A. Bailey Association Schemes
85 James Carlson, Stefan Müller-Stach, & Chris Peters Period Mappings and Period Domains
86 J.J. Duistermaat & J.A.C. Kolk Multidimensional Real Analysis I
87 J.J. Duistermaat & J.A.C. Kolk Multidimensional Real Analysis II
89 M. Golumbic & A.N. Trenk Tolerance Graphs
90 L.H. Harper Global Methods for Combinatorial Isoperimetric Problems
91 I. Moerdijk & J. Mrcun Introduction to Foliations and Lie Groupoids
92 János Kollár, Karen E. Smith, & Alessio Corti Rational and Nearly Rational Varieties
93 David Applebaum Lévy Processes and Stochastic Calculus
95 Martin Schechter An Introduction to Nonlinear Analysis

See http:www.cambridge.org for a complete list of books available in this series

Automorphic Forms and L-Functions for
the Group GL (n, R)

DORIAN GOLDFELD
Columbia University

With an Appendix by Kevin A. Broughan

University of Waikato
  
Cambridge, New York, Melbourne, Madrid, Cape Town, Singapore, São Paulo

Cambridge University Press

The Edinburgh Building, Cambridge  , UK
Published in the United States of America by Cambridge University Press, New York
www.cambridge.org
Information on this title: www.cambridge.org/9780521837712

© D. Goldfeld 2006

This publication is in copyright. Subject to statutory exception and to the provision of

relevant collective licensing agreements, no reproduction of any part may take place
without the written permission of Cambridge University Press.

First published in print format 2006

- ---- eBook (NetLibrary)

- --- eBook (NetLibrary)

- ---- hardback

- --- hardback

Cambridge University Press has no responsibility for the persistence or accuracy of s
for external or third-party internet websites referred to in this publication, and does not
guarantee that any content on such websites is, or will remain, accurate or appropriate.
Dedicated to Ada, Dahlia, and Iris
 Contents

Introduction page xi
1 Discrete group actions 1
1.1 Action of a group on a topological space 3
1.2 Iwasawa decomposition 8
1.3 Siegel sets 15
1.4 Haar measure 19
1.5 Invariant measure on coset spaces 23
1.6 Volume of S L(n, Z)\S L(n, R)/S O(n, R) 27
2 Invariant differential operators 38
2.1 Lie algebras 39
2.2 Universal enveloping algebra of gl(n, R) 42
2.3 The center of the universal enveloping algebra of gl(n, R) 46
2.4 Eigenfunctions of invariant differential operators 50
3 Automorphic forms and L–functions for S L(2, Z) 54
3.1 Eisenstein series 55
3.2 Hyperbolic Fourier expansion of Eisenstein series 59
3.3 Maass forms 62
3.4 Whittaker expansions and multiplicity one for G L(2, R) 63
3.5 Fourier–Whittaker expansions on G L(2, R) 67
3.6 Ramanujan–Petersson conjecture 68
3.7 Selberg eigenvalue conjecture 70
3.8 Finite dimensionality of the eigenspaces 71
3.9 Even and odd Maass forms 73
3.10 Hecke operators 74
3.11 Hermite and Smith normal forms 77
3.12 Hecke operators for L2 (S L(2, Z))\h2 80

vii
viii Contents

3.13 L–functions associated to Maass forms 84

3.14 L-functions associated to Eisenstein series 89
3.15 Converse theorems for S L(2, Z) 91
3.16 The Selberg spectral decomposition 94
4 Existence of Maass forms 99
4.1 The infinitude of odd Maass forms for S L(2, Z) 100
4.2 Integral operators 101
4.3 The endomorphism ♥ 105
4.4 How to interpret ♥: an explicit operator with purely
cuspidal image 106
4.5 There exist infinitely many even cusp forms for S L(2, Z) 108
4.6 A weak Weyl law 110
4.7 Interpretation via wave equation and the role of finite
propagation speed 111
4.8 Interpretation via wave equation: higher rank case 111
5 Maass forms and Whittaker functions for S L(n, Z) 114
5.1 Maass forms 114
5.2 Whittaker functions associated to Maass forms 116
5.3 Fourier expansions on S L(n, Z)\hn 118
5.4 Whittaker functions for S L(n, R) 128
5.5 Jacquet’s Whittaker function 129
5.6 The exterior power of a vector space 134
5.7 Construction of the Iν function using wedge products 138
5.8 Convergence of Jacquet’s Whittaker function 141
5.9 Functional equations of Jacquet’s Whittaker function 144
5.10 Degenerate Whittaker functions 150
6 Automorphic forms and L-functions for S L(3, Z) 153
6.1 Whittaker functions and multiplicity one for S L(3, Z) 153
6.2 Maass forms for S L(3, Z) 159
6.3 The dual and symmetric Maass forms 161
6.4 Hecke operators for S L(3, Z) 163
6.5 The Godement–Jacquet L-function 172
6.6 Bump’s double Dirichlet series 186
7 The Gelbart–Jacquet lift 194
7.1 Converse theorem for S L(3, Z) 194
7.2 Rankin–Selberg convolution for GL(2) 210
7.3 Statement and proof of the Gelbart–Jacquet lift 213
7.4 Rankin–Selberg convolution for G L(3) 223
 Contents ix

8 Bounds for L-functions and Siegel zeros 235

8.1 The Selberg class 235
8.2 Convexity bounds for the Selberg class 238
8.3 Approximate functional equations 241
8.4 Siegel zeros in the Selberg class 245
8.5 Siegel’s theorem 249
8.6 The Siegel zero lemma 251
8.7 Non-existence of Siegel zeros for Gelbart–Jacquet lifts 252
8.8 Non-existence of Siegel zeros on G L(n) 256
9 The Godement–Jacquet L-function 259
9.1 Maass forms for S L(n, Z) 259
9.2 The dual and symmetric Maass forms 261
9.3 Hecke operators for S L(n, Z) 266
9.4 The Godement–Jacquet L-function 277
10 Langlands Eisenstein series 285
10.1 Parabolic subgroups 286
10.2 Langlands decomposition of parabolic subgroups 288
10.3 Bruhat decomposition 292
10.4 Minimal, maximal, and general parabolic Eisenstein series 295
10.5 Eisenstein series twisted by Maass forms 301
10.6 Fourier expansion of minimal parabolic Eisenstein series 303
10.7 Meromorphic continuation and functional equation of
maximal parabolic Eisenstein series 307
10.8 The L-function associated to a minimal parabolic
Eisenstein series 310
10.9 Fourier coefficients of Eisenstein series twisted by
Maass forms 315
10.10 The constant term 319
10.11 The constant term of S L(3, Z) Eisenstein series twisted by
S L(2, Z)-Maass forms 321
10.12 An application of the theory of Eisenstein series to the
non-vanishing of L-functions on the line ℜ(s) = 1 322
10.13 Langlands spectral decomposition for S L(3, Z)\h3 324
11 Poincaré series and Kloosterman sums 337
11.1 Poincaré series for S L(n, Z) 337
11.2 Kloosterman sums 339
11.3 Plücker coordinates and the evaluation of Kloosterman sums 343
11.4 Properties of Kloosterman sums 350
x Contents

11.5 Fourier expansion of Poincaré series 352

11.6 Kuznetsov’s trace formula for S L(n, Z) 354
12 Rankin–Selberg convolutions 365
12.1 The GL(n) × GL(n) convolution 366
12.2 The GL(n) × GL(n + 1) convolution 372
12.3 The G L(n) × G L(n ′ ) convolution with n < n ′ 376
12.4 Generalized Ramanujan conjecture 381
12.5 The Luo–Rudnick–Sarnak bound for the generalized
Ramanujan conjecture 384
12.6 Strong multiplicity one theorem 393
13 Langlands conjectures 395
13.1 Artin L-functions 397
13.2 Langlands functoriality 402

List of symbols 407

Appendix The GL(n)pack Manual Kevin A. Broughan 409
A.1 Introduction 409
A.2 Functions for GL(n)pack 413
A.3 Function descriptions and examples 416
References 473
Index 485
 Introduction

The theory of automorphic forms and L-functions for the group of n × n invert-
ible real matrices (denoted G L(n, R)) with n ≥ 3 is a relatively new subject.
The current literature is rife with 150+ page papers requiring knowledge of a
large breadth of modern mathematics making it difficult for a novice to begin
working in the subject. The main aim of this book is to provide an essentially
self-contained introduction to the subject that can be read by someone with a
mathematical background consisting only of classical analysis, complex vari-
able theory, and basic algebra – groups, rings, fields. Preparation in selected
topics from advanced linear algebra (such as wedge products) and from the
theory of differential forms would be helpful, but is not strictly necessary for
a successful reading of the text. Any Lie or representation theory required is
developed from first principles.
This is a low definition text which means that it is not necessary for the reader
to memorize a large number of definitions. While there are many definitions,
they are repeated over and over again; in fact, the book is designed so that a
reader can open to almost any page and understand the material at hand without
having to backtrack and awkwardly hunt for definitions of symbols and terms.
The philosophy of the exposition is to demonstrate the theory by simple, fully
worked out examples. Thus, the book is restricted to the action of the discrete
group S L(n, Z) (the group of invertible n × n matrices with integer coefficients)
acting on G L(n, R). The main themes are first developed for S L(2, Z) then
repeated again for S L(3, Z), and yet again repeated in the more general case of
S L(n, Z) with n ≥ 2 arbitrary. All of the proofs are carefully worked out over
the real numbers R, but the knowledgeable reader will see that the proofs will
generalize to any local field. In line with the philosophy of understanding by
simple example, we have avoided the use of adeles, and as much as possible
the theory of representations of Lie groups. This very explicit language appears

xi
xii Introduction

particularly useful for analytic number theory where precise growth estimates
of L-functions and automorphic forms play a major role.
The theory of L-functions and automorphic forms is an old subject with roots
going back to Gauss, Dirichlet, and Riemann. An L-function is a Dirichlet series
∞

an
n=1
ns
where the coefficients an , n = 1, 2, . . . , are interesting number theoretic func-
tions. A simple example is where an is the number of representations of n as a
sum of two squares. If we knew a lot about this series as an analytic function of
s then we would obtain deep knowledge about the statistical distribution of the
values of an . An automorphic form is a function that satisfies a certain differ-
ential equation and also satisfies a group of periodicity relations. An example
is given by the exponential function e2πi x which is periodic (i.e., it has the
same value if we transform x → x + 1) and it satisfies the differential equa-
2
tion ddx 2 e2πi x = −4π 2 e2πi x . In this example the group of periodicity relations
is just the infinite additive group of integers, denoted Z. Remarkably, a vast
theory has been developed exposing the relationship between L-functions and
automorphic forms associated to various infinite dimensional Lie groups such
as G L(n, R).
The choice of material covered is very much guided by the beautiful paper
(Jacquet, 1981), titled Dirichlet series for the group G L(n), a presentation of
which I heard in person in Bombay, 1979, where a classical outline of the theory
of L-functions for the group G L(n, R) is presented, but without any proofs. Our
aim has been to fill in the gaps and to give detailed proofs. Another motivating
factor has been the grand vision of Langlands’ philosophy wherein L-functions
are akin to elementary particles which can be combined in the same way as
one combines representations of Lie groups. The entire book builds upon this
underlying hidden theme which then explodes in the last chapter.
In the appendix a set of Mathematica functions is presented. These have
been designed to assist the reader to explore many of the concepts and results
contained in the chapters that go before. The software can be downloaded by
going to the website given in the appendix.
This book could not have been written without the help I have received from
many people. I am particularly grateful to Qiao Zhang for his painstaking read-
ing of the entire manuscript. Hervé Jacquet, Daniel Bump, and Adrian Diaconu
have provided invaluable help to me in clarifying many points in the theory.
I would also like to express my deep gratitude to Xiaoqing Li, Elon Linden-
strauss, Meera Thillainatesan, and Akshay Venkatesh for allowing me to include
their original material as sections in the text. I would like to especially thank
 Introduction xiii

Dan Bump, Kevin Broughan, Sol Friedberg, Jeff Hoffstein, Alex Kontorovich,
Wenzhi Luo, Carlos Moreno, Yannan Qiu, Ian Florian Sprung, C. J. Mozzochi,
Peter Sarnak, Freydoon Shahidi, Meera Thillainatesan, Qiao Zhang, Alberto
Perelli and Steve Miller, for clarifying and improving various proofs, defini-
tions, and historical remarks in the book. Finally, Kevin Broughan has provided
an invaluable service to the mathematical community by creating computer
code for many of the functions studied in this book.

Dorian Goldfeld
 1
Discrete group actions

The genesis of analytic number theory formally began with the epoch making
memoir of Riemann (1859) where he introduced the zeta function,
∞


ζ (s) := n −s , (ℜ(s) > 1),
n=1

and obtained its meromorphic continuation and functional equation

s    ∞
1−s du
π −s/2
Ŵ ζ (s) = π −(1−s)/2
Ŵ ζ (1 − s), Ŵ(s) = e−u u s .
2 2 u
0

Riemann showed that the Euler product representation

 
1 −1
ζ (s) = 1− s ,
p p

together with precise knowledge of the analytic behavior of ζ (s) could be used
to obtain deep information on the distribution of prime numbers.
One of Riemann’s original proofs of the functional equation is based on the
Poisson summation formula



f (ny) = y −1 fˆ(ny −1 ),
n∈Z n∈Z

where f is a function with rapid decay as y → ∞ and

 ∞
fˆ(y) = f (t)e−2πit y dt,
−∞

is the Fourier transform of f . This is proved by expanding the periodic function

F(x) = f (x + n)
n∈Z

1
2 Discrete group actions

in a Fourier series. If f is an even function, the Poisson summation formula

may be rewritten as
∞ ∞


1
f (ny −1 ) = y fˆ(ny) − (y fˆ(0) − f (0)),
n=1 n=1
2

from which it follows that for ℜ(s) > 1,

 ∞  ∞
∞
dy dy
ζ (s) f (y)y s = f (ny)y s
0 y 0 n=1
y
 ∞
∞
 dy
= f (ny)y s + f (ny −1 )y −s
1 n=1
y
 ∞
∞
 dy 1 f (0) fˆ(0)
= f (ny)y s + fˆ(ny)y 1−s − + .
1 n=1
y 2 s 1−s

If f (y) and fˆ(y) have sufficient decay as y → ∞, then the integral above
converges absolutely for all complex s and, therefore, defines an entire function
of s. Let
 ∞
dy
f˜(s) = f (y)y s
0 y
denote the Mellin transform of f , then we see from the above integral rep-
resentation and the fact that fˆˆ(y) = f (−y) = f (y) (for an even function f )
that

ζ (s) f˜(s) = ζ (1 − s) f˜ˆ (1 − s).

