This action might not be possible to undo. Are you sure you want to continue?

# PHYS3001

Classical Mechanics

Robert L. Dewar

Department of Theoretical Physics

Research School of Physical Sciences & Engineering

The Australian National University

Canberra ACT 0200

Australia

robert.dewar@anu.edu.au

Version 1.51

May 20, 2001. c ( R.L. Dewar 1998–2001.

ii

Contents

1 Generalized Kinematics 1

1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

1.2 Generalized coordinates . . . . . . . . . . . . . . . . . . . . . 2

1.3 Example: The ideal ﬂuid . . . . . . . . . . . . . . . . . . . . . 5

1.4 Variational Calculus . . . . . . . . . . . . . . . . . . . . . . . 6

1.4.1 Example: Geodesics . . . . . . . . . . . . . . . . . . . 9

1.4.2 Trial function method . . . . . . . . . . . . . . . . . . 10

1.5 Constrained variation: Lagrange multipliers . . . . . . . . . . 10

1.6 Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

1.6.1 Rigid rod . . . . . . . . . . . . . . . . . . . . . . . . . 13

1.6.2 Ecliptic . . . . . . . . . . . . . . . . . . . . . . . . . . 13

1.6.3 Curvature of geodesics . . . . . . . . . . . . . . . . . . 13

2 Lagrangian Mechanics 15

2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

2.2 Generalized Newton’s 2nd Law . . . . . . . . . . . . . . . . . 16

2.2.1 Generalized force . . . . . . . . . . . . . . . . . . . . . 16

2.2.2 Generalized equation of motion . . . . . . . . . . . . . 20

2.2.3 Example: Motion in Cartesian coordinates . . . . . . . 21

2.3 Lagrange’s equations (scalar potential case) . . . . . . . . . . 21

2.3.1 Hamilton’s Principle . . . . . . . . . . . . . . . . . . . 23

2.4 Lagrangians for some Physical Systems . . . . . . . . . . . . . 24

2.4.1 Example 1: 1-D motion—the pendulum . . . . . . . . 24

2.4.2 Example 2: 2-D motion in a central potential . . . . . 25

2.4.3 Example 3: 2-D motion with time-varying constraint . 26

2.4.4 Example 4: Atwood’s machine . . . . . . . . . . . . . . 27

2.4.5 Example 5: Particle in e.m. ﬁeld . . . . . . . . . . . . 28

2.4.6 Example 6: Particle in ideal ﬂuid . . . . . . . . . . . . 29

2.5 Averaged Lagrangian . . . . . . . . . . . . . . . . . . . . . . . 30

2.5.1 Example: Harmonic oscillator . . . . . . . . . . . . . . 31

2.6 Transformations of the Lagrangian . . . . . . . . . . . . . . . 32

iii

iv CONTENTS

2.6.1 Point transformations . . . . . . . . . . . . . . . . . . . 32

2.6.2 Gauge transformations . . . . . . . . . . . . . . . . . . 34

2.7 Symmetries and Noether’s theorem . . . . . . . . . . . . . . . 35

2.7.1 Time symmetry . . . . . . . . . . . . . . . . . . . . . . 37

2.8 Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

2.8.1 Coriolis force and cyclotron motion . . . . . . . . . . . 37

2.8.2 Anharmonic oscillator . . . . . . . . . . . . . . . . . . 38

3 Hamiltonian Mechanics 41

3.1 Introduction: Dynamical systems . . . . . . . . . . . . . . . . 41

3.2 Mechanics as a dynamical system . . . . . . . . . . . . . . . . 41

3.2.1 Lagrangian method . . . . . . . . . . . . . . . . . . . . 41

3.2.2 Hamiltonian method . . . . . . . . . . . . . . . . . . . 43

3.2.3 Example 1: Scalar potential . . . . . . . . . . . . . . . 45

3.2.4 Example 2: Physical pendulum . . . . . . . . . . . . . 47

3.2.5 Example 3: Motion in e.m. potentials . . . . . . . . . . 48

3.2.6 Example 4: The generalized N-body system . . . . . . 48

3.3 Time-Dependent and Autonomous Hamiltonian systems . . . 50

3.4 Hamilton’s Principle in phase space . . . . . . . . . . . . . . . 50

3.5 Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

3.5.1 Constraints and moving coordinates . . . . . . . . . . . 53

3.5.2 Anharmonic oscillator phase space . . . . . . . . . . . 53

3.5.3 2-D motion in a magnetic ﬁeld . . . . . . . . . . . . . . 53

4 Canonical transformations 55

4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55

4.2 Generating functions . . . . . . . . . . . . . . . . . . . . . . . 56

4.2.1 Example 1: Adiabatic Oscillator . . . . . . . . . . . . . 60

4.2.2 Example 2: Point transformations . . . . . . . . . . . . 62

4.3 Inﬁnitesimal canonical transformations . . . . . . . . . . . . . 63

4.3.1 Time evolution . . . . . . . . . . . . . . . . . . . . . . 64

4.4 Poisson brackets . . . . . . . . . . . . . . . . . . . . . . . . . . 65

4.4.1 Symmetries and integrals of motion . . . . . . . . . . . 66

4.4.2 Perturbation theory . . . . . . . . . . . . . . . . . . . . 66

4.5 Action-Angle Variables . . . . . . . . . . . . . . . . . . . . . . 67

4.6 Properties of canonical transformations . . . . . . . . . . . . . 69

4.6.1 Preservation of phase-space volume . . . . . . . . . . . 69

4.6.2 Transformation of Poisson brackets . . . . . . . . . . . 74

4.7 Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 74

4.7.1 Coriolis yet again . . . . . . . . . . . . . . . . . . . . . 74

4.7.2 Diﬀerence approximations . . . . . . . . . . . . . . . . 75

CONTENTS v

5 Answers to Problems 77

5.1 Chapter 1 Problems . . . . . . . . . . . . . . . . . . . . . . . . 77

5.2 Chapter 2 Problems . . . . . . . . . . . . . . . . . . . . . . . . 80

5.3 Chapter 3 Problems . . . . . . . . . . . . . . . . . . . . . . . . 86

5.4 Chapter 4 Problems . . . . . . . . . . . . . . . . . . . . . . . . 93

6 References and Index 99

vi CONTENTS

Chapter 1

Generalized coordinates and

variational principles

1.1 Introduction

In elementary physics courses you were introduced to the basic ideas of New-

tonian mechanics via concrete examples, such as motion of a particle in a

gravitational potential, the simple harmonic oscillator etc. In this course we

will develop a more abstract viewpoint in which one thinks of the dynamics of

a system described by an arbitrary number of generalized coordinates, but in

which the dynamics can be nonetheless encapsulated in a single scalar func-

tion: the Lagrangian, named after the French mathematician Joseph Louis

Lagrange (1736–1813), or the Hamiltonian, named after the Irish mathe-

matician Sir William Rowan Hamilton (1805–1865).

This abstract viewpoint is enormously powerful and underpins quantum

mechanics and modern nonlinear dynamics. It may or may not be more ef-

ﬁcient than elementary approaches for solving simple problems, but in order

to feel comfortable with the formalism it is very instructive to do some ele-

mentary problems using abstract methods. Thus we will be revisiting such

examples as the harmonic oscillator and the pendulum, but when examples

are set in this course please remember that you are expected to use the ap-

proaches covered in the course rather than fall back on the methods you

learnt in First Year.

In the following notes the convention will be used of italicizing the ﬁrst use

or deﬁnition of a concept. The index can be used to locate these deﬁnitions

and the subsequent occurrences of these words.

The present chapter is essentially geometric. It is concerned with the de-

scription of possible motions of general systems rather than how to calculate

1

2CHAPTER 1. GENERALIZEDCOORDINATES ANDVARIATIONAL PRINCIPLES

physical motions from knowledge of forces. Thus we call the topic generalized

kinematics.

1.2 Generalized coordinates

Suppose we have a system of N particles each moving in 3-space and in-

teracting through arbitrary (ﬁnite) forces, then the dynamics of the total

system is described by the motion of a point q ≡ ¦q

i

[i = 1, 2, . . . , 3N¦ =

¦x

1

, y

1

, z

1

, x

2

, y

2

, z

2

, . . . , x

N

, y

N

, z

N

¦ in a 3N-dimensional generalized conﬁg-

uration space. The number n = 3N of generalized coordinates q

i

is called

the number of degrees of freedom. No particular metric is assumed—e.g.

we could equally as well use spherical polar coordinates (see Fig. 1.1), q

i

=

¦r

1

, θ

1

, φ

1

,r

2

, θ

2

, φ

2

, . . .,r

N

, θ

N

, φ

N

¦, or a more general curvilinear coordinate

system.

In other systems the generalized

x

θ

r

φ

y

z

r

Figure 1.1: Position vector

r in Cartesian and Spherical

Polar coordinates.

“coordinates” need not even be spa-

tial coordinates—e.g. they could

be the charges ﬂowing in an elec-

trical circuit. Thus the convenient

vector-like notation q for the array

of generalized coordinates should not

be confused with the notation r for

the position vector in 3-space. Of-

ten the set of generalized coordi-

nates is simply denoted q, but in

these notes we use a bold font, dis-

tinguishing generalized coordinate

arrays from 3-vectors by using a bold slanted font for the former and a bold

upright font for the latter.

Vectors are entities independent of which coordinates are used to repre-

sent them, whereas the set of generalized coordinates changes if we change

variables. For instance, consider the position vector of a particle in Carte-

sian coordinates x, y, z and in spherical polar coordinates r, θ, φ in Fig. 1.1.

The vector r represents a point in physical 3-space and thus does not change

when we change coordinates,

r = xe

x

+ ye

y

+ze

z

= re

r

(θ, φ) . (1.1)

However its representation changes, because of the change in the unit basis

vectors from e

x

, e

y

, e

z

to e

r

, e

φ

and e

z

. On the other hand, the sets of

1.2. GENERALIZED COORDINATES 3

generalized coordinates r

C

≡ ¦x, y, z¦ and r

sph

≡ ¦r, φ, z¦ are distinct enti-

ties: they are points in two diﬀerent (though related) conﬁguration spaces

describing the particle.

q

1

q

2

q

3

...

q

n

q(t)

q(t

1

)

q(t

2

)

Figure 1.2: Some possible paths in conﬁguration space, each

parametrized by the time, t.

Sometimes the motion is constrained to lie within a submanifold of the

full conﬁguration space.

1

For instance, we may be interested in the motion

of billiard balls constrained to move within a plane, or particles connected

by rigid rods. In such cases, where there exists a set of (functionally inde-

pendent) constraint equations, or auxiliary conditions

f

j

(q) = 0 , j = 1, 2, , m < n , (1.2)

the constraints are said to be holonomic.

Each holonomic constraint reduces the number of degrees of freedom by

one, since it allows us to express one of the original generalized coordinates as

1

A manifold is a mathematical space which can everywhere be described locally by a

Cartesian coordinate system, even though extension to a global nonsingular curvilinear

coordinate system may not be possible (as, e.g. on a sphere). A manifold can always be

regarded as a surface embedded in a higher dimensional space.

4CHAPTER 1. GENERALIZEDCOORDINATES ANDVARIATIONAL PRINCIPLES

a function of the others and delete it from the set. For instance, for a particle

constrained to move in a horizontal plane (an idealized billiard table), the

vertical position z = const is a trivial function of the horizontal coordinates

x and y and the conﬁguration space becomes two dimensional, q = ¦x, y¦.

Consider the set of all conceivable paths through conﬁguration space (see

Fig. 1.2). Each one may be parametrized by the time, t: q = q(t). By

diﬀerentiating eq. (1.2) with respect to time, we ﬁnd a set of constraints on

the generalized velocities, ˙ q

i

≡ dq

i

/ dt, which we write in diﬀerential notation

as

n

¸

i=1

∂f

j

(q)

∂q

i

dq

i

≡

∂f

j

(q)

∂q

· dq = 0 , (1.3)

where ∂f/∂q ≡ ¦∂f/∂q

1

, . . . , ∂f/∂q

n

¦ and we use the shorthand dot-product

notation

a·

∂f

∂b

≡

n

¸

i=1

a

i

∂f

∂b

i

, (1.4)

where a

i

and b

i

are arbitrary conﬁguration space variables.

The condition for functional independence of the m constraints is that

there be m nontrivial solutions of eq. (1.3), i.e. that the rank of the matrix

∂f

j

(q)/∂q

i

be its maximal possible value, m.

Note that not all such diﬀerential constraints lead to holonomic con-

straints. If we are given constraints as a general set of diﬀerential forms

n

¸

i=1

ω

(j)

i

(q) dq

i

= 0 , (1.5)

then we may or may not be able to integrate the constraint equations to the

form eq. (1.2). When we can, the forms are said to be complete diﬀerentials.

When we cannot, the constraints are said to be nonholonomic.

The latter case, where we cannot reduce the number of degrees of freedom

by the number of constraints, will not be considered explicitly in these notes.

Furthermore, we shall normally assume that any holonomic constraints have

been used to reduce ¦q

i

[i = 1, . . . , n¦ to a minimal, unconstrained set. How-

ever, we present in Sec. 1.5 an elegant alternative that may be used when

this reduction is not convenient, or is impossible due to the existence of

nonholonomic constraints.

There are situations where there is an inﬁnite number of generalized

coordinates. For instance, consider a scalar ﬁeld (such as the instaneous

amplitude of a wave), ψ(r, t). Here ψ is a generalized coordinate of the

system and the position vector r replaces the index i. Since r is a continuous

variable it ranges over an inﬁnite number of values.

1.3. EXAMPLE: THE IDEAL FLUID 5

r

0

x(r

0

,t)

Figure 1.3: A ﬂuid element advected from point r = r

0

at time t = 0

to r = x(r

0

, t) at time t.

1.3 Example: The ideal ﬂuid

As an example of a system with both an inﬁnite number of degrees of freedom

and holonomic constraints, consider a ﬂuid with density ﬁeld ρ(r, t), pressure

ﬁeld p(r, t) and velocity ﬁeld v(r, t).

Here we are using the Eulerian description, where the ﬂuid quantities ρ,

p and v are indexed by the actual position, r, at which they take on their

physical values at each point in time.

However, we can also index these ﬁelds by the initial position, r

0

, of

the ﬂuid particle passing through the point r = x(r

0

, t) at time t (see

Fig. 1.3). This is known as the Lagrangian description. (cf. the Schr¨odinger

and Heisenberg pictures in quantum mechanics.) We shall denote ﬁelds in

the Lagrangian description by use of a subscript L: ρ

L

(r

0

, t), p

L

(r

0

, t) and

v

L

(r

0

, t) = ∂

t

x(r

0

, t).

The ﬁeld x(r

0

, t) may be regarded as an inﬁnite set of generalized coor-

dinates, the speciﬁcation of which gives the state of the ﬂuid at time t. The

Jacobian J(r

0

, t) of the change of coordinates r = x(r

0

, t) is deﬁned by

J ≡

∂x

∂x

0

·

∂x

∂y

0

×

∂x

∂z

0

=

∂x

∂x

0

∂y

∂x

0

∂z

∂x

0

∂x

∂y

0

∂y

∂y

0

∂z

∂y

0

∂x

∂z

0

∂y

∂z

0

∂z

∂z

0

, (1.6)

where x

0

, y

0

and z

0

are Cartesian components of r

0

and x, y and z are the

corresponding components of x(r

0

, t). This gives the change of volume of

a ﬂuid element with initial volume dV

0

and ﬁnal volume (at time t) dV

through

dV = J(r

0

, t) dV

0

. (1.7)

To see this, consider dV

0

= dx

0

dy

0

dz

0

to be an inﬁnitesimal rectangular

box, as indicated in Fig. 1.3, with sides of length dx

0

, dy

0

, dz

0

. This ﬂuid

6CHAPTER 1. GENERALIZEDCOORDINATES ANDVARIATIONAL PRINCIPLES

element is transformed by the eﬀect of compression and shear to an inﬁnites-

imal parallelipiped with sides given by dl

x

≡ dx

0

∂x/∂x

0

, dl

y

≡ dy

0

∂x/∂y

0

,

dl

z

≡ dz

0

∂x/∂z

0

. The volume of such a parallepiped is dl

x

· dl

y

×dl

z

=

J dx

0

dy

0

dz

0

2.

Are the ﬁelds ρ(r

0

, t) and p(r

0

, t) additional generalized coordinates which

need to be speciﬁed at each point in time? In an ideal ﬂuid (i.e. one with

no dissipation, also called an Euler ﬂuid) the answer is no, because mass

conservation, ρ dV = ρ

0

dV

0

, allows us to write

ρ

L

(r

0

, t) = ρ

0

(r

0

)/J(r

0

, t) , (1.8)

while the ideal equation of state p( dV )

γ

= p

0

( dV

0

)

γ

, where γ is the ratio of

speciﬁc heats, gives

p

L

(r

0

, t) = p

0

(r

0

)/J

γ

(r

0

, t) , (1.9)

where ρ

0

and p

0

are the initial density and pressure ﬁelds, respectively. These

are, by deﬁnition, ﬁxed in time, so the only time dependence occurs through

the Jacobian J, which we showed in eq. (1.6) to be completely determined

by the Lagrangian displacement ﬁeld x(r

0

, t). Thus eqs. (1.8) and (1.9) have

allowed us to reduce the number of generalized coordinate ﬁelds from 5 to

3 (the three components of x)—mass conservation and the equation of state

have acted as holonomic constraints.

Note: Mass conservation is valid even for nonideal ﬂuids (provided they

are not reacting and thus changing from one state to another). However,

in a ﬂuid with ﬁnite dissipation, heat will be generated by the motion and

entropy will be increased in each ﬂuid element, thus invalidating the use

of the adiabatic equation of state. Further, the entropy increase depends

on the complete path of the ﬂuid through its state space, not just on its

instantaneous state. Thus the pressure cannot be holonomically constrained

in a nonideal ﬂuid.

Remark 1.1 A useful model for a hot plasma is the magnetohydrodynamic

(MHD) ﬂuid—an ideal ﬂuid with the additional property of being a perfect

electrical conductor. This leads to the magnetic ﬁeld B(r, t) being “frozen in”

to the plasma, so that B

L

also obeys a holonomic constraint in the Lagrangian

representation, but as it is a vector constraint it is a little too complicated to

give here.

1.4 Variational Calculus

Consider an objective functional I[q], deﬁned on the space of all diﬀerentiable

paths between two points in conﬁguration space, q(t

1

) and q(t

2

), as depicted

1.4. VARIATIONAL CALCULUS 7

in Fig. 1.2

I[q] ≡

t

2

t

1

dt f(q(t), ˙ q(t), t) . (1.10)

(As we shall wish to integrate by parts later, we in fact assume the paths to

be slightly smoother than simply diﬀerentiable, so that ¨ q is also deﬁned.)

We suppose our task is to ﬁnd a path that makes I a maximum or min-

imum (or at least stationary) with respect to neighbouring paths. Thus we

vary the path by an amount δq(t): q(t) → q(t) + δq(t). Then the ﬁrst

variation, δI, is deﬁned to be the change in I as estimated by linearizing in

δq:

δI[q] ≡

t

2

t

1

dt

¸

δq(t)·

∂f

∂q

+δ ˙ q(t)·

∂f

∂ ˙ q

. (1.11)

Our ﬁrst task is to evaluate eq. (1.11) in terms of δq(t). The crucial step

here is the lemma “delta and dot commute”. That is

δ ˙ q ≡

dδq

dt

. (1.12)

To prove this, simply go back to deﬁnitions: δ ˙ q ≡ d(q +δq)/ dt − dq/ dt =

dδq/ dt 2.

We can now integrate by parts to put δI in the form

δI[q] =

¸

δq·

∂f

∂ ˙ q

t

2

t

1

+

t

2

t

1

dt δq(t)·

δf

δq

. (1.13)

This consists of an endpoint contribution and an integral of the variational

derivative δf/δq, deﬁned by

δf

δq

≡

∂f

∂q

−

d

dt

∂f

∂ ˙ q

. (1.14)

Remark 1.2 The right-hand side of eq. (1.14) is also sometimes called the

functional derivative or Fr´echet derivative of I[q]. When using this termi-

nology the notation δI/δq is used instead of δf/δq so that we can write

eq. (1.13), for variations δq

i

which vanish in the neighbourhood of the end-

points, as

δI[q] ≡

t

2

t

1

dt

n

¸

i=1

δq

i

(t)

δI

δq

i

(t) ,

≡

δq,

δI

δq

, (1.15)

8CHAPTER 1. GENERALIZEDCOORDINATES ANDVARIATIONAL PRINCIPLES

which may be taken as the most general deﬁning equation for δI/δq. The

inner product notation (, ) used above is a kind of inﬁnite-dimensional dot

product where we not only sum over the index i, but integrate over the “index”

t. If we recall that the change in the value of a ﬁeld deﬁned on 3-space, e.g.

ϕ(r), due to an arbitrary inﬁnitesimal change δr is δϕ = δr·∇ϕ, which may

be regarded as the deﬁnition of the gradient ∇ϕ, we see that the functional

derivative δI/δq may be thought of, by analogy, as an inﬁnite-dimensional

gradient deﬁned on the function space of paths.

A typical variational problem is to make I extremal or stationary under

arbitrary variations δq(t) holding the endpoints ﬁxed. That is, we require

δI = 0 ∀ functions δq(t) such that δq(t

1

) = δq(t

2

) = 0 . (1.16)

Note that this condition does not necessarily require I to be a minimum

or maximum—it can be a kind of saddle point in function space, with some

ascending and some descending “directions”. To determine the nature of

a stationary point we would need to expand I to second order in δq—the

second variation.

In the class of variations in eq.

t + ε

q

i

t − ε

Figure 1.4: A time-localized

variation in generalized coor-

dinate q

i

with support in the

range t − to t + .

(1.16), the endpoint contribution in

eq. (1.13) vanishes, leaving only

the contribution of the integral over

t. Since δq(t) is arbitrary, we can,

in particular, consider functions with

arbitrarily localized support in t, as

indicated in Fig. 1.4. (The support

of a function is just the range over

which it is nonzero.) As → 0,

δf/δq(t) becomes essentially con-

stant over the support of q in eq. (1.13)

and we can move it outside the in-

tegral. Clearly then, I can only be

stationary for all such variations if and only if the variational derivative van-

ishes for each value of t and each index i

δf

δq

= 0 . (1.17)

These n equations are known as the Euler–Lagrange equations. Some-

times we encounter variational problems where we wish to extremize I under

1.4. VARIATIONAL CALCULUS 9

variations of the endpoints as well, δq(t

1

) = δq(t

2

) = 0. In such cases we

see from eq. (1.13) that, in addition to eq. (1.17), stationarity implies the

natural boundary conditions

∂f

∂ ˙ q

= 0 (1.18)

at t

1

and t

2

.

1.4.1 Example: Geodesics

In the above development we have used the symbol t to denote the indepen-

dent variable because, in applications in dynamics, paths in conﬁguration

space are naturally parametrized by the time. However, in purely geometric

applications t is simply an arbitrary label for the position along a path, and

we shall in this section denote it by τ to avoid confusion.

The distance along a path is given by integrating the lengths dl of in-

ﬁnitesimal line elements, given a metric tensor g

i,j

such that

( dl)

2

=

n

¸

i,j=1

dq

i

g

i,j

dq

j

. (1.19)

In terms of our parameter τ, we thus have the length l as a functional of the

form discussed above

l =

τ

2

τ

1

dτ

¸

n

¸

i,j=1

˙ q

i

g

i,j

(q, τ) ˙ q

j

¸

1/2

. (1.20)

A geodesic is a curve between two points whose length (calculated using

the given metric) is stationary against inﬁnitesimal variations about that

path. Thus the task of ﬁnding geodesics ﬁts within the class of variational

problems we have discussed, and we can use the Euler–Lagrange equations

to ﬁnd them. Perhaps the best known result on geodesics is the fact that the

shortest path between two points in a Euclidean space (one where g

i,j

= 0

for i = j and g

i,j

= 1 for i = j) is a straight line. Another well-known result

is that the shortest path between two points on the surface of a sphere is a

great circle (see Problem 1.6.3 for a general theorem on geodesics on a curved

surface).

Geodesics are not necessarily purely geometrical objects, but can have

physical interpretations. For instance, suppose we want to ﬁnd the shape of

an elastic string stretched over a slippery surface. The string will adjust its

shape to minimize its elastic energy. Since the elastic potential energy is a

10CHAPTER 1. GENERALIZEDCOORDINATES ANDVARIATIONAL PRINCIPLES

monotonically increasing function of the length of the string, the string will

settle onto a geodesic on the surface.

Geodesics also play an important role in General Relativity, because the

world line of a photon is a geodesic in 4-dimensional space time, with the

metric tensor obeying Einstein’s equations. If the metric is suﬃciently dis-

torted, it can happen that there is not one, but several geodesics between

two points, a fact which explains the phenomenon of gravitational lensing

(multiple images of a distant galaxy behind a closer massive object).

1.4.2 Trial function method

One advantage of the variational formulation of a problem is that we can

use trial function methods to ﬁnd approximate solutions. That is, we can

make a clever guess, q(t) = q

K

(t, a

1

, a

2

, . . . , a

K

) as to the general form of

the solution, using some speciﬁc function q

K

(the trial function) involving a

ﬁnite number of parameters a

k

, k = 1, . . . , K. Then we evaluate the integral

in eq. (1.11) (analytically or numerically) and seek a stationary point of

the resulting function I(a

1

, a

2

, . . . , a

K

) in the K-dimensional space of the

parameters a

k

. Varying the a

k

the variation in I is

δI =

K

¸

k=1

∂I

∂a

k

δa

k

. (1.21)

The condition for a stationary point is thus

∂I

∂a

k

= 0, k = 1, . . . , K , (1.22)

that is, that the K-dimensional gradient of I vanish.

Since the true solution makes the objective functional I stationary with

respect to small variations, if our guessed trial function solution is close to

the true solution the error in I will be small. (Of course, because q

K

may

not be a reasonable guess for the solution in all ranges of the parameters,

there may be spurious stationary points that must be rejected because they

cannot possibly be close to a true solution—see the answer to Problem 2.8.2

in Sec. 5.2.)

1.5 Constrained variation: Lagrange multi-

pliers

As mentioned in Sec. 1.2 we normally assume that the holonomic constraints

have been used to reduce the dimensionality of the conﬁguration space so

1.5. CONSTRAINED VARIATION: LAGRANGE MULTIPLIERS 11

that all variations are allowed. However, it may not be possible to do an

analytic elimination explicitly. Or it may be that some variables appear in

a symmetric fashion, making it inelegant to eliminate one in favour of the

others.

Thus, even in the holonomic case, it is worth seeking a method of han-

dling constrained variations: when there are one or more diﬀerential auxiliary

conditions of the form δf

(j)

= 0. In the nonholonomic case it is mandatory

to consider such variations because the auxiliary conditions cannot be inte-

grated.

We denote the dimension of the conﬁguration space by n. Following

eq. (1.5) we suppose there are m < n auxiliary conditions of the form

δf

(j)

≡ ω

(j)

(q, t)·δq = 0 . (1.23)

The vectors ω

(j)

, j = 1, . . . , m may be assumed linearly independent (else

some of the auxiliary conditions would be redundant) and thus span an m-

dimensional subspace, V

m

(t), of the full n-dimensional linear vector space

V

n

occupied by the unconstrained variations. Thus the equations eq. (1.23)

constrain the variations δq to lie within an (n − m)-dimensional subspace,

V

n−m

(q, t), complementary to V

n

.

The variational problem we seek to solve is to ﬁnd the conditions (the

generalizations of the Euler–Lagrange equations) under which the objective

functional I[q] is stationary with respect to all variations δq in V

m

(q, t).

Apart from this restriction on the variations, the problem is the same as that

described by eq. (1.16). The generalization of eq. (1.17) is

δf

δq

·δq = 0 ∀ δq ∈ V

n−m

(q, t) . (1.24)

If there are no constraints, so that m = 0, then δf/δq is orthogonal to

all vectors in V

n

and the only solution is that δf/δq ≡ 0. Thus eq. (1.24)

and eq. (1.17) are equivalent in this case. However, if m < n, then δf/δq

can have a nonvanishing component in the subspace V

m

and eq. (1.17) is no

longer valid.

An elegant solution to the problem of generalizing eq. (1.17) was found by

Lagrange. Expressed in our linear vector space language, his idea was that

eq. (1.24) can be regarded as the statement that the projection, (δf/δq)

n−m

,

of δf/δq into V

n−m

(q, t) is required to vanish.

However, we can write (δf/δq)

n−m

as δf/δq−(δf/δq)

m

, where (δf/δq)

m

is the projection of δf/δq into V

m

. Now observe that we can write any vector

in V

m

as a linear superposition of the ω

(j)

since they form a basis spanning

12CHAPTER 1. GENERALIZEDCOORDINATES ANDVARIATIONAL PRINCIPLES

this space. Thus we write (δf/δq)

m

= −

¸

λ

j

ω

(j)

, or, equivalently,

δf

δq

m

+

m

¸

j=1

λ

j

ω

(j)

= 0 , (1.25)

where the λ

j

(q, t) coeﬃcients, as yet to be determined, are known as the

Lagrange multipliers. They can be determined by dotting eq. (1.25) with each

of the m basis vectors ω

(j)

, thus providing m equations for the m unknowns.

Alternatively, we can express this variationally as

¸

δf

δq

m

+

m

¸

j=1

λ

j

ω

(j)

¸

·δq = 0 ∀ δq ∈ V

m

(q, t) . (1.26)

Since −

¸

λ

j

ω

(j)

is the projection into V

m

of δf/δq, the projection into

the complementary subspace V

n−m

is found by subtracting (−

¸

λ

j

ω

(j)

) from

δf/δq. That is,

δf

δq

n−m

=

δf

δq

+

m

¸

j=1

λ

j

ω

(j)

= 0 , (1.27)

where the V

n−m

component of the second equality follows by eq. (1.24) and

the V

n

component from eq. (1.25). Thus we have n generalized Euler–

Lagrange equations, but they incorporate the m equations for the, so far

arbitrary, λ

(j)

implicit in eq. (1.25). Thus we really only gain (n −m) equa-

tions from the variational principle, which is at it should be because we also

get m kinematic equations from the constraint conditions—if we got more

from the variational principle the problem would be overdetermined.

The variational formulation of the second equality in eq. (1.27) is

¸

δf

δq

+

m

¸

j=1

λ

j

ω

(j)

¸

·δq = 0 ∀ δq ∈ V

n

(q, t) . (1.28)

That is, by using the Lagrange multipliers we have turned the constrained

variational problem into an unconstrained one.

In the holonomic case, when the auxiliary conditions are of the form in

eq. (1.2), we may derive eq. (1.28) by unconstrained variation of the modiﬁed

objective functional

I

∗

[q] ≡

t

2

t

1

dt (f +

m

¸

j=1

λ

j

f

j

) . (1.29)

The auxiliary conditions also follow from this functional if we require that it

be stationary under variation of the λ

j

.

1.6. PROBLEMS 13

1.6 Problems

1.6.1 Rigid rod

Two particles are connected by a rigid rod so they are constrained to move a

ﬁxed distance apart. Write down a constraint equation of the form eq. (1.2)

and ﬁnd suitable generalized coordinates for the system incorporating this

holonomic constraint.

1.6.2 Ecliptic

Suppose we know that the angular momentum vectors r

k

×m

k

˙ r

k

of a system

of particles are all nonzero and parallel to the z-axis in a particular Cartesian

coordinate system. Write down the diﬀerential constraints implied by this

fact, and show that they lead to a set of holonomic constraints. Hence write

down suitable generalized coordinates for the system.

1.6.3 Curvature of geodesics

Show that any geodesic r = x(τ) on a two-dimensional manifold S : r =

X(θ, ζ) embedded in ordinary Euclidean 3-space, where θ and ζ are arbi-

trary curvilinear coordinates on S, is such that the curvature vector κ(τ) is

everywhere normal to S (or zero).

The curvature vector is deﬁned by κ ≡ de

/ dl, where e

(τ) ≡ dx/ dl is

the unit tangent vector at each point along the path r = x(τ).

Hint: First ﬁnd f(θ, ζ,

˙

θ,

˙

ζ) =

˙

l, the integrand of the length functional,

l =

**f dτ (which involves ﬁnding the metric tensor in θ, ζ space in terms of
**

∂X/∂θ and ∂X/∂ζ). Then show that, for any path on S,

∂f

∂

˙

θ

= e

·

∂X

∂θ

(and similarly for the

˙

ζ derivative) and

∂f

∂θ

= e

·

d

dτ

∂X

∂θ

,

and again similarly for the ζ derivative.

14CHAPTER 1. GENERALIZEDCOORDINATES ANDVARIATIONAL PRINCIPLES

Chapter 2

Lagrangian Mechanics

2.1 Introduction

The previous chapter dealt with generalized kinematics—the description of

given motions in time and space. In this chapter we deal with one formula-

tion (due to Lagrange) of generalized dynamics—the derivation of diﬀerential

equations (equations of motion) for the time evolution of the generalized co-

ordinates. Given appropriate initial conditions, these (in general, nonlinear)

equations of motion specify the motion uniquely. Thus, in a sense, the most

important task of the physicist is over when the equations of motion have

been derived—the rest is just mathematics or numerical analysis (important

though these are). The goal of generalized dynamics is to ﬁnd universal forms

of the equations of motion.

From elementary mechanics we are all familiar with Newton’s Second

Law, F = ma for a particle of mass m subjected to a force F and undergoing

an acceleration a ≡ ¨r. If we know the Cartesian components F

i

(r, ˙ r, t),

i = 1, 2, 3, of the force in terms of the Cartesian coordinates x

1

= x, x

2

= y,

x

3

= z and their ﬁrst time derivatives then the equations of motion are the

set of three second-order diﬀerential equations m¨ x

i

−F

i

= 0.

To give a physical framework for developing our generalized dynamical

formalism we consider a set of N Newtonian point masses, which may be con-

nected by holonomic constraints so the number n of generalized coordinates

may be less than 3N. Indeed, in the case of a rigid body N is essentially

inﬁnite, but the number of generalized coordinates is ﬁnite. For example,

the generalized coordinates for a rigid body could be the three Cartesian

coordinates of the centre of mass and three angles to specify its orientation

(known as the Euler angles), so n = 6 for a rigid body allowed to move freely

in space.

15

16 CHAPTER 2. LAGRANGIAN MECHANICS

Whether the point masses are real particles like electrons, composite par-

ticles like nuclei or atoms, or mathematical idealizations like the inﬁnitesimal

volume elements in a continuum description, we shall refer to them generi-

cally as “particles”.

Having found a very general form of the equations of motion (Lagrange’s

equations), we then ﬁnd a variational principle (Hamilton’s Principle) that

gives these equations as its Euler–Lagrange equations in the case of no fric-

tional dissipation. This variational principle forms a basis for generalizing

even beyond Newtonian mechanics (e.g. to dynamics in Special Relativity).

2.2 Generalized Newton’s 2nd Law

2.2.1 Generalized force

Let the (constrained) position of each of the N particles making up the sys-

tem be given as a function of the n generalized coordinates q by r

k

= x

k

(q, t),

k = 1, . . . , N. If there are holonomic constraints acting on the particles, the

number of generalized coordinates satisﬁes the inequality n ≤ 3N. Thus, the

system may be divided into two subsystems—an “exterior” subsystem de-

scribed by the n generalized coordinates and an “interior” constraint subsys-

tem whose (3N−n) coordinates are rigidly related to the q by the geometric

constraints.

In a naive Newtonian approach we would have to specify the forces act-

ing on each particle (taking into account Newton’s Third Law, “action and

reaction are equal and opposite”), derive the 3N equations of motion for

each particle and then eliminate all the interior subsystem coordinates to

ﬁnd the equations of motion of the generalized coordinates only. This is

clearly very ineﬃcient, and we already know from elementary physics that

it is unnecessary—we do not really need an inﬁnite number of equations to

describe the motion of a rigid body. What we seek is a formulation in which

only the generalized coordinates, and generalized forces conjugate to them,

appear explicitly. All the interior coordinates and the forces required to main-

tain their constrained relationships to each other (the “forces of constraint”)

should be implicit only.

To achieve this it turns out to be fruitful to adopt the viewpoint that

the total mechanical energy (or, rather, its change due to the performance of

external work, W, on the system) is the primitive concept, rather than the

vector quantity force. The basic reason is that the energy, a scalar quantity,

needs only speciﬁcation of the coordinates for its full description, whereas

the representation of the force, a vector, depends also on deﬁning a basis

2.2. GENERALIZED NEWTON’S 2ND LAW 17

set on which to resolve it. The choice of basis set is not obvious when we

are using generalized coordinates. [Historically, force came to be understood

earlier, but energy also has a long history, see e.g. Ernst Mach “The Science

of Mechanics” (Open Court Publishing, La Salle, Illinois, 1960) pp. 309–312,

QA802.M14 Hancock. With the development of Lagrangian and Hamiltonian

methods, and thermodynamics, energy-based approaches can now be said to

be dominant in physics.]

To illustrate the relation between force and work, ﬁrst consider just

one particle. Recall that the work δW done on the particle by a force

F as the particle suﬀers an inﬁnesimal displacement δr is δW = F·δr ≡

F

x

δx+F

y

δy +F

z

δz. A single displacement δr does not give enough informa-

tion to determine the three components of F, but if we imagine the thought

experiment of displacing the particle in the three independent directions,

δr = δx e

x

, δy e

y

and δz e

z

, determining the work, δW

x

, δW

y

, δW

z

, done in

each case, then we will have enough equations to deduce the three compo-

nents of the force vector, F

x

= δW

x

/δx, F

y

= δW

y

/δy, F

z

= δW

z

/δz. [If

δW can be integrated to give a function W(r) (which is not always possi-

ble), then we may use standard partial derivative notation: F

x

= ∂W/∂x,

F

y

= ∂W/∂y, F

z

= ∂W/∂z.]

