Australian

Mathematical
Society

Manifolds
and Mechanics

ARTHUR JONES,
ALISTAIR GRAY
and ROBERT HUTTON

Lecture Series
2
 NUNC COGNOSCO EX PARTE

THOMAS J. BATA LIBRARY

TRENT UNIVERSITY
 Digitized by the Internet Archive
in 2019 with funding from
Kahle/Austin Foundation

https://archive.org/details/manifoldsmechaniOOOOjone
AUSTRALIAN MATHEMATICAL SOCIETY LECTURE SERIES

Editor-in-Chief: Dr. S.A. Morris, Department of Mathematics, La Trobe University

Bundoora, Victoria 3083, Australia

Subject Editors:
Professor C.J. Thompson, Department of Mathematics, University of Melbourne
Parkville, Victoria 3052, Australia
Professor C.C. Heyde, Department of Statistics, University of Melbourne,
Parkville, Victoria 3052, Australia
Professor J.H. Loxton, Department of Pure Mathematics, University of New South Wales
Kensington, New South Wales 2033, Australia

1 Introduction to Linear and Convex Programming, N. CAMERON

2 Manifolds and Mechanics, A. JONES, A. GRAY & R. HUTTON
Australian Mathematical Society Lecture Series. 2

Manifolds and Mechanics

Arthur Jones and Alistair Gray
Mathematics Department, La Trobe University

Robert Hutton
Comalco Ltd

~v
The null I of the

H if Univcrsilv of Cambridge
lo prim anil sell

8 ted all manner of hooks

was granted bv
Hem; I III in ISJ4
The University has printed

% ii and published continuously

since I3S4.

&mi} Unhonslty U'iw**

CAMBRIDGE UNIVERSITY PRESS fSTt**©

Cambridge

London New York New Rochelle

Melbourne Sydney
Published by the Press Syndicate of the University of Cambridge
The Pitt Building, Trumpington Street, Cambridge CB2 1RP
32 East 57th Street, New York, NY 10022, USA
10, Stamford Road, Oakleigh, Melbourne 3166, Australia

© Cambridge University Press 1987

First Published 1987

Printed in Great Britain at the University Press, Cambridge

Library of Congress cataloging-in-publication data applied for

British Library cataloguing-in-publication data

Jones, Arthur
Manifolds and mechanics.—(Australian Mathematical Society
lecture series; 2)
1. Differentiable manifolds
I. Title II. Gray, Alistair III. Hutton, Robert
IV. Series
516.3'6 QA614.3

ISBN 0 521 33375 X hard covers

ISBN 0 521 33650 3 paperback
 CONTENTS

PROLOGUE

1. CALCULUS PRELIMINARIES
1.1. Frechet Derivatives 3
1.2. The Tangent Functor 6
1.3. Partial Differentiation 10
1.4. Componentwise Calculus n
1.5. Variable-free Elementary Calculus 13

2. DIFFERENTIABLE MANIFOLDS
2.1. Charts and Atlases lg
2.2. Definition of a Differentiable Manifold 24
2.3. Topologies 26

3. SUBMANIFOLDS
3.1. What is a Submanifold? 28
3.2. The Implicit Function Theorem 31
3.3. A Test of Submanifolds 33
3.4. Rotations in R3 35

4. DIFFERENTIABILITY
4.1. Local Representatives 39
4.2. Maps to or from the Reals 41
4.3. Diffeomorphisms 42

5. TANGENT SPACES AND MAPS

5.1. Tangent Spaces 44
5.2. Tangent Maps 4g
5.3. Tangent Spaces via Implicit Functions 52

6. TANGENT BUNDLES AS MANIFOLDS

6.1. Charts for TM 55
6.2. Parallelizability 59
6.3. Tangent Maps and Smoothness 60
6.4. Double Tangents 62

7. PARTIAL DERIVATIVES
7.1. Curves in TQ 64
7.2. Traditional Notation 67
7.3. Specialization to TQ 72
7.4. Homogeneous Functions 76
8. DERIVING LAGRANGE'S EQUATIONS
8.1. Lagrange's Equations for Free-Fall 78
8.2. Lagrange's Equations for a Single Particle 81
8.3. Lagrange's Equations for Several Particles 83
8.4. Motion of a Rigid Body 86
8.5. Conservation of Energy 88

9. FORM OF LAGRANGE'S EQUATIONS

9.1. Motion on a Paraboloid 90
9.2. Quadratic Forms 93
9. 3. Lagrange's Equations are Second-Order 95

10. VECTORFIELDS
10.1. Basic Ideas 98
10.2. Maximal Integral Curves 101
10.3. Second-Order Vectorfields 104

11. LAGRANGIAN VECTORFIELDS

11.1. Globalizing Theory 110
11.2. Application to Lagrange's Equations 113
11.3. Back to Newton 117

12. FLOWS
12.1. Flows Generated by Vectorfields 120
12.2. Flows from Mechanics 125
12.3. Existence of Lagrangian Flows 135

13. THE SPHERICAL PENDULUM

13.1. Circular Orbits 138
13.2. Other Orbits, via Charts 141

14. RIGID BODIES

14.1. Motion of a Lamina 147
14.2. The Configuration Manifold of a Rigid Body 150
14.3. Orthogonality of Reaction Forces 154

REFERENCES 159

INDEX 161

SYMBOL TABLE 165

 PROLOGUE

These notes began as an attempt to bridge the gap between

undergraduate advanced calculus texts (such as Spivak(1965)) and graduate

texts on differential topology. One of the major applications of

differentiable manifolds is to the foundations of mechanics. Here a huge

gap exists between the classical literature (see for example Goldstein

(1980) and the more modern differentiable manifolds approach (see Abraham

and Marsden(1978)). Dieudonne(1972) says "The traditional domain of

differential geometry, namely the study of curves and surfaces in three-

dimensional space, was soon realized to be inadequate, particularly under

the influence of mechanics". Two of the authors of this work have been

involved in the writing of papers on problems in Lagrangian mechanics and

they soon realized the need for a simple but modern treatment of the

theoretical background to such problems.

Thus our aim is to make some of the basic ideas about manifolds

readily available to applied mathematicians and theoretical physicists

while at the same time exhibiting applications of an important area of

modern mathematics to mathematicians.

Classical texts use vague ideas such as "virtual work" and

"infinitesimal displacements" in their derivation of Lagrange's equations.

By contrast the modern texts jump straight to Hamiltonian systems and lose

the physical motivation. In these notes, the derivation of Lagrange's

equations makes direct appeal to geometrical principles. As far as the

authors are aware, these ideas do not appear elsewhere.

In his original work, Mecanique analytique, Lagrange used no

diagrams. In fact the geometrical ideas necessary for a thorough under¬

standing of classical mechanics did not exist at that time. It was only

with the development of differential geometry and differential topology

that geometry once again became an essential part of classical mechanics.

2 PROLOGUE

An advantage of starting with the Lagrangian formulation, as opposed to

the Hamiltonian, is that it follows directly from Newton's Laws of Motion

and that the mathematical background required is much less formidable.

These notes were developed primarily with a view to providing

the necessary theory for the study of particle motion as in the papers by

Gray et £1(1982) and Gray(1983). As we demonstrate, however, our treatment

of Lagrangian mechanics is equally applicable to the motion of rigid bodies

which consist of a finite collection of particles.

Our approach can be extended so as to apply to the motion of a

rigid body defined by a continuous mass distribution even in nonholonomic

cases but this work is beyond the scope of the present volume and will

appear separately.

We thank Lindley Scott for the diagrams of page 132, Sid

Morris for his encouragement, the referees for their constructive reports

and Professor Sneddon for reading and commenting on our preliminary notes.

Our debt to Judy Storey is profound and we thank her

accordingly. Without her monumental effort this manuscript would never have

reached its final form.

Arthur Jones

Alistair Gray

Robert Hutton

December, 1986.
 1. CALCULUS PRELIMINARIES

The differential calculus for functions which map one normed

vector space into another normed vector space is the main prerequisite

for the study of manifolds and Lagrangian mechanics in the modem idiom.

The mild degree of abstraction involved helps focus attention on the

simple geometric ideas underlying the basic concepts and results. The

insights and techniques which this approach fosters turn out to be very

worthwhile in the study of many topics in applied mathematics.

The idea of a function as an entity in its own right,

independently of any numerical variables which may be used to define it,

is fundamental for this approach to calculus. This idea, although usually

confined to statements of the theory, can easily be used to provide

computational tools for particular examples. This involves the

development of a "variable free" language of calculus in which the

functions themselves, as distinct from the values that they take are

highlighted. This enables us to formulate and work with many important

and difficult mathematical ideas in a simple manner. Some of the

notation does not appear elsewhere but a good account of the basic ideas
may be found in Spivak(1965) or in Lang(1964).

1.1. FRECHET DERIVATIVES

Here we outline how the idea of a derivative can be formulated for

functions which map one normed vector space to another. The basic idea

is this: a function is differentiable at a point if there is an affine map

which approximates the function very closely. The derivative of the

function at this point is then the linear part of this affine map.

1.1.1. Example. Let f: Rz R2 be given by

£(x,y) = (x2 + y2 + 1, xy + 1)
4 1. CALCULUS PRELIMINARIES

so that

f(l+h, 2+k) = f(1,2) + (2h+hk, 2h+k) + (h2+k2, hk).

As (h,k) approaches (0,0) the quadratic terms become negligible

compared with h and k and so f is approximated very closely by the

affine map A: R2 -*■ R2 with

A(1+h, 2+k) = f(1,2) + (2h+bk, 2h+k).

The linear part of A is the linear map L: R2 ->■ R2 with

L(h,k) = (2h+Uk, 2h+k).

The precise meaning of "approximates f closely" is clarified in the

following definition by the use of limits.

1.1.2. Definition. Let f: U —► F where E and F are normed vector

spaces and U is an open subset of E. By saying that f is

differentiable at a € U we mean that there is a continuous affine map

A: E —>- F such that

f (a) = A (a)

and

lim II f (x) -A (x) II _

x ->- a llar-all

1.1.3. Definition. For a map f as in Definition 1.1.1 we define its

Frechet derivative at a to be the unique continuous linear map

Df(a) : E —► F such that

lim II f (a+li) -f (a)-Df (a) fh) II _ Q B

& + 0 II Till

Now let us return to Example 1.1.2. and verify that in that case

Df (1,2) (h,k) = (2/z+Afe, 2 h+k).

Notice that the norm of a vector (x,y) £ R2 is given by II (a:,!/) II = Sx2+yi

and so in this case we have

lim Ilf (l+h,2+k) - f (1,2) - (2h+bk,2h+k) II

(fc.fcWO.O)
Hh,k) II

= lim II (h2+k2, hk) II

(h,k)+(0,0) '
II (h,k) II

and since the terms in the numerator are quadratic it is easy to show that
 1.1. FRECHET DERIVATIVES 5

this limit is in fact zero.

Notice that the Jacobian matrix for f at the point (1,2)

is just the matrix of the linear map Df(l,2). More about this later.

1.1.4. Theorem. (Chain Rule I). Let maps f: U —>- F and g: F —G

be differentiable at points a £ U and f(a) G F respectively where

E, F and G are normed vector spaces and U is an open subset of E.

The composite map g ° f: U —>- G is then differentiable at a and

D(g o f)(a) = Dg(f(a)) ° Df(a).

A proof of this theorem may be found in Spivak(1965). ■

A variable-free notation for functions which map normed vector

spaces to normed vector spaces (in most applications to mechanics these

normed vector spaces are, or are closely related to, Rn) turns out to be

very useful for our purposes.

1.1.5. Definition. The identity mapping from a normed vector space E

to itself is denoted by idE and defined by

idE(:r) = x for each x G E. ■

1.1.6. Definition. The constant mapping from E to F (E, F normed

vector spaces) which takes the value a G F is denoted by c_ and

defined by

c(x) = c for each x G E. ■

Usually we will omit the subscript "E" from idE when the context is

clear.

EXERCISES 1.1.

1. Show that D idg (a) = idE |[on. each a G E.

2. Show that f tA a conttnuouA ItneaA map between named vector

ApaceA then Df(a) = f.

3. Suppose p: Rn —>■ Rn tA di^enenttable and haA a dt^eaenttable

tnveAAe (that tA, p tA a dt^eomoApklAm). Show that
Dd>—1 (<f>(a)) = (D<j> (a))-1 ion each a £ Rn.

4. Let Mn denote the Aet oft aft aeal n*n matnfceA. It can be
Ahown that Mn can be made tnto a nomed vector Apace In Auch a
way that (,oa oil A, B G W1
DabII < IIAll IIbII .
6 1. CALCULUS PRELIMINARIES

Now dz^tne. f: M71 —> Mn by

f(X) = X2 .

Show that f -C6 dt^eAznttablz at A e If and (,tnd it* Fae.c.het

deAtvattve. Df (A).

1.2. THE TANGENT FUNCTOR

The ideas of this section will later be generalized to sets which are a

little more involved than normed vector spaces.

1.2.1. Definition. For an open set U in the normed vector space E we

define the tangent space TaE at a € U as the set

TaE = {{a,h)i h£ E}. ■

This tangent space can easily and naturally be given a vector space

structure.

1.2.2. Definition. Vector addition and scalar multiplication on TaE are

defined by putting

(a,h) + (a,/c) = (a, h+k)

\(a,h) = (a,\h)

for all vectors (a,h) and (a,k) € TQE and scalars X E R. ■

We may think of a typical vector (a,h) £ TaE as an arrow emanating from

a. Its tail is at a and its tip at a + h. See Figure 1.2.1.

Figure 1.2.1. Tangent Space at a.

 . .
1 2 THE TANGENT FUNCTOR 7

1.2.3. Definition. The tangent bundle of U at a is defined as

U TaE = U x e. ■
a £ U

It is clear from Section 1.1 that the derivative of a map f: U c E —*- F

(E, F normed vector spaces) U open, at a point a € U is the linear

map

Df(a) G C (E,F)

and that in some sense

A 0c) = f(a) + Df(a)(x-a)

is an "affine approximation" to f(x) near a. (See Figure 1.2.2). The two

important parts of this idea of affine approximation are the linear map

Df(a) and the point f(a). In the usual formulation of differentiation the

point f(a) is "lost" and the derivative Df(a) is all that is retained.

It turns out that the tangent map overcomes this problem (and

in doing so, leads to a rather neat version of the chain rule). The idea

is that whereas

the derivative Df(a) maps E to F,

the tangent map Taf should map TaE to Tf(a;F.

Notice that, as in Figure 1.2.3, we can think of the vectors h and

M(a)(h) as being "based at" a and f(a) respectively.

Figure 1.2.2. Affine Approximation at a.

8 1. CALCULUS PRELIMINARIES

1.2.4. Definition. Let f: U —’► F be differentiable at a 6 U where

UcE is open (E,F normed vector spaces). Then the tangent map at a of

f is the map

Taf; T"aE ~'* Tf(a)F

defined by
laf(a,h) = (f(a), Df(a)(h)) for each (a,h) £ TaE. ■

It is clear that the tangent map is linear. Now by allowing the point a
to vary throughout U we obtain a map defined on the whole of the tangent

bundle.

1.2.5. Definition. Let f, U, E, F, a be as in Definition 1.2.4. Then

the tangent functor

Tf: U x E —^ F x F

is defined by

Tf(a,h) = (f(a), Df(a)(ft)). ■

We are now in a position to formulate a very neat version of the chain rule.

You may find the comparison between the theorem below and Theorem 1.1.4

interesting.

1.2.6. Theorem. (Chain Rule II) Suppose g: UcE —>-VcF and

f: V —> G where E, F, G are n.v.s. and U, V are open and f,g are
differentiable. Then T(f°g) = Tf°Tg.

Figure 1.2.3. The tangent map at a.

 1.2. THE TANGENT FUNCTOR 9

Proof. Let a £ U, h £ E. Then

T(f°g)(a,7z) = (f°g (a), D(f°g) (a) (h))

= (f(g(a)), D f(g(a))°Dg(a)(h)) by Theorem 1.1.4.

= Tf(g(a), Dg(a)(h))

= Tf(Tg(a,h)).

Thus T(fog) = Tf°Tg as required. ■

This property of T, as given in Theorem 1.2.6, is usually expressed by

saying that T is a functor.

EXERCISES 1.2

1. Check that

M T - ldE*E

(6) Tf'1 - (Tf)-1

[provided o{j couaac f-1 extAtA and iA dt^eAentiable.)

2. Suppose n : E x e —*■ E : (a,b) a

and, by abuAe notation

II : F x F —>- F : (x,y) i—* x

Shou) that the diagram -In FiguAe 1.2.4 commute.

Tf
EXE -*• F*F

n n
f
E -► F

Figure 1.2.4.

3. Let f: (0,2tt) —*■ R2 be, given by f (a) = (cos(a) ,sin(a)).

(a) Sketch the image undeA f ofa (0, 2tt) .
(b) Compute Tf(a,l).
(c) Sketch Tf(a,l) oa an element o£ ^f(a)R2-
10 1. CALCULUS PRELIMINARIES

1.3 PARTIAL DIFFERENTIATION

The idea of a Frechet derivative is readily generalized to include partial

derivatives.

1.3.1. Definition. Let E^Ez and F be normed vector spaces, Ux ,U2

be open subsets of Ei and E2 respectively and suppose

f: Ux x U2 —*■ F is a differentiable function. Let (a,b) E Ux * U2.

The partial derivative of f with respect to its first

variable, 8xf(a,b)} is the unique continuous linear map from Ex to F

such that

lim Ilf (a+h,b) - f (a,b) -9-, f (g,fc) (h) II -

h 0 \\h\\

The partial derivative 32f(a,b) may be defined in a similar way. A

version of the chain rule for partial derivatives is then given by the

following Theorem.

1.3.2. Theorem. (Chain Rule for Partial Derivatives I).

Let Ex, E2, E and F be normed vector spaces, Ux and U2

be open subsets of Ex and E2 respectively

f: Ux x u2 -> E

be differentiable at (a,b) € Uj. x u2 and

g: E >■ F

be differentiable at t(a,b) £ E. Then for i = 1,2

%i (g°f) (.a,b) = Dg(f (a,i))o8^f (a.h). ■

It is easy to see how Definition 1.3.1. and Theorem 1.3.2. can

be generalized to a "function of n variables".

1.3.3. Lemma. Let f: R>! —► R be differentiable at a £ Rn. Then

n n
Df (a) I I 3jf(a)(Xj)
xjej
J=1

where the ej denote the usual basis vectors for Rn. ■

We will use idj to denote the projection function

 1.3. PARTIAL DIFFERENTIATION 11

EXERCISES 1.3
7. (Lse the chain Kate, panttal deAA.vati.vu to i>hove that ifa
f: R2 —* R2 it di^eAentiable then 3^(idj ° f) = id?- ° a^f fan
each i,3 = 1,2.

2. Let C1([0,1]) denote the 6et ofa once dififieAentiable function* on

the domain [0,1]. We define the evaluation map

ev: C1([0,1]) x [0,1] -»■ R: (f,x) -* f(x).

Find d1ev(f,x)(g)
and 32ev(f,x)(y).

1.4. COMPONENTWISE CALCULUS

Many of the important results in mechanics rely on the chain rule in its

various forms. When we deal with coordinates, which we will later call

"charts", a componentwise but variable free version of the chain rule proves

to be extremely useful. Thus in this section we concentrate on

derivatives of maps between Rn and Rm .

We will begin by introducing some (nonstandard) notation for

partial derivatives of real valued functions.

1.4.1. Definition. Let f: R?1 —► R be differentiable at a £ Rn. Then

the slope of f with respect to its ith variable (i = 1,2,...,n) is

denoted by f^ and defined by f ^ (a) = 3^f(a)(l). ■

The advantages of this notation will become apparent when we are faced with

the task of formulating a readable version of the chain rule for maps

between Rn and Rm. Note that classical texts would write

for f^ (a) and that

x-a
/•£
t (a) is a real number as distinct from dj,ffa) which is a linear map.

A further link with the classical treatment is provided by the following

definition and theorem.

1.4.2. Definition. Let f: Rn —► R^' be differentiable. Then the matrix

of the derivative of f at a £ R , that is the matrix of Df(a), is

called the Jacobian matrix of f at a and is denoted by f'(a). *

 12 1. CALCULUS PRELIMINARIES

1.4.3. Theorem. Let f: R,; —*■ Rm be differentiable. Then the ij-th

component of the Jacobian matrix of f at a is given by f4 (a).

Proof. Denote f by the ordered m-tuple of its component functions,

that is,

f = Cf i ,f2, • • • rri)

and so f^ : R?l —R for each i = 1,2Application of Lemma 1.3.3.

to each of the f^ then gives the required result. ■

1.4.4. Corollary. Let f: R —+■ Rm be differentiable. Then for each

x £ R

Df (rc) (1) = f'(x) = lim f (x+h) - f (tc)

h -* 0 h

Proof. Follows from Theorem 1.4.2. and Definitions 1.4.1. and 1.3.1. ■

We are now in a position to formulate a "component" version of

Theorem 1.3.2.

1.4.5. Theorem. (Chain Rule for Partial Derivatives II).

Let h: Rn -* RP, g: R™ -*• Rn be

differentiable. Then for each i = l,2,...,m

(hog)A = I h/jog gA .
J=1 3

Proof. Let a € Rm. Then for each i = l,2,...,m

(h°g )^(a) = 3^ (hog) (a) (1) (Definition 1.5.1.)

= Dh(g(a))o3^g(a)(1) (Theorem 1.3.2.)

= Dh(g(a)) (gA(a)) (Def inition 1.5.1.)

Now note that gA(a) = (g(a) , gA (a), . .., gA (a) ) and an application of

Theorem 1.4.3. is sufficient to complete the proof. ■

EXERCISES 1.4
1. Reh&i to example 1.1.1. UenthU Theorem 1.4.3. h0>l thti caie.

2. FHZ In the details oh the. pnoofa oh Theoaemi 1.4.3. and 1.4.5.

and CoA.olf.aAy 1.4.4.

3. Let <t>: Rm —> R^, y: R —► (A be cUhh&i&Atlable. Show that

(4>°y)' (t) d<Ky(*)) (y ' C *))

 1.5. VARIABLE-FREE CALCULUS 13

l[ox ca.ck t £ R.

[Hfnt: aie CotiolZany 1.4.4. and Thcoam 1.1.4.).

1.5. VARIABLE-FREE ELEMENTARY CALCULUS

The use of a variable-free notation in elementary calculus provides a means

by which many proofs can be greatly simplified and many ambiguities

removed. The ideas presented here will be used to some effect in later

chapters.

The identity and constant mappings as defined in Section 1.1.

here become

id: R —*- R: x I—*- x and

c_ : R —► R: x i—»• c.

Thus for example id3(#) = x3, and (id-3 + 3)(x) = ^3+ 3 for each x ^ 0.

The derivative of a function f: R —► R will be denoted by f'. For

example cos' = -sin and (id3)' = 3id2. We are now in a position to

formulate some of the basic ideas of elementary calculus in this language.

1.5.1. Lemma. (Chain Rule). For f: R —► R, g: R —► R where f and

g are differentiable

(fog) ' = (f'og) -g' .

Proof. See any standard calculus textbook. ■

1.5.2. Example. (cos°id3) ' = -sinoid3.(3id2)

= -3id2.sinoid3. ■

1.5.3. Definition. The indefinite integral from a of an integrable

function f: R —► R is denoted by f for a £ R and is defined by

'x a
f ](x) f for each x € R.
a

1.5.4. Theorem. (Fundamental Theorem of Calculus).

(i) Let f: I —► R be continuous, where 1 is an interval. Then for

each a £ I
r
f = f.
‘-'a J

(ii) Let ip': I —*■ R be continuous, where I is an interval, Then,

for each x € I and a € I

<p’ = <p(x) - <p(a).

aJ
 14 1. CALCULUS PRELIMINARIES

The proof of this theorem may be found in most respectable books on

elementary calculus. ■

An excellent example of the power and efficacy of a variable-

free approach is to be found in the proof of the next theorem. This is a

result that will prove useful in Chapter 12. It is instructive to compare

the treatment here with others. See for example Abraham & Marsden(1978),
page 63.

1.5.6. Lemma. (Gronwall's Inequality)

Let f: [0,a] —>- R+ and suppose

f < C + Kf on [0,a]

where C, K are each positive constants.

Then f < C&K’±d on [0,a] .

Proof. First we define g: [0,a] —*- R+ by

g = C + Kf

Thus f ^ g on [0,a]. Differentiating the above using Theorem 1.5.5.

gives

g ' = Kf < Kg

~K id
Now using the fact that g(0) = C and an integrating factor e ‘ in

the differential inequation

g' = Kg < 0
K id
gives f ^ g Ce on [0,a] as required. ■

Gronwall's inequality is the cornerstone of the proof that

solutions of certain differential equations vary continuously with the

initial conditions.

EXERCISES 1.5

7.(a) Use the. chain nule in variable ^Aee fioAm to fiind the deAivative
oi a function h defined by

h(x) = arctan(x) + arctan on R\{0}.

1
(Recall that arctan' -—-——
_1 + id

(fa) Show that h 14 not a constant function by evaluating h(l) and

h(-1)•
 1.5. VARIABLE-FREE CALCULUS

Use the fundamental theorem of calculus to thou) that If G' = 0

on an Interval I then G Is constant on I.

Explain the apparent contradiction between (fa) and (c) and

hence s ketch the function h.

Consider the Initial, value problem

<f>' + f.<t> = g. <t> (0) = l

where f,g are continuous on R. Use the fundamental theorem of

calculus to 6how that this problem Is equivalent to the Integral
equation
“ l + (f(f> + g) •
o
[Remember to prove the Implication both ways.)
 2. DIFFERENTIABLE MANIFOLDS

The idea of a differentiable manifold is a combination of ideas

from both analysis and geometry.
In geometry, differentiable manifolds include such things as
curves and surfaces and their higher dimensional analogues. Many of the
ideas to be introduced in this chapter are in fact motivated by the study
of the earth's surface in elementary geography.
In analysis, the role of differentiable manifolds is to provide
a natural setting, generalizing that of normed vector spaces, in which to
study differentiable functions. Thus the theory developed in Chapter 1,
permits you to differentiate functions mapping one normed vector space
(or some open subset thereof) into another normed vector space. But once
you have studied manifolds, you will also be able to differentiate
functions which map, say, a sphere into a torus.

2.1. CHARTS AND ATLASES

Although the surface of the earth is a sphere, small enough regions on it

will appear flat, like a plane. A geographical atlas is in fact just a

collection of maps or pictures, each of which lies in a plane.

Each such picture determines a function $ from some region

U of the earth's surface into the plane R2 as shown in Figure 2.1.1.

This idea leads to our first definition, in which we replace the earth's

surface by an arbitrary set M and ignore nearly everything about the

function <j> except that it maps one-to-one onto some "flat" region

that is, an open set in some Euclidean space R^. The integer n
will be assumed fixed for the rest of this section.

2.1.1. Definition. Let M be a set. A chart for M consists of a subset

U of M together with a one-to-one function <p which maps U onto an

 2.1. CHARTS AND ATLASES 17

open set in

Figure 2.1.1. A chart (U,cj>)

Before collecting together the individual charts to build up an

"atlas" for the entire set M, we introduce a certain compatibility

condition to ensure that the various charts fit together nicely.

To this end, consider a pair of charts for M, say (U,tj)) and

(V,^). Provided these charts overlap, that is, provided U 0 V i □, we can

form the composite map <j)otp 1. It will map the set

ip (U (1 V) c= Rn onto the set tp (U ft V) c Rn

as shown in Figure 2.1.2. The map <poip-1 is called a transition map or an

overlap map since

= (<poip-1) oip on U H V.

Figure 2.1.2. Overlap Maps

 18 2. DIFFERENTIABLE MANIFOLDS

Thus, compatibility as defined below means that we can change

charts smoothly. Note also that the definition involves the concept of

differentiability in a context where it is already familiar to you

— namely, for functions between open sets in Rn.

2.1.2. Definition. By saying that the charts (U, <j>) and (V, i/i) are

compatible we mean that if U fl V ^ □ then

(i) the sets ()>(U fl V) and i|j(U fi V) are open subsets of Rn.