2
Choosing f (y) = e−π y , a function with the property that it is invariant under
Fourier transform, we obtain Riemann’s original form of the functional equa-
tion. This idea of introducing an arbitrary test function f in the proof of the
functional equation first appeared in Tate’s thesis (Tate, 1950).
A more profound understanding of the above proof did not emerge until
2
much later. If we choose f (y) = e−π y in the Poisson summation formula, then
since fˆ(y) = f (y), one observes that for y > 0,
∞ ∞

2
y 1
2
e−πn =√ e−π n /y .
n=−∞ y n=−∞

This identity is at the heart of the functional equation of the Riemann zeta
function, and is a known transformation formula for Jacobi’s theta function
∞

2
θ(z) = e2πin z ,
n=−∞
 1.1 Action of a group on a topological space 3

 
a b
where z = x + i y with x ∈ R and y > 0. If is a matrix with integer
c d
coefficients a, b, c, d satisfiying ad − bc = 1, c ≡ 0 (mod 4), c = 0, then the
Poisson summation formula can be used to obtain the more general transfor-
mation formula (Shimura, 1973)
 
az + b 1
θ = ǫd−1 χc (d)(cz + d) 2 θ(z).
cz + d
Here χc is the primitive character of order ≤ 2 corresponding to the field exten-
1
sion Q(c 2 )/Q,

1 if d ≡ 1 (mod 4)
ǫd =
i if d ≡ −1 (mod 4),
1
and (cz + d) 2 is the “principal determination” of the square root of cz + d, i.e.,
the one whose real part is > 0.
It is now well understood that underlying the functional equation of the
Riemann zeta function are the above transformation formulae for θ(z). These
transformation
  formulae are induced from the action of a group of matrices
a b
on the upper half-plane h = {x + i y | x ∈ R, y > 0} given by
c d
az + b
z→ .
cz + d
The concept of a group acting on a topological space appears to be absolutely
fundamental in analytic number theory and should be the starting point for any
serious investigations.

1.1 Action of a group on a topological space

Definition 1.1.1 Given a topological space X and a group G, we say that G
acts continuously on X (on the left) if there exists a map ◦ : G → Func(X → X )
(functions from X to X ), g → g◦ which satisfies:
r x → g ◦ x is a continuous function of x for all g ∈ G;
r g ◦ (g ′ ◦ x) = (g · g ′ ) ◦ x, for all g, g ′ ∈ G, x ∈ X where · denotes the
internal operation in the group G;
r e ◦ x = x, for all x ∈ X and e = identity element in G.

Example 1.1.2 Let G denote the additive group of integers Z. Then it is easy
to verify that the group Z acts continuously on the real numbers R with group
4 Discrete group actions

action ◦ defined by

n ◦ x := n + x,

for all n ∈ Z, x ∈ R. In this case e = 0.

+
Example
  1.1.3 Let G = G L(2, R) denote the group of 2 × 2 matrices
a b
with a, b, c, d ∈ R and determinant ad − bc > 0. Let
c d
 
h := x + i y  x ∈ R, y > 0
 
a b
denote the upper half-plane. For g = ∈ G L(2, R)+ and z ∈ h define:
c d

az + b
g ◦ z := .
cz + d
Since

az + b ac|z|2 + (ad + bc)x + bd (ad − bc) · y

= +i ·
cz + d |cz + d|2 |cz + d|2

it immediately follows that g ◦ z ∈ h. We leave as an exercise to the reader, the

verification that ◦ satisfies the additional axioms of a continuous action. One
usually extends this action to the larger space h∗ = h ∪ {∞}, by defining
 
a b a/c if c = 0,
◦∞=
c d ∞ if c = 0.

Assume that a group G acts continously on a topological space X . Two

elements x1 , x2 ∈ X are said to be equivalent (mod G) if there exists g ∈ G
such that x2 = g ◦ x1 . We define
 
Gx := g ◦ x  g ∈ G

to be the equivalence class or orbit of x, and let G\X denote the set of equiva-
lence classes.

Definition 1.1.4 Let a group G act continuously on a topological space X .

We say a subset Ŵ ⊂ G is discrete if for any two compact subsets A, B ⊂ X ,
there are only finitely many g ∈ Ŵ such that (g ◦ A) ∩ B = φ, where φ denotes
the empty set.
 1.1 Action of a group on a topological space 5

Example 1.1.5 The discrete subgroup S L(2, Z). Let

   
a b 
Ŵ = S L(2, Z) := a, b, c, d ∈ Z, ad − bc = 1 ,
c d 
and let
   
1 m 
m∈Z
Ŵ∞ :=
0 1 

be the subgroup of Ŵ whichfixes ∞. Note that Ŵ∞ \Ŵ is just a set of coset

a b
representatives of the form where for each pair of relatively prime
c d
integers (c, d) = 1 we choose a unique a, b satisfying ad − bc = 1. This fol-
lows immediately from the identity
     
1 m a b a + mc b + md
· = .
0 1 c d c d
The fact that S L(2, Z) is discrete will be deduced from the following lemma.
Lemma 1.1.6 Fix real numbers 0 < r, 0 < δ < 1. Let Rr,δ denote the
rectangle
 
Rr,δ = x + i y  − r ≤ x ≤ r, 0 < δ ≤ y ≤ δ −1 .
Then for every ǫ > 0, and any fixed set S of coset representatives for
Ŵ∞ \S L(2, Z), there are at most 4 + (4(r + 1)/ǫδ) elements g ∈ S such that
Im(g ◦ z) > ǫ holds for some z ∈ Rr,δ .
 
a b
Proof Let g = . Then for z ∈ Rr,δ ,
c d
y
Im(g ◦ z) = <ǫ
c2 y 2 + (cx + d)2
1 1 1
if |c| > (yǫ)− 2 . On the other hand, for |c| ≤ (yǫ)− 2 ≤ (δǫ)− 2 , we have
y
<ǫ
(cx + d)2
if the following inequalities hold:
1 1
|d| > |c|r + (yǫ −1 ) 2 ≥ |c|r + (ǫδ)− 2 .
Consequently, Im(g ◦ z) > ǫ only if
1 1
|c| ≤ (δǫ)− 2 and |d| ≤ (ǫδ)− 2 (r + 1),
and the total number of such pairs (not counting (c, d) = (0, ±1), (±1, 0)) is at
most 4(ǫδ)−1 (r + 1).

6 Discrete group actions

It follows from Lemma 1.1.6 that Ŵ = S L(2, Z) is a discrete subgroup of

S L(2, R). This is because:

(1) it is enough to show that for any compact subset A ⊂ h there are only
finitely many g ∈ S L(2, Z) such that (g ◦ A) ∩ A = φ;
(2) every compact subset of A ⊂ h is contained in a rectangle Rr,δ for some
r > 0 and 0 < δ < δ −1 ;
(3) ((αg) ◦ Rr,δ ) ∩ Rr,δ = φ, except for finitely many α ∈ Ŵ∞ , g ∈ Ŵ∞ \Ŵ.

To prove (3), note that Lemma 1.1.6 implies that (g ◦ Rr,δ ) ∩ Rr,δ = φ except
for finitely many g ∈ Ŵ∞ \Ŵ. Let S ⊂ Ŵ∞ \Ŵ denote this finite set of such ele-
ments g. If g ∈ S, then Lemma 1.1.6 tells us that it is because Im(gz) < δ for all
z ∈ Rr,δ . Since Im(αgz) = Im(gz) for α ∈ Ŵ∞ , it is enough to show that for each
g ∈ S, there are only finitely many α ∈ Ŵ∞ such that ((αg) ◦ Rr,δ ) ∩ Rr,δ = φ.
This last statement follows from the fact that g ◦ Rr,δ itself
 lies in some other
1 m
rectangle Rr ′ ,δ′ , and every α ∈ Ŵ∞ is of the form α = (m ∈ Z), so
0 1
that
 −1 
α ◦ Rr ′ ,δ′ = x + i y  − r ′ + m ≤ x ≤ r ′ + m, 0 < δ ′ ≤ δ ′ ,

which implies (α ◦ Rr ′ ,δ′ ) ∩ Rr,δ = φ for |m| sufficiently large.

Definition 1.1.7 Suppose the group G acts continuously on a connected topo-

logical space X . A fundamental domain for G\X is a connected region D ⊂ X
such that every x ∈ X is equivalent (mod G) to a point in D and such that no
two points in D are equivalent to each other.

Example 1.1.8 A fundamental domain for the action of Z on R of

Example 1.1.2 is given by

Z\R = {0 ≤ x < 1 | x ∈ R}.

The proof of this is left as an easy exercise for the reader.

Example 1.1.9 A fundamental domain for S L(2, Z)\h can be given as the
region D ⊂ h where
  
 1 1
D = z  − ≤ Re(z) ≤ , |z| ≥ 1 ,

2 2

with congruent boundary points symmetric with respect to the imaginary axis.
 1.1 Action of a group on a topological space 7

-1 -1/2 0 1/2 1

 √ 
Note that the vertical line V ′ := − 21 + i y  y ≥ 23 is equivalent to the
 √ 
vertical line V := 21 + i y  y ≥ 23 under the transformation
 z → z + 1.
′ 1
Furthermore, the arc A :=  z − 2 ≤ Re(z) < 0, |z| = 1 is equivalent to

the reflected arc A := z  0 < Re(z) ≤ 1 , |z| = 1 , under the transformation
2
z → −1/z. To show that D is a fundamental domain, we must prove:

(1) For any z ∈ h, there exists g ∈ S L(2, Z) such that g ◦ z ∈ D;

(2) If two distinct points z, z ′ ∈ D are congruent (mod S L(2, Z)) then
Re(z) = ± 21 and z ′ = z ± 1, or |z| = 1 and z ′ = −1/z.

We first prove (1). Fix z ∈ h. It follows from Lemma 1.1.6 that for every
ǫ > 0, there are at most finitely many g ∈ S L(2, Z) such that g ◦ z lies in the
strip
  
 1 1
Dǫ := w  − ≤ Re(w) ≤ , ǫ ≤ Im(w) .
2 2

Let Bǫ denote the finite set of such g ∈ S L(2, Z). Clearly, for sufficiently small
ǫ, the set Bǫ contains at least one element. We will show that there is at least
one g ∈ Bǫ such that g ◦ z ∈ D. Among these finitely many g ∈ Bǫ ,choose one 
0 −1
such that Im(g ◦ z) is maximal in Dǫ . If |g ◦ z| < 1, then for S = ,
1 0
8 Discrete group actions

 
1 1
T = , and any m ∈ Z,
0 1
 
m −1 Im(g ◦ z)
Im(T Sg ◦ z) = Im = > Im(g ◦ z).
g◦z |g ◦ z|2
This is a contradiction because we can always choose m so that T m Sg ◦ z ∈ Dǫ .
So in fact, g ◦ z must be in D.
To complete the verification that D is a fundamental
  domain, it only remains
a b
to prove the assertion (2). Let z ∈ D, g = ∈ S L(2, Z), and assume
c d
that g ◦ z ∈ D. Without loss of generality, we may assume that
y
Im(g ◦ z) = ≥ Im(z),
|cz + d|2
(otherwise just interchange z and g ◦ z and√ use g −1 ). This implies that
|cz + d| ≤ 1 which implies that 1 ≥ |cy| ≥ 23 |c|. This is clearly impossi-
ble if |c| ≥ 2. So we only have to consider the cases c = 0, ±1. If c = 0
then d = ±1 and g is a translation by b. Since − 21 ≤ Re(z), Re(g ◦ z) ≤ 12 ,
this implies that either b = 0 and z = g ◦ z or else b = ±1 and Re(z) = ± 12
while Re(g ◦ z) = ∓ 12 . If c = 1, then |z + d| ≤ 1 implies that d = 0 unless
z = e2πi/3 and d = 0, 1 or z = eπi/3 and d = 0, −1. The case d = 0 implies
that |z| ≤ 1 which implies |z| = 1. Also, in this case, c = 1, d = 0, we
must have b = −1 because ad − bc = 1. Then g ◦ z = a − 1z . It follows that
a = 0. If z = e2πi/3 and d = 1, then we must have a − b = 1. It follows that
g ◦ e2πi/3 = a − 1+e12πi/3 = a + e2πi/3 , which implies that a = 0 or 1. A similar
argument holds when z = eπi/3 and d = −1. Finally, the case c = −1 can be
reduced to the previous case c = 1 by reversing the signs of a, b, c, d.

1.2 Iwasawa decomposition

This monograph focusses on the general linear group G L(n, R) with n ≥ 2.
This is the multiplicative group of all n × n matrices with coefficients in R
and non-zero determinant. We will show that every matrix in G L(n, R) can be
written as an upper triangular matrix times an orthogonal matrix. This is called
the Iwasawa decomposition (Iwasawa, 1949).
The Iwasawa decomposition, in the special case of G L(2, R), states that
every g ∈ G L(2, R) can be written in the form:
   
y x α β d 0
g= (1.2.1)
0 1 γ δ 0 d
 1.2 Iwasawa decomposition 9

where y > 0, x, d ∈ R with d = 0 and

 
α β
∈ O(2, R),
γ δ
where
 
O(n, R) = g ∈ G L(n, R)  g · gt = I
t
is the orthogonal group. Here I denotes the identity
 matrix on G L(n, R) and g
y x
denotes the transpose of the matrix g. The matrix in the decomposition
0 1
 
α β
(1.2.1) is actually uniquely determined. Furthermore, the matrices
γ δ
   
d 0 ±1 0
and are uniquely determined up to multiplication by .
0 d 0 ±1
Note that explicitly,
   
± cos t − sin t 
O(2, R) = 0 ≤ t ≤ 2π .
± sin t cos t 

We shall shortly give a detailed proof of (1.2.1) for G L(n, R) with n ≥ 2.