The displacements δx e

x

, δy e

y

and δz e

z

are historically called virtual

displacements. They are really simply the same displacements as used in

the mathematical deﬁnition of partial derivatives. Note that the virtual

displacements are done at a ﬁxed instant in time as if by some “invisible

hands”, which perform the work δW: this is a thought experiment—the

displacements are ﬁctitious, not dynamical.

Suppose we now transform from the Cartesian coordinates x, y, z to an

arbitrary curvilinear coordinate system q

1

, q

2

, q

3

(the superscript notation

being conventional in tensor calculus). Then an arbitrary virtual displace-

ment is given by δr =

¸

i

δq

i

e

i

, where the basis vectors e

i

are in general not

orthonormal. The corresponding “virtual work” is given by δW =

¸

i

δq

i

F

i

,

where F

i

≡ e

i

·F. As in the Cartesian case, this can be used to determine the

generalized forces F

i

by determining the virtual work done in three indepen-

dent virtual displacements.

In tensor calculus the set ¦F

i

¦ is known as the covariant representation of

F, and is in general distinct from an alternative resolution, the contravariant

representation ¦F

i

¦. (You will meet this terminology again in relativity

theory.) The energy approach shows that the covariant, rather than the

contravariant, components of the force form the natural generalized forces

conjugate to the generalized coordinates q

i

.

Turning now to the N-particle system as a whole, the example above

suggests we deﬁne the set of n generalized forces, ¦Q

i

¦, conjugate to each

18 CHAPTER 2. LAGRANGIAN MECHANICS

of the n degrees of freedom q

i

, to be such that the virtual work done on the

system in displacing it by an arbitrary inﬁnitesimal amount δq at ﬁxed time

t is given by

δW ≡

n

¸

i=1

Q

i

δq

i

∀ δq . (2.1)

We now calculate the virtual work in terms of the displacements of the N

particles assumed to make up the system and the forces F

k

acting on them.

The virtual work is

δW =

N

¸

k=1

F

k

·δr

k

=

n

¸

i=1

δq

i

N

¸

k=1

F

k

·

∂x

k

∂q

i

. (2.2)

Comparing eq. (2.1) and eq. (2.2), and noting that they hold for any δq, we

can in particular take all but one of the δq

i

to be zero to pick out the ith

component, giving the generalized force as

Q

i

=

N

¸

k=1

F

k

·

∂x

k

∂q

i

. (2.3)

When there are holonomic constraints on the system we decompose the

forces acting on the particles into what we shall call explicit forces and forces

of constraint. (The latter terminology is standard, but the usage “explicit

forces” seems new—often they are called “applied forces”, but this is confus-

ing because they need not originate externally to the system, but also from

interactions between the particles.)

By forces of constraint, F

cst

k

, we mean those imposed on the particles

by the rigid rods, joints, sliding planes etc. that make up the holonomic

constraints on the system. These forces simply adjust themselves to whatever

values are required to maintain the geometric constraint equations and can

be regarded as “private” forces that, for most purposes, we do not need to

know

1

. Furthermore, we may not be able to tell what these forces need to

be until we have solved the equations of motion, so they cannot be assumed

known a priori.

The explicit force on each particle, F

xpl

k

, is the vector sum of any externally

imposed forces, such as those due to an external gravitational or electric ﬁeld,

plus any interaction forces between particles such as those due to elastic

1

Of course, in practical engineering design contexts one should at some stage check that

the constraint mechanism is capable of supplying the required force without deforming or

breaking!

2.2. GENERALIZED NEWTON’S 2ND LAW 19

springs coupling point masses, or to electrostatic attractions between charged

particles. If there is friction acting on the particle, including that due to

constraint mechanisms, then that must be included in F

xpl

k

as well. These

are the “public” forces that determine the dynamical evolution of the degrees

of freedom of the system and are determined by the conﬁguration q of the

system at each instant of time, and perhaps by the generalized velocity ˙ q in

the case of velocity-dependent forces such as those due to friction and those

acting on a charged particle moving in a magnetic ﬁeld.

Figure 2.1 shows a simple system with

N

mg

F

α

Figure 2.1: A body on

an inclined plane as de-

scribed in the text.

a holonomic constraint—a particle slid-

ing on a plane inclined at angle α. It is

subject to the force of gravity, mg, the

normal force N, and a friction force F in

the directions shown. The force of con-

straint is N. It does no work because it

is orthogonal to the direction of motion,

and its magnitude is that required to null

out the normal component of the gravita-

tional force, [N[ = mg cos α, so as to give

no acceleration in the normal direction and thus maintain the constraint.

We now make the crucial observation that, because the constraints are

assumed to be provided by rigid, undeformable mechanisms, no work can

be done on the interior constraint subsystem by the virtual displacements.

That is, no (net) virtual work is done against the nonfriction forces imposed

by the particles on the constraint mechanisms in performing the variations

δq. By Newton’s Third Law, the nonfriction forces acting on the constraint

mechanisms are equal and opposite to the forces of constraint, F

cst

k

. Thus

the sum over F

cst

k

·δx

k

vanishes and we can replace the total force F

k

with

F

xpl

k

in eqs. (2.2) and (2.3). Note: If there are friction forces associated

with the constraints there is work done against these, but this fact does not

negate the above argument because we have included the friction forces in

the explicit forces—any work done against friction forces goes into heat which

is dissipated into the external world, not into mechanical energy within the

constraint subsystem.

That is,

Q

i

=

N

¸

k=1

F

xpl

k

·

∂x

k

∂q

i

. (2.4)

For the purpose of calculating the generalized forces, this is a much more

practical expression than eq. (2.3) because the F

xpl

k

are known in terms of

20 CHAPTER 2. LAGRANGIAN MECHANICS

the instantaneous positions and velocities. Thus Q

i

= Q

i

(q, ˙ q, t).

2.2.2 Generalized equation of motion

We now suppose that eq. (2.4) has been used to determine the Qas functions

of the q (and possible ˙ q and t if we have velocity-dependent forces and time-

dependent constraints, respectively). Then we rewrite eq. (2.3) in the form

N

¸

k=1

F

k

·

∂x

k

∂q

i

= Q

i

(q, ˙ q, t) . (2.5)

To derive an equation of motion we use Newton’s second law to replace

F

k

on the left-hand side of eq. (2.5) with m

k

¨r

k

,

N

¸

k=1

m

k

¨r

k

·

∂x

k

∂q

i

≡

N

¸

k=1

m

k

¸

d

dt

˙ r

k

·

∂x

k

∂q

i

− ˙ r

k

·

d

dt

∂x

k

∂q

i

= Q

i

. (2.6)

We now diﬀerentiate r

k

= x

k

(q(t), t) with respect to t to ﬁnd the function

v

k

such that ˙ r

k

= v

k

(q, ˙ q, t):

˙ r

k

=

∂x

k

∂t

+

n

¸

j=1

˙ q

j

∂x

k

∂q

j

≡ v

k

(q, ˙ q, t) . (2.7)

Diﬀerentiating v

k

wrt ˙ q

i

(treating q and ˙ q as independent variables in partial

derivatives) we immediately have the lemma

∂x

k

∂q

i

=

∂v

k

∂ ˙ q

i

. (2.8)

The second lemma about partial derivatives of v

k

that will be needed is

d

dt

∂x

k

∂q

i

=

∂v

k

∂q

i

, (2.9)

which follows because d/ dt can be replaced by ∂/∂t + ˙ q·∂/∂q, which com-

mutes with ∂/∂q

i

(cf. the “interchange of delta and dot” lemma in Sec. 1.4).

Applying these two lemmas in eq. (2.6) we ﬁnd

N

¸

k=1

¸

d

dt

m

k

v

k

·

∂v

k

∂ ˙ q

i

−m

k

v

k

·

∂v

k

∂q

i

=

N

¸

k=1

¸

d

dt

∂

∂ ˙ q

i

1

2

m

k

v

2

k

−

∂

∂q

i

1

2

m

k

v

2

k

= Q

i

, (2.10)

2.3. LAGRANGE’S EQUATIONS (SCALAR POTENTIAL CASE) 21

In terms of the total kinetic energy of the system, T

T ≡

N

¸

k=1

1

2

m

k

v

2

k

, (2.11)

we write eq. (2.10) compactly as the generalized Newton’s second law

d

dt

∂T

∂ ˙ q

i

−

∂T

∂q

i

= Q

i

. (2.12)

These n equations are sometimes called Lagrange’s equations of motion, but

we shall reserve this term for a later form [eq. (2.24)] arising when we assume

a special (though very general) form for the Q

i

. They are also sometimes

called (e.g. Scheck p. 83) d’Alembert’s equations, but this may be historically

inaccurate so is best avoided.

2.2.3 Example: Motion in Cartesian coordinates

Let us check that we can recover Newton’s equations of motion as a special

case when q = ¦x, y, z¦. In this case

T =

1

2

m( ˙ x

2

+ ˙ y

2

+ ˙ z

2

) (2.13)

so

∂T

∂x

=

∂T

∂y

=

∂T

∂z

= 0 (2.14)

and

∂T

∂ ˙ x

= m˙ x ,

∂T

∂ ˙ y

= m˙ y ,

∂T

∂ ˙ z

= m˙ z . (2.15)

Also, from eq. (2.4) we see that Q

i

≡ F

i

. Substituting in eq. (2.12) we

immediately recover Newton’s 2nd Law in Cartesian form

m¨ x = F

x

m¨ y = F

y

m¨ z = F

y

(2.16)

as expected.

2.3 Lagrange’s equations (scalar potential case)

In many problems in physics the forces F

k

are derivable from a potential ,

V (r

1

, r

2

, , r

N

). For instance, in the classical N-body problem the parti-

cles are assumed to interact pairwise via a two-body interaction potential

22 CHAPTER 2. LAGRANGIAN MECHANICS

V

k,l

(r

k

, r

l

) ≡ U

k,l

([r

k

−r

l

[) such that the force on particle k due to particle l

is given by

F

k,l

= −∇

k

V

k,l

= −

(r

k

−r

l

)

[r

k

−r

l

[

U

k,l

(r

k,l

) , (2.17)

where the prime on U denotes the derivative with respect to its argument,

the interparticle distance r

k,l

≡ [r

k

− r

l

[. Then the total force on particle k

is found by summing the forces on it due to all the other particles

F

k

= −

¸

l=k

∇

k

V

k,l

= −∇

k

V , (2.18)

where the N-body potential V is the sum of all distinct two-body interactions

V ≡

N

¸

k=1

¸

l≤k

V

k,l

=

1

2

N

¸

k,l=1

V

k,l

. (2.19)

In the ﬁrst line we counted the interactions once and only once: noting

that V

k,l

= V

l,k

, so that the matrix of interactions is symmetric we have

kept only those entries below the diagonal to avoid double counting. In the

second, more symmetric, form we have summed all the oﬀ-diagonal entries

of the matrix but have compensated for the double counting by dividing by

2. The exclusion of the diagonal “self-interaction” potentials is indicated by

putting a prime on the

¸

.

Physical examples of such an N-body system with binary interactions

are:

• An unmagnetized plasma, where V

k,l

is the Coulomb interaction

U

k,l

(r) =

e

k

e

l

0

r

, (2.20)

where the e

k

are the charges on the particles and

0

is the permittivity

of free space. We could also allow for the eﬀect of gravity by adding

the potential

¸

k

m

k

gz

k

to V , where z

k

is the height of the kth particle

with respect to a horizontal reference plane and m

k

is its mass.

2.3. LAGRANGE’S EQUATIONS (SCALAR POTENTIAL CASE) 23

• A globular cluster of stars, where V

k,l

is the gravitational interaction

U

k,l

(r) =

Gm

k

m

l

r

, (2.21)

where G is the gravitational constant.

• A dilute monatomic gas, where V

k,l

is the Van der Waal’s interaction.

However, if the gas is too dense (or becomes a liquid) we would have

to include 3-body or higher interactions as the wave functions of more

than two atoms could overlap simultaneously.

Even when the system is subjected to external forces, such as gravity,

and/or holonomic constraints, we can often still assume that the “explicit

forces” are derivable from a potential

F

xpl

k

= −∇

k

V . (2.22)

Taking into account the constraints, we see that the potential V (r

1

, r

2

, , r

N

)

becomes a function, V (q, t), in the reduced conﬁguration space. Then, from

eq. (2.4) we have

Q

i

= −

N

¸

k=1

∂x

k

∂q

i

·∇

k

V

= −

∂

∂q

i

V (q, t) . (2.23)

Substituting this form for Q

i

in eq. (2.12) we see that the generalized

Newton’s equations of motion can be encapsulated in the very compact form

(Lagrange’s equations of motion)

d

dt

∂L

∂ ˙ q

i

−

∂L

∂q

i

= 0 , (2.24)

where the function L(q, ˙ q, t), called the Lagrangian, is deﬁned as

L ≡ T −V . (2.25)

2.3.1 Hamilton’s Principle

Comparing eq. (2.24) with eq. (1.17) we see that Lagrange’s equations of

motion have exactly the same form as the Euler–Lagrange equations for the

variational principle δS = 0, where the functional S[q], deﬁned by

S ≡

t

2

t

1

dt L( ˙ q, q, t) , (2.26)

24 CHAPTER 2. LAGRANGIAN MECHANICS

is known as the action integral . Since the natural boundary conditions

eq. (1.18) are not physical, the variational principle is one in which the end-

points are to be kept ﬁxed.

We can now state Hamilton’s Principle: Physical paths in conﬁguration

space are those for which the action integral is stationary against all inﬁnites-

imal variations that keep the endpoints ﬁxed.

By physical paths we mean those paths, out of all those that are consistent

with the constraints, that actually obey the equations of motion with the

given Lagrangian.

To go beyond the original Newtonian dynamics with a scalar potential

that we used to motivate Lagrange’s equations, we can instead take Hamil-

ton’s Principle, being such a simple and geometrically appealing result, as a

more fundamental and natural starting point for Lagrangian dynamics.

2.4 Lagrangians for some Physical Systems

2.4.1 Example 1: 1-D motion—the pendulum

One of the simplest nonlinear systems is

( 1 − c o s θ ) l

0

mg

z

θ

l

Figure 2.2: Phys-

ical pendulum.

the one-dimensional physical pendulum (so called

to distinguish it from the linearized harmonic

oscillator approximation). As depicted in Fig. 2.2,

the pendulum consists of a light rigid rod of

length l, making an angle θ with the vertical,

swinging from a ﬁxed pivot at one end and

with a bob of mass m attached at the other.

The constraint l = const and the assump-

tion of plane motion reduces the system to one

degree of freedom, described by the general-

ized coordinate θ. (This system is also called

the simple pendulum to distinguish it from the

spherical pendulum and compound pendula,

which have more than one degree of freedom.)

The potential energy with respect to the equi-

librium position θ = 0 is V (θ) = mgl(1 − cos θ), where g is the acceleration

due to gravity, and the velocity of the bob is v

θ

= l

˙

θ, so that the kinetic

energy T =

1

2

mv

2

θ

=

1

2

ml

2

˙

θ

2

. The Lagrangian, T −V , is thus

L(θ,

˙

θ) =

1

2

ml

2

˙

θ

2

−mgl(1 −cos θ) . (2.27)

2.4. LAGRANGIANS FOR SOME PHYSICAL SYSTEMS 25

This is also essentially the Lagrangian for a particle moving in a sinusoidal

spatial potential, so the physical pendulum provides a paradigm for problems

such as the motion of an electron in a crystal lattice or of an ion or electron

in a plasma wave.

From eq. (2.27) ∂L/∂

˙

θ = ml

2

˙

θ and ∂L/∂θ = −mgl sin θ. Thus, the

Lagrangian equation of motion is

ml

2

¨

θ = −mgl sin θ . (2.28)

Expanding the cosine up to quadratic order in θ gives the harmonic os-

cillator oscillator approximation (see also Sec. 2.6.2)

L ≈ L

lin

≡

1

2

ml

2

˙

θ

2

−

1

2

mglθ

2

, (2.29)

for which the equation of motion is, dividing through by ml

2

,

¨

θ + ω

2

0

θ = 0,

with ω

0

≡

g/l.

2.4.2 Example 2: 2-D motion in a central potential

Let us work in plane polar coordinates, q = ¦r, θ¦, such that

x = r cos θ , y = r sin θ , (2.30)

so that

˙ x = ˙ r cos θ −r

˙

θ sin θ ,

˙ y = ˙ r sin θ +r

˙

θ cos θ , (2.31)

whence the kinetic energy T ≡

1

2

( ˙ x

2

+ ˙ y

2

) is found to be

T =

1

2

m

˙ r

2

+r

2

˙

θ

2

. (2.32)

An alternative derivation of eq. (2.32) may be found by resolving v into the

components ˙ re

r

and r

˙

θe

θ

, where e

r

is the unit vector in the radial direction

and e

θ

is the unit vector in the azimuthal direction.

We now consider the restricted two body problem—one light particle

orbiting about a massive particle which may be taken to be ﬁxed at r = 0

(e.g. an electron orbiting about a proton in the Bohr model of the hydrogen

atom, or a planet orbiting about the sun). Then the potential V = V (r)

(given by eq. (2.20) or eq. (2.21)) is a function only of the radial distance

from the central body and not of the angle.

26 CHAPTER 2. LAGRANGIAN MECHANICS

Then, from eq. (2.25) the Lagrangian is

L =

1

2

m

˙ r

2

+r

2

˙

θ

2

−V (r) . (2.33)

First we observe that L is independent of θ (in which case θ is said to be

ignorable). Then ∂L/∂θ ≡ 0 and the θ component of Lagrange’s equations,

eq. (2.24) becomes

d

dt

∂L

∂

˙

θ

= 0 , (2.34)

which we may immediately integrate once to get an integral of the motion,

i.e. a dynamical quantity that is constant along the trajectory

∂L

∂

˙

θ

= const . (2.35)

From eq. (2.33) we see that ∂L/∂

˙

θ = mr

2

˙

θ, which is the angular momentum.

Thus eq. (2.33) expresses conservation of angular momentum.

Turning now to the r-component

u

θ

r

Figure 2.3: Planar mo-

tion with a time-varying cen-

tripetal constraint as de-

scribed in the text.

of Lagrange’s equations, we see from

eq. (2.33)

∂L

∂r

= mr

˙

θ

2

−V

(r) ,

∂L

∂ ˙ r

= m˙ r ,

d

dt

∂L

∂ ˙ r

= m¨ r .(2.36)

From eq. (2.24) we ﬁnd the radial

equation of motion to be

m¨ r −mr

˙

θ

2

= −V

(r) . (2.37)

2.4.3 Example 3: 2-D motion with time-varying con-

straint

Instead of free motion in a central potential, consider instead a weight ro-

tating about the origin on a frictionless horizontal surface (see Fig. 2.3) and

constrained by a thread, initially of length a, that is being pulled steadily

downward at speed u through a hole at the origin so that the radius r = a−ut.

Then the Lagrangian is, substituting for r in eq. (2.32),

L = T =

1

2

m

u

2

+ (a −ut)

2

˙

θ

2

. (2.38)

2.4. LAGRANGIANS FOR SOME PHYSICAL SYSTEMS 27

Now only θ is an unconstrained generalized coordinate. As before, it is

ignorable, and so we again have conservation of angular momentum

m(a −ut)

2

˙

θ ≡ l = const , (2.39)

which equation can be integrated to give θ as a function of t, θ = θ

0

+

(l/mu)[1/(a −ut) −1/a] = θ

0

+ lt/[ma(a −ut)].

Clearly angular momentum is conserved, because the purely radial string

cannot exert any torque on the weight. Thus Lagrange’s equation of motion

gives the correct answer. However, the string is obviously doing work on the

system because T =

1

2

[mu

2

+ (l

2

/m)/(a − ut)

2

] is not conserved. Have we

not therefore violated the postulate in Sec. 2.2.2 of no work being done by

the constraints? The answer is “no” because what we assumed in Sec. 2.2.2

was that no virtual work was done by the constraints. The fact that the

constraint is time-dependent is irrelevant to this postulate, because virtual

displacements are done instantaneously at any given time.

2.4.4 Example 4: Atwood’s machine

Consider two weights of mass m

1

and m

2

− x

x

0

m

1

m

2

Figure 2.4: At-

wood’s machine.

suspended from a frictionless, inertialess pully

of radius a by a rope of ﬁxed length, as de-

picted in Fig. 2.4. The height of weight 1 is x

with respect to the chosen origin and the holo-

nomic constraint provided by the rope allows

us to express the height of weight 2 as −x, so

that there is only one degree of freedom for

this system.

The kinetic and potential energy are T =

1

2

(m

1

+ m

2

) ˙ x

2

and V = m

1

gx − m

2

gx. Thus

L = T −V is given by

L =

1

2

(m

1

+m

2

) ˙ x

2

−(m

1

−m

2

)gx (2.40)

and its derivatives are ∂L/∂x = −(m

1

−m

2

)g

and ∂L/∂ ˙ x = (m

1

+m

2

) ˙ x, so that the equation of motion d(∂L/∂ ˙ x) = ∂L/∂x

becomes

¨ x = −

m

1

−m

2

m

1

+m

2

g . (2.41)

28 CHAPTER 2. LAGRANGIAN MECHANICS

2.4.5 Example 5: Particle in e.m. ﬁeld

The fact that Lagrange’s equations are the Euler–Lagrange equations for the

extraordinarily simple and general Hamilton’s Principle (see Sec. 2.3.1) sug-

gests that Lagrange’s equations of motion may have a wider range of validity

than simply problems where the force is derivable from a scalar potential.

Thus we do not deﬁne L as T −V , but rather postulate the universal validity

of Lagrange’s equations of motion (or, equivalently, Hamilton’s Principle), for

describing non-dissipative classical dynamics and accept any Lagrangian as

valid that gives the physical equation of motion.

In particular, it is obviously of great physical importance to ﬁnd a La-

grangian for which Lagrange’s equations of motion eq. (2.24) reproduce the

equation of motion of a charged particle in an electromagnetic ﬁeld, under

the inﬂuence of the Lorentz force,

m¨r = eE(r, t) +e˙ r×B(r, t) , (2.42)

where e is the charge on the particle of mass m.

We assume the electric and magnetic ﬁelds E and B, respectively, to

be given in terms of the scalar potential Φ and vector potential A by the

standard relations

E = −∇Φ −∂

t

A ,

B = ∇×A . (2.43)

The electrostatic potential energy is eΦ, so we expect part of the La-

grangian to be

1

2

m˙ r

2

− eΦ, but how do we include the vector potential?

Clearly we need to form a scalar since L is a scalar, so we need to dot A

with one of the naturally occurring vectors in the problem to create a scalar.

The three vectors available are A itself, r and ˙ r. However we do not wish to

use A, since A·A in the Lagrangian would give an equation of motion that

is nonlinear in the electromagnetic ﬁeld, contrary to eq. (2.42). Thus we can

only use r and ˙ r. Comparing eqs. (2.42) and (2.43) we see that ˙ r·A has the

same dimensions as Φ, so let us try adding that to form the total Lagrangian

L =

1

2

m˙ r

2

−eΦ + e˙ r·A . (2.44)

Taking q ≡ ¦q

1

, q

2

, q

3

¦ = ¦x, y, z¦ we have

∂L

∂q

i

= −e

∂Φ

∂q

i

+e

3

¸

j=1

˙ q

j

∂A

j

∂q

i

, (2.45)

2.4. LAGRANGIANS FOR SOME PHYSICAL SYSTEMS 29

and

d

dt

∂L

∂ ˙ q

i

= m¨ q

i

+e

dA

i

dt

= m¨ q

i

+ e

¸

∂A

i

∂t

+

3

¸

j=1

˙ q

j

∂A

i

∂q

j

¸

. (2.46)

Substituting eqs. (2.45) and (2.46) in eq. (2.24) we ﬁnd

m¨ q

i

= e

¸

−

∂Φ

∂q

i

−

∂A

i

∂t

+e

3

¸

j=1

˙ q

j

¸

∂A

j

∂q

i

−

∂A

i

∂q

j

. (2.47)

This is simply eq. (2.42) in Cartesian component form, so our guessed La-

grangian is indeed correct.

2.4.6 Example 6: Particle in ideal ﬂuid

In Sec. 1.3 we presented a ﬂuid as a system with an inﬁnite number of degrees

of freedom. However, if we concentrate only on the motion of a single ﬂuid

element (a test particle), taking the pressure p and mass density ρ as known,

prescribed functions of r and t the problem becomes only three-dimensional.

Dividing by ρ we write the equation of motion of a ﬂuid element as

dv

dt

= −

∇p

ρ

−∇V , (2.48)

where V (r, t) is the potential energy (usually gravitational) per unit mass.

To ﬁnd a Lagrangian for this motion we need to be able to combine

the pressure gradient and density into an eﬀective potential. In an ideal

compressible ﬂuid we have the equation of state p( dV )

γ

= p

0

( dV

0

)

γ

, where

γ is the ratio of speciﬁc heats, which we can write as

p

ρ

γ

= const , (2.49)

where the right-hand side is a constant of the motion for the given test

particle. If we further assume that it is the same constant for all ﬂuid

elements in the neighbourhood of the test particle, then we can take the

gradient of the log of eq. (2.49) to get (∇p)/p = γ(∇ρ)/ρ. Thus

∇p

ρ

= ∇

p

ρ

−

p

ρ

2

∇ρ

= ∇

p

ρ

−

∇p

γρ

.

30 CHAPTER 2. LAGRANGIAN MECHANICS

Solving for ∇p/ρ we get

∇p

ρ

= ∇h (2.50)

where the enthalpy (per unit mass) is deﬁned by

h ≡

γ

γ −1

p

ρ

. (2.51)

Using eq. (2.50) in eq. (2.48) we recognize it as the equation of motion

for a particle of unit mass with total potential energy h + V . Thus the

Lagrangian is

L =

1

2

˙ r

2

−h −V . (2.52)

2.5 Averaged Lagrangian

From Sec. 2.3.1 we know that Lagrangian dynamics has a variational for-

mulation, and so we expect that trial function methods (see Sec. 1.4.2) may

be useful as a way of generating approximate solutions of the Lagrangian

equations of motion. In particular, suppose we know that the solutions are

oscillatory functions of t with a frequency much higher than the inverse of

any characteristic time for slow changes in the parameters of the system (the

changes being then said to occur adiabatically). Then we may use a trial

function of the form

q(t) = ¯ q(φ(t), A

1

(t), A

2

(t), . . .) (2.53)

where ¯ q is a 2π-periodic function of φ, the phase of the rapid oscillations, and

the A

k

are a set of slowly varying amplitudes characterizing the waveform

(e.g. see Problem 2.8.2). Thus,

˙ q = ω(t)

∂¯ q

∂φ

+

˙

A

1

∂¯ q

∂A

1

+

˙

A

2

∂¯ q

∂A

2

+ , (2.54)

where the instantaneous frequency is deﬁned by

ω(t) ≡

˙

φ(t) . (2.55)

Since ω d ln A

k

/ dt, to a ﬁrst approximation we may keep only the ﬁrst

term in eq. (2.54). Thus our approximate L is a function of ω, but not of

˙

A

1

,

˙

A

2

etc.

Now take the time integration in the action integral to be over a time

long compared with the period of oscillation, but short compared with the

2.5. AVERAGED LAGRANGIAN 31

timescale for changes in the system parameters. Thus only the phase-average

of L,

¯

L ≡

2π

0

dφ

2π

L (2.56)

contributes to the action,

S ≈

t

2

t

1

dt

¯

L(ω, A

1

(t), A

2

(t), . . .) (2.57)

for the class of oscillatory physical solutions we seek. Note that the averaging

in eq. (2.56) removes all direct dependence on φ, so

¯

L depends only on its

time derivative ω.

Thus, we have a new, approximate form of Hamilton’s Principle in which

the averaged Lagrangian replaces the exact Lagrangian, and in which the

set ¦φ, A

1

, A

2

, . . .¦ replaces the set q

1

, q

2

, . . . as the generalized variables.

Requiring S to be stationary within the class of quasiperiodic trial functions

then gives the new adiabatic Euler–Lagrange equations

d

dt

∂

¯

L

∂

˙

φ

= 0 , (2.58)

and

∂

¯

L

∂A

k

= 0 , k = 1, 2, . . . . (2.59)

The ﬁrst equation expresses the conservation of the adiabatic invariant,

J = const, where J is known as the oscillator action

J ≡

∂

¯

L

∂ω

. (2.60)

The second set of equations gives relations between ω and the A

k

s that give

the waveform and the instantaneous frequency.

2.5.1 Example: Harmonic oscillator

Consider a weight of mass m at the end of a light spring with spring constant

k = mω

2

0

. Then the kinetic energy is T =

1

2

m˙ x

2

and the potential energy is

V =

1

2

kx

2

=

1

2

mω

2

0

x

2

. Thus the natural form of the Lagrangian, T −V , is

L =

1

2

m( ˙ x

2

−ω

2

0

x

2

) . (2.61)

Now consider that the length of the spring, or perhaps the spring constant,

changes slowly with time, so that ω

0

= ω

0

(t), with d ln ω

0

/ dt <ω

0

. Because

32 CHAPTER 2. LAGRANGIAN MECHANICS

the system is changing with time, the amplitude of oscillation is not constant.

We now use the average Lagrangian method to ﬁnd the adiabatic invariant,

which allows us to predict how the amplitude changes with time. (For an

alternative derivation of adiabatic invariance in this case, see Sec. 4.2.1.)

Since the problem is linear, we use a sinusoidal trial function with no

harmonics

x = Asin φ(t) . (2.62)

In response to the slow changes in ω

0

we expect A and ω ≡

˙

φ also to be

slowly varying functions of time. Calculating the averaged Lagrangian as in

eq. (2.56) we have

¯

L =

mA

2

4

ω

2

−ω

2

0

. (2.63)

The Euler–Lagrange equation eq. (2.59) from varying A gives

∂

¯

L

∂A

=

mA

2

ω

2

−ω

2

0

= 0 . (2.64)

Since we assume that the amplitude is not zero, this implies that the insta-

neous frequency tracks the changing natural frequency, ω = ±ω

0

(t) (which

also implies that

¯

L = 0).

The adiabatically conserved action J, eq. (2.60), is

J =

mωA

2

2

. (2.65)

From the adiabatic invariance of J we see that A varies as the inverse of ω

1/2

0

.

Remark 2.1 Our derivation of adiabatic invariance has been carried out

only to ﬁrst order in , the small parameter expressing the ratio between the

short oscillatory timescale and the long slow timescale. However, it can be

shown that an averaged Lagrangian asymptotically independent of φ to all

orders in a power series in can be found. Thus an adiabatic invariant

can be deﬁned to arbitrary order in . (This does not however mean that

adiabatic invariance is exact, since asymptotically small changes, e.g. going

as exp(−1/[[), are not ruled out.)

2.6 Transformations of the Lagrangian

2.6.1 Point transformations

Given an arbitrary Lagrangian L(q, ˙ q, t) in one generalized coordinate sys-

tem, q ≡ ¦q

i

[i = 1, n¦ (e.g. a Cartesian frame), we often want to know

2.6. TRANSFORMATIONS OF THE LAGRANGIAN 33

the Lagrangian L

(Q,

˙

Q, t) in another generalized coordinate system, Q ≡

¦Q

i

[i = 1, n¦ (e.g. polar coordinates as in the examples in Sec. 2.4.2 and

Sec. 2.4.3). Thus, suppose there exists a set g ≡ ¦g

i

[i = 1, n¦ of twice-

diﬀerentiable functions g

i

such that

q

i

= g

i

(Q, t) , i = 1, . . . , n . (2.66)

We require the inverse function of g also to be twice diﬀerentiable, in which

case g : Q → q is said to be a C

2

diﬀeomorphism. Note that we have

allowed the transformation to be time dependent, so transformations to a

moving frame are allowed.

The transformation g maps a path in Q-space to a path in q-space.

However, it is physically the same path—all we have changed is its represen-

tation. What we need in order to discuss how the Lagrangian transforms is a

coordinate-free formulation of Lagrangian dynamics. This is another virtue

of Hamilton’s Principle— δS = 0 on a physical path for all variations with

ﬁxed endpoints—since the action integral S given by eq. (2.26) is an integral

over time only. Thus, if we can deﬁne L

so that

L

(Q(t),

˙

Q(t), t) = L(q(t), ˙ q(t), t) (2.67)

for any path, then S is automatically invariant under the coordinate change

and will be stationary for the same physical paths, irrespective of what co-

ordinates they are represented in.

We can guarantee this trivially, simply by choosing the new Lagrangian

to be the old one in the new coordinates:

L

(Q,

˙

Q, t) = L(g(Q, t), ˙ g(Q,

˙

Q, t), t) , (2.68)

where

˙ g

i

≡

∂g

i

∂t

+

n

¸

j=1

˙

Q

j

∂g

i

∂Q

j

, i = 1, . . . , n . (2.69)

One can prove that eq. (2.68) gives the correct dynamics by calculat-

ing the transformation of the Euler–Lagrange equations explicitly (see e.g.

Scheck, 1990) and showing that eq. (2.24) implies

d

dt

∂L

∂

˙

Q

i

−

∂L

∂Q

i

= 0 , (2.70)

but clearly Hamilton’s Principle provides a much simpler and more elegant

way of arriving at the same result, since eqs. 2.70 are simply the Euler–

Lagrange equations for S to be stationary in the new variables.

We have in fact already run into a special case of this transformation rule,

in Sec. 2.4.2 and Sec. 2.4.3, where we transformed the kinetic and potential

energies from Cartesians to polars in the obvious way.

34 CHAPTER 2. LAGRANGIAN MECHANICS

2.6.2 Gauge transformations

We saw in Sec. 2.4.5 that we cannot always expect that the Lagrangian is

of the form T − V , but nevertheless found a function for which Lagrange’s

equations gave the correct dynamical equations of motion. Thus it is the

requirement that the equations of motion be in the form of Lagrange’s equa-

tions (or, equivalently, that Hamilton’s Principle apply) that is fundamental,

rather than the speciﬁc form of L.

That naturally raises the question: for a given system is there only one

Lagrangian giving the correct equations of motion, or are there many?

Clearly there is a trivial way to generate multiple physically equivalent

Lagrangians, and that is to multiply L by a constant factor. However, we

usually normalize L in a natural way, e.g. by requiring that the part linear

in the mass be equal to the kinetic energy, so this freedom is not encountered

much in practice.

However, there is a more important source of nonuniqueness, known as a

gauge transformation of the Lagrangian in which L is replaced by L

, deﬁned

by

L

(q, ˙ q, t) = L(q, ˙ q, t) +

∂M

∂t

+ ˙ q·

∂M

∂q

(2.71)

[using the shorthand notation deﬁned in eq. (1.4)], where M(q, t) is any C

3

function. We recognize the term in M as the total time derivative,

˙

M, and

thus, from eq. (2.26) and eq. (2.71), the new action integral becomes

S

≡

t

2

t

1

dt L

( ˙ q, q, t)

=

t

2

t

1

dt

¸

L( ˙ q, q, t) +

d

dt

M(q, t)

= S +M(q(t

2

), t

2

) −M(q(t

1

), t

1

) . (2.72)

However, since the endpoints are held ﬁxed in Hamilton’s Principle, the

endpoint terms M(q(t

1,2

), t

1,2

) do not aﬀect the allowed variations of S (i.e.

δS

≡ δS for all admissible variations) and therefore S

is stationary if and

only if S is. Thus both Lagrangians give identical physical paths and L and

L

**are completely equivalent dynamically.
**

Example 1: Harmonic oscillator

Consider the harmonic oscillator Lagrangian, eq. (2.61), for a particle of unit

mass

L =

1

2

( ˙ x

2

−ω

2

0

x

2

) , (2.73)

2.7. SYMMETRIES AND NOETHER’S THEOREM 35

where we take ω

0

to be constant. Since ∂L/∂ ˙ x = ˙ x and ∂L/∂x = ω

2

0

x, we

immediately verify that Lagrange’s equation, eq. (2.24), gives the harmonic

oscillator equation

¨ x = −ω

2

0

x . (2.74)

Now add a gauge term, taking M =

1

2

ω

0

x

2

. Then, from eq. (2.71), the

new Lagrangian becomes

L

=

1

2

( ˙ x

2

+ 2ω

0

x ˙ x −ω

2

0

x

2

) . (2.75)

Calculating ∂L

/∂ ˙ x = ˙ x+ω

0

x, and ∂L

/∂x = ω

2

0

x+ω

0

˙ x and substituting into

the Lagrangian equation of motion, we see that the gauge contributions can-

cel. Thus we do indeed recover the harmonic oscillator equation, eq. (2.74),

and the new Lagrangian is a perfectly valid one despite the fact that it is no

longer in the natural form, T −V .

Example 2: Electromagnetic gauge transformation

It is well known that the scalar and vector potentials in eqs. (2.43) are not

unique, since the electric and magnetic ﬁelds are left unchanged by the gauge

transformation

A(r, t) →A

(r, t) ≡ A(r, t) +∇χ(r, t)

Φ(r, t) →Φ

(r, t) ≡ Φ(r, t) −∂

t

χ(r, t) , (2.76)

where χ is an arbitrary scalar function (a gauge potential ).