(ii) the map ^oi/T1 is a diffeomorphism between these two sets.

p
The charts will be called C compatible if the overlap map is a

diffeomorphism of class Cr (1 ^ r ^ °°) . ■

Finally, the long awaited definition of an "atlas", which will

contain information about the whole set M:

2.1.3. Definition. An atlas for the set M is a collection of mutually

compatible charts

{(11£, 4^)1 i e 1}
such that

U = M. ■
iCL

2.1.4. Example. To get some feeling for what has been going on here think

of your old school atlas. Extract from it a pair of charts (U,t}>) and

(V,tJj) and then think about the map cjioifi'-1. ■

The key to resolving many problems in applied mathematics lies

in a suitable choice of "coordinate system". From the modern viewpoint,

such co-ordinate systems are simply examples of atlases. The following

example shows how a pair of "angular co-ordinates" can be regarded as an

atlas on the unit circle.

2.1.5. Example. Let S1 be the unit circle in R2 given by

S1 = {x £ R2 : IjcI = l}.

Charts (U,<j>) and (V,\p) for S1 may be defined as follows:

Put U = S \{ (— 1,0) } and let <f>: U —*■ R be given by

4>(x) = angle which the vector x makes with

the first axis, chosen so that -ir < 4>(x) < x.

 2.1. CHARTS AND ATLASES 19

Put V = S1\{(1,0)} and let ij; : V —> R be given by

ip (x) = angle which the vector x makes with

the first axis, chosen so that 0 < < 2tt.

Figure 2.1.3. An atlas for S1

With the aid of Figure 2.1.3 it may be seen that

fid on (0,tt)
(jjctlT1 = <
^id-2ir on (tt , 2tt) .

This map tf>o1 is thus a diffeomorphism between the sets

^(U fl V) = (0,2tt)\{ tt } and (J)(U (1 V) = (-it ,tt)\{0} .

Thus the charts (U, cja) and (Vare compatible, where S1 is covered by

U and V. Hence these two charts form an atlas for S1. ■

EXERCISES 2.1

1. Ton each oh the ^oltouoln.g AetA, gtve an atlaA contaA.nA.ng juAt one
chaAt:
(a) M aa an open Aet -in Rn

(fa) M aa the gnaph oh a not neceA^afUZy continuouA function

f : I —»■ R u)ken.e I -La an open -inteAvat on the fieat tine,
(c.) M aa an n-dimenAtonal vecton. Aubipace oh R^ (n < k).
 2. DIFFERENTIABLE MANIFOLDS

Lzt U conbibt o& ail pointb (xx,x2) on thz unit ciActz with
x2 > 0 and lzt v conbibt oj$ thz pointb with > 0. Choobz mapb

cf> = idi |U and <J; = id2 |v

whzaz id: and id2 a>iz thz poojzction mapb on R2, ab An FiguAz
2.1.4.

Figure 2.1.4.

(a) Statz thz domain ofi (pop-1 and dzacue a tfoamula giving ttb
valuzb.
(b) Vzduzz that thz chantb (U,<j>) and (v,\p) faon. S1 aaz
compatiblz.
(c) How maid you gzt an atlab containing thzbz two chaAtb?

izt thz bztb u and V be ab in ExzAcibz 2 and izt mapb

9 : V —*■ R and 0 : U —»- R

be dzfiinzd by putting, fan zach x = (x1,xz) in thz azlzvant domain,

9 Or) = thz anglz a which thz vzcton. x

makzb with thz fainbt axib ("y< a <j)»

ab indicatzd in EiguAZ 2.1.5. Lzt $ be ab in quzbtion 2 abovz.

 2 1 . . CHARTS AND ATLASES 21

-i 1 A- f

Figure 2.1.5.

(a) State the domain ofi 6o<J)-1 and faind a formula fior this map.
(fa) Hence -6how -that -the chart.& (V,e) and (U,cfa) j$on s1 one
compatible.

The unci sphere sn c Rn+1.

In ihci exercise i/ou wkXfa construct an atlas, consisting otwo
charts, ion the. n-dimensional unit sphere

S = { (Xi j • • . 9 | *^1 + . . . + **V2+1 ~ 1) •

(a) Let <j>N be the map obtained by projecting £rom the north pole
N = (0, ..., 0, 1) onto the equatorial plane

{(#! , x
n+i
) \ x ,
n+i
0}

Figure 2.1.6.
Show, by uAing Aimilar triangles, that 6or each x in the set

S^\{N}
22 2. DIFFERENTIABLE MANIFOLDS

(X J » • • • » Xy^ > 0 )
1 - x
n+1
(b) In a Almllan way de;CLne <j>s a* a pnojectlon tfnom the Aouth
pole, s = (0, .... 0,-1) and obtain an analogous fanmula on.
thlA map.

[c] Nou) Identify the. equatorial plane with Rn and Ahow that
1
4>n 4><;0c)
ll<t> s Or) II2

(d) Veduce that ion each y € Rn\{0}

1
° (y) y
»yi
and hence Ahow that

(sn\{N} , 4>n) and (Sn\{s}, <J>S) .

one a pain ofa chantA which fanm an atlaA fan. Sn.

Show that the map

(r,6) : {(a,b) £ R2 : a > 0} —R+ x ( 77 ^

V 2’V
given by
r (a,b) = Sa2+b2

Q(a,b) = arctan

defaneA a chant fan R2 which Ia compatible with the Identity chant

id : R2 —v R2 : (a,i>) —*- (a,b).

Let u be the AubAet otf the unit Aphene s2 c R3 conAlAtlng ofi the
polntA (fcx, x2, x3) with x2 > 0 and let V conAlAt o the polntA
with x3 > 0. Let the mapA

(0,<j>): D+ R2 and ip: V -* R2

be given by putting, fan all x In u on v neApectlvely,

<P(x) = a (-it < a < tt)

0(x) = 0 (0 < 6 < 5jtt)
ip(x) = (X!,XZ)

whene a and 6 one the angleA Ahown In Tlgune 2.1.7.

 2.1. CHARTS AND ATLASES 23

(a) Ex.pA.e2>-!> a and 6 cu> fanctlont (xl3 x2, xz) £ U ft V and

hence fend a ^onmula fan. the map

(6,<p)°ip~1 : u n v -*■ R2 .

(fa) Hence thou) that the chant* (u, (e,<f>)) and (v.iji) fan. s2 one
compatible.

[Note•• the chant (e,<j>) hat been obtained by nettnlctlng "tphenlcal polan.
coon.dlnatej> " (r, 0, cf>) to poJit o{ the unit tphene.)

Figure 2.1.7.
 24 2. DIFFERENTIABLE MANIFOLDS

2.2. DEFINITION OF A DIFFERENTIABLE MANIFOLD

A given set M may have many different atlases. By saying that an atlas

A for the set M is equivalent to another atlas B for the same set, we

mean that every chart in A is compatible with every chart in B.

It is easy to verify that this defines an equivalence relation

on the set of all atlases for the set M. Each equivalence class 3 of

atlases is said to be a differentiable structure for the set M.

A differentiable manifold is an ordered pair (M, 3) where M

is a set and 3 is a differentiable structure for M.

The maximal atlas of a differentiable manifold (M, 3) is the

collection of all charts which belong to at least one atlas in the

differentiable structure 3-

A chart in the maximal atlas is called an admissible chart for

the differentiable manifold.

In speaking of differentiable manifolds we shall often omit the

word "differentiable" and call them simply "manifolds".

Strictly speaking, a manifold consists of a set M together

with a differentiable structure 3- Where there is no risk of ambiguity,

however, we shall often omit reference to 3 and call H itself the

manifold.

The one-chart atlas consisting of the identity function

id: Rn —>■ Rn determines a differentiable structure for R , in which the

admissible charts are the diffeoraorphisms between open subsets of Rn.

When we speak of R as a manifold, we mean it to have this differentiable

structure.

The more generalized concepts of C -equivalent,

-differentiable structure and Cr-differentiable manifold are defined in

an analogous way on replacing the word "compatible" by "C^-compatible"

OO
in the first paragraph of this section. Smooth will mean € .

EXERCISES 2.2.

1. Show that (R, id) and. (R, id3) aAe two chaAtA fioA R which aAe
incompatible.. [Thin AhowA that theAe ate at leant two dl^eAent
dl^eAentiable AiAuctuAeA on R. UeventheleAA, theAe two AtAuctuAeA
will lateA be Ahown to be "dl^eomoAphlc".)

-• -Show that the compatibility between chantn <\ok a &et M Ia not an

equivalence delation.
 2.2. MANIFOLDS - DEFINITION 25

Check that, nevertheless, equivalence of atlases Is an

equivalence delation.

3. Let A be a maximal atlas for a set M and let (u,4>) be a chant

In A.

Consider a set V c u. Shoul that (v, <|>|v) is also a chant In A

If and only If, <j>(v) Is open In Rn.

(This shows that a maximal atlas must contain a lot of chants).

4. Dimension of a manifold.
Let (M, £) be a dlffenentlable manifold and let A and B be two
equivalent atlases In the differentiable structure $. Show that If
all the chants In A map Into Rn and all the chants In B map
Into Rm then m = n.
(This Integer Is called the dimension of the manifold.)
(Hint: Let (U,<|>) £ A, (V,i|0 £ B with U n V ± □.
Then conslden D(<(>1) (a) fon some a £ \p (u n V),)

5. Hot all manifolds have dimension.

Show that the subset of R2 sketched In Flgune 2.2.1. Is a manifold
by constructing a suitable atlas.

Figure 2.2.1. A disconnected manifold.

6. Product manifold.
Let (M, 2) and (N, Z) be two manifolds.
Check that the following collection Is an atlas fon the set M * N:

{(U*V, <j>x<|/) : (U,i}>) is an admissible chart for (M,S)

and (V,^) is an admissible chart for (N,C)}

(The resulting manifold Is denoted by (M*n, 2x%) and Is called the

product of the two original manifolds.)
26 2. DIFFERENTIABLE MANIFOLDS

2.3. TOPOLOGIES

We now show how a differentiable manifold can be made into a topological

space. This will make it meaningful to speak about the continuity of maps

between manifolds.

Before reading further into this section, however, you may wish

to revise your topology. See for example Simmons(1963).

2.3.1. Definition. Let M be a differentiable manifold. By saying that

a set A c M is open we mean that for each x E A there is an admissible

chart (U, 4>) for M such that x € U and U c A. ■

Figure 2.3.1. An open subset of M.

It is easy to verify that the collection of all "open sets" defined in

this way satisfies the axioms of a topology for the set M. We shall say

that this topology is induced by the differentiable structure.

Note that for each chart (U,t(>) the set U is open in this

topology.

We shall suppose that the induced topology is Hausdovff and

second countable, like the usual topology on Rn.

EXERCISES 2.3.

1. Check that the collection ofi "open i>eti>" defined on the manifold M

doet> Indeed AatL&fiy the coelom fiosc a topology.

 2.3. TOPOLOGIES 27

Show that a subset oh the. manihold M is open ih and only ih it can

be mitten as a union oh domains oh admissible chants.

Show that {[on each admissible chant (u,<J>) oh M

[a] 4>-1 is continuous

(fa) <f> is continuous .

Ton. (fa) use Exencise 2.2.3.

Let M be an open subset oh Rn togethen with the usual

dthheAentcable stnuctune, containing the injection map as a chant.
Show that the topology induced on M Is the nelative topology.

Let sn be the unit sphene in Rn+1 with the usual dihheAentiable

stnuctune, induced by the polan pnojeciions oh Exencise 2.1.4.
Show that the topology induced on sn is the nelative topology.
[Hint: Show that the polan pnojections <j>N and <j>s can be extended
to continuous maps dehined on open subsets oh Rn+1).

Show that all connected maniholds have dimension.

 3. SUBMANIFOLDS

It is possible to give some quite nasty looking subsets of R

the structure of a differentiable manifold as in Exercise 2.1.1(b). In
Yl
this chapter, however, we shall single out certain "nice" subsets of R
which correspond to our intuitive ideas of smooth curves, surfaces, etc.
We shall introduce the idea of a submanifold to describe these goemetric

objects precisely.
A submanifold of R^1 is essentially a subset of R which can,
at least locally, be "flattened out" into a subspace of Rn. A useful
criterion for a set to be a submanifold of Rn is given by the implicit

function theorem.

3.1. WHAT IS A SUBMANIFOLD?

We would like to be able to call a sphere S a "submanifold" of R3. In

Figure 3.1.1. the diffeomorphism sends the open set U <= R3 into the

open set (U) c R3 . At the same time it flattens out the piece of the

sphere SOU into a subset of the plane R2 * {0} (which is of course a

Figure 3.1.1. A submanifold chart for S.

 3.1. WHAT IS A SUBMANIFOLD? 29

3.1.1. Definition. Let (U, c|>) be a chart for a manifold M where

4> : U —v R . By saying that this chart has the submanifold property for a

subset S of M we mean that, for some integer k ^

(v x e s n u) <p(x) = .... <pfc (x), o, ..., o)

and, furthermore,

<Ks n u) = <t>(u) n Rk x {o}. ■

We would like to be able to claim that the restricted map

<j)|s 0 U was a chart for the set S. There is, however, no hope of this if

k < n since then the map would not be onto an open set in Rn.

We can easily get over this by dropping the zeros from the end

of the n-tuple <p (%) • The image of this restricted map is therefore the

set 4> (S fl U) . Since <}>(U) is open in Rn this image set may be

identified, on dropping the zeros, with an open set in R^. In this way,

we may regard the map <j> | S fl U as a chart for the set S.

It may happen that there are enough charts for S of the above

form to make up an atlas for S.

3.1.2. Definition. Suppose that S is a subset of a manifold M which

has an atlas consisting of charts of the form

(s n u, no<j>|s n u)

where (U, <£) is an admissible chart for M with the submanifold property
Yl J^
for S and where n: R —*■ R is the natural projection, given by

L (^i > • • • y^k * • * ~ i > • • • y^kf •

The set S, together with the differentiable structure containing these

charts, is then called a submanifold of M. ■

Note that, as a manifold, S then has dimension k.

EXERCISES 3.1.

1. Show that main open i>u.bi>et oh a manthold M a Aubmanthold oh M.

2. Let S be the i>et {(x,y): y = x} tn R2 and let <j>:R2 —> R2 with

4'(.Xyy) = Cc, y-x)

Find the Image undeA. <}> oh the i>et s.

Show that <t> Ia a dlhh&omoaphtAm, hence an adnuAtlble chant hon R2.

deduce that s a nubmanthold oh R2.

30 3. SUBMANIFOLDS

3. ini f: I —> R white I -Li an open Interval -in R.

Show that the. gmph oh f -Li a iubmanlhold oh R2 provided f -Li

dlhh etentLabh.
[Hint: you. mid a chant (u, <f>) for R2 with tki iubmanlhold
pmpeAty h°* thi gmph. A plctute plot a Little experimentation
should lend you to a suitable ho*mula ho* §(x,y) In tinmi oh f(x)
and y.)

4. Prove that every victor iubipace oh Rn otf dlminilon k -Li a

iubmanlhold oh Rn oh duminilon k.

5. Find a dlhhetintlabh hunction 0 : R2 -> R2 which mapi the h^cit

axli onto thi gmph oh thi ab-ioluti value ho-nctlon abs ai In
Figure 3.1.2.
(Hint: thi nixt exetelii -iayi thti can't happen lh 6 -Li a
dlh h^omorphlim.)

Figure 3.1.2. Impossible if 6 is a diffeomorphism.

6. Show that them -Li no dlhh&omorphlim 9: U —»• v, white U and V

ate open ieti In R2, which mapi

U D first axis onto V D graph of abs

Viduci that thi gmph oh abs cannot bi a iubmanlhold oh R2.

1. Compote thi above ixetetii with Exetelii 2.1.1(b).

Givi an ixamph oh thi iubiet oh R2 which can bi given a manlhold
itmetute but which nevertheliii li not a iubmanlhold oh R2.

S. Let M bi a manlhold oh dlmenilon n and let S bi a iubmanlhold

oh M oh dlminilon k. Suppon that (U, 4>) and (v, ate
admliilbh charti ho* M with thi iubmanlhold property ho* S and
let n: Rn —► R^ bi thi natu*al projection.
 3.2. IMPLICIT FUNCTION THEOREM 31

(a) Let I : Rk —*- Rn be the natuAal Injection given by

I (^1 9 • • • 9'^fc) ~ (*^1 » • • • ,3^, 9 0, . . . ,0)

By dnawlng a suitable mapping diagram convince youn*elh That the

following equality between map* hold* on the domain Jloip (u fl v (1 s):

(iio<t>)o (no,},)-1 = n o <j> ° ip 1 ° i

(fa) Show, via the chain fuxle, the compatibility oh the chant* ion. s:

(u n s, n°<j>|u n s) and (v n s, no</i|v n s) .

[Thu* the dlhhenentlable *tnuctune on a *ubmanlhold l* unique.)

9. Show that "being a *ubmanlhold oh" l* a tnan*ltlve nelatlon.

10. Let s be a *ubmanlhold oh M. Show that the topology Induced on

S a* a *ubmanlhold oh Vi l* the *ame a* the nelatlve topology on
S a* a *ub*et oh M.

3.2. THE IMPLICIT FUNCTION THEOREM

To motivate this theorem, consider the following question: given a function

f: R2 —* R, does the set of points

{(*>2/) | f(x,y) = 0}

form the graph of some function?

Unfortunately the answer is no. Consider the function

f : R2 —¥ R defined by

f(x,y) = x1 + y1 - 1.

The set of points f-1(0) is then just the unit circle in R2 and this

cannot be the graph of any function.

So let's replace the above global question by a less ambitious

local question: given a point (a,b) £ f_1(0), is there a neighbourhood

N = 1 x j of (a,b) such that the set

N D f_1(0)

forms the graph of some function?

In the case of the circle the answer if yes provided the point

(a,b) is neither (1,0) nor (-1,0) as is obvious from Figure 3.2.1. Note

that at these excluded points the partial derivative 82f is zero.

The implicit function theorem answers our local question for an

arbitrary smooth function f : pfL+m —>- .

32 3. SUBMANIFOLDS

3.2.1. Theorem. (Implicit function theorem).

Let f: U<=Rnx R777 —k r"7 where U is open and suppose that

f is C*” (r ^ 1).
Suppose there is a point (a,b) in U such that f(a,b) - 0

and suppose that the linear map

32f(a,b) : R™

is a vector space isomorphism.

There is then a neighbourhood N = ixj c u of (a,b), where

I and J are open sets in Rn and R^ respectively, and a function

g : I —*- J such that

V (x,y) € ixj f(x,y) =0 «=» y = g(a:).

function g vs t .

Proof. See Spivak(1965) or Dieudonne(1960). ■

The following corollary is an important step towards our goal

of establishing a criterion for f 1(0) to be a submanifold of R

Figure 3.2.1.

3.2.2. Corollary. Let the map f and the neighbourhood N of (a,b)

be as in Theorem 3.2.1. There is then a map <f> : N R such that

(N,<(>) is a chart for Rn+m with the submanifold property for f_1(0).

Proof. This homes in on the fact that the set f 1(0) D N is the graph

of a function. Details are left as an exercise. ■

 3.3. SUBMANIFOLD TEST 33

EXERCISES 3.2.

1. Ut>e the implicit function theorem to prove the [geometrically

obvious} claims made cut the beginning ofi section 3.2. {,on. the
example

f(xsy') ~~ x 3- y ~ I.

2. Let g : i —>- j be tr where I and j are open sets In Rn

and respectively. With a little experimentation you should be

able to guess a formula fior a map <J> : I * J —*- Rn x Rm which

(a) map6 the graph o& g onto I x {0} as in Figure 3.2.2.

(b) it, a di^eomorphism $rom I * J onto an open t>et in Rn x Rm.

Hence prove Corollary 3.2.2.

Figure 3.2.2.

3.3. A TEST OF SUBMANIFOLDS

The following theorem provides a useful criterion for showing that certain

sets are submanifolds of Rn+m.

3.3.1. Theorem. Let f : U c Rn x |f -)■ Rra where U is open and

suppose that f is C (r > 1).

If at every point (a,b) of f-1(0) the linear map

„ ,s nn+m nm
D f(a,b) : R —* R

has rank m, then the set f—1(0) is a submanifold of Rn+m whose

dimension is n.
34 3. SUBMANIFOLDS

The idea of the proof is to use Corollary 3.2.2. of the

implicit function theorem to get a chart defined on a neighbourhood of the

point (a,b) with the submanifold property for f—1(0). Before applying

this corollary, however, we may need to "change co-ordinates" to ensure

that the point (a,b) satisfies the hypothesis relating to 32f(a,b).

The desired change of co-ordinates will be constructed with the

aid of a technical result from matrix theory. This involves the use of a

permutation matrix to move the columns of a matrix. Consider the following

simple example:

'l 2 4

V.
3 0 0^

This matrix has rank 2, that is 2 of its columns are linearly independent

(for example. and ). If we want to "shift all of the rank" into

the last two columns via a linear diffeomorphism, we could apply the map

whose matrix is

1
0

Clearly the determinant of this matrix is -1 0) and

’l 2 4' ro 1 o' '2 1 4'

3 0 0 1 0 0 — 0 3 0

0 0 1

as required. What we have achieved here is a shifting of all of the rank

into the right hand columns by using a 3x3 matrix whose columns consist

of the linearly independent vectors e and e3.

The 2x2 submatrix

1 4

obtained from picking out those last two columns is the matrix of a vector

space isomorphism from R2 to R2.

Now as an exercise see if you can construct a 4 x 4 matrix

similar to the one above which "moves all the rank" of

3 2 11'

4 4 2 1

V.
12 11 >
 3.3. SUBMANIFOLD TEST 35

to the right (the answer is not unique).

What is required is the following lemma:

3.3.2. Lemma. Let A be an mx(n+m) matrix of rank m. Then there

exists an (n+m) x (n+m) matrix Q, all of whose columns are linearly

independent and consist of the vectors el3 62^ ...., &n+m (not necessarily
in that order) with the property that

A Q P

is a nonsingular m*m matrix and P is the (n+m) x m

'0 0 ... o'

• • •

• • •

0 0 ... 0

I
mx-m

(Here P has the effect of picking out the required mxm submatrix).

The proof is just an exercise in elementary matrix theory. ■

EXERCISES 3.3.

7. U.i>e Theorem 3.3.7. to prove that the. unit Aphere sn Xa a

Aubmanlfiold ofi R ofi dlmenAlon n.

2. Let 0 (n) be the Aet ofi alt nxn orthogonal matrlceA regarded cla
a AubAet otf Rn n. Prove that 0(n) Xa a Aubmanl^old oft Rn*n 0^
dtmenAton %n(n-1).

3. Let S0(n) be the Aet ofi aJUL nxn flotation matrlceA, that tA,
orthogonal matrlceA with determinant 1. Show that S0(n) Ia an
open AubAet o{> 0(n). Hence Ahow that SO(n) Xa clIao a Aubmanl^old
ofi Rn*n ofi dXmenAlon %n(n-l).

4. Here we prove Theorem 3.3.7. by uAlng Lemma 3.3.2.

(a) Ua e the lemma to a how that there Xa a dl^eomorphlAm

q: Rn+m —► Rn+m Audi that 82f°q(q 1(a,b))

Xa a vector Apace XAomorphXAm &rom Rm to R^.

(b) Now apply Corollary 3.2.2. o(, the Implicit function theorem to
the map f°q to deduce the existence ofi a chart which Xa defined on
36 3. SUBMANIFOLDS

cl neA.Qh.bouAh.ood oh the potnt q 1(a,b) ccnd hcu> the tubmanlhold

pAopeAty {,0A (foq)-1(0).

(c) With the aid oh FlguAe 3.3.1. deduce the existence oh a chaAt
which hi defined on a neighbourhood oh the point (a,b) and hat the

tubmanlhold pAopeAty hofl f—1 (°) •

" f"1(0)

q 1 ia,b)

Figure 3. 3.1.

3.4. ROTATIONS IN R3

Exercise 3.3.3. states that S0(n) is a submanifold of H of dimension

%n(n-1). The set of 3x3 rotation matrices SO(3) is important in the

study of rigid body motion, which will be dealt with in detail in Chapter 14.

A summary of the results which will be needed is given below.

3x3
3.4.1. Lemma. SO(3) is a 3-dimensional submanifold of R

Proof. This is a part of Exercise 3.3.3. ■

3.4.2. Lemma. SO(3) under the operation of matrix multiplication is a

group.

Proof. Elementary matrix theory shows that SO(3) is closed under matrix
multiplication and matrix inversion and that it contains the identity. ■

We will denote the linear map with matrix A relative to the

usual bases for R3 by La. Thus for each a £ R3

 3.4. ROTATIONS 37

La (a) = Aa

where on the right we regard a as a column matrix. We will call the map

La a rotation for reasons which will soon be apparent.

3.4.3. Lemma. Rotations preserve distance.

Proof. Let a £ R3 then for A £ S0(3)

Hall2 = aPa

- aPtPka since A2'A = I

= (Aa)^(Aa)

= II Arz II 2 .

Thus for a,b £ R3

Ila-ill = llA(a-&)ll

= IIAa — Ai>II as required. ■

3.4.4. Definition. The orientation of the ordered triple of vectors

(a1}a2Ja3 ), (b^b 2,b3) and {cx>c2. a3)

given by the sign of the determinant

ai bx Cl
det a2 b2 b2
_a3 b3 ?3

Note that the determinant in the above definition is simply the scalar

triple product (ax,a2,a3) x (bx,b2}b3). (21^2^3).

3.4.5. Lemma. Rotations preserve orientations.

Proof. Let A £ SO(3). The orientation of any triple of vectors

(x* (yi*y2*y3)*(3^,22J33)

in R3 is the sign of

y1 21

det 2/2 32 = det X, say.

33 J
L 3

while after rotation their orientation is the sign of

r i
*1 yi 2l
det A A A 22 = det AX
*2 yx

. X3> . yx. . Z3. .

= det A det X

= det X, as det A = 1.
Thus the orientation is the same.
 38 3. SUBMANIFOLDS

Another useful property of elements of S0(3) is as follows:

3.4.6. Lemma. A matrix A £ SO(3) is uniquely determined by its effect

on a linearly independent pair of vectors in R3.

Proof. Let A and A' be two matrices in SO(3) which map a linearly

independent pair of vectors a,b in R3 into the same pair of vectors

x,y respectively, which must also be linearly independent. Since rotations

preserve distance and orientation, axb must be mapped to xxy in each

case. Thus A and A' have the same effect on the basis a, b, a*b for

their domain and hence must be identical. ■

EXERCISES 3.4.

1. LeX A £ SO(3) and let la be the corresponding linear map.

(a) Prove that la maps the. untt sphere s2 in R3 onto Itself.
(b) Prove that the dot product a.b = a}b stays the same when the
the two vectors a and b one mapped under LA.

1 . LeX A € S0(3). Prove that 1^ \ Is a complex eigenvalue o& A

and x a corresponding eigenvector while X and x axe thelx

complex conjugates, then
-T - -T
x x = \X x x

hence XX =1.

3 . Use the fact that the pxoducX o{, the eigenvalues o^ a matrix
[each counted with appropriate multiplicity) is equal to the
determinant ofi the matrix to prove that every matrix A £ SO(3)

has an eigenvalue equal to 1. [Physically, the elgenspace

corresponding to the eigenvalue 1 will be the axis ofa rotation
0& a rigid body.)
 4. DIFFERENTIABILITY

Differentiable manifolds provide the natural context in which

to discuss the differentiability of functions.

In advanced calculus differentiability was studied for maps

between open sets in normed vector spaces. In this chapter the concept of

differentiability will be extended so as to apply to maps between

differentiable manifolds.

4.1. LOCAL REPRESENTATIVES

A manifold is a set which can be locally identified with an open subset of

a Euclidean space.

This property enables us to define smoothness properties of

functions between arbitrary manifolds.

4.1.1. Definition. Let M and N be two differentiable manifolds and let

f : M y N. Suppose furthermore that (U, $) and (V, <Jj) are admissible

charts for M and N respectively, with f(U) c V. The map

ipofcxp-1 : (j)(u) —*- \p(V)

is called the local representative of f relative to the two given charts

and will be denoted by f ^. ■

Figure 4.1.1. Local representatives.

40 4. DIFFERENTIABILITY

Since the local representative f^w, maps an open set in some

Euclidean space into another such set it makes sense to speak about its

differentiability, as understood in advanced calculus. Hence the following

definition.

4.1.2. Definition. Let f: M —*■ N and let a £ M. By saying that f is

differentiable at a we mean:
For each admissible chart (V, ip) for N with t(a) £ V

there is an admissible chart (U, <p) for M with a £ U such that

f(U) c V and the local representative f^ is differentiable at <p('a). ■

Note that the definition has been framed in such a way that

differentiability of f will ensure its continuity.

By saying that f is differentiable on a subset S of M we

mean that f is differentiable at every point of S. By saying that f

is differentiable we mean that it is differentiable on M.

