Integrable Systems 1St Edition Ahmed Lesfari Full Chapter PDF

Integrable Systems 1st Edition Ahmed
Lesfari
Visit to download the full and correct content document:
https://ebookmass.com/product/integrable-systems-1st-edition-ahmed-lesfari/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...
Integrable Systems 1st Edition Ahmed Lesfari
https://ebookmass.com/product/integrable-systems-1st-edition-
ahmed-lesfari-2/
Fractional-order Design: Devices, Circuits, And Systems

Ahmed G. Radwan
https://ebookmass.com/product/fractional-order-design-devices-
circuits-and-systems-ahmed-g-radwan/
Viroids and Satellites 1st Edition Edition Ahmed Hadidi
https://ebookmass.com/product/viroids-and-satellites-1st-edition-
edition-ahmed-hadidi/
Back-Of-The-Envelope Quantum Mechanics: with Extensions

to Many-Body Systems and Integrable Pdes 2nd Edition
Maxim Olshanii
https://ebookmass.com/product/back-of-the-envelope-quantum-
mechanics-with-extensions-to-many-body-systems-and-integrable-
pdes-2nd-edition-maxim-olshanii/
Cybersecurity and Cognitive Science 1st Edition Ahmed
Moustafa
https://ebookmass.com/product/cybersecurity-and-cognitive-
science-1st-edition-ahmed-moustafa/
Religion, Extremism and Violence in South Asia 1st

Edition Imran Ahmed
https://ebookmass.com/product/religion-extremism-and-violence-in-
south-asia-1st-edition-imran-ahmed/
Reservoir Engineering Handbook Fifth Edition 2019 Tarek

Ahmed
https://ebookmass.com/product/reservoir-engineering-handbook-
fifth-edition-2019-tarek-ahmed/
Endodontic Advances and Evidence-Based Clinical

Guidelines 1st Edition Hany M. A. Ahmed
https://ebookmass.com/product/endodontic-advances-and-evidence-
based-clinical-guidelines-1st-edition-hany-m-a-ahmed/
Bionanotechnology: Emerging Applications of

Bionanomaterials Ahmed Barhoum
https://ebookmass.com/product/bionanotechnology-emerging-
applications-of-bionanomaterials-ahmed-barhoum/
Integrable Systems
To my Professor Pierre van Moerbeke
Integrable Systems
Ahmed Lesfari
First published 2022 in Great Britain and the United States by ISTE Ltd and John Wiley & Sons, Inc.
Apart from any fair dealing for the purposes of research or private study, or criticism or review, as
permitted under the Copyright, Designs and Patents Act 1988, this publication may only be reproduced,
stored or transmitted, in any form or by any means, with the prior permission in writing of the publishers,
or in the case of reprographic reproduction in accordance with the terms and licenses issued by the
CLA. Enquiries concerning reproduction outside these terms should be sent to the publishers at the
undermentioned address:
ISTE Ltd John Wiley & Sons, Inc.

27-37 St George’s Road 111 River Street
London SW19 4EU Hoboken, NJ 07030
UK USA
www.iste.co.uk www.wiley.com
© ISTE Ltd 2022

The rights of Ahmed Lesfari to be identified as the author of this work have been asserted by him in
accordance with the Copyright, Designs and Patents Act 1988.
Any opinions, findings, and conclusions or recommendations expressed in this material are those of the
author(s), contributor(s) or editor(s) and do not necessarily reflect the views of ISTE Group.
Library of Congress Control Number: 2022932445
British Library Cataloguing-in-Publication Data

A CIP record for this book is available from the British Library
ISBN 978-1-78630-827-6
Contents
Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix
Chapter 1. Symplectic Manifolds . . . . . . . . . . . . . . . . . . . . . . . 1

1.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2. Symplectic vector spaces . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.3. Symplectic manifolds . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.4. Vectors fields and flows . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.5. The Darboux theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
1.6. Poisson brackets and Hamiltonian systems . . . . . . . . . . . . . . . . 25
1.7. Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
1.8. Coadjoint orbits and their symplectic structures . . . . . . . . . . . . . 35
1.9. Application to the group SO(n) . . . . . . . . . . . . . . . . . . . . . . 37
1.9.1. Application to the group SO(3) . . . . . . . . . . . . . . . . . . . . 39
1.9.2. Application to the group SO(4) . . . . . . . . . . . . . . . . . . . . 41
1.10. Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
Chapter 2. Hamilton–Jacobi Theory . . . . . . . . . . . . . . . . . . . . . 49

2.1. Euler–Lagrange equation . . . . . . . . . . . . . . . . . . . . . . . . . . 49
2.2. Legendre transformation . . . . . . . . . . . . . . . . . . . . . . . . . . 52
2.3. Hamilton’s canonical equations . . . . . . . . . . . . . . . . . . . . . . 53
2.4. Canonical transformations . . . . . . . . . . . . . . . . . . . . . . . . . 55
2.5. Hamilton–Jacobi equation . . . . . . . . . . . . . . . . . . . . . . . . . 57
2.6. Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
2.6.1. Harmonic oscillator . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
2.6.2. The Kepler problem . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
2.6.3. Simple pendulum . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
2.7. Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
vi Integrable Systems
Chapter 3. Integrable Systems . . . . . . . . . . . . . . . . . . . . . . . . . 67

3.1. Hamiltonian systems and Arnold–Liouville theorem . . . . . . . . . . 67
3.2. Rotation of a rigid body about a fixed point . . . . . . . . . . . . . . . . 75
3.2.1. The Euler problem of a rigid body . . . . . . . . . . . . . . . . . . . 78
3.2.2. The Lagrange top . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
3.2.3. The Kowalewski spinning top . . . . . . . . . . . . . . . . . . . . . 83
3.2.4. Special cases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
3.3. Motion of a solid through ideal fluid . . . . . . . . . . . . . . . . . . . . 91
3.3.1. Clebsch’s case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
3.3.2. Lyapunov–Steklov’s case . . . . . . . . . . . . . . . . . . . . . . . . 93
3.4. Yang–Mills field with gauge group SU (2) . . . . . . . . . . . . . . . . 93
3.5. Appendix (geodesic flow and Euler–Arnold equations) . . . . . . . . . 95
3.6. Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
Chapter 4. Spectral Methods for Solving Integrable Systems . . . . 103

4.1. Lax equations and spectral curves . . . . . . . . . . . . . . . . . . . . . 103
4.2. Integrable systems and Kac–Moody Lie algebras . . . . . . . . . . . . 104
4.3. Geodesic flow on SO(n) . . . . . . . . . . . . . . . . . . . . . . . . . . 107
4.4. The Euler problem of a rigid body . . . . . . . . . . . . . . . . . . . . . 108
4.5. The Manakov geodesic flow on the group SO(4) . . . . . . . . . . . . 109
4.6. Jacobi geodesic flow on an ellipsoid and Neumann problem . . . . . . 114
4.7. The Lagrange top . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
4.8. Quartic potential, Garnier system . . . . . . . . . . . . . . . . . . . . . 115
4.9. The coupled nonlinear Schrödinger equations . . . . . . . . . . . . . . 118
4.10. The Yang–Mills equations . . . . . . . . . . . . . . . . . . . . . . . . . 119
4.11. The Kowalewski top . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
4.12. The Goryachev–Chaplygin top . . . . . . . . . . . . . . . . . . . . . . 121
4.13. Periodic infinite band matrix . . . . . . . . . . . . . . . . . . . . . . . 122
4.14. Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
Chapter 5. The Spectrum of Jacobi Matrices and Algebraic

Curves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
5.1. Jacobi matrices and algebraic curves . . . . . . . . . . . . . . . . . . . 129
5.2. Difference operators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
5.3. Continued fraction, orthogonal polynomials and Abelian integrals . . . 137
5.4. Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
Chapter 6. Griffiths Linearization Flows on Jacobians . . . . . . . . . 143

6.1. Spectral curves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
6.2. Cohomological deformation theory . . . . . . . . . . . . . . . . . . . . 144
6.3. Mittag–Leffler problem . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
6.4. Linearizing flows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149
Contents vii
6.5. The Toda lattice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150

6.6. The Lagrange top . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
6.7. Nahm’s equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
6.8. The n-dimensional rigid body . . . . . . . . . . . . . . . . . . . . . . . 155
6.9. Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 156
Chapter 7. Algebraically Integrable Systems . . . . . . . . . . . . . . . 159

7.1. Meromorphic solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
7.2. Algebraic complete integrability . . . . . . . . . . . . . . . . . . . . . . 164
7.3. The Liouville–Arnold–Adler–van Moerbeke theorem . . . . . . . . . . 171
7.4. The Euler problem of a rigid body . . . . . . . . . . . . . . . . . . . . . 173
7.5. The Kowalewski top . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
7.6. The Hénon–Heiles system . . . . . . . . . . . . . . . . . . . . . . . . . 191
7.7. The Manakov geodesic flow on the group SO(4) . . . . . . . . . . . . 200
7.8. Geodesic flow on SO(4) with a quartic invariant . . . . . . . . . . . . . 206
7.9. The geodesic flow on SO(n) for a left invariant metric . . . . . . . . . 210
7.10. The periodic five-particle Kac–van Moerbeke lattice . . . . . . . . . . 212
7.11. Generalized periodic Toda systems . . . . . . . . . . . . . . . . . . . . 213
7.12. The Gross–Neveu system . . . . . . . . . . . . . . . . . . . . . . . . . 214
7.13. The Kolossof potential . . . . . . . . . . . . . . . . . . . . . . . . . . . 215
7.14. Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215
Chapter 8. Generalized Algebraic Completely Integrable

Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221
8.1. Generalities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221
8.2. The RDG potential and a five-dimensional system . . . . . . . . . . . . 225
8.3. The Hénon–Heiles problem and a five-dimensional system . . . . . . . 229
8.4. The Goryachev–Chaplygin top and a seven-dimensional system . . . . 231
8.5. The Lagrange top . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 236
8.6. Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237
Chapter 9. The Korteweg–de Vries Equation . . . . . . . . . . . . . . . 241

9.1. Historical aspects and introduction . . . . . . . . . . . . . . . . . . . . 241
9.2. Stationary Schrödinger and integral Gelfand–Levitan equations . . . . 243
9.3. The inverse scattering method . . . . . . . . . . . . . . . . . . . . . . . 255
9.4. Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 269
Chapter 10. KP–KdV Hierarchy and Pseudo-differential

Operators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275
10.1. Pseudo-differential operators and symplectic structures . . . . . . . . 275
10.2. KdV equation, Heisenberg and Virasoro algebras . . . . . . . . . . . . 279
viii Integrable Systems
10.3. KP hierarchy and vertex operators . . . . . . . . . . . . . . . . . . . . 281

10.4. Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 290
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 293
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 305
Preface
This book is intended for a wide readership of mathematicians and physicists:

students pursuing graduate, masters and higher degrees in mathematics and
mathematical physics. It is devoted to some geometric and topological aspects of the
theory of integrable systems and the presentation is clear and well-organized, with
many examples and problems provided throughout the text. Integrable Hamiltonian
systems are nonlinear ordinary differential equations that are described by a
Hamiltonian function and possess sufficiently many independent constants of motion
in involution. The problem of finding and integrating Hamiltonian systems has
attracted a considerable amount of attention in recent decades. Besides the fact that
many integrable systems have been the subject of powerful and beautiful theories of
mathematics, another motivation for their study is the concepts of integrability that
are applied to an increasing number of physical systems, biological phenomena,
population dynamics and chemical rate equations, to mention but a few applications.
However, it still seems hopeless to describe, or even to recognize with any facility,
the Hamiltonian systems which are integrable, even though they are exceptional.
Chapter 1 is devoted to the study of symplectic manifolds and their connection

with Hamiltonian dynamical systems. We review some interesting properties of
one-parameter groups of diffeomorphisms or of flow, Lie derivative, interior product
or Cartan’s formula, as well as the study of a central theorem of symplectic geometry,
namely, Darboux’s theorem. We also show how to determine explicitly symplectic
structures on adjoint and coadjoint orbits of a Lie group, with particular attention
given to the group SO(n).
Chapter 2 deals with the study of some notions concerning the Hamilton–Jacobi
theory in the calculus of variations. We will establish the Euler–Lagrange differential
equations, Hamilton’s canonical equations and the Hamilton–Jacobi partial
differential equation and explain how it is widely used in practice to solve some
x Integrable Systems
problems. As an application, we will study the geodesics, the harmonic oscillator, the
Kepler problem and the simple pendulum.
In Chapter 3, we study the Arnold–Liouville theorem: the regular compact level

manifolds defined by the intersection of the constants of motion are diffeomorphic to
a real torus on which the motion is quasi-periodic as a consequence of the following
differential geometric fact; a compact and connected n-dimensional manifold on
which there exist n vector fields that commute and are independent at every point is
diffeomorphic to an n-dimensional real torus, and there is a transformation to
so-called action-angle variables, mapping the flow into a straight line motion on that
torus. We give a proof as direct as possible of the Arnold–Liouville theorem and we
make a careful study of its connection with the concept of completely integrable
systems. Many problems are studied in detail: the rotation of a rigid body about a
fixed point, the motion of a solid in an ideal fluid and the Yang–Mills field with
gauge group SU (2).
In Chapter 4, we give a detailed study of the integrable systems that can be

written as Lax equations with a spectral parameter. Such equations have no a priori
Hamiltonian content. However, through the Adler–Kostant–Symes (AKS)
construction, we can produce Hamiltonian systems on coadjoint orbits in the dual
space to a Lie algebra whose equations of motion take the Lax form. We outline an
algebraic-geometric interpretation of the flows of these systems, which are shown to
describe linear motion on a complex torus. The relationship between spectral theory
and completely integrable systems is a fundamental aspect of the modern theory of
integrable systems. This chapter surveys a number of classical and recent results and
our purpose here is to sketch a motivated overview of this interesting subject. We
present a Lie algebra theoretical schema leading to integrable systems based on the
Kostant–Kirillov coadjoint action. Many problems on Kostant–Kirillov coadjoint
orbits in subalgebras of infinite dimensional Lie algebras (Kac–Moody Lie algebras)
yield large classes of extended Lax pairs. A general statement leading to such
situations is given by the AKS theorem, and the van Moerbeke–Mumford
linearization method provides an algebraic map from the complex invariant
manifolds of these systems to the Jacobi variety (or some subabelian variety of it) of
the spectral curve. The complex flows generated by the constants of the motion are
straight line motions on these varieties. This chapter describes a version of the
general scheme, and shows in detail how several important classes of examples fit
into the general framework. Several examples of integrable systems of relevance in
mathematical physics are carefully discussed: geodesic flow on SO(n), the Euler
problem of a rigid body, Manakov geodesic flow on the group SO(4), Jacobi
geodesic flow on an ellipsoid, the Neumann problem, the Lagrange top, a quartic
potential or Garnier system, coupled nonlinear Schrödinger equations, Yang–Mills
equations, the Kowalewski spinning top, the Goryachev–Chaplygin top and the
periodic infinite band matrix.
Preface xi
The aim of Chapter 5 is to describe some connections between spectral theory

in infinite dimensional Lie algebras, deformation theory and algebraic curves. We
study infinite continued fractions, isospectral deformation of periodic Jacobi matrices,
general difference operators, Cauchy–Stieltjes transforms and Abelian integrals from
an algebraic geometrical point of view. These results can be used to obtain insight into
integrable systems.
In Chapter 6, we present in detail the Griffiths’ approach and his cohomological

interpretation of the linearization test for solving integrable systems without reference
to Kac–Moody algebras. His method is based on the observation that the tangent space
to any deformation lies in a suitable cohomology group and on algebraic curves, higher
cohomology can always be eliminated using duality theory. We explain how results
from deformation theory and algebraic geometry can be used to obtain insight into
the dynamics of integrable systems. These conditions are cohomological and the Lax
equations turn out to have a natural cohomological interpretation. Several nonlinear
problems in mathematical physics illustrate these results: the Toda lattice, Nahm’s
equations and the n-dimensional rigid body.
In Chapter 7, the notion of algebraically completely integrable Hamiltonian

systems in the Adler–van Moerbeke sense is explained, and techniques to find and
solve such systems are presented. These are integrable systems whose trajectories are
straight line motions on Abelian varieties (complex algebraic tori). We make, via the
Kowalewski–Painlevé analysis, a study of the level manifolds of the systems, which
are described explicitly as being affine part of Abelian varieties and the flow can be
solved by quadrature, that is to say their solutions can be expressed in terms of
Abelian integrals. We describe an explicit embedding of these Abelian varieties that
complete the generic invariant surfaces into projective spaces. Many problems are
studied in detail: the Euler problem of a rigid body, the Kowalewski top, the
Hénon–Heiles system, Manakov geodesic flow on the group SO(4), geodesic flow
on SO(4) with a quartic invariant, geodesic flow on SO(n) for a left invariant
metric, the periodic five-particle Kac–van Moerbeke lattice, generalized periodic
Toda systems, the Gross–Neveu system and the Kolossof potential.
In Chapter 8, we discuss the study of generalized algebraic completely integrable

systems. There are many examples of differential equations that have the weak
Painlevé property that all movable singularities of the general solution have only a
finite number of branches, and some interesting integrable systems appear as
coverings of algebraic completely integrable systems. The invariant varieties are
coverings of Abelian varieties and these systems are called algebraic completely
integrable in the generalized sense. These systems are Liouville integrable and by the
Arnold–Liouville theorem, the compact connected manifolds invariant by the real
flows are tori, the real parts of complex affine coverings of Abelian varieties. Most of
these systems of differential equations possess solutions that are Laurent series of
t1/n (t being complex time) and whose coefficients depend rationally on certain
xii Integrable Systems
algebraic parameters. We discuss some interesting examples: Ramani–Dorizzi–