The decomposition (1.2.1) allows us to realize the upper half-plane
 
h = x + i y  x ∈ R, y > 0

as the set of two by two matrices of type

   
y x 
x ∈ R, y > 0 ,
0 1 
or by the isomorphism
 
h ≡ G L(2, R) O(2, R), Z 2 , (1.2.2)

where
⎧⎛ ⎞  ⎫
⎨ d 0
⎪  ⎪
 ⎬
⎜
Zn = ⎝ .. ⎟ 
⎠  d ∈ R, d = 0
⎪ .  ⎪
⎩ ⎭
0 d 

is the center of G L(n, R), and O(2, R), Z 2  denotes the group generated by
O(2, R) and Z 2 .
The isomorphism (1.2.2) is the starting point for generalizing the classical
theory of modular forms on G L(2, R) to G L(n, R) with n > 2. Accordingly,
we define the generalized upper half-plane hn associated to G L(n, R).
10 Discrete group actions

Definition 1.2.3 Let n ≥ 2. The generalized upper half-plane hn associated

to G L(n, R) is defined to be the set of all n × n matrices of the form z = x · y
where
⎛ ⎞ ⎛ ′ ⎞
1 x1,2 x1,3 · · · x1,n yn−1
′
⎜
⎜ 1 x2,3 · · · x2,n ⎟ ⎟
⎜
⎜ yn−2 ⎟
⎟
x =⎜
⎜ . .. . ⎟
.. ⎟ , y=⎜
⎜ . .. ⎟
⎟,
⎜ ⎟ ⎜ ⎟
′
⎝ 1 xn−1,n ⎠ ⎝
1y ⎠
1 1
yi′
with xi, j ∈ R for 1 ≤ i < j ≤ n and > 0 for 1 ≤ i ≤ n − 1.
To simplify later formulae and notation in this book, we will always express
y in the form:
⎛ ⎞
y1 y2 · · · yn−1
⎜
⎜ y1 y2 · · · yn−2 ⎟
⎟
y=⎜
⎜ .. ⎟
⎟,
⎜ . ⎟
⎝ y1 ⎠
1
with yi > 0 for 1 ≤ i ≤ n − 1. Note that this can always be done since yi′ = 0
for 1 ≤ i ≤ n − 1.
Explicitly, x is an upper triangular matrix with 1s on the diagonal and y
is a diagonal matrix beginning with a 1 in the lowest right entry. Note that x
is parameterized by n · (n − 1)/2 real variables xi, j and y is parameterized by
n − 1 positive real variables yi .
Example 1.2.4 The generalized upper half plane h3 is the set of all matrices
z = x · y with
⎛ ⎞ ⎛ ⎞
1 x1,2 x1,3 y1 y2 0 0
x = ⎝0 1 x2,3 ⎠ , y=⎝ 0 y1 0⎠,
0 0 1 0 0 1
where x1,2 , x1,3 , x2,3 ∈ R, y1 , y2 > 0. Explicitly, every z ∈ h3 can be written
in the form
⎛ ⎞
y1 y2 x1,2 y1 x1,3
z=⎝ 0 y1 x2,3 ⎠ .
0 0 1
Remark 1.2.5 The generalized upper half-plane h3 does not have a com-
plex structure. Thus h3 is quite different from h2 , which does have a complex
structure.
 1.2 Iwasawa decomposition 11

Proposition 1.2.6 Fix n ≥ 2. Then we have the Iwasawa decomposition:

G L(n, R) = hn · O(n, R) · Z n ,
i.e., every g ∈ G L(n, R) may be expressed in the form

g = z · k · d, (· denotes matrix multiplication)

where z ∈ hn is uniquely determined, k ∈ O(n, R), and d ∈ Z n is a non-zero

diagonal matrix which lies in the center of G L(n, R). Further, k and d are also
uniquely determined up to multiplication by ±I where I is the identity matrix
on G L(n, R).

Remark Note that for every n = 1, 2, 3, . . . , we have Z n ∼

= R× . We shall,
henceforth, write
hn ∼
= G L(n, R)/(O(n, R) · R× ).

Proof Let g ∈ G L(n, R). Then g · tg is a positive definite non–singular

matrix. We claim there exists u, ℓ ∈ G L(n, R), where u is upper triangular with
1s on the diagonal and ℓ is lower triangular with 1s on the diagonal, such that

u · g · tg = ℓ · d (1.2.7)

with
⎛ ⎞
d1
⎜
d=⎝ .. ⎟
⎠, d1 , . . . , dn > 0.
.
dn
 
a b
For example, consider n = 2, and g = . Then
c d
     
t a b a c a 2 + b2 ac + bd
g· g= · = .
c d b d ac + bd c2 + d 2
 
1 t
If we set u = , then u satisfies (1.2.7) if
0 1
   2   
1 t a + b2 ac + bd ∗ 0
· = ,
0 1 ac + bd c2 + d 2 ∗ ∗

so that we may take t = (−ac − bd)/(c2 + d 2 ). More generally, the upper

triangular matrix u will have n(n − 1)/2 free variables, and we will have to
12 Discrete group actions

solve n(n − 1)/2 equations to satisfy (1.2.7). This system of linear equations
has a unique solution because its matrix g · tg is non–singular.
It immediately follows from (1.2.7) that u −1 ℓd = g · tg = d · t ℓ(t u)−1 , or
equivalently

t
d · t ℓ%
"ℓ · d#$· u% = "u · #$ = d.
lower △ upper △

The above follows from the fact that a lower triangular matrix can only equal an
upper triangular matrix if it is diagonal, and that this diagonal matrix must be
d by comparing diagonal entries. The entries di > 0 because g · tg is positive
definite.
Consequently ℓd = d(t u)−1 . Substituting this into (1.2.7) gives

u · g · tg · t u = d = a −1 · (t a)−1

for
⎛ − 21
⎞
d1
⎜ .. ⎟
a=⎜
⎝ . ⎟.
⎠
− 12
dn

Hence aug · (tg · t u · t a) = I so that aug ∈ O(n, R). Thus, we have expressed
g in the form

g = (au)−1 · (aug),

from which the Iwasawa decomposition immediately follows after dividing and
−1
multiplying by the scalar dn 2 to arrange the bottom right entry of (au)−1 to
be 1.
It only remains to show the uniqueness of the Iwasawa decomposition.
Suppose that zkd = z ′ k ′ d ′ with z, z ′ ∈ hn , k, k ′ ∈ O(n, R), d, d ′ ∈ Z n . Then,
since the only matrices in hn and O(n, R) which lie in Z n are ±I where I is the
identity matrix, it follows that d ′ = ±d. Further, the only matrix in hn ∩ O(n, R)
is I . Consequently z = z ′ and k = ±k ′ .

We shall now work out some important instances of the Iwasawa decompo-
sition which will be useful later.
 1.2 Iwasawa decomposition 13

Proposition 1.2.8 Let I denote the identity matrix on G L(n, R), and for every
1 ≤ j < i ≤ n, let E i, j denote the matrix with a 1 at the {i, j}th position and
zeros elsewhere. Then, for an arbitrary real number t, we have
⎛ ⎞
1
⎜ . ⎟
⎜ .. ⎟
⎜ ⎟
⎜ 1 t ⎟
⎜ 1 ··· 2 1 ⎟
⎜ (t 2 + 1) 2 (t + 1) 2 ⎟ 
⎜
I + t E i, j = ⎜ .. .. ⎟ mod (O(n, R) · R× ) ,
⎟
⎜ . . ⎟
⎜ 1 ⎟
⎜ (t 2 + 1) 2 ⎟
⎜ ⎟
⎜ .. ⎟
⎝ . ⎠
1

1 1
where, in the above matrix, 1 occurs at position { j, j}, (t 2 + 1) 2 occurs
(t 2 + 1) 2
t
at position {i, i}, all other diagonal entries are ones, 1 occurs at position
(t 2 + 1) 2
{ j, i}, and, otherwise, all other entries are zero.

Proof Let g = I + t E i, j . Then

g · tg = (I + t E i, j ) · (I + t E j,i ) = I + t E i, j + t E j,i + t 2 E i,i .

If we define a matrix u = I − (t/(t 2 + 1))E j,i , then u · g · tg · t u must be a

diagonal matrix d. Setting d = a −1 · (t a)−1 , we may directly compute:

t2
u · g · tg · t u = I + t 2 E i,i − E j, j ,
t2 +1
t
u −1 = I + E j,i ,
t2 + 1
 
1 &
a −1 =I+ √ − 1 E j, j + t 2 + 1 − 1 E i,i .
t2 + 1

Therefore,
 
1 & t
u −1 a −1 = I + √ − 1 E j, j + t 2 + 1 − 1 E i,i + √ E j,i .
2
t +1 2
t +1

As in the proof of Proposition 1.2.6, we have g = u −1 · a −1

(mod (O(n, R), R× )).

14 Discrete group actions

Proposition 1.2.9 Let n ≥ 2, and let z = x y ∈ hn have the form given in

Definition 1.2.3. For i = 1, 2, . . . , n − 1, define
⎛ ⎞
1
⎜ .. ⎟
⎜
⎜ . ⎟
⎟
⎜ ⎟
⎜ 0 1 ⎟
ωi = ⎜ ⎟,
⎜ 1 0 ⎟
⎜ ⎟
⎜ .. ⎟
⎝ . ⎠
1

to be 
the n ×n identity matrix except for the ith and (i + 1)th rows where we
0 1
have on the diagonal. Then
1 0
⎛ ′ ′ ′ ⎞⎛ ⎞
1 x1,2 x1,3 ··· x1,n y1′ y2′ · · · yn−1
′
′ ′
⎜
⎜ 1 x2,3 ··· x2,n ⎟⎜
⎟⎜
′
y1′ y2′ · · · yn−2 ⎟
⎟
⎜
ωi z ≡ ⎜ .. .. ⎟⎜ .. ⎟
. . ⎟·⎜ . ⎟
⎜ ⎟⎜ ⎟
′
⎝ 1 xn−1,n ⎠⎝ y1′ ⎠
1 1

mod (O(n, R) · R× ) , where yk′ = yk except for k = n − i + 1, n − i, n − i − 1,
in which case
yn−i '
′ ′ 2 2
yn−i = 2 2
, yn−i±1 = yn−i±1 · xi,i+1 + yn−i ,
xi,i+1 + yn−i
′
and xk,ℓ = xk,ℓ except for ℓ = i, i + 1, in which case
2
′ ′ xi− j,i yn−i + xi− j,i+1 xi,i+1
xi− j,i = x i− j,i+1 − x i− j,i x i,i+1 , xi− j,i+1 = 2 2
,
xi,i+1 + yn−i

for j = 1, 2, . . . , i − 2.

Proof Brute force computation which is omitted.

Proposition 1.2.10 The group G L(n, Z) acts on hn .

Proof Recall the definition of a group acting on a topological space given in

Definition 1.1.1. The fact that G L(n, Z) acts on G L(n, R) follows immediately
from the fact that G L(n, Z) acts on the left on G L(n, R) by matrix multiplication
and that we have the realization hn = G L(n, R)/(O(n, R) · R× ), as a set of
cosets, by the Iwasawa decomposition given in Proposition 1.2.6.

 1.3 Siegel sets 15

1.3 Siegel sets

We would like to show that Ŵ n = G L(n, Z) acts discretely on the generalized
upper half-plane hn defined in Definition 1.2.3. This was already proved for
n = 2 in Lemma 1.1.6, but the generalization to n > 2 requires more subtle
arguments. In order to find an approximation to a fundamental domain for
G L(n, Z)\hn , we shall introduce for every t, u ≥ 0 the Siegel set t,u .

Definition 1.3.1 Let a, b ≥ 0 be fixed. We define the Siegel set a,b ⊂ hn to

be the set of all
⎛ ⎞ ⎛ ⎞
1 x1,2 x1,3 ··· x1,n y1 y2 · · · yn−1
⎜
⎜ 1 x2,3 ··· x2,n ⎟ ⎜
⎟ ⎜ y1 y2 · · · yn−2 ⎟
⎟
⎜ .. .. ⎟ ⎜ .. ⎟
⎜ . . ⎟·⎜ . ⎟
⎜ ⎟ ⎜ ⎟
⎝ 1 xn−1,n ⎠ ⎝ y1 ⎠
1 1

with |xi, j | ≤ b for 1 ≤ i < j ≤ n, and yi > a for 1 ≤ i ≤ n − 1.

Let Ŵ n = G L(n, Z) and Ŵ∞ n

⊂ Ŵ n denote the subgroup of upper triangular
matrices with 1s on the diagonal. We have shown in Proposition 1.2.10 that Ŵ n
acts on hn . For g ∈ Ŵ n and z ∈ hn , we shall denote this action by g ◦ z. The
following proposition proves that the action is discrete and that  √3 , 1 is a good
2 2
approximation to a fundamental domain.

Proposition 1.3.2 Fix an integer n ≥ 2. For any z ∈ hn there are only finitely
many g ∈ Ŵ n such that g ◦ z ∈  √3 , 1 . Furthermore,
2 2

(
G L(n, R) = g ◦  √3 , 1 . (1.3.3)
2 2
g∈Ŵ n
√
Remarks The bound 23 is implicit in the work of Hermite, and a proof can
be found in (Korkine and Zolotareff, 1873). The first part of Proposition 1.3.2
is a well known theorem due to Siegel (1939). For the proof, we follow the
exposition of Borel and Harish-Chandra (1962).