We shall also assume the analogous definitions in which the

r
word "differentiable" is everywhere replaced by "is of class C (r ^ 1)".
CO
Smooth means C .

EXERCISES 4.1.

7. Let L, M and N be manifioldA. Show that if f: L —*• M and

g: M —> N are O > I) then So it gof : L —* N.

2. Let M be a set and let N be a manifold. Suppose, furthermore,

that there it, a bijeetive map f: M —N. Show how to define an
atlas for the set M which makes the map f to be Cr.

3. Let M and N be manifolds. Show that if f: M —*■ N it, lr

then for each open t>et u c m and each open t>et v =>f(u) in N, the
restricted map f|u : u —> v it, alto C2", where the open sets one
to be regarded at tsubmanifoldt.
Prove alto the converse.

4. Let Nx and n2 be manifolds and let : Nx * N2 —*■ be the

natural projection (i = 1,2). Show that each it Cr.
(The product of two manifolds wat defined in exercise 2.2.6.)

5. Let M and n be manifolds. Show that if f: M —*■ N it continuous

in the sense of Definition 4.1.2, then f it continuous at a map
between the topological spaces M and N with topologies given by
Definition 2.3.1.
 4.2. MAPS BETWEEN REAL SPACES 41

4.2. MAPS TO OR FROM REAL SPACES

Yl
We may regard R as a manifold with its usual differentiable structure

containing the identity map as a chart. Thus if M is any manifold it

YL
makes sense now to speak about the differentiability of a map f: M —>■ R
r\Vl
or a map g: K —*■ M.

For such maps the definition given in the previous section may

be reduced to a somewhat simpler criterion, as follows:

4.2.1. Theorem. Let f: M —*■ Rn and let a £ M. The map f is Cr

at a if and only if for each open set V in Rn with f(a) £ V there

is an admissible chart (U, <j>) for M with a £ U such that

(i) f(U) c V

(ii) the map f°<J>-1 : 4> (U) —>- Rn is of class tr at if (a).

The proof of this result is left as one of the exercises,along

with the formulation of an analogous result for the differentiability of a

Yl
map g : R —> M. See Figure 4.2.1. ■

Figure 4.2.1.
42 4. DIFFERENTIABILITY

EXERCISES 4.2.

1. Let x and Y be open heth In Rm and Rh h.ehpecttoely, ho that

X and Y can be KegaAded ah mani^oldh.

Shoal that a map f : x —*■ Y ih dti^enenttable ah a map between

mani^oldh ifa and only it ih dt^enenttable ah a map between open
heth.

2. Vfiooe theorem 4.2.1.

3. Formulate a theorem analogouh to the one In the text {,oa a map

g : x —*■ M wheAe M ih a manifold and x ih an open het In R”.

Vaow a diagram and pAove youA theorem.

4. (a) Let M be a manifold. Show that a map f : M —*■ R,! ih C2'

and only ifi each ofi ith component faunctionh f^: M —*■ R ih Cr.

(b) Let S be a hubmantfiold ofi M and huppohe f : H —► N

(n a manifold ) ih Cr.

Show that f|s : S —* N ih alho Cr.

4.3. DIFFEOMORPHISMS

Now that we have a definition of differentiability of functions which map

manifolds to manifolds we are in a position to extend the idea of a

diffeomorphism to include such functions.

4.3.1. Definition. A map f: M —*■ N, where M and N are manifolds is

said to be a diffeomorphism if it is differentiable and has a differentiable

inverse. If f and its inverse are lr it is said to be a Cr

diffeomorphism. ■

4.3.2. Definition. Two manifolds M and N are said to be diffeomorphio

to each other if there exists a diffeomorphism from M to N. ■

We shall regard two manifolds as being "essentiablly the same" if they are

diffeomorphic to each other. This idea is summed up by the following

lemma.

4.3.3. Lemma. Suppose the manifold M is diffeomorphia to the manifold

N and {(U^,^) | i £ 1} is an atlas for M. Let f: M —*■ N be a
diffeomorphism.
Then {(f(U^), <t>^°f *) | £ £ 1} is an atlas for N.
 4.3. DIFFEOMORPHISMS 43

Proof. Follows from the definition of a chart. ■

The above and the following lemma will be useful when we deal
with the motion of a rigid body.

4.3.4. Lemma. Let L : R’ > R be a linear one-to-one map (where

m ^ n). Suppose M is a submanifold of Rn.

Then L(M) is a submanifold of Rm which is diffeomorphic to
M.

Proof. See exercises. ■

We now mention some celebrated results referred to by

Spivak(1970) volume 1, chapter 2 and Stern(1983).

There are, up to diffeomorphism, unique differentiable

structures on each Rn if n ^ 4. There are at least three "exotic"

differentiable structures on R4 and it is speculated that there are

uncountably many.

There are, up to diffeomorphism, unique differentiable

structures on Sn for n < 6. But on S7 there are 28 essentially

different differentiable structures, and for S31 there are over 16

million.

EXERCISES 4.3.

7. Show that ti (U, <j>) Ia an admlAAlble chant ok a manifold M

then <p Ia a dlfaeomonphlAm nom u to <j>(u) <= Rn. Compare

ExenclAe 2.3.3.

2. The two dnaAtA (R, id) and (R, id3) genenaute two distinct
cUfaeAentlable AtAuctuAeA fan. R oa theAe chantA aJie not compatible.
Show that neventheleAA the two AeAultlng manlfaldA one difaeomonphlc.

3. Vtiove Lemma 4.3.4.

 5. TANGENT SPACES AND MAPS

In the previous chapter the idea of differentiability was

generalized so as to apply to maps between manifolds. In this chapter the
idea of the derivative itself will be generalized to this context.
In order to retain the idea of differentiation as a process of
linearization it will be convenient to introduce, at each point of a
manifold, a vector space known as the tangent space at that point. These
vector spaces vn.ll supply the domains and codomains for the linearized maps.
A neat formulation of the chain rule then becomes possible via a
generalization of the tangent functor, from Section 1.2, to manifolds.
Although all the concepts of this chapter can he formulated for
arbitrary manifolds, our discussion will be restricted to submanifolds of
R . This will provide an adequate background for later chapters on
mechanics while avoiding unnecessary abstraction.

5.1. TANGENT SPACES

In elementary geometry one studies tangents to circles and tangent planes

to spheres. The definition of a tangent space will generalize these ideas

to arbitrary curves and surfaces and their higher dimensional analogues.

Let M be a submanifold of Rn and let a 6 M. Intuitively,

we want the tangent space at a to consist of all the arrows emanating

from a in directions which are "tangential" to M at a, as in Figure

5.1.1. The arrows may be regarded as elements of the vector space TaR^,

introduced in Section 1.2. To capture the idea of "tangency" we use

parametrized curves in the manifold.

5.1.1. Definition. A parametrized curve in M at a is a map y: I —► M

with y(0) = a, where I is an open interval containing 0. The image

Y(I) of the map y will be called a curve in M at a and y will be

said to parametrize this curve.
 . .
5 1 TANGENT SPACES 45

The parametrized curve will be called differentiable (or of

class e) if the map y is differentiable (or of class €r). ■

Since M c Rn, the notation in Corollary 1.4.4 is applicable

to y and can be interpreted as follows: If y(£) is the position of a

particle moving along the manifold at time t then y'(£) is the velocity

of the particle. Intuition requires this to be in a direction "tangential"

to the manifold and leads to the following definition, illustrated in

Figure 5.1.2.

Figure 5.1.1. A tangent space at a.

5.1.2. Definition. A tangent vector to M at a is an element of TaR”

of the form (a, y'(0)) where y is a smooth parametrized curve in M

based at a. For brevity, we shall sometimes denote this tangent vector

by Ma- "
 46 5. TANGENT SPACES AND MAPS

5.1.3. Definition. The tangent space to M at a is the set of all

tangent vectors to M at a and will be denoted by TaM. The tangent

bundle of M is the union of all its tangent spaces and will be denoted
by TM so that

TM = U TaM. ■
a £ M
The definition of TaM implies that it is a subset of the

vector space T^R^ from Section 1.2. The following theorem confirms the

expectation that the tangent space should be "flat", rather than "curved"
like a typical manifold.

5.1.4. Theorem. The tangent space TaM is a vector subspace of TaRn

of the same dimension as the manifold M.

Proof, The proof will use a chart (U, cf>) for Rri at a which has the

submanifold property for M. Hence,in the notation of Section 3.1,

<t>(Mnu) = Rfe x {o} n <ku)

as illustrated in Figure 5.1.3. We may assume that <j>(a) = 0.

Figure 5.1.3. A submanifold chart.

We shall check that T^M is closed under vector addition and

leave it as an exercise to prove that it is closed under scalar

multiplication. So let (a, y'(0)) and (a, 6 r(0)) be an arbitrary pair of

vectors in TaM, with y and 6 parametrized curves in M based at a.

As cf> is a submanifold chart it follows that both
 5.1. TANGENT SPACES 47

<t>oy and <j>o(5 map into R x {0}

and hence also

(<J>oy) f and (<}>o6)^ map into x {0}.

From Corollary 1.4.4 this implies that

D(<J>oY)(0)(1) and D(<M)(0)(1) £ R^ x {0}

and hence by the chain rule. Theorem 1.1.4,

D(J>(a) (y ' (0)) and D<j>(a) (<5' (0)) £ R^ x {0}.

Hence by linearity of Dip (a)

Dip (a) (y '(0) + & ' (0)) £ Rk x {0}.

The vectors involved are shown in Figure 5.1.4.

We may thus define a map e from an interval into M by

putting

e(t) = r1(tD<f>(a)(Y,(0) + 6'(0)) .

Since e is a curve in M at a, the vector (a, e'(0)) is an element of

TaM. By the chain rule, however,

e ' (0) = D(j)-1 (0) oD<(> (a) (y ' (0) + 6'(0))

= y'(0) + 6 ’ (0)

and so, by the definition of vector addition in TaRn,

 48 5. TANGENT SPACE AND MAPS

(a,Y'(0)) + (a, 6'(0)) = (a, e'(0)) £ TqM .

Thus T M is closed under vector addition.

a
The proof that T^M is closed under scalar multiplication is
similar and is left as an exercise, as is the statement that T M has the
a
same dimension as M. I

EXERCISES 5.1.

7. Let m be the unit circle {(a:i,X2): Xj2+ xz =1} In R2 and

let a = (a1,a2) 6 M. Show that If y = (yi»Y2) rt a parametrized
curve In M based at a then

Yi(0) Yi'(0) + Y2(0) y2'(0) = 0 .

Deduce that each vector In T^M Is orthogonal to the vector, (a,a)

In T R2 and Illustrate with a sketch,
a
2. Show that If M Is the unit sphere In Rn and a £ M then each
vector In T M Is orthogonal to the vector (a,a) In T Rn .
a a
3. Let M be a submanifold of Rn, let a £ M and let b £ Rn .
Suppose that the distance from b to a point x £ M has a minimum
at x = a. Prove that every vector In T^M Is orthogonal to the
vector (a, a-b) In T Rn and Illustrate with a sketch.
a
4. Complete the proof of the part of Theorem 5.1.4 which says that
Tn as a vector subspace of T^R by proving the closure of T^m
under scalar multiplication.

5. Let M be a submanifold, and u an open subset, of Rn. Show that

M n U Is a submanifold of Rn and that TM D Tu = T(m n u) ,

6. Let M be a submanifold of RP and let a £ M. Prove that If

(U,<j>) Is a submanifold chart for M at a as In the proof of
Theorem 5.1.4, then

TaM = {a} x D4>-1 (0)(Rfc x {o}) .

Do this by showing (l) RHS c LHS and III) LHS c RHS. Techniques
used In the proof of Theorem 5.1.4 with come In handy.

7. Use the result of Exercise 6 to show that the tangent space T^m
has the dimension k (which Is also the dimension of m), thereby
completing the proof of Theorem 5.1.4.
 5.2. TANGENT MAPS 49

5.2. TANGENT MAPS

The idea of a "local linear approximation" to a map between manifolds can

now be formulated in a precise way with the aid of tangent vectors. To

this end let f: M —*■ N be differentiable, where M and N are

submanifolds of and Rn, and let a E M.

Intuitively, a sufficiently small piece of a curve in a

manifold will be indistinguishable from a line segment and hence will

correspond to a tangent vector to the manifold. The map f sends a small

piece of curve through a in M to another small piece of curve through

f(a) in N, as illustrated on the left of Figure 5.2.1.

Figure 5.2.1. The tangent map of f at a.

By now imagining the pieces of curve to be so small that we can regard them

as tangent vectors, we get the map, shown in the figure as Taf, which sends

tangent vectors in TaM across to tangent vectors in This map

is, in a sense, an approximation to f near a and, hopefully, it will

be linear.
By introducing parametrizations for these curves as in Figure

5.2.2 and using their derivatives to define the tangent vectors, we are

led to propose the following definition.

5.2.1. Definition. The tangent map of f at a is the map

 50 5. TANGENT SPACES AND MAPS

with

Taf(a,y'(0)) = (f(a),(f°y)'(0))

or, equivalently, Tf([y] ) = [foY]f. ,

u f (a)
for each differentiable parametrized curve y in M at a. The tangent
map of f is the map

Tf: Tm —► Tn

given by

Tf|TaM - Taf for each a £ M. ■

Figure 5.2.2.

It is necessary to check that the proposed definition is

unambiguous since two distinct curves y and 6 may have the same

derivative at 0. We want to be sure that when this occurs the derivatives

of the curves f°y and fc6 will also be equal at 0, as illustrated in

Figure 5.2.3.
 5.2. TANGENT MAPS 51

5.2.2. Proposition. For all differentiable parametrized curves y and

6 in M at a

[y]. [6], [f °y]

f (a)
[ f o <S ]
f (a)

Proof. It is sufficient to show that

y'(0) = 6'(0) =* (f°y) '(0) = (f o 6) ' (0).

The chain rule. Theorem 1.1.4, cannot be applied directly on the right hand

side of this implication because the domain of f need not be an open set

in a normed vector space. To overcome this difficulty, we introduce a

chart (U, $) for Rn at a with the submanifold property for M and

then write, at least on a neighbourhood of 0 £ R,

foy = (f o <fi-1) oijioy .

Now by Corollary 1.4.4 and the chain rule

(foy)'(0) = D(foy)(0)(1)

= D(focf)-1) (cj> (a) ) °D<j)(y(0)) oDy(0) (1)

= D(f o<t)-1) (<(> (a)) oD(j> (a) (y ' (0)) .

This, together with a similar expression for (f°6)'(0), establishes the

desired implication. ■

Inasmuch as the Frechet derivative of a map at any point of its

domain is linear, the above expression for (f°y) '(0) shows it to be a

linear function of y'(0).

5.2.3. Theorem. The tangent mccp of f at a

V T M t,, ^
f (a)

is linear.

With the aid of the tangent map, the chain rule for maps

between manifolds may now be expressed in the following elegant way.

5.2.4. Theorem. (Chain Rule). Let g: L —*- M and f: M —* N be

differentiable where L, M and N are submanifolds of Euclidean spaces.
The composite fog: L —*- N is then differentiable and

T(fog) = Tf°Tg .

If a £ M then we may also write

Ta(fog) = Tg(a)foTag.
 52 5. TANGENT SPACES AND MAPS

Proof. Proving the differentiability of the composite was set as Exercise

4.1.1.

To complete the proof of the theorem, let y be a

differentiable parametrized curve in L at a so that g°y is a

differentiable parametrized curve in M at g(a). From Definition 5.2.1

of the tangent map

T(f°g) ([y]a) [(f°g)°y]

f(g(a))

[fo(g°y)]
f(g(a))

Tf«8.Vlg(o))

= Tf(Tg([y]a))

= Tf°Tg([y]a). ■

EXERCISES 5.2.

7. Show that TidM = idjM .

2. Cleanly £oa. each Cr (r > 1) chaAt (u,<f>) o{> the manifold M

c(>°<f> 1 = id^^ and <t>~10<p = idjj. U&e the. chatn Ante and queitton
1 above to deduce that T<t>: Tu -+ T((U)) ha& both a le{,t and a
Alght tnveAte and that

(Tcf))— 1 = T(<|>_1) while Tc)) (Tu) = T(<j>(u)).

3. Let f: Rw —Rn and let M and N be Aubmanlfioldi ofi Rm and

Rn Aet>pecllvely, Auch that f(M) c n. PAove that f ti
dl^eAentlable then ao <u> It* AeAtnlction f | M : M —*■ N and that

T(f|M) = Tf|Tm.

5.3. TANGENT SPACES VIA IMPLICIT FUNCTIONS

We now turn to submanifolds of which have the form

M = {x: f(ar) = 0}

where f: R1 * —► Rm is a differentiable function for which Df(a) has

full rank m for each a £ M. For such submanifolds there is a very easy

way to find the tangent space: just linearize the function f defining

the submanifold near the appropriate point. This is the intuitive content
of the following theorem.
 5.3. IMPLICIT FUNCTIONS 53

5.3.1. Theorem. The tangent space to the above submanifold at a £ M

is given by
TaM = {(a,h): h £ with Df(a)(fe) = 0}.

Proof. We will deal only with the case in which the partial derivative

32f(a): Rm —*- Rm has full rank m, leaving it as an exercise to derive

the result in the remaining case.

Suppose first that (a,h) £ TaM so that h = y'(0) for some

curve y in M based at a. From the definition of M, f°y = 0 and

hence by the chain rule

Df(a) (y '(0)) = 0 .

Thus (a,/z) is a member of the set described in the theorem.

Conversely, suppose

Df(a)(h1,h2) = 0 (1)

where h = (h1,h2) with hx £ Rn and hz £ Rm. If we now construct a

curve y in M based at a such that h = y’(0), then it will follow

that (a,h) £ TaM so the theorem will be proved.

To this end note that since 32f(a) has full rank, the implicit

function theorem. Theorem 3.2.1, shows that there is a differentiable

function g such that

f (x) = 0 «=> x2 = gOi) (2)

for all x - (xlsx2) in a neighbourhood of a. We define the curve

Y = (Yi»Y2) by putting
y1(t) = a2 + thx, y2(t) = g(yx(t)) (3)

for t sufficiently small, where a = (a1,a2). Since a2 = g(a±) it

follows from (3) that y(0) = a. From (2) and (3) it follows that

f(Y1(«,Y2(«) = f(Yi(t), g(Y1(t)))= 0

so by the chain rule

Df(a)(Y1'(0), y2'(0)) = 0. (4)
Since, however, 32f(a) has full rank, it follows from (1), (4) and

hx = Y^O) that h2 = Y2'(0) ; hence h = y'(0) as required. ■

54 5. TANGENT SPACES AND MAPS

EXERCISES 5. 3.

7. UAe Tke.oA.rn 5.3.7 to Akon) that tke tangent Apace to the u.ntt Apkene

Sn_1 = {x £ Rn: x.x = 1}

at a point a £ Sn_1 Ia

TaSn = {(a,h): h £ Rn with a-h = 0} .

2. TMe Tke.on.rn 5.3.7 to a koto that the tangent Apace to o(n), defined.
In ExendAe 3.3.2, at the Identity matntx I ,c4

TjOCn) = { (I ,H) : H is a skew-symmetric matrix of order n} .

3. UAe tke IdeaA contained In ExendAe 3.3.4 to extend tke pnooh oh

Tkeonexn 3.3.1 to tke caAe In toklck 32f(a) doeA not neceAAanlly
kave (ja££ nank.
 6. TANGENT BUNDLES AS MANIFOLDS

In this chapter we show that the tangent bundle Tm of a

manifold M is a manifold in its own right. Furthermore it has a

differentiable structure that is induced naturally by the differentiable

structure of M. These ideas can then be extended to form T(TM) which

is the natural setting for second order differential equations which are

fundamental in mechanics and other applications.

6.1. CHARTS FOR TM.

Recall that the tangent bundle TM of a submanifold M of Rn is

defined by

TM = U TQM .
a £ M

This is a subset of TRn and hence may be regarded as a subset of RnxRn.

For example, the tangent bundle TS1 is the collection of all

the tangent spaces to the unit circle. Although it is strictly a subset

of R4, it is nevertheless helpful to vizualize it in the plane by means

of tangent lines attached to S1, as in Figure 6.1.1.

Figure 6.1.1. A tangent bundle TS1

 56 6 . TANGENT BUNDLES

It is even easier to vizualize TS1 if we give each tangent line a

rotation perpendicular to the plane of the circle to form a cylinder, as

illustrated in Figure 6.1.1. This latter interpretation is given formal

justification in Exercise 6.1.1. Meanwhile the following results show

how submanifold charts for M lead, via differentiation, to submanifold

charts for TM.

6.1.1. Lemma. Let M be a Cr(r ^ 2) submanifold of Rn of

dimension k. If (U, <J>) is a chant for Rn with the submanifold

property for M, then (Tu, T<{>) is a chart for R2n such that

T<f>(Tu fl TM) = T(<f>(U)) fl (Rfe x {0} x x {o}).

Proof. Let (U, <J>) be as in the hypothesis of the lemma. By Exercise

5.2.2, the map T<J>: Tu —*■ Tc|>(TU) is one-to-one and onto the set

T«|»(TU) = T(<j>(U)) = <f>(U) x Rn,

which is open in R . Hence (TU, T<j>) is a chart for R2n. Finally,

T(j)(TU 0 TM) = T<j>(T(U fl M)) by Exercise 5.1.5.

= T((U fl M) ) by Exercise 5.2.2.

= T(4>(U) H (R^ x {0})) by Definition 3.1.1.

= (<KU) n (R^ x {0})) n (Rfe x {0} x Rk x {o})

by Definition 1.2.3.

= T<f>(U) n (Rfe x {0} x Rfe x {0})

by Definition 1.2.3. ■

Thus all that remains to get a chart for TM is to shift zeros

to the right. This can be achieved by applying the permutation

P : R2n —> R2n with

P(x, w, y, z) = (x, y, w, z)

for x,y E R and w,z E RS ^ . This gives the following corollaries.

6.1.2. Corollary. If (U,<(>) is a chart for Rn with the submanifold

property for M then (Tu, P°(T<j>)) is a chart for Rzn with the

submanifold property for TM. ■

 6.1. CHARTS FOR TM 57

6.1.3. Corollary. If M is a tr (r ^ 2) submanifold of RM then

Tm is a 1 submanifold of R2n. The dimension of TM is twice the
dimension of M. ■
The following example provides an illustration of Corollary

.. .
6 1 2

6.1.4. Example. A chart (U,(0, r-_l)) on R2 is defined by

U = {(a,b) £ R2: a > 0}

0(a,i) = arctan

(r-1) (a,6) = /a2 + b2 - 1 .

This chart has the submanifold property for S1 since, on U D S1,

a2 + bz = 1 so (0, r-_l){a,b) = (arctan (^j’ 0).

We now compute T(0, r-_l) and show that, to within a

permutation, this is a chart for TR2 with the submanifold property for

TS1. Let (a,b) £ U and (h,k) £ R2. Then

D0 (a,b) (h,k) = D arctan^y0D^-r^-^(a,&) (h,k.)

b idi
+ (h,k)

-bh + ak
a2 + b2

and

D(r-l)(a,2>)(fc,fc) = D(id2° (id^+ida2)-^) (a,b) {h,k)

- Did2(a2+b2) oD(id12+id22)(a,b)(h,k)

ah + bk
~ Ja 2 + b2
Putting this together and identifying TR2 with R1* we obtain

-bh + ak ah + bk
T(0, r-1)(a,b)(h,k) = (arctan
’(!)•/a 'a2+b2-l.
a2 + b2 ‘£2 + b

Restricting this to TS1 gives

-bh + ak
T(0, r-1) (a,b)(h,k) = (arctan —), 0, , 0
TS1 a2 + b2
 58 6. TANGENT BUNDLES

since by Exercise 5.1.1. ah + bk = 0 on S1. Thus (U x R2,T(0, r-1))

has the submanifold property for TS1 as required. ■

6.1.5. Theorem. If (U, 4>) is a (r ^ 2) chart for a submanifold

M of Rn then (Tu, Te}>) is a C21-^ chart for JM.

Proof. Follows from Corollaries 6.1.2 and 6.1.3. ■

EXERCISES 6.1.

7. Show that the. map, faom the. tangent bundle oft the unit cotcie to the
cylinder in R3,

4>: TS1 —► {(a,b,c) £ R3: a2 + b2 = 1}

given by

$((a.b),(h,k)) = (a,b, -bh + ak)

it a di^eomoaphitm.

[Hint: neieti to Example 6.1.4. and then conttauct chaatt fast Ts1
and the cylinder with aspect to which the local aepaet entativ e o{,
$ it the identity mapping).

2. Let U = R2\{(0,b) : b > 0} and

\pN(a,b) =

(r-l)(a,b) = Ja2 + b2 - 1.

(See Figune 6.1.2.)

Figure 6.1.4. Extended stereographic chart.

 6.2. PARALLELIZABILITY 59

(a) Show that (U, (i|>N, r-1)) iA a chant ion R2 with the. Aubmaniiold.
pnopenty t$oa s1.

lb) Use an angument Aimilan to that oi example 6.1.4 to A how that

(Tu, T(^n, r-_l)) iA a chant ion TR2 with the Aubmaniiold pnopenty
ion Ts1.

6.2. PARALLELIZ ABILITY

In the previous section we showed that Ts1 = S1 x R and made the
Yl Yl Yl
observation that TR = R x R , It therefore seems natural to conjecture
k
that in general TM = M x R . This conjecture is false.

6.2.1. Definitition. Let M be Cr (r ^ 2) k-dimensional submanifold of

Rn. If there exists a diffeomorphism

$ : TM —> M x R

which takes each tangent space TaM by a linear isomorphism to {a} x R

then M is called parallelizable. ■

An important example of a manifold that is not parallelizable

and which arises in the study of the spherical pendulum is S2.

6.2.2. Theorem. S2n is not parallelizable. The proof of this theorem

involves the use of the "Hairy Ball Theorem" (see Hirsch(1976)) and for

a sketch of the proof itself see Chillingworth(1976). ■

We do, however, have "local" parallelizability as stated in the

following lemma.

6.2.3. Lemma. Let M be a C2* (r > 2) k-dimensional submanifold of

R22 and let (U, <J>) be an admissible chart for M. Then Tu = U x R^.

Proof. See exercises. ■

We quote a deep result, mentioned in Chillingworth(1976),

whose proof lies outside the scope of this book.

6.2.4. Theorem. T Sn = Sn x Rn only in the cases n = 0, 1, 3, 7. ■

EXERCISES 6.2.

1. Pnove Lemma 6.2.3. by quoting the appnopniate neAuit inom Section 6.1.
 60 6. TANGENT BUNDLES

2. Suppose M t& a. Mob-in6 AtAip (See F-LguAt 6.2.1). 1-6 M

paAaZl.eJLLza.btt? Gtvz gzomztAtc AtaAom ok youA an&weA.

Figure 6.2.1. A Mobius Strip

6.3. TANGENT MAPS AND SMOOTHNESS

Corollary 6.1.3. enables use to give a tangent bundle the structure of a

manifold. Put simply if (U, 4>) is a chart for M then (TU, T<}>) is a

chart for TM. There are clearly other admissible charts for TM which

are not obtained in this way but they are not of interest to us since they

may not preserve the linearity of the tangent spaces and hence may not

preserve the linearity of local representatives of tangent maps.

6.3.1. Definition. Let i G 1} be an atlas for the C2* (r ^ 2)

submanifold M of Rn. Then the collection of charts for TM given by

{(TU^, : i E I) is called the natural atlas for TM. ■

It is now meaningful to discuss differentiability of maps which

have tangent bundles as their domains. An important example of such a

map is given by the following definition.

6.3.2. Definition. Let M be a C2* (n ^ 2) submanifold of R‘. The

natural projection on the tangent bundle of M is the map tm: TM —*■ M

with

xM(a,7i) = a

for each (a,h) E TaM. ■

 6.3. SMOOTHNESS 61

6.3.3. Theorem. The natural projection tm is of class C°°.

Proof. We check the conditions in Definition 4.1.2. Let (a,h) £ TM and

let (U, <f>) be a Cr chart for M. By Theorem 6.1.5, (Tu, Tcf>) is a

chart for TM and x^(TU) c U.

The local representative of relative to these charts is

the map

<f>oTM°(T<f>)

From Figure 6.3.1, which is validated by the definition of the tangent map

T<|>, it is clear that this map just the natural projection 11: R2n —s- Rn,
which is C". ■

Figure 6.3.1. Local representative of Tj^.

Since formation of a tangent map is essentially a process of

differentiation, it is not surprising that it may lead to a loss of one

degree of differentiability.