Using the gauge-transformed potentials Φ

and A

to deﬁne a new La-

grangian L

**in the same way as L was deﬁned by eq. (2.44) we have
**

L

=

1

2

m˙ r

2

−eΦ

+e˙ r·A

. (2.77)

Substituting eqs. (2.76) in eq. (2.77) we ﬁnd

L

= L +

∂(eχ)

∂t

+ ˙ r·∇(eχ) , (2.78)

which is exactly of the form eq. (2.71) with M = eχ. Thus electromagnetic

and Lagrangian gauge transformations are closely related.

2.7 Symmetries and Noether’s theorem

A Lagrangian L is said to have a continuous symmetry if it is invariant under

the transformation q → h

s

(q), where s is a real continuous parameter and

36 CHAPTER 2. LAGRANGIAN MECHANICS

h

s=0

is the identity transformation. That is, given any path (not necessarily

physical) q(t), L has the same value on all members of the continuous family

of paths q(t, s) ≡ h

s

(q(t)). For example, in the case of motion in a central

potential, the continuous symmetry is invariance of L under rotations: θ →

θ +s.

Comparing two inﬁnitesimally close paths we see that δL also possesses

this symmetry, so that if δS ≡

**δLdt = 0 on the path q(t) then it also van-
**

ishes for all members of the family q(t, s). Thus, from Hamilton’s Principle,

if q(t, 0) ≡ q(t) is a physical path, then so are all the paths q(t, s) generated

by the symmetry operation h

s

. Thus, given one solution of the equations

of motion we can use a symmetry to generate a continuous family of other

solutions.

We saw in Sec. 2.4.2 that a system with an ignorable coordinate always

has a conservation equation, giving rise to an integral of motion. This is a

special case of a general theorem proved by Emmy Noether in 1918

2

that

any time-independent continuous symmetry of the Lagrangian, L = L(q, ˙ q, t)

generates an integral of motion. This ability of the Lagrangian formalism to

generate conservation relations is one of its most important features.

Consider a family q(t, s) obeying the Lagrange equation of motion eq. (2.24),

which we write in the form

∂L

∂q

=

d

dt

∂L

∂ ˙ q

, (2.79)

where d/ dt denotes the total derivative along the path s = const.

Since L is constant over this family of paths, its derivative with respect

to s at each point in time vanishes

∂L

∂s

=

∂L

∂q

·

∂q

∂s

+

∂L

∂ ˙ q

·

∂ ˙ q

∂s

. (2.80)

Using eq. (2.79) to eliminate ∂L/∂q in eq. (2.80) and using the time-

independence of h to allow interchange of the order of diﬀerention with re-

spect to s and t we derive the conservation equation

d

dt

∂L

∂ ˙ q

·

∂q

∂s

+

∂L

∂ ˙ q

·

d

dt

∂q

∂s

≡

dI

dt

= 0 , (2.81)

where

I(q, ˙ q, t) ≡

∂L

∂ ˙ q

·

∂q

∂s

(2.82)

is the integral of motion I generated by the symmetry h. 2

2

English translation in Transport Theory and Statistical Physics, 1, 186 (1971).

2.8. PROBLEMS 37

For example, expressing L in polar coordinates we see that the inte-

gral of motion generated by rotational symmetry is the angular momentum,

eq. (2.35), but eq. (2.82) applies in any coordinate system. (Similarly, in-

variance under translations in the x-direction generates the conservation of

linear momentum in the x-direction, and so on.)

2.7.1 Time symmetry

For an autonomous system (i.e. one not subject to external forcing), L is

independent of time. That is, L possesses the symmetry of invariance with

respect to translations in t. At ﬁrst sight this seems very diﬀerent from the

spatial symmetries just discussed, but in fact it is possible to generate an

integral of motion (the energy integral ), using the trick of regarding t as a

generalized coordinate by making it a function of a new “time” parameter τ:

t = t(τ), q = q(τ). (E.g. τ could be the “proper time” in relativity theory.)

In order for the new action integral

¯

Ldτ to be the same as the old one,

the new Lagrangian

¯

L is deﬁned to be

¯

L

q,

dq

dτ

,

dt

dτ

≡ L

q,

1

dt/ dτ

dq

dτ

dt

dτ

. (2.83)

Then, using invariance under t → t + s, q → q, eq. (2.82) gives the

integral of motion

I = L − ˙ q·

∂L

∂ ˙ q

, (2.84)

where we have set t = τ after doing the partial diﬀerentiations, as this applies

on a physical path. For Lagrangians of the form T − V is easily seen that

I = −T − V , so I can be identiﬁed as the negative of the total energy (cf.

Sec. 3.3).

2.8 Problems

2.8.1 Coriolis force and cyclotron motion

As a model of the motion of a ﬂuid element or dust particle in a planetary

(e.g. Earth’s) atmosphere, consider the motion of particle of unit mass con-

strained to move on the surface of a perfectly smooth sphere of radius R

rotating with angular velocity Ω about the z-axis. Suppose the force on the

particle is given by an eﬀective potential V (θ, φ) (see Sec. 2.4.6), where θ and

φ are the latitude and longitude respectively.

38 CHAPTER 2. LAGRANGIAN MECHANICS

(a) Write down the Lagrangian in a frame rotating with the planet, taking the

generalized coordinates to be the latitude and longitude so that z = Rsin θ,

x = Rcos θ cos(φ+Ωt), y = Rcos θ sin(φ+Ωt), where x, y, z is a non-rotating

Cartesian frame. (Draw a diagram illustrating these coordinates, noting that

the origin of the θ coordinate is at the equator, not at the pole as in normal

spherical polar coordinates.)

Write down the equations of motion, and ﬁnd a ﬁrst integral (i.e. con-

stant of the motion) in the case where V is independent of longitude. Among

this class of potentials ﬁnd the special case V = V

0

(θ) required to make equi-

librium possible (i.e. so that the equations of motion admit the solution

˙

θ =

˙

φ = 0 at each latitude).

Now assume that the general form of the potential is V (θ, φ) = V

0

(θ) +

h(θ, φ), where h is the enthalpy (proportional to p

(γ−1)/γ

) deﬁned in Sec. 2.4.6.

Assuming the winds are slow, so that second order time derivatives of θ and φ

(and products of ﬁrst order time derivatives) can be neglected, show that the

velocity of the particle in the rotating frame is at right angles to the pressure

gradient −∇p. (Hint: The pressure gradient is parallel to the enthalpy

gradient. You don’t have to work out the velocity and gradient vectors in

the rotating angular coordinates. It is much simpler to work with the scalar

quantity w·∇p, where w is the “wind velocity”, i.e. the particle velocity

with respect to the rotating frame.)

Draw a sketch of a typical weather-map “low” (i.e. a localized depression

in p) showing the pressure contours, the direction of the force on a ﬂuid ele-

ment and the direction of motion. Hence show that motion in a depression

or low is cyclonic, where “cyclonic” in geophysical ﬂuid dynamics means “in

the direction of the planet’s rotation” (i.e. clockwise in the Southern Hemi-

sphere, counter-clockwise in the Northern Hemisphere for Ω > 0).

(c) Consider a charged particle constrained to move on a non-rotating smooth

insulating sphere, immersed in a uniform magnetic ﬁeld B = Be

z

, on which

the electrostatic potential is a function of latitude and longitude. Write down

the Lagrangian in the same generalized coordinates as above and show it is

the same as that for the particle on the rotating planet with appropriate

identiﬁcations of Ω and V .

2.8.2 Anharmonic oscillator

Consider the following potential V , corresponding to a particle of mass m

oscillating along the x-axis under the inﬂuence of a nonideal spring (i.e. one

2.8. PROBLEMS 39

with a nonlinear restoring force),

V (x) =

mω

2

0

2

x

2

+ σ

x

4

l

2

0

,

where the constant ω

0

is the angular frequency of oscillations having ampli-

tude small compared with the characteristic length l

0

, and σ = ±1 depends

on whether the spring is “soft” (σ = −1) or “hard” (σ = +1).

Consider the trial function

x = l

0

[Acos ωt +Bcos 3ωt +C sin 3ωt] ,

where A, B, C are the nondimensionalized amplitudes of the fundamental

and third harmonic, respectively, and ω is the nonlinearly shifted frequency.

By using this trial function in the time-averaged Hamilton’s Principle, ﬁnd

implicit relations giving approximate expressions for ω, B and C as functions

of A. Show that C ≡ 0. The trial function is strictly appropriate only to the

case A < 1, but plot ω/ω

0

and B vs. A from 0 to 1 in the case of both a

hard and a soft spring. (You are encouraged to use Maple or Mathematica

and/or MatLab in this problem.)

40 CHAPTER 2. LAGRANGIAN MECHANICS

Chapter 3

Hamiltonian Mechanics

3.1 Introduction: Dynamical systems

Mathematically, a continuous-time dynamical system is deﬁned to be a sys-

tem of ﬁrst order diﬀerential equations

˙ z = f(z, t) , t ∈ R , (3.1)

where f is known as the vector ﬁeld and R is the set of real numbers. The

space in which the set of time-dependent variables z is deﬁned is called phase

space.

Sometimes we also talk about a discrete-time dynamical system. This is

a recursion relation, diﬀerence equation or iterated map

z

t+1

= f(z, t) , t ∈ Z , (3.2)

where f is known as the map (or mapping) and Z is the set of all integers

¦. . . , −2, −1, 0, 1, 2, . . .¦.

Typically, such systems exhibit long-time phenomena like attracting and

repelling ﬁxed points and limit cycles, or more complex structures such as

strange attractors. In this chapter we show how to reformulate nondissipative

Lagrangian mechanics as a dynamical system, but shall ﬁnd that it is a very

special case where the above-mentioned phenomena cannot occur.

3.2 Mechanics as a dynamical system

3.2.1 Lagrangian method

Lagrange’s equations do not form a dynamical system, because they implic-

itly contain second-order derivatives, ¨ q. However, there is a standard way to

41

42 CHAPTER 3. HAMILTONIAN MECHANICS

obtain a system of ﬁrst-order equations from a second-order system, which

is to double the size of the space of time-dependent variables by treating

the generalized velocities u as independent of the generalized coordinates, so

that the dynamical system is ˙ q = u, ˙ u = ¨ q(q, u, t). Then the phase space

is of dimension 2n. This trick is used very frequently in numerical problems,

because the standard numerical integrators require the problem to be posed

in terms of systems of ﬁrst-order diﬀerential equations.

In the particular case of Lagrangian mechanics, we have from eq. (2.24),

expanding out the total derivative using the chain rule and moving all but

the highest-order time derivatives to the right-hand side,

n

¸

j=1

∂

2

L

∂ ˙ q

i

∂ ˙ q

j

¨ q

j

=

∂L

∂q

i

−

∂

2

L

∂ ˙ q

i

∂t

−

n

¸

j=1

∂

2

L

∂ ˙ q

i

∂q

j

˙ q

j

. (3.3)

The matrix H acting on ¨ q, whose elements are given by

H

i,j

≡

∂

2

L

∂ ˙ q

i

∂ ˙ q

j

, (3.4)

is called the Hessian matrix. It is a kind of generalized mass tensor (see

Sec. 3.2.6), and for our method to work we require it to be nonsingular, so

that its inverse, H

−1

, exists and we can ﬁnd ¨ q. Then our dynamical system

becomes

˙ q = u ,

˙ u = H

−1

·

¸

∂L

∂q

−

∂

2

L

∂ ˙ q∂t

−

∂

2

L

∂ ˙ q∂q

· ˙ q

. (3.5)

Remark 3.1 The Lagrangian method per se does not break down if the Hes-

sian is singular, only our attempt to force it into the standard dynamical

system framework. We can still formally solve the dynamics in the follow-

ing manner: Suppose H is singular, with rank n − m. Then, within the n-

dimensional linear vector space ^ on which H acts, there is an m-dimensional

subspace ´ (the nullspace) such that H·¨ q ≡ 0 for all ¨ q ∈ ´. We can solve

eq. (3.3) for the component of ¨ q lying in the complementary subspace ^ `´

provided the right-hand side satisﬁes the solubility condition that it have no

component in ´. The component of ¨ q lying in ´ cannot be found directly,

but the solubility condition provides m constraints that complete the deter-

mination of the dynamics.

As a simple example, suppose L does not depend on one of the generalized

velocities, ˙ q

i

. Then ∂L/∂ ˙ q

i

≡ 0 and the ith component of the Lagrange

equations of motion eq. (2.24) reduces to the constraint ∂L/∂q

i

= 0. 2

3.2. MECHANICS AS A DYNAMICAL SYSTEM 43

3.2.2 Hamiltonian method

We can achieve our aim of ﬁnding 2n ﬁrst-order diﬀerential equations by

using many choices of auxiliary variables other than u. These will be more

complicated functions of the generalized velocities, but the extra freedom of

choice may also bring advantages.

In particular, Hamilton realised that it is very natural to use as the new

auxiliary variables the set p = ¦p

i

[i = 1, . . . , n¦ deﬁned by

p

i

≡

∂

∂ ˙ q

i

L(q, ˙ q, t) , (3.6)

where p

i

is called the generalized momentum canonically conjugate to q

i

.

We shall always assume that eq. (3.6) can be solved to give ˙ q as a function

of q and p

˙ q = u(q, p, t) . (3.7)

Remark 3.2 We have in eﬀect changed variables from u to p, and such a

change of variables can only be performed if the Jacobian matrix ∂p

i

/∂u

j

=

∂

2

L/∂ ˙ q

i

∂ ˙ q

j

is nonsingular. From eq. (3.4) we recognize this matrix as being

the Hessian we encountered in the Lagrangian approach to constructing a

dynamical system. Thus in either approach we require the Hessian to be

nonsingular (i.e. for its determinant to be nonzero). This condition is usually

trivially satisﬁed, but there are physical problems (e.g. if the Lagrangian does

not depend on one of the generalized velocities) when this is not the case.

However, in the standard Hamiltonian theory covered in this course it is

always assumed to hold.

The reason for deﬁning p as in eq. (3.6) is that ∂L/∂ ˙ q occurs explicitly

in Lagrange’s equations, eq. (2.24), so we immediately get an equation of

motion for p

˙ p =

∂L(q, ˙ q, t)

∂q

˙ q=u(q,p,t)

. (3.8)

Equations (3.7) and (3.8) do indeed form a dynamical system, but so far

it looks rather unsatisfactory: now u is deﬁned only implicitly as a function

of the phase-space variables q and p, yet the right-hand side of eq. (3.8)

involves a partial derivative in which the q-dependence of u is ignored!

We can ﬁx the latter problem by holding p ﬁxed in partial derivatives

with respect to q (because it is an independent phase-space variable) but

then subtracting a correction term to cancel the contribution coming from

44 CHAPTER 3. HAMILTONIAN MECHANICS

the q-dependence of u. Applying the chain rule and then using eqs. (3.6)

and (3.7) we get

˙ p =

∂L(q, u, t)

∂q

−

n

¸

i=1

∂L

∂u

i

∂u

i

∂q

=

∂L(q, u, t)

∂q

−

n

¸

i=1

p

i

∂u

i

∂q

=

∂

∂q

[L(q, u, t) −p·u]

≡ −

∂H

∂q

, (3.9)

where we have deﬁned a new function to replace the Lagrangian, namely the

Hamiltonian

H(q, p, t) ≡ p·u −L(q, u, t) . (3.10)

Given the importance of ∂H/∂q it is natural to investigate whether

∂H/∂p plays a signiﬁcant role as well. Diﬀerentiating eq. (3.10) we get

∂H

∂p

= u(q, p, t) +

n

¸

i=1

¸

p

i

−

∂

∂u

i

L(q, u, t)

∂u

i

∂p

= ˙ q , (3.11)

where the vanishing of the expression in the square bracket and the identiﬁ-

cation of u with ˙ q follows from eqs. (3.6) and (3.7).

Summarizing eqs. (3.9) and (3.11),

˙ q =

∂H

∂p

˙ p = −

∂H

∂q

, (3.12)

These equations are known as Hamilton’s equations of motion. As with

Lagrange’s equations they express the dynamics of a system with an arbitrary

number of degrees of freedom in terms of a single scalar function! Unlike the

Lagrangian dynamical system, the phase-space variables are treated on a

completely even footing—in Hamiltonian mechanics both the conﬁguration-

space variables q and momentum-space variables p are generalized coordi-

nates. We deﬁne canonical coordinates as phase-space coordinates such that

the equations of motion can be expressed in the form of eq. (3.12) and a

canonical system as one for which canonical coordinates exist.

3.2. MECHANICS AS A DYNAMICAL SYSTEM 45

Remark 3.3 The transition from the Lagrangian to the Hamiltonian in or-

der to handle the changed meaning of partial derivatives after a change of

variable is a special case of a technique known as a Legendre transformation.

It is encountered quite often in physics and physical chemistry, especially in

thermodynamics.

3.2.3 Example 1: Scalar potential

Consider the Lagrangian for a particle in Cartesian coordinates, so q =

¦x, y, z¦ may be replaced by r = xe

x

+ye

y

+ze

z

. Also assume that it moves

under the inﬂuence of a scalar potential V (r, t) so that the natural form of

the Lagrangian is

L = T −V =

1

2

m[ ˙ r[

2

−V (r, t) . (3.13)

Then, from eq. (3.6)

p →p ≡

∂L

∂˙ r

= m˙ r , (3.14)

so that in this case the canonical momentum is the same as the ordinary

kinematic momentum. Equation (3.14) is solved trivially to give ˙ q = u(p)

where u(p) = p/m. Thus, from eq. (3.10) we have

H =

[p[

2

m

−

[p[

2

2m

−V (r, t)

=

[p[

2

2m

+V (r, t)

= T +V . (3.15)

That is, the Hamiltonian is equal to the total energy of the system, ki-

netic plus potential. The fact that the Hamiltonian is an important physical

quantity, whereas the physical meaning of the Lagrangian is more obscure,

is one of the appealing features of the Hamiltonian approach. Both the

Lagrangian and Hamiltonian have the dimensions of energy, and both ap-

proaches can be called energy methods. They are characterized by the use

of scalar quantities rather than the vectors encountered in the direct use of

Newton’s second law. This has both the theoretical advantage of leading to

very general formulations of mechanics and the practical beneﬁt of avoiding

some vector manipulations when changing between coordinate systems (in

fact, Lagrangian and Hamiltonian methods were developed before modern

vector notation was invented).

46 CHAPTER 3. HAMILTONIAN MECHANICS

Harmonic Oscillator

An example is the harmonic oscillator Hamiltonian corresponding to the

Lagrangian, eq. (2.61)

H =

p

2

2m

+

mω

0

x

2

2

. (3.16)

From eq. (3.12) the Hamiltonian equations of motion are

˙ x =

p

m

˙ p = −mω

0

x . (3.17)

Gauge-transformed Harmonic Oscillator

Now consider the gauge-transformed harmonic oscillator Lagrangian eq. (2.75)

L

=

1

2

m( ˙ x

2

+ 2ω

0

x ˙ x −ω

2

0

x

2

) .

The canonical momentum is thus

p = ∂L

/∂ ˙ x = m( ˙ x +ω

0

x) (3.18)

and we see that the gauge transformation has eﬀected a transformation of

the canonical momentum, even though the generalized coordinate x remains

the same. This is an example of a canonical transformation, about which we

will have more to say later.

Solving eq. (3.18) for ˙ x we ﬁnd u(p) = (p −mω

0

x)/m. Hence

H = p

(p −mω

0

x)

m

−

(p −mω

0

x)

2

2m

+ω

0

x(p −mω

0

x) −

1

2

mω

2

0

x

2

=

(p −mω

0

x)

2

2m

+

1

2

mω

2

0

x

2

= T +V . (3.19)

Thus, even though L was not of the natural form T − V in this case, the

Hamiltonian remains equal to the total energy, thus conﬁrming that it is a

quantity with a more direct physical signifance than the Lagrangian. (The

functional form of the Hamiltonian changes under the gauge transformation

however, because the meaning of p changes.)

3.2. MECHANICS AS A DYNAMICAL SYSTEM 47

3.2.4 Example 2: Physical pendulum

A nonlinear one-dimensional case is provided by the physical pendulum, in-

troduced in Sec. 2.4.1. Using eq. (2.27) in eq. (3.6) we get p

θ

= ml

2

˙

θ. Thus

˙

θ = p

θ

/ml

2

and so the Hamiltonian, H = p

θ

˙

θ −L, becomes

H(θ, p

θ

) =

p

2

θ

2ml

2

+ mgl(1 −cos θ) , (3.20)

which again is of the form T +V .

By conservation of energy (see also Sec. 3.3),

0 0.2 0.4 0.6 0.8 1

-3

-2

-1

0

1

2

3

θ/2π

p

θ

X O O

Figure 3.1: Phase

space of the phys-

ical pendulum.

the Hamiltonian H = T + V is a constant

of the motion so the nature of the orbits in

phase space can be found simply by plotting

the contours of H as in Fig. 3.1 (which is in

units such that m = l = g = 1). We see that

the structure of the phase space is rather more

complicated than in the case of the harmonic

oscillator since there are two topologically dis-

tinct classes of orbit. One class is the rotating

orbits for which the pendulum has enough en-

ergy, H > 2mgl, to swing entirely over the top

so θ increases or decreases secularly (though

the physical position x = l sin θ does not).

The other class is the librating orbits for H < 2mgl implying that the

pendulum is trapped in the gravitational potential well and oscillates like a

pair of scales (hence the name). For [θ[ < 1 we may expand the cosine so

V ≈ mglθ

2

/2 and the system is approximately a harmonic oscillator. The

equilibrium point θ = 0 or 2π, labelled O in Fig. 3.1, is a ﬁxed point. The

orbits in its neighbourhood, like those of the harmonic oscillator, remain in

the neighbourhood for all time (i.e. the ﬁxed point is stable or elliptic).

The dividing line H = 2mgl between the two topological classes of orbit is

called the separatrix, and on the separatrix lies another ﬁxed point, labelled

X in Fig. 3.1. This corresponds to the case where the pendulum is just

balanced “upside down”. Almost all orbits in the neighbourhood of an X

point eventually are repelled far away from it, and thus it is referred to as

an unstable or hyperbolic ﬁxed point.

48 CHAPTER 3. HAMILTONIAN MECHANICS

3.2.5 Example 3: Motion in e.m. potentials

Now consider the case of a charged particle in an electromagnetic ﬁeld with

magnetic vector potential A and electrostatic potential Φ. The Lagrangian

is given by eq. (2.44) and thus, from eq. (3.6),

p ≡

∂L

∂˙ r

= m˙ r +eA(r, t) . (3.21)

Thus u(p) = p/m−eA/m and, from eq. (3.10) we have

H =

(p −eA)·p

m

−

[p −eA[

2

2m

+

e(p −eA)·A

m

−eΦ(r, t)

=

[p −eA(r, t)[

2

2m

+eΦ(r, t)

= T +V . (3.22)

Thus we ﬁnd again that, although the Lagrangian cannot be put into the

natural form T −V (r), the Hamiltonian is still the total energy, kinetic plus

electrostatic potential energy.

3.2.6 Example 4: The generalized N-body system

Let us now revisit the general case which led to the original deﬁnition of the

Lagrangian in Sec. 2.3, the case of N particles interacting via a scalar N-body

potential, possibly with constraints (which we here assume to be independent

of time), so that the number of generalized coordinates is n ≤ 3N. Then, by

eq. (2.25) and eq. (2.7) (assuming the function x

k

(q) to be independent of

time) the natural form of the Lagrangian is

L = T −V

=

1

2

n

¸

i,j=1

N

¸

k=1

m

k

˙ q

i

∂x

k

∂q

i

·

∂x

k

∂q

j

˙ q

j

−V (q)

=

1

2

n

¸

i,j=1

˙ q

i

µ

i,j

˙ q

j

−V (q) , (3.23)

where the symmetric mass matrix

µ

i,j

(q) ≡

N

¸

k=1

m

k

∂x

k

∂q

i

·

∂x

k

∂q

j

(3.24)

3.2. MECHANICS AS A DYNAMICAL SYSTEM 49

is the metric tensor for a conﬁguration-space mass-weighted “length” element

ds deﬁned by ( ds)

2

≡

¸

k

m

k

dl

2

k

. From eq. (3.4), we see that it is the Hessian

matrix for this system.

Then, using eq. (3.23) in eq. (3.6)

p

i

=

n

¸

j=1

µ

i,j

(q) ˙ q

j

. (3.25)

Assuming none of the particles is massless, µ

i,j

is a positive-deﬁnite ma-

trix, so its inverse µ

−1

i,j

exists and we can formally solve eq. (3.25) for ˙ q

i

to

give

u

i

(q, p, t) =

n

¸

j=1

µ

−1

i,j

p

j

. (3.26)

Then, from eq. (3.10) we have

H =

n

¸

i,j=1

p

i

µ

−1

i,j

p

j

−

1

2

n

¸

i,j,i

,j

=1

p

i

µ

−1

i,i

µ

i

,j

µ

−1

j

,j

p

j

−V (q)

=

1

2

n

¸

i,j=1

p

i

µ

−1

i,j

p

j

+V (q)

= T +V . (3.27)

Thus, the Hamiltonian is again equal to the total energy of the system. This

result does not hold in the case of a time-dependent representation, x

k

(q, t).

[See Problems 3.5.1(c) and 4.7.1.]

Particle in a central potential

As a simple, two-dimensional example of a problem in non-Cartesion coor-

dinates we return to the problem of motion in a central potential, expressed

in plane polar coordinates in Sec. 2.4.2. Recapitulating eq. (2.33),

L =

1

2

m

˙ r

2

+r

2

˙

θ

2

−V (r) .

Comparing with eq. (3.23) we see that the mass matrix is diagonal

µ =

¸

m 0

0 mr

2

, (3.28)

and thus can be inverted simply taking the reciprocal of the diagonal ele-

ments. Hence, from eq. (3.27)

H =

p

2

r

2m

+

p

2

θ

2mr

2

+V (r) . (3.29)

50 CHAPTER 3. HAMILTONIAN MECHANICS

3.3 Time-Dependent and Autonomous Hamil-

tonian systems

An autonomous dynamical system is one in which the vector ﬁeld f [see

eq. (3.1)] depends only on the phase-space coordinates and has no explicit

dependence on t. (The reason for the name is that explicit time dependence

would come from external forcing of the system, whereas “autonomous”

means independent or self-governing.) In the Hamiltonian case this means

that H has no explicit time dependence. Conversely, if there is an external

time-varying perturbation, then H = H(p, q, t).

Consider the time rate of change of the Hamiltonian,

˙

H ≡ dH/ dt, follow-

ing the phase-space trajectory. Using the Hamiltonian equations of motion,

eq. (3.12),

dH

dt

=

∂H

∂t

+ ˙ q·

∂H

∂q

+ ˙ p·

∂H

∂p

=

∂H

∂t

+

∂H

∂p

·

∂H

∂q

−

∂H

∂q

·

∂H

∂p

≡

∂H

∂t

. (3.30)

Thus, if H does not depend explicitly on time, so its partial time deriva-

tive vanishes, then its total time derivative also vanishes. That is, in an

autonomous Hamiltonian system the Hamiltonian is an integral of motion.

We have seen above that H can often be identiﬁed as the total energy.

When this is the case we may interpret the completely general result above as

a statement of conservation of energy. More generally, comparing eq. (3.10)

and eq. (2.84) we recognize H as the negative of the energy integral I for

autonomous systems as predicted by Noether’s theorem.

3.4 Hamilton’s Principle in phase space

Let us express the action integral S in terms of q and p and show that we

may derive Hamilton’s equations of motion by making S stationary under

variations of the trajectory in phase space, rather than in conﬁguration space.

To ﬁnd a suitable deﬁnition for the action integral, we ﬁrst rearrange

eq. (3.10) to get L on the left-hand side. Then we replace u(q, p, t) by ˙ q in

the term p·u and thus deﬁne the phase-space Lagrangian

L

ph

(q, ˙ q, p, t) ≡ p· ˙ q −H(q, p, t) , (3.31)

3.4. HAMILTON’S PRINCIPLE IN PHASE SPACE 51

If ˙ q = u(q, p, t) were identically satisﬁed, even on arbitrarily varied phase-

space paths, then L

ph

would simply be L expressed in phase-space coordi-

nates. However, one can easily construct a counter example to show that

this is not the case: consider a variation of the path in which we can vary

the direction of its tangent vector, at some point (q, p), while keeping this

point ﬁxed (see Sec. 4.2 for an illustration of this). Then ˙ q changes, but u

remains the same. Thus L

ph

and L are the same value only on the subset of

paths (which includes the physical paths) for which p· ˙ q = p·u(q, p, t).

Replacing L by L

ph

in eq. (2.26) we deﬁne the phase-space action integral

S

ph

[q, p] =

t

2

t

1

dt L

ph

(q, p, ˙ q, t) . (3.32)

We shall now show that requiring δS

ph

= 0 for arbitrary variations about a

given path implies that that path is such that Hamilton’s equations of motion

are satisﬁed at all points along it. Since we concluded above that L

ph

and L

were not the same on arbitrary nonphysical paths, this variational principle

is subtly diﬀerent from the Lagrangian version of Hamilton’s principle, and

is sometimes called the modifed Hamilton’s Principle.

We know from variational calculus that S

ph

is stationary under arbitrary

variations of the phase-space path (with endpoints ﬁxed) if and only if the

Euler–Lagrange equations δL

ph

/δq = δL

ph

/δp = 0 are satisﬁed at all points

on the path, where the variational derivatives are as deﬁned in eq. (1.14),

with p now being regarded as an independent coordinate

δL

ph

δq

=

∂L

ph

∂q

−

d

dt

∂L

ph

∂ ˙ q

= −

∂H

∂q

− ˙ p = 0 ,

δL

ph

δp

=

∂L

ph

∂p

= ˙ q −

∂H

∂p

= 0 . (3.33)

These are identical to Hamilton’s equations of motion, eq. (3.12), as claimed.

Remark 3.4 Note that, although we ﬁx the endpoints in the modiﬁed Hamil-

ton’s Principle, we will not be able to satisfy Hamilton’s equations of motion

at all points on the path unless the endpoints are both chosen to lie on a

physical path. This is because the ﬁxing of the initial phase-space point speci-

ﬁes 2n initial conditions, which completely determines the physical path (i.e.

the one that obeys the equations of motion). This is in contrast with the

the Lagrangian version of Hamilton’s Principle, where we ﬁxed only the n

conﬁguration-space coordinates q. The problem of ﬁnding the physical path

between two given conﬁguration-space points thus involves specifying only n

initial and n ﬁnal data, which is generally exactly the right number to allow

52 CHAPTER 3. HAMILTONIAN MECHANICS

us to ﬁnd one (or more) solutions of the equations of motion passing through

arbitrarily chosen initial and ﬁnal points.

3.5. PROBLEMS 53

3.5 Problems

3.5.1 Constraints and moving coordinates

(a) This revisits Problem 1.6.1. Write down a Lagrangian for the problem of

two particles of mass m

1

and m

2

connected by a light rigid rod of length l in

a gravitational ﬁeld g. Take the generalized coordinates of the system to be

q = ¦x, y, z, θ, φ¦, with the coordinates of the two particles being given by

x

1

= x −α

1

l sin θ cos φ

y

1

= y −α

1

l sin θ sin φ

z

1

= z −α

1

l cos θ

x

2

= x +α

2

l sin θ cos φ

y

2

= y +α

2

l sin θ sin φ

z

2

= z + α

2

l cos θ ,

where α

1

≡ m

2

/(m

1

+ m

2

) and α

2

≡ m

1

/(m

1

+ m

2

) [so that (x, y, z) is the

centre of mass].

(b) Use this Lagrangian to construct the Hamiltonian for the system. Is it

of the form T +V ?

(c) Find the Hamiltonian corresponding to the Coriolis Lagrangian in Prob-

lem 2.8.1,

L =

1

2

R

2

˙

θ

2

+ cos

2

θ (Ω +

˙

φ)

2

−V (θ, φ) .

Is it of the form T +V ?

3.5.2 Anharmonic oscillator phase space

Find the Hamiltonian corresponding to the Lagrangian of the anharmonic

oscillator in Problem 2.8.2. Sketch contours of H(x, p) in the soft potential

case for (a) H < mω

2

0

l

2

0

/8 and (b) H > mω

2

0

l

2

0

/8. Indicate representative

trajectories of the system in phase-space, showing the direction of travel.

Try to include representatives of each qualitatively distinct orbit.

3.5.3 2-D motion in a magnetic ﬁeld

Consider the motion of a particle of charge e and mass m in a straight, in-

ﬁnitely long magnetic conﬁnement system with vector potential A = ψ(x, y)e

z

.

54 CHAPTER 3. HAMILTONIAN MECHANICS

(a) Show that B

x

= ∂ψ/∂y, B

y

= −∂ψ/∂x, B

z

= 0 and thus that contours

of ψ(x, y) deﬁne magnetic ﬁeld lines.

(b) Show that the Hamiltonian is

H =

p

2

x

2m

+

p

2

y

2m

+

(p

z

−eψ)

2

2m

,

write down the Hamiltonian equations of motion and give two integrals of

the motion.

(c) It can be shown that in any region with no electric current, ψ(x, y) can be

represented as the real part of an analytic function of the complex variable

ζ ≡ x + iy, ψ(x, y) = Re Ψ(ζ). Show that the equations of motion for ζ and

p

ζ

≡ p

x

+ ip

y

are

˙

ζ =

1

m

p

ζ

,

˙ p

ζ

=

e

m

[p

z

−eRe Ψ(ζ)] [Ψ

(ζ)]

∗

,

where the prime on Ψ means derivative with respect to its argument, and ∗

means complex conjugate.

Chapter 4

Canonical transformations

4.1 Introduction

A canonical transformation is a transformation of the phase-space coordi-

nates q and p to new coordinates Q and P that obey Hamilton’s equations

of motion, eq. (3.12), with a new Hamiltonian K(Q, P, t).

The discussion in Sec. 3.2.6 shows the general form of p for an N-particle

system in an arbitrary, stationary generalized coordinate system. In this case,

if we change from one conﬁguration-space coordinate system, q, to another,

Q, then the transformation of p to P follows from comparing eq. (3.25) in the

two coordinate systems. For instance, if we go from coordinates x, y in a two-

dimensional system to polar coordinates r, θ, the canonical transformation

is

x = r cos θ

y = r sin θ

p

x

= m˙ x = m( ˙ r cos θ −r

˙

θ sin θ) = p

r

cos θ −

p

θ

sin θ

r

p

y

= m˙ y = m( ˙ r sin θ + r

˙

θ cos θ) = p

r

sin θ +

p

θ

cos θ

r

K(r, θ, p

r

, p

θ

) = H(x, y, p

x

, p

y

) . (4.1)

Thus one obvious class of canonical transformation is produced by a

change of conﬁguration-space coordinates, or point transformation, q =

g(Q, t) as discussed in the context of the general Lagrangian formalism in

Sec. 2.6. Applying eq. (2.68) and using eq. (2.69) we ﬁnd the general rela-

tionship

P

i

=

∂

∂

˙

Q

i

L(g(Q, t), ˙ g(Q,

˙

Q, t), t)

55

56 CHAPTER 4. CANONICAL TRANSFORMATIONS

= p·

∂ ˙ g

∂

˙

Q

i

=

n

¸

j=1

∂g

j

∂Q

i

p

j

. (4.2)

Thus by inverting the matrix T

i,j

≡ ∂g

j

/∂Q

i

, we complete the construction

of the canonical transformation by ﬁnding p = T

−1

·P. We can also show

by direct calculation that K(Q, P) = H(q, p) for time-independent trans-

formations as in eq. (4.1). The discussion of the time dependent case will be

left to the next section.

We have also encountered another special case of a canonical transforma-

tion (for which p was changed but q remained the same) in Sec. 3.2.3 where

we compared the Hamiltonian for the harmonic oscillator derived from the

natural form for the Lagrangian with that from a gauge-transformed version.

Likewise, in the case of the particle in an electromagnetic ﬁeld treated in

Sec. 3.2.5 an electromagnetic gauge transformation eﬀects a change in the

momenta but leaves the conﬁguration-space coordinates unaﬀected.

This suggests that gauge transformations may hold the clue to construct-

ing a general theory of canonical transformations and we show in the next

section that this is indeed the case.

4.2 Generating functions

To discuss gauge transformations in the Hamiltonian context we use the mod-

iﬁed Hamilton’s Principle based on the phase-space action integral eq. (3.32),

S

ph

[q, p] =

t

2

t

1

dt L

ph

(q, ˙ q, p, ˙ p, t) , (4.3)

which we showed to be stationary on physical paths with the form of phase-

space Lagrangian given in eq. (3.31),

L

(1)

ph

≡ p· ˙ q −H(q, p, t) . (4.4)

However, as in Sec. 2.6.2, we can ﬁnd a physically equivalent Lagrangian

by adding to L

(1)

ph

the total time derivative of an arbitrary function, M(q, p, t)

say. Although it clearly has no eﬀect on Hamilton’s equations of motion, and

thus does not generate a canonical transformation, we shall have use of the

particular choice M = −q·p. This gives the alternative, gauge-transformed

phase-space Lagrangian

L

(2)

ph

≡ −q· ˙ p −H(q, p, t) . (4.5)

4.2. GENERATING FUNCTIONS 57

Now assume we have a phase-space transformation (q, p) →(Q, P) where

the mapping between the two coordinates is invertible and suﬃciently smooth

that we can evaluate the required derivatives. Since we are seeking a canon-

ical transformation, there must also exist an action integral S

ph

of the form

S

ph

[Q, P] =

t

2

t

1

dt L

ph

(Q,

˙

Q, P,

˙

P, t) (4.6)

that is stationary under variations vanishing at the endpoints for the same

paths (represented in the new coordinates) as S

ph

is, where L

ph

is a phase-

space Lagrangian in either of the forms given in eqs. (4.4) or (4.5) with q, p

replaced by Q, P and H replaced by K. For example,

L

(1)

ph

(Q,

˙

Q, P, t) ≡ P·

˙

Q−K(Q, P, t) . (4.7)

As for the gauge transformation discussed above, we ensure simultaneous

stationarity of the two functionals by requiring that the old and new phase-

space Lagrangians diﬀer only by the total time derivative of some function,

which we denote by F,

L

ph

(q, ˙ q, p, ˙ p, t) = L

ph

(Q,

˙

Q, P,

˙

P, t) +

dF

dt

. (4.8)

This equality is to hold on all phase-space paths, not only the physical ones.