Proof of Proposition 1.3.2 In order to prove (1.3.3), it is enough to show that

(
S L(n, R) = g ◦  ∗√3 1 , (1.3.4)
2 ,2
g∈S L(n,Z)

∗
where t,u denotes the subset of matrices t,u · Z n which have determinant 1
∗ ∗
and ◦ denotes the action of S L(n, Z) on 0,∞ . Note that every element in a,b
16 Discrete group actions

is of the form
⎛ ⎞ ⎛ ⎞
1 x1,2 x1,3 ··· x1,n dy1 y2 · · · yn−1
⎜
⎜ 1 x2,3 ··· x2,n ⎟ ⎜
⎟ ⎜ dy1 y2 · · · yn−2 ⎟
⎟
⎜ .. .. ⎟ ⎜ . .. ⎟
⎜ . . ⎟·⎜ ⎟
⎜ ⎟ ⎜ ⎟
⎝ 1 xn−1,n ⎠ ⎝ dy1 ⎠
1 d
(1.3.5)
where the determinant
⎛ ⎞
dy1 y2 · · · yn−1
⎜
⎜ dy1 y2 · · · yn−2 ⎟
⎟
⎜
Det ⎜ .. ⎟
⎟ = 1,
⎜ . ⎟
⎝ dy1 ⎠
d
so that
n−1 −1/n

d= yin−i .
i=1

In view of the Iwasawa decomposition of Proposition 1.2.6, we may identify

∗
0,∞ as the set of coset representatives S L(n, R)/S O(n, R), where S O(n, R)
denotes the subgroup O(n, R) ∩ S L(n, R).

In order to prove (1.3.4), we first introduce some basic notation. Let

e1 = (1, 0, . . . , 0), e2 = (0, 1, . . . , 0), ..., en = (0, 0, . . . , 1),

denote the canonical basis for Rn . For 1 ≤ i ≤ n and any matrix g ∈

G L(n, R), let ei · g denote the usual multiplication of a 1 × n matrix with
an n × n'matrix. For an arbitrary v = (v1 , v2 , . . . , vn ) ∈ Rn , define the norm:
||v|| := v12 + v22 + · · · + vn2 . We now introduce a function

φ : S L(n, R) → R>0

from S L(n, R) to the positive real numbers. For all g = (gi, j )1≤i, j≤n in S L(n, R)
we define
'
2 2 2 .
φ(g) := ||en · g|| = gn,1 + gn,2 + · · · + gn,n

Claim The function φ is well defined on the quotient space

S L(n, R)/S O(n, R).
 1.3 Siegel sets 17

To verify the claim, note that for k ∈ S O(n, R), and v ∈ Rn , we have
& √ √
||v · k|| = (v · k) · t (v · k) = v · k · t k · t v = v · t v = ||v||.

This immediately implies that φ(gk) = φ(g), i.e., the claim is true.
∗
Note that if z ∈ 0,∞ is of the form (1.3.5), then

n−1 −1/n

φ(z) = d = yi(n−i) . (1.3.6)
i=1

∗
Now, if z ∈ 0,∞ is fixed, then

en · S L(n, Z) · z ⊂ Ze1 + · · · + Zen − {(0, 0, . . . , 0)} · z, (1.3.7)

where · denotes matrix multiplication. The right-hand side of (1.3.7) consists

of non–zero points of a lattice in Rn . This implies that φ achieves a positive
minimum on the coset S L(n, Z) · z. The key to the proof of Proposition 1.3.2
will be the following lemma from which Proposition 1.3.2 follows immediately.
∗
Lemma 1.3.8 Let z ∈ 0,∞ . Then the minimum of φ on S L(n, Z) ◦ z is
∗√
achieved at a point of  3 1 .
2 ,2

Proof It is enough to prove that the minimum of φ is achieved at a point of

 ∗√3 because we can always translate by an upper triangular matrix
2 ,∞

⎛ ⎞
1 u 1,2 u 1,3 ··· u 1,n
⎜
⎜ 1 u 2,3 ··· u 2,n ⎟
⎟
⎜
u=⎜ .. .. ⎟
. . ⎟ ∈ S L(n, Z)
⎜ ⎟
⎝ 1 u n−1,n ⎠
1

to arrange that the minimum of φ lies in  ∗√3 1 . This does not change the
2 ,2
value of φ because of the identity φ(u · z) = ||en · u · z|| = ||en · z||. We shall
use induction on n. We have already proved a stronger statement for n = 2
in Example 1.1.9. Fix γ ∈ S L(n, Z) such that φ(γ ◦ z) is minimized. We set
γ ◦ z = x · y with
⎛ ⎞ ⎛ ⎞
1 x1,2 x1,3 · · · x1,n dy1 y2 · · · yn−1
⎜
⎜ 1 x2,3 · · · x2,n ⎟
⎟
⎜
⎜ dy1 y2 · · · yn−2 ⎟
⎟
x =⎜
⎜ .. . ⎟
.. ⎟ , y = ⎜ ⎜ .. ⎟
. . ⎟,
⎜ ⎟ ⎜ ⎟
⎝ 1 xn−1,n ⎠ ⎝ dy1 ⎠
1 d
18 Discrete group actions

)n−1 n−i −1/n √

3
with d = ( i=1 yi ) as before. We must show yi ≥ 2
for i = 1, 2, . . . ,
n − 1. The proof proceeds in 3 steps.
√
3
Step 1 y1 ≥ 2
.
⎛ ⎞
In−2
This follows from the action of α := ⎝ 0 −1 ⎠ on γ ◦ z. Here In−2
1 0
denotes the identity (n − 2) × (n − 2)–matrix. First of all

φ(α ◦ γ ◦ z) = ||en · α ◦ γ ◦ z|| = ||en−1 · x · y|| = ||(en−1 + xn−1,n en ) · y||

'
= d y12 + xn−1,n
2
.

Since |xn−1,n | ≤ 12 we see that φ(αγ z)2 ≤ d 2 (y12 +41 ). On the other hand, the
assumption of minimality forces φ(γ z)2 = d 2 ≤ d 2 y12 + 14 . This implies that
√
3
y1 ≥ 2
.
 
′ g′ 1
Step 2 Let g ∈ S L(n − 1, Z), g = . Then φ(gγ z) = φ(γ z).
0 1
This follows immediately from the fact that en · g = en .
√
3
Step 3 yi ≥ 2
for i = 2, 3, . . . , n − 1.
 ′ ′ 
z ·d ∗
Let us write γ ◦ z = with z ′ ∈ S L(n − 1, R) and d ′ ∈ Z n−1
d
a suitable diagonal matrix. By induction, there exists g ′ ∈ S L(n − 1, Z) such
that g ′ ◦ z ′ = x ′ · y ′ ∈  ∗√3 1 ⊂ hn−1 , the Siegel set for G L(n − 1, R). This is
2 ,2
equivalent to the fact that
⎛ ⎞
an−1
⎜ an−2 ⎟
y′ = ⎜
⎜ ⎟
.. ⎟
⎝ . ⎠
a1

and
√
a j+1 3
≥ for j = 1, 2, . . . , n − 2. (1.3.9)
aj 2
 
g′ 0
Define g := ∈ S L(n, Z). Then
0 1
   ′ ′   
g′ 0 z ·d ∗ g′ ◦ z′ · d ′ ∗
g◦γ ◦z = ◦ = = x ′′ · y ′′ ,
0 1 d 0 d
 1.4 Haar measure 19

   
y′d ′ 0 x′ ∗
where y ′′ = , x ′′ = . The inequalities (1.3.9) applied to
0 d 0 1
⎛ ⎞
y1 y2 · · · yn−1 d

y′d ′ 0
 ⎜ .. ⎟
y ′′ =
⎜
=⎜ . ⎟
⎟,
0 d ⎝ y1 d ⎠
d
√
3
imply that yi ≥ 2
for i = 2, 3, . . . , n − 1. Step 2 insures that multiplying by
√
3
g on the left does not change the value of φ(γ z). Step 1 gives y1 ≥ 2
.

1.4 Haar measure

Let n ≥ 2. The discrete subgroup S L(n, Z) acts on S L(n, R) by left multipli-
cation. The quotient space S L(n, Z)\S L(n, R) turns out to be of fundamental
importance in number theory. Now, we turn our attention to a theory of inte-
gration on this quotient space.
We briefly review the theory of Haar measure and integration on locally
compact Hausdorff topological groups. Good references for this material are
(Halmos, 1974), (Lang, 1969), (Hewitt and Ross, 1979). Excellent introductary
books on matrix groups and elementary Lie theory are (Curtis, 1984), (Baker,
2002), (Lang, 2002).
Recall that a topological group G is a topological space G where G is also
a group and the map

(g, h) → g · h −1

of G × G onto G is continuous in both variables. Here · again denotes the

internal group operation and h −1 denotes the inverse of the element h. The
assumption that G is locally compact means that every point has a compact
neighborhood. Recall that G is termed Hausdorff provided every pair of distinct
elements in G lie in disjoint open sets.

Example 1.4.1 The general linear group G L(n, R) is a locally compact

Hausdorff topological group.

Let gl(n, R) denote the Lie algebra of G L(n, R). Viewed as a set, gl(n, R)
is just the set of all n × n matrices with coefficients in R. We assign a topology
20 Discrete group actions

to gl(n, R) by identifying every matrix

⎛ ⎞
g1,1 g1,2 ··· g1,n
⎜ g2,1 g2,2 ··· g2,n ⎟
⎜ ⎟
g=⎜ . .. ⎟
⎝ .. ··· . ⎠
gn,1 gn,2 ··· gn,n
with a point
2
(g1,1 , g1,2 , . . . , g1,n , g2,1 , g2,2 , . . . , g2,n , . . . , gn,n ) ∈ Rn .

This identification is a one–to–one correspondence. One checks that gl(n, R) is

a locally compact Hausdorff topological space under the usual Euclidean topol-
2
ogy on Rn . The determinant function Det : gl(n, R) → R is clearly continuous.
It follows that

G L(n, R) = gl(n, R) − Det−1 (0)

must be an open set since {0} is closed. Also, the operations of addition and
multiplication of matrices in gl(n, R) are continuous maps from

gl(n, R) × gl(n, R) → gl(n, R).

The inverse map

Inv : G L(n, R) → G L(n, R),

given by Inv(g) = g −1 for all g ∈ G L(n, R), is also continuous since each entry
of g −1 is a polynomial in the entries of g divided by Det(g). Thus, G L(n, R)
is a topological subspace of gl(n, R) and we may view G L(n, R) × G L(n, R)
as the product space. Since the multiplication and inversion maps: G L(n, R) ×
G L(n, R) → G L(n, R) are continuous, it follows that G L(n, R) is a topological
group.
By a left Haar measure on a locally compact Hausdorff topological group
G, we mean a positive Borel measure (Halmos, 1974)

µ : {measurable subsets of G} → R+ ,

which is left invariant under the action of G on G via left multiplication. This
means that for every measurable set E ⊂ G and every g ∈ G, we have

µ(g E) = µ(E).

In a similar manner, one may define a right Haar measure. If every left invariant
Haar measure on G is also a right invariant Haar measure, then we say that G
is unimodular.
 1.4 Haar measure 21

Given a left invariant Haar measure µ on G, one may define (in the usual
manner) a differential one-form dµ(g), and for compactly supported functions
f : G → C an integral

f (g) dµ(g),
G

which is characterized by the fact that


dµ(g) = µ(E)
E

for every measurable set E. We shall also refer to dµ(g) as a Haar measure.
The fundamental theorem in the subject is due to Haar.

Theorem 1.4.2 (Haar) Let G be a locally compact Hausdorff topological

group. Then there exists a left Haar measure on G. Further, any two such Haar
measures must be positive real multiples of each other.

We shall not need this general existence theorem, because in the situations
we are interested in, we can explicitly construct the Haar measure and Haar
integral. For unimodular groups, the uniqueness of Haar measure follows easily
from Fubini’s theorem. The proof goes as follows. Assume we have two Haar
measures µ, ν on G, which are both left and right invariant. Let h : G → C be
a compactly supported function satisfying

h(g) dµ(g) = 1.
G

Then for an arbitrary compactly supported function f : G → C,

  
f (g)dν(g) = h(g ′ )dµ(g ′ ) f (g)dν(g)
G
G  G

= h(g ′ ) f (g)dν(g)dµ(g ′ )
G G
 
= h(g ′ ) f (g · g ′ ) dν(g) dµ(g ′ )
G G
= h(g ′ ) f (g · g ′ ) dµ(g ′ ) dν(g)
G G
 
= h(g −1 · g ′ ) f (g ′ ) dµ(g ′ ) dν(g)
G G
 
= h(g −1 · g ′ ) f (g ′ ) dν(g) dµ(g ′ )
G G

=c· f (g ′ ) dµ(g ′ )
G
* −1
where c = G h(g )dν(g).
22 Discrete group actions

Proposition 1.4.3 For n = 1, 2, . . . , let

⎛ ⎞
g1,1 g1,2 · · · g1,n
⎜ g2,1 g2,2 · · · g2,n ⎟
⎜ ⎟
g=⎜ . .. ⎟ ∈ G L(n, R),
⎝ .. ··· . ⎠
gn,1 gn,2 ··· gn,n
where g1,1 , g1,2 , . . . , g1,n , g2,1 , . . . , gn,n are n 2 real variables. Define
)
dgi, j
1≤i, j≤n
dµ(g) := , (wedge product of differential one-forms)
Det(g)n
where dgi, j denotes the usual differential one–form on R and Det(g) denotes
the determinant of the matrix g. Then dµ(g) is the unique left–right invariant
Haar measure on G L(n, R).
Proof Every matrix in G L(n, R) may be expressed as a product of a diagonal
matrix in Z n and matrices of the form x̃r,s (with 1 ≤ r, s ≤ n) where x̃r,s denotes
the matrix with the real number xr,s at position r, s, and, otherwise, has 1s on
the diagonal and zeros off the diagonal. It is easy to see that
dµ(g) = dµ(ag)
for a ∈ Z n . To complete the proof, it is, therefore, enough to check that
dµ(x̃r,s · g) = dµ(g · x̃r,s ) = dµ(g),
for all 1 ≤ r, s ≤ n. We check the left invariance and leave the right invariance
to the reader.
It follows from the definition that in the case r = s,
⎛ ⎞
⎜ ) ⎟ )
⎝ dgi, j ⎠ d(gr, j + gs, j xr,s )
1≤i, j≤n 1≤ j≤n
i =r
dµ(x̃r,s · g) = .
Det(x̃r,s · g)n
First of all,
Det(x̃r,s · g) = Det(x̃r,s ) · Det(g) = Det(g)
because Det(x̃r,s ) = 1.
Second, for any 1 ≤ j ≤ n,
⎛ ⎞
⎜  ⎟
⎝ dgi, j ⎠ ∧ dgs, j = 0
1≤i, j≤n
i =r
 1.5 Invariant measure on coset spaces 23

⎛ ⎞
⎜ ) ⎟
because gs, j also occurs in the product ⎝ dgi, j ⎠ and dgs, j ∧ dgs, j = 0.
1≤i< j≤n
i =r
Consequently, the measure is invariant under left multiplication by x̃r,s .