Grammaticos (RDG) potential, the Hénon–Heiles system, the Goryachev–Chaplygin
top, a seven-dimensional system and the Lagrange top.
Chapter 9 covers the stationary Schrödinger equation, the integral

Gelfand–Levitan equation and the inverse scattering method used to solve exactly the
Korteweg–de Vries (KdV) equation. The latter is a universal mathematical model for
the description of weakly nonlinear long wave propagation in dispersive media. The
study of this equation is the archetype of an integrable system and is one of the most
fundamental equations of soliton phenomena.
In Chapter 10, we study some generalities on the algebra of infinite order

differential operators. The algebras of Virasoro, Heisenberg and nonlinear evolution
equations such as the KdV, Boussinesq and Kadomtsev–Petviashvili (KP) equations
play a crucial role in this study. We make a careful study of some connection between
pseudo-differential operators, symplectic structures, KP hierarchy and tau functions
based on the Sato–Date–Jimbo–Miwa–Kashiwara theory. A few other connections
and ideas concerning the KdV and Boussinesq equations and the Gelfand–Dickey
flows, the Heisenberg and Virasoro algebras are given. The study of the KP and KdV
hierarchies, the use of tau functions related to infinite dimensional Grassmannians,
Fay identities, vertex operators and the Hirota’s bilinear formalism led to obtaining
remarkable properties concerning these algebras such as, for example, the existence
of an infinite family of first integrals functionally independent and in involution.
It is well known that when studying integrable systems, elliptic functions and
integrals, compact Riemann surfaces or algebraic curves, Abelian surfaces (as well as
the basic techniques to study two-dimensional algebraic completely integrable
systems) play a crucial role. These facts, which may be well known to the algebraic
reader, can be found, for example, in Adler and van Moerbeke (2004); Fay (1973);
Griffiths and Harris (1978); Lesfari (2015b) and Vanhaecke (2001).
I would like to thank and am grateful to P. van Moerbeke and L. Haine, from whom
I learned much of this subject through conversations and remarks. I would also like to
thank the editors for their interest, seriousness and professionalism. Finally my thanks
go to my wife and our children for much encouragement and undeniable support, who
helped bring this book into being.
Ahmed L ESFARI
September 2021
1
Symplectic Manifolds
1.1. Introduction
This chapter is devoted to the study of symplectic manifolds and their connection
with Hamiltonian systems. It is well known that symplectic manifolds play a crucial
role in classical mechanics, geometrical optics and thermodynamics, and currently
have conquered a rich territory, asserting themselves as a central branch of
differential geometry and topology. In addition to their activity as an independent
subject, symplectic manifolds are strongly stimulated by important interactions with
many mathematical and physical specialties, among others. The aim of this chapter is
to study some properties of symplectic manifolds and Hamiltonian dynamical
systems, and to review some operations on these manifolds.
This chapter is organized as follows. In the second section, we begin by briefly

recalling some notions about symplectic vector spaces. The third section defines and
develops explicit calculation of symplectic structures on a differentiable manifold
and studies some important properties. The forth section is devoted to the study of
some properties of one-parameter groups of diffeomorphisms or flow, Lie derivative,
interior product and Cartan’s formula. We review some interesting properties and
operations on differential forms. The fifth section deals with the study of a central
theorem of symplectic geometry, namely Darboux’s theorem: the symplectic
manifolds (M, ω) of dimension 2m are locally isomorphic to (R2m , ω). The sixth
section contains some technical statements concerning Hamiltonian vector fields.
The latter form a Lie subalgebra of the space vector field and we show that the matrix
associated with a Hamiltonian system forms a symplectic structure. Several
properties concerning Hamiltonian vector fields, their connection with symplectic
manifolds, Poisson manifolds or Hamiltonian manifolds as well as some interesting
examples are studied in the seventh section. We will see in the eight section how to
determine a symplectic structure on the orbit of the coadjoint representation of a Lie
group. Section nine is dedicated to the explicit determination of symplectic structures
Integrable Systems,
First Edition. Ahmed Lesfari.
© ISTE Ltd 2022. Published by ISTE Ltd and John Wiley & Sons, Inc.
2 Integrable Systems
on adjoint and coadjoint orbits of a Lie group SO(n). Some exercises are proposed
in the last section.
1.2. Symplectic vector spaces
D EFINITION 1.1.– A symplectic space (E, ω) is a finite dimensional real vector space
E with a bilinear form ω : E × E −→ R, which is alternating (or antisymmetric),
that is, ω(x, y) = −ω(y, x), ∀x, y ∈ E, and non-degenerate, that is, ω(x, y) = 0,
∀y ∈ E =⇒ x = 0. The form ω is referred to as symplectic form (or symplectic
structure).
The dimension of a symplectic vector space is always even. We show (using a

reasoning similar to the Gram–Schmidt orthogonalization process) that any
symplectic vector space (E, ω) has a base (e1 , ..., e2m ) called symplectic basis (or
canonical basis), satisfying the following relations: ω(em+i , ej ) = δij and
ω(ei , ej ) = ω(em+i , em+j ) = 0. Note that each em+i is orthogonal to all base
vectors except ei . In terms of symplectic basic vectors (e1 , ..., e2m ), the matrix (ωij )
where ωij ≡ ω(ei , ej ) has the form
⎛ ⎞
ω11 ... ω12m
⎜ .. . . .. ⎟ 0 −Im
⎝ . . . ⎠ =
Im 0
,
ω2m1 ... ω2m2m
where Im denotes the m × m unit matrix.

m
E XAMPLE 1.1.– R2m with the form ω(x, y) = k=1 (xm+k yk −xk ym+k ), x ∈ R2m ,
y ∈ R2m , is a symplectic vector space. Let (e1 , ..., em ) be an orthonormal basis of
Rm . Then, ((e1 , 0), ..., (em , 0), (0, e1 ), ..., (0, em )) is a symplectic basis of R2m .
Let (E, ω) be a symplectic vector space and F a vector subspace of E. Let F ⊥ =

{x ∈ E : ∀y ∈ F, ω(x, y) = 0} be the orthogonal (symplectic) of F .
D EFINITION 1.2.– The subspace F is isotropic if F ⊂ F ⊥ , coisotropic if F ⊥ ⊂ F ,

Lagrangian if F = F ⊥ and symplectic if F ∩ F ⊥ = {0}.
If F , F1 and F2 are subspaces of a symplectic space (E, ω), then
dim F + dim F ⊥ = dim E, (F ⊥ )⊥ = F,
F1 ⊂ F2 =⇒ F2⊥ ⊂ F1⊥ , (F1 ∩ F2 )⊥ = F1⊥ + F2⊥ , F1⊥ ∩ F2⊥ = (F1 + F2 )⊥ ,
F is coisotropic if and only if F ⊥ is isotropic and F is Lagrangian if and only if

F is isotropic and coisotropic.
Symplectic Manifolds 3
1.3. Symplectic manifolds
D EFINITION 1.3.– Let M be an even-dimensional differentiable manifold. A

symplectic structure (or symplectic form) on M is a closed non-degenerate
differential 2-form ω on M . The non-degeneracy condition means that: ∀x ∈ M ,
∀ξ = 0, ∃η: ω (ξ, η) = 0, (ξ, η ∈ Tx M ). The pair (M, ω) (or simply M ) is called a
symplectic manifold.
At a point p ∈ M , we have a non-degenerate antisymmetric bilinear form on the

tangent space Tp M , which explains why the dimension of the manifold M is even.
m
E XAMPLE 1.2.– R2m with the 2-form ω = k=1 dxk ∧ dyk is a symplectic
manifold. The vectors ∂
∂x1 , ..., ∂
∂xm , ∂
∂y1 , ..., ∂y∂m , p ∈ R2m ,
p p p p
constitute a symplectic basis of the tangent space T R = R . Similarly, Cm with
2m 2m
m
the form ω = 2i k=1 dzk ∧ dz k is a symplectic manifold. This form coincides with
the previous form by means of the identification Cm R2m , zk = xk + iyk .
Riemann surfaces, Kählerian manifolds and complex projective manifolds are
symplectic manifolds. Another class of symplectic manifolds consists of the
coadjoint orbits (see section 1.8).
We will see that the cotangent bundle T ∗ M (i.e. the union of all cotangent spaces
of M ) admits a natural symplectic structure. The phase spaces of the Hamiltonian
systems studied below are symplectic manifolds and often they are cotangent bundles
equipped with the canonical structure.
T HEOREM 1.1.– Let M be a differentiable manifold of dimension m and T ∗ M its

cotangent bundle. Then T ∗ M possesses a symplectic structure and in a local
m
coordinate (x1 , . . . , xm , y1 , . . . , ym ), the form ω is given by ω = k=1 dxk ∧ dyk .
P ROOF.– Let (U, ϕ) be a local chart in the neighborhood of p ∈ M , and consider the
m
application ϕ : U ⊂ M −→ Rm , p −→ ϕ(p) = k=1 xk ek , where ek are the vectors
basis of R . Consider the canonical projections T M −→ M , and T (T ∗ M ) −→
m
T ∗ M , of tangent bundles, respectively, to M and T ∗ M on their bases. We note π ∗ :

T ∗ M −→ M , the canonical projection and dπ ∗ : T (T ∗ M ) −→ T M , its linear
tangent application. We have ϕ∗ : T ∗ M −→ R2m , α −→ ϕ∗ (α) = k=1 (xk ek +
m
∗ m ∗
yk εk ), where εk are the basic forms of T R and α denotes αp ∈ T M . So, if α is a
1-form on M and ξα is a vector tangent to T ∗ M , then dϕ∗ : T (T ∗ M ) −→ T R2m =
R2m , ξα −→ dϕ∗ (ξα ) =
m
k=1 (βk ek + γk εk ), where βk , γk are the components
of ξα in the local chart of R . Consider λα (ξα ) = α(dπ ∗ ξα ) = α(ξ), where ξ is
2m
a tangent vector to M . Let (x1 , ..., xm , y1 , ..., ym ) be a system of local coordinates

compatible with a local trivialization of the tangent bundle T ∗ M . Let us show that:
⎛ ⎞

m
m
m
m
λα (ξα ) = α βk e k = (xk ek + yk εk ) ⎝ βj e j ⎠ = βk yk .
k=1 k=1 j=1 k=1
Indeed, let (x1 , ..., xm ) be a system of local coordinates around p ∈ M . Since

∀α ∈ T ∗ M , α =
m
k=1 αk dxk , then by defining local coordinates y1 , ..., ym by
m
yk (α) = yk , k = 1, ..., m, the 1-form λ is written as λ = k=1 yk dxk . The form λ
∗
on the cotangent bundle T M (doing correspondence λα to α) is called Liouville
m m
form. We have λ(α) = k=1 yk (α)dxk (α), λ(α)(ξα ) = k=1 yk (α)dxk (α)
m m m
j=1 βj ej + γj εj = k=1 yk βk = λα (ξα ), λ = k=1 yk dxk . The symplectic
structure of T ∗ M is given by the exterior derivative of λ, that is, the 2-form
ω = −dλ. The forms λ and ω are called canonical forms on T ∗ M . We can visualize
all this with the help of the following diagram:
T ∗ (T ∗ M )
↑λ
λα (ξ) ϕ∗
R ←− T (T ∗ M ) −→ T ∗M −→ R2m
↓dπ ∗ ↓π ∗
α(ξ) ϕ
R ←− TM −→ M −→ Rm
The form ω is closed: dω = 0 since d ◦ d = 0 and it is non-degenerate. To

show this last property, just note that the form is well defined independently of the
chosen coordinates but we can also show it using a direct calculation. Indeed, let
ξ = (ξ1 , ..., ξ2m ) ∈ Tp M and η = (η1 , ..., η2m ) ∈ Tp M . We have

m
m
ω(ξ, η) = dxk ∧ dyk (ξ, η) = (dxk (ξ)dyk (η) − dxk (η)dyk (ξ)) .
k=1 k=1
Since dxk (ξ) = ξm+k is the (m + k)th component of ξ and dyk (ξ) = ξk is the
kth component of ξ, then
⎛ ⎞

m η1
O −I ⎜ .. ⎟
ω(ξ, η) = (ξm+k ηk − ηm+k ξk ) = (ξ1 ...ξ2m ) ⎝ . ⎠,
I O
k=1 η2m
with O the null matrix and I the unit matrix of order m. Then, for all x ∈ M and for
all ξ = (ξ1 , ..., ξ2m ) = 0, it exists η = (ξm+1 , ..., ξ2m , −ξ1 , ..., −ξm ) such that:
m 2
ω(ξ, η) = k=1 ξm+k − ξk2 = 0, because ξk = 0, ∀k = 1, ..., 2m. In the local
coordinate system (x1 , ..., xm , y1 , ..., ym ), this symplectic form is written as

n
ω = k=1 dxk ∧ dyk , which completes the proof.
A manifold M is said to be orientable if there exists on M an atlas such that the

Jacobian of any change of chart is strictly positive or if M has a volume form (i.e. a
differential form that does not vanish anywhere). For example, Rn is oriented by the
volume form dx1 ∧ ... ∧ dxn . The circle S 1 is oriented by dθ. The torus T 2 = S 1 × S 1
is oriented by the volume form dθ ∧ dϕ. All holomorphic manifolds are orientable.
T HEOREM 1.2.– (a) A closed differential 2-form ω on a differentiable manifold M of

dimension 2m is symplectic, if and only if, ω m is a volume form. (b) Any symplectic
manifold is orientable. (c) Any orientable manifold of dimension two is symplectic.
However, in even dimensions larger than 2, this is no longer true.
P ROOF.– (a) This is due to the fact that the non-degeneracy of ω is equivalent to the
fact that ω m is never zero. (b) We have ω = dx1 ∧ dxm+1 + · · · + dxm ∧ dx2m in a
system of symplectic charts (x1 , ..., x2m ). Therefore, ω m = dx1 ∧dxm+1 ∧...∧dxm ∧
m(m−1)
dx2m = (−1) 2 dx1 ∧ dx2 ∧ ... ∧ dx2m , which means that the 2m-form ω m is a
volume form on the manifold M and therefore this one is orientable. The orientation
associated with the differential form ω is the canonical orientation of R2m . (c) This
results from the fact that any differential 2-form on a 2-manifold is always closed.
T HEOREM 1.3.– Let α be a differential 1-form on the manifold M and α∗ λ the

reciprocal image of the Liouville form λ on the cotangent bundle T ∗ M . Then,
α∗ λ = α.
P ROOF.– Since α : M −→ T ∗ M , we can consider the reciprocal image that we

note α∗ : T ∗ T ∗ M −→ T ∗ M , of λ : T ∗ M −→ T ∗ T ∗ M (Liouville form), such
that, for any vector ξ tangent to M , we have the relation α∗ λ(ξ) = λ(α)(dαξ). Since
dα is an application T M −→ T T ∗ M , then α∗ λ(ξ) = λ(α)(dαξ) = λα (dαξ) =
αdπ ∗ dα(ξ) = αd(π ∗ α)(ξ) = α(ξ), because π ∗ α(p) = p where p ∈ M and the result
follows.
A submanifold N of a symplectic manifold M is called Lagrangian if for all p ∈

N , the tangent space Tp N coincides with the following configuration space {η ∈
Tp M : ωp (ξ, η) = 0, ∀ξ ∈ Tp N }. On this space, the 2-form dxk ∧ dyk that defines
the symplectic structure is identically zero. Lagrangian submanifolds are considered
among the most important submanifolds of symplectic manifolds. Note that dim N =
2 dim M and that for all vector fields X, Y on N , we have ω(X, Y ) = 0.
1
E XAMPLE 1.3.– If (x1 , ..., xm , y1 , ..., ym ) is a local coordinate system on an open

U ⊂ M , then the subset of U defined by y1 = · · · = ym = 0 is a Lagrangian
submanifold of M . The submanifold α(M ) is Lagrangian in T ∗ M if and only if the
form α is closed because 0 = α∗ ω = α∗ (−dλ) = −d(α∗ λ) = −dα.
Let M be a differentiable manifold, T ∗ M its cotangent bundle with the symplectic

form ω, sα : U −→ T ∗ M , p −→ α(p), a section on an open U ⊂ M . From the
local expression of ω (theorem 1.6), we deduce that the null section of the bundle
T ∗ M is a Lagrangian submanifold of T ∗ M . If sα (U ) is a Lagrangian submanifold
of T ∗ M , then sα is called the Lagrangian section. We have (theorem 1.8), s∗α λ = α,
and according to the previous example, sα (U ) is a Lagrangian submanifold of T ∗ M
if and only if the form α is closed. Let (M, ω), (N, η) be two symplectic manifolds
of the same dimension and f : M −→ N , a differentiable application. We say that
f is a symplectic morphism if it preserves the symplectic forms, that is, if f satisfies
f ∗ η = ω. When f is a diffeomorphism, we say that f is a symplectic diffeomorphism
or f is a symplectomorphism.
T HEOREM 1.4.– (a) A symplectic morphism is a local diffeomorphism. (b) A

symplectomorphism preserves the orientation.
P ROOF.– (a) Indeed, since the 2-form ω is non-degenerate, then the differential df (p) :
Tp M −→ Tp N , p ∈ M , is a linear isomorphism and according to the local inversion
theorem, f is a local diffeomorphism. Another proof is to note that f ∗ η m = (f ∗ η)m =
ω m . The map f has constant rank 2m because ω m and η m are volume forms on M and
N , respectively. And the result follows. (b) It is deduced from (a) that the symplectic
diffeomorphisms or symplectomorphisms preserve the volume form and therefore the
orientation. The Jacobian determinant of the transformation is +1.
R EMARK 1.1.– Note that the inverse f −1 : N −→ M of a symplectomorphism f :

M −→ N is also a symplectomorphism.
Let (M, ω), (N, η) be two symplectic manifolds, pr1 : M × N −→ M , pr2 :