Do we take F to be a function of the old or of the new variables? The

traditional answer is to take F to be a mixed function of n of the old and

n of the new (since there are only 2n independent variables), giving four

possibilities

F

1

(q, Q, t), F

2

(q, P, t), F

3

(p, Q, t), F

4

(p, P, t) . (4.9)

These functions are known as the generating functions of types 1 to 4. We

require the two sets of coordinates used to be functionally independent, so

that F really is a function of 2n independent phase-space coordinates. This

condition rules out, for example, the use of the type 1 generating function for

point transformations, including the identity transformation. Otherwise the

choice of generating function is somewhat arbitrary and we choose whichever

is convenient for the problem.

The total time derivatives in the four cases are given by

dF

1

dt

= ˙ q·

∂F

1

∂q

+

˙

Q·

∂F

1

∂Q

+

∂F

1

∂t

(4.10)

dF

2

dt

= ˙ q·

∂F

2

∂q

+

˙

P·

∂F

2

∂P

+

∂F

2

∂t

(4.11)

58 CHAPTER 4. CANONICAL TRANSFORMATIONS

dF

3

dt

= ˙ p·

∂F

3

∂p

+

˙

Q·

∂F

3

∂Q

+

∂F

3

∂t

(4.12)

dF

4

dt

= ˙ p·

∂F

4

∂p

+

˙

P·

∂F

4

∂P

+

∂F

4

∂t

. (4.13)

The four types of generating function correspond to choosing the two

alternative forms, eqs. (4.4) and (4.5), on either side of the gauge transfor-

mation, eq. (4.8), as indicated in the following table:

L

(1)

ph

(q, ˙ q, p) = L

(1)

ph

(q, ˙ q, p) =

L

(1)

ph

(Q,

˙

Q, P) +

˙

F

1

(q, Q) , L

(2)

ph

(Q, P,

˙

P) +

˙

F

2

(q, P) ,

L

(2)

ph

(q, p, ˙ p) = L

(2)

ph

(q, p, ˙ p) =

L

(1)

ph

(Q,

˙

Q, P) +

˙

F

3

(p, Q) , L

(2)

ph

(Q, P,

˙

P) +

˙

F

4

(p, P) ,

(4.14)

where

˙

F

i

denotes dF

i

/ dt and we have suppressed explicit time dependences

for brevity.

q

1

p

1

q

2

...

p

n

z(t)

z(t

1

)

z(t

2

)

Figure 4.1: Some possible paths z(t) ≡ (q(t), p(t)) in phase space,

each passing through the same point at time t but with diﬀerent ˙ z.

As indicated in eq. (4.14), Type 1 generating functions combine most

naturally with the choice L

ph

= L

(1)

ph

, L

ph

= L

(1)

ph

, since both the Lagrangians

4.2. GENERATING FUNCTIONS 59

and

˙

F

1

involve only ˙ q and

˙

Q, but not ˙ p or

˙

P. Substituting eq. (4.10) in

eq. (4.8) with these choices and collecting terms in ˙ q and

˙

Q we get

p −

∂F

1

∂q

· ˙ q −

¸

H(q, p, t) +

∂F

1

∂t

=

P +

∂F

1

∂Q

·

˙

Q−K(Q, P, t) . (4.15)

Since eq. (4.8) is to hold on all paths, it must in particular hold at any given

phase-space point (q, p) for all possible paths passing through this point. As

illustrated in Fig. 4.1 the possible total time derivatives ( ˙ q, ˙ p) span the full

2n-dimensional space of tangent directions through the given point. Since we

require q and Q to be functionally independent, the possible choices of ˙ q and

˙

Q also span a 2n-dimensional space, and thus eq. (4.15) can hold identically

only if the coeﬃcients of ˙ q and

˙

Q vanish separately

p =

∂F

1

∂q

(4.16)

P = −

∂F

1

∂Q

, (4.17)

ﬁnally leaving

K = H +

∂F

1

∂t

. (4.18)

These equations deﬁne, albeit implicitly, a canonical transformation.

For type 2 generating functions we choose L

ph

= L

(1)

ph

, L

ph

= L

(2)

ph

so

that eq. (4.8) becomes

p −

∂F

2

∂q

· ˙ q −

¸

H(q, p, t) +

∂F

2

∂t

= −

Q−

∂F

2

∂P

·

˙

P −K(Q, P, t) . (4.19)

Equating the coeﬃcients of ˙ q and

˙

P to zero we get the implicit equations

deﬁning type 2 canonical transformations

p =

∂F

2

∂q

(4.20)

Q =

∂F

2

∂P

(4.21)

and

K = H +

∂F

2

∂t

. (4.22)

60 CHAPTER 4. CANONICAL TRANSFORMATIONS

For type 3 we choose L

ph

= L

(2)

ph

, L

ph

= L

(1)

ph

and ﬁnd

q = −

∂F

3

∂p

(4.23)

P = −

∂F

3

∂Q

(4.24)

and

K = H +

∂F

3

∂t

. (4.25)

For type 4 we choose L

ph

= L

(2)

ph

, L

ph

= L

(2)

ph

, to ﬁnd

q = −

∂F

4

∂p

(4.26)

Q =

∂F

4

∂P

(4.27)

and

K = H +

∂F

4

∂t

. (4.28)

4.2.1 Example 1: Adiabatic Oscillator

Consider the harmonic oscillator Lagrangian eq. (2.61) and Hamiltonian

eq. (3.16), taking ω

0

to be a slowly varying function of time which never

vanishes, ω

0

= ω

0

(t) = 0, where is small parameter. (Introducing is

a useful device for the formal asymptotic theory, in which we consider the

scaling of error terms as →0.)

Now consider the Hamiltonian, H = p

2

/2m + mω

2

0

x

2

/2 and try the type

1 generating function

F

1

=

1

2

mω

0

x

2

cot φ , (4.29)

with x playing the role of the old generalized coordinate, q, and φ playing

the role of the new generalized coordinate, Q. Using eq. (4.29) in eqs. (4.16)

and (4.17) we have

p = mω

0

x cot φ (4.30)

P =

1

2

mω

0

x

2

cosec

2

φ . (4.31)

Eliminating x between these equations we ﬁnd the canonical transformation

between x, p and φ, P

p = (2mω

0

P)

1/2

cos φ , (4.32)

4.2. GENERATING FUNCTIONS 61

x =

2P

mω

0

1/2

sin φ . (4.33)

Comparing eq. (4.33) with eq. (2.62) we identify P = mω

0

A

2

/2. Com-

paring with eq. (2.65) we see that P is in fact J, the action. Thus we should

be able to show its adiabatic invariance from the Hamiltonian equations of

motion. First we need the new Hamiltonian, which from eq. (4.18) is given

by

K =

p

2

2m

+

mω

2

0

x

2

2

+

m˙ ω

0

x

2

cot φ

2

= ω

0

P +

1

2

d ln ω

0

dt

P sin 2φ . (4.34)

The equations of motion are thus

˙

φ =

∂K

∂P

= ω

0

+

1

2

d ln ω

0

dt

sin 2φ (4.35)

and

˙

P = −

∂K

∂φ

=

d ln ω

0

dt

P cos 2φ . (4.36)

So far the results are exact, but to proceed further (unless we solve these

equations numerically) we need to make use of the expansion parameter .

Observe that d ln ω

0

/ dt = ω

0

(t)/ω

0

is small, O(). Thus to leading order

we may drop the sin 2φ term in eq. (4.35) because it is small in comparison

with ω

0

, so that we are making the same approximation, ω ≡

˙

φ ≈ ω

0

, as we

were led to in the averaged Lagrangian approach. Integrating the equation

to get the phase as a function of time we have

φ(t) ≈

t

t

0

ω

0

(t) , (4.37)

where t

0

is an arbitrary starting time. Thus the phase increases with time

without bound. Functions with this property are known as secular terms.

The right-hand side of eq. (4.36) is O() so we see that P is approximately

constant over short times. However we can say more than that since we can

estimate the long-term change by integrating the equation,

ln P(t) −ln P

0

≈

t

t

0

dt

d ln ω

0

dt

cos 2φ . (4.38)

Using eq. (4.37), we see that the integrand averages to zero over times long

compared with 1/ω

0

and thus the integral will also be oscillatory, with no

secular increase. Thus the canonical transformation method conﬁrms the

adiabatic invariance of the action.

62 CHAPTER 4. CANONICAL TRANSFORMATIONS

Remark 4.1 As implied by the remark at the end of Sec. 2.5, we could go on

to construct a formal asymptotic procedure in which the canonical transfor-

mation was successively reﬁned to remove the oscillatory terms in P at each

order in to produce an adiabatic invariant that is conserved to arbitrarily

high order (but not exactly, because of the possibility of asymptotically small

corrections “beyond all orders”).

Note that, apart from the small correction term ∂F

1

/∂t, K is still ap-

proximately the total energy of the oscillator, E ≡ T +V , so eq. (4.34) gives

P ≈ E/ω

0

. The quantum mechanical interpretation of this was pointed out

around 1911 by Einstein: If there are N >> 1 quanta of energy in the os-

cillator, then E = Nω

0

(ignoring the zero-point energy

1

2

ω

0

). Thus the

adiabatic invariance of P = N can be interpreted as saying the number of

quanta is conserved during adiabatic changes in the system.

4.2.2 Example 2: Point transformations

As an example of a type 2 generating function, take

F

2

(q, P, t) = G(q, t)·P , (4.39)

where G(q, t) is a function such that the matrix ∂G

j

/∂q

i

is nonsingular.

Then eqs. (4.20) and (4.21) give

p

i

=

n

¸

j=1

∂G

j

∂q

i

P

j

,

Q = G(q, t) (4.40)

The second equation shows that eq. (4.39) is the generating function for a

point transformation, and the ﬁrst equation is the same as eq. (4.2) with

the roles of the old and new variables interchanged. However, in the case

of a time-dependent transformation, the generating function method now

provides us with the transformation of the Hamiltonian, a useful result that

we did not have easily before. That is, substituting eq. (4.39) in eq. (4.22)

gives

K(Q, P, t) = H(q, p, t) +

∂G(q, t)

∂t

·P . (4.41)

Note that included among the point transformations is the identity trans-

formation. Thus, the class of canonical transformations that can be generated

by a type 2 generating function includes the identity, whereas the type 1 class

does not [though a type 1 transformation may come arbitrarily close to the

identity—see Problem 4.7.2(c) and consider the limit ∆t →0].

4.3. INFINITESIMAL CANONICAL TRANSFORMATIONS 63

4.3 Inﬁnitesimal canonical transformations

The class of canonical transformations continuously connected to the identity

transformation is important for several reasons—it includes time evolution,

continuous symmetry operations and canonical perturbation theory.

Since type 1 transformations do not include the identity, we consider the

type 2 generating functions F

2

(q, P, t) and note from eqs. (4.20) and (4.21)

that F

2

= q·P gives p = ∂F

2

/∂q = P and Q = ∂F

2

/∂P = q. That is, q·P

generates the identity transformation.

We now move continuously away from the identity among a family of

canonical transformations parameterized by , say, such that = 0 is the

identity. The generating function of such a family can be written in the form

F

2

= q·P +σ(q, P, t) +O(

2

) , (4.42)

where the notation O(

2

) means a term scaling like

2

in the limit →0, and

thus negligibly small compared with the O() term. The term σ is called the

inﬁnitesimal generator of the family of transformations.

From eqs. (4.20) and (4.21) we have

Q =

∂F

2

∂P

= q +

∂σ(q, P, t)

∂P

+O(

2

) ,

p =

∂F

2

∂q

= P +

∂σ(q, P, t)

∂q

+O(

2

) . (4.43)

Due to the nature of the generating function method these equations are

mixed between the perturbed and unperturbed canonical variables, but they

can easily be solved by iteration to give, up to ﬁrst order in ,

Q(q, p, t, ) = q +

∂σ(q, p, t)

∂p

+ O(

2

) ,

P(q, p, t, ) = p −

∂σ(q, p, t)

∂q

+ O(

2

) . (4.44)

Also, inserting eq. (4.42) in eq. (4.22),

K(Q, P, t, ) = H(q, p, t) +

∂σ(q, p, t)

∂t

+O(

2

) . (4.45)

This is again a mixed expression in the perturbed and unperturbed variables

so we use eq. (4.44) to get an explicit expression

K(q, p, t, ) = H(q, p, t) +

∂σ(q, p, t)

∂q

·

∂H

∂p

−

∂σ(q, p, t)

∂p

·

∂H

∂q

+

∂σ(q, p, t)

∂t

+ O(

2

) . (4.46)

64 CHAPTER 4. CANONICAL TRANSFORMATIONS

4.3.1 Time evolution

Normally is a parameter that we imagine varies at ﬁxed time, but it is also

of great interest to consider the case where the parameter is time itself. That

is, we consider transformations at times t + δt where t is ﬁxed but the time

variation comes from taking = δt. Now consider the choice σ(q, p, t) =

H(q, p, t). Then we see by comparing eq. (4.44) with Hamilton’s equations

of motion, eq. (3.12), that

Q(q, p, t, δt) = q(t) + ˙ q(t)δt +O((δt)

2

) = q(t +δt) +O((δt)

2

) ,

P(q, p, t, δt) = p(t) + ˙ p(t)δt +O((δt)

2

) = p(t +δt) + O((δt)

2

) . (4.47)

That is, we have shown that the Hamiltonian is the inﬁnitesimal generator

for dynamical evolution in time.

Since the result is true at all times t, we can integrate it to obtain a

statement valid over ﬁnite time intervals. First we introduce some dynami-

cal systems jargon by deﬁning the Hamiltonian ﬂow, which is the function

ϕ

t,s

(q

s

, p

s

) evolving any phase space point (q

s

, p

s

) at time s to the corre-

sponding point on the physical path at time t: ϕ

t,s

(q

s

, p

s

) = (q(t), p(t)),

with ˙ q = ∂H/∂p and ˙ p = −∂H/∂q. Then we can state the integrated result

in the following form: Hamiltonian ﬂows are canonical transformations.

A commonly encountered example of such a ﬂow is what we shall call a

stroboscopic map, ϕ

t

n+1

,t

n

(q

n

, p

n

), which evolves a phase-space point at time

t

n

= n∆t to the corresponding point at time t

n+1

= (n +1)∆t as if the state

of the system were captured periodically in the ﬂashes of a stroboscope.

The main use of of such a stroboscopic map is the analysis of the dynamics

of a nonautonomous, periodically forced system, in which one takes ∆t to

be the period of the forcing. In this case the map is called a period map. If

we visualize the motion in the (2n +1)-dimensional space formed by adding

the time dimension to phase space, and evaluate t modulo ∆t (i.e. make the

space topologically toroidal), we may think of the period map as giving the

intersections of orbits with the Poincar´e surface of section, t = 0. (This is

most useful when n = 1, for then the surface of section is two-dimensional

and easy to visualize.) For this reason systems where the Hamiltonian has

an explicit periodicity in time are sometimes called (2n + 1)-dimensional

Hamiltonian systems.

Another example is the construction of a numerical integrator of the

ODEs making up Hamilton’s equations of motion. Of necessity we replace

the inﬁnite number of points in time of the exact problem with a ﬁnite num-

ber of points in a numerical approximation. Although we do not know the

ﬂow exactly (otherwise there would be no point in integrating the equations

4.4. POISSON BRACKETS 65

numerically) it is sometimes very important to require the approximate stro-

boscopic map to be exactly canonical. (The full signiﬁcance of the fact that

a map is canonical will become apparent when we have established some

more properties of canonical transformations.) In this case the numerical

algorithm is called a symplectic integrator. An example is the discrete-time

dynamical system derived in Problem 4.7.2(c).

4.4 Poisson brackets

We deﬁne the canonical Poisson bracket ¦f, g¦ between any two phase-space

functions f and g by

1

¦f, g¦ ≡

∂f

∂q

·

∂g

∂p

−

∂f

∂p

·

∂g

∂q

. (4.48)

Note that the bracket is an antisymmetric bilinear form in f and g:

¦f, g¦ = −¦g, f¦, so that ¦f, f¦ ≡ 0. We observe that this algebraic prop-

erty is analogous to the properties of the commutator of two operators, A

and B say: [A, B] ≡ AB − BA. Indeed there is a deep physical connection

between the commutator and the Poisson bracket through the quantum-

mechanical correspondence principle: if a

op

and b

op

are quantum-mechanical

observables, then the classical limit may be obtained by replacing [a

op

, b

op

]

with i¦a, b¦, where a and b are the classical counterparts of the respective

quantum-mechanical operators.

A property shared with the commutator is the Jacobi identity

¦u, ¦v, w¦¦ +¦v, ¦w, u¦¦ +¦w, ¦u, v¦¦ ≡ 0 . (4.49)

The utility of the Poisson bracket notation lies in the fact that Hamilton’s

equations of motion, eq. (3.12), can be written in the symmetric fashion

˙ q = ¦q, H¦

˙ p = ¦p, H¦ , (4.50)

which allows the more compact notation

˙ z = ¦z, H¦ , (4.51)

where z ≡ (q, p) is the 2n-dimensional phase-space position deﬁned by con-

catenating q and p.

1

Do not confuse these curly brackets with the similar notation we have been using for

sets of variables. We follow the sign convention used by Goldstein (see Sec. 6) whereas

Scheck deﬁnes the Poisson bracket with the opposite sign.

66 CHAPTER 4. CANONICAL TRANSFORMATIONS

We can generalize this result to the time derivative of any function,

f(z, t),

df

dt

≡ ˙ z·

∂f

∂z

+

∂f

∂t

= ¦f, H¦ +

∂f

∂t

. (4.52)

For example, if we take f = H we immediately get

dH

dt

=

∂H

∂t

, (4.53)

which agrees with eq. (3.30). Note that, in an autonomous system, f is an

integral of the motion if it does not depend explicitly on time and its Poisson

bracket with the Hamiltonian vanishes.

We can also write eqs. (4.44) and (4.46) in the more compact forms

Z(z, t, ) = z +¦z, σ(z, t)¦ + O(

2

) , (4.54)

where Z ≡ (Q, P), and

K(z, t, ) = H(z, t) +

¦σ(z, t), H(z, t)¦ +

∂σ(z, t)

∂t

+O(

2

) . (4.55)

4.4.1 Symmetries and integrals of motion

In this section we consider only autonomous systems. From eq. (4.55) the

Hamiltonian is invariant under the canonical transformations generated by

σ if and only if its Poisson bracket with σ(z) vanishes. But eq. (4.52) shows

that this implies that σ is an integral of the motion. That is, the inﬁnitesimal

generator of a symmetry of the Hamiltonian is an integral of the motion.

For example, take σ = p

φ

, the momentum canonically conjugate to q = φ,

the angle measured about the z-axis. Then eq. (4.44) implies that φ is

mapped to φ + . Thus we see that p

φ

is the inﬁnitesimal generator for ro-

tations about the z-axis. If the system is rotationally symmetric then the

Hamiltonian is invariant under increments in φ. Then the above result shows

that p

φ

is an integral of the motion. This is a Hamiltonian form of Noether’s

theorem. (Of course the result is obvious if we use φ as a generalized coordi-

nate, since the absence of φ from the Hamiltonian immediately gives ˙ p

φ

= 0,

but we now do not need to use such coordinates—all we need to do is evaluate

the Poisson bracket of H with p

φ

.)

4.4.2 Perturbation theory

Often we are given a system for which the Hamiltonian may be resolved

into the form H

0

+ H

1

, where the case = 0 is exactly solvable and H

1

4.5. ACTION-ANGLE VARIABLES 67

represents the eﬀect of a small perturbation. (An example is the planets

moving around the sun—the motion of a single planet was exactly solved by

Newton, and shown to give rise to Kepler’s Laws. Thus we can take H

0

to be

the Hamiltonian for the planets with only the gravitational ﬁeld of the sun

included, and put the small gravitational interactions between the planets

into H

1

. This is known as a problem in celestial mechanics.)

One strategy for solving the perturbed problem is to successively trans-

form the new Hamiltonian into an exactly integrable one order by order in a

power series in . This is known as canonical perturbation theory.

The classical method of generating the canonical transformations is called

Von Zeipel’s method, and this consists of extending eq. (4.42) to higher order

in . The problem with this method is the awkward mixture of old and new

variables that has to be unscrambled. A more modern approach, amenable

to computerized algebraic manipulation, is known as the Lie transform. This

is an alternative to the traditional generating function method that can be

motivated by noting that time evolution generates a canonical transformation

continously connected to the identity. If we regard as pseudo-time, then we

can use a pseudo-Hamiltonian, W say, to generate a canonical transformation

that is an explicit function of only the old variables.

There are various variants of the Lie transform method, but they all

make use of the idea that we can represent the Poisson bracket with an

inﬁnitesimal generator W as an operator L

W

such that its action on any

phase-space function f is given by

L

W

f ≡ ¦f, W¦ . (4.56)

In the case that W does not depend on , the canonical transformation is

eﬀected by exponentiating L

W

exp(L

W

) = 1 +L

W

+

1

2

2

L

2

W

+. . . . (4.57)

4.5 Action-Angle Variables

It is clear from the equations of motion that if the Hamiltonian does not

depend on one of the generalized coordinates q

i

then the canonical momentum

conjugate to it, p

i

, is conserved. (We can as before refer to q

i

as ignorable,

though the terminology cyclic coordinate is also used in this context.)

This suggests a general strategy for solving Hamiltonian dynamics problems—

make a canonical transformation to new canonical coordinates Q, P such

that the new Hamiltonian is cyclic in all the new generalized coordinates

68 CHAPTER 4. CANONICAL TRANSFORMATIONS

K = K(P) (we consider the autonomous case only here). Then the dynam-

ics is very simple since all the new momenta are constants of the motion

and thus the generalized velocities ∂K/∂P

i

are also constant, so all the new

generalized coordinates evolve linearly in time

Q

i

= Q

i0

+

∂K

∂P

i

t . (4.58)

A system for which this strategy can, in principle at least, be carried out is

called integrable. Clearly, since there are n of the P

i

, a necessary condition for

integrability is that there must exist n independent integrals of the motion

that are well behaved as functions of the old canonical coordinates for all

time, including the limits t → ±∞. (For an autonomous system in fact the

integrals of motion have no explicit dependence on t at all.)

If the problem is such that the orbits remain in a bounded region of phase

space, then the coordinates have to oscillate in a quasiperiodic fashion, rather

than increasing secularly. In this, rather general, case the orbits are conﬁned

to topological tori labelled by the new momenta, and the new coordinates Q

i

can be taken to be angles deﬁned on the tori. The dimensions of the P

i

are

then those of action, and thus such a set of canonical coordinates is called

action-angle variables. Since the Hamiltonian ﬂow maps the P

i

= const tori

onto themselves, they are invariant sets of the dynamical system (just as

a ﬁxed point is an invariant set consisting of just one point) and are often

called invariant tori.

We have already come across a transformation to action-angle coordi-

nates, eq. (4.29), for the harmonic oscillator problem (in the case when

ω

0

= const). Other well known integrable systems are the particle moving

under the inﬂuence of an inverse-square law force, and the physical pen-

dulum, but time does not permit the development of explicit action-angle

transformations for these systems.

Although much of the classical mechanics literature is concerned with in-

tegrable problems it must be emphasized that integrability is not generic—

given an arbitrarily chosen Hamiltonian, from the set of all possible ones, it

will almost always not be integrable. Nevertheless we are very often inter-

ested in cases which are not far integrability, in the sense that the Hamil-

tonian can be decomposed into the form H

0

+ H

1

required to apply per-

turbation theory (see Sec. 4.4.2), where H

0

is integrable. Thus the study

of integrable systems remains an important one. Of particular interest are

inﬁnite-dimensional integrable Hamiltonian systems, such as the Korteweg–

de Vries equation and nonlinear Schr¨ odinger equation that give rise to soliton

phenomena.

4.6. PROPERTIES OF CANONICAL TRANSFORMATIONS 69

If the problem is not integrable, perturbation theory will in general give

results useful over only a ﬁnite time interval as the long-time behaviour may

be chaotic and impossible to describe in detail. However by choosing the

region of phase space carefully one may be able to construct an invariant

torus by perturbation theory, even in a nonintegrable system, as shown by

Kolmogorov, Arnol’d and Moser (the KAM theorem).

The essential idea in this theory is to ﬁnd an invariant torus of the un-

perturbed system which is as far from resonance as possible and to keep

changing parameters in such as way as is increased away from zero that the

torus remains nonresonant.

Although we shall not go into this further, some idea of the meaning of

resonance can be had by considering the pendulum problem. For H 2mgl

the potential energy term is small compared with the kinetic energy and we

can use H

0

= p

2

θ

/2m and take g as the expansion parameter. This will give

a perturbation theory involving inverse powers of p

θ

, which will obviously

blow up as [p

θ

[ → 0, i.e. as the rotation frequency approaches zero and we

approach the separatrix. This blowup is an example of a resonance between

the frequency of rotation and the frequency of the perturbation (in this case

zero frequency).

4.6 Properties of canonical transformations

4.6.1 Preservation of phase-space volume

One of the most important properties of canonical transformations is that

their Jacobian determinants det(∂Z

i

/∂z

j

) are always unity

2

. That is, a region

of phase space has the same volume whether represented in the old or new

canonical coordinates.

This is most easily seen (in the case of transformations continuously con-

nected with the identity at least) by showing that Hamiltonian ﬂows are

incompressible (see Fig. 4.2). To do this we calculate the divergence of

˙ z = (∂H/∂p, −∂H/∂q):

div ˙ z =

∂

∂q

·

∂H

∂p

−

∂

∂p

·

∂H

∂q

≡ 0 . (4.59)

Then the conservation of the phase-space volume follows from the application

of Gauss’ theorem since the surface integral of ˙ z is the rate of change of the

volume within the surface. This result obviously applies to any ﬂows built

up from inﬁnitesimal canonical transformations, not only to Hamiltonian

2

Unless we also include time reversal, x →x, p →−p, where the Jacobian is −1.

70 CHAPTER 4. CANONICAL TRANSFORMATIONS

q

1

p

1

q

2

...

p

n

z(t

1

)

z(t)

Figure 4.2: Incompressible motion of a volume in phase space.

time evolution. (In fact the conservation of phase-space volume holds for all

canonical transformations.)

This result has profound implications both in statistical mechanics and

in dynamical systems theory.

Liouville’s Theorem

In statistical mechanics we exploit the ﬂuid dynamical analogy implied by

the above “proof” by deﬁning a distribution function, f(q, p, t), which is the

phase-space density of a statistical ensemble of systems normalized so that

the probability of ﬁnding a system in a small region of phase space of volume

d

2n

z is f(z, t) d

2n

z. (Think of phase space as being ﬁlled with a dust cloud,

each particle of which represents a system of the ensemble.)

Then, since the number of systems in the ensemble is conserved, f obeys

the continuity equation

∂f

∂t

+ div( ˙ zf) = 0 , (4.60)

4.6. PROPERTIES OF CANONICAL TRANSFORMATIONS 71

or, using the incompressibility result eq. (4.59),

∂f

∂t

+ ˙ z·

∂f

∂z

= 0 . (4.61)

That is, the distribution function is constant on a physical path in phase

space. This is Liouville’s theorem of statistical mechanics.

Using ˙ z = (∂H/∂p, −∂H/∂q) we can write the result in terms of Poisson

brackets

∂f

∂t

+¦f, H¦ = 0 (4.62)

As a consequence we see that statistical equilibrium, ∂f/∂t = 0, is obtained

when ¦f, H¦ = 0, which can be achieved by taking f to be a function of

the integrals of the motion. A general closed N-body system in statistical

mechanics is nearly always assumed to be completely chaotic (the ergodic

hypothesis), so that the only integral of motion in a closed system is H itself.

In this case the equilibrium distribution function is a function of H only

[usually the Maxwellian distribution, proportional to exp(−H/kT) where k

is Boltzmann’s constant and T is the temperature, though in a “collisionless”

plasma it can be appropriate to consider other possibilities].

Hamiltonian mechanics as a dynamical system

The conservation of phase-space volume provides a strong constraint on the

long-time behaviour of a system in that it rules out the possibility of purely

attracting or repelling sets. For instance, if the basin of attraction of an

attracting ﬁxed point had the same dimensions as phase space we could

draw a surface entirely enclosing the ﬁxed point which would contract with

time, thus contradicting the conservation of phase-space volume. This does

not mean that ﬁxed points that are attracting in some directions cannot exist

in Hamiltonian systems, merely that for any attracting direction there must

be a repelling direction to balance it.

The simplest system in which to see some of the consequences of the

conservation of phase-space volume is provided by the period map of a peri-

odically forced one-degree-of-freedom system (sometimes called a 1

1

2

degree

of freedom Hamiltonian system). Then the phase-space is only two dimen-

sional and the “volume” is in fact an area. Thus the period map is area

preserving.

A standard example is the “kicked rotor”—a pendulum (see Fig. 2.2) in

which gravity is varied periodically (say by oscillating the support vertically

72 CHAPTER 4. CANONICAL TRANSFORMATIONS

and transforming to the accelerating frame in which the support is station-

ary). Furthermore we assume the gravity to be applied purely impulsively

g(t) = g

0

∆t

∞

¸

n=−∞

δ(t −t

n

) (4.63)

where t

n

≡ n∆t, with n an integer, and δ(t) is the Dirac δ function, so that g

0

is the time average of g(t). The form of the equations of motion is unaﬀected

by the fact that g depends on time, so the Hamiltonian is the same as in

eq. (3.20)

H(θ, p) =

p

2

2ml

2

+mg(t)l(1 −cos θ) . (4.64)

Between the kicks the motion is

t

ε

t

n−1

t

n

t

n+1

p

n

θ

n

g(t)

∆t

p(t)

θ(t)

Figure 4.3: Trial function for

the kicked rotor.

one of uniform rotation with con-

stant angular velocity p/ml. Thus

we can give a potentially exact rep-

resentation of the phase-space path

with θ a continuous, piecewise lin-

ear function of t and p a piecewise

constant step function with only the

values θ

n

at, and p

n

immediately

before, t = t

n

as yet unknown. As

illustrated in Fig. 4.3,

θ(t) =

1

∆t

[(t

n+1

−t)θ

n

+ (t −t

n

)θ

n+1

] ,

p(t) = p

n+1

(4.65)

for t in each range t

n

≡ n∆t < t < t

n+1

≡ (n + 1)∆t. Inserting this

trial function in the phase-space action integral eq. (3.32) over the interval

t

−N

+ε ≤ t ≤ t

N

+ε, N ≥ 1, ε →0+, we get

S

ph

=

N−1

¸

n

=−N

¸

p

n

+1

(θ

n

+1

−θ

n

) −

p

2

n

+1

∆t

2ml

2

−mg

0

l∆t (1 −cos θ

n

+1

)

.

(4.66)

The conditions for S

ph

to be stationary under variations of θ

n

and p

n+1

,

−N < n < N, (noting that θ

n

occurs in the terms of the sum for both n

= n

and n

= n −1) are

∂S

ph

∂p

n+1

= θ

n+1

−θ

n

−

p

n+1

∆t

ml

2

= 0 , (4.67)

4.6. PROPERTIES OF CANONICAL TRANSFORMATIONS 73

and

∂S

ph

∂θ

n

= p

n

−p

n+1

−mg

0

l∆t sin θ

n

= 0 . (4.68)

These equations deﬁne an area-preserving map known as the Standard

Map. It is clearly seen that they approximate the Hamiltonian equations

of motion for the physical pendulum as ∆t → 0. Commonly the map is

re-expressed in terms of an angle coordinate x with period normalized to

unity and a nondimensionalized momentum y, which is p expressed in units

of 2πml

2

/∆t,

θ = 2πx ,

p =

2πml

2

∆t

y . (4.69)

Then the map becomes

y

n+1

= y

n

−

k

2π

sin 2πx

n

,

x

n+1

= x

n

+y

n+1

, (4.70)

where the chaos parameter k is deﬁned by

k ≡

g

0

(∆t)

2

l

. (4.71)

Figure 4.4 shows some typical iterated orbits of the map, showing both

chaotic and regular regions. The remnants of the librating pendulum orbits

around the O points in Fig. 3.1 are seen as large “islands” around (x, y) =

(0, 0) and (0, 1) [also (1, 0) and (1, 1) because the phase space is periodic in

y as well as x], but the region corresponding to the X point in Fig. 3.1 is

highly chaotic. There are also new islands due to resonances not present in

the physical pendulum problem (in fact an inﬁnite number of them), each

with chaotic separatrices.

The value of k used in Fig. 4.4 is k

c

= 0.971635406 . . ., which is the value

at which the last KAM invariant curves with the topology of a rotating pen-

dulum orbit become unstable and break up into invariant Cantor sets. Two

of these KAM curves are shown, being the only curves completely crossing

the ﬁgure from left boundary to right. For k < k

c

the motion is bounded in

y, while for k greater than this value a phase-space point can diﬀuse without

bound in the positive or negative y-direction.

From eq. (4.71) we see that the integrable limit ∆t → 0 corresponds to

k →0 [noting that eq. (4.69) shows that the width of the physical pendulum’s

separatrix shrinks to zero as k →0 when represented in terms of y].

74 CHAPTER 4. CANONICAL TRANSFORMATIONS

1

0

0 1

y

x

Figure 4.4: Some representative orbit segments for the Standard Map

showing both chaotic and regular motion.

4.6.2 Transformation of Poisson brackets

Finally, we note the important result the Poisson bracket is invariant under

canonical transformation

¦f, g¦ =

∂f

∂q

·

∂g

∂p

−

∂f

∂p

·

∂g

∂q

=

∂f

∂Q

·

∂g

∂P

−

∂f

∂P

·

∂g

∂Q

. (4.72)

4.7 Problems

4.7.1 Coriolis yet again

Given the Hamiltonian for a particle on a sphere in nonrotating latitude and

longitude coordinates, θ and φ

, respectively,

K =

p

2

θ

2mR

2

+

p

2

φ

2mR

2

cos

2

θ

+V

(θ, φ

, t) ,

4.7. PROBLEMS 75

ﬁnd the transformation to the θ, p

θ

, φ, p

φ

canonical coordinates, where φ is

the usual latitude measured in the rotating frame deﬁned in problem 2.7.1.

That is, φ

**= Ωt + φ. Find the Hamiltonian H in the rotating coordinates.
**

4.7.2 Diﬀerence approximations

Using the harmonic oscillator Lagrangian L =

1

2

m( ˙ x

2

−ω

2

0

x

2

)

(a) calculate an approximate action integral S using the piecewise-linear trial

function

x(t) =

1

∆t

[(t

n+1

−t)x

n

+ (t −t

n

)x

n+1

]

for t in each range t

n

≡ n∆t < t < t

n+1

≡ (n + 1)∆t.

(b) Show that the approximate action integral from t

−N

to t

N

(N ≥ 1 being

an arbitrary integer) is stationary for −N < n < N if x

n

obeys the second-

order diﬀerence equation

x

n−1

−2x

n

+x

n+1

= −

ω

2

0

(∆t)

2

6

(x

n−1

+ 4x

n

+x

n+1

)

(c) Using the approximate action integral

¯

S(x

n

, x

n+1

) evaluated over the

range t

n

< t < t

n+1

, as a type 1 generating function, F

1

(q, Q) = −

¯

S(x

n

, x

n+1

),

with q = x

n

and Q = x

n+1

, ﬁnd the linear canonical transformation from

(x

n

, p

n

≡ p) to (x

n+1

, p

n+1

≡ P). Represent this linear transformation as a

matrix and show that the determinant is unity. (It would be advisable to

use Maple or Mathematica for the algebra.)

This provides a discrete-time dynamical system approximation to the true

continuous-time system for the harmonic oscillator.

(d) Iterate the map obtained in (c) 100 times taking the initial point as

x

0

= 1, p

0

= 0 and plot the result. Use units such that m = ω

0

= 1 and use

three timesteps: ∆t = 0.1, ∆t = 3.45 and ∆t = 3.47.

For a bonus mark verify one of the solutions by showing that the second-

order diﬀerence equation derived in (b) above (when arranged to give x

n+1

in

terms of x

n

and x

n−1

) gives the same sequence of x-values as the discrete-time

dynamical system in (c) if two successive values of x given by the dynam-

ical system are used to start the second-order diﬀerence equation. I.e. the

discrete time “Lagrangian” and “Hamiltonian” discriptions are dynamically

equivalent. For another bonus mark calculate the eigenvalues for the three

values of ∆t given above. (Again, use Maple or Mathematica.)

