Professional Documents
Culture Documents
Integrable Systems 1St Edition Ahmed Lesfari Full Chapter PDF
Integrable Systems 1St Edition Ahmed Lesfari Full Chapter PDF
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 ...
https://ebookmass.com/product/integrable-systems-1st-edition-
ahmed-lesfari-2/
https://ebookmass.com/product/fractional-order-design-devices-
circuits-and-systems-ahmed-g-radwan/
https://ebookmass.com/product/viroids-and-satellites-1st-edition-
edition-ahmed-hadidi/
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/
https://ebookmass.com/product/religion-extremism-and-violence-in-
south-asia-1st-edition-imran-ahmed/
https://ebookmass.com/product/reservoir-engineering-handbook-
fifth-edition-2019-tarek-ahmed/
https://ebookmass.com/product/endodontic-advances-and-evidence-
based-clinical-guidelines-1st-edition-hany-m-a-ahmed/
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:
www.iste.co.uk www.wiley.com
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.
Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 293
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 305
Preface
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.
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.
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.
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).
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.
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 ∗ (T ∗ M )
↑λ
λα (ξ) ϕ∗
R ←− T (T ∗ M ) −→ T ∗M −→ R2m
↓dπ ∗ ↓π ∗
α(ξ) ϕ
R ←− TM −→ M −→ Rm
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
Symplectic Manifolds 5
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.
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.
f ◦ π∗ ◦ f ∗ = π∗ [1.1]
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.
X
M −→ T M
idM ↓π
M
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.
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 .
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
Symplectic Manifolds 11
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,
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
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
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
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
Symplectic Manifolds 13
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
14 Integrable Systems
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
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
∂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
Symplectic Manifolds 15
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
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.
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)
.
iX iX ω = 0, iX (f ω) = f (iX ω),
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
Symplectic Manifolds 17
The following properties are often involved in solving practical problems using
Lie derivatives.
k
(LX ω)(X1 , ..., Xk ) = LX (ω(X1 , ..., Xk )) − ω(X1 , ..., LX Xj , ..., Xk ).
j=1
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
d) Just consider ω = f dxi1 ∧ ... ∧ dxik and λ = gdxj1 ∧ ... ∧ dxjl . We have
ω ∧ λ = f gdxi1 ∧ ... ∧ dxik ∧ dxj1 ∧ ... ∧ dxjl , and
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
We will now establish a fundamental formula for the Lie derivative, which can be
used as a definition.
Moreover, we have
and
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 ROOF.– Taking into account the previous theorem and the fact that iX iX = 0, we
obtain
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.
Similarly, we have
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
∂
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 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
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.
22 Integrable Systems
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
Symplectic Manifolds 23
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).
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
24 Integrable Systems
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
Symplectic Manifolds 25
T HEOREM 1.11.– The matrix that is associated with a Hamiltonian system determines
a symplectic structure.
⎛ ⎞
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 .
26 Integrable Systems
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.
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.
Symplectic Manifolds 27
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.
T HEOREM 1.12.– If F and G are two first integrals of a Hamiltonian system, then
{F, G} is also 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.
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
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
∂ 2 H ∂F ∂G
+ Jkl Jij [1.8]
i,j
∂xk ∂xi ∂xj ∂xl
k,l
∂H ∂ 2 F ∂G
+ Jkl Jij [1.9]
i,j
∂xi ∂xk ∂xj ∂xl
k,l
∂Jij ∂G ∂H ∂F
+ Jkl
i,j
∂xk ∂xi ∂xj ∂xl
k,l
∂ 2 G ∂H ∂F
+ Jkl Jij [1.10]
i,j
∂xk ∂xi ∂xj ∂xl
k,l
∂G ∂ 2 H ∂F
+ Jkl Jij [1.11]
i,j
∂xi ∂xk ∂xj ∂xl
k,l
∂Jij ∂F ∂G ∂H
+ Jkl
i,j
∂xk ∂xi ∂xj ∂xl
k,l
30 Integrable Systems
∂ 2 F ∂G ∂H
+ Jkl Jij [1.12]
i,j
∂xk ∂xi ∂xj ∂xl
k,l
∂F ∂ 2 G ∂H
+ Jkl Jij . [1.13]
i,j
∂xi ∂xk ∂xj ∂xl
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
Since the Jacobi identity must be identically zero, the expression to prove follows
immediately, ending the proof of theorem.
1.7. Examples
∂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
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
32 Integrable Systems
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
1
H= λ1 m21 + λ2 m22 + λ3 m23 , (Hamiltonian).
2
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.
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.”