On the other hand, if r = s, then

⎛ ⎞
⎜ ) ⎟ )
⎝ dgi, j ⎠ (xr,s · dgr, j )
1≤i, j≤n 1≤ j≤n
i =r
dµ(x̃r,s · g) =
Det(x̃r,s · g)n
n
xr,s
= dµ(g) ·
Det(x̃r,s )n
= dµ(g).

1.5 Invariant measure on coset spaces

This monograph focusses on the coset space
G L(n, R)/(O(n, R) · R× ).
We need to establish explicit invariant measures on this space. The basic prin-
ciple which allows us to define invariant measures on coset spaces, in general,
is given in the following theorem.
Theorem 1.5.1 Let G be a locally compact Hausdorff topological group, and
let H be a compact closed subgroup of G. Let µ be
* a Haar measure on G, and let
ν be a Haar measure on H , normalized so that H dν(h) = 1. Then there exists
a unique (up to scalar multiple) quotient measure µ̃ on G/H . Furthermore
   
f (g) dµ(g) = f (gh) dν(h) d µ̃(g H ),
G G/H H

for all integrable functions f : G → C.

Proof For a proof see (Halmos, 1974). We indicate, however, why the formula
in Theorem 1.5.1 holds. First of all note that if f : G → C, is an integrable
function on G, and if we define a new function, f H : G → C, by the recipe

f H (g) := f (gh) dν(h),
H

then f (gh) = f (g) for all h ∈ H. Thus, f H is well defined on the coset
H H

space G/H. We write f H (g) = f H (g H ), to stress that f H is a function on

24 Discrete group actions

the coset space. For any measurable subset E ⊂ G/H , we may easily choose
a measurable function δ E : G → C so that

1 if g H ∈ E,
δ E (g) = δ EH (g H ) =
0 if g H ∈ E.

We may then define an H –invariant quotient measure µ̃ satisfying:

 
µ̃(E) = δ E (g) dµ(g) = δ EH (g H ) d µ̃(g H ),
G G/H

and
 
f (g) dµ(g) = f H (g H ) d µ̃(g H ),
G G/H

for all integrable functions f : G → C.

Remarks There is an analogous version of Theorem 1.5.1 for left coset spaces
H \G. Note that we are not assuming that H is a normal subgroup of G. Thus
G/H (respectively H \G) may not be a group.

Example 1.5.2 (Left invariant measure on G L(n, R)/(O(n, R) · R× ))

For n ≥ 2, we now explicitly construct a left invariant measure on the

generalized upper half-plane hn = G L(n, R)/(O(n, R) · R× ). Returning to the
Iwasawa decomposition (Proposition 1.2.6), every z ∈ hn has a representation
in the form z = x y with
⎛ ⎞ ⎛ ⎞
1 x1,2 x1,3 · · · x1,n y1 y2 · · · yn−1
⎜
⎜ 1 x2,3 · · · x2,n ⎟⎟
⎜
⎜ y1 y2 · · · yn−2 ⎟
⎟
⎜
x =⎜ .. .. ⎟ , ⎜
y=⎜ .. ⎟
. . ⎟ . ⎟,
⎜ ⎟ ⎜ ⎟
⎝ 1 xn−1,n ⎠ ⎝ y1 ⎠
1 1

with xi, j ∈ R for 1 ≤ i < j ≤ n and yi > 0 for 1 ≤ i ≤ n − 1. Let d ∗ z denote

the left invariant measure on hn . Then d ∗ z has the property that

d ∗ (gz) = d ∗ z

for all g ∈ G L(n, R).

 1.5 Invariant measure on coset spaces 25

Proposition 1.5.3 The left invariant G L(n, R)–measure d ∗ z on hn can be

given explicitly by the formula
d∗z = d∗x d∗ y
where
 n−1

d∗x = d xi, j , d∗ y = yk−k(n−k)−1 dyk . (1.5.4)
1≤i< j≤n k=1

 
y x d xd y
For example, for n = 2, with z = , we have d ∗ z = y2
, while for
0 1
n = 3 with
⎛ ⎞
y1 y2 x1,2 y1 x1,3
z= ⎝ 0 y1 x2,3 ⎠ ,
0 0 1

we have
dy1 dy2
d ∗ z = d x1,2 d x1,3 d x2,3 .
(y1 y2 )3

Proof We sketch the proof. The group G L(n, R) is generated by diagonal

matrices, upper triangular matrices with 1s on the diagonal, and the Weyl group
Wn which consists of all n × n matrices with exactly one 1 in each row and
column and zeros everywhere else. For example,
   
1 0 0 1
W2 = , ,
0 1 1 0
⎧⎛ ⎞ ⎛ ⎞ ⎛ ⎞
⎨ 1 0 0 1 0 0 0 1 0
W3 = ⎝ 0 1 0 ⎠ , ⎝ 0 0 1 ⎠ , ⎝ 1 0 0 ⎠ ,
⎩
0 0 1 0 1 0 0 0 1
⎛ ⎞ ⎛ ⎞ ⎛ ⎞⎫
0 1 0 0 0 1 0 0 1 ⎬
⎝0 0 1⎠,⎝1 0 0⎠,⎝0 1 0⎠ .
⎭
1 0 0 0 1 0 1 0 0

Note that the Weyl group Wn has order n! and is simply the symmetric group on
n symbols. It is clear that d ∗ (gz) = d ∗ z if g is an upper triangular matrix with
1s on the diagonal. This is because the measures d xi, j (with 1 ≤ i < j ≤ n) are
all invariant under translation. It is clear that the differential d ∗ z is Z n -invariant
where Z n ∼= R× denotes the center of G L(n, R). So, without loss of generality,
26 Discrete group actions

we may define a diagonal matrix a with its lower-right entry to be one:

⎛ ⎞
a1 a2 · · · an−1
⎜
⎜ a1 a2 · · · an−2 ⎟
⎟
⎜
a=⎜ .. ⎟
⎟.
⎜ . ⎟
⎝ a1 ⎠
1

Then

az = ax y = (axa −1 ) · ay
⎛ ⎞
1 an−1 x1,2 an−1 an−2 x1,3 · · · an−1 · · · a1 x1,n
⎜
⎜ 1 an−2 x2,3 · · · an−2 · · · a1 x2,n ⎟
⎟
=⎜
⎜ .. .. ⎟
⎜ . . ⎟
⎟
⎝ 1 a1 xn−1,n ⎠
1
⎛ ⎞
a1 y1 · · · an−1 yn−1
⎜ .. ⎟
⎜
×⎜ . ⎟
⎟.
⎝ a1 y1 ⎠
1
n−1 
)
∗ −1
Thus d (axa ) = akk(n−k) d ∗ x. It easily follows that
k=1

d ∗ (az) = d ∗ (axa −1 · ay) = d ∗ z.

It remains to check the invariance of d ∗ z under the Weyl group Wn . Now, if

w ∈ Wn and
⎛ ⎞
dn
⎜
⎜ dn−1 ⎟
⎟
d=⎜ .. ⎟ ∈ G L(n, R)
⎝ . ⎠
d1

is a diagonal matrix, then wdw −1 is again a diagonal matrix whose diagonal

entries are a permutation of {d1 , d2 , . . . , dn }. The Weyl group is generated
by the transpositions ωi (i = 1, 2, . . . n − 1) given in Proposition 1.2.9 which
interchange (transpose) di and di+1 when d is conjugated by ωi . After a tedious
calculation using Proposition 1.2.9 one checks that d ∗ (ωi z) = d ∗ z.

 1.6 Volume of S L(n, Z)\S L(n, R)/S O(n, R) 27

1.6 Volume of S L(n, Z)\S L(n, R)/S O(n, R)

Following earlier work of Minkowski, Siegel (1936) showed that the volume
of

S L(n, Z)\S L(n, R)/S O(n, R) ∼

= S L(n, Z)\G L(n, R)/(O(n, R) · R× )
∼
= S L(n, Z)\hn ,
can be given in terms of

ζ (2) · ζ (3) · · · ζ (n)

where ζ (s) is the Riemann zeta function. The fact that the special values (taken
at integral points) of the Riemann zeta function appear in the formula for the
volume is remarkable. Later, Weil (1946) found another method to prove such
results based on a direct application of the Poisson summation formula. A vast
generalization of Siegel’s computation of fundamental domains for the case of
arithmetic subgroups acting on Chevalley groups was obtained by Langlands
(1966). See also (Terras, 1988) for interesting discussions on the history of this
subject.
The main aim of this section is to explicitly compute the volume

d ∗ z,
S L(n,Z)\S L(n,R)/S O(n,R)
∗
where d z is the left–invariant measure given in Proposition 1.5.3. We follow
the exposition of Garret (2002).

Theorem 1.6.1 Let n ≥ 2. As in Proposition 1.5.3, fix

 n−1

d∗z = d xi, j yk−k(n−k)−1 dyk
1≤i< j≤n k=1

to be the left S L(n, R)–invariant measure on hn = S L(n, R)/S O(n, R). Then
 n
 ζ (ℓ)
d ∗ z = n 2n−1 · ,
S L(n,Z)\hn ℓ=2
Vol(S ℓ−1 )

where
√
2( π)ℓ
Vol(S ℓ−1 ) =
Ŵ (ℓ/2)
∞
+
denotes the volume of the (ℓ − 1)–dimensional sphere S ℓ−1 and ζ (ℓ) = n −ℓ
n=1
denotes the Riemann zeta function.
28 Discrete group actions

Proof for the case of S L(2, R) We first prove the theorem for S L(2, R). The
more general result will follow by induction. Let K = S O(2, R) denote the
maximal compact subgroup of S L(2, R). We use the Iwasawa decomposition
which says that

  1
 ,

1 x y2 0
S L(2, R)/K ∼

= z = 1  x ∈ R, y > 0 .
0 1 0 y− 2 

Let f : R2 /K → C be an arbitrary smooth compactly supported function.

Then, by definition, f ((u, v) · k) = f ((u, v)) for all (u, v) ∈ R2 and all k ∈ K .
We can define a function F : S L(2, R)/K → C by letting

F(z) := f ((m, n) · z).
(m,n)∈Z2

 
a b
If γ = ∈ S L 2(Z), then
c d
   

a b
F(γ z) = f (m, n) · ·z
c d
(m,n)∈Z2


= f (ma + nc, mb + nd) · z
(m,n)∈Z2
= F(z).

Thus, F(z) is S L(2, Z)–invariant.

Note that we may express

-  .
{(m, n) ∈ Z2 } = (0, 0) ∪

ℓ · (0, 1) · γ  0 < ℓ ∈ Z, γ ∈ Ŵ∞ \S L(2, Z) ,
(1.6.2)

where
   
1 r 
 r ∈Z .
Ŵ∞ =
0 1 

We now integrate F over Ŵ\h2 , where h2 = S L(2, R)/K , Ŵ = S L(2, Z),

and d xd y/y 2 is the invariant measure on h2 given in Proposition 1.5.3. It
 1.6 Volume of S L(n, Z)\S L(n, R)/S O(n, R) 29

immediately follows from (1.6.2) that


d xd y
F(z) 2 = f ((0, 0)) · Vol(Ŵ\h2 )
y
Ŵ\h2



 d xd y
+ f ℓ(0, 1) · γ · z
ℓ>0 γ ∈Ŵ∞ \Ŵ
y2
Ŵ\h2

  d xd y
= f ((0, 0)) · Vol(Ŵ\h2 ) + 2 f ℓ(0, 1) · z .
ℓ>0
y2
Ŵ∞ \h2

 
The factor 2 occurs because −1 −1 acts trivially on h2 . We easily observe
that
1  
 y2 0 1
f ℓ(0, 1) · z = f ℓ(0, 1) · − 12
= f 0, ℓy − 2 .
0 y

It follows, after making the elementary transformations

y → ℓ2 y, y → y −2

that
 ∞
d xd y
F(z) 2 = f ((0, 0)) · Vol(Ŵ\h2 ) + 22 ζ (2) f ((0, y)) ydy. (1.6.3)
y
Ŵ\h2 0

Now, the function

 f ((u, v)) isinvariant under multiplication by k ∈ K on the
sin θ − cos θ
right. Since ∈ K , we see that
cos θ sin θ

f ((0, y)) = f ((y cos θ, y sin θ))

for any 0 ≤ θ ≤ 2π. Consequently

 ∞  2π  ∞
1
f ((0, y)) ydy = f ((y cos θ, y sin θ)) dθ ydy
0 2π 0 0

1
= f ((u, v)) dudv
2π
R2
1 ˆ
= f ((0, 0)). (1.6.4)
2π
30 Discrete group actions

Here fˆ denotes the Fourier transform of f in R2 . If we now combine (1.6.3)

and (1.6.4), we obtain

d xd y 2ζ (2) ˆ
F(z) 2 = f ((0, 0)) · Vol(Ŵ\h2 ) + f ((0, 0)). (1.6.5)
y π
Ŵ\h2

To complete the proof, we make use of the Poisson summation formula (see
appendix) which states that for any z ∈ G L(2, R)

1

F(z) = f ((m, n)z) = fˆ (m, n) · (t z)−1
(m,n)∈Z2
|Det(z)| (m,n)∈Z2


= fˆ (m, n) · (t z)−1 ,
(m,n)∈Z2

1 1
y2 y− 2 x
since z = 1 and Det(z) = 1. We now repeat all our computations
0 y− 2
with the roles of f and fˆ reversed. Since the group Ŵ is stable under transpose–
inverse, one easily sees (from the Poisson summation formula above), by letting
z → (t z)−1 , that the integral

d xd y
F(z) 2
y
Ŵ\h2

is unchanged if we replace f by fˆ.