76 CHAPTER 4. CANONICAL TRANSFORMATIONS

Chapter 5

Answers to Problems

5.1 Chapter 1 Problems

Problem 1.6.1

Two particles are connected by a rigid rod so they are constrained to move a

ﬁxed distance apart. Write down a constraint equation of the form eq. (1.2)

and ﬁnd suitable generalized coordinates for the system incorporating this

holonomic constraint.

Answer

Let the position of particle 1 with respect to a stationary Cartesian frame

be ¦x

1

, y

1

, z

1

¦ and that of particle 2 be ¦x

2

, y

2

, z

2

¦. The rigid rod constraint

equation is then

(x

1

−x

2

)

2

+ (y

1

−y

2

)

2

+ (z

1

−z

2

)

2

= l

2

. (5.1)

This reduces the number of degrees of freedom from 6 to 5, which we could

take to be position of particle 1, ¦x

1

, y

1

, z

1

¦, and the spherical polar angles

θ and φ depicted in Fig. 1.1 to specify the orientation of the rod. However,

a better choice for the three position coordinates would be the coordinates

of the centre of mass [see Problem 3.5.1(a)], whose position vector is

r

CM

=

m

2

r

1

+m

1

r

2

m

1

+m

2

. (5.2)

Problem 1.6.2

Suppose we know that the angular momentum vectors r

k

×m

k

˙ r

k

of a system

of particles are all nonzero and parallel to the z-axis in a particular Cartesian

77

78 CHAPTER 5. ANSWERS TO PROBLEMS

coordinate system. Write down the diﬀerential constraints implied by this

fact, and show that they lead to a set of holonomic constraints. Hence write

down suitable generalized coordinates for the system.

Answer

The diﬀerential constraints are provided by the vanishing of the x and y

components of the angular momentum vectors (multiplied by dt)

y

k

dz

k

−z

k

dy

k

= 0

z

k

dx

k

−x

k

dz

k

= 0 . (5.3)

for k = 1, 2, . . . , N.

These are satisﬁed by taking

z

k

≡ 0 , (5.4)

which form a set of holonomic constraints. Thus the motions are constrained

to lie in the x, y plane and we can either take q = ¦x

1

, y

1

, x

2

, y

2

, . . . , x

N

, y

N

¦

or transform to polar coordinates using x

k

= r

k

cos θ

k

, y

k

= r

k

sin θ

k

, so that

q = ¦r

1

, θ

1

, r

2

, θ

2

, . . . , r

N

, θ

N

¦.

Remark 5.1 This problem has been called “Ecliptic” because this is the as-

tronomical name of the plane to which the planets are approximately “con-

strained”. It suggests that the reason for this constraint is because the planets

were all created by condensation from the same spinning disk of dust and gas

and thus have parallel angular momenta.

Note: Although eq. (5.4) clearly solves eq. (5.3) we might ask whether this

is the only solution. Thus, suppose that z

k

is not equal to zero. We can

assume that, except for isolated instants, x

k

and y

k

are nonzero also. Thus

we can divide the ﬁrst eq. (5.3) by z

k

x

k

, the second by y

k

z

k

, and add to

show (since the terms dz

k

/z

k

cancel) that dx

k

/x

k

− dy

k

/y

k

= 0. However

we are told that the z components of the angular momenta are nonzero, so

that x

k

dy

k

− y

k

dx

k

= 0. Dividing by x

k

y

k

we clearly have a contradiction,

so the assumption that z

k

= 0 must be wrong. 2

Problem 1.6.3

Show that any geodesic r = x(τ) on a two-dimensional manifold S : r =

X(θ, ζ) embedded in ordinary Euclidean 3-space, where θ and ζ are arbi-

trary curvilinear coordinates on S, is such that the curvature vector κ(τ) is

everywhere normal to S (or zero).

5.1. CHAPTER 1 PROBLEMS 79

n

S

∂

ζ

X(θ,ζ)

∂

θ

X(θ,ζ)

ζ = const

θ = const

κ

e

| |

Figure 5.1: Surface S with curvilinear coordinate system, a geodesic,

the three tangent-plane vectors e

, ∂

θ

X and ∂

ζ

X, and the normal

vectors n and κ.

The curvature vector is deﬁned by κ ≡ de

/ dl, where e

(τ) ≡ dx/ dl is

the unit tangent vector at each point along the path r = x(τ).

Answer

The vector length element is dl = dθX

θ

+ dζX

ζ

, where X

θ

≡ ∂

θ

X ≡ ∂X/∂θ

and X

ζ

≡ ∂

ζ

X ≡ ∂X/∂ζ. Thus l =

f(θ, ζ,

˙

θ,

˙

ζ) dτ with

f ≡

dl

dτ

=

X

2

θ

˙

θ

2

+ 2X

θ

·X

ζ

˙

θ

˙

ζ +X

2

ζ

˙

ζ

2

1/2

(5.5)

and

e

≡

1

f

dl

dτ

=

˙

θX

θ

+

˙

ζX

ζ

f

. (5.6)

From eq. (5.5), we have, for any path

∂

˙

θ

f =

1

f

X

2

θ

˙

θ +X

θ

·X

ζ

˙

ζ

=

1

f

X

θ

˙

θ +X

ζ

˙

ζ

·X

θ

≡ e

·X

θ

(5.7)

Similarly, by interchanging θ and ζ we have

∂

˙

ζ

f = e

·X

θ

. (5.8)

Also,

∂

θ

f =

1

2f

˙

θ

2

∂

θ

X

2

θ

+ 2

˙

θ

˙

ζ∂

θ

(X

θ

·X

ζ

) +

˙

ζ

2

∂

θ

X

2

ζ

**80 CHAPTER 5. ANSWERS TO PROBLEMS
**

=

1

f

˙

θ

2

X

θ

·X

θθ

+

˙

θ

˙

ζX

θθ

·X

ζ

+

˙

θ

˙

ζX

θ

·X

θζ

+

˙

ζ

2

X

ζ

·X

θζ

=

1

f

˙

θX

θ

+

˙

ζX

ζ

·

˙

θX

θθ

+

˙

ζX

θζ

≡ e

·

dX

θ

dτ

. (5.9)

Similarly, interchanging θ and ζ, we have

∂

ζ

f ≡ e

·

dX

ζ

dτ

. (5.10)

Now consider the Euler–Lagrange equation for a geodesic

d

dτ

∂

˙

θ

f −∂

θ

f = 0 . (5.11)

Using the identities eq. (5.7) and eq. (5.9) we can write this as

d

dτ

e

·X

θ

−e

·

dX

θ

dτ

= 0 . (5.12)

Or, expanding out and cancelling,

X

θ

·

d

dτ

e

= 0 . (5.13)

Similarly, the Euler–Lagrange equation for ζ variations becomes

X

ζ

·

d

dτ

e

= 0 . (5.14)

Recognizing that κ ≡ ( d/f dτ)e

**we can write eqs. (5.13) and (5.14) as
**

X

ζ

·κ = X

θ

·κ = 0 . (5.15)

Since the vectors X

θ

and X

ζ

span the tangent plane at the point r = X(θ, ζ)

(see Fig. 5.1), we have shown that the curvature vector κ has no component

in the tangent plane. Thus either κ = 0 or it is parallel to the unit normal,

n. 2

5.2 Chapter 2 Problems

Problem 2.8.1(a)

As a model of the motion of a ﬂuid element or dust particle in a planetary

(e.g. Earth’s) atmosphere, consider the motion of particle of unit mass con-

strained to move on the surface of a perfectly smooth sphere of radius R

5.2. CHAPTER 2 PROBLEMS 81

rotating with angular velocity Ω about the z-axis. Suppose the force on the

particle is given by an eﬀective potential V (θ, φ) (see Sec. 2.4.6), where θ and

φ are the latitude and longitude respectively.

(a) Write down the Lagrangian in a frame rotating with the planet, taking the

generalized coordinates to be the latitude and longitude so that z = Rsin θ,

x = Rcos θ cos(φ+Ωt), y = Rcos θ sin(φ+Ωt), where x, y, z is a non-rotating

Cartesian frame.

Write down the equations of motion, and ﬁnd a ﬁrst integral (i.e. constant

of the motion) in the case where V is independent of longitude. Among

this class of potentials ﬁnd the special case V = V

0

(θ) required to make

equilibrium possible (i.e. so that the equations of motion admit the solution

˙

θ =

˙

φ = 0 at each latitude).

Answer

˙ x = −R

˙

θ sin θ cos(φ + Ωt) −R(Ω +

˙

φ) cos θ sin(φ + Ωt)

˙ y = −R

˙

θ sin θ sin(φ + Ωt) +R(Ω +

˙

φ) cos θ cos(φ + Ωt)

˙ z = R

˙

θ cos θ (5.16)

Thus the kinetic energy is given by

T =

1

2

˙ x

2

+ ˙ y

2

+ ˙ z

2

=

1

2

R

2

˙

θ

2

sin

2

θ

cos

2

(φ + Ωt) + sin

2

(φ + Ωt)

+ (Ω +

˙

φ)

2

cos

2

θ

sin

2

(φ + Ωt) + cos

2

(φ + Ωt)

+

˙

θ

2

cos

2

θ

=

1

2

R

2

˙

θ

2

+ (Ω +

˙

φ)

2

cos

2

θ

, (5.17)

and the Lagrangian, L = T −V , by

L =

1

2

R

2

˙

θ

2

+ (Ω +

˙

φ)

2

cos

2

θ

−V (θ, φ) . (5.18)

Thus

∂L

∂

˙

θ

= R

2

˙

θ (5.19)

and

∂L

∂

˙

φ

= R

2

(Ω +

˙

φ) cos

2

θ (5.20)

82 CHAPTER 5. ANSWERS TO PROBLEMS

and thus the Lagrange equations of motion ( d/ dt)∂L/∂

˙

θ = ∂L/∂θ and

( d/ dt)∂L/∂

˙

φ = ∂L/∂φ become

R

2

¨

θ = −R

2

(Ω +

˙

φ)

2

sin θ cos θ −

∂V

∂θ

(5.21)

and

R

2

¨

φcos

2

θ −2R

2

˙

θ(Ω +

˙

φ) sin θ cos θ = −

∂V

∂φ

. (5.22)

In the case where V = V

0

(θ), φ is an ignorable coordinate and thus ∂L/∂

˙

φ,

given by eq. (5.20), is an integral of the motion.

If in this latter case we further demand that V

0

be such that

˙

θ ≡

˙

φ = 0

for all θ and φ, and use the identity sin 2θ ≡ 2 sin θ cos θ, then eq. (5.21) gives

∂V

0

∂θ

= −

1

2

R

2

Ω

2

sin 2θ , (5.23)

or, integrating,

V

0

(θ) =

1

4

R

2

Ω

2

cos 2θ + const =

1

2

R

2

Ω

2

cos

2

θ , (5.24)

where the second equality follows from the identity cos 2θ ≡ 2 cos

2

θ −1 and

the choice const =

1

4

R

2

.

Problem 2.8.1(b)

Now assume that the general form of the potential is V (θ, φ) = V

0

(θ) +

h(θ, φ), where h is the enthalpy (proportional to p

(γ−1)/γ

) deﬁned in Sec. 2.4.6.

Assuming the winds are slow, so that second order time derivatives of θ and

φ (and products of ﬁrst order time derivatives) can be neglected, show that

the velocity of the particle in the rotating frame is at right angles to the

pressure gradient −∇p.

Draw a sketch of a typical weather-map “low” (i.e. a localized depression

in p) showing the pressure contours, the direction of the force on a ﬂuid ele-

ment and the direction of motion. Hence show that motion in a depression

or low is cyclonic, where “cyclonic” in geophysical ﬂuid dynamics means “in

the direction of the planet’s rotation” (i.e. clockwise in the Southern Hemi-

sphere, counter-clockwise in the Northern Hemisphere for Ω > 0).

5.2. CHAPTER 2 PROBLEMS 83

Answer

Making the slow approximation, and taking into account eq. (5.23) the equa-

tions of motion, eqs. (5.21) and (5.22), become

R

2

Ω

˙

φsin 2θ ≈ −

∂h

∂θ

(5.25)

and

R

2

Ω

˙

θ sin 2θ ≈

∂h

∂φ

. (5.26)

Thus the pressure gradient causes a wind proportional to spatial derivatives

of h, and the slow approximation will be best satisﬁed when h is small.

These two equations give the wind speed in angular velocities. We now

calculate the time rate of change of the enthalpy along a stream line

˙

h ≡ w·∇h

≡

˙

θ

∂h

∂θ

+

˙

φ

∂h

∂φ

≈

1

R

2

Ωsin 2θ

∂h

∂φ

∂h

∂θ

−

∂h

∂θ

∂h

∂φ

≡ 0 (5.27)

from eqs. (5.25) and (5.26). Thus the eﬀect of the Coriolis forces is to make

the air move along the pressure contours, even though the forces acting on a

ﬂuid element act at right angles to the contours. It is for this reason that the

vortices caused by highs and lows of pressure are so long lived—the air does

not rush in to equalize the pressure diﬀerences but, rather, rotates around

the maxima and minima of pressure.

In order to establish the direction of rotation it is suﬃcient to draw a

sketch in a Mercator’s projection (i.e. in the θ-φ plane) and use eqs. (5.25)

and use (5.26) to show the wind direction at a few representative points on

a typical pressure contour.

Note: It is easy to show from eq. (2.49) that p/ρ = constp

(γ−1)/γ

, and so,

since γ > 1, the enthalpy is an increasing function of the pressure.

Problem 2.8.1(c)

Consider a charged particle constrained to move on a non-rotating smooth

insulating sphere, immersed in a uniform magnetic ﬁeld B = Be

z

, on which

the electrostatic potential is a function of latitude and longitude. Write down

the Lagrangian in the same generalized coordinates as above and show it is

the same as that for the particle on the rotating planet with appropriate

identiﬁcations of Ω and V .

84 CHAPTER 5. ANSWERS TO PROBLEMS

Answer

The Lagrangian is L = T+e˙ r·A−eΦ, where T is given by eq. (5.17) with Ω set

to zero. We now need a vector potential A such that B

z

≡ ∂

x

A

y

−∂

y

A

x

= B.

A suitable choice is A

x

= −

1

2

By, A

y

=

1

2

Bx, A

z

= 0. Thus, from eq. (5.16),

e˙ r·A =

1

2

eB(x ˙ y −y ˙ x)

=

1

2

eBR

2

cos θ

cos φ(−

˙

θ sin θ sin φ +

˙

φcos θ cos φ)

+ sin φ(

˙

θ sin θ cos φ +

˙

φcos θ sin φ)

=

1

2

eBR

2

˙

φcos

2

θ . (5.28)

Thus the Lagrangian is

L =

1

2

mR

2

˙

θ

2

+

˙

φ

2

cos

2

θ

+

1

2

eBR

2

˙

φcos

2

θ −eΦ(θ, φ) . (5.29)

Completing the square we can rewrite the Lagrangian in the form

L =

1

2

mR

2

¸

˙

θ

2

+

˙

φ +

eB

2m

2

cos

2

θ

¸

−

e

2

B

2

R

2

8m

cos

2

θ −eΦ(θ, φ) . (5.30)

This is the same as eq. (5.18) with the identiﬁcations

Ω =

eB

2m

(5.31)

and

V (θ, φ) =

e

2

B

2

R

2

8m

cos

2

θ +eΦ(θ, φ) . (5.32)

Note that eq. (5.31) can also be written Ω =

1

2

ω

c

, where ω

c

≡ eB/m is

the cyclotron frequency.

Problem 2.8.2

Consider the following potential V , corresponding to a particle of mass m

oscillating along the x-axis under the inﬂuence of a nonideal spring (i.e. one

with a nonlinear restoring force),

V (x) =

mω

2

0

2

x

2

+ σ

x

4

l

2

0

,

where the constant ω

0

is the angular frequency of oscillations having ampli-

tude small compared with the characteristic length l

0

, and σ = ±1 depends

on whether the spring is “soft” (σ = −1) or “hard” (σ = +1).

5.2. CHAPTER 2 PROBLEMS 85

Consider the trial function

x = l

0

[Acos ωt +Bcos 3ωt +C sin 3ωt] ,

where A, B, C are the nondimensionalized amplitudes of the fundamental

and third harmonic, respectively, and ω is the nonlinearly shifted frequency.

By using this trial function in the time-averaged Hamilton’s Principle, ﬁnd

implicit relations giving approximate expressions for ω, B and C as functions

of A. Show that C ≡ 0. The trial function is strictly appropriate only to the

case A < 1, but plot ω/ω

0

and B vs. A from 0 to 1 in the case of both a

hard and a soft spring.

Answer

See the following Maple notebook.

86 CHAPTER 5. ANSWERS TO PROBLEMS

5.3 Chapter 3 Problems

Problem 3.5.1(a)

Write down a Lagrangian for the problem of two particles of mass m

1

and

m

2

connected by a light rigid rod of length l in a gravitational ﬁeld g. Take

the generalized coordinates of the system to be q = ¦x, y, z, θ, φ¦, with the

coordinates of the two particles being given by

x

1

= x −α

1

l sin θ cos φ

y

1

= y −α

1

l sin θ sin φ

z

1

= z −α

1

l cos θ

x

2

= x +α

2

l sin θ cos φ

y

2

= y + α

2

l sin θ sin φ

z

2

= z + α

2

l cos θ ,

where α

1

≡ m

2

/(m

1

+ m

2

) and α

2

≡ m

1

/(m

1

+ m

2

) (so that (x, y, z) is the

centre of mass).

Answer

˙ x

1

= ˙ x +α

1

l

˙

φsin θ sin φ −α

1

l

˙

θ cos θ cos φ

˙ x

2

= ˙ x −α

2

l

˙

φsin θ sin φ +α

2

l

˙

θ cos θ cos φ

(5.33)

Because m

1

α

1

= m

2

α

2

the cross terms cancel when we expand the x-contribution

to the kinetic energy,

1

2

m

1

˙ x

2

1

+

1

2

m

2

˙ x

2

2

=

m

1

+m

2

2

˙ x

2

+

m

1

m

2

l

2

2(m

1

+m

2

)

(

˙

φsin θ sin φ −

˙

θ cos θ cos φ)

2

.

(5.34)

Similarly

1

2

m

1

˙ y

2

1

+

1

2

m

2

˙ y

2

2

=

m

1

+m

2

2

˙ y

2

+

m

1

m

2

l

2

2(m

1

+m

2

)

(

˙

φsin θ cos φ +

˙

θ cos θ sin φ)

2

,

(5.35)

and

1

2

m

1

˙ z

2

1

+

1

2

m

2

˙ z

2

2

=

m

1

+m

2

2

˙ z

2

+

m

1

m

2

l

2

2(m

1

+ m

2

)

˙

θ

2

sin

2

θ . (5.36)

5.3. CHAPTER 3 PROBLEMS 87

Thus, adding, the kinetic energy is

T =

m

1

+m

2

2

( ˙ x

2

+ ˙ y

2

+ ˙ z

2

) +

m

1

m

2

l

2

2(m

1

+m

2

)

(

˙

θ

2

+ sin

2

θ

˙

φ

2

) . (5.37)

This is the kinetic energy of a free particle of mass (m

1

+m

2

) plus that of a

particle with the reduced mass µ ≡ m

1

m

2

/(m

1

+m

2

) constrained to move on

the surface of a sphere of radius l. [Cf. eq. (5.17) of the Coriolis problem with

Ω set to zero, and recalling that we are here using spherical polar coordinates

where the origin of θ is at the pole, in contrast to the latitude angle whose

origin was at the equator—hence sin θ and cos θ are interchanged.] This is a

consequence of the well-known result that the kinetic energy of a system of

particles is the kinetic energy of the centre of mass plus the kinetic energy

in the centre-of-mass frame.

The potential energy is

V = (m

1

z

1

+m

2

z

2

)g = (m

1

+m

2

)zg , (5.38)

and the Lagrangian is L = T −V . We see from eqs. (5.37) and (5.38) that, by

using the centre of mass as the origin for our polar coordinates, we have made

L ignorable in x, y and φ and that the centre-of-mass coordinates x, y, z form

a subsystem completely independent from that of the angular coordinates.

Problem 3.5.1(b)

Use this Lagrangian to construct the Hamiltonian for the system. Is it of the

form T +V ?

Answer

Answering the second part of the question ﬁrst, the Hamiltonian must be

of the form T + V because the constraint [r

2

− r

1

[ = l does not depend

explicitly on time and thus the system comes under the general case treated

in Sec. 3.2.6. From eq. (5.37) the mass matrix is

µ =

M 0 0 0 0

0 M 0 0 0

0 0 M 0 0

0 0 0 µl

2

0

0 0 0 0 µl

2

sin

2

θ

¸

¸

¸

¸

¸

¸

, (5.39)

88 CHAPTER 5. ANSWERS TO PROBLEMS

where M ≡ m

1

+ m

2

is the total mass and µ is the reduced mass deﬁned

earlier. This is trivially inverted to give, from Sec. 3.2.6, eq. (3.27),

H =

p

2

x

2M

+

p

2

y

2M

+

p

2

z

2M

+

p

2

θ

2µl

2

+

p

2

φ

2µl

2

sin

2

θ

+Mgz . (5.40)

Note: Because x, y and φ are cyclic (ignorable) coordinates, p

x

, p

y

and

p

φ

are constants of the motion. The motion of the rod can thus be found

by studying two independent, one-dimensional dynamical problems in z and

θ, with Hamiltonians p

2

z

+ Mgz and p

2

θ

/2µl

2

+ p

2

φ

/2µl

2

sin

2

θ respectively.

The z-dynamics is just uniformly accelerated motion, with solution z =

z

0

+(p

z,0

/m)t −

1

2

gt

2

, p

z

= p

z,0

−mgt. A qualitative understanding of the θ-

dynamics can be had by considering that the “pseudo-potential” p

2

φ

/2µl

2

sin

2

θ

forms a well about θ = π/2 (assuming p

φ

= 0) with sides becoming inﬁnitely

high at θ = 0 and π. Thus θ is an oscillatory function of time. (In fact we

know from elementary physics that the motion can be integrated in closed

form because the rod spins at right angles to a constant angular momentum

vector with arbitrary, ﬁxed direction.)

Problem 3.5.1(c)

Find the Hamiltonian corresponding to the Coriolis Lagrangian in Problem

2.8.1,

L =

1

2

R

2

˙

θ

2

+ cos

2

θ (Ω +

˙

φ)

2

−V (θ, φ) .

Is it of the form T +V ?

Answer

From the above,

p

θ

≡

∂L

∂

˙

θ

= R

2

˙

θ (5.41)

so

˙

θ =

p

θ

R

2

. (5.42)

Also

p

φ

≡

∂L

∂

˙

φ

= R

2

cos

2

θ (Ω +

˙

φ) . (5.43)

so

˙

φ =

p

φ

R

2

cos

2

θ

−Ω . (5.44)

5.3. CHAPTER 3 PROBLEMS 89

Thus

H ≡ p

θ

˙

θ + p

φ

˙

φ −L

=

p

2

θ

R

2

+

p

2

φ

R

2

cos

2

θ

−Ωp

φ

−

p

2

θ

2R

2

−

p

2

φ

2R

2

cos

2

θ

+V (θ, φ)

=

p

2

θ

2R

2

+

p

2

φ

2R

2

cos

2

θ

−Ωp

φ

+ V (θ, φ) . (5.45)

Given that

T =

1

2

R

2

˙

θ

2

+ cos

2

θ (Ω +

˙

φ)

2

=

p

2

θ

2R

2

+

p

2

φ

2R

2

cos

2

θ

(5.46)

it is clear that the Hamiltonian diﬀers from T +V by the term −Ωp

φ

.

Problem 3.5.2

Find the Hamiltonian corresponding to the

-1 -0.5 0.5 1

x

-0.1

-0.05

0.05

0.1

H

Figure 5.2: H(x, 0).

Lagrangian of the anharmonic oscillator in Prob-

lem 2.8.2. Sketch contours of H(x, p) in the

soft potential case for (a) H < mω

2

0

l

2

0

/8 and

(b) H > mω

2

0

l

2

0

/8. Indicate representative tra-

jectories of the system in phase-space. Try

to include representatives of each qualitatively

distinct orbit.

Answer

The Lagrangian is L =

1

2

m˙ x

2

+ V (x), which is in the standard form treated

in Sec. 3.2.3. Thus ˙ x = p/m and the Hamiltonian is

H =

p

2

2m

+

mω

2

0

2

x

2

−

x

4

l

2

0

. (5.47)

Before sketching the phase space we ﬁrst get a qualitative feel for its

salient features by looking at the behaviour of H as a function of x and p

and ﬁnding its stationary points, i.e. where ∂H/∂x = ∂H/∂p = 0. (From

Hamilton’s equations of motion we see that such points correspond to ˙ x =

˙ p = 0 and are thus equilibrium points, where the system does not move—ﬁxed

90 CHAPTER 5. ANSWERS TO PROBLEMS

points of the dynamical system. Points where the Hamiltonian is a minimum

are stable, whereas saddle points correspond to unstable equilibria.)

For any ﬁxed x the graph of H is a parabola, concave upwards, with its

minimum at p = 0. Thus H approaches +∞ as [p[ →∞.

We now examine the behaviour of H in the limits of large and small x.

For [x[ <l

0

the term quartic in x is very small compared with the quadratic

term so the graph of H is a parabola, concave upwards, with its minimum

at x = 0.

On the other hand, for [x[ l

0

the negative quartic term dominates over

the quadratic term and H approaches −∞as [x[ →∞. The behaviour of H

on the x-axis is graphed in Fig. 5.2, using units such that l

0

= ω

0

= m = 1.

It is seen that there are maxima of H at ﬁnite values of [x[. The maxima

and minima of H as a function of x are the roots of

∂H

∂x

= mω

2

0

x

1 −

2x

2

l

2

0

, (5.48)

which are at x = 0 (the minimum) and x = ±l

0

/

√

2 (the maxima).

Since the origin is the intersection of the

-1 -0.5 0 0.5 1

x

-0.6

-0.4

-0.2

0

0.2

0.4

0.6

p X O X

Figure 5.3: Phase

space of the an-

harmonic oscilla-

tor.

lines of minima, x = 0, p = 0, there is a stable

equilibrium point at (x, p) = (0, 0). There are

mimimax or saddle points at the intersections

of the line of minima p = 0 and the lines of

maxima x = ±l

0

/

√

2. Thus the points (x, p) =

(±l

0

/

√

2, 0) are also equilibrium points, but

are unstable towards small perturbations. At

both these saddle points, H = mω

2

0

l

2

0

/8, so

that they are linked by the contour H = const =

mω

2

0

l

2

0

/8.

Contours of H in units such that l

0

= ω

0

=

m = 1 are plotted in Fig. 5.3. As in Fig. 3.1

we have marked the stable and unstable ﬁxed

points by O and X, respectively, and have in-

dicated the contour H = mω

2

0

l

2

0

/8 by a dashed line. Directions of motion (to

the right in the upper half plane because ˙ x > 0 and to the left in the lower

half plane) are marked by arrows. It is seen that the contour H = mω

2

0

l

2

0

/8

forms the separatrix between ﬁve classes of qualitatively diﬀerent orbit:

• bounded oscillatory (librating) orbits contained within the separatrix,

so that H < mω

2

0

l

2

0

/8, corresponding to a particle trapped in the po-

tential well;

5.3. CHAPTER 3 PROBLEMS 91

• unbounded orbits with H < mω

2

0

l

2

0

/8 which cross the x-axis (p = 0) at a

value x < l

0

/

√

2 or x > l

0

/

√

2, corresponding to particles approaching

from the left or right, respectively, and being reﬂected by a potential

barrier;

• unbounded orbits with H > mω

2

0

l

2

0

/8, which cross the p-axis (x = 0) at

a value x < mω

0

l

0

/2 or x > mω

0

l

0

/2, corresponding to particles with

suﬃcient energy to cross the potential barriers (from right to left, or

left to right, respectively) without being reﬂected.

Problem 3.5.3

Consider the motion of a particle of charge e and mass m in a straight, in-

ﬁnitely long magnetic conﬁnement system with vector potential A = ψ(x, y)e

z

.

(a) Show that B

x

= ∂ψ/∂y, B

y

= −∂ψ/∂x, B

z

= 0 and thus that contours

of ψ(x, y) deﬁne magnetic ﬁeld lines.

Answer

Given A

x

= A

y

= 0, A

z

= ψ(x, y) we calculate B ≡ ∇×A:

B

x

= ∂

y

A

z

−∂

z

A

y

= ∂

y

ψ ,

B

y

= ∂

z

A

x

−∂

x

A

z

= −∂

x

ψ . (5.49)

(Alternatively, we can observe that B = ∇ψ×e

z

and take the x and y com-

ponents.) From eq. (5.49) we have B·∇ψ = ∂

y

ψ ∂

x

ψ − ∂

x

ψ ∂

y

ψ ≡ 0. Thus

ψ = const on a magnetic ﬁeld line and so plotting contours of ψ maps out

the ﬁeld.

(b) Show that the Hamiltonian is

H =

p

2

x

2m

+

p

2

y

2m

+

[p

z

−eψ(x, y)]

2

2m

,

write down the Hamiltonian equations of motion and give two integrals of

the motion.

Answer

The given expression for H follows directly by writing eq. (3.2.5) in Carte-

sian components and using A

x

= A

y

= 0, A

z

= ψ(x, y). The Hamiltonian

92 CHAPTER 5. ANSWERS TO PROBLEMS

equations of motion are

˙ x =

∂H

∂p

x

=

p

x

m

,

˙ y =

∂H

∂p

y

=

p

y

m

,

˙ z =

∂H

∂p

z

=

(p

z

−eψ)

m

(5.50)

and

˙ p

x

= −

∂H

∂x

=

e(p

z

−eψ)

m

∂ψ

∂x

,

˙ p

y

= −

∂H

∂y

=

e(p

z

−eψ)

m

∂ψ

∂y

,

˙ p

z

= −

∂H

∂z

= 0 . (5.51)

From the last equation above, p

z

is clearly an integral of the motion. Also,

since H does not depend explicitly on time, H itself constitutes a second

integral of motion as shown in Sec. 3.3.

(c) It can be shown

1

that in any region with no electric current, ψ(x, y) can

be represented as the real part of an analytic function of the complex variable

ζ ≡ x + iy. Thus, write ψ(x, y) = Re Ψ(ζ) and show that the equations of

motion for ζ and p

ζ

≡ p

x

+ ip

y

are

˙

ζ =

1

m

p

ζ

,

˙ p

ζ

=

e

m

[p

z

−eRe Ψ(ζ)] [Ψ

(ζ)]

∗

,

where the prime on Ψ means derivative with respect to its argument, and ∗

means complex conjugate.

Answer

Equation (5.50) implies that

˙

ζ ≡ ˙ x + i ˙ y = (p

x

+ ip

y

)/m ≡ p

ζ

/m, as was to

be shown.

1

This is because Amp`ere’s law ∇B = µ

0

j = 0 leads to Laplace’s equation, ∇

2

ψ = 0,

and it is a standard result of complex analysis that the real and imaginary parts of analytic

functions obey the two-dimensional Laplace’s equation.

5.4. CHAPTER 4 PROBLEMS 93

We are given that ψ(x, y) = Re Ψ(ζ). Therefore

∂

x

ψ = Re [∂

x

ζ Ψ

(ζ)] = Re Ψ

(ζ) ,

∂

y

ψ = Re [∂

y

ζ Ψ

(ζ)] = Re [iΨ

(ζ)] ≡ −ImΨ

(ζ) . (5.52)

Thus, from eq. (5.51),

˙ p

ζ

=

e

m

(p

z

−eψ)(∂

x

ψ + i∂

y

ψ)

=

e

m

(p

z

−eRe Ψ)[Re Ψ

(ζ) −i ImΨ

(ζ)]

≡

e

m

(p

z

−eRe Ψ)[Ψ

(ζ)]

∗

(5.53)

as was to be shown.

5.4 Chapter 4 Problems

Problem 4.7.1

Given the Hamiltonian for a particle of unit mass on a sphere in nonrotating

latitude and longitude coordinates, θ and φ

, respectively,

K =

p

2

θ

2R

2

+

p

2

φ

2R

2

cos

2

θ

+V

(θ, φ

, t) , (5.54)

ﬁnd the transformation to the θ, p

θ

, φ, p

φ

canonical coordinates, where φ is

the usual latitude measured in the rotating frame deﬁned in Problem 2.8.1.

That is, φ

**= Ωt + φ. Find the Hamiltonian H in the rotating coordinates.
**

Answer

We treat this as a “point transformation”, as in Sec. 4.2.2, with q = ¦θ, φ¦,

Q = ¦θ

, φ

¦, p = ¦p

θ

, p

φ

¦ and P = ¦p

θ

, p

φ

¦.

We thus have the generating function

F

2

(θ, φ, p

θ

, p

φ

, t) = θ

p

θ

+φ

p

φ

, (5.55)

where θ

≡ θ and φ

≡ Ωt +φ. That is,

F

2

(θ, φ, p

θ

, p

φ

, t) = θp

θ

+ (Ωt +φ)p

φ

. (5.56)

[Note that in eq. (5.55) we have committed the usual physicist’s “abuse of

notation” in taking θ

and φ

**to denote functions of θ and φ rather than
**

94 CHAPTER 5. ANSWERS TO PROBLEMS

independent variables. We avoided this abuse during the formal development

in Sec. 4.2.2 by denoting the corresponding function by G rather than Q.

In practice it is convenient to take it as understood from the context that Q

is a function of q, rather than generate a new notation such as G. This is

convenient, but one needs to be aware of what one is doing.]

The transformation equations are, from Sec. 4.2,

p

θ

=

∂F

2

∂θ

= p

θ

p

φ

=

∂F

2

∂φ

= p

φ

θ

=

∂F

2

∂p

θ

= θ

φ

=

∂F

2

∂p

φ

= Ωt +φ , (5.57)

and

K = H +

∂F

2

∂t

= H + Ωp

φ

. (5.58)

That is, from eq. (5.54),

H = K −Ωp

φ

=

p

2

θ

2R

2

+

p

2

φ

2R

2

cos

2

θ

−Ωp

φ

+V

(θ, Ωt + φ, t) . (5.59)

With the identiﬁcation V

(θ, Ωt + φ, t) = V (θ, φ) we recognize eq. (5.59) as

the same Hamiltonian derived in Problem 3.5.1(c). This derivation makes it

clear that the reason H is not of the form T +V is because the transformation

to the rotating frame is time dependent.

Problem 4.7.2

Using the harmonic oscillator Lagrangian L =

1

2

m( ˙ x

2

−ω

2

0

x

2

)

(a) calculate an approximate action integral S by evaluating the action inte-

gral exactly using the piecewise-linear trial function

x(t) =

1

∆t

[(t

n+1

−t)x

n

+ (t −t

n

)x

n+1

]

for t in each range t

n

≡ n∆t < t < t

n+1

≡ (n + 1)∆t.

5.4. CHAPTER 4 PROBLEMS 95

Answer

In the interval t

n

≤ t ≤ t

n+1

the velocity is constant, ˙ x = (x

n+1

− x

n

)/∆t.

Thus the contribution to the action from this interval is given by

¯

S(x

n

, x

n+1

) ≡

t

n+1

t

n

dt L

=

m∆t

2(x

n+1

−x

n

)

x

n+1

x

n

dx

¸

(x

n+1

−x

n

)

2

(∆t)

2

−ω

2

0

x

2

=

m∆t

2(x

n+1

−x

n

)

¸

(x

n+1

−x

n

)

3

(∆t)

2

−ω

2

0

(x

3

n+1

−x

3

n

)

3

=

m

2

¸

(x

n+1

−x

n

)

2

∆t

−

ω

2

0

∆t

3

(x

2

n+1

+x

n+1

x

n

+ x

2

n

)

. (5.60)

(b) Show that the approximate action integral from t

−N

to t

N

(N > 1 being

an arbitrary integer) is stationary for −N < n < N if x

n

obeys the second-

order diﬀerence equation

x

n−1

−2x

n

+x

n+1

= −

ω

2

0

(∆t)

2

6

(x

n−1

+ 4x

n

+x

n+1

)

Answer

The total action S is a sum, . . . +

¯

S(x

n−1

, x

n

) +

¯

S(x

n

, x

n+1

) + . . ., in which

x

n

occurs in two terms. The condition for S to be stationary with respect

to variations in x

n

is ∂S/∂x

n

= 0. From eq. (5.60), after multiplying by

−∆t/m, this may be written

−

∆t

m

∂

¯

S(x

n−1

, x

n

)

∂x

n

+

∂

¯

S(x

n

, x

n+1

)

∂x

n

= −(x

n

−x

n−1

) +

(ω

0

∆t)

2

6

(2x

n

+x

n−1

)

+ (x

n+1

−x

n

) +

(ω

0

∆t)

2

6

(x

n+1

+ 2x

n

)

= x

n−1

−2x

n

+x

n+1

+

(ω

0

∆t)

2

6

(x

n−1

+ 4x

n

+x

n+1

) = 0 2 .(5.61)

(c) Using the approximate action integral

¯

S(x

n

, x

n+1

) evaluated over the

range t

n

< t < t

n+1

, as a type 1 generating function, F

1

(q, Q) = −

¯

S(x

n

, x

n+1

),

with q = x

n

and Q = x

n+1

, ﬁnd the linear canonical transformation from

(x

n

, p

n

≡ p) to (x

n+1

, p

n+1

≡ P). Show that the determinant of the map is

unity.

This provides a discrete-time dynamical system approximation to the true

continuous-time system for the harmonic oscillator.