M × N −→ N , the projections of M × N on its two factors. The forms pr1∗ ω + pr2∗ η
and pr1∗ ω − pr2∗ η on the product M × N are symplectic forms. Take the case where
dim M = dim N = 2m and consider a differentiable map f : M −→ N , as well
as its graph defined by the set A = {(x, y) ∈ M × N : y = f (x)}. The application
g defined by g : M −→ A, x −→ (x, f (x)) is a diffeomorphism. The set A is a
2m-dimensional Lagrangian submanifold of (M × N, pr1∗ ω − pr2∗ η) if and only if
the reciprocal image of pr1∗ ω − pr2∗ η by application g is the identically zero form on
M . For the differentiable map f to be a symplectic morphism, it is necessary and
sufficient that the graph of f is a Lagrangian submanifold of the product manifold
(M × N, pr1∗ ω − pr2∗ η).
T HEOREM 1.5.– (a) Let f : M −→ M be a diffeomorphism. Then, the application

f ∗ : T ∗ M −→ T ∗ M is a symplectomorphism. (b) Let g : T ∗ M −→ T ∗ M be a
diffeomorphism such that: g ∗ λ = λ. Then, there is a diffeomorphism f : M −→ M
such that: g = f ∗ .
P ROOF.– (a) Let us show that f ∗∗ ω = ω. We have
f ∗∗ λ(α)(ξα ) = λ(f ∗ (α))(df ∗ ξα ) = f ∗ (α)dπ ∗ df ∗ (ξα ) = α(df dπ ∗ df ∗ (ξα )),
and therefore, f ∗∗ λ(α)(ξα ) = α(d(f ◦ π ∗ ◦ f ∗ )(ξα )). Since f ∗ α = αf −1 (p) and

π ∗ f ∗ α = f −1 (p), then f ◦ π ∗ ◦ f ∗ (α) = p = π ∗ α, that is,
f ◦ π∗ ◦ f ∗ = π∗ [1.1]
and f ∗∗ λ(α)(ξα ) = α(dπ ∗ (ξα )) = λα (ξα ) = λ(α)(ξα ). Consequently, f ∗∗ λ = λ

and f ∗∗ ω = ω. (b) Since g ∗ λ = λ, then g ∗ λ(η) = λ(dgη) = ω(ξ, dgη) = λ(η) =
ω(ξ, η). Moreover, we have g ∗ ω = ω, hence ω(dgξ, dgη) = ω(ξ, η) = ω(ξ, dgη) and
ω(dgξ −ξ, dgη) = 0, ∀η. Since the form ω is non-degenerate, we deduce that dgξ = ξ
and that g preserves the integral curves of ξ. On the null section of the tangent bundle
(i.e. on the manifold), we have ξ = 0 and then g|M is an application f : M −→ M .
Let us show that: f ◦ π ∗ ◦ g = π ∗ = f ◦ π ∗ ◦ f ∗ . Indeed, taking the differential, we get
df ◦dπ ∗ ◦dg(ξ) = df ◦ dπ ∗ (ξ) = df (ξp ), because dg(ξ) = ξ and ξp ≡ dπ ∗ (ξ)), hence
df ◦ dπ ∗ ◦ dg(ξ) = ξp = dπ ∗ (ξ). Therefore, df ◦ dπ ∗ ◦ dg = dπ ∗ , f ◦ π ∗ ◦ g = π ∗ .
Since f ◦ π ∗ ◦ f ∗ = π ∗ (according to [1.1]), so g = f ∗ .
T HEOREM 1.6.– Let I : Tx∗ M −→ Tx M , ωξ1 −→ ξ, where ωξ1 (η) = ω (η, ξ),
∀η ∈ Tx M . Then I is an isomorphism generated by the symplectic form ω.
P ROOF.– Denote by I −1 the map I −1 : Tx M −→ Tx∗ M , ξ −→ I −1 (ξ) ≡ ωξ1 , with

I −1 (ξ)(η) = ωξ1 (η) = ω(η, ξ), ∀η ∈ Tx M . The form ω being bilinear, then we
have I −1 (ξ1 + ξ2 )(η) = I −1 (ξ1 )(η) + I −1 (ξ2 )(η), ∀η ∈ Tx M . To show that I −1
is bijective, it suffices to show that it is injective (because dim Tx M = dim Tx∗ M ).
The form ω is non-degenerate, and it follows that KerI −1 = {0}. Hence, I −1 is an
isomorphism and consequently I is also an isomorphism.
1.4. Vectors fields and flows

Let M be a differentiable manifold of dimension m. Let T M = x∈M Tx M ,
be the bundle tangent to M (union of spaces tangent to M at all its points x). This
bundle has a structure of a differentiable manifold of dimension 2m and allows us to
immediately transfer to manifolds the theory of ordinary differential equations.
D EFINITION 1.4.– A vector field (also called a tangent bundle section) on M is an

application, noted by X, which at any point x ∈ M associates a tangent vector
Xx ∈ Tx M . In other words, it is an application: X : M −→ T M , such that if
π : T M −→ M , is the natural projection, then we have π ◦ X = idM .
Figure 1.1. Vector field
Figure 1.2. Tangent space
Note that the diagram
X
M −→ T M
idM ↓π
M
is commutative. In a local coordinate system (x1 , ..., xm ) in a neighborhood U ⊂ M ,

m
the vector field X is written in the form X = k=1 fk (x) ∂x∂ k , x ∈ U , where the
functions f1 , . . . , fm : U −→ R, are the components of X with respect to
(x1 , ..., xm ). A vector field X is differentiable if its components fk (x) are
differentiable functions. This definition of differentiability does not obviously depend
on the choice of the local coordinate system. Indeed, if (y1 , ..., ym ) is another local
m
coordinate system in U , then X = ∂
k=1 hk (x) ∂yk , x ∈ U , where
h1 , . . . , hm : U −→ R, are the components of X in relation to (y1 , ..., ym ) and the
m
result follows from the fact that hk (x) = l=1 ∂x∂y k
l fl (x
), x ∈ U . To the vector field
X corresponds to a system of differential equations
dx1 dxm
= f1 (x1 , ..., xm ) , ..., = fm (x1 , ..., xm ) . [1.2]
dt dt
D EFINITION 1.5.– A differentiable vector field X over M is called a dynamical
system.
A vector field is written locally in the form [1.2].

D EFINITION 1.6.– An integral curve (or trajectory) of a vector field X is a

differentiable curve γ : I −→ M , t −→ γ(t), such that ∀t ∈ I, dγ(t)
dt = X (γ(t)),
where I is an interval of R.
m ∂
If k=1 fk (x) ∂xk is the local expression of X, then the integral curves (or
trajectories) of X are the solutions γ(t) = {xk (t)} of [1.2]. We assume in the
following that the vector field X is differentiable (of class C ∞ ) and with compact
support (i.e. X is zero outside of a compact of M ). This will especially be the case if
the manifold M is compact. Given a point x ∈ M , we denote by gtX (x) (or quite
simply gt (x)) the position of x after a displacement of a duration t ∈ R.
Figure 1.3. Flow
We therefore have an application gtX : M −→ M , t ∈ R, which is a

diffeomorphism (a one-to-one differentiable mapping with a differentiable inverse),
by virtue of the theory of differential equations. More precisely, to the vector field X
we associate a one-parameter group of diffeomorphisms gtX on M , that is, a
differentiable application (of class C ∞ ): M × R −→ M , verifying a group law:
(i) ∀t ∈ R, gtX : M −→ M is a diffeomorphism of M on M . (ii) ∀t, s ∈ R,
X
gt+s = gtX ◦ gsX . Condition (ii) means that the correspondence t −→ gtX is a
homomorphism of the additive group R in the group of diffeomorphisms from M to
X
−1
M . It implies that g−t = gtX , because g0X = idM is the identical transformation
that leaves each point invariant.
D EFINITION 1.7.– The one-parameter group of diffeomorphism gtX on M is called

d X
flow. It admits the vector field X for velocity field dt gt (x) = X gtX (x) , with the
X
initial condition: g0 (x) = x.

Obviously, dtd X
gt (x)t=0 = X(x). So through these formulas gtX (x) is the curve
on the manifold,
which passes through x such that the tangent at each point is the
vector X gtX (x) . We will now see how to construct the flow gtX over the whole
variety M .
T HEOREM 1.7.– The vector field X generates a unique one-parameter group of

diffeomorphism of M .
X
d X
X of gt for small t. For x fixed, the differential
P ROOF.– a) Construction
X
equation
g
dt t (x) = X g t (x) , function of t with the initial condition: g 0 (x) = x, admits a
unique solution gtX defined in the neighborhood of the point x0 and depending on the
initial condition C ∞ . So gtX is locally a diffeomorphism. Therefore, for each point
x0 ∈ M, we can find a neighborhood U (x0 ) ⊂ M , a positive real number ε ≡ ε (x0 )
such that for all t ∈ ]−ε, ε[, the differential equation in question with its initial
condition admits a unique solution differentiable gtX (x) defined in U (x0 ) and
X
verifying the group relation gt+s (x) = gtX ◦ gsX (x), with t, s, t + s ∈ ]−ε, ε[. Indeed,
X
let us pose x1 = gt (x), t fixed and consider the solution of the differential equation
X
satisfying in the neighborhood of the point x0 to the initial condition gs=0 = x1 . This
solution satisfies the same differential equation and coincides at a point gtX (x) = x1 ,
X
with the function gt+s . Therefore, by uniqueness of the solution of the differential
equation, the two functions are locally equal. Therefore, the application gtX is locally
a diffeomorphism. We recall that the vector field X is supposed to be differentiable
(of class C ∞ ) and with compact support K. From the open cover of K formed by
U (x), we can extract a finite subcover (Ui ), since K is compact. Let us denote εi by
the numbers ε corresponding to Ui and put ε0 = inf (εi ), gtX (x) = x, x ∈ / K.
Therefore, the differential equation in question admits a unique solution gtX on
M × ]−ε0 , ε0 [ verifying the group relation: gt+s
X
= gtX ◦ gsX , the inverse of gtX being
X X
g−t and so gt is a diffeomorphism for t small enough.
b) Construction of gtX for all t ∈ R. According to (a), it suffices to construct gtX for
t ∈ ]−∞, −ε0 [∪]ε0 , ∞[. We will see that the applications gtX are defined according to
the multiplication law group. Note that t can be written
ofε0the in the form t = k ε20 + r,
with k ∈ Z and r ∈ 0, 2 . Let us consider, for t ∈ R+ and for t ∈ R∗− ,
∗
ε0 ◦ · · · ◦ g ε0 ◦ g
gtX = g X ε0 ◦ · · · ◦ g ε0 ◦ g
X X
r , gtX = g−
X X
− 2
X
r ,
2 2
2
k−times k−times
X X
respectively. The diffeomorphisms g± ε0 and gr were defined in (a), and we deduce
2
that for any real t, gtX is a diffeomorphism defined globally on the manifold M .
C OROLLARY 1.1.– Every solution of the differential equation dx(t) dt = X(x(t)), x ∈

M , with the initial condition x (for t = 0), can be extended indefinitely. The value of
the solution gtX (x) at the instant t is differentiable with respect to t and x.
With a slight abuse of notation, we can write the preceding differential equation
in the form of the system of differential equations [1.2] with the initial conditions
x1 , ..., xm for t = 0. With the vector field X, we associate the first-order differential
operator LX . We refer here to the differentiation of functions in the direction
of the

field X. We have LX : C ∞ (M ) −→ C ∞ (M ), F −→ LX F (x) = dt d
F gtX (x) t=0 ,
x ∈ M . Here, C ∞ (M ) designates the set of functions F : M −→ R of class C ∞ . The
operator LX is linear: LX (α1 F1 + α2 F2 ) = α1 LX F1 + α2 LX F2 , (α1 , α2 ∈ R),
and satisfies Leibniz’s formula: LX (F1 F2 ) = F1 LX F2 + F2 LX F1 . Since LX F (x)
only depends on the values of F in the neighborhood of x, we can therefore apply the
operator LX without the need to extend them to the whole manifold M . Let
(x1 , ..., xm ) be local coordinates on M . In this coordinate system, the vector X is
given by its components f1 , . . . , fm and the flow gtX is given by the system of
differential equations [1.2]. So the derivative of the function F = F (x1 , ..., xm ) in
the direction X is LX F = f1 ∂x ∂F
1
+ · · · + fm ∂x
∂F
m
. In other words, in the coordinates
(x1 , ..., xm ) the operator LX has the form LX = f1 ∂x ∂
1
+ · · · + fm ∂x∂m .
D EFINITION 1.8.– We say that two vector fields X1 and X2 on a manifold M commute
(or are commutative) if and only if the corresponding flows commute,
gtX1 1 ◦ gtX2 2 (x) = gtX2 2 ◦ gtX1 1 (x), ∀x ∈ M.
Figure 1.4. Commutative flows
T HEOREM 1.8.– Two vector fields X1 and X2 on a manifold M commute if and only
if, [LX1 , LX2 ] ≡ LX1 LX2 − LX2 LX1 = 0.
P ROOF.– a) Let us first show that the condition is necessary. Note that,

∂2
F gtX2 2 ◦ gtX1 1 (x) − F gtX1 1 ◦ gtX2 2 (x)
∂t1 ∂t2
t2 =t1 =0
= (LX1 LX2 − LX2 LX1 ) F (x), ∀F ∈ C ∞ (M ), ∀x ∈ M.
Indeed, according to the definition of LX2 , we find,

∂
F gt2 ◦ gt1 (x)
X2 X1
= LX2 F gtX1 1 (x) .
∂t2 t2 =0
Thus,

∂2 ∂
F gt2 ◦ gt1 (x)
X2 X1
= LX2 F gtX1 1 (x)
,
∂t1 ∂t2 t2 =t1 =0 ∂t 1 t1 =0

∂
= G gtX1 1 (x) where G ≡ LX2 F,
∂t1 t1 =0
= LX1 G(x) by definition of LX1 ,

= LX1 LX2 F (x) .
Likewise, we have

∂2
F gtX1 1 ◦ gtX2 2 (x) = LX1 F gtX2 2 (x) ,
∂t2 ∂t1 t1 =0
and

∂2
F gtX1 1 ◦ gtX2 2 (x) = LX2 LX1 F (x) .
∂t2 ∂t1 t2 =t1 =0
Therefore,

∂2 ∂2
F gt1 ◦ gt2 (x)
X1 X2
− F gtX1 1 ◦ gtX2 2 (x)

∂t2 ∂t1 t1 =0 ∂t2 ∂t1 t2 =t1 =0

∂ 2
= F gtX2 2 ◦ gtX1 1 (x) − F gtX1 1 ◦ gtX2 2 (x) ,
∂t1 ∂t2
t2 =t1 =0
= LX1 LX2 F (x) − LX2 LX1 F (x) .
t1
So if X1 and X2 commute on M , that is, gX 1
◦ gtX2 2 (x) = gtX2 2 ◦ gtX1 1 (x), ∀x ∈
M , then according to the above formula, (LX1 LX2 − LX2 LX1 ) F (x) = 0, ∀F ∈
C ∞ (M ), ∀x ∈ M . Consequently, LX1 LX2 = LX2 LX1 .
b) Let us show that gtX1 1 ◦ gtX2 2 (x) = gtX2 2 ◦ gtX1 1 (x), ∀x ∈ M , that is, that the
condition is sufficient, or that: F gtX1 1 ◦ gtX2 2 (x) = F gtX2 2 ◦ gtX1 1 (x) ,
∀F ∈ C ∞ (M ), ∀x ∈ M . Let us pose ξ = gtX1 1 ◦ gtX2 2 (x), ζ = gtX2 2 ◦ gtX1 1 (x), and
develop in Taylor series the function F (ξ) − F (ζ) at the neighborhood of

t1 = t2 = 0. We have,

∂
F (ξ) − F (ζ) = F (x) − F (x) + t1 (F (ξ) − F (ζ))
∂t1 t1 =t2 =0
2
∂ 2
t1 ∂
+t2 (F (ξ) − F (ζ)) + (F (ξ) − F (ζ))
∂t2 2 ∂t1 2
t1 =t2 =0 t1 =t2 =0

2
t2 ∂ 2 ∂ 2
+ (F (ξ) − F (ζ)) + t1 t2 (F (ξ) − F (ζ))
2 ∂t2 2 ∂t1 ∂t2
t1 =t2 =0 t1 =t2 =0
3 3 2 2

+o t1 , t2 , t1 t2 , t1 t2 .
Let us calculate the different terms. We have

∂ ∂
F (ξ) = F gtX1 1 ◦ gtX2 2 (x)

∂t1 t1 =t2 =0 ∂t1 t1 t2 =0

= = Lx1 F gtX2 2 (x) = Lx1 F (x),
t2 =0
and

∂ ∂
F (ζ) = F gtX2 2 ◦ gtX1 1 (x)
,
∂t1 t1 =t2 =0 ∂t 1 t1 =t2 =0

∂
= G gt1 (x)
X1
where G = F gtX2 2 ,
∂t1 t1 =0 t2 =0

= Lx1 G(x) = Lx1 F gtX2 2 = Lx1 F (x).
t2 =0

Therefore, we have ∂
∂t1 (F (ξ) − F (ζ)) = 0. By symmetry, we also have
t1 =t2 =0

∂
∂t2 (F (ξ) − F (ζ)) = 0. Likewise, we have
t1 =t2 =0

∂2 ∂2
2 (F (ξ) − F (ζ)) = X1
2 F gt1 ◦ gt2 (x)
X2
∂t1 t1 =t2 =0 ∂t 1

−F gtX2 2 ◦ gtX1 1 (x) .
t1 =t2 =0
Now ∂
∂t1 F gtX1 1 ◦ gtX2 2 (x) = ∂
∂t1 F gtX1 1 (y) = LX1 F gtX1 1 (y) , where y =
gtX2 2 (x), so
∂2 ∂
F gtX1 1 ◦ gtX2 2 (x) = LX1 F gtX1 1 (y) = LX1 LX1 F gtX1 1 (y) ,
∂t21 ∂t1
= LX1 LX1 F gtX1 1 ◦ gtX2 2 (x) −→ LX1 LX1 F (x).

t1 =t2 =0
Likewise, we have ∂
∂t1 F gtX2 2 ◦ gtX1 1 (x) = ∂
∂t1 G gtX1 1 (x) =
LX1 G gtX1 1 (x) , where G = F gtX2 2 , hence
∂2 ∂
2 F gtX2 2 ◦ gtX1 1 (x) = LX1 G gtX1 1 (x) = LX1 LX1 G gtX1 1 (x) ,
∂t1 ∂t1
= LX1 LX1 F gtX2 2 ◦ gtX1 1 (x) −→ LX1 LX1 F (x).

t1 =t2 =0

∂2
Thus, (F (ξ) − F (ζ))
∂t21
= 0. It follows, by symmetry, that
t1 =t2 =0
∂2
∂t22
(F (ξ) − F (ζ)) = 0. Moreover, we deduce from the necessary condition
t1 =t2 =0
and from the fact that the vector fields X1 and X2 commute the following relation:
∂2
(F (ξ) − F (ζ))|t1 =t2 =0
∂t1 ∂t2

∂2
= F gtX1 1 ◦ gtX2 2 (x) − F gtX2 2 ◦ gtX1 1 (x) ,
∂t1 ∂t2
t1 =t2 =0
∂2
= (LX2 LX1 − LX1 LX2 ) F (x) = 0.
∂t1 ∂t2

Therefore, F gtX1 1 ◦ gtX2 2 (x) − F gtX2 2 ◦ gtX1 1 (x) = o t31 , t32 , t21 t2 , t1 t22 .
Consider the times t1 and t2 of the order ε. We find a difference between the two
new points of the manifold, depending on whether we apply the field X1 before the
field X2 or the inverse, of the order of ε3 . F gtX1 1 ◦ gtX2 2 (x) −

F gtX2 2 ◦ gtX1 1 (x) = o ε3 . Now, if t1 and t2 are arbitrary fixed times, let us square
the space between the two paths with squares of sides ε. Each square represents the
small space traveled during a small time ε, either according to the field X1 or
according to the field X2 . We have found that when the space between two paths
differs from that of a square, we get a difference ε3 . By modifying the path traveled
by a square in successive stages, we obtain
t1 t2 3
F gtX1 1 ◦ gtX2 2 (x) − F gtX2 2 ◦ gtX1 1 (x) ≤ o ε ,
ε2
by the fact that we have tε1 × tε2 steps intermediaries.