6.3.3. Theorem. Let M and N be (r ^ 2) submanifolds of Rm and

Rn respectively. If f : M —*■ N is of class then Tf: TM —> Tn is
of class e 1.
62 6. TANGENT BUNDLES

Proof. This again involves straightforward checking of the conditions in

Definition 4.1.2. ■

EXERCISES 6.3.

I. Let f : M —*■ N be a* in The.on.em 6.3.3. Show that the local

n.epn.e* entativ e oi Tf i>ati*iie*

Tf T6 Tip = T fw

uiheae (u, 40 i* a chant ion. M, (v, \p) i* a chant ion. N and

f(U) c v. See TlguAe 6.3.2.

Figure 6.3.2. Local representative of Tf.

2. Une question 1 to iill in the detail* oi the pnooi oi Theonem 6.3.3.

6.4. DOUBLE TANGENTS.

In Section 6.1 we showed that if M was a manifold then TM was also a

manifold with a differentiable structure which is induced naturally by that

of M. It is possible to repeat this process and obtain an atlas for

T(TM). We will denote by T2M the manifold T(TM).

 6.4. DOUBLE TANGENTS 63

6.4.1. Lemma. Let M be a C (r ^ 3) submanifold of of dimension

k and suppose (U, <J>) is an admissible chart for M. Then (T2U, T2<j>)

is a chart for T2M. Furthermore T2M is a submanifold of (R”)“* of

dimension 4k.

Proof. Follows from Lemma 6.1.1, Corollary 6.1.2. and Corollary 6.1.3. ■

Various other results follow from those of previous sections for example:

T2 (f°g) = T2f °T2g

and T2f = T2f

' r T24> T2^ 1

when each of the above expressions is properly defined. We will not fill

in the details but the following lemma gives a little insight into the
structure of double tangent maps.

6.4.2. Lemma. Let g: Rn -»• Rm be twice differentiable. Then

T2g: (Rn)4 — (R"V and

l2g(a,h,k,l) = (g(a) ,Dg(a) (?z)JDg(a) (k) ,D2g(a) (Jn,k) + Dg (a) (-£)).

Proof. Follows from Definitions 1.1.3 and 1.2.4. ■

EXERCISES 6.4.

7. Fill in the. detailt, oh the poooh oft Lemma 6.4.2.

2. Notice that xM: Tm —+■ M and xM(a,?2) = a. Thai Tr : T2m —*■ Tm.
M
(a) Suppose (a,h,k,l) € T2M. Find ln.ta,h,k,l).

(fa) Show that = xM ° XjM>

(c) Figure 6.3.1. thorn the local ne.pn.eo> entativ e. oh xM with netpcct to
national chantt. Find the local nepneoentativet oh TxM and x^
with netpect to suitable, national chanti.

(dj Fon what tubtet oh T2m -a xT = TxM ?

IM
 7. PARTIAL DERIVATIVES

The tangent bundle of a differentiable manifold provides a

natural setting for the study of particle motion. Before deriving
Lagrange 's equations of motion we establish some technical results on
partial differentiation with respect to charts. The resulting formulas
have a classical look but are here given a precise and useful meaning.

7.1. CURVES IN TQ

To specify the state of a particle moving on a manifold Q at any given

time one specifies both the position and the velocity of the particle.

Hence it is natural to regard the particle as tracing out a curve in TQ

rather than in Q. The motion of the particle is thus described by a

parametrized curve

c : I TQ

where I is an open interval in R. We may write c in terms of its

components as (ci, c2) where C;l is the projection of c onto Q,

Tqoc : I —*■ Q,

which traces out only the position of the particle, as in figure 7.1.1.

Figure 7.1.1. The position of a particle.

 7.1. PATHS IN TQ 65

Intuitively

c : I —*■ TQ

can be thought of as the parametrized curve tq°c together with an

attached "arrow" at each point of Tq°c(I), as in Figure 7.1.2.

o c

Figure 7.1.2. The position and velocity.

In this interpretation c(t) consists of Tq0c(tj (the position of the

particle on Q) together with an "arrow" (the velocity of the particle).

The arrow should point in the direction of motion of the particle, that

is, it should be tangential to the image of Tqoc at the point in

question and the length of the arrow should give the speed of the particle.

These requirements may be expressed in terms of the components of c by

the condition that

c2 (t) = ci ' (£)

and hence that

c(t) = (c1(t),c1 ' (t)) .

This leads to the following definitions.

7.1.1. Definition. Let Q be a manifold and c = (c1,c2): I —*■ TQ a

differentiable parametrized curve; We define

Cl;(t) = (Cl(t), Cl'(t))

( = TCl(t,l)). ■

7.1.2. Definition. A parametrized curve c: I —*■ TQ is said to be

self-eonsistent if (Tq°c)’ = c. ■
66 7. PARTIAL DERIVATIVES

7.1.3. Example. Let c : R —*■ TR be given by c(£) = (t2, 21) for each

t € R. Here

(tr°c)(t) = t2

(tr°c):(t) = (t2, 2t)

and hence

( T R° C ) = C

so that the parametrized curve c is self-consistent. The effect of c on

Figure 7.1.3.

some elements of R is indicated in Figure 7.1.3 (c.f. Figure 7.1.2).

Identifying TR with R2 gives the diagram shown in Figure 7.1.4. ■

 7.2. TRADITIONAL NOTATION 67

EXERCISES 7.1.

7. Show that c : R —>- TR g-iv&n by c(t) = (sin(t), cos(£)) -ci a

■helfi-con-hthtent path -in TR and h ketch -Oth tAajectoay -in TR -in
the. home mannea ah FtguAe-h 7.1.3 and 7.1.4.

2. Show that c : R —»- TR2 g-ivzn by

c(t) = ((sin(t), cos(t)), (cos(t), -sin(t))

-i-h a he-ttf-conhthtent path -in TR2. Sketch an "ciaaoiw cUagaam" {,on.

thLh path -in R2.

7.2. TRADITIONAL NOTATION

The traditional notation for partial derivatives is often used in an

ambiguous way. We give a rigorous definition which is valid in the context

of partial differentiation with respect to the component functions of a

chart.

To this end let Q be a manifold of dimension k and let

f : Q —*■ R be differentiable. If (U, <j>) is a chart for Q then the

local representative f^ of f satisfies

f = f^otf)

as shown in Figure 7.2.1. Now fj maps an open set in Euclidean space

/ •
into R, hence the Definition 1.4.1 of f^ ^ is applicable.

Figyre 7.2.1. Local representative of real-valued f.

68 7. PARTIAL DERIVATIVES

7.2.1. Definition. Let (U, <t>) be a chart for the manifold Q and, in

terms of components, let

^ — (4*i » <t)2 » • • • > ‘f’feX

The partial derivative of f with respect to <(>£, denoted by

3f
: U —> R
3<t>£
is defined by putting

— = tJi
*
where fi = (f°<}> 1) is the local representative of f.

Notice that in the above definition all of the components of the chart

function <j> are involved. This is often emphasised in the literature by

writing

3f
as
3H
•»»‘t’i+i.•

A good account of the confusion which can arise when this point is over¬

looked is given in Munroe(1963) Chapter 5.

In the classical approach the "coordinate functions" <)>.£ were
3f
called "real-variables". We can still think of —— as being the
3H
derivative of f with respect to <j>y keeping the other <j>-,-,s fixed. The

traditional rules for calculating with partial derivatives, some of which

are contained in the exercises, are also valid in this newer context. The

following example illustrates this.

7.2.2. Example. Refer to Exercise 2.1.5 where the chart (U,(r,9)) for

Rz was defined with {(a,fc) £ R2 : a > 0} for U.

It is easy to show that

(r, 0) (U) = R+ x (- |, |)

(r,0)-1 = (idx.cosoid2, id1.sinoid2)

Now suppose f: R2 —► R : (a,b) t—> a2 + b2 . Thus f = idx2 + id22 (= r2 )

so on U from Definition 7.2.1. we have

|| = (f°(r,0)-1)/lo(r,0)

= ((idi2+id22)o(idx.cosoid2 ,idx.sinoid2))' o(r,0)

 . .
7 2 TRADITIONAL NOTATION 69

= (id12)/1o(r,e)

= 2id1»(r,e)

= 2r.

and ||- = (id12)/2°(r,0)

= 0. ■

It will be convenient to define the partial derivative of a

function g which maps into RW as a column vector.

7.2.3. Definition. Let Q be a manifold of dimension k with chart

(U, <}>) and let g be a £r (r ^ 1) map

g: Q — Rn •

We define the derivative of g with respect to 4>i as the column vector

(gi ° 4>—1) ^°<

3t-

(gno4»-1)/ro4»j

Notice that if h: R —* R is differentiable then h :R —>- R can be

thought of as the row vector of functions (hx ' ,h2',...,hn '). The

following lemma should be read with this in mind.

7.2.4. Lemma. (Semi-classical chain rule).

Suppose g: Q —>- R and h: R —> R are differentiable. If

(U,4>) is a chart for the manifold Q then

n /• 3g
9h°i = i h' h og lg_ .
3H 3<f>i Hi
j=i

Proof.

3h °g-(h
——2 _ /t_ o _ o
g 4)x-is/i
) by Definition 7.2.1.
3 4> i
n /■
= I (h/Jo g ° <p 1 (g °
J=1 3

by the chain rule. Theorem 1.4.5,

70 7. PARTIAL DERIVATIVES

A further generalization of Definition 7.2.1. which involves the partial

derivative of a real valued function with respect to all of a chart's

component functions is as follows:

7.2.5. Definition. Suppose Q is a manifold of dimension k and

f: Q —>- R is differentiable. For each chart (U, iJO for Q we define

the partial derivative of f with respeet to ip as the row vector

3f_ = /3f 3f \ m
3<p dip^J

We can now formulate a classical looking chain rule.

7.2.6. Theorem. (Classical chain rule). Let Q he a k-dimensional

manifold, and f: Q —R be differentiable. If (U, tf>) and (V, \p) are
charts for Q with U D V ^ , then

3f = d£_ m dip
3<Pi dip d(p£

Proof.

rrr “ (f ^o<t> by Definition 7.2.1

= ((foirVojM-1))7^
k
= I ( (folp~ 1)/°o (ip,*-1) by the chain rule for
\ 3 / partial derivatives
3=1
Theorem 1.4.5.
M ^

(fo^-1 )^'olp (ipo^1)^ o<p

II

3
j=i

k
3f_
= I by Definition 7.2.1
diPj
J=1

3f df m
3iji dQi

The following definition generalizes Definitions 7.2.1 and

7.2.3.

7.2.7. Definition. Let (U, <f>) and (V, ip) where U D V { □ be charts

for the fc-dimensional manifold Q. Then the derivative of i|j with respect
to <|> > denoted by , is the k x k matrix whose i-jth component is
 7.2. TRADITIONAL NOTATION 71

The above definition will be useful later as will be the following obvious
Corollary to Theorem 7.2.6.

8f = 3f 3^
7.2.8. Corollary.
3(J) d\Jj 3cJ)

The classical notion of a "differential" has a modern analogue

as shown by the following theorem.

7.2.9. Theorem. If (U, <J>) is a chart for Rn and. f : u —*- R is

differentiable, then
n 3f
Df" £ W ■

Proof. Let a € U.
Df (a) = D(fo cj)—1o (jj) (a)

= D(fo tf)k ) (<p(a))°Dtp(a) by Theorem 1.1.4.

nI r
= (fotj) 1) t(<p(a)) (D(p(a))^ by Theorem 1.4.3.
i= 1
n
I (g^ D<Pi) (a) by Definition 7.2.1.
^=l

The following version of the chain rule expresses what is

sometimes referred to in more traditional books as the rule for the "total
derivative" of a function.

7.2.10 Theorem. Let (U, <j>) be a chart for a k-dimensional manifold Q.

If Y : I —*■ Q is a differentiable parametrized curve and f : Q —>- R is

differentiable then

(fey)' = I (|:o Y) (4>7* 0 Y) ' .

3-1 ^ °

Proof. (f»y)' = (fo1 ° 4>°y) '

= Z (fo<j> 1)/’Jo<poY (<j>»y) ! by Theorem 1.4.5.

J=1
k gf
= I (|t7« Y)(4>7-oY)' by Definition 7.2.1. ■
j-1 3
 72 7. PARTIAL DERIVATIVES

EXERCISES 7.2.

In thete exehcitet (u,(J)) will denote a chant on. a manifold

Q and we let ((> = (<(>!» 4>2, ...

7. Show that {^oh i,j £ {1,2, ... ,fe)

r\ li i=j

0 .
2. Let f:Q —> R and g:Q —R fae dt^ehentixible.
Phove each othe. following £amiltah looking hulet:

5(f+g) 3f 3g 3(fg) _ f 9g , 3f .
9<t>i 34>i 3<J»i ’ 3^ 3<|>i 8 34)^

(A44u/ne -the cohhetponding hulet fioh (f+g)7^ and (fg)^).

3. Let (x,y) denote, the Identity chant {[on. R2, that it, let

(x,y) (a,b) = (a,d) eacA (a,L>) £ R2. Show that, fioh (u,(r,9))
ai tn Example 7.2.2,
3r _ x 3r _ y
9X /x2+y2 ’ 9y /x2+y2

39 _ -y 39 _ x
^ ~ x2+y2 * ~ x=4-y2

3(r,9) .
Hence white out the mathix exphet&lon £oh
9(x,y)

7.3 SPECIALIZATION TO TQ

The notation introduced in the previous section is applicable to functions

defined on any manifold hence, in particular, it is applicable to functions

defined on the tangent bundle TQ of a submanifold Q of Rn. In this

more specialized context additional results on partial differentiation can

be obtained. We assume throughout that Q has dimension k.

The dot notation is used ambiguously in the traditional books

on mechanics. Here we assign a rigorous meaning to one of its two possible

interpretations (and ignore the other in which it is regarded as

differentiation with respect to time). A key role will be played by the

idea of a self-consistent path, introduced in Section 7.1.

7.3.1. Definition. For a differentiable map f : Q —*■ R we define

f : TQ —> R by putting, for [y]a £ Tq,
 7.3. SPECIALIZATION TO TQ 73

f([y]a) = TT2oTf ([y]a) = (foY) '(0)

where tt2: R x R —»- R denotes projection onto the second factor. ■

The maps involved in the definition of f are illustrated in

Figure 7.3.1.

Figure 7.3.1. The f map

Now let (U, <J)) be a chart for a manifold Q. Since each

component cf). maps U into R we may apply the above definition to

obtain a map
: TU —* R

such that for [Y]a E TU

$i([Y]a) = (^-Y) '(0).

Hence, by Definition 5.2.1, we may write

T<J> = (<f>°tg, 4>) .

But since, by Definition 6.3.1, (TU, T«J>) is a chart on the tangent

bundle TQ, we may partially differentiate with respect to the components

of the map (4>oTq, <j>). To avoid repeated use of cumbersome symbolism, we

introduce the following definition.

7.3.2. Definition. Let g: TQ —*■ R be differentiable and let

(TU, (({)oTq, $)) be a natural chart on TQ. For !<-£<&, we define

R
 74 7. PARTIAL DERIVATIVES

to be the map - • ■
9(Vtq)

Note that the earlier Definition 7.2.1 is not applicable here because g

and <J) do not have a common domain.

7.3.3. Theorem. For f as in Definition 7.3.1 and <p as above,

f =

Proof. For e Tq,

f([y]a) = F2oTf([y]a) by Definition 7.3.1.

= TT2oT(fo4)-1o4>) ([Y]^)

= 7T2oT(fo(f) 1)oT0([Y]a) by Theorem 5.2.4.

= tt2oT(fo1) (<j>(a) » (4>°Y) '(0) ) by Definition 5.2.1.

= D(fo(f,“1)(<Ka))(((|,.Y) '(0))

= I (focj) 1)/ (<p(a))((<poy)'(O)). by Theorem 1.4.3.

i=1 *
k
= I (fo<t>“1)^(<})(a))(<j).oY)'(0)
i=l ^

by Definition 7.2.1."

7.3.4. Corollary. (Cancellation of dots rule)

3f 3f
o T,
Q •
3^. d<Pj
t '"rz

7.3.5. Corollary.
3f k dzf
32i
I r. ±- O T,
^i j=1

7.3.6. Theorem. (Interchange of dot and dash 1). If c : I —Tq is a

self-consistent path in Tq then, for f: Q —R differentiable,

f»C = (foTgoc) ' .

Proof. f(c(t)) = 1T2oTf(c(*)) by Definition 7.3.1.

= 7T2oTf (TxQoc(t,l)) by Definition 7.1.1.
= F2oT(foTpOC)(t,l) by the chain rule
= (foTqoc) '(£). ■
 7.3. SPECIALIZATION TO TQ 75

7.3.7. Corollary. For c: I TQ a self-consistent path, (U,4») a

chart for Q,
k 3f
(foTq°c) ' = i ° T °C (<p .°c) .
d

Proof. See exercises. ■

7.3.8. Corollary. (Interchange of dot and dash 2).

3f 3f
■o C
dip. 9^°tQoc

for each self-consistent path c : I —*■ TQ.

Proof. See exercises. ■

7.3.9. Theorem. Let Q be a submanifold of Rn and let (Tu, T<j>) be a

natural chart for IQ. If c : I —*■ TU is any differentiable curve (not
necessarily self-consistent) and f : Tq —R is differentiable then

(foe) ' = I (« c)(^.otqoc) ' + I (-^-« c)(i.o C) ' .

3-1 J=1 34>j J

Proof. This follows as a special case of Theorem 7.2.10, on use of

Definition 7.3.2. ■

The geometric interpretation of the next theorem is indicated

in Figure 7.3.2. for the special case of a two-dimensional submanifold of

R3. The theorem has a traditional counterpart which is usually considered

to be geometrically obvious.

Figure 7.3.2.
76 7. PARTIAL DERIVATIVES

7.3.10. Theorem. If Q is a submanifold of Rn of dimension k3 X the

Yl
identity map on R and (U, q) a chart for Q at each a £ Q, then

(»• % <“>)6 TaQ

for 1 < i < k.

Proof. Define y : I —*■ U by putting

Y(t) = q_1(q(a) + t e-i)

with I an open interval chosen so that 0 £ I and y(I) c u. Hence

[y]a £ TaQ. Since, moreover, y maps I into Rn and q-1 maps an

open set in R^ into Rn we get

Y' (0) = (q-1)^ (q(a))

= (X.q-1)/^°q(a)

EXERCISES 7.3.
1. Pnove Conollany 7.3.7.
You will need to use Theonem 1.1.6. and 7.3.3.

2. Pnove Cofwllan.Lj 7.3.8.

3. Note that the. chant q = (0, r-_l) as defined in Example 6.1.4, ii

a chant fon R2 with the submanifold pnopenty fon

Q = {(a,b) : a2 + b2 = 1}.

Venify by dined calculation that Theonem 7.3.10. holds in this

special case.

4. Complde the pnoof of Theonem 7.3.9.

7.4. HOMOGENEOUS FUNCTIONS

This topic is traditionally discussed in the context of functions which

map R to R as, for example, in Courant & John(1974). For

applications to mechanics, however, it is more useful to define homogeneity

for functions having the tangent bundle of a manifold as domain. We

suppose Q is a submanifold of Rn of dimension k.

 7.4. HOMOGENEOUS FUNCTIONS 77

7.4.1. Definition. A map g: TQ —»- R is said to be homogeneous of degree

r > l if, for all (a,v) € TQ and t € R

g(a, tv) = tr g(a,y). ■

For example, the map g: TR2 —>- R with

g(x,y, h,k) = cos(x)h2 + 2hk + sin(y)k2

is homogeneous of degree 2.

The following theorem extends the familiar result known as

"Euler's relation for homogeneous functions" to the new context.

7.4.2. Theorem . (Euler's relation). If g: TQ —>- R is homogeneous of

degvee v and (Tu, T<|>) is a natural chart for TQ then

^ . gg

i= 1 d<t>i
Proof. Let (a,h) E TQ and define f: R —*■ TQ by putting

f (t) = g (a,th).

Next define a curve c: R —>- TQ by putting

c(t) = (a,th).

Thus f = g°c and so we may apply Theorem 7.3.9, to get

k - k ^

f'-Z Orjjj- ° c) (<t>j o Tq O c) ' + I (-1- O c) (<L o c) ' .

j=l J Q j= 1 J

But o Tq o c is a constant function while <f>^. is linear on the tangent

space T„Q, hence (<f>7* o c) ' = <j> • o (a,h) . Thus

f ' = I (^4- o c)i. « (a,?z). (1)

.7=1 3*,- J ~ "

Now by homogeneity, f(t) = trg(a,h) so that

f'(1) = rg(a,h) (2)

Evaluating (1) at 1 and comparing with (2) shows that

k
*. = g
J=1 3*.
at each point (a,h) in the common domain of these two functions. ■
 8. DERIVING LAGRANGE’S EQUATIONS

In this chapter we begin our account of classical mechanics.

The physics background required is minimaZ. We wiZZ assume that the reader
is famiZiar with the concepts of velocity, acceZeration, mass and force at
about the ZeveZ of high schooZ physics.

The resuZts on partiaZ derivatives estabZished in the previous

chapter wiZZ be used to derive Lagrange's equations for particZe motion
directZy from Newton's Zaws. Our derivation foZZows the same generaZ pZan
as the traditionaZ one, which may be found in books such as GoZdstein(1980),
Synge & Griffith(1959) and Whittaker(1952). The use of manifolds,
however, enables us to give a more geometrical slant to this topic in that
we use ideas such as 'orthogonality with respect to a tangent space ' in
place of 'infinitesimal displacements' and 'virtual work'.
To give our calculations a conventional Zook we will use
X = (x,y,z) for the identity chart on R3 and (X ° Tq, X) for the
naturally induced chart on TR3.

8.1. LAGRANGE'S EQUATIONS FOR FREE-FALL

To provide some physical motivation we first derive Lagrange's equations

for the very simple mechanical system consisting of a particle falling

freely under gravity. Regard the force F acting on a particle of mass m

and the acceleration a which it produces as vectors. Newton's laws then
imply that
F = ma

provided measurements are made relative to an 'inertial frame of reference'.

From the study of motion near the earth's surface it is often useful to

suppose that axes fixed relative to the earth provide such a frame of

reference; for the study of planetary motion, however, it would be more

appropriate to fix the axes relative to the distant stars.

 8.1. FREE-FALL 79

Consider now a particle of mass m moving near the surface of

the earth and choose a cartesian coordinate system fixed relative to the

earth with the x and y axes horizontal and the z axis vertical. We

will assume that the particle is acted on by a constant force of magnitude

mg vertically downwards. See Figure 8.1.1.

mg

Figure 8.1.1. Motion under gravity.

It can be shown that the position and velocity of the particle

at any instant determine its subsequent motion uniquely. Hence it is

natural to represent its state at time t by an element c(t) £ TR3 with

c: I —*■ TR3 a self-consistent path, which is defined on some interval I.

The cartesian coordinates of the particle at time t are

(X»Tp3oc)(t) and so Newton's laws of motion imply that

m(x°Tn3°c)"(i) = 0

m(yoTR3oc)"(£) = 0 (1)

m(zoTp3<>c) "(t) = -mg.

Dynamically significant quantities associated with the motion of the

particle are the kinetic and potential energies, given by the formulas

T = Jgw(x2 + y2 + z2)

V = mgz .

We regard T and V as maps from TR3 and R3, respectively, into R

and then define the Lagrangian function L : TR3 —*■ R as
 80 8. DERIVING LAGRANGE'S EQUATIONS

V o T
R3

where t^3 : TR3 —*■ R3 is the natural projection. The equations (1) may

be written in terms of L as

Ik o c ' _ 9L o

o
o
ii
3x , 3x
r
o c = 0 (2)
,3y 8y °
• r
3L
C c = 0 .
3z 3z

The equations (2) are Lagrange's equations for the motion of

the free-falling particle. They have an advantage over the Newtonian

equations of motion (1) in that their form is preserved under change of

coordinates, as will become apparent in the next section. The gravitational

force in the example above is the simplest example of a conservative field

of force, as defined below.

8.1.1. Definition. If the force acting on a particle has the form

w x / /aV 3V 3V,. .
F(a) =K-{jp 97* £ T R3
a
at each a £ U, an open subset of R3, for some differentiable function

V : U —► R we say that the map F : U —*■ TR3 is a conservative field of

force. ■
YL
This definition has an obvious generalization to R . Further

examples of conservative fields of force are given in the exercises.

EXERCISES 8.1.

1. VeJiioe Lagaange’* equation* (2) hofl the fanee-putting paAticle

theia Newtonian hoAm (1). Vou will hxtti need to apply theorem
7.3.6 to the leht hand Aide* oh the equation* (1).

2. Show that the gAavitational {ieid oh ^(Vme de*ctibed in the text i*

comeAvative.

3. Show that the hoxee hi^-d defined on R1 by F(a) = (a, -ka)

i* conteAuative. (Thi* fix-eid oh h0>xoe aAi*e*, h°;1 k > 0, when

Hooke’* law i* u*ed to model the motion oh a paAticle attached to
an elastic Apning and lend* to the Atudy oh Aimple haAmonic motion.)
 8.1. SINGLE PARTICLES 81

4. Show that the. Notice, fiield de^tned on R3\{0} by

(/a2+bz+a2)3

it conteAvative. [Thit ic.e &ieZd eoAAetpondt to the gAaoitatonal

attAaetton pAodueed by a paAticle o& unit matt at the OAlgtn).

8.2. LAGRANGE'S EQUATIONS FOR A SINGLE PARTICLE

Lagrange's equations will now be derived for the motion of a particle of

mass m which is constrained to move on a submanifold Q of R3, the

dimension of Q being k where 1 < k < 3. Again we take X = (x,y,z)

to be the identity chart on R3.

It will be assumed, furthermore, that the particle is subject

to a conservative field of force together with a reaction force which acts

orthogonally to the manifold at each point - as would be the case if there

were no friction between the particle and the manifold. In mathematical

terms this means that the force acting on the particle at each point a £ Q
is a sum

F(a) + R([y]a) = (F°tq + R)([y]a) £ TR3

where F(a) and R([y] ) have the following forms:

(1)

for some function V : U —*■ R where U is open, while the reaction force

R([y]Q) is orthogonal to TaQ (2)

in the sense of Definition 8.2.1. below.

Figure 8.2.1.
82 8. DERIVING LAGRANGE'S EQUATIONS

8.2.1. Definition. A vector (a,v) £ TR1 is orthogonal to the submanifold

TQ of TR at a £ Q if the dot product V.h = 0 for every vector

(a,h) £ TaQ. ■

At time t the position of the particle will be Tq°c(t) and

its state will be c(t) for some self-consistent path c : I —*■ TQ, where

I is an interval. Relative to the cartesian coordinate system defined by

the x,y and z axes the position and velocity of the particle will be

X°Tqoc(£) and X°c(t) respectively. Hence Newton's laws imply that

m(XoTQoc)"(£) = tt2o(F°tq + R) o (X°Tq,X) q c(t)

= TT2o(FoTq + R) o c(t)

where tt2: R3 x R3 —>- R3 denotes projection onto the second factor. Hence
by Theorem 7.3.6.

m(X°c) ’ (t) = tt2o(FoTq + R) ° c(t). (3)

To convert (3) to Lagrangian form we first introduce the

kinetic energy
T = isrn(x2 + y2 + z2)

= %mX2

where X2 denotes X*X. After restricting T to Tq as domain, we

introduce the Lagrangian function L : Tq —► R by putting

L = T - V°Tq (4)

8.2.2. Theorem. (Lagrange's equations). Under the above assumptions the

curve c : I —*■ Tq satisfies the equations

r 3l -° c
5L
-oC 0 (i-i.k)
3q,'

for each chart (Tu, ((qi,...,qfc) o tq, (4x.qfc))) for TQ.

Proof. Partial differentiation of the kinetic energy T with respect to

each of q■ an<^ ^i on use °f Lemma 7.2.4, the semi-classical
chain rule.
9T 9X
= mX (5)
3q. 3q,-
9T 9X 9X
= mX = mX • (6)
9q.
9qi
 8.3. SEVERAL PARTICLES 83

where the last equation follows from Corollary 7.3.4, the cancellation of

dots rule.