96 CHAPTER 5. ANSWERS TO PROBLEMS

Answer

From eq. (5.60), the generating function is

F

1

(x

n

, x

n+1

) = −

m

2

¸

(x

n+1

−x

n

)

2

∆t

−

ω

2

0

∆t

3

(x

2

n+1

+ x

n+1

x

n

+x

2

n

)

. (5.62)

Thus, from eqs. (4.16) and (4.17) the transformation is given by

p

n

=

∂F

1

∂x

n

=

m

2

¸

2(x

n+1

−x

n

)

∆t

+

ω

2

0

∆t

3

(2x

n

+x

n+1

)

,

p

n+1

= −

∂F

1

∂x

n+1

=

m

2

¸

2(x

n+1

−x

n

)

∆t

−

ω

2

0

∆t

3

(2x

n+1

+x

n

)

. (5.63)

For further analysis see the following Maple printout where it is veriﬁed that

the transformation matrix has unit determinant. That this must be the

case follows both from the general principal that canonical transformations

preserve phase-space volume (or, in this case, area) and from the fact that

the determinant in this case is also the Poisson bracket between q and p,

which is unity and conserved under canonical transformation. Thus we have

constructed a symplectic integrator—a discrete-time dynamical system that

preserves some of the Hamiltonian character of the continuous-time dynam-

ical system.

(d) Iterate the map obtained in (c) 100 times taking the initial point as

x

0

= 1, p

0

= 0 and plot the result. Use units such that m = ω

0

= 1 and use

three timesteps: ∆t = 0.1, ∆t = 3.45 and ∆t = 3.47.

For a bonus mark verify one of the solutions by showing that the second-

order diﬀerence equation derived in (b) above (when arranged to give x

n+1

in

terms of x

n

and x

n−1

) gives the same sequence of x-values as the discrete-time

dynamical system in (c) if two successive values of x given by the dynam-

ical system are used to start the second-order diﬀerence equation. I.e. the

discrete time “Lagrangian” and “Hamiltonian” discriptions are dynamically

equivalent. For another bonus mark calculate the eigenvalues for the three

values of ∆t given above.

Again, see the following Maple printout. For ∆t = 0.1 the diﬀerence

approximation gives results very close to the true continuous-time harmonic

5.4. CHAPTER 4 PROBLEMS 97

oscillator solution, which is a circle in phase space. For large enough ∆t

the iteration goes unstable, which occurs somewhere between ∆t = 3.45

and ∆t = 3.47. This can be predicted by evaluating the eigenvalues of the

transformation matrix—because the determinant is unity they must satisfy

λ

1

λ

2

= 1, which means they are either complex-conjugate pairs on the unit

circle in the complex plane (the stable case), or they are real, mutually

reciprocal numbers, one of which must be greater than unity, leading to

instability (exponential growth).

98 CHAPTER 5. ANSWERS TO PROBLEMS

Chapter 6

References and Index

SOME TEXTS on classical dynamics are listed below.

1. The most referred to text is undoubtedly Classical Mechanics (Addison-

Wesley, 1st ed. 1953; 2nd ed. 1980) by Herbert Goldstein, although

it does not have much on modern developments in nonlinear Hamil-

tonian dynamics. The call number is QA805.G6 in the Hancock and

Earth Sciences libraries.

2. Another classic, concise text is Mechanics (Pergamon 1st ed. 1960, 2nd

ed. 1969) by L. D. Landau and E. M. Lifshitz; tr. from the Russian

by J. B. Sykes and J. S. Bell. The call number is QA805.L283 in the

Hancock, Stromlo, Chemistry and Earth Sciences libraries.

3. A clearly written text of similar vintage is Principles of Mechanics by

J.L. Synge and B.A. Griﬃths (McGraw-Hill, New York, 1959). The call

number is QA807.S9 in the Hancock, Stromlo, Chemistry and Physical

Sciences libraries.

4. A good, more modern text is Mechanics : from Newton’s Laws to De-

terministic Chaos (Springer, 1990) by Florian Scheck. The call number

is TA350.S35313 in the Physical Sciences library.

5. Another interesting, though not as carefully written, text is Dynamics

(Wiley, 1983) by S. Neil Rasband. The call number is QA845.R35 in

the Hancock library.

6. A good introductory book is I. Percival and D. Richards, Introduction to

Dynamics (Cambridge University Press, 1982). By limiting attention

to systems with one degree of freedom, the authors have space to devote

to the qualitative analysis of nonlinear dynamical systems, including

99

100 CHAPTER 6. REFERENCES AND INDEX

such modern topics as chaos. The call number is QC133.P46 in the

Hancock library.

7. For a great deal of detail on nonlinear dynamics, written in a style easily

accessible by physicists, see Regular and Chaotic Dynamics (Springer,

2nd ed. 1992; 1st ed. 1983 called Regular and stochastic motion) by

Allan J. Lichtenberg and Michael A. Lieberman. The call number is

QA1.A647 in the Hancock and Stromlo libraries.

8. Classical Dynamics: a Contemporary Approach Jorge V. Jos´e and Eu-

gene J. Saletan (Cambridge U. P., New York, 1998) is a large modern

text. See Physics Today May 1999, p. 66, for a favourable review. The

call number is QA805.J73 in the Hancock library.

9. Fundamentals of Applied Dynamics by James H. Williams (New York,

Wiley, c1996) is a large text used in the ANU Engineering Dynamics

course. It includes the Lagrangian approach and Hamilton’s Principle,

but not Hamilton’s equations. It is written in an interesting manner

with many physical examples having a modern ﬂavour, including treat-

ment of electrical circuits using Lagrangian methods. Despite the use of

the word “applied” in the title, it is actually a very scholarly work and

contains an unusual amount of thoughtful historical discussion. The

call number is QA845.W68 in the Hancock library.

10. A more mathematically oriented book by one of the modern grand

masters in the ﬁeld is Mathematical Methods of Classical Mechanics

by V.I. Arnol’d, translated from the Russian by K. Vogtmann and A.

Weinstein (Springer–Verlag, New York, 1980).

11. An older mathematically oriented classic is A Treatise on the Analytical

Dynamics of Particles and Rigid Bodies by E.T. Whittaker (Cambridge

University Press, Cambridge UK, 1927). QA845.W62 Hancock.

Index

N-body problem, 21

abuse of notation, 93

action, 31

action integral, 24

action integral, phase space, 51

action-angle coordinates, 68

adiabatic invariance, 30, 31

angular momentum, 26, 27, 37

anharmonic oscillator, 38, 84, 89

asymptotic expansion, 32, 62

attractor, 41

Atwood’s machine, 27

autonomous system, 37, 50, 66

auxiliary condition, 3, 10

averaged Lagrangian, 30

billiards, 3

canonical coordinates, 44

canonical momentum, 43, 45, 46

canonical system, 44

canonical transformation, 46, 55

inﬁnitesimal, 63

time evolution, 64

celestial mechanics, 67

chaos parameter, 73

commutator, 65

conﬁguration space, 2

conservation equation, 36

constant of motion, 38, 81

constraint, 42

holonomic, 3, 5, 6, 13, 16, 23,

24, 48, 53, 77, 86

nonholonomic, 4

time-varying, 20, 26

coordinate-free formulation, 33

Coriolis force, 37, 74, 80, 82, 87,

88, 93

curvature vector, 13, 79

cycle, 41

cyclic coordinate, 67

cyclotron motion, 37, 83

d’Alembert

equations, 21

degrees of freedom, 2

diﬀeomorphism, 33

diﬀerence equation, 75, 94

distribution function, 70

dynamical system, 41

continuous time, 41

discrete time, 41

Hamiltonian, 43, 71

dynamics, 15

ecliptic, 77

energy integral, 37, 50

energy methods, 17, 45

enthalpy, 30

equation of motion, 15, 20, 23, 25

Coriolis, 82

discretization, 64, 73

Hamiltonian, 44

harmonic oscillator, 25, 35, 46

particle in compressible ﬂuid,

29

particle in e.m. ﬁeld, 28, 92

101

102 INDEX

Poisson-bracket form, 65

radial, 26

ergodic hypothesis, 71

Euler angles, 15

Euler–Lagrange equation, 8, 23, 51

Eulerian description, 5

expontially small term, 32, 62

extremum condition, 8

ﬁeld, 5

ﬁxed point, 41, 47

elliptic, 47

hyperbolic, 47

stable, 47

unstable, 47

ﬂow, Hamiltonian, 64

ﬂuid

Euler, 6

ideal, 5, 6, 29, 37, 80

ideal equation of state, 6, 29

test-particle Lagrangian, 29

force, 16

applied, 18

constraint, 16, 18

explicit, 18, 23

friction, 19

implicit, 16

Fr´echet derivative, 7

functional derivative, 7

gauge potential, 35

gauge transformation, 34, 35, 46,

56

generalized coordinates, 2

generalized force, 17

generating function, 56, 57

geodesic, 9

curvature, 13, 78

gradient

deﬁnition, 8

inﬁnite-dimensional, 8

Hamilton’s equations of motion, 44

Hamilton’s Principle, 24, 28, 31, 33,

50

Hamilton’s principle

modiﬁed, 51

Hamiltonian, 1, 43, 44

particle in e.m. ﬁeld, 48, 91

Hamiltonian system

(2n + 1)-dimensional, 64

harmonic oscillator, 25, 34, 46

Hessian matrix, 42, 43, 49

ignorable coordinate, 26, 27, 35, 67

inﬁnitesimal generator, 63

inner product, 7

instantaneous frequency, 30

integrability, 68

integral of motion, 26, 36, 38, 50,

66, 81

integrator, numerical, 64

integrator, symplectic, 65

interchange of delta and dot, 7

invariant tori, 68

Jacobian, 5, 43, 69

KAM theorem, 69, 73

kicked rotor, 71

kinematic momentum, 45

kinematics, 1, 15

Korteweg–de Vries equation, 68

Lagrange multiplier, 10, 12

Lagrange’s equations, 21, 23

Lagrangian, 1, 23, 41

natural form, 31, 35, 45

particle in e.m. ﬁeld, 28

particle in ideal ﬂuid, 29

Lagrangian description, 5

Legendre transformation, 45

libration, 47

Lie transform, 66, 67

INDEX 103

Liouville’s Theorem, 70

Lorentz force, 28

manifold, 3, 13, 78

map

area preserving, 71

iterated, 41

period, 64, 71

stroboscopic, 64

mass conservation, 6

mass matrix, 48

Maxwellian distribution, 71

metric tensor, 9, 13, 49

motion in a central potential, 25

natural boundary conditions, 9

natural form, 46, 48

Newton’s second law, 15, 20, 21, 45

generalized, 21

Newton’s third law, 16, 19

Noether’s theorem, 35, 66

nonlinear Schr¨odinger equation, 68

nullspace, 42

objective functional, 6, 10

particle, 16

path, 3, 4, 6, 58

physical, 24

pendulum

physical, 24

simple, 24

perturbation theory, canonical, 66,

67

phase space, 41, 43, 44

phase-space Lagrangian, 50, 57

physical pendulum, 47

Poincar´e section, 64

point transformations, 32, 55, 57,

62

Poisson bracket, 65

polar coordinates, 25, 49

potential

N-body, 21

two-body, 21

quantum mechanics, 62

quasiperiodic motion, 30

reduced mass, 87

reduction to ﬁrst-order ODEs, 42

representation, 2, 33

resonance, 69

rigid rod, 3, 13, 18, 24, 53, 77, 86

rotation, 47

secular behaviour, 47, 61, 68

separatrix, 47

solubility condition, 42

spurious solution, 10

Standard Map, 73

statistical mechanics, 70

subsystem, 16, 87

support of a function, 8

symmetry, 35, 66

test particle, 29

texts on classical dynamics, 99

time reversal, 69

total energy, 37, 45, 46, 48, 50, 87

trial function method, 10, 30, 39,

85

variation

constrained, 10

ﬁrst, 7

second, 8

variational calculus, 6, 51

variational derivative, 7

vector ﬁeld, 41

virtual displacement, 17

virtual work, 17

ii

Contents

1 Generalized Kinematics 1.1 Introduction . . . . . . . . . . . 1.2 Generalized coordinates . . . . 1.3 Example: The ideal ﬂuid . . . . 1.4 Variational Calculus . . . . . . 1.4.1 Example: Geodesics . . 1.4.2 Trial function method . 1.5 Constrained variation: Lagrange 1.6 Problems . . . . . . . . . . . . . 1.6.1 Rigid rod . . . . . . . . 1.6.2 Ecliptic . . . . . . . . . 1.6.3 Curvature of geodesics . 1 1 2 5 6 9 10 10 13 13 13 13 15 15 16 16 20 21 21 23 24 24 25 26 27 28 29 30 31 32

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . multipliers . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

. . . . . . . . . . .

2 Lagrangian Mechanics 2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2 Generalized Newton’s 2nd Law . . . . . . . . . . . . . . . . 2.2.1 Generalized force . . . . . . . . . . . . . . . . . . . . 2.2.2 Generalized equation of motion . . . . . . . . . . . . 2.2.3 Example: Motion in Cartesian coordinates . . . . . . 2.3 Lagrange’s equations (scalar potential case) . . . . . . . . . 2.3.1 Hamilton’s Principle . . . . . . . . . . . . . . . . . . 2.4 Lagrangians for some Physical Systems . . . . . . . . . . . . 2.4.1 Example 1: 1-D motion—the pendulum . . . . . . . 2.4.2 Example 2: 2-D motion in a central potential . . . . 2.4.3 Example 3: 2-D motion with time-varying constraint 2.4.4 Example 4: Atwood’s machine . . . . . . . . . . . . . 2.4.5 Example 5: Particle in e.m. ﬁeld . . . . . . . . . . . 2.4.6 Example 6: Particle in ideal ﬂuid . . . . . . . . . . . 2.5 Averaged Lagrangian . . . . . . . . . . . . . . . . . . . . . . 2.5.1 Example: Harmonic oscillator . . . . . . . . . . . . . 2.6 Transformations of the Lagrangian . . . . . . . . . . . . . . iii

. . . . . . . . . . . . . . . . .

3. . . . . . . . . . . . . . . . . 3. . . . . . . . .2 Perturbation theory . . . 3. . . . . . .7. . . . . .1 Example 1: Adiabatic Oscillator . . . . . .1 Symmetries and integrals of motion 4. . 4. . . . .2 Generating functions . . . . . . . . . . . . . . 4. . . . . . . . .4. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1 Coriolis yet again . . . . . . . . . . . . . .3 Example 1: Scalar potential . . . . . . . . .2. . 4. . . . . 3. . . . . .5 Example 3: Motion in e. . . . . . . . . . . . . . . . . . . . . 4.4 Example 2: Physical pendulum .iv 2. . .1 Introduction . . . . . . . . . .2 Anharmonic oscillator . . .6. . . 4. . . .6. . . . . .7. . . . . . . . . .2 Transformation of Poisson brackets 4. . . 4. . . . . . . . . . . . . .5. . . . . .3 Inﬁnitesimal canonical transformations . .1 Coriolis force and cyclotron motion 2.2 Example 2: Point transformations . . . . . . . . . . . . .6 Properties of canonical transformations . . . . . . .6 Example 4: The generalized N -body system . . . . . . . . .m. . .5 Action-Angle Variables . . . . . . . . . . . . . . . . 4. . . . . . 2. . . .4 Hamilton’s Principle in phase space . . . .4 Poisson brackets . . . . . . . . . . . . . . . . . . . . . . . .2. . . .1 Preservation of phase-space volume 4. . . . . . . . . . . . . . . . . . . . . . . . . . . . .3 Time-Dependent and Autonomous Hamiltonian systems 3. . . . . . . . . . . . . 4 Canonical transformations 4. . . . . . . . . . . . 3.8. . . . . . . . . 2. . . . .3. 3. . . . . . . . . .2 Diﬀerence approximations . . . . . .2 Hamiltonian method . . . . . . . . . . . . . . .1 Point transformations .2 Anharmonic oscillator phase space . 2. . . . . . . . . . . 32 34 35 37 37 37 38 41 41 41 41 43 45 47 48 48 50 50 53 53 53 53 55 55 56 60 62 63 64 65 66 66 67 69 69 74 74 74 75 3 Hamiltonian Mechanics 3. . . . . . .5. . . 4. . . . .7. . . . 3. . . . . 4. .1 Time symmetry . . . .2 Gauge transformations . . . .1 Lagrangian method . . . . . . . . . . . . .2. .2. . . 4. . . . .2. . . . . . . . . . . . . .1 Time evolution . . . 3. . . . . . . . . . . . . . . . . . . . . . . . . .5 Problems . .6. .6.2. . . . . . . . . . .1 Constraints and moving coordinates . . . . . . . 4. . . . . . . . .8. . . . . . . . . 3. . . . . . . . . . . . . . . . . . . . . . . .1 Introduction: Dynamical systems . . . . . . . . . . . . . . . . . CONTENTS . . . . . .2. . . 2. . . . . . 3.2. . . . . . . 4. . . . . . potentials . . .2 Mechanics as a dynamical system . . . . . . . . . . . 2. . 3.5.3 2-D motion in a magnetic ﬁeld . . . 3. .8 Problems . .7 Problems . .7 Symmetries and Noether’s theorem . . . . . . . . . . .4. .

. . . . . . . . . . . . 5. . . . . . . . . . . . . . .2 Chapter 2 Problems . . . . . . . . . . . . . .1 Chapter 1 Problems . . . 6 References and Index v 77 77 80 86 93 99 . . . . . . . . . . . . 5. . . . . . . . 5. . .CONTENTS 5 Answers to Problems 5. . . . . . . . . . . . . . . . . .4 Chapter 4 Problems . . . . . .3 Chapter 3 Problems . . . . . . . . . .

vi CONTENTS .

In the following notes the convention will be used of italicizing the ﬁrst use or deﬁnition of a concept. but in which the dynamics can be nonetheless encapsulated in a single scalar function: the Lagrangian. or the Hamiltonian. Thus we will be revisiting such examples as the harmonic oscillator and the pendulum. This abstract viewpoint is enormously powerful and underpins quantum mechanics and modern nonlinear dynamics. It is concerned with the description of possible motions of general systems rather than how to calculate 1 . The present chapter is essentially geometric. In this course we will develop a more abstract viewpoint in which one thinks of the dynamics of a system described by an arbitrary number of generalized coordinates. such as motion of a particle in a gravitational potential. named after the French mathematician Joseph Louis Lagrange (1736–1813). It may or may not be more efﬁcient than elementary approaches for solving simple problems. named after the Irish mathematician Sir William Rowan Hamilton (1805–1865).Chapter 1 Generalized coordinates and variational principles 1. the simple harmonic oscillator etc. The index can be used to locate these deﬁnitions and the subsequent occurrences of these words.1 Introduction In elementary physics courses you were introduced to the basic ideas of Newtonian mechanics via concrete examples. but when examples are set in this course please remember that you are expected to use the approaches covered in the course rather than fall back on the methods you learnt in First Year. but in order to feel comfortable with the formalism it is very instructive to do some elementary problems using abstract methods.

ez to er . y.rN . because of the change in the unit basis vectors from ex . Thus we call the topic generalized kinematics. The number n = 3N of generalized coordinates qi is called the number of degrees of freedom. z2 . qi = {r1 . . or a more general curvilinear coordinate system. x2 . xN . No particular metric is assumed—e.g. we could equally as well use spherical polar coordinates (see Fig. . . y1 . then the dynamics of the total system is described by the motion of a point q ≡ {qi |i = 1.. ey . φ) . In other systems the generalized “coordinates” need not even be spatial coordinates—e. zN } in a 3N -dimensional generalized conﬁguration space.1) However its representation changes. On the other hand. 1. distinguishing generalized coordinate arrays from 3-vectors by using a bold slanted font for the former and a bold upright font for the latter. φ1 . GENERALIZED COORDINATES AND VARIATIONAL PRINCIPLES physical motions from knowledge of forces. .1). they could r y be the charges ﬂowing in an elecθ trical circuit.2CHAPTER 1. OfFigure 1. yN . θ. φN }. 1. (1. θ1 . 2. Thus the convenient φ vector-like notation q for the array x of generalized coordinates should not be confused with the notation r for the position vector in 3-space. 1.1: Position vector ten the set of generalized coordir in Cartesian and Spherical nates is simply denoted q. The vector r represents a point in physical 3-space and thus does not change when we change coordinates. z r r = xex + yey + zez = rer (θ. . . eφ and ez . .r2 . but in Polar coordinates. θN . y2 . . whereas the set of generalized coordinates changes if we change variables. . For instance. Vectors are entities independent of which coordinates are used to represent them. the sets of .1. 3N } = {x1 . θ2 . consider the position vector of a particle in Cartesian coordinates x.2 Generalized coordinates Suppose we have a system of N particles each moving in 3-space and interacting through arbitrary (ﬁnite) forces. these notes we use a bold font.g. z1 . φ in Fig. z and in spherical polar coordinates r. . φ2 . .

1 For instance. or particles connected by rigid rods.2: Some possible paths in conﬁguration space. qn Figure 1. q2 q(t2) q(t1) q(t) q1 q3 . In such cases. · · · . φ. Sometimes the motion is constrained to lie within a submanifold of the full conﬁguration space. t. GENERALIZED COORDINATES 3 generalized coordinates r C ≡ {x. y.2) the constraints are said to be holonomic. even though extension to a global nonsingular curvilinear coordinate system may not be possible (as. where there exists a set of (functionally independent) constraint equations. 2.. 1 .2. e. j = 1. m < n .1. z} and r sph ≡ {r. Each holonomic constraint reduces the number of degrees of freedom by one. A manifold can always be regarded as a surface embedded in a higher dimensional space.. we may be interested in the motion of billiard balls constrained to move within a plane. (1. on a sphere). or auxiliary conditions fj (q) = 0 .g. since it allows us to express one of the original generalized coordinates as A manifold is a mathematical space which can everywhere be described locally by a Cartesian coordinate system. z} are distinct entities: they are points in two diﬀerent (though related) conﬁguration spaces describing the particle. each parametrized by the time.

4CHAPTER 1. GENERALIZED COORDINATES AND VARIATIONAL PRINCIPLES a function of the others and delete it from the set. For instance, for a particle constrained to move in a horizontal plane (an idealized billiard table), the vertical position z = const is a trivial function of the horizontal coordinates x and y and the conﬁguration space becomes two dimensional, q = {x, y}. Consider the set of all conceivable paths through conﬁguration space (see Fig. 1.2). Each one may be parametrized by the time, t: q = q(t). By diﬀerentiating eq. (1.2) with respect to time, we ﬁnd a set of constraints on the generalized velocities, qi ≡ dqi / dt, which we write in diﬀerential notation ˙ as n ∂fj (q) ∂fj (q) dqi ≡ · dq = 0 , (1.3) ∂qi ∂q i=1 where ∂f /∂q ≡ {∂f /∂q1 , . . . , ∂f /∂qn } and we use the shorthand dot-product notation n ∂f ∂f a· ≡ ai , (1.4) ∂b ∂bi i=1 where ai and bi are arbitrary conﬁguration space variables. The condition for functional independence of the m constraints is that there be m nontrivial solutions of eq. (1.3), i.e. that the rank of the matrix ∂fj (q)/∂qi be its maximal possible value, m. Note that not all such diﬀerential constraints lead to holonomic constraints. If we are given constraints as a general set of diﬀerential forms

n

ωi (q) dqi = 0 ,

i=1

(j)

(1.5)

then we may or may not be able to integrate the constraint equations to the form eq. (1.2). When we can, the forms are said to be complete diﬀerentials. When we cannot, the constraints are said to be nonholonomic. The latter case, where we cannot reduce the number of degrees of freedom by the number of constraints, will not be considered explicitly in these notes. Furthermore, we shall normally assume that any holonomic constraints have been used to reduce {qi |i = 1, . . . , n} to a minimal, unconstrained set. However, we present in Sec. 1.5 an elegant alternative that may be used when this reduction is not convenient, or is impossible due to the existence of nonholonomic constraints. There are situations where there is an inﬁnite number of generalized coordinates. For instance, consider a scalar ﬁeld (such as the instaneous amplitude of a wave), ψ(r, t). Here ψ is a generalized coordinate of the system and the position vector r replaces the index i. Since r is a continuous variable it ranges over an inﬁnite number of values.

1.3. EXAMPLE: THE IDEAL FLUID

5

x(r0,t) r0

Figure 1.3: A ﬂuid element advected from point r = r0 at time t = 0 to r = x(r0 , t) at time t.

1.3

Example: The ideal ﬂuid

As an example of a system with both an inﬁnite number of degrees of freedom and holonomic constraints, consider a ﬂuid with density ﬁeld ρ(r, t), pressure ﬁeld p(r, t) and velocity ﬁeld v(r, t). Here we are using the Eulerian description, where the ﬂuid quantities ρ, p and v are indexed by the actual position, r, at which they take on their physical values at each point in time. However, we can also index these ﬁelds by the initial position, r0 , of the ﬂuid particle passing through the point r = x(r0 , t) at time t (see Fig. 1.3). This is known as the Lagrangian description. (cf. the Schr¨dinger o and Heisenberg pictures in quantum mechanics.) We shall denote ﬁelds in the Lagrangian description by use of a subscript L: ρL (r0 , t), pL (r0 , t) and vL (r0 , t) = ∂t x(r0 , t). The ﬁeld x(r0 , t) may be regarded as an inﬁnite set of generalized coordinates, the speciﬁcation of which gives the state of the ﬂuid at time t. The Jacobian J(r0 , t) of the change of coordinates r = x(r0 , t) is deﬁned by J ≡ ∂x ∂x ∂x · × ∂x0 ∂y0 ∂z0

∂x ∂x0 ∂x ∂y0 ∂x ∂z0 ∂y ∂x0 ∂y ∂y0 ∂y ∂z0 ∂z ∂x0 ∂z ∂y0 ∂z ∂z0

=

,

(1.6)

where x0 , y0 and z0 are Cartesian components of r0 and x, y and z are the corresponding components of x(r0 , t). This gives the change of volume of a ﬂuid element with initial volume dV0 and ﬁnal volume (at time t) dV through dV = J(r0 , t) dV0 . (1.7) To see this, consider dV0 = dx0 dy0 dz0 to be an inﬁnitesimal rectangular box, as indicated in Fig. 1.3, with sides of length dx0 , dy0 , dz0 . This ﬂuid

6CHAPTER 1. GENERALIZED COORDINATES AND VARIATIONAL PRINCIPLES element is transformed by the eﬀect of compression and shear to an inﬁnitesimal parallelipiped with sides given by dlx ≡ dx0 ∂x/∂x0 , dly ≡ dy0 ∂x/∂y0 , dlz ≡ dz0 ∂x/∂z0 . The volume of such a parallepiped is dlx · dly × dlz = J dx0 dy0 dz0 2. Are the ﬁelds ρ(r0 , t) and p(r0 , t) additional generalized coordinates which need to be speciﬁed at each point in time? In an ideal ﬂuid (i.e. one with no dissipation, also called an Euler ﬂuid) the answer is no, because mass conservation, ρ dV = ρ0 dV0 , allows us to write ρL (r0 , t) = ρ0 (r0 )/J(r0 , t) , (1.8)

while the ideal equation of state p( dV )γ = p0 ( dV0 )γ , where γ is the ratio of speciﬁc heats, gives pL (r0 , t) = p0 (r0 )/J γ (r0 , t) , (1.9) where ρ0 and p0 are the initial density and pressure ﬁelds, respectively. These are, by deﬁnition, ﬁxed in time, so the only time dependence occurs through the Jacobian J, which we showed in eq. (1.6) to be completely determined by the Lagrangian displacement ﬁeld x(r0 , t). Thus eqs. (1.8) and (1.9) have allowed us to reduce the number of generalized coordinate ﬁelds from 5 to 3 (the three components of x)—mass conservation and the equation of state have acted as holonomic constraints. Note: Mass conservation is valid even for nonideal ﬂuids (provided they are not reacting and thus changing from one state to another). However, in a ﬂuid with ﬁnite dissipation, heat will be generated by the motion and entropy will be increased in each ﬂuid element, thus invalidating the use of the adiabatic equation of state. Further, the entropy increase depends on the complete path of the ﬂuid through its state space, not just on its instantaneous state. Thus the pressure cannot be holonomically constrained in a nonideal ﬂuid. Remark 1.1 A useful model for a hot plasma is the magnetohydrodynamic (MHD) ﬂuid—an ideal ﬂuid with the additional property of being a perfect electrical conductor. This leads to the magnetic ﬁeld B(r, t) being “frozen in” to the plasma, so that BL also obeys a holonomic constraint in the Lagrangian representation, but as it is a vector constraint it is a little too complicated to give here.

1.4

Variational Calculus

Consider an objective functional I[q], deﬁned on the space of all diﬀerentiable paths between two points in conﬁguration space, q(t1 ) and q(t2 ), as depicted

(1. deﬁned by δf ∂f d ∂f ≡ − . we in fact assume the paths to ¨ be slightly smoother than simply diﬀerentiable.15) ≡ δI δq. The crucial step here is the lemma “delta and dot commute”.13). is deﬁned to be the change in I as estimated by linearizing in δq: t2 ∂f ∂f ˙ δI[q] ≡ dt δq(t)· + δ q(t)· . δI.13) This consists of an endpoint contribution and an integral of the variational derivative δf /δq. (1.1. for variations δqi which vanish in the neighbourhood of the endpoints. as t2 n δI[q] ≡ t1 dt i=1 δqi (t) . dt (1. Thus we vary the path by an amount δq(t): q(t) → q(t) + δq(t). (1.11) in terms of δq(t). δI (t) . ˙ δq ∂q dt ∂ q (1. That is ˙ δq ≡ dδq .) We suppose our task is to ﬁnd a path that makes I a maximum or minimum (or at least stationary) with respect to neighbouring paths. 1.2 The right-hand side of eq. (1.4. We can now integrate by parts to put δI in the form ∂f δI[q] = δq· ˙ ∂q t2 t2 + t1 t1 dt δq(t)· δf . Then the ﬁrst variation.2 t2 7 I[q] ≡ t1 ˙ dt f (q(t). δq .11) ˙ ∂q ∂q t1 Our ﬁrst task is to evaluate eq. δq (1. q(t). When using this termie nology the notation δI/δq is used instead of δf /δq so that we can write eq. (1.14) Remark 1.12) ˙ To prove this. simply go back to deﬁnitions: δ q ≡ d(q + δq)/ dt − dq/ dt = dδq/ dt 2. δqi (1. VARIATIONAL CALCULUS in Fig. t) .10) (As we shall wish to integrate by parts later.14) is also sometimes called the functional derivative or Fr´chet derivative of I[q]. so that q is also deﬁned.

To determine the nature of a stationary point we would need to expand I to second order in δq—the second variation. I can only be stationary for all such variations if and only if the variational derivative vanishes for each value of t and each index i qi δf =0. ·) used above is a kind of inﬁnite-dimensional dot product where we not only sum over the index i. That is. but integrate over the “index” t. and we can move it outside the integral. (1. δq (1. (The support of a function is just the range over which it is nonzero.13) dinate qi with support in the range t − to t + . by analogy.16). (1. Clearly then. consider functions with arbitrarily localized support in t. as an inﬁnite-dimensional gradient deﬁned on the function space of paths. with some ascending and some descending “directions”. the endpoint contribution in eq. A typical variational problem is to make I extremal or stationary under arbitrary variations δq(t) holding the endpoints ﬁxed.4: A time-localized δf /δq(t) becomes essentially convariation in generalized coorstant over the support of q in eq.4. we require δI = 0 ∀ functions δq(t) such that δq(t1 ) = δq(t2 ) = 0 . we see that the functional derivative δI/δq may be thought of. (1. Sometimes we encounter variational problems where we wish to extremize I under .16) Note that this condition does not necessarily require I to be a minimum or maximum—it can be a kind of saddle point in function space. If we recall that the change in the value of a ﬁeld deﬁned on 3-space.) As → 0. Figure 1. we can.8CHAPTER 1. 1.13) vanishes. in particular. which may be regarded as the deﬁnition of the gradient ϕ. leaving only the contribution of the integral over t. In the class of variations in eq. ϕ(r). (1. as t−ε t+ε indicated in Fig. due to an arbitrary inﬁnitesimal change δr is δϕ = δr· ϕ. GENERALIZED COORDINATES AND VARIATIONAL PRINCIPLES which may be taken as the most general deﬁning equation for δI/δq. e.g. The inner product notation (·.17) These n equations are known as the Euler–Lagrange equations. Since δq(t) is arbitrary.

4.j such that n ( dl) = i. given a metric tensor gi. δq(t1 ) = δq(t2 ) = 0. Another well-known result is that the shortest path between two points on the surface of a sphere is a great circle (see Problem 1. in addition to eq.17). stationarity implies the natural boundary conditions ∂f =0 (1.20) A geodesic is a curve between two points whose length (calculated using the given metric) is stationary against inﬁnitesimal variations about that path. in purely geometric applications t is simply an arbitrary label for the position along a path.19) In terms of our parameter τ . VARIATIONAL CALCULUS 9 variations of the endpoints as well.13) that. Perhaps the best known result on geodesics is the fact that the shortest path between two points in a Euclidean space (one where gi.1. (1. (1. τ )qj ˙ ˙ .18) ˙ ∂q at t1 and t2 . paths in conﬁguration space are naturally parametrized by the time. 1.j = 0 for i = j and gi. we thus have the length l as a functional of the form discussed above τ2 n 1/2 l= τ1 dτ i.4. but can have physical interpretations.j = 1 for i = j) is a straight line. suppose we want to ﬁnd the shape of an elastic string stretched over a slippery surface. The string will adjust its shape to minimize its elastic energy.j=1 2 dqi gi.6. The distance along a path is given by integrating the lengths dl of inﬁnitesimal line elements.1 Example: Geodesics In the above development we have used the symbol t to denote the independent variable because. Thus the task of ﬁnding geodesics ﬁts within the class of variational problems we have discussed.3 for a general theorem on geodesics on a curved surface). Geodesics are not necessarily purely geometrical objects. In such cases we see from eq. However.j dqj . (1. (1. in applications in dynamics. and we shall in this section denote it by τ to avoid confusion.j=1 qi gi. For instance.j (q. and we can use the Euler–Lagrange equations to ﬁnd them. Since the elastic potential energy is a .

2. ∂ak k = 1. 1. . K . Varying the ak the variation in I is K δI = k=1 ∂I δak . that the K-dimensional gradient of I vanish.8. aK ) as to the general form of the solution. . . (Of course.2 Trial function method One advantage of the variational formulation of a problem is that we can use trial function methods to ﬁnd approximate solutions. (1.11) (analytically or numerically) and seek a stationary point of the resulting function I(a1 . a2 . but several geodesics between two points. . a fact which explains the phenomenon of gravitational lensing (multiple images of a distant galaxy behind a closer massive object). 5.2 in Sec. . (1. . there may be spurious stationary points that must be rejected because they cannot possibly be close to a true solution—see the answer to Problem 2. . k = 1. aK ) in the K-dimensional space of the parameters ak . GENERALIZED COORDINATES AND VARIATIONAL PRINCIPLES monotonically increasing function of the length of the string. because the world line of a photon is a geodesic in 4-dimensional space time. the string will settle onto a geodesic on the surface. 1. If the metric is suﬃciently distorted. if our guessed trial function solution is close to the true solution the error in I will be small.5 Constrained variation: Lagrange multipliers As mentioned in Sec. Then we evaluate the integral in eq. . .4.22) that is. K. Geodesics also play an important role in General Relativity.) 1. using some speciﬁc function q K (the trial function) involving a ﬁnite number of parameters ak . . with the metric tensor obeying Einstein’s equations. . a1 .2 we normally assume that the holonomic constraints have been used to reduce the dimensionality of the conﬁguration space so . . a2 . we can make a clever guess. . Since the true solution makes the objective functional I stationary with respect to small variations. ∂ak (1. That is.21) The condition for a stationary point is thus ∂I = 0. q(t) = q K (t. .10CHAPTER 1. it can happen that there is not one. . . because q K may not be a reasonable guess for the solution in all ranges of the parameters.