This is valid for all ε; just take
ε small enough, tending to zero, so that t1ε2t2 o ε3 = t1 t2 o (ε) −→ 0.
ε→0
Figure 1.5. Proof of the commutativity of flows
Since every first-order linear differential operator is given by a vector field,

LX2 LX1 − LX1 LX2 being a first-order linear differential operator, the latter also
corresponds to some vector field that we denote by X3 .
D EFINITION 1.9.– The Poisson bracket or commutator of two vector fields X1 and
X2 on the manifold M , denoted by X3 = {X1 , X2 } or X3 = [X1 , X2 ], is the vector
field X3 for which LX3 = LX2 LX1 − LX1 LX2 .
E XAMPLE 1.4.– The Poisson bracket transforms the vector space of vector fields over
a manifold into Lie algebra.
Let X be a vector field on a differentiable manifold M . We have shown (theorem

1.7) that X generates a unique one-parameter group of diffeomorphism gtX (which we
d X
also denote by gt ) on M , solution to the differential equation: dt gt (p) = X(gtX (p)),
p ∈ M , with the initial condition g0 (p) = p. Let ω be a k-form differential on M .
X
D EFINITION 1.10.– The Lie

derivative of ωg∗with respect to X is the k-form differential
d ∗ (ω(gt (p)))−ω(p)
defined by LX ω = dt gt ω t=0 = limt→0 t t .
In general, for t = 0, we have

d ∗ d ∗ ∗ d ∗

gt ω =
gt+s ω = gt gs ω = gt∗ (LX ω). [1.3]
dt ds s=0 ds s=0
For all t ∈ R, the application gt : R −→ R being a diffeomorphism then dgt and

dg−t are the applications, dgt : Tp M −→ Tgt (p) M , dg−t : Tgt (p) M −→ Tp M .
D EFINITION 1.11.– The Lie derivative of a vector field Y in the direction X is defined
d
by LX Y = dt g−t Y t=0 = limt→0 g−t (Y (gt (p)))−Y
t
(p)
.
In general, for t = 0, we have

d d d
g−t Y =
g−t−s Y = g−t g−s Y = g−t (LY ).
dt ds s=0 ds s=0
D EFINITION 1.12.– The interior product of a k-form differential ω by a vector field

X on the differentiable manifold M is a (k − 1)-form differential, iX ω, defined by
(iX ω)(X1 , ..., Xk−1 ) = ω(X, X1 , ..., Xk−1 ), where X1 , ..., Xk−1 are vector fields.
It is easy to show that if ω is a k-differential form, λ a differential form of any

degree, X and Y two vector fields, f a linear map and a a constant, then
iX+Y ω = iX ω + iY ω, iaX ω = aiX ω, iX iY ω = −iY iX ω,
iX iX ω = 0, iX (f ω) = f (iX ω),
iX (ω ∧ λ) = (iX ω) ∧ λ + (−1)k ω ∧ (iX λ), iX f ∗ ω = f ∗ (if X ω),
where f ∗ ω denote the pull-back by f .
E XAMPLE 1.5.– Let us calculate the expression of the interior product in local
m ∂
coordinates. If X = j=1 Xj (x) ∂x j
is the local expression of the vector field on the
manifold M of dimension m and ω = i1 <i2 <...<ik fi1 ...ik (x)dxi1 ∧ ... ∧ dxik is a
k-differential form, then

m
iX ω = ω(X, ·) = fji2 ...ik Xj dxi2 ∧ ... ∧ dxik
i2 <i3 <...<ik j=1

m
− fi1 j...ik Xj dxi1 ∧ dxi3 ∧ ... ∧ dxik + · · ·
i1 <i3 <...<ik j=1

m
+(−1)k−1 fi1 i2 ...j Xj dxi1 ∧ dxi2 ∧ ... ∧ dxik−1 ,
i1 <i2 <...<ik−1 j=1

m
= k fji2 ...ik Xj dxi2 ∧ ... ∧ dxik .
i2 <i3 <...<ik j=1
∂
Hence, i ∂ ω= ∂(dxj ) ω, where we put dxj in the first position in ω.
∂xj
The following properties are often involved in solving practical problems using
Lie derivatives.
P ROPOSITION 1.1.– a) If f : M −→ R is a differentiable function, then the Lie

derivative of f is the image of X by the differential of f , LX f = df (X) = X.f .
b) LX and d commute, LX ◦ d = d ◦ LX . c) Let X, X1 , ..., Xk be vector fields on M
and ω a k-form differential. So

k
(LX ω)(X1 , ..., Xk ) = LX (ω(X1 , ..., Xk )) − ω(X1 , ..., LX Xj , ..., Xk ).
j=1
d) For all differential forms ω and λ, LX (ω ∧ λ) = LX ω ∧ λ + ω ∧ LX λ.
P ROOF.– a) Indeed, we have

d ∗ d
dgt

LX f = gt f = f ◦ gt = df = df (X),
dt t=0 dt t=0 dt t=0
and (see theorem 1.9), LX f = iX df = X.f , hence the result.
b) Indeed, as the differential and the inverse image commute, then

d ∗ d ∗
d ◦ LX ω = d ◦ g t ω = gt ◦ dω = LX ◦ dω.
dt t=0 dt t=0
c) We have

d ∗
(LX ω)(X1 , ..., Xk ) = gt ω(X1 , ..., Xk ) ,
dt t=0

d
= ω(gt )(dgt X1 , ..., dgt Xk ) ,
dt t=0
= LX ω(gt )(dgt X1 , ..., dgt Xk )|t=0

k
d
+ ω(gt ) dgt X1 , ..., dgt Xj , ..., dgt Xk ,
dt
j=1
t=0
and the result is deduced from the fact that

d d
dgt Xj = − dg−t Xj = −LX Xj .
dt t=0 dt t=0
d) Just consider ω = f dxi1 ∧ ... ∧ dxik and λ = gdxj1 ∧ ... ∧ dxjl . We have
ω ∧ λ = f gdxi1 ∧ ... ∧ dxik ∧ dxj1 ∧ ... ∧ dxjl , and
LX (ω ∧ λ)(X1 , ..., Xk , Xk+1 , ..., Xk+l )

= (LX f ).gdxi1 ∧ ... ∧ dxik ∧ dxj1 ∧ ... ∧ dxjl (X1 , ..., Xk , Xk+1 , ..., Xk+l )
+f (LX g)dxi1 ∧ ... ∧ dxik ∧ dxj1 ∧ ... ∧ dxjl (X1 , ..., Xk , Xk+1 , ..., Xk+l ),
= ((LX ω) ∧ λ + ω ∧ (LX λ))(X1 , ..., Xk , Xk+1 , ..., Xk+l ),
and the result follows.
P ROPOSITION 1.2.– Let X and Y be two vector fields on M . Then, the Lie derivative
of LX Y is the Lie bracket [X, Y ].
dg−t Y −Y
P ROOF.– We have LX Y (f ) = limt→0 t (f ) = limt→0 dg−t Y −dg
t
tY
(f ),
Y (f )−Y (f ◦g )◦g −1
hence, LX Y (f ) = limt→0 Y (f )−dg
t
t Y (f )
=
limt→0 t
t t
. Put
gt (x) ≡ g(t, x), and apply to g(t, x) the Taylor formula with integral remainder.
So there is h(t, x) such that: f (g(t, x)) = f (x) + th(t, x), with
∂
h(0, x) = ∂t f (g(t, x))(0, x). According to the definition of the tangent vector, we
∂
have X(f ) = ∂t f ◦ gt (x)(0, x), hence h(0, x) = X(f )(x). Therefore,

Y (f ) − Y (f ) ◦ gt−1
LX Y (f ) = lim − Y (h(t, x)) ◦ gt−1 ,
t→0 t

(Y (f ) ◦ gt − Y (f )) ◦ gt−1
= lim − Y (h(t, x)) ◦ gt−1 .
t→0 t
Since limt→0 gt−1 (x) = g0−1 (x) = id., we deduce that:

Y (f ) ◦ gt − Y (f )
LX Y (f ) = lim − Y (h(0, x)) ,
t→0 t
∂
= Y (f ) ◦ gt (x) − Y (X(f )),
∂t
= X(Y (f )) − Y (X(f )) = [X, Y ],
which completes the proof.
We will now establish a fundamental formula for the Lie derivative, which can be
used as a definition.
T HEOREM 1.9.– Let X be a vector field on M and ω a differential k-form. Then

LX ω = d(iX ω) + iX (dω). We have the Cartan homotopy formula LX = d ◦ iX +
iX ◦ d.
P ROOF.– We will reason by induction on the degree k of the differential form ω.

Let DX ≡ d ◦ iX + iX ◦ d. For a differential 0-form, that is, a function f , we have
DX f = d(iX f ) + iX (df ), or iX f = 0, hence d(iX f ) = 0, iX df = df (X), and so
DX f = df (X). We know (proposition 1.1, (a)) that LX f = df (X) = X.f , so
DX f = LX f . Assume that the formula in question is true for a differential
(k − 1)-form and is proved to be true for a differential k-form. Let λ be a differential
(k − 1)-form and let ω = df ∧ λ, where f is a function. We have
LX ω = LX (df ∧ λ) = LX df ∧ λ + df ∧ LX λ, (proposition 1.1, (d)),

= dLX f ∧ λ + df ∧ LX λ, (because LX df = dLX f, proposition 1.1, (b)),
= d(df (X)) ∧ λ + df ∧ LX λ, (because LX f = df (X), proposition 1.1, (a)).
Hypothetically, we have LX λ = d(iX λ) + iX (dλ). Since iX df = df (X), then
LX ω = d(iX df ) ∧ λ + df ∧ d(iX λ) + df ∧ iX (dλ). [1.4]
Moreover, we have
iX dω = iX d(df ∧ λ) = −iX (df ∧ dλ) = −(iX df )d ∧ dλ + df ∧ iX (dλ),
and
d(iX ω) = diX (df ∧ λ) = d ((iX df ) ∧ λ − df ∧ (iX λ)) ,

= d(iX df ) ∧ λ + (iX df ) ∧ dλ + df ∧ d(iX λ),
because d(df ) = 0. Hence,
diX (df ∧ λ) + iX d(df ∧ λ) = d(iX df ) ∧ λ + df ∧ d(iX λ) + df ∧ iX (dλ).
Comparing this expression with that obtained in [1.4] above, we finally obtain
LX ω = d(iX ω) + iX (dω), and the theorem is proved.
P ROPOSITION 1.3.– For a differential form ω, we have iX LX ω = LX iX ω.
P ROOF.– Taking into account the previous theorem and the fact that iX iX = 0, we
obtain
iX LX ω = iX (diX ω) + iX (iX dω) = iX (diX ω),

LX iX ω = (d ◦ iX + iX ◦ d)iX ω = iX (diX ω),
hence iX LX ω − LX iX ω = 0 and the proof ends.

P ROPOSITION 1.4.– Let X and Y be two vector fields on M and ω a differential form.
Then, LX+Y ω = LX ω + LY ω, Lf X ω = f LX ω + df ∧ iX ω, where f : M −→ R is
a differentiable function.
P ROOF.– Indeed, just use theorem 1.9,
LX+Y ω = d(iX+Y ω) + iX+Y (dω) = d(iX ω + iY ω) + iX (dω) + iY (dω),

= d(iX ω) + iX (dω) + d(iY ω) + iY (dω) = LX ω + LY ω.
Similarly, we have
Lf X ω = d(if X ω) + if X (dω) = d(f iX ω) + f iX (dω),

= df ∧ iX ω + f d(iX ω) + f iX (dω) = df ∧ iX ω + f LX ω,
which completes the proof.
E XAMPLE 1.6.– The expression of the Lie derivative of the differential form ω =
i1 <...<ik fi1 ...ik dxi1 ∧ ... ∧ dxik in local coordinates is given by
m
∂fi 1 ...ik ∂Xj
LX ω = Xj + kfji2 ...ik dxi1 ∧ ... ∧ dxik .
i1 <...<ik j=1
∂xj ∂xi1
m∂
Indeed, if X = j=1 Xj (x) ∂x j
is the local expression of the vector field on the
m-dimensional manifold M , then

m m

LX ω = L Xj ∂ ω= dXj ∧ i ∂ ω + Xj L ∂ ω .
∂xj ∂xj ∂xj
j=1 j=1
According to example 1.5, we know that
∂
i ∂ ω= ω=k fji2 ...ik dxi2 ∧ dxi3 ∧ ... ∧ dxik ,
∂xj ∂(dxj ) i 2 <i3 <...<ik
∂X
hence, dXj ∧ i ∂ ω=k i1 <i2 <...<ik fji2 ...ik ∂xij dxi1 ∧ ... ∧ dxik . Similarly, using
∂xj 1
∂fi1 ...ik
proposition 1.1, (c), we obtain L ∂ ω = i1 <...<ik ∂xj dxi1 ∧ ... ∧ dxik . Since
∂xj
∂ ∂
,
∂xj ∂xl = 0, we finally get the result.
P ROPOSITION 1.5.– We have [LX , iY ] = i[X,Y ] , [LX , LY ] = L[X,Y ] , where X and

Y are two vector fields on M .
P ROOF.– The proof is to show that for a differential k-form ω, we have [LX , iY ] ω =
i[X,Y ] ω and [LX , LY ] ω = L[X,Y ] ω. We reason by induction assuming first that k =
1, that is, ω = df . We have [LX , iY ]df = LX iY df −iY LX df = LX (Y.f )−iY dLX f ,
because LX ◦ d = d ◦ LX ). Since LX f = X.f , we have [LX , iY ]df = X.(Y.f ) −
iY d(X.f ) = X.(Y.f ) − Y.(X.f ), = [X, Y ].f = i[X,Y ] df . Suppose the formula in
question is true for a form ω of degree less than or equal to k − 1. Let λ and θ be two
forms of degree less than or equal to k − 1, so that ω = λ ∧ θ is a form of degree k.
We have
[LX , iY ]ω = LX iY ω − iY LX ω = LX iY (λ ∧ θ) − iY LX (λ ∧ θ),
= LX (iY λ ∧ θ + (−1)degλ λ ∧ iY θ) − iY (LX λ ∧ θ + λ ∧ LX θ),
= LX iY λ ∧ θ + iY λ ∧ LX θ + (−1)degλ LX λ ∧ iY θ
+(−1)degλ λ ∧ LX iY θ − iY LX λ ∧ θ − (−1)degλ LX λ ∧ iY θ
−iY λ ∧ LX θ − (−1)degλ λ ∧ iY LX θ,
= (LX iY λ − iY LX λ) ∧ θ + (−1)degλ λ ∧ (LX iY θ − iY LX θ),
= i[X,Y ] λ ∧ θ + (−1)degλ λ ∧ i[X,Y ] θ,
= i[X,Y ] (λ ∧ θ) = i[X,Y ] ω.
Similarly, we have
[LX , LY ]ω = LX LY ω − LY LX ω = LX diY ω + LX iY dω − diY LX ω − iY dLX ω,