Differentiation of (6) along the curve c : I —>■ TQ now gives

3T ' r . 3X ax n
o c = m^(X°c) ' • ° In ° c + (X°c)1 ° XQ 0 C

9q,' 3q . J
ax ax
= mj^ (X°c) ' • o Tq 0 c + (X°c) o c
3q.
3q;
by Corollary 7.3.8, the interchange of dot and dash rule. Hence

subtraction of (5) composed with c gives

8T 3T
o c o c
3q.

t .
ax
m(X ° c) ° Tn 0 c
3q,-
ax ° Tq 0 c by Newton (3)
= TT2o(F oTq+R) “c
3q.'
ax o Tq o c by (2) and Theorem 7.3.10.
7T o o
aq.

av ax
=Tq°c by (1)
3X 3q .

av Tq o c by Theorem 7.2.6.
3q •
1
3T 8T_ av
Thus ° TQoC*
3q .
v 3qi 3qi

Lagrange's equations now follow from (4) since

3(V.tq) 3V _ 3 (VoTq) __
—— — o tq and U»

9q; 3qi

8.3. LAGRANGE'S EQUATIONS FOR SEVERAL PARTICLES

Lagrange's equations will now be set up for a system of n particles each

of which moves in R3 subject to certain constraints. (One obvious

constraint, for example, is that two particles may not simultaneously occupy

the same position.) Here only those constraints will be considered which

place restrictions upon the positions which the particle may occupy - and

not those which place constraints upon their velocities, for example.
84 8. DERIVING LAGRANGE'S EQUATIONS

8.3.1. Definition. The configuration set of the system of particles will

be defined as the set of all elements

a ~ *^i^n^ ^ R

such that for i = 1,...,n the particle may occupy the position

E R3, in conformity with the constraints. ■

In this way, instead of considering n particles moving in R3

we consider one particle moving in R3?1. Only those systems will be

considered in which the configuration set can be made into a submanifold

Q of R3n .

8.3.2. Definition. If the configuration set of a system of n particles

is a submanifold Q of R3n then Q will be called the configuration

manifold and the tangent bundle TQ will be called the velocity phase

space of the system. ■

Let the identity chart of R3n be denoted by

X = (xi’yi’zi’.’Vyn*zn}-
The history of the particles will then be described by a curve c: I —*■ TQ

such that the position of the iLn particle at time t is

(x^,y^,z^)°Tq°c(£) and its velocity is (x^,y^,z^)oc(t). We now generalize

some of the ideas of Section 8.1.

8.3.3. Definition. The field of force for a system of n particles in

R3 is the map F : U —*■ R3n with

F (&) (a , TT 2 o F1 ((2^ ^ 2 ° F^j (Oy[ yhyi yDyi) )

where U c R3n is open and F^ : U —>■ TR3 is the field of force acting on

the i*"*1 particle. The field of force F is said to be conservative if

there exists a potential function V : U —*■ R such that

(9V 9V 9V 9V 9V 9V_\
F(a) = (a, (a)). ■
\3Xl* 9yi’ 9zx.
3V V 3z J

8.3.4. Definition. Let be the respective masses of the n

particles of a system with configuration manifold Q c R3n. The kinetic

energy of the system is the map T : TR3?1 —*- R given by

n
T = h I m.% (x.'
v
+ z,2).
i-l
 8.3. SEVERAL PARTICLES 85

The above formula for the kinetic energy can be written more

succinctly as a matrix product

T = %XT MX

where M is the diagonal matrix given in terms of the masses by

M = diag(m1,m1,m1,m1,m2,m2,.,mn,mn,mn)
•p
and where X is the transpose of the column matrix X.

We are now able to define the Lagrangian function of the

system of particles.

8.3.5. Definition. Restricting the maps T and V in Definitions 8.3.4.

and 8.3.3. to TQ we define the Lagrangian function L : TQ —>- R by

putting

L = T - VoTq. ■

8.3.6. Theorem. Consider a system of n particles in R3 which are

constrained to move in such a way that the equivalent single particle moves
on a submanifold Q of R3n of dimension k. Let the field of force
acting on this particle be the sum of a conservative field of force and a
reaction which is orthogonal to TQ at each point. Let L : TQ —*■ R be
the Lagrangian function of the system. For each chart

(TUj ((q^ > • • • >qjf) °^q> (q^»• • •»q^)))

for TQ, each self-consistent curve c: I —► TQ describing the history

of the particle satisfies the equations

/8L
(- ° c V 9L
j - - n
o c = 0 n < I <
'9q. ' 9qi

Proof. The proof of this theorem is very similar to that of Theorem 8.2.2.

Formally, one just replaces the real number " m " occurring in the proof

of Theorem 8.2.2. by the diagonal matrix " M " . The validity of the

resulting steps in the proof is established in the exercises. ■

EXERCISES 8.3.

The notation uAed In theAe zxeACAAeA Ia explained above In the

text.

1. Chech that the kinetic energy oh the AyAtrn, oa given in Vehinttion

8.3.4, can be coAttten in the tfcDun cZcumed above,
 86 8. DERIVING LAGRANGE'S EQUATIONS

2. Chuck that Mewton'A law fan. the single panticle equivalent to the
AyAtem can be wnitten in the fanm

M(XoC) ' (t) = tt2 o(F°Tq + R)°c(£).

3. LUe the Aemi-claAAical chain nuie, Lemma 7.2.4, to Ahow that fan
1 < i < k,

, * 3T *T w 3X
(a) - = X M -
3qi

(b) T,
Q*

Show that (Toe) ' = (X°c) ,T M(x o c)

5. UAe the neAultA o^ ExenciAe 2 and 3 to give a pnoofi ofi Theonem 8.3.6.

8.4. MOTION OF A RIGID ROD

In the previous section we derived Lagrange's equations for a conservative

system for which the resulting single particle was constrained to move on

a submanifold Q of R3W. We now show how it is possible to fit a

prototype rigid body into this scheme.

It will consist of two particles constrained to move in such a

way that the distance between them remains constant. The reaction forces

between the two particles are assumed to be equal in magnitude but opposite

in direction and to act along the line joining the two particles, as in

Figure 8.4.1. This system is thus a mathematical model of a pair of

particles joined by a rigid rod of negligible mass.

(n2>b2>02)

(a 1,bi ,c1)

Figure 8.4.1. A rigid rod.

The configuration set for this system is a set of points of the

form
a = (ai, blt o1, a2, b2, c2) e R6

where (a1,2?1,c1) and (a2Jjb2,e2) are the positions of the two particles
 8.4. RIGID ROD 87

The constraint that the two particles lie at opposite ends of a

rigid rod of length d > 0, can then be expressed by the condition that

f(a) = 0

where f: R6 —>■ R with

f(a) = (a1-a2)2 + (b1-b2)2 + (c^-Cj)2 - d2 (1)

We now show that the configuration set is a submanifold of R6.

Let a £ R6 be as above and let h £ R6. By (1) the Frechet derivative of

f is given by the dot product

Df(tO(/z) — <22 , b i~b 2 » b2—bXi <22—<22) (2)

Clearly Df(cz) has full rank 1 and so Theorem 3.3.1, shows that the

configuration set f-1(0) is a submanifold of R6, of dimension 5.

The following proposition states a somewhat surprising fact,

which is crucial for the valid application of Theorem 8.3.6.

8.4.6. Proposition. The reaction force due to the rigid rod connecting
two particles is orthogonal to the configuration manifold at each point.

Proof. Let Q denote the configuration manifold f-1(0), with f given

by (1) above, and let (a,h) £ TaQ. By Theorem 5.3.1,

Df(a) (h) = 0

which by (2), is equivalent to

2(a1-a2, b1-bZ3 cx-c23 az-al3 b2-bl3 Cz-ox) . h = 0. (3)

Since the reaction forces on the two particles dct along the

rigid rod and are equal in magnitude but opposite in direction, we may

write their directions as

\(ax-a23 bx-b23 cx-c2) and -\{ax-aZ3 bx-b23 cx-c2)

respectively. These two reaction forces are combined to give a single

reaction force in TaR6

R = (a3 \(.ax-a2, bx-b23 ox-c23 a2-aX3 bz-bl3 c2-ox))

and so, by (3), B..(a,h) = 0. Thus the reaction R at the point a is

orthogonal to TaQ.
If now we are given a potential function V : R6 —> R for the

pair of particles and a submanifold chart (U,(qx,q2,q3,q4,q5)), for Q

we may apply Theorem 8.3.6, to obtain Lagrange's equations for the system.
88 8. DERIVING LAGRANGE'S EQUATIONS

EXERCISE 8.4.
Find a submanifold chaAt Suitable foA the configuration manifold
[Hint: fix one. end of the fwd and then find a submanifold chaAt
foA S2 <= R3.)

8.5. CONSERVATION OF ENERGY

Physical systems are subject to the law of conservation of energy. For

systems of the type discussed in Section 8.3, it will be shown from

Newton's laws that the total mechanical energy - kinetic energy plus

potential energy - remains constant with the lapse of time. From a

physical viewpoint, these systems conserve their mechanical energy because

they are free from the effects of such forces as friction and air-

resistance, which dissipate mechanical energy into other forms of energy

such as heat.

Returning to the situation described in Section 8.3, we

consider a single particle moving on a submanifold Q of R3n subject to

a conservative field of force together with a reaction force orthogonal to

TQ at each point. As before, we let V : Q —*■ R be a potential function

for, and T : TQ —*- R be the kinetic energy of, the particle. The total

energy function E : TQ —► R for the system is defined by putting

E = T + V°Tq .

On the basis of Newton's laws, the following theorem can now be proved.

8.5.1. Theorem. (Conservation of energy). If c: I —*- TQ is a self-

consistent curve describing the history of the above particle then E « c

is a constant function.

Proof. The notation being as in Section 8.3,

fav
(V°Tq°c) ' = o Tn 0 C (X°c) by Corollary 7.3.7.
ax " LQ

tt2o F o Tn O c| . (X°c) by Definition 8.3.3.

= — (tt2 ° (F°tq + R)oc)*(X°c)

where the last step is valid since, by Theorem 7.3.6, (X°c) equals

(X°Tq°c)', which lies in TQ and hence is orthogonal to the reaction R

at each point. Note also that

 8.5. CONSERVATION OF ENERGY 89

(Toe) ' = (X 0 c)'^M (X 0 c) by Exercise 8.3.4.

= 2
(tt °(F°Tq + R)oc)'(X°c) by Exercise 8.3.2.

Thus (E°c) ' = _0 in the interval I and E°c is a constant. ■

The above proof is a rigorous version of one of the two proofs

given in Whittaker(1952) section 41. In the alternative approach Whittaker

first of all proves a conservation law for a system of Lagrangian equations

defined in terms of an arbitrary Lagrangian function L and then

specializes the result to the case where L may be expressed in terms of

kinetic and potential energies. In the exercises, the reader is invited

to follow through the details of the alternative approach by way of

comparison.

EXERCISES 8.5.

In tliwa exenctAeA (u, q) denotes a chant ion. a k-dimenAZonal

AubmanZioZd Q oi R3n.

7. Let L : TQ —*- R be any diiieAe.ntia.bZe. iunction and AuppoAe that

c : I —>- Tu Za a AeZi~conAZAtent path AatiAiyZng La.gn.ange.'A equationA

(2k. . c\ 2k. c = 0 ion. 1 < i < 7c.

W ) 3,.
Show, by uAZng TheonemA 7.3.9. and 7.3.6, that the derivative oi the

iunction ^
9L
-LoC
(V i=l
1 qi
9q.1

Za zero on I and hence that thtA iunction Za a conAtant.

2. Let T: Tq —* R and V : Q —► R be dZHenentZabte iunctionA and

AuppoAe that T Za homogeneouA oi degnee 2 and let

L = T - V°Tq .

(a) Show, inom Theonm 7.4.2, that

X q . —-L = T + V o Tq
i=1 t 9q•
n

(b) Hence pnove Theonem 8.5.1.

 9. FORM OF LAGRANGE’S EQUATIONS

This chapter begins with a simple example of a mechanical

system. By explicit calculations, the Lagrange equations for this system
can be proved equivalent to a system of second-order differential equations.
For arbitrary mechanical systems of the type studied in Chapter
8, a similar fact can be proved by using standard results about quadratic
forms. In subsequent chapters the theory of differential equations will be
used to show how Lagrange 's equations determine the motion of mechanical
systems.

9.1. MOTION ON A PARABOLOID

This section investigates the explicit form of Lagrange's equations for a

mechanical system consisting of a particle of mass m moving under

gravity on a paraboloid in the absence of friction. The paraboloid is

assumed to be the subset Q of R3 consisting of all the points on which

the equation

z = 4(x2 + y2) (1)

holds, where (x, y, z) is the identity map on R3. Thus Q is a

paraboloid of circular cross-section with its axis vertical, as illustrated

in Figure 9.1.1.

A chart for R3 with the submanifold property for Q is

(R3. (x, y, z -^(x2 + y2))).

Hence Q is a submanifold of R3 and has (Q, (x,y)) as a chart.

Incidentally, this chart forms a one-chart atlas for Q and so Lagrange's

equations with respect to this chart apply to the whole of the tangent
bundle TQ.
 9.1. MOTION ON A PARABOLOID 91

Figure 9.1.1. A paraboloid Q.

With x,y and z restricted to Q it follows from (1) and

Theorem 7.3.3. that

z = (x°tq)x + (yoTq)y .

Hence, on temporarily writing x,y in place of xoTq and yoxq for

simplicity, we find that the kinetic energy is given by

T = V?(x2 + y2 + (xx + yy)2).

Since the potential function, moreover, is simply V = mg z it follows

that the Lagrangian function is given by

L = %rn ((l+x2)x2 + 2xy xy + (l+y2)y2) -%mg(x2+y2) . (2)

Now let c: I —► TQ be a self-consistent curve describing the

history of the particle. We assume c is of class C2 and introduce the

notation

X = XOTqOC, y = yoXqOC (3)

for the local representative ^Tq°c^(x y) t'le curve Tq°c respect

to the chart (Q, (x,y)). From the interchange of dot and dash rule,

Theorem 7.3.6, it follows that

x' = x»c, y ' = yoc. (4)

An application of Theorem 8.2.2. shows that Lagrange's

equations are satisfied by the curve c and from (2), (3) and (4) it
92 9. FORM OF LAGRANGE'S EQUATIONS

follows that these equations can be written in terms of the local

representatives as

(1 + x2)x" + xyy" + x(g + x'2 + y'2) = 0

(5)
xy x" + (1 + y2)y" + y(g + x'2 + y'2) = 0

These equations can be solved to give the acceleration x" and y"

explicitly:

x" = -x[g + (x')2 + (y')2]/[ 1 + x2 + y2]

(6)
y" = -y[g + (x')2 + (y')2]/[i + x2 + y2].

Thus Lagrange's equations lead to a pair of second-order

differential equations for the local representative of the curve c.

Conversely, the steps can be reversed to show that these equations imply

Lagrange's equations for the system.

EXERCISES 9.1.

7. Fon the problem dircussed tn the text concerning motion on the

paraboloid Q:

Show that the kinetic energy is given by

x
T + VKx y)a°Tq
.y.
for a Suitable matrix-valued. function A Q R2*2 with
pointwise inverse A-1 : Q —>- R2X2.

(fa) Show that Lagrange's equations (5) for the paraboloid can be written
as:
X o'
(A°Xq°c) + (.g + x'2 + y'2)
y"j .y. 0

and then deduce the differential equations (6) with the aid of k -l

2. Show that the system of second-order differential equations (6),

describing motion on the paraboloid, admits the family of periodic
solutions given by

x(t) = a cos(/g t), y(t) = a sin(/g t)

for each a £ R. Note that the period is constant along the family.
Use the relations 13) in the text to help you sketch the set of
points
 9.2. QUADRATIC FORMS 93

{tq°c(£): t £ R} c Q

which. 1i tnaced out by the pantlcle on Q.

3. [The iphenlcal pendulum). Comlden a pantlcle o& man m moving

undeA gnavlty on the unit iphene S2 c R3 In the absence otf rfnlctlon
Let (u, (6,<j>)) be the sphenlcal-polan chant fion s2 defined In
Exendie 2.1.6.

[a] Show that the kinetic and potential enehglei may be expn.ei.ied In thli
chant at
T = hqn(Q2 + sin2o0 <fi2)

V = mg cos°0.

(fa) Denary that, with 0 = 0oTq°c and cj> = <})oTq°c, Lagnange'i equations
one expllclty

m(^9"-(sin°-6) (coso0)<jj'2 - g sin<>0^ = 0

(jn (sin2o0)$r)' = 0.

(c) Deduce that Lagnange'i equatlom one equivalent to a iyitem oft

second-onden. dl^enentlal equatlom In 0 and

[d) Show that while the pantlcle itayi within the chant domain u,
(sin2o0)<J)' nemalm comtant.

9.2. QUADRATIC FORMS

The discussion of Lagrange's equations in the next section will use the

basic ideas about quadratic forms which are summarized below. A fuller

account of this topic is contained in standard linear algebra texts such

as Nering(1964). Our definition of the matrix of a quadratic form differs

from the usual one, however, and is more natural in the context of
manifolds.

A quadratic form on a vector space V over R is a mapping

g: V —*- R whose values are given by

g(y) = f(y,u)

for some bilinear map f: VxV —>- R. If g is a quadratic form then there

is only one such bilinear form which is symmetric and we call it the

bilinear form generating g.

A quadratic form g is said to be positive definite if

g(y) > 0 for all y ^ 0. It is clear that the restriction of a positive

94 9. FORM OF LAGRANGE'S EQUATIONS

definite quadratic form to a vector subspace of its domain is also a

positive definite quadratic form.

9.2.1. Example. We claim that the map g: R2 —* R with values given by

g (a,£0 = a2 + 6ab + 10b2

is a quadratic form and is positive definite. To see this it is helpful

to use matrices and write

'l 3 ’ a
g(a,2>) = (a b)
3 10 y b,

We construct the bilinear map f generating g by putting

1 3 ' c
f((a,2>). (c,d)) = (a b)
3 10 y dy

Hence g is a quadratic form. Completing the square shows that

g(a,6) = (a + 3b)2 + b2 > 0

and this vanishes only if (a,b) = (0,0). Hence g is positive

definite. ■

The representation of quadratic forms by matrices will play an

A
important role in the next section. The relevant definitions and theorems

will be stated first for the special case of quadratic forms on Rn. The

usual basis for Rn will be denoted by (e1,... .e,,) .

9.2.2. Definition. The matrix of a quadratic form g on R?i has as its

i-jth element

f (e-f.ej)

where f is the bilinear function generating g. ■

9.2.3. Theorem. If g is a quadratic form on Rn with matrix A then

for each v 6 Rn, regarded as a column matrix,
T
g(y) = v Ay.

Conversely if this relation holds for all v then g is a quadratic form

and A, if symmetric, is the matrix of g. ■

More generally, we now consider quadratic forms on an arbitrary

vector space V of dimension n over R. There is then a vector-space

isomorphism 4>: V —► Rn, which we may regard as a chart for V.

 9.3. LAGRANGE'S EQUATIONS SECOND-ORDER 95

9.2.4. Theorem. If g: V —*- R is a quadratic form on V then its

local representative g^ (shown in Figure 9.2.1.) is a quadratic form
on Rn. If A denotes the matrix of g^ then the values of g are given
by
g(u) = <J>(u)T Acf>(y)

where we regard <j>(y) as a column matrix. Conversely if the values of g

are given by this formula then g is a quadratic form on V. ■

Figure 9.2.1. Local representative of a quadratic form.

The key theorem for our purposes is as follows:

9.2.5. Theorem. If a quadratic form g: V —>• R is positive definite

then the matrix of each local representative is non-singular. ■

9.3. LAGRANGE'S EQUATIONS ARE SECOND-ORDER.

It will now be shown how Lagrange's equations for a system of the type

discussed in Section 8.3 can be written as a system of second-order

differential equations. The kinetic energy of the system of particles was

there defined in terms of the identity chart (X°x, X) on TR3^ as

T = h xtm X

where M = diagC^, mlt m1, m2, mz, m2,.mn,mn,mn) with each mass

m^ > 0. Thus for each a € R3n the restriction of T to TQR3?1 is a

positive-definite quadratic form.

96 9. FORM OF LAGRANGE'S EQUATIONS

Now suppose that the particle representing the system is

constrained to move on a smooth fc-dimensional submanifold Q of R3n.

For each a £ Q the restriction of T to TaQ is the restriction of

a positive-definite quadratic form to a vector subspace of its domain,

hence it is again a positive-definite quadratic form. Suppose furthermore,

that (q°Tq, q) is a chart for TQ so that q|TaQ is a linear chart on

TaQ. It then follows from Theorem 9.2.4. that the kinetic energy for the

particle moving on Q is given by

.t
T = %q AoTq q (1)

kx-k
for some matrix-valued function A:Q —*■ R . A simple example

illustrating this is contained in Exercise 9.1.1.

To work out Lagrange's equations explicitly, we will need the

smoothness of the matrix valued function A occurring in (1).

9.3.1. Lemma. The i-jth entry a^ of the matrix-valued function A

in (1) above is given by

32T
aijoTQ
9q3q
^ 3

for each i}j = 1,2,...,k.

Proof. Notice that equation (1), written out in full, is

k k

1 ■ \z, E, a«"°T« ’i ^ •
£=lm=l

Differentiating this expression gives the required result. ■

9.3.2. Corollary. The matrix-valued function A is smooth and

^21 o = Aolqoc (qoTqoc)" + (A°Tq°c) '(qoTqoc) '

for each smooth self-consistent path c: I —*- TQ. ■

We can now prove the main result of this chapter.

9.3.3. Theorem. Lagrange's equations are a system of second-order

differential equations in the local representative q = q « » c .

Proof. We may suppose the Lagrangian function L : TQ —>- R is given by

•T *
L = hq A°Tq q - VoTq

where A is as in Lemma 9.3.1. By Theorem 8.3.6. Lagrange's equations in

a chart (q°Tq, q) become

 9.3. LAGRANGE'S EQUATIONS SECOND-ORDER 97

8
— /l‘T.
(hq AoTq q)°cj
V -
8L ° c = 0
A

where c: I —*■ TQ is a smooth self-consistent path.

Application of Corollary 9.3.2. then gives

(AoTqoc) (qoTgoc)" + (A°Tq°c)' (q°Tqoc)' - ° c = 0.

By Theorem 9.2.5. the matrix-valued function A°Tq°c is

pointwise invertible so the above equation becomes

(qclpoc)" = (AoTqoc) 1 (-(AoTqoC) '(qoTqoC) '

as required.

EXERCISES 9.3.

1 TUJL in the detatt6 o& the. pnooi o{ Lemma 9.3.1.

2. Compare the content* oft Section 9.3 with Section* 27 and 28 o{,
WhittakeA{1 952).
 10. VECTORFIELDS

In modem terminology, the right-hand side of a differential

equation determines a "vectorfield", which means assigning an "arrow" to
each point of the domain, while the solutions of the differential equation
core referred to as "integral curves", which cere tangential to the arrows.
In the context of manifolds, the integral curves are regarded as maps into
the manifold while the "arrows" are elements of its tangent spaces. The
modem terminology thus invites us to think of differential equations in a
very geometrical way. In this chapter, enough of the theory of vector-
fields will be developed to enable Lagrange's and Newton's equations to be
put into proper geometric and analytic perspective.
Among the most basic results in the theory of differential
equations are those concerning the eocistence and uniqueness of solutions
and these results translate easily into corresponding results about
integral curves. When studying vectorfields and their integral curves
on a manifold, however, it is important to distinguish between "local" and
"global" propertiesj the former refer to what happens in a particular
chart domain whereas the latter refer to what happens on the whole
manifo Id.
Finally, some special properties of vectorfields which arise
from second-order differential equations, such as Lagrange 's equations,
will be studied.

10.1. BASIC IDEAS

In this section we introduce the ideas of vectorfields and integral curves

and give a vectorfield version of the classical existence and uniqueness

theorem for the solutions of differential equations.

A vectorfield may be thought of as a mapping from a manifold

M to its tangent space TM which preserves the base points. The formal

definition is as follows.
 10. BASIC IDEAS 99

10.1.1. Definition. Let M be a smooth manifold. A mapping

Y : M —*■ TM

with the property o Y = id^ is called a vectorfield on M. ■

Figure 10.1.1. A vectorfield on M.

10.1.2. Example. The mapping

Y : R2 —* TR2 given by

YGv,y) = (-z/,x))

is a vectorfield on R2 since x^2 ° Y = id^2. The assignment of the

various arrows to points in R2 is illustrated in Figure 10.1.2. ■

Figure 10.1.2.
100 10. VECTORFIELDS

Before reading on, recall (see Definition 7.1.1.) that if c: I —► M is

differentiable (I an interval, M a manifold), then we define

c;(t) = Tc(t,1) = (c(*), c'(i)).

The manifold version of a solution to a differential equation is given by

the following definition.

10.1.3. Definition. Let M be a smooth manifold and suppose Y : M —*• TM

is a vectorfield on M. An integral curve of Y at a £ M is a mapping

c: I —*■ M where I is an interval containing 0 such that c(0) -a

and

cJ(t) = Y(c(£))

for each £ € I. ■

The relationship between a differential equation on R

written as

c'(*) = f(c(t)) (1)

for some function f : R —*■ R and its vectorfield counterpart is then

given by including the "base point" as follows:

c;(t) = (c(t), c'(£)) = (c(t), f(c<*))).

Thus the vectorfield for the differential equation (1) is then given by

Y : R^ —+■ TR^ where

Y(x) = (x, f(x)) for each x £ R^.

10.1.4. Example. Let Y : R —* TR be the vectorfield defined by

Y(x) = (x,3x) for each x £ R.

Thus c : I —*■ R is an integral curve at a £ R for Y if

and only if

c(0) = a and c'(t) = (c(t), c'(t)) = (c(t), 3c(t)).

By elementary differential equation techniques this means that c: I -*■ R

with
3t
c(t) = ae for each t £ I.

Notice that the maximal choice for I is R. ■

The key existence-uniqueness theorem which we state below in the

language of vectorfields on R will, under certain smoothness assumptions,

be generalized in the next section to arbitrary smooth manifolds.

 10.2. MAXIMAL INTEGRAL CURVES 101

10.1.5. Theorem. Let U be an open subset of and let

F : u ~* T U be a vectorfield on U. Suppose that there exists a K > 0
suoh that for each y,z E U

H'H'20F(zy)-7J2o F(s) II < K lly - z II. (2)

Then for each a 6 U, there is an e > 0 and a unique c: (-e,e) —> u

suoh that for each t £ (-£,£)

c(0) = a and c’(t) = F(c(i)).

Proof. See Chillingworth(1976) or any standard analysis text. ■

10.1.6. Lemma. The condition (2) in Theorem 10.1.5. can be replaced

by "F is C1 on U". ■

EXERCISES 10.1.

7. Show that the. vectoAfiield Y : R —*■ TR given by Y(x) = (x,x2 ) hat

a. unique integAal cuAve cut 1 which, it defined on a maximal open
inteAval. show neveAthele64 that the solution it not defined on
the whole ofi R.

2 . Show that the vectoAfield Y given by Y(x) = (x, -x)

hot at integAal cnAvet at 0 the zzao function 0^ and the function
c: R —* R+ given by
ia+ht)2 , t < -2
c(t)
0 t > -2.

Explain why ihit do el not contAadict TheoAem 10.1.5.

3. Eind the integAal cuAve ioA the vectoA^ield o& Example 10.1.2. at the
point (a,b) Oj$ R2. What it the laAgeit open inteAval in R on
which it may be defined?

10.2. MAXIMAL INTEGRAL CURVES

In Section 10.1. we showed that a smooth vectorfield on an open subset of

R gave rise to integral curves which were defined on some interval I.

In this section we show how that theory may be applied to obtain integral

curves on the manifold itself. Finally we show how integral curves may be

patched together in an unambiguous way. The idea is as follows:

102 10. VECTORFIELDS

take local representatives of the vectorfield on the manifold

♦
obtain integral curves in Rk of the resulting vectorfields
on Rk I
lift these integral curves up to the manifold via the inverses
of the chart mappings ^

show that these integral curves may be patched together to give

unique maximal integral curves on the manifold.

We begin this process with the manifold version of the existence-uniqueness

Theorem 10.1.5. The relationship between a vectorfield and its local

representative is illustrated in Figure 10.2.1.

Figure 10.2.1. Local representative Y^ of a vectorfield Y

10.2.1. Theorem. Let M be a smooth k-dimensional submanifold of Rn

and suppose Y : M —*■ TM is a smooth vectorfield on M.

 10.2. MAXIMAL INTEGRAL CURVES 103

Then for each a E M there is ccn £ > 0 and a unique integral

curve c : (-£,£) —>■ M of Y at a.

Proof. Suppose (U,(j>) is a chart for M with a € U. Then by Theorem

10.1.5. and Lemma 10.1.6. there is an integral curve £ : (-£,£) —> cf)(U)

such that

vet) = t<(> o y . ip-1 act))

with 5(0) = c)>(a).