Also, since fˆˆ (x) = f (−x), the formula (1.6.5) now becomes

d xd y 2ζ (2)
F(z) 2 = fˆ((0, 0)) · Vol(Ŵ\h2 ) + f ((0, 0)). (1.6.6)
y π
Ŵ\h2

If we combine (1.6.5) and (1.6.6) and solve for the volume, we obtain
  2ζ (2)
f ((0, 0)) − fˆ((0, 0)) · vol(Ŵ\h2 ) = f ((0, 0)) − fˆ((0, 0)) · .
π
Since f is arbitrary, we can choose f so that f ((0, 0)) − fˆ((0, 0)) = 0. It
follows that
2ζ (2) π
Vol(Ŵ\h2 ) = = .

π 3
Proof for the case of S L(n, R) We shall now complete the proof of
Theorem 1.6.1 using induction on n.

The proof of Theorem 1.6.1 requires two preliminary lemmas which we

straightaway state and prove. For n > 2, let Un (R) (respectively Un (Z)) denote
 1.6 Volume of S L(n, Z)\S L(n, R)/S O(n, R) 31

the group of all matrices of the form

⎛ ⎞
1 u1
⎜ .. .. ⎟
⎜
⎜ . . ⎟
⎟
⎝ 1 u n−1 ⎠
1
with u i ∈ R (respectively, u i ∈ Z), for i = 1, 2, . . . , n − 1.
Lemma 1.6.7 Let n > 2 and fix an element γ ∈ S L(n − 1, Z). Consider
n−1
  of Un (Z) on R
the action given by left matrix multiplication of Un (Z) on
γ 0
· Un (R). Then a fundamental domain for this action is given by the
0 1
set of all matrices
⎛ ⎞
1 u1
  ⎜ .. .. ⎟
γ 0 ⎜ . . ⎟
·⎜ ⎟
0 1 ⎝ 1 u n−1 ⎠
1
with 0 ≤ u i < 1 for 1 ≤ i ≤ n − 1. In particular,
/ 
γ 0
Un (Z) · Un (R) ∼
= (Z\R)n−1 .
0 1

Proof of Lemma 1.6.7 Let m be a column vector with (m 1 , m 2 , . . . , m n−1 ) as

entries. Then one easily checks that
       
In−1 m γ γ I γ −1 m
· = · n−1 ,
1 1 1 1
where In−1 denotes the (n − 1) × (n − 1) identity matrix. It follows that
(  In−1 m  γ

I (Z\R)n−1

· · n−1
1 1 1
m∈Zn−1
     
( γ In−1 γ −1 m In−1 (Z\R)n−1
= · ·
1 1 1
m∈Zn−1
  (  
γ In−1 (Z\R)n−1 + γ −1 m
= ·
1 1
m∈Zn−1
 
γ
= · Un (R).
1
It is also clear that the above union is over non-overlapping sets. This is because
γ −1 Zn−1 = Zn−1 for γ ∈ S L(n − 1, Z).

The second lemma we need is a generalization of the identity (1.6.4).

32 Discrete group actions

Lemma 1.6.8 Let n > 2 and let f : Rn → C be a smooth function, with

sufficient decay at ∞, which satisfies f (u 1 , . . . , u n ) = f (v1 , . . . , vn ) whenever
u 21 + · · · + u 2n = v12 + · · · + vn2 . Then
 ∞ 
1
f (0, . . . , 0, t) t n−1 dt = f (x1 , . . . , xn ) d x1 · · · d xn
0 Vol(S n−1 ) Rn
fˆ(0)
= ,
Vol(S n−1 )
where
√
n−1 2( π )n
Vol(S )=
Ŵ (n/2)

denotes the volume of the (n − 1)–dimensional sphere S n−1 .

Proof of Lemma 1.6.8 For n ≥ 2 consider the spherical coordinates:
x1 = t · sin θn−1 · · · sin θ2 sin θ1 ,
x2 = t · sin θn−1 · · · sin θ2 cos θ1 ,
x3 = t · sin θn−1 · · · sin θ3 cos θ2 , (1.6.9)
..
.
xn−1 = t · sin θn−1 cos θn−2 ,
xn = t · cos θn−1 ,
with
0 < t < ∞, 0 ≤ θ1 < 2π, 0 ≤ θ j < π, (1 < j < n).
Clearly x12 + · · · + xn2 = t 2 . One may also show that the invariant measure on
the sphere S n−1 is given by

dµ(θ) = (sin θ j ) j−1 dθ j ,
1≤ j<n

and that d x1 d x2 · · · d xn = t n−1 dt dµ(θ). Then the volume of the unit sphere,
Vol(S n−1 ), is given by
 √
n−1 2( π)n
Vol(S ) = dµ(θ) = .
S n−1 Ŵ (n/2)
Since f is a rotationally invariant function, it follows that

1
f (0, . . . , 0, t) = f (x1 , . . . , xn ) dµ(θ)
Vol(S n−1 ) S n−1
 1.6 Volume of S L(n, Z)\S L(n, R)/S O(n, R) 33

with x1 , . . . , xn given by (1.6.9). Consequently

 ∞  ∞
1
f (0, . . . , 0, t) t n−1 dt = f (x1 , . . . , xn )t n−1 dµ(θ)dt
0 Vol(S n−1 ) 0 S n−1

1
= f (x1 , . . . , xn ) d x1 · · · d xn .
Vol(S n−1 ) Rn

We now return to the proof of Theorem 1.6.1. Let K n = S O(n, R) denote

the maximal compact subgroup of S L(n, R). In this case, the Iwasawa decom-
position (Proposition 1.2.6) says that every z ∈ S L(n, R)/K n is of the form
z = x y with
⎛ ⎞
1 x1,2 x1,3 ··· x1,n
⎜
⎜ 1 x2,3 ··· x2,n⎟
⎟
⎜
x=⎜ .. .. ⎟
. . ⎟,
⎜ ⎟
⎝ 1 xn−1,n ⎠
1
⎛ ⎞ (1.6.10)
y1 y2 · · · yn−1 t
⎜
⎜ y1 y2 · · · yn−2 t ⎟
⎟
⎜
y=⎜ .. ⎟
⎟,
⎜ . ⎟
⎝ y1 t ⎠
t

n−1 −1/n
)
with t = Det(y)−1/n = yin−i .
i=1
In analogy to the previous proof for S L(2, R) we let f : Rn /K n → C be an
arbitrary smooth compactly supported function. We shall also define a function
F : S L(n, R)/K n → C by letting


F(z) := f (m · z).
m∈Zn

As before, the function F(z) will be invariant under left multiplication by

S L(n, Z).
Let
 
∗
Pn = ∈ S L(n, Z)
0 0 ··· 0 1
34 Discrete group actions

denote the set of all n × n matrices in S L(n, Z) with last row (0, 0, . . . , 0, 1).
Let Ŵn = S L(n, Z). Then we have as before:



F(z) = f (0) + f (ℓen · γ · z),
0<ℓ∈Z γ ∈Pn \Ŵn

where f (0) denotes f ((0, 0, . . . , 0)) and en = (0, 0, . . . , 0, 1).

We now integrate F(z) over a fundamental domain for Ŵn \hn . It follows that


∗ n
F(z) d z = f (0) · Vol(Ŵn \h ) + 2 f (ℓen · z) d ∗ z. (1.6.11)
ℓ>0
Ŵn \hn Pn \hn

The factor 2 occurs because −In (In = n × n identity matrix) acts trivially on
hn . The computation of the integral above requires some preparations.
We may express z ∈ hn in the form
⎛ ⎞
y1 y2 · · · yn−1 t
⎜
⎜ y1 y2 · · · yn−2 t ⎟ 
⎟ 1   − 1 
⎜ .. ⎟ t n − 1 · In−1 t n − 1 · In−1
z = x ·⎜ . ⎟· · ,
⎜ ⎟ t −1 t
⎝ y1 t ⎠
t

where x and t are given by (1.6.10). It follows that

⎛ ⎞ ⎛ ⎞
1 x1,n 1 x1,2 x1,3 ··· x1,n−1 0
⎜
⎜ 1 x2,n ⎟ ⎜
⎟ ⎜ 1 x2,3 ··· x2,n−1 0⎟ ⎟
z=⎜
⎜ . . .
. ⎟ ⎜
·⎜ .. .. .. ⎟
⎜ . . ⎟ ⎟ ⎜ . . .⎟⎟
⎝ 1 xn−1,n ⎠ ⎝ 1 0⎠
1 1
⎛ n ⎞
y1 y2 · · · yn−1 · t n−1
n
⎜ y1 y2 · · · yn−2 · t n − 1 ⎟ 
1 
⎜ ⎟
⎜ .. ⎟ t − n − 1 · In−1
×⎜ . ⎟·
⎜ n
⎟ t
⎝ y1 · t n − 1 ⎠
1
⎛ ⎞
1 x1,n
⎜ 1 x2,n ⎟   1
z′ t − n − 1 · In−1
⎜ ⎟
⎜
=⎜ .. .. ⎟
. . ⎟· · ,
⎜ ⎟ 1 t
⎝ 1 xn−1,n ⎠
1
(1.6.12)
 1.6 Volume of S L(n, Z)\S L(n, R)/S O(n, R) 35

where
⎛ ⎞
1 x1,2 x1,3 ··· x1,n−1
⎜
⎜ 1 x2,3 ··· x2,n−1 ⎟
⎟
⎜
z =⎜′ .. .. ⎟
⎜ . . ⎟
⎟
⎝ 1 xn−2,n−1 ⎠
1
⎛ n ⎞
y1 y2 · · · yn−1 · t n−1
n
⎜
⎜ y1 y2 · · · yn−2 · t n−1 ⎟
⎟
×⎜ .. ⎟.
⎝ . ⎠
n
y1 · t n−1

Now z ′ represents the Iwasawa coordinate for S L(n − 1, R)/S O(n − 1, R) =

hn−1 , and the Haar measure d ∗ z ′ can be computed using Proposition 1.5.3 and
is given by
 n−2
 −k(n−1−k)−1
d ∗ z′ = d xi, j yk+1 dyk+1 .
1≤i< j≤n−1 k=1

If we compare this with

 n−1

d∗z = d xi, j yk−k(n−k)−1 dyk
1≤i< j≤n k=1

 n−2
 −(k+1)(n−1−k)−1
= d xi, j yk+1 dyk ,
1≤i< j≤n k=0

we see that
n−1
 dy1
d ∗ z = d ∗ z′ d x j,n t n . (1.6.13)
j=1
y1

Here, the product of differentials is understood as a wedge product satisfying

the usual rule: du ∧ du = 0, given by the theory of differential forms. Since
n−1

−(n−1)/n −(n−i)/n
t = y1 yi ,
i=2

we see that
dt n − 1 dy1
=− + ,
t n y1
where  is a differential form involving dy j for each j = 2, 3, . . . , n − 1, but
Another random document with
no related content on Scribd:
party inside. These signs, as they could easily understand them,
excited defiance.
They believed that in the few words exchanged at the door of the
carriage, a project for a rescue had been broached; they closed
round the carriage, and the words “Be on the alert!” circulated
through the ranks of the National Guards.
M. de Dampierre tried to approach the carriage once again, and
was repulsed, not only with murmurs, but with menaces; the guards
crossing their muskets across the door to prevent his holding any
communication with the King.
This almost insolent persistence on his part had exasperated even
the most temperate.
Seeing that his efforts were useless, M. de Dampierre resolved to
finish with an act of bravado.
Having accomplished two-thirds of the descent, at a spot called La
Grevières, M. de Dampierre called out a second time “Vive le Roi!”
fired off his gun in the air, and plunging his rowels into his steed,
darted off at a gallop.
A wood was situated about half a league from the road. They
believed that some troops were in ambuscade there, and that the
discharge of the gun was a preconcerted signal for them.
Five or six horsemen dashed off in pursuit of M. de Dampierre; ten
or twelve shots were fired at him at the same time, but none of the
bullets touched him.
M. de Dampierre, still at a gallop, waved his arm in a triumphant
manner in the air.
I rushed off like the others, though on foot, not to capture M. de
Dampierre—heaven forbid!—but, on the contrary, to help him if
needful.
M. de Dampierre had already galloped more than five hundred
yards, and he had almost escaped from his pursuers, when his
horse, in leaping a ditch, stumbled, and fell.
But, with the aid of the bit and bridle, he managed to raise him up
again, and once more set off at a gallop. His gun was left in the
ditch.
At this moment a solitary gun was discharged.
It was fired by a peasant, mounted on a horse belonging to one of
the hussars, which he had captured the evening before.
It was easy to see that M. de Dampierre was wounded. He fell
backwards on the croup of his horse, which reared.
Then, in a moment, with the rapidity of lightning, on the little
bridge of St. Catherine, by the borders of the ditch, the waters of
which pass under the bridge, a horrible scene took place, which I
saw in all its dreadful details, but was unable to oppose.
The peasant who had fired the shot, followed by about forty men,
caught up the Comte de Haus, dealt him a blow with his sabre, and
then unhorsed him. I saw no more. I heard the report of about
twenty guns, into the suffocating smoke of which I dashed.
They were firing at M. de Dampierre.
I arrived too late. Had I reached the mob sooner, it would have
been to have died with him, for I could not have saved him.
His body was riddled with bullets, and gashed with bayonets; his
face, scratched by the peasants’ hob-nailed boots, was
unrecognisable.
His watch was dashed to pieces by a ball which had penetrated his
fob.
There was nothing to be done. I threw my gun over my shoulder,
and, with tears in my eyes and sweat on my brow, I rejoined my
rank.
The royal berlin continued its route slowly and sorrowfully under a
sweltering sun, along that unbending route which crosses like a
pencil line that sorrowful portion of France called the Paltry Land.
 C H A P T E R X X X I V.
THE CRITICS CRITICISED.

Not only does it seem to me sufficient to relate what I have seen;

I desire also, as an eye-witness, to rectify history and to combat, on
sure grounds, the mistake of historians.
In order to give a slight idea of the intense excitement of
Republican France against the King, and more particularly against
the Queen, I quote the following letter, the original of which was
sent by the citizens of Counien to the municipal officers of Varennes:
—

“27th of June, 2nd year of Liberty.