= dLX iY ω + LX iY dω − diY LX ω − iY LX dω, (LX dω = dLX ω)
= di[X,Y ] ω + i[X,Y ] dω, (according to [LX , iY ] = i[X,Y ] ) = L[X,Y ] ω,
and the proof ends.
R EMARK 1.2.– Using the results above, we give a quick proof of the Poincaré lemma:
in the neighborhood of a point of a manifold, any closed differential form is exact.
Consider the differential equation in Rn , ẋ = Xt (x) = xt , whose solution gt (x0 ) =
x0 t is a one-parameter group of diffeomorphisms. That proof is based on the properties
of this differential equation, whose solution is a family of maps gt (x) = x0 t, t ∈]0, 1].
It can also be regarded as a time dependent family of vector fields X(t). Note that
they are only defined for t > 0. For t = 0, the right part of that equation is undefined.
Nevertheless, that family gt (x) = x0 t can be defined for all t ∈ [0, 1], while for
t = 0, the map g0 (x) = 0 is a constant map. We have g0 (x0 ) = 0, g1 (x0 ) = x0 ,
1 d ∗ 1
g0∗ ω = 0, g1∗ ω = ω, and ω = g1∗ ω − g0∗ ω = 0 dt gt ωdt = 0 gt∗ (LX ω)dt, according
1
to [1.3]. By theorem 1.9 and the fact that dω = 0, we have ω = 0 gt∗ (diX ω)dt =
1 ∗
dgt iX ωdt, because df ∗ ω = f ∗ dω). We can, therefore, find a differential form λ
0 1
such that: ω = dλ, where λ = 0 gt∗ iX ωdt, which completes the proof.
1.5. The Darboux theorem
We will study a central theorem of symplectic geometry, namely Darboux’s

theorem, which states that every point in a symplectic manifold has a neighborhood
with the so-called Darboux coordinates. The Darboux theorem plays a central role in
symplectic geometry; the symplectic manifolds (M, ω) of dimension 2m are locally
isomorphic to (R2m , ω). More precisely, if (M, ω) is a symplectic manifold of
dimension 2m, then in the neighborhood of each point of M , there exist local
m
coordinates (x1 , ..., x2m ) such that: ω = k=1 dxk ∧ dxm+k . We call such
coordinates Darboux coordinates. In particular, there is no local invariant in
symplectic geometry, analogous to the curvature in Riemannian geometry. The
classical proof given by Darboux is by induction on the dimension of the manifold
and is accessible in different versions in the basic literature (see, for example, Arnold
1989; Sternberg 1983). We will give an overview in remark 1.3. The proof presented
in this section, however, is due to J. Moser (consult Guillemin and Sternberg 1984)
and is based on standard differential calculus. The idea of proof, known as the
Moser’s trick, works in many situations. The Moser’s trick involves constructing an
appropriate isotopy gt generated by a time-dependent vector field Xt on M such that
gt∗ ωt = ω0 . Let us also point out the information that Givental (see Arnold and
Givental 1990) proved, that a germ of a submanifold in a symplectic manifold is
defined, up to a symplectomorphism, by the restriction of the symplectic structure to
the tangent bundle of the submanifold. This theorem constitutes a generalization of
Darboux’s theorem. The latter corresponds to the particular case when the
submanifold is a point. Givental’s theorem is closely related to the following
generalization of Darboux’s theorem obtained by Weinstein (1971): a submanifold of
a symplectic space is defined, up to a symplectomorphism of a neighborhood of it, by
the restriction of the symplectic form to the bundle of tangent vectors of the ambient
space at the point of the submanifold. Weinstein’s theorem is global in the sense that
it gives a conclusion on the whole submanifold. Givental’s theorem is local; however,
it requires only interior information, whereas in order to be able to apply Weinstein’s
theorem we must know some exterior information (namely, the values of the
symplectic form on non-tangential elements). We will first give a proof of Moser
(1965) and then prove Darboux’s theorem based on this lemma.
L EMMA 1.1.– Let {ωt }, 0 ≤ t ≤ 1, be a family of symplectic forms, differentiable in

t. Then, for all p ∈ M , there exists a neighborhood U of p and a function gt : U −→ U ,
such that: g0∗ = identity and gt∗ ωt = ω0 .
P ROOF.– Let us first remember that we have shown in theorem 1.7 that the vector
field generates a unique one-parameter group of diffeomorphism of M . Looking for a
family of vector fields Xt on U such that these fields generate locally a one-parameter
d
group of diffeomorphisms gt with dt gt (p) = Xt (gt (p)), g0 (p) = p. First note that
d d d
the form ωt is closed (i.e. dωt = 0) as the form dt ωt (since d dt ωt = dt dωt = 0).
∗
Therefore, by deriving the relationship gt ωt = ω0 and using the Cartan homotopy
formula LXt = iXt d + diXt , taking into account

dωt depends ontime, we obtain
that
d ∗
the expression dt gt ωt = gt∗ dt
d
ωt + LXt ωt = gt∗ dt ωt + diXt ωt . By Poincaré’s
lemma (in the neighborhood of a point, any closed differential form is exact), the form
∂
∂t ωt is exact in the neighborhood of p. In other words, we can find a form λt such
d
that: dt ωt = dλt . Hence,
d ∗
g ωt = gt∗ d(λt + iXt ωt ). [1.5]
dt t
We want to show that for all p ∈ M , there exists a neighborhood U of p and a
function gt : U −→ U , such that: g0∗ = identity and gt∗ ωt = ω0 , therefore,
d ∗
dt gt ωt = 0. By [1.5], the problem amounts to finding Xt such that: λt + iXt ωt = 0.
Since the form ωt is non-degenerate, the above equation is solvable with respect to
the vector field Xt and (according to theorem 1.7) defines the family {gt } for
0 ≤ t ≤ 1. In local coordinates (xk ) of the 2m-dimensional manifold M , with
∂
∂xk a basis of T M and (dxk ) the dual basis of ∂x∂ k , k = 1, ..., 2m, we have
2m 2m ∂
λt = k=1 λk (t, x)dxk , Xt = k=1 Xk (t, x) ∂xk , ωt =
2m 2m 2m
k,l=1 ωk,l (t, x)dxk ∧ dxl , iXt ωt = 2 l=1 k=1 ωk,l Xk dxl . We, therefore,
k<l
solve the system of equations in xk (t, x) according to:
2m
λl (t, x) + 2 k=1 ωk,l (t, x)Xk (t, x) = 0. The form ωt is non-degenerate and the
matrix (ωk (t, x)) is non-singular. Then the above system has a unique solution. This
determines the vector field Xt and thus functions gt∗ such that: gt∗ ωt = ω0 , which
completes the proof.
Using the above lemma, we give a proof (we will proceed with “Moser’s trick”) of
the Darboux theorem (sometimes also referred to as the Darboux–Weinstein theorem).
T HEOREM 1.10.– Any symplectic form on a manifold M of dimension 2m is locally

diffeomorphic to the standard form on R2m . In other words, if (M, ω) is a symplectic
manifold of dimension 2m, then in the neighborhood of each point p of M , there exist
m
local coordinates (x1 , ..., x2m ) such that: ω = k=1 dxk ∧ dxm+k .
P ROOF.– In theorem 1.7, we have seen that the vector field X generates a unique one-
d
parameter group of diffeomorphism gtX of M with dt gt (p) = Xt (gt (p)), g0 (p) =
p. We have given a detailed proof on the construction of the flow gtX for small t
and for all t ∈ R. In the proof of Darboux’s theorem, one can be satisfied with the
symplectic forms defined on an open neighborhood of 0 ∈ R2m . Indeed, since the
statement is local, without loss of generality, we may assume that M = R2m and
p = 0. If we have a symplectic manifold (M, ω) and a point p, we can take a smooth
coordinate chart about p and then use the coordinate function to push ω forward to
a symplectic form on a neighborhood of 0 in R2m . If the result holds on R2m , we
can compose the coordinate chart with the resulting symplectomorphism to get the
theorem in general. Let {ωt }, 0 ≤ t ≤ 1, be a family of 2-differential forms that
m
depends differentiably on t and let ωt = ω0 + t(ω − ω0 ), ω0 = k=1 dxk ∧ dm+k ,
where (x1 , ..., x2m ) are local coordinates on R . These ωt ’s are volume forms. Note
2m
that these 2-forms are closed. We have ω0 (0) = ω(0), so ωt (0) = ω0 (0). Since the
function det :[0, 1] × R2m −→ R, (t, p) −→ det ωt (p) is continuous and as ω0 is
non-degenerate, we can find a small neighborhood of [0, 1] × {0} in R × R2m where
the form ωt (0) is non-degenerate for all t ∈ [0, 1] (in short, ωt is non-degenerate in
a neighborhood of 0 and independent of t at 0). Since the interval [0, 1] is compact
and R2m is locally compact, we can find a neighborhood of 0 on which all the ωt
are symplectic forms. In other words, ωt are symplectic forms and by lemma 1.1,
for all 0 ∈ R2m , there exists a neighborhood U of 0 and a function gt : U −→ U
such that: gt∗ = identity and gt∗ ωt = ω0 . Differentiating this relation
d with respect
to
d ∗
t, we obtain (as in the proof of lemma 1.1) dt gt ωt = 0, gt∗ dt ωt + LXt ωt = 0,
d
gt∗ dt ωt + diXt ωt = 0. Therefore, diXt ωt = − dt d
ωt and since the form dt d
ωt is
exact in the neighborhood of 0 (indeed, the forms ω0 and ω are closed, so it is the
same for ω0 − ω and we can therefore, according to Poincaré’s lemma, find in the
neighborhood of 0 a 1-form θt such that: dt d
ωt = ω0 − ω = dθt on U , and by
adding an exact 1-form if necessary, we can achieve that θt (0) = 0). Since θt (0) = 0,
we see that Xt (0) = 0, that is 0 is a fixed point of gt . Then diXt ωt = dθt , where
θt is a 1-differential form. In addition, with ωt being non-degenerate, the equation
iXt ωt = θt is solvable and determines solely the vector field Xt depending on t. Note
that for t = 1, ω1 = ω and for t = 0, ω0 = ω0 , and also we can find g1∗ such
that: g1∗ ω = ω0 . According to theorem 1.7, vector fields Xt generate one-parameter
families of diffeomorphisms {gt }, 0 ≤ t ≤ 1. In other words, you can make a change
m
of coordinates such as: ω = k=1 dxk ∧ dm+k , and the proof is completed.
R EMARK 1.3.– As we pointed out at the beginning of this section, we will give an
overview on the classical proof given by Darboux of his theorem. We proceed by
induction on m. Suppose the result is true for m − 1 ≥ 0 and show that it is also for
m. Fix x and let xm+1 be a differentiable function on M whose differential dxm is a
non-zero point x. Let X be the unique differentiable vector field satisfying the
relation iX ω = dxm+1 . As this vector field does not vanish at x, we can find a
function x1 in a neighborhood U of x such that X(x1 ) = 1. Consider a vector field Y
on U satisfying the relation iY ω = −dx1 . Since dω = 0, then LX ω = LY ω = 0,
according to the Cartan homotopy formula (theorem 1.9). Therefore,
i[X,Y ] ω = LX iY ω = LX (iY ω) − iY (LX ω) = LX (−dx1 ) = −d(X(x1 )) = 0, from
which we have [X, Y ] = 0, since any point in the form ω is of rank equal to 2m. By
the Recovery theorem1, it follows that there exist local coordinates

x1 , xm+1 , z1 , z2 , ..., z2m−2 on a neighborhood U1 ⊂ U of x such that: X = ∂x
∂
1
,
∂
Y = ∂xm+1 . Consider the differential form λ = ω − dx1 ∧ dxm+1 . We have dλ = 0
and iX λ = LX λ = iY λ = LY λ = 0. So λ is expressed as a 2-differential form
based only on variables z1 , z2 , ..., z2m−2 . In particular, we have λm+1 = 0.
Furthermore, we have 0 = ω m = mdx1 ∧ dxm+1 ∧ λm−1 . The 2-form λ is closed
and of maximal rank (rank half) m − 1 on an open set of R2m−2 . It is, therefore,
sufficient to apply the induction hypothesis to λ.
1.6. Poisson brackets and Hamiltonian systems
As a result of what was previously mentioned, the symplectic form ω induces a

Hamiltonian vector field IdH : M −→ Tx M , x −→ IdH(x), where H : M −→
R is a differentiable function (Hamiltonian). In others words, the differential system
defined by ẋ(t) = XH (x(t)) = IdH(x) is a Hamiltonian vector field associated with
the function H. The Hamiltonian vector fields form a Lie subalgebra of the vector
t
field space. The flow gX leaves invariant the symplectic form ω.
D EFINITION 1.13.– A Hamiltonian system is a triple (M, ω, H) where (M, ω) is a

symplectic manifold and H ∈ C ∞ (M ) a smooth function (Hamiltonian).
T HEOREM 1.11.– The matrix that is associated with a Hamiltonian system determines
a symplectic structure.
P ROOF.– Let (x1 , . . . , xm ) be a local coordinate system on M, (m = dim M ). We

have

m
∂H
m
∂H k
ẋ(t) = I (dxk ) = ξ , [1.6]
∂xk ∂xk
k=1 k=1
where I (dxk ) = ξ k ∈ Tx M is defined such that: ∀η ∈ Tx M ,

k
ηkk = dx k(η) = ω η, ξ (kth-component of η). Define (η1 , . . . , ηm ) and
ξ1 , . . . , ξm to be, respectively, the components of η and ξ k , then
k
⎛ ⎞
ξ1k

m
∂ ∂ ⎜ ⎟
ηk = ηi , ξjk = (η1 , . . . , ηm ) J −1 ⎝ ... ⎠ ,
i=1
∂xi ∂xj k
ξm
1 Let X1 , ..., Xr be differentiable vector fields on a manifold M and x ∈ M . Assume that for
all k, l = 1, ..., r, [xk , Xl ] = 0 and X1 (x), ..., Xr (x) are linearly independent. We show that
there is an open U of M containing x and a local coordinate system on U such that: X1 |U =
∂
∂x1
, ..., Xr |U = ∂x∂ r .
where J −1 is the matrix defined by J −1 ≡ ω ∂ ∂

∂xi , ∂xj . Note that this
1≤i,j≤m
−1
matrix is invertible. Indeed, it suffices to show that the matrix J has maximal rank.
Suppose this were not possible, that is, we assume that rank(J −1 ) = m. Hence
m
i=1 ai ω ∂ ∂
∂xi , ∂xj = 0, ∀1 ≤ j ≤ m, with ai not all null and
m
ω i=1
∂
ai ∂x , ∂
i ∂xj
= 0, ∀1 ≤ j ≤ m. In fact, since ω is non-degenerate, we have
m
i=1
∂
ai ∂x i
= 0. Now ∂ ∂
∂x1 , . . . , ∂xm is a basis of Tx M, then ai = 0, ∀i,
contradiction. Since this matrix is invertible, we can search ξ k such that:
⎛ ⎞ ⎛ ⎞
0 0
⎜ .. ⎟ ⎜ .. ⎟
⎛ k⎞ ⎜ ⎜.
⎟
⎟ ⎛ k⎞ ⎜.⎟
⎜ ⎟
ξ1 ⎜0 ⎟ ξ1 ⎜0⎟
−1 ⎜ .. ⎟
⎜ ⎟ ⎜ .. ⎟ ⎜ ⎟
⎜ th ⎟
J ⎝ . ⎠ = ⎜ 1 k -place ⎟ =⇒ ⎝ . ⎠ = J ⎜
⎜1⎟,
⎟
⎜0 ⎟ ⎜0⎟
ξmk
⎜ ⎟ ξmk
⎜ ⎟
⎜. ⎟ ⎜.⎟
⎝.. ⎠ ⎝ .. ⎠
0 0
m
so ξ k = (kth-column of J), that is, ξik = Jik , 1 ≤ i ≤ m and ξ k = i=1 Jik ∂x∂
i
.
2
The matrix J is skew-symmetric . From [1.6], we deduce that
m ∂H m ∂ m m ∂H ∂
ẋ(t) = k=1 ∂xk i=1 Jik ∂xi = i=1 k=1 Jik ∂xk ∂xi . Writing
m m
ẋ(t) = i=1 dxdti (t) ∂
∂xi , it is seen that ẋi (t) = k=1 Jik ∂xk , where 1 ≤ i ≤ j ≤ m,
∂H
which can be written in more compact form ẋ(t) = J(x) ∂H ∂x , this is the Hamiltonian
vector field associated with the function H.
Let (M, ω) be a symplectic manifold. To any pair of differentiable functions

(F, G) over M , we associate the function {F, G} = du F (XG ) = XG F (u) =
ω(XG , XF ), where XF and XG are the Hamiltonian vector fields associated with the
functions F and G, respectively. We say that {F, G} is a Poisson bracket (or Poisson
structure) of the functions F and G. It is easily verified that the Poisson bracket on
the space C ∞ (M ), that is, the bilinear application {, } : C ∞ (M ) × C ∞ (M ) −→
C ∞ (M ), (F, G) −→ {F, G}, defined above (where C ∞ (M ) is the commutative
algebra of regular functions on M ) is skew-symmetric {F, G} = −{G, F }, obeys
the Leibniz rule {F G, H} = F {G, H} + G{F, H}, and satisfies the Jacobi identity
{{H, F }, G} + {{F, G}, H} + {{G, H}, F } = 0, for all F, G, H ∈ C ∞ (M ).