Setting c = <J>-1°5 gives the existence of a suitable integral curve on M.

For uniqueness see the exercises. ■

We now show how the domain of each integral curve may be extended to yield

"maximal" integral curves.

10.2.2. Lemma. Let Y : M —► TM he a smooth vectorfield on a smooth

k-dimensional submanifold M of Rn and suppose

{cx : IA — M|X £ J}

is the set of all integral curves of Y at a € M. Then for each pair

X,y € J and each t £ IA n Iy, cA(t) = Cy(t).

Proof. The fact that M is a submanifold of Rn and hence is Hausdorff

is the key to the proof. See exercises for the details. ■

10.2.3. Corollary. The map c : I —>■ M where

U IA defined by I =
AEJ
cCt) = c^(t) for some X £ J is an integral curve for Y at a £ M. ■

10.2.4. Definition. Let Y, M, I and c be as in Lemma 10.2.2. and

Corollary 10.2.3. The map c : I —M is called the maximal integral

curve for Y at a € M. ■

EXERCISES 10.2.

J. Lett M be a smooth Aubmantfiold ofi Rn, Y : M —*• Tm a smooth

uectonfileld and c : I —>• M a smooth panametntzejd canoe. Shoal that
c t6 an tntegnal canve ofi Y tfi and only tfi, fion each chant (U,4>)

fion M, c^ h> an Integral canve ofi Y<j,.

2. A&Aume the hypothecei ofi Theonem 10.2.1. and AappoAe that fion each
Intenval I aitth 0 e I thene one dlitinct tntegnal canveA
c : I —*- M and d : I —* M fion Y at a. Shoal that thiA
contnadictA Theonm 10.1.5.
104 10. VECTORFIELDS

3. HeJle we pnooe Lemma 10.2.2. AAAume the hypotheAeA o{> the lemma..
Let I be the -4u.fa.6et 0($ the. Interval IA n iy conAlAting ofi all
t Auch that cA(t) = Cy(t). The atm 1a to pfwoe I = IA fl Iy.

Thti Ia done by showing:

[a] I t □. (EaAy! Txy t = 0.)

(fa) I lit open In I\ n iy. FlAAt AuppoAe s 6 1 and 4et

b = cA(s) = c^(s). Now conAldeA an Integral cuAve
d : (-£,e) —*■ M oft Y at b wheAe d(t) = cA(s+t) {,oa each
t £ (-e.e). Ute Theorem 10.2.1 to Ahow that fioA each
t £ (-£,£) we obtain cA(s+t) = cy(s+t) = d(£), which glveA
the. AequlAed azauIX.

(c) 1 Ia cloAed In I\ n Iy. [ilAe the £act that M Ia HauAdoA^

and the continuity o{> cA and cy).

(d) IA n Iy = I.

10.3. SECOND-ORDER VECTORFIELDS.

This section is about the special properties of the vectorfields which

arise from the study of second-order differential equations, such as

Lagrange's equations. In constructing a vectorfield from a second-order

differential equation, we first apply the usual "reduction of order"

procedure. The ideas of Section 10.1. can then be applied to the resulting

pair of first-order equations.

k b k
10.3.1. Example. Given f : R x R^ —y R , consider the second-order

differential equation

5" = f°(5,5') (l)

for the unknown function 5:1 —>- R^. To reduce the order put n = 5'

and thus get the pair of first-order equations

V = n
(2)
n' = f«(5>n)•

These may be combined into the single first-order equation

(5,n)' = Fo(5,n) (3)

where F : R^ x R^ —>- R2^ with F(x,y) = (y, f(x,z/)).
 10.3. SECOND-ORDER VECTORFIELDS 105

To obtain the vectorfield Y corresponding to (3) we simply adjoin the

"base point" (x,y) corresponding to (2) to the "right-hand side" F(x,y),

as in Section 10.1, to get

Y(x,y) = ((x,y), F(x,y))

= (.y, f(x,y)) (4)

Thus a vectorfield Y has been constructed in a natural way from the

original second-order differential equation (2). ■

The following remarks concerning the above example are intended

to motivate our definition below of a "second-order vectorfield".

10.3.2. Remarks. (a) The above vectorfield Y has the rather special

property that, of the four components occurring on the right-hand side

of (4), the second and third components are equal.

k k
(b) If (£,U): I —> R x R is any integral curve of the

vectorfield Y then, in accordance with Section 10.1, it is a solution

of the differential equation (3) and hence also of (2). This means in

particular that

(£,n) = (£,£')

so that (Cil) is a self-consistent parametrized curve. ■

In generalizing these ideas to manifolds, we note that if a

vectorfield Y : M —*■ TM is to have self-consistent integral curves then

the manifold M should itself be the tangent bundle TQ of some

submanifold Q of Rn. In the following definition, the elements of TQ

and T2Q are ordered pairs and ordered quadruples, respectively, of

elements of R .

10.3.3. Definition. Let Q be a smooth submanifold of R . A vectorfield

Y : TQ —>- T2Q is called a second-order vectorfield on TQ if for each

(x,y) 6 TQ there is z € Rn such that

Y(x,z/) = ((x,y), (y,z)). ■

It is possible to express the idea that Y is a second-order

vectorfield by using projection maps. This depends on the following lemma.

10.3.4. Lemma. If Q is a submanifold of Rn and Tq : TQ —+ Q is

the natural projection then the map

Ttq : T2Q —*■ TQ

106 10. VECTORFIELDS

has its value at ((a,b), (o,d)) £ T2Q given by

Ttq (.(a,b), (a,d)) = (a,a).

Proof. We use Definition 5.2.1. To this end, choose a differentiable

parametrized curve y in TQ at (a,b) with

Y'(0) = (o,d).

Now write y componentwise as (yx,y2) with yx = Tqoy so that

(tq°y) '(0) = yx'(0) = a.

Hence,

Ttq((a,i),(c,d)) = Tiq((a,i), y'(0))

= (tq( a,b), (TqoY) '(0))

= (a,c). ■

The following corollary gives the desired characterization of

a second-order vectorfield in terms of projections.

10.3.5. Corollary. Let Cl be a submanifold of Rn. A veotorfield

Y : TQ — T2Q is second-order if and only if

TiqoY = idyq .

Proof. Given a vectorfield Y : TQ —y T2Q, let (a,b) £ TQ and let

Y(a,b) = ((a,b),(e,d)) £ T2Q .

By Lemma 10.3.4, the condition

(TTqoY)(a,b) (idjq)(a,i)

is equivalent to

(a,a) (a,b),
and hence to

Y(a,£>) = ((a,b) ,(b,d)) .

This establishes the desired equivalence of the conditions in the

corollary. ■

The next result shows that second-order vectorfields may be

characterized in terms of their integral curves - a possibility which might

be guessed from Example 10.3.1.

 10.3. SECOND-ORDER VECTORFIELDS 107

10.3.6. Theorem. A smooth, vectorfield Y : TQ —*■ T2Q is second—order

if and only if each of its integral curves is self-consistent.

Proof. Suppose first that Y : TQ —*■ T2Q is a smooth second-order

vectorfield. Let c: I —*■ TQ be an integral curve for Y so that

c; = Y»c
and hence

(Ttq)oc’ = (TTq)oY°C

= idjqoc by Corollary 10.3.5.

= c.

It now follows from Exercise 10.3.3 that (Tq°c) ’ = c and hence c is

self-consistent.

Conversely, suppose that Y : TQ —»■ T2Q is a smooth vector-

field with the property that each of its integral curves is self-

consistent. Let (a,b) £ TQ. By Theorem 10.2.1, there is an integral

curve c : I —> TQ of Y at (a,i>). Hence

c5 = YoC
and so

(Ttq)oc’ = (Ttq)°Yoc .

From Exercise 10.3.3 it now follows that

(Tqoc)’ = (TTq)oYoc

and since c is self-consistent this implies that

c = (Ttq)°Y°c.

Evaluating both sides at 0 gives

(a,b) = (Ttq)°Y(a,b).

Thus TTq»y = id-j-Q, showing that Y is second-order by Corollary 10.3.5. ■

Thus the significant part of the integral curve c of a

second-order vectorfield is the first component since the second is simply

the derivative of the first.

10.3.7. Definition. If c : I —*■ TQ is a (self-consistent) integral

curve for a second-order vectorfield Y : TQ —*■ T2Q then Tq°c: I —*- Q

is called a base integral curve of Y. ■

The following lemma will provide a useful tool for establishing

a further criterion for second-order vectorfields, involving local

representatives.
108 10. VECTORFIELDS

10.3.8. Lemma. Let Cl be a smooth submanifold of Rn. A parametrized

curve c: I —*■ TQ is self-consistent if and only if, for each chart
(U, <t>) for Q, the local representative c-^ = T<}>°c is self-consistent.

Proof. See the exercises. ■

10.3.9. Theorem. Let Q be a smooth submanifold of R‘. A smooth

vectorfield Y : TQ —*■ T2Q is second-order if and only if, for each chart
(U, <j>) for Q, the local representative Yj^: Tu —► T2U is second-order.

Proof. Suppose that Y is second-order and let r|: I —*■ TQ be an

integral curve of the vectorfield Y^. Put c = T<J>-1°r| so that c is an

integral curve of Y by Exercise 10.2.1. By Theorem 10.3.6, c is

self-consistent. Hence, by Lemma 10.3.8, the local representative

t| is also self-consistent. It now follows from Theorem 10.3.6. that

T<)>
the vectorfield is second-order.
T4>
The proof in the converse direction is similar.

EXERCISES 10.3.

1. UAlte down the. oeetoAfleld on TR coAAesponding to the. following

seeond-oAdeA dlffeAentlal equation foA E, : I —>- R:

i" + 2E,' + E, = 0.

Ftnd the. maximal base IntegAal cuAve of thl& vectoAfield at

Ca,b) e TR.

2. Let Q be a submanifold of Rn and let Y : TQ —* T2Q be any map.

Show that Y ti> a second-oAdeA veeto Afield on TQ If and only If

tTq-y ■ idTQ ■ TVY

wheAe t^:T2q —*■ TQ and Tq:Tq —► Q aAe the natuAal pAojeetlons.

3. Let Q be a submanifold of Rn, let c : I —> TQ be a Smooth

paAametAlzed euAve and let Tq:TQ —»■ q be the natuAal pAojectlon.
Show fxorn Veflnltlon 7.1.1 that

(Ttq)°C’ = (Tqoc)’ .

4. Let Q be a submanifold of Rn with (U,$) as a ehaAt and let

Tq : TQ —Q and x^y) : T4>(u) —► d>(U) be the pAojectlons. Show

fAom Veflnltlon 5.2.1 that, fox each paAametAlzed cuAve
c : I -*■ TQ,

t+(u)-t»’° ■ **TQ'C •
 10.3. SECOND-ORDER VECTORFIELDS 109

Vnooe Lemma 10.3.8. (Hint: the preceding exeticiAe.)

Let Q be a Aubmanlfiold ofi Rn and let Y : TQ —* T2q be a smooth

vectonfileld. Skou) that Y Ia Aecond-oAdeA tfi and only l^ each ofi
ItA Integral euAveA c : I —*- Tq can be uvvitten aA c = d’ wkene
d : I —»- Q AattA^leA d»» = Y°d’ .
 11. LAGRANGIAN VECTORFIELDS

In this chapter the theory of veotorfields and their integral

curves will he related hack to the study of the motion of a particle on a
submanifold Q of R . Whereas in Section 8.3 we assumed the existence
of a self-consistent differentiable curve c: I —*■ TQ satisfying Newton's
second law of motion, in this chapter one of the aims is to prove the
existence of such curves - thereby verifying that the theory developed in
Section 8.3 is not vacuous.
To achieve this aim we shall use the idea that Lagrange 's
equations, with respect to each chart in an atlas for TQ., define "local"
veotorfields which can be patched together to form a "global" vectorfield
on the whole of Tq. The existence-uniqueness theory of the preceding
chapter can then he used to establish the existence of integral curves for
this vectorfield. Finally, the integral curves so obtained can be shown
to satisfy Newton 's equations of motion.

11.1. GLOBALIZING THEORY

This section sets up some language which is useful for studying how local

vectorfields may be patched together to form a global vectorfield on a

Throughout this section H will denote a k-dimensional

submanifold of R .

11.1.1. Definition. Let (U,4>) and (V,ip) be two compatible charts for
M. A pair of local vectorfields

Y : <fi(U) —*■ T<(>(U) and Z : \p(V) —* T^(V)

is said to be patchable with respect to these charts if, on the domain

U n V,

(T<J>) Y o <p = W1 Z o \p ,
 11.1. GLOBALIZING THEORY 111

Figure 11.1.1. Patchable local vectorfields.

The motivation for this definition is that Y and Z will

satisfy it whenever there is a global vectorfield on the manifold M of

which Y and Z are the local representatives with respect to the charts

(U,cj)) and (V,ip) as is illustrated in Figure 11.1.1.

In applications to specific examples it will sometimes be

possible to avoid awkward computations by using the following theorem -

which shifts attention away from the vectorfields to their integral curves.

11.1.2. Lemma. Let (U,<J>) and (V,^) be too compatible charts for M.

A pair of local vectorfields

Y : <}>(U) —*■ T0(U) and z : ip(V) —* T^(V)

 112 11. LAGRANGIAN VECTORFIELDS

is patchable with respect to these charts if the following condition holds:

if ri : I —*■ |i(U (1 V) is an integral curve for Y

then C : I —*- ip(U n V) is an integral curve for Z

where £ = (^°<}r1)on.

Proof. Suppose the condition stated in the lemma holds. Let a £ 4>(U fl V) .
By the local existence-uniqueness theorem, Theorem 10.1.5, there is an

integral curve at a for Y, say

0 : I —*■ 0 (U n V)

where I is an interval containing 0. Hence

n; = Yoq (1)

and

n(0) = a. (2)

The assumed condition in the lemma now implies that

C = (4* ° 0 0 (3)

is an integral curve for Z so that

= Z o C. (4)

Now substitute (3) into(4) and use Definition 7.1.1 and then the chain rule

to get

T(^o<j>-1)on’ = z o (iM_1) °n.

Hence

T(i^ot()-1)oYon = z o 4>—1)°n by (l)

T(iJ;o(j>-1)oY(a) = Z ° (ip°<p~1)(a) by (2)

and so
T(f>-1oY(a) = T^_1 o z o ip o <j>-1(a).

But as a is an arbitrary element of <f>(U D V) this implies that, on the

domain cf>(U D V),

T$~l.Y = T^_1 o Z ° \poQ~1

and hence, on the domain U (1 V,

T4>-1.Yo<J) = Tij;-1 o z ° ip. ■

The next theorem is a trivial consequence of our definition of

patchability.
 11.2. GLOBALIZING LAGRANGE 113

11.1.3. Theorem. Let {(u\<|>^): X 6 J} be an atlas for M and, for each

I. € J, let be a smooth local vectorfield on (Ku^). Suppose that, for
each X, y £ J, the vectorfields YX and Yw are patchable with respect
to the charts (u\ and (Uy, <)>y). There is then a unique smooth
vectorfield Y on M such that each Y^ is the local representative of
Y in the chart (u\ (J)''1).

Proof. We may define Y on each by putting

y|ux = (T4>A)_1 o yx o cpx .

This definition is unambiguous by the definition of patchability. This,

moreover, is the only possible choice of Y. ■

11.1.4. Corollary. If in Theorem 11.1.3 the manifold M is the tangent

bundle of some other manifold and each local vectorfield Y^ is second-
order then the vectorfield Y on M is also second-order.

Proof. This is an immediate consequence of Theorem 10.3.9. ■

11.2. APPLICATION TO LAGRANGE'S EQUATIONS

The study of the motion of a particle on a manifold leads, via Lagrange's

equations with respect to the charts on the manifold, to local second-order

vectorfields. In this section it will be shown that these vectorfields can

be patched together to give a vectorfield defined on the whole of the

tangent bundle of the manifold.

In order to be able to apply the globalizing theory of Section

II. 1, we must first check that the local vectorfields obtained from

Lagrange's equations are patchable. For this task, the next lemma is

crucial. Note that the function L which occurs in the lemma is not

restricted to being the Lagrangian for any particle motion. Birkhoff(1927),

page 21, has an alternative albeit more coordinatewise proof.

Throughout this section, Q will denote a -dimensional

YL
submanifold of R .

11.2.1. Lemma. Consider any smooth function L : TQ —*■ R. If (U, q)

and (V, r) are charts for Q and if c : I —► U fl V is any self-

consistent smooth curve then
 114 11. LAGRANGIAN VECTORFIELDS

Proof. A mild extension of Corollary 7.2.8, shows that

3L = 3L _3r 3l _3r
3q 3^ 3q 3r 3q ° TQ (1)
and also

3L = 3L 3r
(2)
3q 3r 3q
3r
since r depends only on q so that — - We also have by Corollary
7.3.4,

3r 3r
- O c = -r— o Tr (3)
3q
3q
and hence by Corollary 7.3.8.

3r 3r
oC 0 c . (4)
'■Sq ' 3q

Now suppose, in accordance with the hypotheses of the theorem,

that

3L V 3L
(5)
7° >

where by (2)

'%• c)’ .fit. cVii. c + ^.c fii. CV

‘3q ‘ '3r ' 3q 9r '9q '

= (*k «CV ll _ . 3L 3r
° xQoC + 7T oc ° c
3r M

by (3) and (4)

while by (1)

3L _ _ 3L 3r 3l 3r
oc-oC — o C + - oC o To c
3q 3q 3r 3q Q

Hence (5) becomes

3L V 3L \ 3r
3r
Cj - 37 * 3? • Vc ■ °-
But as (U,q) and (V,r) are both charts for Q, it follows that the

matrix on the right is non-singular and hence

0 c = 0. ■
eH'-t
 11.2. GLOBALIZING LAGRANGE 115

Our next result recasts Lemma 11.2.1 into a form which refers

to local representatives of curves rather than to the curves themselves.

Motivation for this comes from Chapter 9 where we studied the

explicit form of Lagrange's equations for the local representatives of the

curves representing the history of the particle. Recall that if

c: I —»- TQ is a curve and (Tu, Tq) is a natural chart for TQ then the

local representative c-^ of c is defined by

cTq = Tq°C 0r c = (Tq)_1°cTq •

In addition to the notation (q°Tq°c, q°c) for the local representative

of c, we also used the notation (q, q') to give our answers in Chapter 9

a more customary appearance.

Since, however, we now wish to change tack and take the local

representatives themselves as our starting point, we shall use letters such

as and "n" to denote them. In this way we avoid the possibility of

begging the question as to whether we can construct a suitable global curve

c : I —»■ TQ which is independent of charts.

11.2.2. Lemma. Consider any smooth function L : TQ —*■ R. Let (U,q)

and (V,r) be charts for Q and let

H: I —► (Tq) (U n V) and £:!->- (Tr)(U fl V)

be any self-consistent smooth curves.

If
/9L -t ,-1
(Tq)
V 3L
° (Tq) oq = 0 (1)

then ° (Tr)"1.^) ~ Tr ° (Tr) 1=C = 0 (2)

'9r '

where C = Tr ° (Tq) n . (3)

Proof. Define a parameterized curve c: I —► D (1 V by putting

c = (Tq) 1or) . (4)

From the hypothesis (1) of the lemma

9L Y 9L .
9q >

Since r| is self-consistent, however, the same is true of c, by Lemma

10.3.8. Hence, by Lemma 11.2.1,

/9l Y 9L
O c = 0. (5)
 116 11. LAGRANGIAN VECTORFIELDS

But from (A) and the hypothesis (3) in the lemma,

c = (Tr) 1 ° C-

Substituting this in (5) gives the desired conclusion (2) of the lemma. ■

The stage has now been set for an application of the globalizing

theory of Section 11.1 to the local vectorfields defined by Lagrange's

equations. The assumption that the function L : TQ —*■ R is the

Lagrangian of a mechanical system will now be used for the first time in

this section.

11.2.3. Theorem. Let L : TQ —► R be the Lagrangian of the particle

moving on the submanifold Q as described in Section 8. 3. If
{(UX, qX):A £ J} is an atlas for Q then
(a) for each A £ J there is a smooth local second-order vectorfield

YX : TqX(UX) —► T2qX(UX)

whose integral curves are the solutions qX of Lagrange's equations

-1 3l
(TqX) °n (TqA)"1onA = o ,
” °

(b) there is a unique smooth second-order vectorfield

Y : TQ —T2Q

for which the local representative in the chart (Tu\ TqX) is YX .

Proof. By Theorem 9.3.3, for each A £ J, Lagrange's equations in the

form written above are second-order differential equations for the function

qX where q X = (qX,qXf). Thus, as in Section 10.3, they define a smooth

local second-order vectorfield YX with the stated integral curves.

By Lemmas 11.2.2 and 11.1.2, each pair of vectorfields YX

and Y^ is locally patchable with respect to the charts (TUX, TqX) and

(TU^, Tq^), so Theorem 11.1.3 gives the existence of the desired

vectorfield Y. ■

The vectorfield Y occurring in Theorem 11.2.3 will be called

the Lagrangian vectorfield of the mechanical system described in Section

8.3.

11.2.4. Corollary. If Y is the vectorfield given by Theorem 11.2.3

then for each a £ Tq there is a unique integral curve c : I —>- Tq for
Y at a having maximal domain. This curve is self-consistent and for any
 11.3. BACK TO NEWTON 117

chart (U, q) for Q it satisfies Lagrange's equations

3L 3l
o c ° c = 0 .
3q
3q

Proof. Left for the exercises. ■

EXERCISE 11.2.

Phova Conottasiy 11.2.4.

11.3. BACK TO NEWTON.

It will now be shown that the arguments leading from Newton's second law

of motion to Lagrange's equations can be reversed. A result in this reverse

direction is given in Landau & Liftshitz(1960), page 9, although their

result applies only in the more rudimentary context in which there are no

geometric constraints on the motion. The following lemma will be used in

the proof of our result.

11.3.1. Lemma. Let X be the identity map on R , let Q be a sub-

manifold of R of dimension k and let a £ Q. If (U,q) is a chart

for Q at a then the set of vectors

is a basis for TaQ.

Proof. For 1 < i define y£ : I —>■ U by putting

Yi<.t) = q X(q(a) + t e.f)

so that, as in the proof of Theorem 7.3.10,

v<°> - It <»> •
But by the construction of y^

q " Y{ = q(a) + idy H

and hence by the chain rule

Tq 0 Ty.£ = T(q(a) + idx e^)

Tq(a, y .'(0))= (q(a), e^).

But since Tq|TaQ is linear and the set of vectors (ei,e2»•••»e^) is

 118 11. LAGRANGIAN VECTORFIELDS

linearly independent the same is true of the set of vectors

(Yi,(0)» Y2 ' (0)>•••>Y^ ' (0))» which therefore is a basis for the
/c-dimensional vector space TaQ. ■

Although our theorem is stated and proved only in the context

of Section 8.2, where the submanifolds lie in R3, it can easily be

extended to the more general situation discussed in Section 8.3.

11.3.2. Theorem. Consider a particle of mass m moving on a k-dimensional

submanifold Q of R3 subject to a conservative field of force. Let
T : TQ —y R and V : Q —*■ R be the kinetic and potential energies of the
particle and let L = T-V°Tq be the Lagrangian.
The reaction force R, constraining the particle to stay on the
manifoldcan then be chosen in just one way at each point of TQ in order
that the following condition holds:
if c: I —*■ TQ is an integral curve for the Lagrangian
vectorfield on Tq determined by L then Newton's second law holds for
each t £ I:

m(X° tq° c) "(£) = 0 Tq +tt2 ° R ^ 0 c(t) .

At each point the reaction force R is orthogonal to Q.

Proof. Let c : I —>- Tq be an integral curve for the Lagrangian vector-

field so that, as in Corollary 11.2.4, c is self-consistent. Let t £1

and let (U,q) be a chart for Q such that c(t) £ U. The argument

contained in the proof of Theorem 8.2.2 which shows that for 1 < i < k

Id T 9T .. ./ (9X
0 c = m(Xoc) • !- ° Tqo c (1)
9q.

remains valid. By the hypotheses of the theorem, moreover,

9L
(*-. '' c = 0
V9q. 9q i
and since L = T - Vot^ this implies that

/ 9T / 9T ^ 9V
c ]-°c + ” T C - 0 . (2)
9q. 9ck
^i
Q

Subtracting (2) from (1) now gives

9V
m(i°c),'(lr0 tqoc) 9^7 ° XQ ° c
 11.3. BACK TO NEWTON 119

(9V 3X \
' \dx dqiJ°TQ.°c
so that

(”«•<=)' + H" Vc)-%“ tq"c - ° •

But since this holds for 1 ^ i- ^ k we may apply Lemma 11.3.1 after

evaluating both sides at t, to deduce that the vector

m(Xoc)'(i) + ~ ° xpoC(t)

is orthogonal to the tangent space Tc(t)Q. Hence there is a function

R : TQ —* TQ such that

• 9V
m(X° c) (t) + <= Tq°c(£) = 7T2oRoc(t)

and R(c(t)) is orthogonal to T Q. ■

c(t)

EXERCISE 11.3.

Extend Tke.oA.em 11.3.2. to the context oh Se.ctton 8.3 and. then

pAjove. tt.
 12. FLOWS

At any particular time the state of a system is given by some

collection of parameters. A deterministic mathematical model would hope

to predict future states of that system. So if say a is the state at

time t = 0 we may have a state A*(a) at time t. After another time

lapse of say s we would have a state As(A^(a)). We would therefore want

• S f-
our model to satisfy the relation A (A‘'(a)) = A (a), that is,
si • •
A = A °A . In other words the "time evolution" of our system would have

to satisfy certain group properties. We could thus think of the "time

evolution" operator A^ as acting on all points of a manifold of states

at once and think of this as giving a "flow" on the manifold with increasing

time.
In the case of a mechanical system our manifold is TQ and the

flow is a collection of integral curves in Tq of a Lagrangian vectorfield,

each such curve corresponding to a set of initial conditions.

12.1. FLOWS GENERATED BY VECTORFIELDS

Given a smooth vectorfield on a manifold, it was shown in Chapter 10 that

there is a unique integral curve at each point of the manifold. Here our

concern is with the overall structure of the set of all such integral

curves obtained as the point ranges over the manifold.

It is helpful to regard the integral curves as the paths traced

out by particles of a fluid moving with velocities which are prescribed

at each point by the vectorfield. In this sense the vectorfield is said to

determine a "flow" on the manifold. Intuitively, we may think of the flow

as the set of all the integral curves of the vectorfield. The formal

definition given later, however, expresses the idea of a flow in terms of

certain mappings associated with the integral curves. An example will be

given below to show how these mappings arise and to illustrate their

special properties.
 12.1. FLOWS AND VECTORFIELDS 121

In the study of flows, attention is usually restricted to

vectorfields which satisfy the following definition.

12.1.1. Definition. A smooth vectorfield Y on a manifold M is said to

be complete if every integral curve for Y has the whole of R as its

domain. ■

Not all vectorfields have this property (as Exercise 10.1.1. shows) but it

can often be shown to hold for the vectorfields arising in Lagrangian

mechanics.

12.1.2. Example. The vectorfield on R2 of Example 10.1.2. given by

Y(x,y) = {(x,y), (~y,x))

has at each point ([a,b) £ R2 the integral curve c: R —R2 with

c(A) = (a cos(t) - b sin(t), a sin(t) + b cos(t)) (1)

Thus each integral curve is defined on the whole of R and so the

vectorfield is complete.

With a change of viewpoint we may regard c(t), as given by (1),

as function of the point (a,b) £ R2. This determines a map A7' : R2 —*- R2

for each t £ R with

A^(a,b) = (a cos(t) - b sin(t), a sin(t) + b cos(£))

This map A^ can be regarded as a "time evolution" operator which assigns

to each initial state of the system its final state after the lapse of a

time t. Thus A2' is a map which acts on the whole of R2, rotating each

point around the origin anticlockwise through an angle t. Hence it is

clear that

At+s(a,b) = A*(AS(a,b))

= A*°A S(a,2>)

or, more simply,

At+S = A* • AS (2)

Intuitively, this means that the time evolution operator can be applied to

initial conditions (a,£>) given at an arbitrarily assigned initial instant.

The property (2) is illustrated in Figure 12.1.1.

Other properties of the maps (A2' : t £ R} are

122 12. FLOWS

so A can be regarded as the operator which moves back along the

integral curves for a time t. Most importantly, for each t £ R,

A* : R2 -*■ R2

is a diffeomorphism. ■

In order to set the above properties of the time-evolution

operators in a natural algebraic setting, it is appropriate at this stage

to introduce some notation for the set of diffeomorphisms from a manifold

into itself.