1.5. CONSTRAINED VARIATION: LAGRANGE MULTIPLIERS

11

that all variations are allowed. However, it may not be possible to do an analytic elimination explicitly. Or it may be that some variables appear in a symmetric fashion, making it inelegant to eliminate one in favour of the others. Thus, even in the holonomic case, it is worth seeking a method of handling constrained variations: when there are one or more diﬀerential auxiliary conditions of the form δf (j) = 0. In the nonholonomic case it is mandatory to consider such variations because the auxiliary conditions cannot be integrated. We denote the dimension of the conﬁguration space by n. Following eq. (1.5) we suppose there are m < n auxiliary conditions of the form δf (j) ≡ ω (j) (q, t)·δq = 0 . (1.23)

The vectors ω (j) , j = 1, . . . , m may be assumed linearly independent (else some of the auxiliary conditions would be redundant) and thus span an mdimensional subspace, Vm (t), of the full n-dimensional linear vector space Vn occupied by the unconstrained variations. Thus the equations eq. (1.23) constrain the variations δq to lie within an (n − m)-dimensional subspace, Vn−m (q, t), complementary to Vn . The variational problem we seek to solve is to ﬁnd the conditions (the generalizations of the Euler–Lagrange equations) under which the objective functional I[q] is stationary with respect to all variations δq in Vm (q, t). Apart from this restriction on the variations, the problem is the same as that described by eq. (1.16). The generalization of eq. (1.17) is δf ·δq = 0 ∀ δq ∈ Vn−m (q, t) . δq (1.24)

If there are no constraints, so that m = 0, then δf /δq is orthogonal to all vectors in Vn and the only solution is that δf /δq ≡ 0. Thus eq. (1.24) and eq. (1.17) are equivalent in this case. However, if m < n, then δf /δq can have a nonvanishing component in the subspace Vm and eq. (1.17) is no longer valid. An elegant solution to the problem of generalizing eq. (1.17) was found by Lagrange. Expressed in our linear vector space language, his idea was that eq. (1.24) can be regarded as the statement that the projection, (δf /δq)n−m , of δf /δq into Vn−m (q, t) is required to vanish. However, we can write (δf /δq)n−m as δf /δq −(δf /δq)m , where (δf /δq)m is the projection of δf /δq into Vm . Now observe that we can write any vector in Vm as a linear superposition of the ω (j) since they form a basis spanning

**12CHAPTER 1. GENERALIZED COORDINATES AND VARIATIONAL PRINCIPLES this space. Thus we write (δf /δq)m = − δf δq
**

m

λj ω (j) , or, equivalently, (1.25)

+

m j=1

λj ω (j) = 0 ,

where the λj (q, t) coeﬃcients, as yet to be determined, are known as the Lagrange multipliers. They can be determined by dotting eq. (1.25) with each of the m basis vectors ω (j) , thus providing m equations for the m unknowns. Alternatively, we can express this variationally as δf δq

m

+

m j=1

λj ω (j) ·δq = 0 ∀ δq ∈ Vm (q, t) .

(1.26)

Since − λj ω (j) is the projection into Vm of δf /δq, the projection into the complementary subspace Vn−m is found by subtracting (− λj ω (j) ) from δf /δq. That is, δf δq δf = + δq = 0,

m

λj ω (j)

j=1

n−m

(1.27)

where the Vn−m component of the second equality follows by eq. (1.24) and the Vn component from eq. (1.25). Thus we have n generalized Euler– Lagrange equations, but they incorporate the m equations for the, so far arbitrary, λ(j) implicit in eq. (1.25). Thus we really only gain (n − m) equations from the variational principle, which is at it should be because we also get m kinematic equations from the constraint conditions—if we got more from the variational principle the problem would be overdetermined. The variational formulation of the second equality in eq. (1.27) is δf + δq

m

λj ω (j) ·δq = 0 ∀ δq ∈ Vn (q, t) .

j=1

(1.28)

That is, by using the Lagrange multipliers we have turned the constrained variational problem into an unconstrained one. In the holonomic case, when the auxiliary conditions are of the form in eq. (1.2), we may derive eq. (1.28) by unconstrained variation of the modiﬁed objective functional

t2 m

I∗ [q] ≡

t1

dt (f +

j=1

λj fj ) .

(1.29)

The auxiliary conditions also follow from this functional if we require that it be stationary under variation of the λj .

1.6. PROBLEMS

13

1.6

1.6.1

Problems

Rigid rod

Two particles are connected by a rigid rod so they are constrained to move a ﬁxed distance apart. Write down a constraint equation of the form eq. (1.2) and ﬁnd suitable generalized coordinates for the system incorporating this holonomic constraint.

1.6.2

Ecliptic

˙ Suppose we know that the angular momentum vectors rk ×mk rk of a system of particles are all nonzero and parallel to the z-axis in a particular Cartesian coordinate system. Write down the diﬀerential constraints implied by this fact, and show that they lead to a set of holonomic constraints. Hence write down suitable generalized coordinates for the system.

1.6.3

Curvature of geodesics

Show that any geodesic r = x(τ ) on a two-dimensional manifold S : r = X(θ, ζ) embedded in ordinary Euclidean 3-space, where θ and ζ are arbitrary curvilinear coordinates on S, is such that the curvature vector κ(τ ) is everywhere normal to S (or zero). The curvature vector is deﬁned by κ ≡ de / dl, where e (τ ) ≡ dx/ dl is the unit tangent vector at each point along the path r = x(τ ). ˙ ˙ ˙ Hint: First ﬁnd f (θ, ζ, θ, ζ) = l, the integrand of the length functional, l = f dτ (which involves ﬁnding the metric tensor in θ, ζ space in terms of ∂X/∂θ and ∂X/∂ζ). Then show that, for any path on S, ∂X ∂f =e · ˙ ∂θ ∂θ ˙ (and similarly for the ζ derivative) and ∂f d ∂X =e · , ∂θ dτ ∂θ and again similarly for the ζ derivative.

GENERALIZED COORDINATES AND VARIATIONAL PRINCIPLES .14CHAPTER 1.

For example. In this chapter we deal with one formulation (due to Lagrange) of generalized dynamics—the derivation of diﬀerential equations (equations of motion) for the time evolution of the generalized coordinates. 2. If we know the Cartesian components Fi (r. the most important task of the physicist is over when the equations of motion have been derived—the rest is just mathematics or numerical analysis (important though these are). x2 = y. From elementary mechanics we are all familiar with Newton’s Second Law. 3. The goal of generalized dynamics is to ﬁnd universal forms of the equations of motion. nonlinear) equations of motion specify the motion uniquely. t). F = ma for a particle of mass m subjected to a force F and undergoing ˙ an acceleration a ≡ ¨. the generalized coordinates for a rigid body could be the three Cartesian coordinates of the centre of mass and three angles to specify its orientation (known as the Euler angles). Indeed. r i = 1. r. which may be connected by holonomic constraints so the number n of generalized coordinates may be less than 3N . Thus. x To give a physical framework for developing our generalized dynamical formalism we consider a set of N Newtonian point masses. x3 = z and their ﬁrst time derivatives then the equations of motion are the set of three second-order diﬀerential equations m¨i − Fi = 0. of the force in terms of the Cartesian coordinates x1 = x. in the case of a rigid body N is essentially inﬁnite. in a sense. but the number of generalized coordinates is ﬁnite. so n = 6 for a rigid body allowed to move freely in space. 15 . these (in general.1 Introduction The previous chapter dealt with generalized kinematics—the description of given motions in time and space. Given appropriate initial conditions.Chapter 2 Lagrangian Mechanics 2.

. This variational principle forms a basis for generalizing even beyond Newtonian mechanics (e. Having found a very general form of the equations of motion (Lagrange’s equations). rather than the vector quantity force. depends also on deﬁning a basis .16 CHAPTER 2. to dynamics in Special Relativity). .2 2. The basic reason is that the energy. . and generalized forces conjugate to them.g. needs only speciﬁcation of the coordinates for its full description. and we already know from elementary physics that it is unnecessary—we do not really need an inﬁnite number of equations to describe the motion of a rigid body. rather. the number of generalized coordinates satisﬁes the inequality n ≤ 3N . LAGRANGIAN MECHANICS Whether the point masses are real particles like electrons.2. on the system) is the primitive concept. whereas the representation of the force. a scalar quantity. a vector. 2. W . we then ﬁnd a variational principle (Hamilton’s Principle) that gives these equations as its Euler–Lagrange equations in the case of no frictional dissipation. . In a naive Newtonian approach we would have to specify the forces acting on each particle (taking into account Newton’s Third Law. derive the 3N equations of motion for each particle and then eliminate all the interior subsystem coordinates to ﬁnd the equations of motion of the generalized coordinates only. its change due to the performance of external work. we shall refer to them generically as “particles”.1 Generalized Newton’s 2nd Law Generalized force Let the (constrained) position of each of the N particles making up the system be given as a function of the n generalized coordinates q by rk = xk (q. If there are holonomic constraints acting on the particles. or mathematical idealizations like the inﬁnitesimal volume elements in a continuum description. composite particles like nuclei or atoms. “action and reaction are equal and opposite”). t). the system may be divided into two subsystems—an “exterior” subsystem described by the n generalized coordinates and an “interior” constraint subsystem whose (3N − n) coordinates are rigidly related to the q by the geometric constraints. All the interior coordinates and the forces required to maintain their constrained relationships to each other (the “forces of constraint”) should be implicit only. To achieve this it turns out to be fruitful to adopt the viewpoint that the total mechanical energy (or. What we seek is a formulation in which only the generalized coordinates. This is clearly very ineﬃcient. appear explicitly. Thus. k = 1. N .

and is in general distinct from an alternative resolution. Illinois. conjugate to each . the example above suggests we deﬁne the set of n generalized forces. The choice of basis set is not obvious when we are using generalized coordinates. q 2 . With the development of Lagrangian and Hamiltonian methods. 309–312. δWx . Turning now to the N -particle system as a whole. Note that the virtual displacements are done at a ﬁxed instant in time as if by some “invisible hands”. but if we imagine the thought experiment of displacing the particle in the three independent directions.] The displacements δx ex .] To illustrate the relation between force and work. see e. Fz = δWz /δz. q 3 (the superscript notation being conventional in tensor calculus). δr = δx ex .M14 Hancock. determining the work. y. GENERALIZED NEWTON’S 2ND LAW 17 set on which to resolve it. which perform the work δW : this is a thought experiment—the displacements are ﬁctitious. Fx = δWx /δx. [Historically. force came to be understood earlier. ﬁrst consider just one particle. the contravariant representation {F i }. done in each case.2. rather than the contravariant. δWz . Ernst Mach “The Science of Mechanics” (Open Court Publishing. Fz = ∂W/∂z. Fy = δWy /δy.g. A single displacement δr does not give enough information to determine the three components of F. (You will meet this terminology again in relativity theory. and thermodynamics. then we will have enough equations to deduce the three components of the force vector. QA802. δy ey and δz ez . The corresponding “virtual work” is given by δW = i δq i Fi .) The energy approach shows that the covariant. Recall that the work δW done on the particle by a force F as the particle suﬀers an inﬁnesimal displacement δr is δW = F·δr ≡ Fx δx + Fy δy + Fz δz.2. 1960) pp. La Salle. δy ey and δz ez are historically called virtual displacements. this can be used to determine the generalized forces Fi by determining the virtual work done in three independent virtual displacements. energy-based approaches can now be said to be dominant in physics. As in the Cartesian case. They are really simply the same displacements as used in the mathematical deﬁnition of partial derivatives. δWy . where Fi ≡ ei ·F. [If δW can be integrated to give a function W (r) (which is not always possible). In tensor calculus the set {Fi } is known as the covariant representation of F. not dynamical. z to an arbitrary curvilinear coordinate system q 1 . where the basis vectors ei are in general not orthonormal. Fy = ∂W/∂y. {Qi }. components of the force form the natural generalized forces conjugate to the generalized coordinates q i . Then an arbitrary virtual displacement is given by δr = i δq i ei . then we may use standard partial derivative notation: Fx = ∂W/∂x. Suppose we now transform from the Cartesian coordinates x. but energy also has a long history.

) By forces of constraint. we do not need to know1 . but this is confusing because they need not originate externally to the system. sliding planes etc. plus any interaction forces between particles such as those due to elastic Of course.18 CHAPTER 2. LAGRANGIAN MECHANICS of the n degrees of freedom qi . but the usage “explicit forces” seems new—often they are called “applied forces”. Furthermore. we can in particular take all but one of the δqi to be zero to pick out the ith component. such as those due to an external gravitational or electric ﬁeld. joints. but also from interactions between the particles. ∂qi (2. (2. for most purposes. to be such that the virtual work done on the system in displacing it by an arbitrary inﬁnitesimal amount δq at ﬁxed time t is given by n δW ≡ i=1 Qi δqi ∀ δq . The virtual work is N δW = k=1 n Fk ·δrk N = i=1 δqi k=1 Fk · ∂xk ∂qi . and noting that they hold for any δq. so they cannot be assumed known a priori.1) and eq.2).3) When there are holonomic constraints on the system we decompose the forces acting on the particles into what we shall call explicit forces and forces of constraint. (2.1) We now calculate the virtual work in terms of the displacements of the N particles assumed to make up the system and the forces Fk acting on them. The explicit force on each particle. we mean those imposed on the particles k by the rigid rods. (2. is the vector sum of any externally k imposed forces. giving the generalized force as N Qi = k=1 Fk · ∂xk . we may not be able to tell what these forces need to be until we have solved the equations of motion. (2. Fxpl .2) Comparing eq. in practical engineering design contexts one should at some stage check that the constraint mechanism is capable of supplying the required force without deforming or breaking! 1 . (The latter terminology is standard. These forces simply adjust themselves to whatever values are required to maintain the geometric constraint equations and can be regarded as “private” forces that. that make up the holonomic constraints on the system. Fcst .

no work can be done on the interior constraint subsystem by the virtual displacements. so as to give no acceleration in the normal direction and thus maintain the constraint. undeformable mechanisms. the nonfriction forces acting on the constraint mechanisms are equal and opposite to the forces of constraint. not into mechanical energy within the constraint subsystem. (2. It is subject to the force of gravity. We now make the crucial observation that. and perhaps by the generalized velocity q in the case of velocity-dependent forces such as those due to friction and those acting on a charged particle moving in a magnetic ﬁeld. If there is friction acting on the particle. an inclined plane as deand its magnitude is that required to null scribed in the text. That is. It does no work because it Figure 2. ∂qi (2. |N| = mg cos α. GENERALIZED NEWTON’S 2ND LAW 19 springs coupling point masses.3) because the Fxpl are known in terms of k . Fcst . this is a much more practical expression than eq.1 shows a simple system with N a holonomic constraint—a particle slidF ing on a plane inclined at angle α. These k are the “public” forces that determine the dynamical evolution of the degrees of freedom of the system and are determined by the conﬁguration q of the ˙ system at each instant of time.1: A body on is orthogonal to the direction of motion. Note: If there are friction forces associated k with the constraints there is work done against these. The force of constraint is N. Figure 2. Thus k the sum over Fcst ·δxk vanishes and we can replace the total force Fk with k Fxpl in eqs. By Newton’s Third Law.3).2. no (net) virtual work is done against the nonfriction forces imposed by the particles on the constraint mechanisms in performing the variations δq. including that due to constraint mechanisms.4) For the purpose of calculating the generalized forces. the normal force N. out the normal component of the gravitational force. and a friction force F in mg α the directions shown. or to electrostatic attractions between charged particles. N Qi = k=1 Fxpl · k ∂xk . mg. That is.2) and (2. but this fact does not negate the above argument because we have included the friction forces in the explicit forces—any work done against friction forces goes into heat which is dissipated into the external world. because the constraints are assumed to be provided by rigid. then that must be included in Fxpl as well.2. (2.

q. 2. respectively). ∂qi (2.9) ˙ which follows because d/ dt can be replaced by ∂/∂t + q·∂/∂q.7) ˙ Diﬀerentiating vk wrt qi (treating q and q as independent variables in partial ˙ derivatives) we immediately have the lemma ∂xk ∂vk = . t). (2. (2. ∂qi ∂ qi ˙ d ∂xk ∂vk = . Applying these two lemmas in eq. the “interchange of delta and dot” lemma in Sec.6) ≡ k=1 mk We now diﬀerentiate rk = xk (q(t). Thus Qi = Qi (q. LAGRANGIAN MECHANICS ˙ the instantaneous positions and velocities. t) . which commutes with ∂/∂qi (cf.3) in the form N Fk · k=1 ∂xk ˙ = Qi (q.20 CHAPTER 2. r N mk ¨k · r k=1 N ∂xk ∂qi d dt ˙ rk · ∂xk ∂qi ˙ − rk · d ∂xk = Qi . q. t) . dt ∂qi (2.8) The second lemma about partial derivatives of vk that will be needed is (2. t) with respect to t to ﬁnd the function ˙ ˙ vk such that rk = vk (q. t): ∂xk ˙ rk = + ∂t n qj ˙ j=1 ∂xk ˙ ≡ vk (q. 1.4).5) To derive an equation of motion we use Newton’s second law to replace Fk on the left-hand side of eq. ∂qj (2. dt ∂qi ∂qi (2.10) . (2. q.5) with mk ¨k .4) has been used to determine the Q as functions ˙ of the q (and possible q and t if we have velocity-dependent forces and timedependent constraints. (2.6) we ﬁnd N k=1 N d dt mk vk · d ∂ dt ∂ qi ˙ ∂vk ∂ qi ˙ − mk vk · − ∂ ∂qi ∂vk ∂qi 2 1 2 mk vk = k=1 2 1 2 mk vk = Qi . q.2 Generalized equation of motion We now suppose that eq. Then we rewrite eq. (2.2.

2. · · · . For instance. in the classical N -body problem the particles are assumed to interact pairwise via a two-body interaction potential . ˙ ∂y ˙ ∂T = mz . Scheck p. LAGRANGE’S EQUATIONS (SCALAR POTENTIAL CASE) In terms of the total kinetic energy of the system. rN ). ˙ ∂x ˙ ∂T = my . (2. T N 21 T ≡ k=1 2 1 2 mk vk .3. (2. m¨ = Fy y m¨ = Fy z (2. but this may be historically inaccurate so is best avoided.g. V (r1 . (2. z}. 83) d’Alembert’s equations.10) compactly as the generalized Newton’s second law d ∂T ∂T − = Qi . They are also sometimes called (e. (2. but we shall reserve this term for a later form [eq. from eq.13) (2. In this case T = 1 m(x2 + y 2 + z 2 ) ˙ ˙ ˙ 2 so ∂T ∂T ∂T = = =0 ∂x ∂y ∂z ∂T = mx . (2.24)] arising when we assume a special (though very general) form for the Qi .2.12) These n equations are sometimes called Lagrange’s equations of motion. r2 . ˙ ∂z ˙ (2.16) 2. y. dt ∂ qi ∂qi ˙ (2.15) Also.3 Example: Motion in Cartesian coordinates Let us check that we can recover Newton’s equations of motion as a special case when q = {x.3 Lagrange’s equations (scalar potential case) In many problems in physics the forces Fk are derivable from a potential .12) we immediately recover Newton’s 2nd Law in Cartesian form m¨ = Fx x as expected.11) we write eq.14) and (2.4) we see that Qi ≡ Fi . 2. Substituting in eq.

20) where the ek are the charges on the particles and 0 is the permittivity of free space. . form we have summed all the oﬀ-diagonal entries of the matrix but have compensated for the double counting by dividing by 2. |rk − rl | k. the interparticle distance rk.l (|rk − rl |) such that the force on particle k due to particle l is given by Fk.l ≡ |rk − rl |.l (rk − rl ) U (rk.l = − = − k Vk.19) In the ﬁrst line we counted the interactions once and only once: noting that Vk. where zk is the height of the kth particle with respect to a horizontal reference plane and mk is its mass. so that the matrix of interactions is symmetric we have kept only those entries below the diagonal to avoid double counting. The exclusion of the diagonal “self-interaction” potentials is indicated by putting a prime on the .k .l (2.l ) . rl ) ≡ Uk. Then the total force on particle k is found by summing the forces on it due to all the other particles Fk = − l=k k Vk.l kV = − .17) where the prime on U denotes the derivative with respect to its argument. (2. (2. more symmetric.l=1 = Vk.l . 0r (2.l is the Coulomb interaction Uk.l (r) = ek el .l = Vl.l (rk . Physical examples of such an N -body system with binary interactions are: • An unmagnetized plasma.18) where the N -body potential V is the sum of all distinct two-body interactions N V ≡ k=1 l≤k N Vk. where Vk.22 CHAPTER 2. LAGRANGIAN MECHANICS Vk. We could also allow for the eﬀect of gravity by adding the potential k mk gzk to V . In the second.l 1 2 k.

q.l (r) = (2. from eq. such as gravity.12) we see that the generalized Newton’s equations of motion can be encapsulated in the very compact form (Lagrange’s equations of motion) d dt ∂L ∂ qi ˙ − ∂L =0.24) with eq.17) we see that Lagrange’s equations of motion have exactly the same form as the Euler–Lagrange equations for the variational principle δS = 0.4) we have N Qi = − k=1 ∂xk · ∂qi kV = − ∂ V (q. where Vk. (2. LAGRANGE’S EQUATIONS (SCALAR POTENTIAL CASE) 23 • A globular cluster of stars. (2. Then.1 Hamilton’s Principle Comparing eq. r2 . ∂qi (2.24) ˙ where the function L(q. if the gas is too dense (or becomes a liquid) we would have to include 3-body or higher interactions as the wave functions of more than two atoms could overlap simultaneously. rN ) becomes a function. · · · .25) 2. Uk. we see that the potential V (r1 . Even when the system is subjected to external forces. r where G is the gravitational constant.21) • A dilute monatomic gas. ∂qi (2. V (q.26) .3. where the functional S[q]. where Vk. (2.23) Substituting this form for Qi in eq.l is the Van der Waal’s interaction. is deﬁned as L≡T −V . t). we can often still assume that the “explicit forces” are derivable from a potential Fxpl = − k kV . in the reduced conﬁguration space.22) Taking into account the constraints. (2. (2. t). (1. t) .l is the gravitational interaction Gmk ml . t) . q. (2. However. deﬁned by t2 S≡ t1 ˙ dt L(q.3. and/or holonomic constraints. called the Lagrangian.2.

where g is the acceleration ˙ due to gravity. θ) = 2 ml2 θ2 − mgl(1 − cos θ) .2: Physspherical pendulum and compound pendula. To go beyond the original Newtonian dynamics with a scalar potential that we used to motivate Lagrange’s equations. and the velocity of the bob is vθ = lθ.2. swinging from a ﬁxed pivot at one end and (1−cosθ)l with a bob of mass m attached at the other. is thus z 1 ˙ ˙ L(θ.1 Lagrangians for some Physical Systems Example 1: 1-D motion—the pendulum One of the simplest nonlinear systems is the one-dimensional physical pendulum (so called to distinguish it from the linearized harmonic oscillator approximation). (This system is also called the simple pendulum to distinguish it from the Figure 2. We can now state Hamilton’s Principle: Physical paths in conﬁguration space are those for which the action integral is stationary against all inﬁnitesimal variations that keep the endpoints ﬁxed. out of all those that are consistent with the constraints. making an angle θ with the vertical.24 CHAPTER 2. By physical paths we mean those paths. (2. that actually obey the equations of motion with the given Lagrangian.27) . T − V .4. described by the generalized coordinate θ.4 2. Since the natural boundary conditions eq. so that the kinetic 2 2 ˙2 1 1 energy T = 2 mvθ = 2 ml θ . the variational principle is one in which the endpoints are to be kept ﬁxed. LAGRANGIAN MECHANICS is known as the action integral . ical pendulum.) The potential energy with respect to the equilibrium position θ = 0 is V (θ) = mgl(1 − cos θ). (1. As depicted in Fig. we can instead take Hamilton’s Principle. being such a simple and geometrically appealing result. The constraint l = const and the assumption of plane motion reduces the system to one 0 mg degree of freedom. 2. as a more fundamental and natural starting point for Lagrangian dynamics. which have more than one degree of freedom.18) are not physical. The Lagrangian. the pendulum consists of a light rigid rod of θ l length l. 2.

(2.4. so the physical pendulum provides a paradigm for problems such as the motion of an electron in a crystal lattice or of an ion or electron in a plasma wave.28) Expanding the cosine up to quadratic order in θ gives the harmonic oscillator oscillator approximation (see also Sec.29) 2 ¨ for which the equation of motion is. an electron orbiting about a proton in the Bohr model of the hydrogen atom. (2. (2.2) ˙ L ≈ Llin ≡ 1 ml2 θ2 − 1 mglθ2 .30) (2. We now consider the restricted two body problem—one light particle orbiting about a massive particle which may be taken to be ﬁxed at r = 0 (e. where er is the unit vector in the radial direction ˙ and eθ is the unit vector in the azimuthal direction. 2. (2. ˙ ˙ ˙ y = r sin θ + rθ cos θ . 2 2 (2.31) An alternative derivation of eq. so that ˙ x = r cos θ − rθ sin θ .4. (2. (2. the Lagrangian equation of motion is ¨ ml2 θ = −mgl sin θ . dividing through by ml2 . ˙ ˙ whence the kinetic energy T ≡ 1 (x2 + y 2 ) is found to be ˙ 2 ˙ ˙ T = 1 m r2 + r2 θ2 ˙ 2 .27) ∂L/∂ θ = ml2 θ and ∂L/∂θ = −mgl sin θ. such that x = r cos θ . LAGRANGIANS FOR SOME PHYSICAL SYSTEMS 25 This is also essentially the Lagrangian for a particle moving in a sinusoidal spatial potential.21)) is a function only of the radial distance from the central body and not of the angle.2.20) or eq. with ω0 ≡ g/l. ˙ ˙ From eq. q = {r.2 Example 2: 2-D motion in a central potential Let us work in plane polar coordinates.g. 2. θ}. .32) may be found by resolving v into the ˙ components rer and rθeθ . or a planet orbiting about the sun). Then the potential V = V (r) (given by eq. θ + ω0 θ = 0.32) y = r sin θ .6. Thus. (2.

34) ˙ dt ∂ θ which we may immediately integrate once to get an integral of the motion. (2. (2. 1 ˙ L = T = 2 m u2 + (a − ut)2 θ2 . ˙ ∂θ (2.33) ∂L ˙ = mrθ2 − V (r) . r (2. ˙ 2 (2. Turning now to the r-component r of Lagrange’s equations. (2. (2. (2.4. ∂r d ∂L ∂L = mr .e.33) expresses conservation of angular momentum.26 CHAPTER 2. from eq.36) r ∂r ˙ dt ∂ r ˙ From eq. (2. that is being pulled steadily downward at speed u through a hole at the origin so that the radius r = a−ut. Then ∂L/∂θ ≡ 0 and the θ component of Lagrange’s equations.25) the Lagrangian is ˙ L = 1 m r2 + r2 θ2 − V (r) . eq. a dynamical quantity that is constant along the trajectory ∂L = const .3: Planar motion with a time-varying centripetal constraint as described in the text. consider instead a weight rotating about the origin on a frictionless horizontal surface (see Fig. Thus eq.24) becomes d ∂L =0.32).(2. we see from θ eq. ˙ = m¨ .3) and constrained by a thread. substituting for r in eq. 2. which is the angular momentum.33) we see that ∂L/∂ θ = mr2 θ. Then the Lagrangian is. (2.33) First we observe that L is independent of θ (in which case θ is said to be ignorable). i.35) ˙ ˙ From eq. (2.3 Example 3: 2-D motion with time-varying constraint Instead of free motion in a central potential.38) . initially of length a. (2.37) u Figure 2.24) we ﬁnd the radial equation of motion to be ˙ m¨ − mrθ2 = −V (r) . LAGRANGIAN MECHANICS Then. 2.

2. The fact that the constraint is time-dependent is irrelevant to this postulate.4. LAGRANGIANS FOR SOME PHYSICAL SYSTEMS 27 Now only θ is an unconstrained generalized coordinate. as depicted in Fig. The kinetic and potential energy are T = 1 ˙2 2 (m1 + m2 )x and V = m1 gx − m2 gx. Thus L = T − V is given by L = 1 (m1 + m2 )x2 − (m1 − m2 )gx ˙ 2 (2. 2. it is ignorable.2. The height of weight 1 is x with respect to the chosen origin and the holonomic constraint provided by the rope allows us to express the height of weight 2 as −x.39) which equation can be integrated to give θ as a function of t. 2.4. and so we again have conservation of angular momentum ˙ m(a − ut)2 θ ≡ l = const . the string is obviously doing work on the 1 system because T = 2 [mu2 + (l2 /m)/(a − ut)2 ] is not conserved.4. inertialess pully of radius a by a rope of ﬁxed length. (2.2 of no work being done by the constraints? The answer is “no” because what we assumed in Sec. because virtual displacements are done instantaneously at any given time. θ = θ0 + (l/mu)[1/(a − ut) − 1/a] = θ0 + lt/[ma(a − ut)]. 2.2 was that no virtual work was done by the constraints. Clearly angular momentum is conserved.4 Example 4: Atwood’s machine Consider two weights of mass m1 and m2 suspended from a frictionless. so that there is only one degree of freedom for this system.4: Atwood’s machine.41) m1 + m2 . because the purely radial string cannot exert any torque on the weight.40) m1 m2 Figure 2. x 0 −x and its derivatives are ∂L/∂x = −(m1 − m2 )g and ∂L/∂ x = (m1 +m2 )x. However. (2. Have we not therefore violated the postulate in Sec. 2. Thus Lagrange’s equation of motion gives the correct answer. so that the equation of motion d(∂L/∂ x) = ∂L/∂x ˙ ˙ ˙ becomes m1 − m2 x=− ¨ g.2. As before.

However we do not wish to use A.24) reproduce the equation of motion of a charged particle in an electromagnetic ﬁeld.42) and (2. Thus we can ˙ ˙ only use r and r. equivalently. it is obviously of great physical importance to ﬁnd a Lagrangian for which Lagrange’s equations of motion eq. ﬁeld The fact that Lagrange’s equations are the Euler–Lagrange equations for the extraordinarily simple and general Hamilton’s Principle (see Sec. q3 } = {x. (2. q2 .43) The electrostatic potential energy is eΦ. respectively. ˙ The three vectors available are A itself. t) + e˙ ×B(r. z} we have ∂L ∂Φ ∂Aj = −e +e qj ˙ .m. r 2 r Taking q ≡ {q1 . since A·A in the Lagrangian would give an equation of motion that is nonlinear in the electromagnetic ﬁeld.42) where e is the charge on the particle of mass m. Thus we do not deﬁne L as T − V .1) suggests that Lagrange’s equations of motion may have a wider range of validity than simply problems where the force is derivable from a scalar potential.28 CHAPTER 2. In particular. but rather postulate the universal validity of Lagrange’s equations of motion (or. so let us try adding that to form the total Lagrangian L = 1 m˙ 2 − eΦ + e˙ ·A . under the inﬂuence of the Lorentz force. to be given in terms of the scalar potential Φ and vector potential A by the standard relations E = − Φ − ∂t A . so we need to dot A with one of the naturally occurring vectors in the problem to create a scalar.3. contrary to eq.4. m¨ = eE(r. but how do we include the vector potential? Clearly we need to form a scalar since L is a scalar.44) (2. y. so we expect part of the La1 r grangian to be 2 m˙ 2 − eΦ.5 Example 5: Particle in e.43) we see that r·A has the same dimensions as Φ. (2. 2. for describing non-dissipative classical dynamics and accept any Lagrangian as valid that gives the physical equation of motion. r and r. LAGRANGIAN MECHANICS 2. (2.42). B = ×A . Hamilton’s Principle). Comparing eqs. ∂qi ∂qi ∂qi j=1 3 (2. (2.45) . We assume the electric and magnetic ﬁelds E and B. t) . r r (2.

γρ . Dividing by ρ we write the equation of motion of a ﬂuid element as dv p =− − dt ρ V .4.45) and (2. where γ is the ratio of speciﬁc heats. Thus p = ρ = p ρ p ρ − − p ρ2 ρ p .6 Example 6: Particle in ideal ﬂuid In Sec. (2. so our guessed Lagrangian is indeed correct.47) This is simply eq.46) Substituting eqs. prescribed functions of r and t the problem becomes only three-dimensional.48) where V (r. (2. However.42) in Cartesian component form.2. then we can take the gradient of the log of eq.3 we presented a ﬂuid as a system with an inﬁnite number of degrees of freedom.46) in eq. (2. taking the pressure p and mass density ρ as known. ργ (2. 1. In an ideal compressible ﬂuid we have the equation of state p( dV )γ = p0 ( dV0 )γ . LAGRANGIANS FOR SOME PHYSICAL SYSTEMS and dAi d ∂L ∂Ai = m¨i + e q = m¨i + e q + dt ∂ qi ˙ dt ∂t 3 29 qj ˙ j=1 ∂Ai ∂qj .49) where the right-hand side is a constant of the motion for the given test particle. if we concentrate only on the motion of a single ﬂuid element (a test particle). (2.24) we ﬁnd ∂Φ ∂Ai ∂Aj ∂Ai m¨i = e − q − +e qj ˙ − ∂qi ∂t ∂qi ∂qj j=1 3 . t) is the potential energy (usually gravitational) per unit mass. (2. If we further assume that it is the same constant for all ﬂuid elements in the neighbourhood of the test particle. (2. 2. which we can write as p = const . (2.4. To ﬁnd a Lagrangian for this motion we need to be able to combine the pressure gradient and density into an eﬀective potential.49) to get ( p)/p = γ( ρ)/ρ.

51) Using eq.3. 1.55) Since ω d ln Ak / dt.30 Solving for p/ρ we get CHAPTER 2. Then we may use a trial function of the form q(t) = q(φ(t).) (2. In particular. . but not of ˙ ˙ A1 . to a ﬁrst approximation we may keep only the ﬁrst term in eq. γ −1ρ (2. Now take the time integration in the action integral to be over a time long compared with the period of oscillation. but short compared with the . 2.53) where q is a 2π-periodic function of φ. see Problem 2.2).52) L = 1 r2 − h − V . Thus our approximate L is a function of ω. .50) where the enthalpy (per unit mass) is deﬁned by h≡ γ p . and so we expect that trial function methods (see Sec. the phase of the rapid oscillations.4. Thus. and the Ak are a set of slowly varying amplitudes characterizing the waveform (e.1 we know that Lagrangian dynamics has a variational formulation. ˙ + A1 ∂φ ∂A1 ∂A2 (2. 2˙ 2.8. A1 (t).2) may be useful as a way of generating approximate solutions of the Lagrangian equations of motion.48) we recognize it as the equation of motion for a particle of unit mass with total potential energy h + V .g.54) where the instantaneous frequency is deﬁned by ˙ ω(t) ≡ φ(t) . (2. (2. (2. Thus the Lagrangian is (2. A2 etc. A2 (t). suppose we know that the solutions are oscillatory functions of t with a frequency much higher than the inverse of any characteristic time for slow changes in the parameters of the system (the changes being then said to occur adiabatically).5 Averaged Lagrangian From Sec. . LAGRANGIAN MECHANICS p = ρ h (2.50) in eq.54). (2. ˙ q = ω(t) ∂q ˙ ∂q + A2 ∂q + · · · .

q2 . . Thus. A1 . .61) Now consider that the length of the spring. ∂ω (2. . A2 .57) for the class of oscillatory physical solutions we seek. ˙ 2 (2. 2. with d ln ω0 / dt ω0 . . k = 1.5. we have a new. A1 (t).2.5. Then the kinetic energy is T = 1 mx2 and the potential energy is ˙ 2 2 1 V = 2 kx2 = 1 mω0 x2 . approximate form of Hamilton’s Principle in which the averaged Lagrangian replaces the exact Lagrangian. . T − V . J = const.) (2.} replaces the set q1 . so that ω0 = ω0 (t). Thus only the phase-average of L.1 Example: Harmonic oscillator Consider a weight of mass m at the end of a light spring with spring constant 2 k = mω0 . . (2. changes slowly with time. .60) The second set of equations gives relations between ω and the Ak s that give the waveform and the instantaneous frequency. Requiring S to be stationary within the class of quasiperiodic trial functions then gives the new adiabatic Euler–Lagrange equations ¯ d ∂L =0.56) 2π 0 contributes to the action. 2.59) ∂Ak The ﬁrst equation expresses the conservation of the adiabatic invariant. . AVERAGED LAGRANGIAN 31 timescale for changes in the system parameters. . is 2 2 L = 1 m(x2 − ω0 x2 ) . so L depends only on its time derivative ω. A2 (t). . and in which the set {φ. . . t2 S≈ t1 ¯ dt L(ω. as the generalized variables. Because . (2. ˙ dt ∂ φ and (2.58) ¯ ∂L = 0 . . Note that the averaging ¯ in eq. or perhaps the spring constant. where J is known as the oscillator action J≡ ¯ ∂L .56) removes all direct dependence on φ. Thus the natural form of the Lagrangian. 2π dφ ¯ L≡ L (2.

(2. (2. q ≡ {qi |i = 1. LAGRANGIAN MECHANICS the system is changing with time.1 Our derivation of adiabatic invariance has been carried out only to ﬁrst order in . 2 (2. going as exp(−1/| |). 4. (2. (For an alternative derivation of adiabatic invariance in this case.g. q.1 Transformations of the Lagrangian Point transformations ˙ Given an arbitrary Lagrangian L(q. eq. ω = ±ω0 (t) (which ¯ also implies that L = 0). since asymptotically small changes.) 2. a Cartesian frame). (This does not however mean that adiabatic invariance is exact.2. Thus an adiabatic invariant can be deﬁned to arbitrary order in . see Sec.6. is mωA2 J= . this implies that the instaneous frequency tracks the changing natural frequency.6 2.) Since the problem is linear.65) 1/2 From the adiabatic invariance of J we see that A varies as the inverse of ω0 . the amplitude of oscillation is not constant.59) from varying A gives ¯ ∂L mA 2 2 = ω − ω0 = 0 .32 CHAPTER 2. are not ruled out.g. we often want to know . t) in one generalized coordinate system. e. Calculating the averaged Lagrangian as in eq. (2. it can be shown that an averaged Lagrangian asymptotically independent of φ to all orders in a power series in can be found. the small parameter expressing the ratio between the short oscillatory timescale and the long slow timescale.62) ˙ In response to the slow changes in ω0 we expect A and ω ≡ φ also to be slowly varying functions of time. we use a sinusoidal trial function with no harmonics x = A sin φ(t) . However. Remark 2. which allows us to predict how the amplitude changes with time. We now use the average Lagrangian method to ﬁnd the adiabatic invariant. (2.1.63) 4 The Euler–Lagrange equation eq.60). The adiabatically conserved action J. L= (2. n} (e.64) ∂A 2 Since we assume that the amplitude is not zero.56) we have 2 2 ¯ mA ω 2 − ω0 .