2 Indeed, since ω ∂
, ∂
∂xi ∂xj
= −ω ∂
∂xj
, ∂
∂xi
, that is, ω is symmetric, it follows that J −1

−1
is skew-symmetric. Then, I = J.J −1 = J .J = −J −1 .J and consequently J = J.
D EFINITION 1.14.– A Poisson structure on a manifold M is a Lie algebra structure

{., .} on the space C ∞ (M ) that satisfies the Leibniz rule. The pair (M, {., .}) is called
a Poisson manifold.
The Leibniz formula ensures that the mapping G −→ {G, F } is a derivation. The
antisymmetry and identity of Jacobi ensure that {, } is a Lie bracket, and they provide
C ∞ (M ) an infinite-dimensional Lie algebra structure. When this Poisson structure is
non-degenerate, we obtain the symplectic structure discussed above. So any
symplectic manifold is a Poisson manifold.
m
E XAMPLE 1.7.– Consider M = R2n with symplectic form ω = k=1 dxk ∧ dyk
n ∂H ∂H
and let H be the Hamiltonian function. We have dH = i=1 ∂xi dxi + ∂y i
dyi ,
and it follows that in the local coordinate system (x1 , . . . , xn , y1 , . . . , yn ), the vector
n
i=1 ∂xi ∂yi − ∂yi ∂xi , and XH F = {H, F },
∂H ∂ ∂H ∂
field XH is given by XH =
∞
∀F ∈ C (M ). In particular, the Poisson brackets of the Darboux coordinate
functions are {xi , xj } = {yi , yj } = 0, {xi , yj } = [δij . The manifold M with the
local coordinates and the canonical Poisson bracket mentioned above is a Poisson
manifold. The Hamiltonian systems form a Lie algebra.
A non-constant function F is called a first integral (or constant of motion) of XF ,

if XH F = 0; F is constant on the trajectories of XH . In particular, H is integral.
Two functions F , G are said to be in involution or to commute if {F, G} = 0. An
interesting result is given by the following Poisson theorem:
T HEOREM 1.12.– If F and G are two first integrals of a Hamiltonian system, then
{F, G} is also a first integral.
P ROOF.– Jacobi’s identity is written as {{H, F } , G} + {{F, G} , H} +

{{G, H} , F } = 0, where H is the Hamiltonian. Since {H, F } = {H, G} = 0, then
we have {{F, G} , H} = 0, which shows that {F, G} is a first integral.
R EMARK 1.4.– If we know two first integrals, we can, according to Poisson’s theorem,
find new integrals, but we often fall back on known first integrals or a constant.
Let M and N be two differentiable manifolds and f ∈ C ∞ (M, N ). The linear

tangent map to f at the point p is the induced mapping between the tangent spaces
Tp M and Tf (p) N , defined by f∗ : Tp M −→ Tf (p) N , f∗ v(ϕ) = v(ϕ ◦ f ), where
v ∈ Tp M and ϕ ∈ C ∞ (N, R). Let L : T M −→ R be a differentiable function
(Lagrangian) on the tangent bundle T M . We say that (M, L) is invariant under the
differentiable application g : M −→ M if for all v ∈ T M , we have L(g∗ v) = L(v).
The Noether theorem below expresses the existence of a first integral associated with
a symmetry of the Lagrangian. In other words, each parameter of a group of
transformations corresponds to a conserved quantity. One of the consequences of the
invariance of the Lagrangian with respect to a group of transformations is the

conservation of the generators of the group. For example, the first integral associated
with rotation invariance is the kinetic moment. Similarly, the first integral associated
with the invariance with respect to the translations is the velocity. The Noether
theorem applies to certain classes of theories, described either by a Lagrangian or a
Hamiltonian. We will give below the theorem in its original version, which applies
to the theories described by a Lagrangian. There is also a version (see Adler and
van Moerbeke 2004) that applies to theories described by a Hamiltonian.
T HEOREM 1.13.– If (M, L) is invariant under a parameter group of diffeomorphisms

gs : M −→ M , s ∈ R, g0 = E, then the system of Lagrange equations,
dt ∂ q − ∂q = 0, corresponding to L admits a first integral I : T M −→ R with
d ∂L ∂L
.

. ∂L dgs (q)
I(q, q) = ∂ q. ds , the q being local coordinates on M .
s=0
P ROOF.– The first integral I is independent of the choice of local coordinates q over
M and so we consider the case M = Rn . Let f : R −→ M , t −→ q = f (t), be
a solution of the system of Lagrange equations above. By hypothesis, g∗s leaves L
invariant, so gs ◦ f : R −→ M , t −→ gs ◦ f (t), also satisfies the system of Lagrange
equations. We translate the solution f (t) considering the application F : R × R −→
Rn , (s, t) −→ q = gs (f (t)). The fact that gs leaves invariant L implies that:
. . .
∂L(F, F ) ∂L ∂F ∂L ∂ F ∂L ∂q ∂L ∂ q
0= = + . =⇒ + . = 0. [1.7]
∂s ∂q ∂s ∂ q ∂s ∂q ∂s ∂ q ∂s
Since F is also a solution of the system of Lagrange equations, that is,

d ∂L . ∂L .
. F (s, t), F (s, t) = F (s, t), F (s, t) ,
dt ∂q ∂q
.
∂q d ∂q
so noting that: ∂s = dt ∂s , and equation [1.7] is written in the form

∂q d ∂L . ∂L d ∂q
0= . F (s, t), F (s, t) + .
∂s dt ∂ q ∂ q dt ∂s

d ∂L . ∂q
= . F (s, t), F (s, t) ,
dt ∂ q ∂s
which completes the proof of the theorem.
We now give the following definition of the Poisson bracket:

{F, G} = ∂F ∂G
∂x , J ∂x =
∂F ∂G
i,j Jij ∂xi ∂xj . We will look for conditions on the matrix
J for Jacobi’s identity to be satisfied. This is the purpose of the following theorem.
T HEOREM 1.14.– The matrix J satisfies the Jacobi identity, if
2n
∂Jli ∂Jjl ∂Jij
Jkj + Jki + Jkl = 0, ∀1 ≤ i, j, l ≤ 2n.
∂xk ∂xk ∂xk
k=1
P ROOF.– Consider the Jacobi identity: {{H, F }, G}+{{F, G}, H}+{{G, H}, F } =
0. We have

∂ {H, F } ∂G ∂{H, F } ∂G
{{H, F }, G} = ,J = Jkl ,
∂x ∂x ∂xk ∂xl
k,l
∂Jij ∂H ∂F ∂G ∂ 2 H ∂F ∂G
= Jkl + Jkl Jij
i,j
∂xk ∂xi ∂xj ∂xl i,j
∂xk ∂xi ∂xj ∂xl
k,l k,l
∂H ∂ 2 F ∂G
+ Jkl Jij .
i,j
∂xi ∂xk ∂xj ∂xl
k,l
By symmetry, we have immediately {{F, G}, H} and {{G, H}, F }. Then
{{H, F }, G} + {{F, G}, H} + {{G, H}, F }

∂Jij ∂H ∂F ∂G
= Jkl
i,j
k,l
∂ 2 H ∂F ∂G
+ Jkl Jij [1.8]
i,j
k,l
∂H ∂ 2 F ∂G
+ Jkl Jij [1.9]
i,j
k,l
∂Jij ∂G ∂H ∂F
+ Jkl
i,j
k,l
∂ 2 G ∂H ∂F
+ Jkl Jij [1.10]
i,j
k,l
∂G ∂ 2 H ∂F
+ Jkl Jij [1.11]
i,j
k,l
∂Jij ∂F ∂G ∂H
+ Jkl
i,j
k,l
∂ 2 F ∂G ∂H
+ Jkl Jij [1.12]
i,j
k,l
∂F ∂ 2 G ∂H
+ Jkl Jij . [1.13]
i,j
k,l
Note that the indices i, j, k and l play a symmetric roll. Applying in the term
[1.11] the permutation i ← l, j ← k, k ← i, l ← j, and add the term [1.8], with
the understanding that Jlk = −Jkl , we get k,l i,j (Jij Jlk + Jkl Jij )
∂G ∂ 2 H ∂F
∂xl ∂xi ∂xk ∂xj = 0, as a consequence of Schwarz’s lemma. Again, applying in the
term [1.12] the permutation i ← k, j ← l, k ← j, l ← i, and adding the term [1.9]
2
yields k,l i,j (Jji Jkl + Jkl Jij ) ∂x∂j ∂x
F ∂G ∂H
k ∂xl ∂xi
= 0. By the same argument as
above, applying in the term [1.13] the permutation i ← l, j ← k, k ← i, l ← j, and
∂F ∂ 2 G ∂H
adding the term [1.10], we obtain k,l i,j (Jij Jlk + Jkl Jij ) ∂x l ∂xi ∂xk ∂xj
= 0,
and thus
{{H, F }, G} + {{F, G}, H} + {{G, H}, F }

∂Jij ∂H ∂F ∂G
= Jkl [1.14]
i,j
k,l
∂Jij ∂G ∂H ∂F
+ Jkl [1.15]
i,j
k,l
∂Jij ∂F ∂G ∂H
+ Jkl .
i,j
k,l
Under permuting the indices i ← l, j ← i, k ← k, l ← j, for [1.14] and i ← j,

j ← l, k ← k, l ← i, for [1.15], we obtain the following:
{{H, F }, G} + {{F, G}, H} + {{G, H}, F }

!
∂Jli ∂Jjl ∂Jij ∂H ∂F ∂G
= Jkj + Jki + Jkl .
∂xk ∂xk ∂xk ∂xl ∂xi ∂xj
i,j,l k
Since the Jacobi identity must be identically zero, the expression to prove follows
immediately, ending the proof of theorem.
C ONCLUSION.– Consequently, we have a complete characterization of the

Hamiltonian vector field
∂H
ẋ(t) = XH (x(t)) = J , x ∈ M, [1.16]
∂x
where H : M −→ R is a smooth function (Hamiltonian) and J = J(x) is a

skew-symmetric matrix, for which the corresponding Poisson bracket satisfies the
Jacobi identity:
{{H,
F }, G} +∂H{{F, G}, H} + {{G, H}, F } = 0, with
{H, F } = ∂H∂x , J ∂F
∂x = i,j J ij
∂F
∂xi ∂xj (Poisson bracket).
1.7. Examples
The following examples will be studied in detail in other chapters.

O I
E XAMPLE 1.8.– An important special case is when J = , where O is the
−I O
n × n zero matrix and I is the n × n identity matrix. The condition on J is trivially
satisfied. Indeed, here the matrix J does not depend on the variable x and we have

∂H
2n 2n n
∂F ∂H ∂F ∂H ∂F
{H, F } = Jij = − .
i=1
∂xi j=1 ∂xj i=1
∂xn+i ∂xi ∂xi ∂xn+i
∂H ∂H
Moreover, equations [1.16] are transformed into q̇1 = ∂p 1
, . . . , q̇n = ∂p n
, ṗ1 =
− ∂q
∂H
1
= − ∂qn , where q1 = x1 , . . . , qn = xn , p1 = xn+1 , . . . , pn = x2n .
, . . . , ṗn ∂H
These are exactly the well-known differential equations of classical mechanics in the
canonical form. They show that it suffices to know the Hamiltonian function H to
determine the equations of motion. They are often interpreted by considering that
the variables pk and qk are the coordinates of a point that moves in a space with 2n
dimensions, called phase space. The flow associated with the system above obviously
leaves invariant each hypersurface of constant energy H = c. The Hamilton equations
mentioned above can still be written in the form q̇i = {H, qi } = ∂p ∂H
i
, ṗi = {H, pi } =
− ∂qi , where 1 ≤ i ≤ n. Note that the functions 1, qi , pi (1 ≤ i ≤ n) verify the
∂H
following commutation relations: {qi , qj } = {pi , pj } = {qi , 1} = {pi , 1} = 0,

{pi , qj } = δij , 1 ≤ i, j ≤ n. These functions constitute a basis of a real Lie algebra
(Heisenberg algebra) of dimension 2n + 1.
E XAMPLE 1.9.– The Hénon–Heiles differential equations are defined by
ẋ1 = y1 , ẏ1 = −Ax1 − 2x1 x2 ,

ẋ2 = y2 , ẏ2 = −Bx2 − x21 − εx22 ,
where x1 , x2 , y1 , y2 are the canonical coordinates and moments and A, B, ε are free
constant parameters. The above equations can be rewritten as a Hamiltonian vector
field, ẋ = J ∂H
∂x , x = (x1 , x2 , y1 , y2 ) , where
1 2 ε
H= (y + y22 + Ax21 + Bx22 ) + x21 x2 + x32 , (Hamiltonian),
2 1 3

0 I
and J = is the matrix associated with the vector field (where O is the 2 × 2
−I 0
zero matrix and I is the 2 × 2 identity matrix).
E XAMPLE 1.10.– The Euler equations (usually referred to as Euler top) of motion of
a rotating rigid body around a fixed point, taken as the origin of the reference bound
to the solid, when no external force is applied to the system, are written as
ṁ1 = (λ3 − λ2 ) m2 m3 , ṁ2 = (λ1 − λ3 ) m1 m3 , ṁ3 = (λ2 − λ1 ) m1 m2 ,
where (m1 , m2 , m3 ) is the angular momentum of the solid and λi ≡ Ii−1 , I1 , I2 , I3

are moments of inertia. These equations can be written in the form of a Hamiltonian
vector field: ẋ = J ∂H
∂x , x = (m1 , m2 , m3 ) , with
1
H= λ1 m21 + λ2 m22 + λ3 m23 , (Hamiltonian).
2
To determine the matrix J = (Jij )1≤i,j≤3 , we proceed as follows: since J is

antisymmetric, then obviously Jii = 0, Jij = −Jji , 1 ≤ i, j ≤ 3, hence
⎛ ⎞
0 J12 J13
J = ⎝ −J12 0 J23 ⎠ .
−J13 −J23 0
Therefore,
⎛ ⎞ ⎛ ⎞⎛ ⎞
ṁ1 0 J12 J13 λ1 m 1
⎝ ṁ2 ⎠ = ⎝ −J12 0 J23 ⎠ ⎝ λ2 m2 ⎠ , [1.17]
ṁ3 −J13 −J23 0 λ3 m 3
⎛ ⎞
(λ3 − λ2 ) m2 m3
= ⎝ (λ1 − λ3 ) m1 m3 ⎠ . [1.18]
(λ2 − λ1 ) m1 m2
Comparing expressions [1.17] and [1.18], we deduce that: J12 = −m3 , J13 = m2
and J23 = −m1 . Finally,
⎛ ⎞
0 −m3 m2
J = ⎝ m3 0 −m1 ⎠ ∈ so(3),
−m2 m1 0
is the matrix of the Hamiltonian vector field. It satisfies the Jacobi identity.
Another random document with
no related content on Scribd:
“Yes,” she said; and I knew that she had satisfied herself. “Will you
take me away now, at once, please?”
The idea of escape, of liberty once realized, it would have been
dangerous to balk her by a moment. I had acquainted mamma that I
might possibly bring her a visitor. Well, it simply meant that the
suggested visit must be indefinitely prolonged.
Miss Gray accompanied me home, where certain surprises, in
addition to the tenderest of ministrations, were awaiting her. All that
becomes private history, and outside my story. I am not a man of
sentiment; and if people choose to write poems and make general
asses of themselves, why—God bless them!
The problem I had set myself to unravel was what looked deucedly
like a tough psychologic poser. But I was resolute to face it, and had
formed my plan. It was no unusual thing for me to be out all night.
That night, after dining, I spent in the “converted” flat in West
Kensington.
I had brought with me—I confess to so much weakness—one of
your portable electric lamps. The moment I was shut in and
established, I pulled out the paper Miss Gray had typed for me,
spread it under the glow and stared at it. Was it a copy of my
circular? Would a sober “First Aid Society” Secretary be likely, do
you think, to require circulars containing such expressions as
“William! William! Come back to me! O, William, in God’s name!
William! William! William!”—in monstrous iteration—the one cry, or
the gist of it, for lines and lines in succession?
I am at the other end from humour in saying this. It is heaven’s
truth. Line after line, half down the page, went that monotonous,
heart-breaking appeal. It was so piercingly moving, my human terror
of its unearthliness was all drowned, absorbed in an overflowing pity.
I am not going to record the experiences of that night. That
unchanging mood of mine upheld me through consciousnesses and
sub-consciousnesses which shall be sacred. Sometimes, submerged
in these, I seemed to hear the clack of the instrument in the window,
but at a vast distance. I may have seen—I may have dreamt—I
accepted it all. Awaking in the chill grey of morning, I felt no surprise
at seeing some loose sheets of paper lying on the floor. “William!
William!” their text ran down, “Come back to me!” It was all that same
wail of a broken heart. I followed Miss Gray’s example. I took out my
match-box, and reverently, reverently burned them.
An hour or two later I was at Paul’s Exchange, privately
interviewing my manager.
“Did you ever employ a Miss Lucy Rivers?”
“Certainly we did. Poor Lucy Rivers! She rented a machine of us.
In fact——”
He paused.
“Well?”
“Well—it is a mere matter of business—she ‘flitted,’ and we had to
reclaim our instrument. As it happens, it was the very one purchased
by the young lady who so interested you here two days ago.”
“The first machine, you mean?”
“The first—and the second.” He smiled. “As a matter of fact, she
took away again what she brought.”
“Miss Rivers’s?”
He nodded.
“There was absolutely nothing wrong with it—mere fad. Women
start these fancies. The click of the thing gets on their nerves, I
suppose. We must protect ourselves, you see; and I’ll warrant she
finds it perfection now.”
“Perhaps she does. What was Miss Rivers’s address?”
He gave me, with a positive grin this time, the “converted” flat.
“But that was only latterly,” he said. “She had moved from——”
He directed me elsewhere.
“Why,” said I, taking up my hat, “did you call her ‘poor Lucy
Rivers’?”
“O, I don’t know!” he said. “She was rather an attractive young
lady. But we had to discontinue our patronage. She developed the
most extraordinary—but it’s no business of mine. She was one of the
submerged tenth; and she’s gone under for good, I suppose.”
I made my way to the other address—a little lodging in a shabby-
genteel street. A bitter-faced landlady, one of the “preordained” sort,
greeted me with resignation when she thought I came for rooms, and
with acerbity when she heard that my sole mission was to inquire
about a Miss Lucy Rivers.
“I won’t deceive you, sir,” she said. “When it come to receiving
gentlemen privately, I told her she must go.”
“Gentlemen!”
“I won’t do Miss Rivers an injustice,” she said. “It was ha
gentleman.”
“Was that latterly?”
“It was not latterly, sir. But it was the effects of its not being latterly
which made her take to things.”
“What things?”
“Well, sir, she grew strange company, and took to the roof.”
“What on earth do you mean?”
“Just precisely what I say, sir; through the trap-door by the steps,
and up among the chimney-pots. He’d been there with her before,
and perhaps she thought she’d find him hiding among the stacks. He
called himself an astronomer; but it’s my belief it was another sort of
star-gazing. I couldn’t stand it at last, and I had to give her notice.”
It was falling near a gloomy midday when I again entered the flat,
and shut myself in with its ghosts and echoes. I had a set conviction,
a set purpose in my mind. There was that which seemed to scuttle,
like a little demon of laughter, in my wake, now urging me on, now
slipping round and above to trip me as I mounted. I went steadily on
and up, past the sitting-room door, to the floor above. And here, for
the first time, a thrill in my blood seemed to shock and hold me for a
moment. Before my eyes, rising to a skylight, now dark and choked
with snow, went a flight of steps. Pulling myself together, I mounted
these, and with a huge effort (the bolt was not shot) shouldered the
trap open. There were a fall and rustle without; daylight entered; and,
levering the door over, I emerged upon the roof.
Snow, grim and grimy and knee-deep, was over everything,
muffling the contours of the chimneys, the parapets, the irregularities
of the leads. The dull thunder of the streets came up to me; a fog of
thaw was in the air; a thin drizzle was already falling. I drove my foot
forward into a mound, and hitched it on something. In an instant I
was down on my knees, scattering the sodden raff right and left, and
—my God!—a face!
She lay there as she had been overwhelmed, and frozen, and
preserved these two months. She had closed the trap behind her,
and nobody had known. Pure as wax—pitiful as hunger—dead! Poor
Lucy Rivers!
Who was she, and who the man? We could never learn. She had
woven his name, his desertion, her own ruin and despair into the
texture of her broken life. Only on the great day of retribution shall he
answer to that agonized cry.
THE FAIR WITH GOLDEN HAIR
Ho! bring me some lovers, fat or lean,
That I may crunch ’em my teeth between!
I could eat so many, so many, so many,
That in the wide world there would not be left any.
Ho! Here is Avenant to be seen,