“G entlemen ,—
“Allow the patriotic women of the State, who have the honor of
being members of the Club of the Society of the Friends of the
Constitution, to present to you their best congratulations at the
capture in your city of the execrable traitors, Citizen and Citizeness
Capet, whose traitorous machinations have so long tried to crush
freedom in France. Our only prayer is that both may speedily be
humiliated. Vive la France! Vive la Liberté! A bas les Capets!
“For the Citizenesses of Counien,
(Signed) “Citoyenne Marie Benoit.
“To the Municipal Officers of Varennes.”

After this specimen of the feeling of the women of France can it

be wondered that the fate of the King and Queen seemed assured.
Either they must die, or France must sink lower than ever. This, of
course, was only my opinion; but events have proved whether I was
right or wrong.
 C H A P T E R X X X V.
IS LOVE ETERNAL?

The route from St. Menehould to Châlons is long and fatiguing—

nine apparently never-ending leagues, traversing flat and arid plains
under a leaden sky, with a sun darting his scorching rays with
reflected lustre on the musket barrels and sword blades.
The royal family arrived at Châlons fatigued, dispirited and worn
out, at ten in the evening.
Half the original followers of the royal escort had thrown
themselves down under hedges and in ditches, unable any longer to
proceed.
But the actual escort was as strong on arriving at Châlons as when
leaving St. Menehould, since it was recruited by the National Guard
of every village through which it passed; and the villages were pretty
thickly scattered on the right and left of that road.
The authorities, of whom the Mayor took the lead, conducted the
prisoners to the gate of the Dauphin. I use the word prisoners
advisedly, as the royal family were in fact, at that time, prisoners of
the nation.
Strange coincidence!—the gate through which they passed was
the triumphal arch raised by the French people in commemoration of
the entry of Madame the Dauphine into France.
It still bore the inscription, “May it stand eternal like our love.”
The arch, in fact, still stood, but the love which prompted it had
fallen away.
At Châlons, especially, opinion changes.
 The bluffness of the national party was lessened. The old town
where Attila lost himself, and which now preserved its trade only in
the wines of Champagne, was inhabited by Royalists of the better
class, and by poor gentlemen. These good people were sorely vexed
to see their unhappy King in such doleful plight.
They expected his arrival; consequently a great supper was
prepared.
The King and Queen partook of the meal in public, as they did at
Varennes. A sort of royal drawing-room was held. The ladies bore
with them immense bouquets. The Queen was positively
overwhelmed with flowers.
They determined to start on the morrow, feeling an increase of
confidence on account of the reception they had met with.
Before they set out, mass was celebrated at ten o’clock by M.
Charber, perpetual-curate of Notre Dame. The King was present,
accompanied by the Queen and the royal family; but hardly had the
solemn service commenced, before a disturbance was made.
It was the National Guard of Rheims, who wished the King to set
out at once. The time spent in mass appeared to them wasted, as
they had come solely to gloat over the downfall of monarchy, and
the ruin of their King. They broke open, therefore, the doors of the
chapel, despite the resistance offered to them by the National
Guard.
The King and Queen were advised to show themselves at the
balcony. They did so; but the sight of their august persons
exasperated, in place of calming, the turbulence of the excited
populace, who shouted for the royal family to leave their city, and
actually drew the carriages to the door, harnessed the horses, and
did, in fact, all they could do to accelerate the departure of the King.
The King appeared again at the balcony, and pronounced the
following words:—
“Since you oblige me to leave you, I go!”
 Although this was a reproach more than anything else, it satisfied
the people.
At eleven exactly, the royal family re-entered their carriage, and
put themselves en route.
The heat was dreadfully oppressive. Their journey was made, as it
were, through a blast furnace, and their eyes were incessantly
tormented by a penetrating dust.
I happened to know the situation of a cool spring. I approached
the royal carriage, and demanded respectfully of her Majesty the
Queen if she desired a glass of fresh water, as we were near to
some of a most excellent quality.
“Thank you,” replied the Queen.
“Oh, do have some, mamma—do have some! I am so thirsty!” said
the Dauphin.
“I wish it not; but give me some for my children,” said the Queen.
Madame de Tourzel handed me a silver cup.
“Fill this up for me, also,” said Madame Elizabeth.
She handed me another cup.
In fancy, after a lapse of sixty years, I can still see her angelic face
—still hear that charming voice, whose entreaties were more than
commands.
I leant my gun against the trunk of a tree, rushed to the fountain,
and brought back the two cups filled with the sparkling water, which,
through my rapidity, had not had time to lose its freshness.
The Dauphin and Madame Royale shared one cup between them.
Madame Elizabeth, after offering the other cup to the Queen, who
refused it, drank it herself.
“Oh, what delicious water it is,” cried the Dauphin. “Why does the
world drink aught else?”
 “Because they have drinks they like better,” replied the King.
“My son thanks you, sir,” said the Queen.
“I also thank you,” said Madame Elizabeth, with her sweet smile.
I seized my gun, which had been left at the foot of a tree.
“I saw you once run after M. de Dampierre,” said the Queen.
“With what intention?”
“With the hope of saving him, if possible, madame.”
“You have the same opinions, then, as M. de Dampierre,” said the
Queen.
“I agree with him in the respect which he feels towards your
enemies.”
“Do you know that you give an ambiguous answer, young man?”
said the King.
“Yes, sire,” I replied.
“Ha, ha!” said he.
Then to the Queen: “The minds of these people are poisoned
against us, from their very childhood.”
“Oh, papa!” cried the Dauphin, “what a beautiful gun he has!”
I was the person referred to. To the Queen and Madame Elizabeth
I was “monsieur,” but to the Dauphin I was simply “he.”
The King looked at my gun.
“It is,” said he, “a gun manufactured at Versailles. Where did you
procure it?”
“The Duc d’Enghien gave it to me, sire.”
“Yes,” said the King, “the Condés have all the benefits on this side,
—‘the department of the Meuse,’ as they call it.”
Then looking towards me: “Have you ever served princes?”
 “Sire,” said I, smiling, “is it necessary to have served princes in
order to receive a present from them?”
The Queen bent her regards upon the King.
“Strange!” said she.
I retreated a pace.
The King beckoned me, but not knowing how to address me, he
said, “My young friend, you say that the Duc d’Enghien gave you
that gun?”
“Yes, sire. I understand,” said I, “that the King wishes to know
upon what occasion this gun was presented to me. I was the
nephew of a park-keeper of the Forest of Argonne, whose name was
Father Descharmes. The Duc de Condé and the Duc d’Enghien often
hunted in this forest. The Duc d’Enghien took a fancy to me, and
gave me this gun.”
The King, for a moment, appeared buried in thought.
“Your uncle is still alive?” he then asked.
“Sire, he is dead?”
“Why do you not solicit his place?”
“Because, sire, the keepers wear livery. I am a free man.”
“Children suck in republican ideas, even with their mothers’ milk!”
murmured the King.
He then threw himself back in his carriage.
I know not if the King spoke again; but the carriage at the
moment stopped; and, perchance, with it stopped something of
importance.
We had arrived at Port Bassion.
Suddenly was heard a cry: “The commissaires! the commissaires!”
At this moment, a man on horseback dashed up to the door of the
King’s carriage. The King put out his head, to see what had caused
the halt.
“Sire,” said the horseman, “here are three deputies, who wish to
direct the return of your Majesty.”
“Aha!” said the King, “Can you tell me the names of these
estimable gentlemen?”
“Their names, sire, are Citizens Latour-Maubourg, Barnave, and
Pétion.”
The three deputies represented the three different parties of the
Assembly. Latour-Maubourg was Royalist, Barnave was
Constitutional, Pétion was Republican.
The crowd respectfully drew back. Three men approached the
royal carriage, stopped at the door, and saluted the King, who
returned their inclination.
One of them held in his hand a paper, which he read in a loud
voice. It was the decree of the National Assembly.
The man who read it was Pétion.
This decree ordered them to proceed to the King, not only to
ensure his safety, but also as a mark of respect due to royalty, as
represented in the persons of Louis XVI and Marie Antoinette.
The King knew that M. Latour-Maubourg was a Royalist.
He therefore desired that, as two deputies had to sit in the
carriage with him, he would name the two. The Queen expressed
the same desire.
M. Latour-Maubourg replied, in a whisper:—
“I accepted the sad mission which introduces me to your Majesty
only in the hopes of being of some service to you. Your Majesty can,
then, count upon me as a faithful follower. But I have not the power
of Barnave, who exercises an enormous influence over the Assembly.
He is vain as an advocate, and will be flattered by having a seat in
the carriage of the King. It is, therefore, needful that he should
occupy a place, and that the Queen should take the opportunity of
improving his acquaintance. I, therefore, beg your Majesties to
excuse my surrendering my seat.”
The Queen bowed her head. She wished to again assume her
womanly properties, and to seduce Barnave, as she had Mirabeau.
To be sure, it was humiliating, but, at the same time, it was a
distraction.
Strange contradiction! It was the King who had most repugnance
to Barnave’s occupying a seat in the royal carriage. Barnave, a little
Dauphinois advocate, pride upon his face, his nose perked up in the
air, and his tout ensemble proclaiming insufferable conceit, took his
place. Pétion likewise, his rosy cheeks glowing with satisfaction,
disposed of himself to his perfect content.
Barnave and Pétion, therefore, as we have said, entered the royal
carriage.
Madame de Tourzel had resigned her place, and entered, with M.
Latour-Maubourg, the carriage set apart for the attendants.
Pétion at once proclaimed his discourtesy by claiming, as
representative of the National Assembly, a seat with his face to the
horses. The King and Queen made a sign to Madame Elizabeth, who
at once changed places with him.
At last, all inside the royal carriage were satisfactorily arranged.
On the back seat were the King, Pétion, the Queen; and on the
front, Madame Elizabeth, face to face with Pétion, Madame Royale
and the Dauphin face to face and knee to knee with the Queen, who
was opposite, also, to Barnave.
At the first glance the Queen fancied that Barnave was dry, cold,
and wicked.
Barnave had hoped to take the place of Mirabeau at the Assembly.
He had succeeded in part; could not the Queen confer the rest?
Why not?
 Had she not, at St. Cloud, given a secret interview to Mirabeau?
Why should not he, Barnave, be accorded a similar favor.
But then, public rumor spread abroad that one of the three
gentlemen on the box of the carriage, “M. Fersen,” was the accepted
lover of the Queen.
Strange thing! As I have told you the good self-opinion of
Barnave, he was yet jealous of M. de Fersen.
With the admirable shrewdness of women, the Queen discovered
this before a quarter of an hour had elapsed.
She managed to get the three guards, named respectively MM. de
Malden, de Valory, and de Moustier.
No Fersen!
Barnave breathed, smiled, and became positively charming.
Barnave was young, handsome, polished, of fascinating manners,
and felt great commiseration for the unfortunate royal party.
In place of the Queen seducing Barnave, Barnave almost seduced
the Queen.
 CHAPTER XXXVI.
BARNAVE AND PETION.

One naturally asks how I became acquainted with all this.