12.1.3. Definition. Let M be a manifold. We denote by Diff (M) the set

00

of all C diff eomorphisms from M to M. ■

Figure 12.1.1.

12.1.4. Example. In each case a subset of the given set of diffeomorphisms

is shown:

(a) {X id : A € R and A i 0} <= Diff°°(Rfe)

r*
(b) {id a: o e Rfe} c Diff°°(Rfe)

(c) {idp2 + t c : t € R, c £ R2} c Diffc°(R2)

(d) {e* id : t 6 R) c Diff°°(R) . ■

12.1.5. Lemma. The set Diff (M) is a group under the operation of
composition. ■
 12.1. FLOWS AND VECTORFIELDS 123

The property (2) of the time evolution operators in Example

12.1.2. can now be expressed in terms of group theory. Since the map

t —y tv'

has the group (R,+) as its domain and the group (Diff(R2), o) as its

codomain, property (2) simply says that this map preserves the group

structure and hence is a group homomorphism.

The formal definition of a flow can now be given. It makes

no direct reference to vectorfields, although they provide the motivation

for introducing the concept.

12.1.6. Definition. A flow on a manifold M is a map

A : R —Diff°°(M)

such that, for each s and t £ R, A(t+s) = A(t)°A(s). We also require

that the map (a,t) >—* A(t)(a) be smooth. ■

Thus the image A(R) of R under a flow A is a copy of the real line in
„ 00

Dlff (M) with the same group properties as R and is sometimes called a

one-parameter group of diffeomorphisms.

Figure 12.1.2.

12.1.7. Examples. The following maps are flows on the manifold R:

oo ~h
(a) A: R -> Diff (R) : t e. id
 124 12. FLOWS

00 rt
(b) A: R —► Diff (R2)

: t (cos(A)id1-sin(t)id2, sin(t)id1 + cos(t)id2). ■

The following definition recovers the idea of an integral

curve, or more precisely its image, from the flow.

12.1.8. Definition. Let A be a flow on a manifold M. The orbit of a

point a £ M under the flow A is the set of points from M

{A(t)(a) : t £ R}. ■

This idea is illustrated in Figure 12.1.3.

Figure 12.1.3. The orbit of a under A

The orbit of each point of R2 under the flow in Example 12.1.7(b) is a

circle whose centre is the origin.

To complete this section we state a theorem which brings

vectorfields back into the picture and is illustrated by Example 12.1.2.

12.1.9. Theorem. Let Y be a smooth vectorfield on a manifold M. If

Y is complete then putting

A (t)(a) = c(t),

where c is the maximal integral curve of Y at a, defines a flow on M.

Proof. This involves ideas and results from Abraham & Marsden(1978),
Section 2.1. ■
 12.2. FLOWS FROM MECHANICS 125

EXERCISES 12.1.
7. Check that each ofi the AubAetA tn Example 12.1.4. Ia a Aubgnoup ofi
the Indicated gnoup ofi dl^eomonphlAmA, undeA compoAltton.

2. Let A : R —> Diff (M)

oo
be a falow on a manifold M. Show that
A(0) = idM and, (,oA each t £ R, (A(t))-1 = A(-t) .

3. Vetl&y that the map A oi Example 12.1.7(a) Ia a &low on R. Unite

down ItA one-pan.ameteA gnoup ofi dl^eomoAphlAmA. Give the onblt o&
the point a £ R In each o& the caAeA a < 0, a = 0 and a > 0.

4. VeAliy that the map A : R —* Diff^tJR) with

(V \
A(t)(a,u) = ^ sin(kt) + a cos(kt), v cos(Tct) - ak sin(kt) j

wheAe k > 0, Ia a low on TR. Sketch Aome typical onbltA.

12.2. FLOWS FROM MECHANICS

Some familiar problems from mechanics will be used here to illustrate the

idea of the flow of a vectorfield. The theoretical background for these

examples is very straightforward. The vectorfields which arise in these

problems can be proved complete by an application of the results in the

next section; the integral curves of these vectorfields then generate flows

in accordance with Theorem 12.1.9. Hence our discussion will concentrate

on the geometrical ideas associated with the flows.

C°
Recall that a flow A : R —> Dlff (M) generated by a complete

vectorfield Y on a manifold M determines two families of mappings

(a) for each t £ R, the time evolution operator for lapse of time t

A* : M —y H : a —>■ A (t) (a)

(b) for each a £ M, the integral curve at a

Aa : R —* M : t —>- A(t)(a).

In addition, there is for each a £ M the subset of M consisting of the

image of the integral curve through a and called the orbit of the flow

through a.

In the mechanical problems which follow, the configuration

manifolds are 1-dimensional and hence their tangent bundles, which contain

the orbits, are 2-dimensional. This permits direct diagrammatic

representation of the orbits and also, but less directly, of the integral
126 12. FLOWS

curves and the time evolution operators. Thus in Figure 12.2.1, the

curves with arrows are the orbits of a flow A. From a knowledge of the

times at which the particle assumes the various states, it is possible to

infer what is doing.

t
A
states
initial after
states time t

Figure 12.2.1. Time evolution operator.

12.2.1. Example. (Motion under gravity in one dimension.) For a particle

of mass m moving under gravity and with the x-coordinate measuring the

height, the Lagrangian L is given by

L = h mx2 - mg

and hence Lagrange's equation becomes

(x°c) = -g .

When expressed as a vectorfield, as in Example 10.3.1, this gives

Y : R —*■ TR with

Y(x,W) = ((x,u) , (.W,-g))

for each (x,u) E TR. The vectorfield Y is shown in Figure 12.2.2.

Elementary differential equations techniques show that the

integral curve for Y at the point (a,y) € TR is given by

c(t) = (a + vt - hgt2, v-gt)

oo
and hence the flow A : R —* Diff (TR) is given by

A(t) (a,y) = (a+yt - hgt2, v-gt)

 12.2. FLOWS FROM MECHANICS 127

Figure 12.2.2. Gravitational vectorfield on phase space.

Note that each integral curve is self-consistent and that the orbits consist

of parabolas which are tangent to the vectorfield. See Figure 12.2.3.

Figure 12.2.3. Gravitational orbits in phase space.

12.2.2. Example. (The simple pendulum.) This consists of a particle of

mass m constrained to move under gravity on a circle lying in a vertical

plane, in the absence of friction.

Distances in the horizontal and vertical directions will be

measured by x and y-coordinates, respectively, and the circle on which

the particle moves with be taken as S1. See Figure 12.2.4.

 128 12. FLOWS

Figure 12.2.4. A simple pendulum

Thus the configuration manifold for the particle is S1 and the velocity

phase space is TS1, which is diffeomorphic to the cylinder s1 x R in

view of Exercise 6.1.1.

e2

b 0

o it

6i

O-TT

Figure 12.2.5. Charts for S1 .

We shall use the two charts (Ux, 0X) and (U2, 02) for S1

which are defined by Figure 12.2.5. These charts are compatible with the
submanifold structure of S1 and have the property that
 12.2. FLOWS FROM MECHANICS 129

on one connected component of Ux fl U2

{6*
02—2tt on the other.

Hence
Sl= 02 on TUx fl TU2 .

Thus we may define a smooth map, called the angular velocity map, by putting

co : Ts1 —*- R with to = {-1 on I^1

[02 on IU2

In addition to its obvious kinematic interpretation, the angular velocity

has an important geometrical role as a component of the diffeomorphism,

defined in Exercise 6.1.1,

$ : TS1 —>- S1 x R

between the tangent bundle of the circle and the cylinder. It is set as

an exercise to show that, in fact, this diffeomorphism satisfies

$ = (x i, to) (1)
S

The geometrical quantities needed for the study of the simple

pendulum have now been defined. Before giving a formal discussion, however,

we use physical intuition to guess at the possible types of orbit which

can appear in the velocity phase space, as in Irwin(1980). Thus the orbits

shown in Figure 12.2.7, shown in both chart and cylinder representation,

can arise in the following ways:

(a) If placed with zero initial velocity at either the lowest or

the highest point of the circle, the particle will stay there

indefinitely. These two "equilibrium points" give rise to the

two single-point orbits where the cylinder meets the y-axis.

These two orbits are said to be "stable" and "unstable",

respectively, for obvious reasons.

(b) If started with zero initial velocity at some point

intermediate in height between the two equilibrium points, the

particle will oscillate to and fro, rising indefinitely often

to the initial height - first on one side of the origin, then

on the other. These motions are represented on the cylinder

by closed orbits which encircle the y-axis.

130 12. FLOWS

(c) If the particle is started with sufficiently large initial

velocity it will pass through the highest point of the circle

and then continue to make complete circuits. The corresponding

orbits are the closed orbits which go right around the cylinder

There are two distinct families of these orbits because the

particle may traverse the full circle either clockwise or

anticlockwise.

It is left as an exercise to show that there exists an orbit having the

unstable equilibrium point orbit as a limit point. A particle traversing

this orbit approaches arbitrarily close to the equilibrium point without

ever actually reaching it.

To complete our discussion of the motion of the simple

pendulum, it remains to show how the formal theory from previous chapters

leads to the sketches of (a) the vectorfields in Figure 12.2.6. and (b)

the orbits in Figure 12.2.7.

As to the vectorfields, we first apply the theory of Section

8.2 to the particle constituting the bob of the pendulum. In terms of

the x and y-coordinates the kinetic and potential energies are

T = lsm(x2 + y2)

V = mgy.

After restricting these functions to TS1 we get the Lagrangian function

L : TS1 —y R by putting L = T - V«x We will need to use both of the

charts (l^, 0;^) and (U2, 02) for S1 so let i. = 1 or 2 and note
that

x = sin°0.
on U,'
y = -coso0£

and hence

L = %m02 + mg cos°0£<>t 1 on U^.

Thus, in terms of the local representative = 0{ ° l i"c of an integral

curve c: I —► TS with respect to the chart (U^, 0^) , Lagrange's

equation becomes

0 . "+ tisinoO . = 0. (2)

1 v V —

Hence, by the procedure given in Example 10.3.1, we find for i = 1 that

the local vectorfield corresponding to this second-order differential

equation is
 12.2. FLOWS FROM MECHANICS 131

YT6 : T(~TT’Tr) ~* T2R

: (<f>,co) i y ((4>,(jj) , (0),-sin(4>) ) )

This vectorfield is sketched on the left in Figure 12.2.6 and lies flat

in the plane. The sketch for Yjg is similar. Now, by the

construction in Section 11.1, the global vectorfield for the problem,

Y: TS1 —* T2S1,

is related to the above vectorfields by

Y = T(T0y) XoY »T6„ on Ui

T0,'
Finally, the vectorfield sketched on the cylinder in Figure 12.2.6. is

Y^: S1 x R —> T(S1 x R)

given by

Y$ = T<J) o Y o O-1

where 0 is the diffeomorphism from TS1 to the cylinder. It is possible

to calculate an explicit formula for Y$, although we shall not give the

details here. Intuitively, this vectorfield is obtained by wrapping the

local vectorfields and Y. around the cylinder, at the same

T0! T9,
-2
time allowing their arrows to poke out into R3 while remaining tangent

to the cylinder.

Figure 12.2.6. The vectorfield for the simple pendulum.

 132 12. FLOWS

Turning now to the orbits, our aim is to formally justify the

claims made earlier on the basis of physical intuition. In other words

we are going to show how to derive mathematically the sort of orbits shown
in Figure 12.2.7.

co

Figure 12.2.7. Pendulum orbits in velocity phase space.

Recall that an orbit is a subset of Ts1 of the form

(c(t): t £ R} (3)

where c : R *■ Ts1 is an integral curve of the Lagrangian vectorfield of

the pendulum. In Figure 12.2.7 the images of such orbits are shown under

(a) the natural chart T0! and (b) the diffeomorphism $ onto the
cylinder.

Under the natural chart T0X the image of the orbit (3) is the
set

{(01,0! ')(*): t £ R} £ TR ,

where (0lf0!') denotes the local representative of c with respect to

the natural chart T01. A similar set is obtained under T02. It is

left as an exercises to check that the image of the orbit (3) under the

diffeomorphism $ is obtained by wrapping the images of the orbit under

T0! and T02 around the cylinder.

 12.2. FLOWS FROM MECHANICS 133

To complete our discussion of the orbits we use energy

considerations. Given e 6 R, the set of all points in Ts1 at which

the sum of the kinetic and potential energies assume the value e turns

out to be a curve in TS1 - called a constant energy curve. Since by

Theorem 8.5.1 the total energy stays constant along an integral curve

c: R —y TS1, it follows that the orbits are subsets of the constant energy

curves. The images of the constant energy curves under the charts T0-L

and T02 may be obtained by putting

h(Q')2 - g coso0^ = e_ (4)

'Is

for i = 1,2 and then sketching the graph of 0^ against 0£. This is,

in fact, the way in which the curves shown in Figure 12.2.7 were obtained.

Each orbit must lie inside a constant energy curve, although some constant

energy curves contain more than one orbit. Since (4) implies that

e > -g, the following cases exhaust the possibilities.

(a) e = -g. Here the constant energy curve is a single

point corresponding to the stable equilibrium point.

(b) -g < e < g. Here the constant energy curves are closed

curves encircling the stable equilibrium point. Each

curve is a single orbit.

(c) e = g. The constant energy curve contains the unstable

equilibrium point. Removal of this point leaves two

connected components, each of which is a single orbit.

(d) e > g. The constant energy curves encircle the

cylinder. Each curve is a single orbit.

These results will be verified in the exercises.

EXERCISES 12.2.

1. V-Licum the advantageA o(5 nepaeAenttng the oabltA ion the pendulum
problem on a cyLinden nathen than tn the plane via chantA
(Aee Inwln(l980), page 4).

2. Venliy that the diHeomonphiAm 0: Ts1 —*■ s1 x R defined In

ExenclAe 6.1.1 can be ivnltten In the ionm [1] In the text.
 12. FLOWS

LeX $ be as in the previous exercise and let (u^.6^) be the

chant ion s1 Introduced In the text (i = 1,2). Show that ii
c: R *■ Ts1 is an Integral curve ion the Lagnanglan vectonileld oi
the pendulum then

<f°c(t) = (sin°0£, cos°0^, 0^)(£)

ion all t £ R Auch that c(t) £ U^.

(Thlc> veniileA the claim made In the text that the Image oi an onbtt
under $ Ia obtained by snapping anound the cyllnden the swages oi
that onbtt under each oi the chantA T0! and T02.)

Verciy inom (4) In the text that the conAtant energy cunveA have the
general Ahape claimed in the text ion the vaniouA nangeA oi valueA
oi the parameter e.

Let c: R )- Ts1 be an integral curve oi the Lagnanglan vectonileld

oi the pendulum. Mote that li e > g then the angular velocity,
woe, along the integral curve Ia bounded away inom zero. Veduce
that in thlA caAe the conAtant energy curve iA a Aingle orbit.

Let 0 : R —* R be a dlHerentiable iunctlon Auch that

[a.) 0(t) Ia bounded oa t —► °°

(b) lim 0'(t) exlAtA.

t -*■ °°

Show inom the mean value theorem that

lim 0'(t) = 0.
t -*■ 00

figure 12.2.& AhowA a conAtant energy curve ion the pendulum in the
caAe -g < e < g, which cnoAAeA the Q1-axis at the points (-a,0)
and (a,0) where 0 < a < it. Let (6x,e' ) be the local
representative oi an integral curve and suppose that (j^.e')(£)

lies on the constant energy curve and in the upper hali plane when
t = t0.

This exercise is to prove that at some later time t > t0 the point
(5i.V)(*> into the lower hali-plane. Suppose, on the
contrary that it stays in the upper hali-plane ion all t > t0 and
then dereve a contradication by showing, with the aid oi Exercise 6,
that
 12.3. LAGRANGIAN FLOWS 135

(a) lim d1(t) exltit

t 00

(b) lim d' (t) exitit

t -*■ 00

(c) lim 01 (t) = 0

t 00

(d) lim 6i (t) exittt

t -*■ 00

(e) lim 6"(t) = 0

t °°

(<$) lim 9"(t) * 0 Figure 12.2.8.

t ■* °°

s. Veduce {,Aom the pAevlout exeAcite that In the catet -g < e < g each

conttant eneAgy cuAve tn the pendulum pAoblem contittt ofa a single

oAbit.

9. Suppose, that Y: M —*- Tm it a smooth vectoAfiletd on the. manifold M

and let a £ M with Y(a) =0 (2.0 that a it an equillbntum point

ofi the vecto Afield). Let c: R —>■ Tm be an Integral cuAve ofi the

vectoAfileld. Show that l/5 c(£0) $ a faoA i,ome tQ £ R then

c(t) f a ion. all t £ R.

10. PAOve the btatmentt made In the text about the oAbitt oft the

pendulum In the cate e = g. (lie the Aetultt ofa ExeAcitet 6 and 9.

12.3. EXISTENCE OF LAGRANGIAN FLOWS

In this section the major result is a theorem which can be used to

establish the completeness of many Lagrangian vectorfields arising from

classical problems such as the simple pendulum, motion on a paraboloid

(see Section 9.1) and the spherical pendulum (dealt with in Chapter 13).

As noted in Section 12.1, completeness of a vectorfield implies the

existence of a flow.
Yl
12.3.1. Theorem. Let Q be a smooth k-dimensional submanifold of R

and suppose Q is a closed subset of Rn. Suppose Y : TQ —> T2Q is a

smooth second-order vectorfield on TQ and that

c: (-6,e) —*■ TQ

is a maximal integral curve for Y with the property llTr2oc(t)H ^ K for

some K and all t £ (-6,c). Then 5 = e = 00.

 136 12. FLOWS

3 I CO
Proof. The first part of the proof involves showing that for tn = e -
the sequence c(.tyf) has an accumulation point in TQ. The proof given by

Chillingworth(1976) page 188 can then be modified to complete the proof. ■

The energy integral of Theorem 8.5.1 can be used to establish

the boundedness condition of Theorem 12.3.1 in problems of classical

mechanics where the potential energy is bounded below.

12.3.2. Theorem. Suppose L : Tq —>- R is a smooth Lagrangian function

where Q is a smooth submanifold of R3n which is closed and let

L = T - V°Tq with T = %XTMX

where M = diagC^, , mlt mz, m2, m2,.,mn, mn, mn) , mi > 0.

If the potential energy map V : Q —* R is bounded below, then

the Lagrangian vectorfield generated by L is complete.

Proof. By Theorem 8.5.1, for each integral curve c and each t € I,

(T + VoxQ)(c(t) = (T + V°tq)(c(0))

= e say.

Thus, if d is a lower bound for V,

(*sXT MX)(c(t)) < e - d

which gives the required result. ■

12.3.3. Corollary. Let L and Q be as in Theorem 12.3.2. and suppose

Q is a closed subset of R3n. Then there exists a flow on TQ generated
by the Lagrangian vectorfield corresponding to L. ■

Finally we state a theorem in the context of a vectorfield on

k
R which tells us that integral curves vary continuously with initial

conditions. This property can readily be extended to vectorfields on

manifolds.

12.3.4. Theorem. Let c: (-e,e) u, d: (-e,e) -> u where U is

an open subset of R^, be integral curves for the vectorfield
F : U —*■ TU. Suppose F satisfies

1It72oF(x) - 7t2oF(z/) U < Kllx - y\\

for some K > 0 and all x, y € U.

 12.3. LAGRANGIAN FLOWS 137

Then for each t 6 (-e,e)

llc(t) - d (t) II < llc(O) - d(0)lleKltl.

Proof. See exercises. ■

EXERCISES 12.3.

Explain why theae exlbtb a Lagnanglan filou) the pnoblemb detctlbed

In Example. 12.2.2 and Section 9.1.

Vnove Theorem 12.3.4 flubt note that c’(t) = F°c(t) and bo we

can apply Theorem 1.5.4 [ll] to bhow that

ft

7T2oFoC.

Wow bhoul that faoA t 6 (0,e)

■t
llc(t) - d(t)ll < llc(O) - d(0)ll + K normo (c-d)
, J o
wheAe "norm" lb given by norm: R*" —*- R: x —> Hall. Then apply
Gnowiall' b Inequality, Lemma 1.5.6, to get the AequAJied nebult. finally
bhou) hou) to extend the aebult to all o£ (-c,e).
 13. THE SPHERICAL PENDULUM

The spherical pendulum is typical of the sorts of problems

which are traditionally studied in Lagrangian mechanics in that it has a
configuration manifold which is 2-dimensional. In addition it has the
very special simplifying property of being "integrable": in suitable charts
Lagrange's equations "uncouple" leading, in effect, to a pair of
1-dimensional problems. A fairly extensive list of integrable problems is
contained in Whittaker(1952) and a recent addition to this list is given
in Gray et_ al_(1982). Our discussion of the spherical pendulum will
illustrate the way in which the preceding theory can be applied to such
problems.

13.1. CIRCULAR ORBITS

A spherical pendulum consists of a particle of mass m constrained to move

under gravity on a sphere, in the absence of friction. The constraint can

be achieved, for example, by attaching the particle to one end of a light

rod while keeping the other end fixed. The sphere on which the particle
moves will be taken as

S2 = {(a,b,c) £ R3: a2 + b2 + c2 = 1}

which is a submanifold of R3. The kinetic and potential energies are

given in terms of the identity chart (x, y, z) on R3 by

T = %m(x2 + y2 + z2)

V = gz .

The Lagrangian L : TS2 —*- R is then the map T - Voi^ restricted to the
domain Ts2, so L is smooth.
 13.1. CIRCULAR ORBITS 139

By Theorem 11.2.3, Lagrange's equations, with respect to the

various charts in any atlas for S2, yield a second-order vectorfield on

TS2 - the Lagrangian vectorfield for the spherical pendulum. Since S2 is

compact, moreover, Theorem 12.3.2. shows that this vectorfield is complete.

Hence it generates a flow on Ts2 by Theorem 12.1.9.

Thus there is a unique integral curve of the vectorfield at

each point of Ts2 with R as domain. The integral of curves, further¬

more, are self-consistent. Since there is no realistic way to sketch

orbits on the 4-dimensional manifold TS2, we shall have to be content to

sketch the projections of these orbits on S2. These projected orbits are

traced out by the base integral curves.

The existence of the following families of orbits is almost

obvious on the basis of physical intuition:

(a) The equilibrium 'points at the north pole N = (0,0,1)

and the south pole S = (0,0,-l). If plaoed at rest

at either of these points, the particle stays there
indefinitely.

(b) Orbits along each meridian} obtained by intersecting

S2 with a vertical plane through the poles. The
motion of the particle along a meridian is the same as
for a simple pendulum.

(c) Orbits around each parallel of latitude below the

equator3 obtained by intersecting S2 with a plane
perpendicular to the polar axis of the sphere. A
particle moving in these orbits is called a "conical
pendulum".

These orbits are illustrated in Figure 13.1.1.

It is very easy to formally establish the existence of these

orbits by using the equivalence we have established between Newton's

second law and Lagrange's equations. To illustrate the method, we prove

the following physically plausible result:

13.1.1. Proposition. Let c: R —*• TS2 be an integral curve for the

Lagrangian vectorfield of the spherical pendulum at the point (a,u) 6 Ts2

where a is either of the two poles. For all t £ R it then follows that
(a) if v = 0 then x 2°c(t) = a

(b) if v i 0 then xs2°c(t) lies in the vertical plane

through the poles containing the vector v £ R3.

 140 13. THE SPHERICAL PENDULUM

Figure 13.1.1. Motion on a meridian and on a parallel of

latitude.

Proof. We leave the proof of part (a) as an exercise and prove part (b).

The theory for the simple pendulum given in Example 12.2.2 may

be adapted to the meridian M as configuration manifold. There is then

an integral curve d: R —► TM at (a,V) for that problem. By Theorem

11.3.2, the reaction force R is orthogonal at each point to the circle M.

It also lies in the plane of M and hence is orthogonal to S2. Thus

the integral curve d satisfies Newton's second law, as formulated in

Section 8.2, for the spherical pendulum problem and hence, by Theorem 8.2.2

and Theorem 11.2.3, d is an integral curve for the Lagrangian vectorfield

of the spherical pendulum at the point (a,V) £ Ts2. By uniqueness, c = d

and so xg2°c maps into the meridian M. ■

EXERCISES 13.1.

1. Paove pcuut [a] o(, paopoAitlon 13.1.1.

2 . Let d: R —*- Ts2 be an IntegAal cuAve &oa the LagAanglan vectoA^letd

ot\le 6ph.eAA.cal pendulum. Shoo) that i{> the. base IntegAal cuAve
t °d pai>i>eA thaough a pole at tome time t0 then the baie tntegnal

cxiAve map6 Into a meAldlan.

(Hint: Put c(t) = d(t + t0) and thou) c 16 an IntegAal cxiAve 0&
 13.2. OTHER ORBITS 141

0A the. vecXorhleld at some point (a,v) o$ the. tout described In

Proposition 13.1.1.)

3. Read about the conical pendulum In Synge 6 Grlhhlthll959), pages

336-337, and then deduce the existence ofi corresponding Integral
curves ior the Lagranglan vectorfileld oh the spherical pendulum.

13.2. OTHER ORBITS, VIA CHARTS

The orbits of the spherical pendulum to be discussed in this section are

not as obvious physically as those introduced in the previous section. To

establish their existence, we shall choose charts for S2 with respect to

which Lagrange's equations "uncouple". Since orbits through the poles of

S2 have already been fully discussed in Section 13.1, moreover, we lose

nothing by choosing charts whose domains exclude these points.

To define the desired charts put

Ux = S2\{(a,i,c) £ S2 : a < 0 and b = 0}

U2 = S2\{(a,b,c) £ S2 : a > 0 and b = 0}

and let the maps 4^, <J>2» z be as shown in Figure 13.2.1 where

(-TT.TT)

2 -+ (0,2tt)

z (-1,1) .
These charts are then (U^, (<j>£, z|u^)) for i = 1,2.

x y

Figure 13.2.1. Charts for S2

142 13. THE SPHERICAL PENDULUM

Note that Ux U U2, which is the domain of z, consists of

the unit sphere in S2 with both poles removed. On the other hand,

U-l fl U2 consists of two open hemispheres and

on one hemisphere

$2
4>, +2tt on the other

and hence we may define a smooth map u) : T(Ui D U2) R by putting

4>x on TU-l

CD .
<j>2 on TU2

In this way we get a pair of maps, z and oo, which have a sort of

global significance throughout the part of the manifold which now concerns

us.

The differential equation introduced in the next proposition

will play a key role in our discussion of the possible orbits for the

spherical pendulum. As usual, we write

z = zoi^2°c, <jK = <Pi°Ts2°c

for the local representatives of an integral curve c.

13.2.1. Proposition. Let c: R —»■ Ts2 he an integral curve for the

Lagrangian vectorfield of the spherical pendulum satisfying the condition

that the corresponding base integral curve never passes through a pole.

There are then numbers e, K £ R, depending on the initial value c(0),

such that

(z')2 = 2g((l-z2)((e/gO-z) - K2/2g0

= foz, say. (1)

Proof. When expressed in terms of the charts the Lagrangian L is given

on the domain U.£ by

• * 2
L = S*n((l-z2)<J>_? H— -) - mgz
^ 1-z2

for i = 1,2. Hence by Theorem 8.2.2 the integral curve satisfies the

Lagrangian equations

((1-z2)
%
= 0 (2)
 13.2. OTHER ORBITS 143

+ Z^')2 + Z&'J2 + g = 0
(3)
1-z2 ' (1-S2)2 -

(It is perhaps reassuring to note that our assumptions ensure that z2

can never assume the value 1.) From (2) it follows that there is a
constant K £ R such that

(1 - z2)^' = K (4)

Since and <p2 are restrictions of the smooth map u) it follows,

moreover, that K is the same for both choices of i. We may assume,

furthermore, that K ^ 0 since otherwise (4) would imply that (J) • r = 0

'Is -
and hence the base integral curve would be an orbit along a meridian,
contrary to hypothesis.

Now since, by Theorem 8.5.1, the total energy of the system is

conserved, there is an e £ R such that

(1-z2) ($■£ ')2 + ~Z J ^ + gz = e (5)

v 1-S2 /

and, moreover, e is independent of i. Solving (4) for and then

substituting in (5) gives, after some manipulation, the required equation

(1). ■

13.2.2. Proposition. The function f appearing in the previous proposition

has a graph which looks like one of those in Figure 12.2.2. In particular,
f must have three zeros alt a2, a3 such that

1 < ax < a2 < 1 < a3 .

Both of the cases = a2 and < a2 actually occur, for suitable choice
of the initial value c(0) for the integral curve.