Who comes to draw your teeth so keen;
He’s not the greatest man to view,
But he’s big enough to conquer you.
Planché’s “D’Aulnoy,” slightly misquoted.
Sir Richard Avenant came home from Abyssinia to an interesting

notoriety. He had been associated—a sort of explorative free-lance
—with the expedition of Mr. Bruce, who was not yet returned from his
adventures up the Nile in quest of the sources of that bewildering
water; and, upon his arrival in London, he found himself engaged to
a romance which was certainly remote from his deserts.
Now he was a strong, saturnine man, but apt to whimsical
decisions, whose consequences, the fruits of whatever odd
impulses, he never had a thought but to hold by; and as the self-
reserved must suffer the character accorded to their appearance (the
only side of them confessed), Sir Richard found himself accredited,
by anticipation, with deeds adapted to the countenance he had
always addressed to the world.
He was strolling, some days after his return, through the streets,
when he was accosted by an acquaintance, a preux chevalier of the
highest ton, curled, be-ruffed, and imperturbably self-assured.
“Why, strike me silly, Dick!” cried this exquisite, “what do you,
wandering unsociable in a shag coat, and all London by the ears to
lionize ye?”
“Well, I know not, George. What have I done to be lionized?”
“Done! Done? asks the man that will not devour a steak but ’tis cut
raw from the buttock of the living beast! Done? asks Bluebeard (and
stap me, Dick, but your chin is as blue as a watchman’s!)—done, he
says, that brings grass-petticoats in his train enough to furnish the
Paradise of the Grand Turk! Prithee, Dick, where hast stowed ’em
all? O, thou hast a great famous reputation, I assure thee, to justify
thyself of with the women! Such is the report of thy peris—their teeth,
their raven hair, their eyes like stars of the night—there’s no virtue in
town could resist, if asked, to be thy queen and theirs.”
He was chuckling, and taking a delicate pinch of martinique, with
his little finger cocked to display a glittering stone, when his eyes
lighted on a house over against which they were standing.
“Hist!” said he, pointing with his cane; “pan my honour, the single
reservation.”
“Single reservation?” repeated the explorer. “To what? To this
London of frailties?”
“To be sure,” said the other. “The one party, I’ll dare swear, that
would not put her nose in a ring for thy sake.”
“Indeed!” said Avenant. “Then she’s the one I must wed.”
The elegant cocked his head, squinting derisive.
“I lay you a double pony to a tester you don’t, within the decade.”
“Done! Tell me about her.”
“I’ll do more. I’ll carry you in to her, here and at once. Tell me about
her, quotha! She’s the Fair with Golden Hair, and a guinea and a
suitor to every thread of it.”
“Whence comes she?”
“From Arcadia, man, with a fortune of gold and roses. She cuts out
hearts raw, as you do steaks, and devours them by the dozen. O,
you shall know her!”
“But by what name, George, by what name?”
“Have I not told you? It shall suffice for all your needs. Thou shalt
take a pack of Cabriolles, and never hunt her to the death. Come,
my friend!”
He led Sir Richard to the house, and had himself announced. They
ascended a flight of stairs, going up into a heaven of floating
fragrance and melodious sounds. Their feet moved noiseless over
silken carpets. They crossed an anteroom ruffling with lackeys, and
were ushered into the Fair’s boudoir.
She sat at her mirror, in the hands of her perruquier. She was the
most beautiful insolent creature Sir Richard had ever seen. There
was not an inch of her which Nature could have altered to its
improvement. The very patch on her cheek was a theft from
perfection. But to so much loveliness her hair was the glory, a
nimbus which, condensing in the heavy atmosphere of adoration,
dropped in a melting flood of gold, which, short of the ground only,
shrank and curled back from its gross contact.
All round and about her hummed her court—poets, lords, minstrels
—suitors straining their wits and their talents for her delectation,
while they bled internally. Many of them greeted Sir Richard’s
chaperon, many Sir Richard himself—good-humouredly, jealously,
satirically, as the case might be—as the two pushed by. A stir went
round, however, when the rough new-comer’s name was put about;
and some rose in their seats, and all dwelt inquisitively on the
explorer’s reception.
It was condescending enough; as was that of his friend, who loved
himself too well, and too wittily, to show a heart worth the beauty’s
discussing.
“Have you got back your appetite, sir,” said the Fair to Avenant,
“for dressed meats?”
“And ladies?” whispered Sir Richard’s friend.
“O, fie!” said madam.
“I will return the question on you,” said Sir Richard, in a low voice.
The Fair lifted her brows.
“Why, I am told, madam,” said Avenant, “that you feed on raw
hearts; but I am willing to believe that the one lie is as certain as the
other.”
The imperious beauty bit her underlip, and laughed.
“I perceive, Sir Richard,” she said, “that you do not court by
flattery.”
“I do not court at all, madam,” he answered.
“Ah, true!” she replied. “You buy in the open market. It must be
simpler; though, in the plain lodging where I hear you lie at present,
the disposal of so responsible an establishment must exercise your
diplomacy.”
She spoke aloud, evoking a general titter; and so aloud Avenant
answered her.
“By no means, madam. I have in my sleeping-room a closet with
three shelves. On one of these lies Beauty, unspoiled by adulation;
on another lies Virtue, that respects her sex too well to traduce it; on
the third lies feminine Truth, loveliest of her sisters. These are my
whole establishment; and as they are shadows all, existing only in
the imagination, they exercise nothing but my fondness for
unattainable ideals.”
The company broke into much laughter over this Jeremiad; and
the girl joined her young voice to theirs. But a little glow of colour
was showing in her cheek, verily as if Sir Richard had flicked that fair
surface with his glove.
“O!” she said, “this is a sad regale! Sure, sir, does the climate of
Abyssinia breed no hotter than Leicestershire Quakers? Why, I have
heard a lion roar fiercer in a caravan. Now, pray, Sir Richard, put off
your civilities, and give us news instead of lessons. They say there is
a form of lawless possession in the women of the country you
visited.”
“It is very true there is, madam. It is called the Tigrétier—a seizure
of uncontrollable vanity, during which the victim is so self-centred
that she is unable to attend to the interests, or even to distinguish the
sexes of those about her. She will, for instance, surround herself with
a circle of male admirers, assuming all the time, apparently, that they
are the gossips of her own sex, with whom, like a decent woman,
she would wont ordinarily, of course, to consort in private.”
The Fair cried out, “Enough! Your stories are the most intolerable
stuff, sir. I wish Mr. Bruce joy of your return, as I hear you are not to
remain in England.”
Then she turned her shoulder to him, her flush deepening to fire;
and Sir Richard, bowing and moving away, fell into conversation with
one or two of his acquaintances. Presently, looking up, he was
surprised to see the room near empty. Goldenlocks had, in fact,
issued her wilful mandate, and her court was dismissing itself.
The explorer was pressing out after the rest, when a maidservant
touched his sleeve, and begged him to return to her lady, who
desired a word with him. Sir Richard acquiesced immediately. He
found the Fair standing solitary by her dressing-table, frowning, her
head bent, her fingers plucking at a wisp of lace. Her hair, still
undressed, hung down deep over her shoulders, mantling them with
heavy gold, like a priest’s chasuble.
“Did you seek my acquaintance, sir,” she said imperatively, “with
the sole purpose to insult me?”
“Nay, madam,” he answered, as cool as tempered steel; “but
because you was described to me as the one woman in London that
I might not marry, if I had the will to.”
“Why not?” was on the tip of her tongue; he saw it there. But she
caught at herself, and answered, “So, sir, like sour reynard, I
suppose, you would spite what you found it useless to covet.”
“I covet, madam!” he said, in a tone of astonishment. “I aspire to
wrest this wealth and beauty from a hundred worthier candidates!
Believe me, my ambition halted far short of such attainment.”
Her lips smiled, despite herself. What were the value, she
suddenly thought, in a world of suitors that did not include this
shagg’d and rugged Jeremiah? Her speech fell as caressing as the
sound of water in a wood.
“Yet you confess to some ambition?” she murmured.
“True,” he answered; “the virtuoso’s.”
She lifted her beautiful brows.
“I will be candid, madam,” he said. “I have the collector’s itch.
Whithersoever I visit, I lay my toll on the most characteristic
productions of the tribes—robes, carvings, implements of war—even
scalps. Madam, madam, you must surely be of the sun children!
Your hair is the most lovely thing! I would give my soul—more, I
would give a thousand pounds to possess it.”
“I see, sir,” she said; “to carry your conquest at your belt.”
“Nay,” he answered, with feigned eagerness. “Not a soul need
know. The thing is done constantly. You have but to subscribe to the
fashion of powder, and you gain a novel beauty, and I a secret I
swear to hold inviolate.”
“O!” she said softly “This is Samson come with the shears to turn
the tables on poor Delilah!”
And on the instant she flashed out, breaking upon him in a storm
of passion. That he dared, that he dared, on no warrant but his
reputation for inhumanity, so to outrage and insult her.
“Go, sir!” she cried. “Return to your Nubians and Dacoits—to
countries where head-hunting is considered an honourable proof of
manliness!”
He stood, as outwardly insensate as a bull.
“Then you decline to deal?”
Her only answer was to throw herself into a chair, and to abandon
herself to incomprehensible weeping. But even her sobs seemed to
make no soft impression on him. He took a step nearer to her, and
spoke in the same civil and measured tone he had maintained
throughout.
“Take care, madam. I never yet set my will upon a capture that in
the long run escaped me.”
She checked her tears, to look up at him with a little furious laugh.
“Poor boaster!” she said. “I think, perhaps, that recounting of your
Tigrétier hath infected you with it.”
“By my beard, madam,” said he, “I will make that hair my own!”
“See,” she cried jeeringly, “how a boaster swears by what he has
not!”
Sir Richard felt to his chin.
“That is soon remedied,” said he. “And so, till my oath is
redeemed, to consign my razors to rust!” And with these words,
bowing profoundly, he turned and left the room.
Shortly after this he sailed to rejoin his expedition, and was not
again in England during a period of eighteen months.
At the end of that time, being once more in London, he devoted
himself—his affairs having now been ordered with the view to his
permanent residence in the country—to some guarded inquiries
about the Fair with Golden Hair. For some days, the season of the
town being inauspicious, he was unable to discover anything definite
about her. And then, suddenly, the news which he sought and
desired came in a clap.
He was walking, one day, down a street of poor and genteel
houses, when he saw her before him. He stood transfixed. There
was no doubting his own eyesight. It was she: tall, slender, crowned
with her accustomed glory, the flower of her beauty a little wan, as if
seen by moonlight. But what confounded him was her condition. Her
dress was mean, her gloves mended; every tag of cheap ribbon
which hung upon her seemed the label to a separate tragedy. Thus
he saw her again, the Fair with Golden Hair; but how deposed and
fallen from her insolent estate!
She mounted a step to a shabby door. While she stood there,
waiting to be admitted, an old jaunty cavalier came ruffling it down
the street, accosted her, and accompanied her within. She might
have glanced at Avenant without recognizing him. The rough dark
beard he wore was his sufficient disguise.
Sir Richard made up his mind on the spot, and acted promptly.
Having no intention to procure himself a notoriety in this business, he
rigidly eschewed personal inquiry, and employed an official informer,
at a safe figure, to ferret out the truth for him. This, epitomized,
discovered itself as follows:—
Cytherea—Venus Calva—Madonna of the magic girdle, who had
once reigned supreme between wealth and loveliness, who had
once eaten hearts raw for breakfast, feeding her roses as vampires
do, was desolate and impoverished—and even, perhaps, hungry. A
scoundrelly guardian had eloped with trust funds: the crash had
followed at a blow. Robbed of her recommendation to respect;
deposed, at once, from the world’s idolatry to its vicious solicitation,
she had fled, with her hair and her poor derided virtue, into squalid
oblivion; that, at least, she hoped. But, alas for the fateful recoils on
Vanity! She drives with a tight rein; and woe to her if the rein snap! A
certain libidinous and crafty nobleman, of threescore or so years,
had secured, in the days of the Fair’s prosperity, some little bills of
paper bearing that beauty’s signature. These he had politicly
withheld himself from negotiating, on the mere chance that they
might serve him some day for a means to humiliate one who, in the
arrogance of her power, had scoffed at his amatory, and perfectly
honourable, addresses. That precaution had justified itself. The peer
was now come to woo again, and less scrupulously, with his hand on
a paper weapon, one stroke from which alone was needed to give
the Fair’s poor drabbled fortune its quietus. She was at bay, between
ruin and dishonour.
Sir Richard came immediately to a resolve, and lost no time in
giving it effect. He wrote a formal note to the Fair, recalling himself
courteously to her remembrance, reminding her of his original offer,
and renewing it in so many words. He would do himself the honour,
he said, to wait upon her for her answer on such and such a day.
To this he received no reply; nor, perhaps, expected one. He went,
nevertheless, to his self-made appointment with the imperturbable
confidence of a strong man.
Passing, on his way, by a perruquier’s, he checked himself, and
stood for some moments at gaze in a motionless reverie. Then he
entered the shop, made a purchase, and, going to a barber’s,
caused himself to be shorn, shaved, and restored to the
conventional aspect. Thus conditioned, he knocked at the Fair’s
door, and was ushered up—bawled up, rather, by a slattern landlady
—into her presence.
She rose to face him as he entered. She had his letter in her hand.
Her beautiful hair, jealous, it seemed, to withdraw itself from the
curioso’s very appraisement, was gathered into and concealed under
a cap. Her features, thus robbed of their dazzling frame, looked
curiously, sadly childish and forlorn. There were dark marks round
her eyes—the scarce dissipated clouds of recent tears. Who can tell
what emotions, at sight of this piteous, hard-driven loveliness, stirred
the heart of the man opposite, and were repressed by his iron will?
“This letter, sir,” said the Fair, holding out the paper in a hand
which shook a little. “I have tacitly permitted you to presume a right
to a personal answer to that which it proposes, because such a
course appeared to me the least compromising. I cannot write my
name, sir, nowdays—as scandal doubtless hath informed you—but
Fortune will be using it to my discredit.”
Sir Richard bowed.
“There is this difference only, madam: my word is the bond of a
gentleman. I vowed you secrecy.”
“That is to assume, on your part,” she said quietly, “a
confidentialness which, in its insult to misfortune, is at least not the
act of a gentleman. Moreover, a gentleman, surely, had not taken
advantage of circumstances to propose to destitution what affluence
had once refused him.”
“Beware, madam!” said Avenant. “Pride must make some
sacrifices to virtue. If, in renewing a pure business offer, I, a simple
instrument in the hands of Providence, give you an opportunity to
maintain that priceless possession unimpaired, would it not be the
truer self-respect to secure your honour at whatever cost to your
sentiments?”
“I thank you, sir,” she said. “I have not forgotten, nor forgotten to
resent, my self-constituted Mentor. I will assure him that, for the
matter of my virtue, it is safe in my hands, though I have to arm
those against myself.”
“Good heavens, madam!” cried Avenant. “You are not at that
resource?”
“Give yourself no concern, sir,” she answered coldly. “The moral I
learned of your insult, was to save myself in its despite.”
His deep eyes glowed upon her.
“You have sold your hair?” he said.
“Yes,” she answered; “to pay my debts. ’Twas your letter decided
me.”
“At a thousand pounds?”
“At a hundred.”
Then she added, as if irresistibly, because she was still little more
than a child, “And now, sir, how is the boaster vindicated? But your
oath, I perceive, still goes beardless.”
“Within the hour only,” said he; and, thrusting his hand into his
breast, he drew out the long tresses of the Fair With Golden Hair.
She stared, amazed a moment; then threw herself upon her knees
by a chair, weeping and crying out—
“O, I hate you, I hate you!”
He strode, and stood over her.
“I saw them through a window as I came. How could I mistake
them? There is not their like in the world. But now, my oath
redeemed, it is for you to say if I am to destroy them.”
“O, my hair!” she wept; “my one beauty!”
“I have staked all on this,” he cried. “If your hair was your one
beauty, my beard alone redeemed me from appalling ugliness by so
much as it hid of me. Well, I have lost on both counts, if the net result
is your hatred.”
She looked up, with drowned bewildered eyes, and held out her
hand blindly.
“Give me back my hair,” she said, “and you shall have the hundred
pounds.”
“Nay, sweet Delilah,” quoth he; “for that would be to return you
your strength, and I want you weak.”
Her arm dropped to her side.
“That you may insult me with impunity!” she said bitterly.
“Ah, Delilah!” he cried; “is it so bad, that the offer of my hand and
heart is an insult to a woman?”
She sank back, sitting on her heels. From under her cap, fallen
awry, curled shavings of gold hung out—the residue of a squandered
wealth. Her eyes were wide with amazement.
“So bad?” she whispered. “Are you asking me to marry you?”
He was not a conformable wooer. The love-wise sex shall say if he
was a diplomatic one. He threw himself on his knees beside the Fair,
seized her in his bear-like grip, and kissed her lips.
“Now,” he said, “it is neck or nothing. None but a parson can wipe
out the stain. Hate me now, and put Love to bed for by and by.”
She smiled suddenly—like the rainbow; like an angel.
“Yes,” she said, “if you insist. But the poor thing has slept so long
in my heart, that it would fain wake up at last, and confess itself.”
The peer took his settlement with a very bad grace; but he had to
take it, and there was an end of him.
“Avenant,” whispered the Fair, on the evening of their wedding
day, “I have been vain, spoiled, perhaps untruthful. But I wished to
tell you—you can put me to sleep on the middle shelf of your
cupboard.”
“It has been converted into a closet for skeletons,” he said. “I was
a bachelor then.”
THE LOST NOTES
The faculty of music is generally, I believe, inimical to the
development of all the other faculties. Sufficient to itself is the
composing gift. There was scarcely ever yet a born musician, I do
declare, who, outside his birthright, was not a born ass. I say it with
the less irreverence, because my uncle was patently one of the rare
exceptions which prove the rule. He knew his Shakespeare as well
as his musical-glasses—better than, in fact; for he was a staunch
Baconian. This was all the odder because—as was both early and
late impressed upon me—he had a strong sense of humour.
Perhaps an eternal study of the hieroglyphics of the leger lines was
responsible for his craze; for craze I still insist it was, in spite of the
way he took to convince me of the value of cryptograms. I was an
obstinate pupil, I confess, and withstood to the end the fire of all the
big guns which he—together with my friend, Chaunt, who was in the
same line—brought to bear upon me.
Well, I was honest, at least; for I was my uncle’s sole provisional
legatee, and heir presumptive to whatever small fortune he had
amassed during his career. And day by day, as the breach between
us widened, I saw my prospect of the succession attenuating, and
would not budge from my position. No, Shakespeare was
Shakespeare, I said, and Bacon, Bacon; and not all the cyphers in
the world should convince me that any profit was to be gained by
either imagining or unravelling a single one of them.
“What, no profit!” roared my uncle. “But I will persuade you, young
man, of your mistake before I’m done with you. Hum-ti-diddledidee!
No profit, hey? H’m—well!”
Then I saw that the end was come. And, indeed, it was an open
quarrel between us, and I was forbidden to call upon him again.
I was sorry for this, because, in his more frolicsome and
uncontroversial moments, he was a genial companion, unless or
until one inadvertently touched on the theme, when at once he
exploded. Professionally, he could be quite a rollicking blade, and his
settings of plantation songs were owned to be nothing less than lyric
inspirations. Pantomime, too, in the light of his incidental music, had
acquired something more than a classical complexion; and, in the
domain of knockabout extravaganza, not only did the score of “The
Girl who Knew a Thing or Two” owe to him its most refined numbers,
but also the libretto, it was whispered, its best Attic bonnes-bouches.
However, all that good company I must now forgo—though Chaunt
tried vainly to heal the breach between us—and in the end the old
man died, without any visible relenting towards me.
I felt his loss pretty keenly, though it is no callousness in me to
admit that our long separation had somewhat dulled the edge of my
attachment. I expected, of course, no testamentary consideration
from him, and was only more surprised than uplifted to receive one
morning a request from his lawyers to visit them at my convenience.
So I went, soberly enough, and introduced myself.
“No,” said the partner to whom I was admitted, in answer to a
question of mine: “I am not in a position to inform you who is the
principal beneficiary under our friend’s will. I can only tell you—what
a few days before his death he confided to us, and what, I think,
under the circumstances, you are entitled to learn—that he had quite
recently, feeling his end approaching, realized on the bulk of his
capital, converted the net result into a certain number—five, I think
he mentioned—of Bank of England notes, and… burned ’em, for all
we know to the contrary.”
“Burned them!” I murmured aghast.
“I don’t say so,” corrected the lawyer drily. “I only say, you know,
that we are not instructed to the contrary. Your uncle” (he coughed
slightly) “had his eccentricities. Perhaps he swallowed ’em; perhaps
gave ’em away at the gate. Our dealings are, beyond yourself, solely
with the residuary legatee, who is, or was, his housekeeper. For her
benefit, moreover, the furniture and effects of our late client are to be
sold, always excepting a few more personal articles, which, together
with a sealed enclosure, we are desired to hand over to you.”
He signified, indeed, my bequest as he spoke. It lay on a table
behind him: A bound volume of minutes of the Baconian Society; a
volume of Ignatius Donnelly’s Great Cryptogram; a Chippendale tea-
caddy (which, I was softened to think, the old man had often known
me to admire); a large piece of foolscap paper twisted into a cone,
and a penny with which to furnish myself with a mourning ring out of
a cracker.
I blushed to my ears, regarding the show; and then, to convince
this person of my good-humoured sanity, giggled like an idiot. He did
not even smile in reply, the self-important ass, but, with a manner of
starchy condescension, as to a wastrel who was getting all his
deserts, rose from his chair, unlocked a safe, took an ordinary sealed
envelope from it, handed it to me, and informed me that, upon giving
him a receipt, I was at liberty to remove the lot.
“Thanks,” I said, grinding my astral teeth. “Am I to open this in your
presence?”
“Quite inessential,” he answered; and, upon ascertaining that I
should like a cab called, sent for one.
“Good morning,” he said, when at last it was announced (he had
not spoken a word in the interval): “I wish you good morning,” in the
morally patronizing tone of a governor discharging a prisoner.
I responded coldly; tried, for no reason at all, to look threatening;
failed utterly, and went out giggling again. Quite savagely I threw my
goods upon the seat, snapped out my address, closed the apron
upon my abasement, and sat slunk into the cavity behind, like a
salted and malignant snail.
Presently I thumped the books malevolently. The dear old man
was grotesque beyond reason. Really he needn’t have left life cutting
a somersault, as it were.
But, as I cooled outwardly, a warmer thought would intrude. It
drew, somehow, from the heart of that little enclosure lying at the
moment in my pocket. It was ridiculous, of course, to expect anything
of it but some further development of a rather unkind jest. My uncle’s
professional connexion with burlesque had rather warped, it would
appear, his sense of humour. Still, I could not but recall that story of
the conversion of his capital into notes: and an envelope——!
Bah! (I wriggled savagely). It was idiotic beyond measure so to
flatter myself. Our recent relations had precluded for ever any such
possibility. The holocaust, rather! The gift to a chance passer-by, as
suggested by that fool of a lawyer! I stared out of the window,
humming viciously, and telling myself it was only what I ought to
expect; that such a vagary was distinctly in accordance with the
traditions of low comedy. It will be observed that I was very
contemptuous of buffoonery as a profession. Paradoxically, a joke is
never played so low as when it is played on our lofty selves.
Nevertheless, I was justified, it appeared. It may be asked, Why
did I not at once settle the matter by opening the envelope in the
cab? Well, I just temporized with my gluttony, till, like the greedy boy,
I could examine my box in private—only to find that the rats had
devoured all my cake. It was not till I was shut into my sitting-room
that I dared at length to break the seal, and to withdraw——
Even as it came out, with no suggestion of a reassuring crackle, I
realized my fate. And this was it: please to examine it carefully—
Now, what do you make of it? “Ex nihilo nihil fit,” I think you will say
with me. It was literally thus, carefully penned in the middle of a
single sheet of music-paper—a phrase, or motif, I suppose it would
be called—an undeveloped memorandum, in fact—nothing else
whatever. I let the thing drop from my hand.
No doubt there was some capping jest here, some sneer, some
vindictive sarcasm. I was not musician enough to tell, even had I had
spirit for the endeavour. It was unworthy, at least, of the old man—
much more, or less, than I deserved. I had been his favourite once.
Strange how the idée fixe could corrode an otherwise tractable
reason. In justice to myself I must insist that quite half my
disappointment was in the realization that such dislike, due to such a
trifle, could have come to usurp the old affection.
By and by I rose dismally, and carried the jest to the piano. (Half a
crown a day my landlady exacted from me, if I so much as thumped
on the old wreck with one finger, which was the extent of my talent.)
Well, I was reckless, and the theme appeared ridiculously simple.
But I could make nothing of it—not though Mrs. Dexter came up in
the midst, and congratulated me on my performance.
When she was gone I took the thing to my chair again, and
resumed its study despondently. And presently Chaunt came in.
“Hullo!” he said: “how’s the blooming legatee?”
“Pretty blooming, thanks,” I said. “Would you like to speculate in
my reversion? Half a crown down to Mrs. Dexter, and the use of the
tin kettle for the day.”
“Done,” he said, “so far as the piano’s concerned. Let’s see what
you’ve got there.”
He had known of my prospective visit to the lawyers, and had
dropped in to congratulate me on that performance. I acquainted him
with the result; showed him the books, and the tea-caddy, and the
penny, and the remnants of foolscap—finally, handed him the
crowning jest for inspection.
“Pretty thin joke, isn’t it?” I growled dolefully. “Curse the money,
anyhow! But I didn’t think it of the old man. I suppose you can make
no more of that than I can?”
He was squinting at the paper as he held it up, and rubbing his
jaw, stuck out at an angle, grittily.
“H’m!” he said, quite suddenly, “I’d go out for a walk and revive
myself, if I were you. I intend to hold you to that piano, for my part;
and you wouldn’t be edified.”
“No,” I said: “I’ve had enough of music for a lifetime or so! I fancy
I’ll go, if you won’t think me rude.”
“On the contrary,” he murmured, in an absorbed way; and I left
him.
I took a longish spin, and returned, on the whole refreshed, in a
couple of hours. He was still there; but he had finished, it appeared,
with the piano.
“Well,” he said, rising and yawning, “you’ve been a deuce of a time
gone; but here you are”—and he held out to me indifferently a little
crackling bundle.
Without a word I took it from his hand—parted, stretched, and
explored it.
“Good God!” I gasped: “five notes of a thousand apiece!”
He was rolling a cigarette.
“Yes,” he drawled, “that’s the figure, I believe.”
“For me?”
“For you—from your uncle.”
“But—how?”
He lighted, took a serene puff or two, drew the jest from his
pocket, and, throwing it on a chair, “You’ll have to allow some value
to cryptograms at last,” he said, and sat down to enjoy himself.
“Chaunt!”
“O,” he said, “it was a bagatelle. An ass might have brayed it out at
sight.”
“Please, I am something less than an ass. Please will you interpret
for me?” I said humbly.
He neighed out—I beg his pardon—a great laugh at last.
“O,” he cried: “your uncle was true blue; he stuck to his guns; but I
never really supposed he meant to disinherit you, Johnny. You
always had the first place in his heart, for all your obstinacy. He took
his own way to convince you, that was all. Pretty poor stuff it is, I’m
bound to confess; but enough to run your capacities to extinction.
Here, hand it over.”
“Don’t be hard on me,” I protested, giving him the paper. “If I’m all
that you say, it was as good as cutting me off with a penny.”
“No,” he answered: “because he knew very well that you’d apply to
me to help you out of the difficulty.”
“Well, help me,” I said, “and, in the matter of Bacon, I’ll promise to
be a fool convinced against my will.”
“No doubt,” he answered drily, and came and sat beside me. “Look
here,” he said; and I looked:—
“You know your notes, anyhow,” said he. “Well, you’ve only got to
read off these into their alphabetical equivalents, and cut the result
into perfectly obvious lengths. It’s child’s play so far; and, indeed, in
everything, unless this rum-looking metronome beat, or whatever it
may be, bothers you for a moment.”