I have already said that, on leaving Varennes, I had taken a place
on the back of the carriage of the King. Happily, I had managed to
retain my position, despite the heat, the fatigue, and the dust. Twice
only, for a few minutes, had I quitted my location; firstly, to try and
assist M. de Dampierre, and, secondly, to procure the water for
Madame Elizabeth and the Dauphin. Both times, on my return, I
recovered my place. The glass windows of the berlin were let down
on account of the heat, and the royal family, not speaking in very
low voices, I managed to hear pretty well all that was said.
This explanation given, I will continue my story, with the history of
the rudeness of Pétion, and the courtesy of Barnave.
There was placed between Madame Elizabeth and Madame
Royale, a bottle of lemonade and a glass. Pétion was thirsty, and felt
inclined to drink. He took the glass, and handed it to Madame
Elizabeth, who took up the lemonade, and filled it.
“Enough!” said Pétion, lifting his glass as he would have done at a
cabaret.
The Queen’s eyes flashed with anger.
The Dauphin, with the impatience of a youngster, shifted in his
seat; Pétion seized him, and imprisoned him between his legs.
The Queen said nothing, but again darted a look of menace at
Pétion; who, remembering that it might be politic to gain the favor
of the King, caressed the Dauphin’s white locks with apparent
affection.
 The Dauphin made a grimace expressive of grief.
The Queen snatched him from Pétion’s legs.
Barnave, smiling, immediately opened his arms to him.
The boy seemed willing, and was, therefore, soon installed on
Barnave’s knees.
His instinct shewed him that he would find in Barnave a protector.
Playing with a button on the coat of the representative, he
discovered that a device was inscribed upon it, and, after many
efforts, succeeded in reading it.
The device was, “Live free, or die.”
The Queen sighed, and regarded Barnave, her eyes filled with
tears.
Barnave’s heart smote him.
This was his position. He followed his own individual romance in
the midst of a royal and terrible history, when suddenly a noise was
heard some paces behind the royal carriage.
The cries and tumult drew Barnave from the magic circle which
surrounded him.
A venerable ecclesiastic approached the carriage, much in the
same manner as M. de Dampierre had done, and uplifted his hands
and blessed the royal martyr.
The mob, unsatiated by one murder, rushed upon the priest, and
drew him away, to slaughter him in the ditch by the roadside.
I was on the opposite side of the carriage to where this affair was
taking place.
“M. Barnave, M. Barnave!” cried I; “help, help!”
At the same moment, M. Barnave, putting his head out of the
window, saw what was taking place.
 He placed the Dauphin in the arms of his aunt, and opened the
carriage door with such violence and rapidity, that he almost fell out;
in fact, he would have fallen, had not Madame Elizabeth caught and
retained him by his coat-tails.
“Oh, Frenchmen!” cried he; “ye are a nation of brave men—would
ye become a horde of assassins?”
At this eloquent appeal, the people let go the priest, who escaped,
protected by the outspread arms and eloquent gestures of Barnave.
The door was again shut, Barnave retook his place, and the Queen
said to him, “I thank you, M. Barnave.”
He bowed his head.
Before the arrival of the commissaires, the King had eaten alone
with his family; but now, after consulting the Queen, he invited them
to share his repast.
Pétion accepted the invitation; Latour-Maubourg and Barnave
declined.
Barnave insisted, however, on waiting on the royal family; but the
Queen made him a sign, and he yielded.
I was one of the guard at the door of the dining-room.
In the evening, MM. Drouet and Guillaume set out at full speed, to
inform the Assembly of what had taken place.
Drouet came to bid me good-bye.
“M. Drouet,” said I to him, “you know me, as I am your pupil. I
take the greatest interest in that which is going on. It will be
something to talk about for the rest of my life. Give the order, before
you leave, to have me always placed close to their Majesties. The
fatigue will be nothing, and I wish to see all that goes on.”
“Be it as you wish,” said he, without making the least objection.
That was the reason why I had been appointed one of the guards
that day at the door of the dining-room.
 This is what happened at Dormans.
After dinner, the three commissaires went into the neighboring
room—that is to say, the one at which I mounted guard.
“Citizens,” said Barnave to them, “we are commissaires of the
National Assembly, and not the executioners of the royal family; and
to make them proceed under this burning sun is simply to conduct
them to the scaffold.”
“Good!” said Pétion. “What has happened to them has been
brought on by their own follies.”
“Still they are no less King and Queen,” replied M. de Latour-
Maubourg.
“If affairs keep progressing as they do now, it is extremely
probable that they will not long even have that title to console
them.”
“Quite right,” said Barnave. “But still I think that, as long as they
retain the titles of King and Queen, they ought to be treated as
such.”
“I have no objection,” said Pétion, in an indifferent tone. “Do as
you like, most loyal gentlemen.”
Saying these words, he left the room.
Barnave and M. de Latour-Maubourg, being alone, decided that
the royal carriage should be accompanied only by a cavalry escort,
so that it might proceed at a trot, and on the third evening arrive at
Meaux.
At that moment, they relieved guard. I ran to the postmaster at
Dormans, who was a friend of M. Drouet’s, and with whom we had
lodged on our way to the federation, and prayed him to lend me a
horse, to go as far as Meaux, where the royal family would halt, to
pass the night in repose.
In these critical times, paternal feeling elevated itself. The
postmaster had seen M. Drouet the evening before, who had
announced to him my arrival to-day. He would not let me hire the
steed—he gave it to me.
They arrived at Meaux about six in the evening.
The King again invited the commissaires to sup with him, as he
had before invited them to dine. Pétion accepted the invitation; M.
de Latour-Maubourg and Barnave refused it.
But the Queen, with charming grace, turning towards Barnave,
said, “Pray accept it, M. Barnave, as, after the meal, I shall have
need of you.”
Barnave bowed, the King signed to M. de Latour-Maubourg and
the two took their places at the royal table.
They were located in the palace of the Bishop of Meaux, a
melancholy-looking place enough, with its dark oak staircase and
mysterious and dusty passages.
I was on guard at the garden gate.
After dinner, the Queen, who, as she had said to Barnave, had
need of him, took his arm, and mounted the staircase to the
apartments above, under pretext of seeing a chamber once occupied
by Bossuet.
As for the King, he descended into the gardens with Pétion. Pétion
it was who desired the tête-à-tête.
Pétion, who, apart from his folly, was a good man, and had a good
heart, had formed an idea of escape for the King. It was, to allow
the three body-guards to go, so that they might disguise themselves
as National Guards, and so facilitate their entrance into Paris.
But, extraordinary to relate, the King could not understand this
idea of Pétion’s; and not wishing to be under an obligation to Pétion,
and having the absurd suspicion that he wished to assassinate the
guards, he refused.
And yet, on the day when he could have caused Lafayette to be
proclaimed Mayor of Paris, he nominated Pétion.
 It was because the Queen hated Lafayette more than the King
detested Pétion.
As for the Queen, no one knows what passed between her and
Barnave, except through the account which she afterwards gave to
Madame Campan.
The impression which the young representative produced on the
Queen may be summed up in those words.
“If ever power returns into our hands, the pardon of Barnave is
assured in our hearts.”
The Queen was ready to pardon Barnave for his rebellion; France
did not pardon him for his weakness.
The unhappy orator paid with his head for the few moments of
happiness he spent with this second Marie Stuart.
Perchance he had the same honor as Mirabeau, of kissing her
hand.
 CHAPTER XXXVII.
PARIS.

Day dawned.
It was the 25th of June. They returned to Paris after five days
absence.
Five days! What terrible events had come to pass in the space of
five days!
As they approached Paris, Barnave retook his seat at the back.
No longer was it a seat of honor, but the place of danger.
If a fanatic should fire on the King, which was, indeed, probable; if
on the Queen, which was more than probable;—Barnave was there,
to arrest with his own body the fell bullet aimed at royalty.
M. Mathieu Damas had been charged by Lafayette, Royalist
though he was, to protect their entry.
This able strategist had drawn from all parties in order to diminish
the danger. He confided the guarding of the carriage to the
grenadiers, whose tall hats hid entirely the doors; a line of horse
grenadiers formed a second ring.
As for the three guards whom Louis XVI had not wished to go,
two grenadiers, with their muskets bayoneted, sat a little behind the
box-seat, ready to suppress any attempt at rescue or flight.
The heat was tremendous. The carriage, the nearer it approached
Paris, appeared to be entering the mouth of a furnace.
The Queen, whom nothing hitherto had conquered, was beaten by
the heat. Twice or thrice she cried, “I suffocate!”
 At Bourget, the King asked for wine.
Broken down by fatigue, Madame Elizabeth slept.
The change of places had brought Pétion close by her. The face of
the future Mayor of Paris had a remarkable expression of joy. The
Queen, who cared not for sleeping herself, shook her by the arm in
order to awaken her.
“Let her alone,” cried Pétion. “Nature must take its course.”
They passed the barrier, and entered into the midst of a moving
and agitated people.
From time to time the crowd gave a tremendous yell. The King,
trying to show sang-froid began, apparently, to read.
“Suppose one were to applaud the King!”
“He shall be scourged!”
“Suppose one were to insult him?”
“He shall be hanged!”
The crowd kept pace with the carriage.
Mathieu Damas, commanding the escort, did not wish to enter
Paris by the Faubourg St. Martin. He was nearer the Faubourg St.
Antoine, of terrible memory, on account of the attack and seizure of
the Bastille.
He asked himself if he had a human barrier strong enough to
protect the royal family from the crowd who had virtually sentenced
them to death. He went round Paris by the external Boulevards, and
entered it by the Champ Elysées and the Place Louis XV.
On the Place Louis XV stood, at that period, the statue of the
monarch whose name the place bore.
They had bandaged the eyes of the statue with a handkerchief.
This allusion, though ignored by the King, still disquieted him.
 “Why this bandage on the eyes of my predecessor?” inquired the
King.
“To show the blindness of the monarchy, sire,” replied Pétion.
In the progress from the Champs Elysées to the Place Louis XV,
the barrier of grenadiers was often broken.
Then the Queen saw appear at the windows hideous faces,
expressive of satisfaction and revenge.
What caused those devils to turn away and bow?
It was a kiss which the Dauphin sent them, and a bow from his
sister; those white-winged angels hovering over the royal family.
Lafayette, with his etat major, passed by the Queen.
As soon as she perceived them, she cried out, “M. Lafayette,
above all things save my three body-guards; their crime has been
but to obey me.”
The same cry was uttered by her at Versailles on the 6th of
October. Their danger was really great.
The carriages passed through the gate of the Tuileries, which was
vainly endeavored to be shut after them. They proceeded along the
grand promenade of the garden, and halted only at the end of the
great terrace which stretched along the front of the palace.
It was there that the crowd, greater than ever, awaited them. It
was impossible to go farther; they must get out of their carriage.
The Assembly was not present, but it had sent twenty deputies.
Lafayette cleared a pathway from the terrace to the palace door.
He constructed an iron arch with the muskets and bayonets of the
National Guard.
“M. Barnave,” again cried the Queen, “I ask you to protect my
three guards.”
The children first descended, and entered the palace without
opposition. It was then the turn of the three guards, for whom the
Queen had asked protection from M. Lafayette and M. Barnave.
Then there came a terrible outcry.
I had left my horse at the top of the Champs Elysées, and
marched with the grenadiers on foot. At first, they tried to turn me
out, but the King said, “Let him alone; he is a friend.”
They did leave me alone. M. Pétion gave me a side glance; M.
Barnave smiled.
The King and Queen looked to see what would happen to the
three guards; the King gazed with his usual apathy, the Queen with
intense interest.
The sabres and pikes of the National Guard waved over them as
they shouted, “Death to the traitors!”
All of a sudden, I saw a stream of blood running down M. de
Malden’s cheek.
Being in the circle, I drew him, with a vigorous effort, towards me,
crying, “Peace! peace! I am the friend of M. Drouet.”
Five hundred voices shouted, “Long live Drouet! Long live
Guillaume!”
I drew M. de Malden under the arch of the Grand Pavilion, but he
would proceed no farther until assured of the safety of the King and
Queen.
During this time, in the midst of the most terrible murmurs, they
saved M. de Valory and M. de Moustier.
Like M. de Malden, M. de Valory was wounded; but also like M. de
Malden, his wound was but slight.
At this moment, the Queen cried, in a suffocating voice, “Help!
help!”
In getting out of the carriage, she found herself in the arms of two
men, who regarded her with looks of mortal enmity, and at the same
time held her fast.
 These two men were M. de Aguillon and M. de Noailles. The
Queen seemed likely to faint with terror. Both said to her, “Fear
nothing, madame; we protect you.”
At the peril of their lives, they conducted her to her room. There
she was seized with agony. She called the Dauphin—she looked for
the Dauphin, but no Dauphin was there.
Madame Royale took her by the hand and led her into the bed-
room, and pointed out to her the Dauphin, who, overcome with
fatigue, slept.
She could not believe, after the threats she had heard, that the
whole of the royal family could re-enter their palace safe and sound.
I returned to the carriage, where still remained Madame Elizabeth
and the King.
Barnave thought that it would not be too much for him and Pétion
to safeguard the King.
“Some one,” cried he—“some one to offer Madame Elizabeth an
arm.”
Madame Elizabeth descended from the carriage with her usual
angelic calmness.
“Monsieur,” said she to me, “will you give me your arm?”
I was frightened out of my wits.
“Oh, madame,” said I; “this dress?”
“The dress that you wear is far better than a royal robe. And
besides,” continued she, “I have watched you: you are a young man
of a good heart.”
I threw my gun over my shoulder, and took my hat in my hand.
“Madame,” said I, “if you desire one ready to die for you—to throw
down his life in your behalf, your choice could not fall on one better
than myself.”
 They saw Madame Elizabeth take the arm of a simple National
Guard, and they clapped their hands.
Arrived at the foot of the staircase, I wished to retire.
“My brother?” said she, trying to see.
I looked back.
“He is coming,” said I, “between M. Barnave and M. Pétion.”
I then bowed to Madame Elizabeth a second time.
“Will you not return to see us, sir?” asked Madame Elizabeth.
“I fear, madame, that I shall not again have the opportunity of
being of service to you.”
“Perhaps so, but you have been; and, whatever people may say,
we are a family that never forget.”
At this moment the King arrived.
“Thank you, gentlemen; thank you,” said he to Barnave and
Pétion. “I need not say to you that if you like to come up-stairs——”
“Sire,” replied Barnave, “your Majesty and her Majesty the Queen
are at present in safety. We must go to render an account of our
mission to the Assembly.”
They bowed to the King, and retired.
I did the same; that is to say, I bowed; but as I was retiring,
Madame Elizabeth, pointing me out to the King, said, “My brother,
this young man?”
She evidently, in her noble heart, did not wish me to go without
some recompense.
“’Tis true,” said the King; “I forgot that he was your protegé.”
“Say, rather, that I am his protegé.”
He took me by the collar of my coat.
 “Look here, young man; unhappy as we are, can we do nothing to
help you?”
I felt wounded that the King should think that I required to be
paid for what I had done.
“Sire,” replied I, “if you make a promise to the nation, keep it;
and, as a citizen, you will have done all for me that I can ask.”
“You see, sister,” said the King, “he is a savage.”
“What is your name, sir?” asked Madame Elizabeth.
“Réné Besson.”
“Whence come you?”
“From the Forest of Argonne.”
“I told you he was a savage,” said the King. “What else could you
expect?”
“What trade are you?”
“A carpenter.”
“My brother, you know the fable of the Lion and the Rat,” said
Madame Elizabeth.
“My friend,” said the King, “you see that I must enter my house. If
you have need of me, ask for Cléry, my valet-de-chambre.”
“Sire,” replied I, “a man who has an occupation has need of no
one, much less of a King.”
The King shrugged his shoulders, and mounted the staircase.
Madame Elizabeth stayed behind.
“But, on the other hand, my friend,” said she, “suppose that we
have need of you?”
“Ah, madame,” cried I, “that is another affair!”
“In that case, M. Réné Besson, ask for Cléry.”
 She followed her brother, whilst I stood there motionless,
regarding that angel who knew how to recompense one in asking.
On the morrow, the journalist, Prudhomme, wrote:—

“Certain good patriots, in whom the sentiment of loyalty has not

extinguished that of compassion, appear uneasy concerning the moral
and physical state of Louis XVI and his family, after a journey so
fatiguing in all respects as that from St. Menehould.
“Let them reassure themselves. Our friend, on entering his
apartments, on his return, felt no more fatigue than if he had been
indulging in the pleasures of the chase.
“He ate his chicken as usual, and the next day played after dinner
with his son.
“As for the mother, she took a bath on her arrival. Her first request
was for boots; she having remarked with sorrow, that hers had been
destroyed by travelling. She acted with hauteur to the officers picked
out especially to guard her, and said that it was ridiculous and
indecent to have the door of her bath-room and bed-chamber left
open.”

We quote these four paragraphs to show to what an extent party

spirit can blind men.
The Citizen Prudhomme, who, after having written “The
Revolutions of Paris in ’91,” was to write “The Crimes of the
Revolution of ’98,” wrongfully describes four incidents:—“That the
King ate a fowl, and that he played with his son; that the Queen had
a bath, and shut her door when taking it.”
It is always so. There can never be a revolution without a
Prudhomme: first, to glorify them: and then to grossly insult.