Figure 13. 2.2.

144 13. THE SPHERICAL PENDULUM

Proof. Note first that, since K f 0, f(-l) < 0 and f(1) < 0. It is

clear also that

lim f (a) = and lim f (a) = 00 .

ct->- -00 a-*00

For motion to be at all possible, we must have f«z ^ 0_ and hence there

must be at least one a £ (-1,1) such that f(a) > 0. Hence the graph of

f must be of one of the possible types shown.

To show that the case < cx2 can be realized by suitable

choice of initial conditions note that from (1)

f(0) = 2e - K2 .

But it is clear from (4) and (5) that we can keep K fixed while making

e arbitrarily large by suitably choosing the initial conditions. This

gives f(0) > 0 and hence a3 < a2.

Our argument that the case = a2 can actually occur is

indirect: from Section 13.1 there are orbits of the "conical pendulum"

type; as shown later, these are not the sort of orbits obtained in the

case ax < a2; hence they must belong to the case 0^ = 0t2. ■

The following result completes our discussion of the possible

types of orbits for the spherical pendulum.

13.2.3. Proposition. Suppose that the numbers a1 and a2 in the

previous proposition satisfy the condition ot1 < a2. There is then a

family of orbits which oscillate between the parallels of latitude

("apsidal circles") on which the z-coordinate assumes the values a3and a2

respectively, as in Figure 13.2.3.

Proof. If z lies between a3 and cx2 then z' cannot be zero since

f°z >0 on (o^, a2). Thus z must increase or decrease monotonically

in this range. Now suppose z(0) = a3. Thus z'(0) =0. We need to

show that z"(0) / 0 for if this is not the case the particle would

remain at z = a3.

From Lagrange's equations (3) we obtain the expression for

z"(0) subject to our initial conditions

ct K2
-„(0) = _i-+ (a,2 - 1)g . (6)
ai 2-l

Now f(a3) =0 so from (1) we obtain

 13.2. OTHER ORBITS 145

20(cxi - f )
a12-l

and on substituting this into (6) we obtain

z"(0) = g-(2a1(a1 - - ) + (o^2-!)). (7)

Figure 13.2.3. Oscillatory base integral curve.

It is easy to see that the right hand side of (7) is just ^'(04). But

ff (cx^) cannot be zero otherwise ot2 would be a double root of f which

would contradict our assumption that ax < a2 < 1 < a3. Thus z"(0) is

non-zero so that particle cannot remain at z = .

Now suppose we have initial conditions given by

ax < z(s) < a2, z'(s) > 0.

Suppose by way of contradiction that z does not reach z = a2 in finite

time. Thus z(£) < ot2 whenever t > S. We deduced above that z

increases monotonically in this range if z' > 0 initially. So here we

have

z' = /fog t 0 on ( alta2).

146 13. THE SPHERICAL PENDULUM

Thus
z'
t - s
/ foz

and so zXt)
dx
= t (8)
z(s)
Jig(x-ax) (x-a2) (x-a3)
The right hand side of (8) can be made arbitrarily large but the left hand
side is bounded above by
J rv

(9)
Jlg(x-ax)(x-a2)(x-a3)
ai

which can be shown to be finite. Thus z reaches a2 in finite time.

The argument which shows z reaches ax is the same and the expression (9)

gives the half-period of the motion between ax and a2. ■

EXERCISE 13.2.

A&Aume. -1 < ax < a2 < 1 < a3 and that

f(a) = 2g(a-a1)(a-a2)(a-a3)

■ r> - 0
U4e elementary algebra to thou) that ax + a2 < 0 and ute tkii to t>hou)
that the. mean o{ the two "apiidal circlet" ihown in Figure 13.2.3 it
below the x-y plane. What doet that tell ai about the "conical pendulum"
bate integral curvet?
 14. RIGID BODY MOTION

In section 8.4 it was shown that the configuration set for the

motion of a rigid rod was a 5-dimensional submanifold of R6 and

furthermore that the reaction (or "internal") forces were orthogonal to

this submanifold. Thus the problem could be treated as one of a single

particle moving on a 5-dimensional submanifold of R6.

In this chapter the motion of a rigid body moving in R3 is

considered and it is shown that the configuration space can be thought of

as R3 x SO(3). Thus this problem reduces essentially to the motion of a

single particle on a 6-dimensional configuration manifold. Furthermore the

reaction forces are shown to be orthogonal to the configuration manifold.

The derivation of Lagrange's equations given in Section 8.3. thus applies

to rigid body motion when the "external" forces are due to a conservative

field of force.

14.1. MOTION OF A LAMINA.

In section 8.4. we showed that Lagrange's equations were applicable to the

problem of the motion of a rigid rod. There the "constraint condition"

that the rod's length remained constant was sufficient to define the

configuration manifold. By way of introduction to the problem of the rigid

body we look at the motion on a plane of a "lamina" which we think of as

being defined by three "point masses". Its configuration set will clearly

be a subset of R6 and we will again require that the distances between

the particles remain constant. These three constraints, however, do not

tell the whole story.

In Figure 14.1.1. we have two configurations which satisfy the

same distance constraint conditions but if our lamina cannot "flip (which

is certainly the case because it is constrained to move in a plane) then

these configurations cannot both be on the same configuration manifold.

148 14. RIGID BODY MOTION

Figure 14.1.1.

Thus there must be another constraint which implies the preservation of the

"orientation" of the lamina. This is obtained by the use of the vector or

cross-product:

(b1-a1, b2-a2, 0) x (b1-o1, b2-e2, 0)

which defines a vector normal to the plane in which the lamina moves. We

will insist on this cross product also remaining constant.

The "distance preservation" constraint is

1^1>^2) =: 0 where

ftel ,a2, b 1 ,b 2 ,<2i > c2) = te.u ,w)

with

u = Q>i - ai)2 + tea - a2)2 -d

(asb)

V = tei " ai)2 + te 2 - a2)2 -d 2

1(a,c)

w - tei - Ci)2 + tea - o2)z -d 2

(e,b)

Thus our configuration set Q satisfies Q c: f—1 (0) . It turns out that

f_1(0) is a three dimensional submanifold of R6 which consists of the

union of two disjoint open subsets, one of which is Q. The details are
left as exercises.

An alternative way of describing the configuration manifold Q,

which uses rotation matrices, is hinted at in the exercises. This

alternative approach will be exploited in the next section where a direct

description of the configuration manifold of a rigid body via constraints

on distance and orientations becomes unweildy.

 14.1. MOTION OF A LAMINA 149

EXERCISES 14.1.
Read youA solutions to ExeAcises 3.3.2. and 3.3.3. and then pAove
the statement!) made tn the text conceAning the sets q and f—1 (0).

ConsideA the following [distance pA.eieA.vtng) tAans^oAmation ofa a

lamina which consists o^ an anticlockwise notation thAough an angle
0 about one o^ the ventices, which we place at the oAigin as in
FiguAe 14.1.2.

Figure 14.1.2.

(a) Find a matAix A Such that

K
= A
V
K \

ci
t = A
°2\ C2_

(b) Vo these conditions detenmine the matAix A uniquely?

(c) Show that det(A) = 1.

(d) Voes this toonsfaoAmotion poeseove oAientation?

Now consideA the (distance pA.eAeA.ving) toons (>oAmation o& a lamina

which consists ofa a oefilection about one o<J its Aides, assumed to
lie on the fiiAAt axis, as in FiguAe 14.1.3.
150 14. RIGID BODY MOTION

(b1,-b2)

Figure 14.1.3.

[a] Find the. matnix ofa this t/uwA^onmation and. xXa deietminant.

(b) Voza thxA tnanAfioAmation pAzAZAve onizntation?

14.2. THE CONFIGURATION MANIFOLD OF A RIGID BODY

Our mathematical model of a rigid body is a collection of n (> 4)

particles in R3, at least four of which are not coplanar. They are to

be constrained in such a way that the distances between, and the relative

orientation of, the particles remain constant throughout the motion. We

regard these quantities as being preserved by means of rigid rods, of

negligible mass, connecting each pair of particles.

To help us define the configuration set of the rigid body

precisely, the following definition is introduced.

14.2.1. Definition. A rigid motion is a map p : R3 —y R3 which preserves

(a) distances between pairs of points in R3,

(b) the orientation of triples of vectors in R3. ■

Algebraically, this definition means that the map p is to satisfy the

conditions that

(a) llp(a)-p(£>) II = IIa - b\\

(b) det(p(a),p(£),p(c)) = det(a,b,c)

for all a, b and a in R3 where II II is the usual norm on R3.

The configuration set can now be defined as follows.

 14.2. CONFIGURATION MANIFOLD 151

14.2.2. Definition. The configuration set Q of a rigid body is the set

of n- tuples of vectors from R3,

{ (p(a1), p(a2).p(an)): p is a rigid motion} ,

where a , a .. are the respective initial positions in R3 of the

n particles constituting the rigid body. ■

We now aim at showing that the configuration set is a sub¬

manifold of R3n. Intuitively we can think of the configuration set of a

rigid body as consisting of a combination of translations of one of its

points together with rotations about axes passing through this point, as
illustrated in Figure 14.2.1.

s, \
translation-► |-rotation-»■

THIS THIS

SIDE SIDE

UP UP

rigid motion

Figure 14.2.1. A rigid motion.

Thus it would seem that the configuration set may well look like

{translations in R3} x {rotations in R3},

which can be represented as the manifold R3 x S0(3). Our task then is to

show that this is so. The key result is as follows:

14.2.3. Theorem. A map p: R3 —>■ R3 is a rigid motion if and only if

there is some c £ R3 and A £ SO(3) such that, for all a £ R3,

p(a) = c + Aa.
152 14. RIGID BODY MOTION

Proof. It is rather long and technical and appears at the end of this
section. ■

In the following lemma, the element A £ R9 is to be regarded

as a 3x3 matrix.

14.2.4. Lemma. Let L : R3 x R3x3 —* (R3)n with

L(c, A) = (c + Aa1, 0 + Aa2.a + kan)

where n ^ 4 ccnd a1, a2,... ,an are the position vectors of points in

R3 not all lying in the same plane. The map L is then linear and one-
to-one.

Proof. This is left as an exercise. ■

By using the above map L we can easily derive the following

theorem, which is the main result of this chapter.

14.2.5. Theorem. The configuration set Q for a rigid body consisting

of n particles is a submanifold of R3n diffeomorphic to the

6-dimensional manifold R3 x SO (3).

Proof. The idea is that the one-to-one linear map

L ; R3 x R3X3 R3n

gives a diffeomorphism

L' : R3 x SO(3) -> Q

when restricted to the set R3 x S0(3), as illustration in Figure 14.2.2.

Figure 14.2.2.
 14.2. CONFIGURATION MANIFOLD 153

The remaining details of the proof are left as an exercise. ■

We now go back and give a proof of Theorem 14.2.3, thereby

completing this section.

Proof of Theorem 14.2.3. First, assume p : R3 —R3 is given by

p(a) = c + Aa

where e £ R3 and A £ SO(3). It follows from Lemmas 3.4.3. and 3.4.5.

that p preserves distances between points and orientation of triples of

vectors. Hence p is a rigid motion.

Conversely, assume p : R3 —s- R3 is a rigid motion. Let

L = p - p(0). It is easily checked that L preserves distances and

orientations and that L(0) = 0. Thus, for each a £ R3,

IIL(a) II = Hall

and since, for all a,b £ R3,

lla-ill2 = Hall2 - 2a.b + ll£>ll2

we have llL(a)2ll-2L(a)-L(6) + llL(&) II2 = Hall2 - 2a.b + \\b\\2

and so L(a)-L(i) = a-b .

Now let u^ = L(e^) for i = 1,2, . Then for each i ,j, llw^ll = 1
and Uj = and so } is an orthonormal basis for R3.
ui-

Now for each i = 1,2, and a,b £ R3,

L(a) .U£ = a- e.£ = a^

L (b).U£ = b'&i = bi

and so

(L(a + b) - L(a) - L{b))-u^ = 0

which gives

L(a + b) = L(a) + L(b).

Similarly L(Xa) = AL(a) for each X £ R. Thus L is linear and can be

represented by some 3 x 3 matrix A whose columns are u1,uz,u3. Thus

T T
(A A). . = u.’Uj and so A A = I.
i0 % 3
Now since L preserves orientations we must have, for three linearly

independent vectors (ax, a2, a3), (b1, b2, b3) and (cx, o2 , c3),
154 14. RIGID BODY MOTION

( f
ai V °1 ax bx o1
sgn A a2 A bz = sgn
C2 a2 b2 °2
C
a3 3
a
3
2>
3 3
- .

Now letting
a1 b1 ox

&2 bz ‘-'2

<23 C3

we have

sgn(|AZ|) = sgn(|A|) sgn(|z|) = sgn(|z|).

Thus sgn(|A|) = 1 since sgn(|z|) ^0 and so A € SO (3). But

p(a) = p(0) + L(a) = p(0) + Aa

which shows that the map p has the desired form. ■

EXERCISES 14.2.

1. Skou) that lh pj^ : R3 —>- R3 and p2 : R3 —>- R3 axe xigld motion*

then <60 1* the composite px°p2 : R3 —R3.

2. Show that lh an n-tuple oh vectox* (b1, b2,...,bn) h^om R3

belong* to the conhlguxatlon *et oh a xigld body then the
conhlguxatlon *et may be written a*

{(p(b1), p(b2),..., p(bn): p is a rigid motion}.

3. Pxove Lemma 14.2.4. To *how L l* one-to-one, u*e the hac^- that lh

houA element* oh R3, *ay a1, a2, a3, a4, axe the po*ltlon vectox*
oh houx non-coplanax point* then the dlhhzxencex a2-a2, a2-a3 , a3-a1
hoxm a llneaxly Independent *et oh vectox*.

4. Complete the pxooh oh Theoxem 14.2.5. by u*lng Lemma 4.3.4.

14.3. ORTHOGONALITY OF REACTION FORCES

The methods employed here to study the reaction forces within a rigid body

are basically the same as those used to study the corresponding problem for

the rigid rod in Section 8.4. Here, however, the computational details

are more complicated and it will be convenient to introduce some extra

notation to describe the constraints between the n particles

constituting the rigid body.
 14.3. REACTION FORCES 155

To this end we write a typical element a £ R3n in the form

a = (a1, a2,...,an)

where,for 1 < i < n, a1 is an element of R3 which we write as

i , i i is
a = (a1, a2, a3).

We think of o' as the point occupied by the i^ particle in R3. For

1 < i < j < n, furthermore, let d. . denote the distance between the
^3
i^h and jth particles and let

f . . : R3n r with

f (a) = (a^-a^)2 + (a\-a{)2 + (a\-a^)2 - d (1)

"Z- 3 L 1 ^ 0 3- 2. j

w ___ "I \

Finally, let f: R3n —*- R2 be the map whose respective real-valued

components are the %rc(n-1) maps

f12 » f13’’

f2 3 » ' >f 2n

f nn-i •

Since rigid motions preserve distance, it follows that the configuration

manifold Q of the rigid body satisfies

Q <= f-1 (0) . (2)

In the following geometrical lemma, notation is used for the

components of h £ R3n which is analogous to that used for a above.

14.3.1. Lemma. If (a,h) £ TQ then} for 1 < i < j < n,

(a^ - aP)‘h^ = (a^ - a.3 )'h?

Proof. Assume ([a,h) £ TQ so that h = y'(0) for some smooth

parametrized curve y in Q at a, by Definition 5.1.2. But by (2),

for each t in an interval containing 0,

(f °y) (t) = 0

and hence by the chain rule

Df (a)(h) = Df (y(0)) (y' (0)) = D(f = y)(t)(l) = 0.

156 14. RIGID BODY MOTION

In terms of the components of f this means that, for 1 < i < j < n,

Df • j (a) W) = 0
I'd

The explicit form (1) of the map f. . now gives the result. ■

As to the reaction forces between the particles constituting

the rigid body, observe that the force exerted on the particle by

the (1 ^ i ^ n, will lie along the line joining these

particles, as in Figure 14.3.1, and hence have the form

{a y {a? - a}'))

for some X^j € R. The reactions being equal in magnitude but opposite in
direction, moreover, it follows that

(3)

aC

Figure 14.3.1. Equal and opposite reaction.

Now the total reaction force R2- on the £th particle will be the sum

of such forces exerted by each of the other n-1 particles. Hence

R* ~Ui, 2 A# (rf'-aM (4)

V C=1 '
C+1
Hence, if we regard the rigid body as a single particle moving on the

configuration manifold Q then the total reaction force acting on this

particle at the point a. 6 Q is given by

R = (a, (tt2 (r1).7r2(Rn) )) (5)

We can now state and prove the desired orthogonality result.

 14.3. REACTION FORCES 157

14.3.2. Theorem. The reaction force R acting at any point of the

configuration manifold Q is orthogonal to Q at the point.

Proof. Let (a,?z) £ Tq.

(a,R) • (a,h) = n2(R1)‘h1 + ... + ir2(Rn)-hn by (3)

= E E A • • (aP - aV)-hl by (4)

i=1 J=1
qH

= E Ai<7- ((aJ'-aV^ - (J-a)'h^) by (3)

•£,<7=1
i<3

=0 by Lemma 14.3.1. ■
 REFERENCES

Abraham, R. & Marsden, J.E. (1978). Foundations of mechanics. (Second

Edition). Reading, Mass.: Benjamin-Cummings.

Arnold, V.I. (1978). Mathematical methods of classical mechanics. New

York: Springer Verlag.

Birkhoff, G.D. (1927). Dynamical systems. New York: American

Mathematical Society.

Chillingworth, D. (1976). Differential topology with a view to

applications. London: Pitman.

Courant, R. & John, J. (1974). Introduction to calculus and analysis

volume 2. New York: Wiley

Dieudonne, J. (1960). Foundations of analysis. New York and London:

Academic Press.

Dieudonne, J. (1972). Treatise on analysis. New York and London:

Academic Press.

Goldstein, H. (1980). Classical mechanics. Reading, Mass.: Addison-

Wesley.

Gray, A., Jones, A. & Rimmer, R. (1982). Motion under gravity on a

paraboloid. J. Diff. Equations, 4^5, 168-181.

Gray, A. (1985). Motion under gravity on a saddle. J. Diff. Equations,

57, 248-257.

Guillemin, V. & Pollack, A. (1974). Differential topology. New Jersey:

Prentice-Hall.

Hirsch, M.W. (1976). Differential topology. Graduate texts in

Mathematics, Volume 33. New York: Springer.

Irwin, M.C. (1980). Smooth dynamical systems. New York and London:

Academic Press.

Landau, L.D. & Lifshitz, E.M. (1960). Mechanics. Oxford: Permagon.

Lang, S. (1964). A second course in calculus. Reading, Mass.:

Addison-Wesley.

Loomis, L.H. & Sternberg, S. (1968). Advanced Calculus. Reading,

Mass.: Addison-Wesley.
 REFERENCES 160

Munroe, M.E. (1963). Modern multidimensional calculus. Reading, Mass.:

Addison-Wesley.

Nering, E.D. (1964). Linear algebra and matrix theory. New York and
London: Wiley.

Simmons, G.F. (1963). Introduction to topology and modern analysis.

New York: McGraw-Hill.

Spivak, M. (1965). Calculus on manifolds. Reading, Mass.: Benjamin-

Cummings.

Spivak, M. (1970). Differential geometry. Volume 1, Boston, Mass.:

Publish or Perish.

Stern, R.J. (1983). Instantons and the topology of 4-manifolds. The

Mathematical Intelligencer, _5, 39-44.

Synge, J.L. & Griffiths, B.A. (1959). Principles of Mechanics. New York:
McGraw-Hill.

Whittaker, E.T. (1952). A treatise on the analytical dynamics of particles

and rigid bodies. Cambridge: Cambridge University Press.

 INDEX

Admissible chart, definition of: 24

Angular co-ordinates: 18
Angular velocity map: 129
Apsidal circles: 146
Atlas, definition of: 18
Base integral curve, definition of: 107
Basis for a tangent space: 117
Bilinear form: 93
Cancellation of dots rule: 74
Chain rule: 5, 12, 13, 51, 69
Chain rule, classical: 70
Chain rule for partial derivatives: 10, 12
Chain rule, semi-classical: 69
Chart for TM, theorem on: 58
Chart, definition of: 16
Classical chain rule: 70
Compatible, definition of: 18
Complete vectorfield, definition of: 121
Configuration manifold, definition of: 84
Configuration set, definition of: 84
Conical pendulum: 139, 141, 146
Connected manifolds: 27
Conservation of energy: 88
Conservation of energy, theorem on: 88
Conservative field of force, definition of: 80
Conservative, definition of: 84
Constant energy curve: 133
Constant mapping: 13
Constant mapping, definition of: 5
Constraints: 83
Coordinate functions: 68
Coordinate system: 18
Curve, definition of: 44
Derivative: 7
Derivative of g with respect to <J>£, definition of: 69
Derivative, total: 71
Diffeomorphism: 43
Diffeomorphic, definition of: 42
Diffeomorphism, definition of: 42
Diffeomorphism on a manifold: 122
Differentiable: 45
Differentiable, definition of: 4, 40
Differentiable on a subset, definition of: 40
Differentiable manifold, definition of: 24
Differentiable structure, definition of: 24
Differential: 71
Dimension of a manifold: 25
Dimension of TM, corollary on: 57
Dimension of TM, theorem on: 46
Dimension of TM, lemma on: 63
Dot notation: 72
Dot product: 82
162 INDEX

Double tangents: 62
Energy function: 88
Energy, kinetic: 79
Energy, potential: 79
Equivalence of atlases, definition of: 24
Euler's relation: 77
Existence-uniqueness theorem: 101, 103
Field of force, definition of: 84
Flow, definition of: 123
Flow, theorem on: 124
Force: 78
Force, conservative: 80
Force, reaction: 81
Frechet derivative, definition of: 4
Functor: 9
Fundamental theorem of calculus: 13
Global Lagrangian vectorfield: 116
Gronwall's inequality: 14
Group homomorphism: 123
Homogeneous map, definition of: 77
Identity chart: 95
Identity mapping: 13
Identity mapping, definition of: 5
Implicit function theorem: 32
Indefinite integral, definition of: 13
Integral curve: 125
Integral curve, definition of: 100
Interchange of dot and dash: 74, 75
Jacobian matrix of, theorem on: 12
Kinetic energy: 79, 82
Kinetic energy, definition of: 84
Lagrange's equations: 82
Lagrange's equations, theorem on: 85, 96
Lagrangian flows, existence of: 135
Lagrangian function: 82
Lagrangian function, definition of: 85
Lagrangian vectorfield: 116
Lagrangian vectorfield, theorem on: 116
Lamina: 147
Local representative: 62
Local representative, definition of: 39
Matrix of a quadratic form, definition of:
Maximal atlas: 25
Maximal atlas, definition of: 24
Maximal integral curve, corollary on: 116
Maximal integral curve, definition of: 103
Motion on a paraboloid: 90
Motion under gravity: 126, 127
Natural atlas, definition of: 60
Natural projection, definition of: 60
Natural projection, theorem on: 61
Newton's laws: 79
Newton's second law: 117
Newton's second law, theorem on: 118
One-parameter group: 123
Open, definition of: 26
 INDEX 163

Orbit: 125
Orbit of a point, definition of: 124
Orientation: 148
Orientation, definition of: 37
Orthogonal: 81
Orthogonal, definition of: 82
Orthogonal matrices: 35
Overlap Map: 17
Paraboloid, motion on: 90
Parallelizable, definition of: 59
Parametrized curve: 65
Parametrized curve, definition of: 44
Parametrized curves, proposition on: 51
Partial derivative, definition of: 10, 68, 70
Partial derivatives of real valued functions: 11
Partial differentiation: 67
Patchable, definition of: 110
Patchable, lemma on: 111
Patchable, theorem on: 113
Paths in TQ: 64
Positive definite: 93
Potential energy: 79
Potential function, definition of: 84
Product manifold: 25
Projection function: 10
Quadratic form: 93, 94
Quadratic form, theorem on: 95
Reaction force: 81, 86
Rigid body: 43
Rigid body, definition of: 150
Rigid rod: 86
Rigid motion, definition of: 150
Rigid motion, theorem on: 151
Rotation matrices: 35
Rotations: 36
Second-order vectorfield, definition of: 105
Second-order vectorfields: 106, 108
Self-consistent, definition of: 65
Self-consistent integral curves: 108
Semi-classical chain rule: 69
Simple pendulum: 127
Simple pendulum, orbits for: 132
Simple pendulum, vectorfield for: 131
Slope, definition of: 11
Smooth manifold, definition of: 24
Spherical polar co-ordinates: 23
Submanifold charts: 56
Submanifold, definition of: 29
Submanifold property, definition of: 29
Submanifold property, lemma on: 56
Submanifold test: 33
Tangent bundle, definition of: 7, 46
Tangent bundle: 55
Tangent functor, definition of: 8
Tangent map: 7, 50
Tangent map, definition of: 8, 49
164 INDEX

Tangent map, theorem on: 51

Tangent space, definition of: 6, 46
Tangent space, theorem on: 46, 53
Tangent vector, definition of: 45
Time evolution operator: 121, 125
Total derivative: 71
Total energy function, definition of: 88
Transition map: 17
Unit sphere: 21, 22
Variable-free notation: 5, 13
Vectorfield, definition of: 99
Vectorfield, Lagrangian: 116
Velocity phase space, definition of: 84
Virtual work: 78
 SYMBOL TABLE

at transpose of the matrix A

D empty set
fog composite of f with g, i.e. (fog)(x) = f(g(x))

C(E,F) set of linear maps from E to F

N natural numbers {1,2,3,... }

Rn
euclidean n-space

R+ strictly positive real numbers

llxll norm of x in

X identity map on Rn

Df (a) Frechet derivative of f at a : 4

identity function on E : 5
ldE
a constant mapping : 5

T E tangent space of E at a : 6
a
3.f partial Frechet derivative of f with respect to the -ith
V
variable : 10

id. jth projection function : 10

3
f '(a) jacobian matrix of f at a : 11

fH slope of f in the ith direction : 11

(fl.f2,-.,fm)components of vector valued function f : 12

indefinite integral of f from a : 13

, , . nn\-1
sn unit sphere m K : 21

e5 r-times continuously differentiable : 24

n natural projection map : 29

0(n) nxn orthogonal matrices : 35

S0(n) nxn rotation matrices : 35

tangent vector at a : 45

T M tangent space of M at a : 46
a
TM tangent bundle of M : 46

Tf tangent map of f : 49

natural projection from the tangent bundle : 60

tm

T2M double tangent bundle of M : 62

c5(t) 65
166 SYMBOL TABLE

9f
partial derivative of f with respect to d>. : 68
34>.
T,
3f
partial derivative of f with respect to : 70
dip

f f dot : 72

(4>°TQ,<j>) natural chart for TQ : 73

local representative of f with respect to charts <(),^ : 62

f#
CO
Diff°°(M) set of all C diffeomorphisms from M to M : 122
A flow : 123
688059
DATE DUE
QA 614.3 .J66 1987
Jones, Arthur, 1934-
010101 000
anc* mechanics / Arth

63 0024525 9
TRENT UNIVERSITY
 This book provides an easy introduction to the theory of
differentiable manifolds. Theauthors then show howthe theory can
be used to develop, simply but rigorously, the theory of Lagrangian
mechanicsdirectlyfrom Newton's laws. Unnecessary abstraction has
been avoided to produce an account suitable for students in
mathematics or physics who have taken courses in advanced
calculus. Clearly but intuitively defined geometric notions have
replaced those concepts of Lagrangian mechanics that are usually
vaguely defined. Consequently, the book is also a thorough but basic
account of modern mechanics; at the same time applications of the
theory to some classical problems are also given.

AUSTRALIAN MATHEMATICAL SOCIETY LECTURE SERIES

Editor-in-Chief: Dr. S.A. Morris, Department of Mathematics, LaTrobe

University, Bundoora, Victoria3083, Australia

Editors:
Professor C.}. Thompson, Department of Mathematics, University of
Melbourne, Parkville, Victoria3052, Australia
Professor C.C. Heyde, Department of Statistics, University of Melbourne,
Parkville, Victoria 3052, Australia
Professor J.H. Loxton, Department of Pure Mathematics, University of New
South Wales, Kensington, New South Wales 2033, Australia

The Australian Mathematical Society Lecture Series is intended to operate at

the frontiers of mathematics itself and of its teaching, and therefore will
contain both research monographs and textbooks suitable for graduate or
undergraduate students.

Cover design: Ken Vail

Cambridge University Press

ISBN 0-521-33b50-3