Integrable Systems 1St Edition Ahmed Lesfari Full Chapter PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Integrable Systems 1St Edition Ahmed Lesfari Full Chapter PDF

Uploaded by

Copyright:

Available Formats

Integrable Systems 1st Edition Ahmed

Integrable Systems 1st Edition Ahmed Lesfari

Fractional-order Design: Devices, Circuits, And Systems

Viroids and Satellites 1st Edition Edition Ahmed Hadidi

Back-Of-The-Envelope Quantum Mechanics: with Extensions

Religion, Extremism and Violence in South Asia 1st

Reservoir Engineering Handbook Fifth Edition 2019 Tarek

Endodontic Advances and Evidence-Based Clinical

Bionanotechnology: Emerging Applications of

ISTE Ltd John Wiley & Sons, Inc.

© ISTE Ltd 2022

Library of Congress Control Number: 2022932445

British Library Cataloguing-in-Publication Data

Chapter 1. Symplectic Manifolds . . . . . . . . . . . . . . . . . . . . . . . 1

Chapter 2. Hamilton–Jacobi Theory . . . . . . . . . . . . . . . . . . . . . 49

Chapter 3. Integrable Systems . . . . . . . . . . . . . . . . . . . . . . . . . 67

Chapter 4. Spectral Methods for Solving Integrable Systems . . . . 103

Chapter 5. The Spectrum of Jacobi Matrices and Algebraic

Chapter 6. Grifﬁths Linearization Flows on Jacobians . . . . . . . . . 143

6.5. The Toda lattice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150

Chapter 7. Algebraically Integrable Systems . . . . . . . . . . . . . . . 159

Chapter 8. Generalized Algebraic Completely Integrable

Chapter 9. The Korteweg–de Vries Equation . . . . . . . . . . . . . . . 241

Chapter 10. KP–KdV Hierarchy and Pseudo-differential

10.3. KP hierarchy and vertex operators . . . . . . . . . . . . . . . . . . . . 281

This book is intended for a wide readership of mathematicians and physicists:

Chapter 1 is devoted to the study of symplectic manifolds and their connection

In Chapter 3, we study the Arnold–Liouville theorem: the regular compact level

In Chapter 4, we give a detailed study of the integrable systems that can be

The aim of Chapter 5 is to describe some connections between spectral theory

In Chapter 6, we present in detail the Grifﬁths’ approach and his cohomological

In Chapter 7, the notion of algebraically completely integrable Hamiltonian

In Chapter 8, we discuss the study of generalized algebraic completely integrable

algebraic parameters. We discuss some interesting examples: Ramani–Dorizzi–

Chapter 9 covers the stationary Schrödinger equation, the integral

In Chapter 10, we study some generalities on the algebra of inﬁnite order

This chapter is organized as follows. In the second section, we begin by brieﬂy

1.2. Symplectic vector spaces

The dimension of a symplectic vector space is always even. We show (using a

where Im denotes the m × m unit matrix.

Let (E, ω) be a symplectic vector space and F a vector subspace of E. Let F ⊥ =

D EFINITION 1.2.– The subspace F is isotropic if F ⊂ F ⊥ , coisotropic if F ⊥ ⊂ F ,

If F , F1 and F2 are subspaces of a symplectic space (E, ω), then

dim F + dim F ⊥ = dim E, (F ⊥ )⊥ = F,

F1 ⊂ F2 =⇒ F2⊥ ⊂ F1⊥ , (F1 ∩ F2 )⊥ = F1⊥ + F2⊥ , F1⊥ ∩ F2⊥ = (F1 + F2 )⊥ ,

F is coisotropic if and only if F ⊥ is isotropic and F is Lagrangian if and only if

1.3. Symplectic manifolds

D EFINITION 1.3.– Let M be an even-dimensional differentiable manifold. A

At a point p ∈ M , we have a non-degenerate antisymmetric bilinear form on the

T HEOREM 1.1.– Let M be a differentiable manifold of dimension m and T ∗ M its

T ∗ M , of tangent bundles, respectively, to M and T ∗ M on their bases. We note π ∗ :

a tangent vector to M . Let (x1 , ..., xm , y1 , ..., ym ) be a system of local coordinates

Indeed, let (x1 , ..., xm ) be a system of local coordinates around p ∈ M . Since

The form ω is closed: dω = 0 since d ◦ d = 0 and it is non-degenerate. To

coordinate system (x1 , ..., xm , y1 , ..., ym ), this symplectic form is written as

A manifold M is said to be orientable if there exists on M an atlas such that the

T HEOREM 1.2.– (a) A closed differential 2-form ω on a differentiable manifold M of

T HEOREM 1.3.– Let α be a differential 1-form on the manifold M and α∗ λ the

P ROOF.– Since α : M −→ T ∗ M , we can consider the reciprocal image that we

A submanifold N of a symplectic manifold M is called Lagrangian if for all p ∈

E XAMPLE 1.3.– If (x1 , ..., xm , y1 , ..., ym ) is a local coordinate system on an open

Let M be a differentiable manifold, T ∗ M its cotangent bundle with the symplectic

T HEOREM 1.4.– (a) A symplectic morphism is a local diffeomorphism. (b) A

D EFINITION 1.7.– The one-parameter group of diffeomorphism gtX on M is called