66) We require the inverse function of g also to be twice diﬀerentiable. then S is automatically invariant under the coordinate change and will be stationary for the same physical paths. t) = L(g(Q. . We have in fact already run into a special case of this transformation rule. 2. 2. t) = L(q(t). . The transformation g maps a path in Q-space to a path in q-space. in which case g : Q → q is said to be a C 2 diﬀeomorphism. Q ≡ {Qi |i = 1. g(Q.70) but clearly Hamilton’s Principle provides a much simpler and more elegant way of arriving at the same result. where we transformed the kinetic and potential energies from Cartesians to polars in the obvious way.g. .67) for any path. so transformations to a moving frame are allowed. q(t). i = 1. . n} of twicediﬀerentiable functions gi such that qi = gi (Q.3).69) One can prove that eq. t) .4. n .4.26) is an integral over time only.2.4. (2.70 are simply the Euler– Lagrange equations for S to be stationary in the new variables. t) . (2. t) (2. 1990) and showing that eq. Q. This is another virtue of Hamilton’s Principle— δS = 0 on a physical path for all variations with ﬁxed endpoints—since the action integral S given by eq. Qj ∂Qj j=1 n i = 1. (2. 2. (2. .68) gives the correct dynamics by calculating the transformation of the Euler–Lagrange equations explicitly (see e. suppose there exists a set g ≡ {gi |i = 1. We can guarantee this trivially. Q. simply by choosing the new Lagrangian to be the old one in the new coordinates: ˙ ˙ ˙ L (Q. However.2 and Sec. n . . t) in another generalized coordinate system. it is physically the same path—all we have changed is its representation. irrespective of what coordinates they are represented in.2 and Sec. since eqs.68) where ∂gi gi ≡ ˙ + ∂t ˙ ∂gi . t). Q. 2. (2. if we can deﬁne L so that ˙ ˙ L (Q(t).g. Scheck. Thus. . Q(t). TRANSFORMATIONS OF THE LAGRANGIAN 33 ˙ the Lagrangian L (Q. n} (e. (2. .3. t). in Sec. ∂Qi (2.6. .4. polar coordinates as in the examples in Sec.24) implies d dt ∂L ˙ ∂ Qi − ∂L =0. What we need in order to discuss how the Lagrangian transforms is a coordinate-free formulation of Lagrangian dynamics. Note that we have allowed the transformation to be time dependent. 2. Thus.

2 ). That naturally raises the question: for a given system is there only one Lagrangian giving the correct equations of motion. 2. t2 ) − M (q(t1 ). q. and that is to multiply L by a constant factor. eq. δS ≡ δS for all admissible variations) and therefore S is stationary if and only if S is. there is a more important source of nonuniqueness.2 ) do not aﬀect the allowed variations of S (i. t) is any C 3 ˙ function. and thus. the new action integral becomes t2 S ≡ t1 t2 ˙ dt L (q. t1 ) . t1. but nevertheless found a function for which Lagrange’s equations gave the correct dynamical equations of motion. known as a gauge transformation of the Lagrangian in which L is replaced by L .71) L (q. the endpoint terms M (q(t1. ˙ (2.71). rather than the speciﬁc form of L. LAGRANGIAN MECHANICS 2. from eq. or are there many? Clearly there is a trivial way to generate multiple physically equivalent Lagrangians.2 Gauge transformations We saw in Sec. we usually normalize L in a natural way. q.4)].e. so this freedom is not encountered much in practice. that Hamilton’s Principle apply) that is fundamental. = (2.4. Example 1: Harmonic oscillator Consider the harmonic oscillator Lagrangian.34 CHAPTER 2.5 that we cannot always expect that the Lagrangian is of the form T − V . (2. M . deﬁned by ∂M ∂M ˙ ˙ ˙ + q· (2. equivalently.6. by requiring that the part linear in the mass be equal to the kinetic energy. t) ˙ dt L(q. Thus both Lagrangians give identical physical paths and L and L are completely equivalent dynamically. t) + ∂t ∂q [using the shorthand notation deﬁned in eq. t) + d M (q. Thus it is the requirement that the equations of motion be in the form of Lagrange’s equations (or. t) = L(q. where M (q. e. We recognize the term in M as the total time derivative.26) and eq.61). (2. (1.72) However. (2. q. t) dt t1 = S + M (q(t2 ). since the endpoints are held ﬁxed in Hamilton’s Principle. for a particle of unit mass 2 1 L = 2 (x2 − ω0 x2 ) . However. q.73) .g. However.

74) 1 Now add a gauge term. eq. taking M = 2 ω0 x2 . eq. the new Lagrangian becomes 2 L = 1 (x2 + 2ω0 xx − ω0 x2 ) . (2.71) with M = eχ.43) are not unique. T − V . (2. gives the harmonic oscillator equation 2 x = −ω0 x .74). t) ≡ Φ(r. (2. where s is a real continuous parameter and .24).2. r 2 r Substituting eqs. (2.7. Using the gauge-transformed potentials Φ and A to deﬁne a new Lagrangian L in the same way as L was deﬁned by eq. t) → A (r. Since ∂L/∂ x = x and ∂L/∂x = ω0 x. we see that the gauge contributions cancel. ¨ (2. and ∂L /∂x = ω0 x+ω0 x and substituting into ˙ ˙ ˙ the Lagrangian equation of motion. (2.7 Symmetries and Noether’s theorem A Lagrangian L is said to have a continuous symmetry if it is invariant under the transformation q → hs (q). t) + χ(r. (2. Then. t) − ∂t χ(r.78) (2. SYMMETRIES AND NOETHER’S THEOREM 35 2 where we take ω0 to be constant. Example 2: Electromagnetic gauge transformation It is well known that the scalar and vector potentials in eqs.71). and the new Lagrangian is a perfectly valid one despite the fact that it is no longer in the natural form.77) which is exactly of the form eq. ∂t (2. from eq.44) we have L = 1 m˙ 2 − eΦ + e˙ ·A . t) Φ(r. ˙ 2 ˙ (2.77) we ﬁnd L =L+ ∂(eχ) ˙ + r· (eχ) . since the electric and magnetic ﬁelds are left unchanged by the gauge transformation A(r. Thus we do indeed recover the harmonic oscillator equation.76) in eq. 2.76) where χ is an arbitrary scalar function (a gauge potential ).75) 2 Calculating ∂L /∂ x = x+ω0 x. (2. (2. (2. we ˙ ˙ immediately verify that Lagrange’s equation. t) . t) ≡ A(r. t) → Φ (r. Thus electromagnetic and Lagrangian gauge transformations are closely related.

Thus. which we write in the form d ∂L = ∂q dt ∂L ˙ ∂q .36 CHAPTER 2. 2 2 (2.2 that a system with an ignorable coordinate always has a conservation equation. This is a special case of a general theorem proved by Emmy Noether in 19182 that ˙ any time-independent continuous symmetry of the Lagrangian. This ability of the Lagrangian formalism to generate conservation relations is one of its most important features. That is. (2. t) generates an integral of motion. For example. so that if δS ≡ δL dt = 0 on the path q(t) then it also vanishes for all members of the family q(t. 0) ≡ q(t) is a physical path. t) ≡ ∂q ∂s ≡ dI =0. q. giving rise to an integral of motion. 1. q. given any path (not necessarily physical) q(t). s) generated by the symmetry operation hs .80) Using eq. in the case of motion in a central potential.79) where d/ dt denotes the total derivative along the path s = const. then so are all the paths q(t. the continuous symmetry is invariance of L under rotations: θ → θ + s. 2. ˙ ∂s ∂q ∂s ∂ q ∂s (2. (2. Since L is constant over this family of paths. dt (2. given one solution of the equations of motion we can use a symmetry to generate a continuous family of other solutions. .81) ∂L ∂q · ˙ ∂ q ∂s is the integral of motion I generated by the symmetry h. Consider a family q(t. Comparing two inﬁnitesimally close paths we see that δL also possesses this symmetry. s). We saw in Sec. LAGRANGIAN MECHANICS hs=0 is the identity transformation.4. L = L(q. 186 (1971). its derivative with respect to s at each point in time vanishes ˙ ∂L ∂L ∂q ∂L ∂ q = · + · .80) and using the timeindependence of h to allow interchange of the order of diﬀerention with respect to s and t we derive the conservation equation d dt where ∂L ˙ ∂q · ∂q ∂L d + · ˙ ∂s ∂ q dt ˙ I(q. s) ≡ hs (q(t)).24). (2. (2. s) obeying the Lagrange equation of motion eq. L has the same value on all members of the continuous family of paths q(t. if q(t. from Hamilton’s Principle. Thus.82) English translation in Transport Theory and Statistical Physics.79) to eliminate ∂L/∂q in eq.

¯ the new Lagrangian L is deﬁned to be dq dt ¯ . (2. q → q.6). (2. L q. At ﬁrst sight this seems very diﬀerent from the spatial symmetries just discussed. one not subject to external forcing).) 2.8. L is independent of time. (2. as this applies on a physical path. Sec. τ could be the “proper time” in relativity theory.3). consider the motion of particle of unit mass constrained to move on the surface of a perfectly smooth sphere of radius R rotating with angular velocity Ω about the z-axis. L possesses the symmetry of invariance with respect to translations in t. and so on.83) Then.g. so I can be identiﬁed as the negative of the total energy (cf. q = q(τ ).1 Problems Coriolis force and cyclotron motion As a model of the motion of a ﬂuid element or dust particle in a planetary (e. φ) (see Sec. eq.7. That is. Earth’s) atmosphere. but eq. 1 dq dt/ dτ dτ dt . PROBLEMS 37 For example. For Lagrangians of the form T − V is easily seen that I = −T − V .82) gives the integral of motion ∂L ˙ I = L − q· .84) ˙ ∂q where we have set t = τ after doing the partial diﬀerentiations. eq.4. dτ (2. 2. where θ and φ are the latitude and longitude respectively.e. dτ dτ ≡ L q.g. 2. expressing L in polar coordinates we see that the integral of motion generated by rotational symmetry is the angular momentum. (E.8. but in fact it is possible to generate an integral of motion (the energy integral ). invariance under translations in the x-direction generates the conservation of linear momentum in the x-direction. (2.82) applies in any coordinate system. using the trick of regarding t as a generalized coordinate by making it a function of a new “time” parameter τ : t = t(τ ).8 2. (Similarly. Suppose the force on the particle is given by an eﬀective potential V (θ. 3.1 Time symmetry For an autonomous system (i. . using invariance under t → t + s.35).) ¯ In order for the new action integral L dτ to be the same as the old one.2.

e.e. so that the equations of motion admit the solution ˙ ˙ θ = φ = 0 at each latitude). (c) Consider a charged particle constrained to move on a non-rotating smooth insulating sphere.) Draw a sketch of a typical weather-map “low” (i. the particle velocity with respect to the rotating frame. corresponding to a particle of mass m oscillating along the x-axis under the inﬂuence of a nonideal spring (i. 2. (Hint: The pressure gradient is parallel to the enthalpy gradient. y = R cos θ sin(φ+Ωt). clockwise in the Southern Hemisphere.38 CHAPTER 2. where w is the “wind velocity”. Among this class of potentials ﬁnd the special case V = V0 (θ) required to make equilibrium possible (i.e. taking the generalized coordinates to be the latitude and longitude so that z = R sin θ. φ) = V0 (θ) + h(θ. Assuming the winds are slow. It is much simpler to work with the scalar quantity w· p. counter-clockwise in the Northern Hemisphere for Ω > 0). noting that the origin of the θ coordinate is at the equator. i. and ﬁnd a ﬁrst integral (i. Now assume that the general form of the potential is V (θ. one . φ).e.8. so that second order time derivatives of θ and φ (and products of ﬁrst order time derivatives) can be neglected. immersed in a uniform magnetic ﬁeld B = Bez . where “cyclonic” in geophysical ﬂuid dynamics means “in the direction of the planet’s rotation” (i.e. 2. constant of the motion) in the case where V is independent of longitude. Hence show that motion in a depression or low is cyclonic. show that the velocity of the particle in the rotating frame is at right angles to the pressure gradient − p. You don’t have to work out the velocity and gradient vectors in the rotating angular coordinates. LAGRANGIAN MECHANICS (a) Write down the Lagrangian in a frame rotating with the planet. on which the electrostatic potential is a function of latitude and longitude. where h is the enthalpy (proportional to p(γ−1)/γ ) deﬁned in Sec. x = R cos θ cos(φ+Ωt). y. not at the pole as in normal spherical polar coordinates. (Draw a diagram illustrating these coordinates. z is a non-rotating Cartesian frame. a localized depression in p) showing the pressure contours.e.4.6.) Write down the equations of motion. where x.2 Anharmonic oscillator Consider the following potential V . Write down the Lagrangian in the same generalized coordinates as above and show it is the same as that for the particle on the rotating planet with appropriate identiﬁcations of Ω and V . the direction of the force on a ﬂuid element and the direction of motion.

and σ = ±1 depends on whether the spring is “soft” (σ = −1) or “hard” (σ = +1).) . respectively. By using this trial function in the time-averaged Hamilton’s Principle. 2 mω0 V (x) = 2 39 x4 x +σ 2 l0 2 . and ω is the nonlinearly shifted frequency.2. B. ﬁnd implicit relations giving approximate expressions for ω. but plot ω/ω0 and B vs. PROBLEMS with a nonlinear restoring force).8. The trial function is strictly appropriate only to the case A 1. where the constant ω0 is the angular frequency of oscillations having amplitude small compared with the characteristic length l0 . Consider the trial function x = l0 [A cos ωt + B cos 3ωt + C sin 3ωt] . where A. C are the nondimensionalized amplitudes of the fundamental and third harmonic. (You are encouraged to use Maple or Mathematica and/or MatLab in this problem. B and C as functions of A. A from 0 to 1 in the case of both a hard and a soft spring. Show that C ≡ 0.

LAGRANGIAN MECHANICS .40 CHAPTER 2.

. (3.}. This is a recursion relation.2 3. diﬀerence equation or iterated map z t+1 = f (z. 0. q . . t) . 1. t) . t∈R.1) where f is known as the vector ﬁeld and R is the set of real numbers.1 Introduction: Dynamical systems Mathematically. . −1. because they implic¨ itly contain second-order derivatives. .2) where f is known as the map (or mapping) and Z is the set of all integers {.1 Mechanics as a dynamical system Lagrangian method Lagrange’s equations do not form a dynamical system. The space in which the set of time-dependent variables z is deﬁned is called phase space. such systems exhibit long-time phenomena like attracting and repelling ﬁxed points and limit cycles. a continuous-time dynamical system is deﬁned to be a system of ﬁrst order diﬀerential equations ˙ z = f (z. . 2. However. Typically. (3. In this chapter we show how to reformulate nondissipative Lagrangian mechanics as a dynamical system. there is a standard way to 41 . or more complex structures such as strange attractors.2. 3. −2. but shall ﬁnd that it is a very special case where the above-mentioned phenomena cannot occur. . t∈Z.Chapter 3 Hamiltonian Mechanics 3. Sometimes we also talk about a discrete-time dynamical system.

u. (2.6). within the ndimensional linear vector space N on which H acts. there is an m-dimensional ¨ subspace M (the nullspace) such that H·¨ ≡ 0 for all q ∈ M. with rank n − m. expanding out the total derivative using the chain rule and moving all but the highest-order time derivatives to the right-hand side. qi . (3. We can solve q ¨ eq. ∂ qi ∂ qj ˙ ˙ (3. which is to double the size of the space of time-dependent variables by treating the generalized velocities u as independent of the generalized coordinates.1 The Lagrangian method per se does not break down if the Hessian is singular. so ˙ ˙ ¨ that the dynamical system is q = u. so ¨ that its inverse. t). Then.3) ¨ The matrix H acting on q . 2 . In the particular case of Lagrangian mechanics. and for our method to work we require it to be nonsingular. H−1 . exists and we can ﬁnd q . because the standard numerical integrators require the problem to be posed in terms of systems of ﬁrst-order diﬀerential equations.24). but the solubility condition provides m constraints that complete the determination of the dynamics.3) for the component of q lying in the complementary subspace N \ M provided the right-hand side satisﬁes the solubility condition that it have no ¨ component in M. HAMILTONIAN MECHANICS obtain a system of ﬁrst-order equations from a second-order system. (2.42 CHAPTER 3. whose elements are given by Hi. As a simple example.2. Then our dynamical system becomes ˙ q = u. We can still formally solve the dynamics in the following manner: Suppose H is singular. Then ∂L/∂ qi ≡ 0 and the ith component of the Lagrange ˙ ˙ equations of motion eq. ˙ u = H−1 · ∂2L ∂2L ∂L ˙ − − ·q ˙ ˙ ∂q ∂ q∂t ∂ q∂q . only our attempt to force it into the standard dynamical system framework. This trick is used very frequently in numerical problems. suppose L does not depend on one of the generalized velocities.j ≡ ∂2L . The component of q lying in M cannot be found directly. Then the phase space is of dimension 2n.4) is called the Hessian matrix . n j=1 ∂2L ∂L ∂2L qj = ¨ − − ∂ qi ∂ qj ˙ ˙ ∂qi ∂ qi ∂t ˙ n j=1 ∂2L q˙j . ∂ qi ∂qj ˙ (3. It is a kind of generalized mass tensor (see Sec. (3.5) Remark 3. we have from eq.24) reduces to the constraint ∂L/∂qi = 0. u = q (q. 3.

6) is that ∂L/∂ q occurs explicitly in Lagrange’s equations.8) do indeed form a dynamical system. but the extra freedom of choice may also bring advantages. . ˙ We shall always assume that eq.p. eq. yet the right-hand side of eq. but so far it looks rather unsatisfactory: now u is deﬁned only implicitly as a function of the phase-space variables q and p. t) . (3.6) where pi is called the generalized momentum canonically conjugate to qi . p. . ∂ qi ˙ (3.2 Hamiltonian method We can achieve our aim of ﬁnding 2n ﬁrst-order diﬀerential equations by using many choices of auxiliary variables other than u.8) ∂q q=u(q. n} deﬁned by pi ≡ ∂ ˙ L(q. q.4) we recognize this matrix as being ˙ ˙ the Hessian we encountered in the Lagrangian approach to constructing a dynamical system. in the standard Hamiltonian theory covered in this course it is always assumed to hold.t) ˙ Equations (3.2.g. (3.24). (2.e. . .7) Remark 3. Thus in either approach we require the Hessian to be nonsingular (i. (3.8) involves a partial derivative in which the q-dependence of u is ignored! We can ﬁx the latter problem by holding p ﬁxed in partial derivatives with respect to q (because it is an independent phase-space variable) but then subtracting a correction term to cancel the contribution coming from . ˙ The reason for deﬁning p as in eq. (3. These will be more complicated functions of the generalized velocities. q. MECHANICS AS A DYNAMICAL SYSTEM 43 3.2 We have in eﬀect changed variables from u to p. and such a change of variables can only be performed if the Jacobian matrix ∂pi /∂uj = ∂ 2 L/∂ qi ∂ qj is nonsingular. This condition is usually trivially satisﬁed. but there are physical problems (e. (3. Hamilton realised that it is very natural to use as the new auxiliary variables the set p = {pi |i = 1. t) ˙ p= . so we immediately get an equation of motion for p ˙ ∂L(q.3. From eq.6) can be solved to give q as a function of q and p ˙ q = u(q.7) and (3. if the Lagrangian does not depend on one of the generalized velocities) when this is not the case. (3. t) . In particular. for its determinant to be nonzero). However.2.

Applying the chain rule and then using eqs. t) ≡ p·u − L(q. t) − p·u] ∂q ∂H ≡ − . t) ˙ − p = ∂q = ∂L(q.11) ˙ = q.10) Given the importance of ∂H/∂q it is natural to investigate whether ∂H/∂p plays a signiﬁcant role as well. (3. u. t) − ∂q n i=1 n ∂L ∂ui ∂ui ∂q pi ∂ui ∂q i=1 ∂ = [L(q.9) and (3. (3.7). (3. ∂q (3. HAMILTONIAN MECHANICS the q-dependence of u.9) where we have deﬁned a new function to replace the Lagrangian.12) These equations are known as Hamilton’s equations of motion.10) we get ∂H ∂p n = u(q. Summarizing eqs. ∂H ∂p ∂H ˙ p = − . p. t) . u.6) and (3. u. t) ∂ui ∂p (3. u. namely the Hamiltonian H(q.11). u. (3.44 CHAPTER 3.6) and (3. p. We deﬁne canonical coordinates as phase-space coordinates such that the equations of motion can be expressed in the form of eq. (3.7) we get ∂L(q. where the vanishing of the expression in the square bracket and the identiﬁ˙ cation of u with q follows from eqs. . t) + i=1 pi − ∂ ∂ui L(q.12) and a canonical system as one for which canonical coordinates exist. Diﬀerentiating eq. (3. ∂q ˙ q = (3. As with Lagrange’s equations they express the dynamics of a system with an arbitrary number of degrees of freedom in terms of a single scalar function! Unlike the Lagrangian dynamical system. the phase-space variables are treated on a completely even footing—in Hamiltonian mechanics both the conﬁgurationspace variables q and momentum-space variables p are generalized coordinates.

kinetic plus potential. They are characterized by the use of scalar quantities rather than the vectors encountered in the direct use of Newton’s second law. especially in thermodynamics. r ˙ ∂r (3. from eq.2. y. MECHANICS AS A DYNAMICAL SYSTEM 45 Remark 3. Equation (3. The fact that the Hamiltonian is an important physical quantity. (3.14) is solved trivially to give q = u(p) where u(p) = p/m. Both the Lagrangian and Hamiltonian have the dimensions of energy. the Hamiltonian is equal to the total energy of the system.6) p→p≡ ∂L = m˙ . t) m 2m |p|2 = + V (r. (3.14) (3. Thus. t) so that the natural form of the Lagrangian is L = T − V = 1 m|˙ |2 − V (r. t) 2m = T +V . 3.3 The transition from the Lagrangian to the Hamiltonian in order to handle the changed meaning of partial derivatives after a change of variable is a special case of a technique known as a Legendre transformation. (3. r 2 Then. and both approaches can be called energy methods.3. is one of the appealing features of the Hamiltonian approach. t) .10) we have H = |p|2 |p|2 − − V (r. so q = {x. z} may be replaced by r = xex + yey + zez . It is encountered quite often in physics and physical chemistry. This has both the theoretical advantage of leading to very general formulations of mechanics and the practical beneﬁt of avoiding some vector manipulations when changing between coordinate systems (in fact. whereas the physical meaning of the Lagrangian is more obscure. .2.13) so that in this case the canonical momentum is the same as the ordinary ˙ kinematic momentum. Also assume that it moves under the inﬂuence of a scalar potential V (r. Lagrangian and Hamiltonian methods were developed before modern vector notation was invented).3 Example 1: Scalar potential Consider the Lagrangian for a particle in Cartesian coordinates.15) That is. from eq.

46 Harmonic Oscillator CHAPTER 3. the Hamiltonian remains equal to the total energy. Hence ˙ H = p (p − mω0 x) (p − mω0 x)2 2 1 − + ω0 x(p − mω0 x) − 2 mω0 x2 m 2m (p − mω0 x)2 1 2 = + 2 mω0 x2 2m = T +V . about which we will have more to say later. ˙ (3. (3. HAMILTONIAN MECHANICS An example is the harmonic oscillator Hamiltonian corresponding to the Lagrangian. Solving eq. even though the generalized coordinate x remains the same. (The functional form of the Hamiltonian changes under the gauge transformation however. (2.) .75) 2 L = 1 m(x2 + 2ω0 xx − ω0 x2 ) .17) Gauge-transformed Harmonic Oscillator Now consider the gauge-transformed harmonic oscillator Lagrangian eq. (3. because the meaning of p changes. even though L was not of the natural form T − V in this case.12) the Hamiltonian equations of motion are x = ˙ p m p = −mω0 x . (3.16) H= 2m 2 From eq.18) for x we ﬁnd u(p) = (p − mω0 x)/m. (3. (2. ˙ ˙ 2 The canonical momentum is thus p = ∂L /∂ x = m(x + ω0 x) ˙ ˙ (3.19) Thus. This is an example of a canonical transformation. eq.61) mω0 x2 p2 + . thus conﬁrming that it is a quantity with a more direct physical signifance than the Lagrangian.18) and we see that the gauge transformation has eﬀected a transformation of the canonical momentum.

This corresponds to the case where the pendulum is just balanced “upside down”. is a ﬁxed point. Using eq. For |θ| 1 we may expand the cosine so V ≈ mglθ2 /2 and the system is approximately a harmonic oscillator.e. The dividing line H = 2mgl between the two topological classes of orbit is called the separatrix.2 0.1. 3. (3. One class is the rotating orbits for which the pendulum has enough enFigure 3. The other class is the librating orbits for H < 2mgl implying that the pendulum is trapped in the gravitational potential well and oscillates like a pair of scales (hence the name). 3. The equilibrium point θ = 0 or 2π. By conservation of energy (see also Sec.6 0. 3. MECHANICS AS A DYNAMICAL SYSTEM 47 3. H > 2mgl. 2.1 (which is in -1 units such that m = l = g = 1).1. (2. labelled X in Fig. so θ increases or decreases secularly (though the physical position x = l sin θ does not).4. Almost all orbits in the neighbourhood of an X point eventually are repelled far away from it.4 0. becomes ˙ θ H(θ. 3. H = pθ θ − L.8 1 complicated than in the case of the harmonic θ/2π oscillator since there are two topologically distinct classes of orbit.6) we get pθ = ml2 θ. Thus ˙ = pθ /ml2 and so the Hamiltonian.4 Example 2: Physical pendulum A nonlinear one-dimensional case is provided by the physical pendulum. the ﬁxed point is stable or elliptic). and on the separatrix lies another ﬁxed point. 2ml2 (3. The orbits in its neighbourhood. like those of the harmonic oscillator.3. labelled O in Fig.3). to swing entirely over the top ical pendulum.2.2. 3 the Hamiltonian H = T + V is a constant 2 of the motion so the nature of the orbits in 1 phase space can be found simply by plotting pθ 0 O X O the contours of H as in Fig. remain in the neighbourhood for all time (i.1.20) which again is of the form T + V . in˙ troduced in Sec.1: Phase space of the physergy. and thus it is referred to as an unstable or hyperbolic ﬁxed point.27) in eq. We see that -2 the structure of the phase space is rather more -3 0 0. . pθ ) = p2 θ + mgl(1 − cos θ) .

t) 2m = T +V .2.j=1 k=1 n 1 2 i. (2. possibly with constraints (which we here assume to be independent of time). (3.10) we have H = (p − eA)·p |p − eA|2 e(p − eA)·A − + − eΦ(r.3.44) and thus. by eq. t) . kinetic plus electrostatic potential energy.j (q) ≡ k=1 mk ∂xk ∂xk · ∂qi ∂qj (3. 3.6 Example 4: The generalized N -body system Let us now revisit the general case which led to the original deﬁnition of the Lagrangian in Sec.5 Example 3: Motion in e.j=1 m k qi ˙ ∂xk ∂xk · qj − V (q) ˙ ∂qi ∂qj (3. t) m 2m m |p − eA(r. The Lagrangian is given by eq.7) (assuming the function xk (q) to be independent of time) the natural form of the Lagrangian is L = T −V n N = = 1 2 i.22) Thus we ﬁnd again that.6).2. p≡ ∂L = m˙ + eA(r. HAMILTONIAN MECHANICS 3. Then. r ˙ ∂r (3.48 CHAPTER 3. (3. (3. the Hamiltonian is still the total energy.j qj − V (q) . so that the number of generalized coordinates is n ≤ 3N .m.25) and eq. from eq.24) . ˙ ˙ where the symmetric mass matrix N µi. (2. although the Lagrangian cannot be put into the natural form T − V (r). the case of N particles interacting via a scalar N -body potential. 2. from eq.21) Thus u(p) = p/m − eA/m and. t)|2 = + eΦ(r.23) qi µi. (2. potentials Now consider the case of a charged particle in an electromagnetic ﬁeld with magnetic vector potential A and electrostatic potential Φ.

1(c) and 4.j i. using eq. (2. expressed in plane polar coordinates in Sec.10) we have n n H = = pi µ−1 pj − i.j µ−1 pj − V (q) i. ˙ (3.2.j + V (q) (3. t).29) .25) for qi to ˙ i. ˙ L = 1 m r2 + r2 θ2 − V (r) . we see that it is the Hessian k matrix for this system. from eq. (3.2. (3. from eq. the Hamiltonian is again equal to the total energy of the system. 2.1. p.4). (3.33). µi. two-dimensional example of a problem in non-Cartesion coordinates we return to the problem of motion in a central potential. This result does not hold in the case of a time-dependent representation.25) Assuming none of the particles is massless.j =1 pi µ−1 µi .3.7.23) in eq.27) p2 p2 r H= + θ 2 + V (r) .j give n ui (q. [See Problems 3. ˙ 2 Comparing with eq.j (q)qj . (3.26) Then.4. so its inverse µ−1 exists and we can formally solve eq. (3. MECHANICS AS A DYNAMICAL SYSTEM 49 is the metric tensor for a conﬁguration-space mass-weighted “length” element ds deﬁned by ( ds)2 ≡ k mk dl2 .28) and thus can be inverted simply taking the reciprocal of the diagonal elements. (3.j=1 n 1 pi µ−1 pj i.j (3. Recapitulating eq.6) n pi = j=1 µi. t) = j=1 µ−1 pj . Hence.] Particle in a central potential As a simple.i j . 2m 2mr (3. From eq.23) we see that the mass matrix is diagonal µ= m 0 0 mr2 . (3.j is a positive-deﬁnite matrix. xk (q.i .27) = T +V . Then. i.j=1 1 2 i.5.j. Thus. (3.j 2 i.

if H does not depend explicitly on time. then H = H(p. 3.) In the Hamiltonian case this means that H has no explicit time dependence. t) .50 CHAPTER 3. (3. following the phase-space trajectory. q. HAMILTONIAN MECHANICS 3.12). then its total time derivative also vanishes.10) and eq.1)] depends only on the phase-space coordinates and has no explicit dependence on t. in an autonomous Hamiltonian system the Hamiltonian is an integral of motion. When this is the case we may interpret the completely general result above as a statement of conservation of energy. t) by q in the term p·u and thus deﬁne the phase-space Lagrangian ˙ ˙ Lph (q. whereas “autonomous” means independent or self-governing. t) ≡ p·q − H(q. Using the Hamiltonian equations of motion. if there is an external time-varying perturbation. To ﬁnd a suitable deﬁnition for the action integral. (2. (3.10) to get L on the left-hand side. comparing eq. q. ∂t = (3. dH dt ∂H ∂H ∂H ˙ ˙ + q· + p· ∂t ∂q ∂p ∂H ∂H ∂H ∂H ∂H = + · − · ∂t ∂p ∂q ∂q ∂p ∂H ≡ . (3. p.30) Thus. we ﬁrst rearrange ˙ eq. t). We have seen above that H can often be identiﬁed as the total energy. That is. (3.84) we recognize H as the negative of the energy integral I for autonomous systems as predicted by Noether’s theorem. H ≡ dH/ dt. p. Then we replace u(q. eq. so its partial time derivative vanishes.31) . Conversely. (3.3 Time-Dependent and Autonomous Hamiltonian systems An autonomous dynamical system is one in which the vector ﬁeld f [see eq. p.4 Hamilton’s Principle in phase space Let us express the action integral S in terms of q and p and show that we may derive Hamilton’s equations of motion by making S stationary under variations of the trajectory in phase space. rather than in conﬁguration space. More generally. (The reason for the name is that explicit time dependence would come from external forcing of the system. ˙ Consider the time rate of change of the Hamiltonian.

(3.32) We shall now show that requiring δSph = 0 for arbitrary variations about a given path implies that that path is such that Hamilton’s equations of motion are satisﬁed at all points along it. this variational principle is subtly diﬀerent from the Lagrangian version of Hamilton’s principle. (3. ˙ δq ∂q dt ∂ q ∂q δLph ∂Lph ∂H ˙ = = q− =0. one can easily construct a counter example to show that this is not the case: consider a variation of the path in which we can vary the direction of its tangent vector. This is in contrast with the the Lagrangian version of Hamilton’s Principle. p. while keeping this ˙ point ﬁxed (see Sec.14). δp ∂p ∂p (3. at some point (q.12). and is sometimes called the modifed Hamilton’s Principle. although we ﬁx the endpoints in the modiﬁed Hamilton’s Principle. (1. as claimed.3. Then q changes. This is because the ﬁxing of the initial phase-space point speciﬁes 2n initial conditions. (2. where we ﬁxed only the n conﬁguration-space coordinates q.e. t) . t). with p now being regarded as an independent coordinate ∂Lph d ∂Lph ∂H δLph ˙ = − = − −p=0. 4. the one that obeys the equations of motion). We know from variational calculus that Sph is stationary under arbitrary variations of the phase-space path (with endpoints ﬁxed) if and only if the Euler–Lagrange equations δLph /δq = δLph /δp = 0 are satisﬁed at all points on the path. Remark 3. where the variational derivatives are as deﬁned in eq. t) were identically satisﬁed.4. we will not be able to satisfy Hamilton’s equations of motion at all points on the path unless the endpoints are both chosen to lie on a physical path.2 for an illustration of this). p. p] = t1 ˙ dt Lph (q. Since we concluded above that Lph and L were not the same on arbitrary nonphysical paths. p). then Lph would simply be L expressed in phase-space coordinates. q. which is generally exactly the right number to allow . Replacing L by Lph in eq.33) These are identical to Hamilton’s equations of motion.4 Note that. Thus Lph and L are the same value only on the subset of ˙ paths (which includes the physical paths) for which p·q = p·u(q. which completely determines the physical path (i. The problem of ﬁnding the physical path between two given conﬁguration-space points thus involves specifying only n initial and n ﬁnal data. even on arbitrarily varied phasespace paths. eq. HAMILTON’S PRINCIPLE IN PHASE SPACE 51 ˙ If q = u(q. but u remains the same. However. p.26) we deﬁne the phase-space action integral t2 Sph [q.

.52 CHAPTER 3. HAMILTONIAN MECHANICS us to ﬁnd one (or more) solutions of the equations of motion passing through arbitrarily chosen initial and ﬁnal points.

5 3.6. Take the generalized coordinates of the system to be q = {x. y.8. y)ez . ˙ ˙ L = 1 R2 θ2 + cos2 θ (Ω + φ)2 − V (θ. showing the direction of travel. 3. (b) Use this Lagrangian to construct the Hamiltonian for the system. 2 Is it of the form T + V ? 3. Write down a Lagrangian for the problem of two particles of mass m1 and m2 connected by a light rigid rod of length l in a gravitational ﬁeld g.2 Anharmonic oscillator phase space Find the Hamiltonian corresponding to the Lagrangian of the anharmonic oscillator in Problem 2. θ. p) in the soft potential 2 2 2 2 case for (a) H < mω0 l0 /8 and (b) H > mω0 l0 /8.5.1 Problems Constraints and moving coordinates (a) This revisits Problem 1. inﬁnitely long magnetic conﬁnement system with vector potential A = ψ(x.2. Sketch contours of H(x.3 2-D motion in a magnetic ﬁeld Consider the motion of a particle of charge e and mass m in a straight. Try to include representatives of each qualitatively distinct orbit.1. φ) .8. where α1 ≡ m2 /(m1 + m2 ) and α2 ≡ m1 /(m1 + m2 ) [so that (x. . Indicate representative trajectories of the system in phase-space. Is it of the form T + V ? (c) Find the Hamiltonian corresponding to the Coriolis Lagrangian in Problem 2. with the coordinates of the two particles being given by x1 y1 z1 x2 y2 z2 = = = = = = x − α1 l sin θ cos φ y − α1 l sin θ sin φ z − α1 l cos θ x + α2 l sin θ cos φ y + α2 l sin θ sin φ z + α2 l cos θ . y. z) is the centre of mass].5.5. z.5.1. PROBLEMS 53 3. φ}.3.

and ∗ means complex conjugate. HAMILTONIAN MECHANICS (a) Show that Bx = ∂ψ/∂y. (b) Show that the Hamiltonian is p2 p2 (pz − eψ)2 y x H= + + . Show that the equations of motion for ζ and pζ ≡ px + ipy are ˙ ζ = pζ ˙ 1 pζ . (c) It can be shown that in any region with no electric current. y) can be represented as the real part of an analytic function of the complex variable ζ ≡ x + iy. m where the prime on Ψ means derivative with respect to its argument. ψ(x. 2m 2m 2m write down the Hamiltonian equations of motion and give two integrals of the motion. By = −∂ψ/∂x. Bz = 0 and thus that contours of ψ(x. . ψ(x.54 CHAPTER 3. m e = [pz − eRe Ψ(ζ)] [Ψ (ζ)]∗ . y) = Re Ψ(ζ). y) deﬁne magnetic ﬁeld lines.

if we go from coordinates x.1) Thus one obvious class of canonical transformation is produced by a change of conﬁguration-space coordinates. the canonical transformation is x = r cos θ y = r sin θ pθ sin θ ˙ px = mx = m(r cos θ − rθ sin θ) = pr cos θ − ˙ ˙ r pθ cos θ ˙ py = my = m(r sin θ + rθ cos θ) = pr sin θ + ˙ ˙ r K(r. px . or point transformation. stationary generalized coordinate system. t).12). Q.69) we ﬁnd the general relationship Pi = ∂ ˙ ˙ L(g(Q. (4. (2. q. (3.25) in the two coordinate systems.6 shows the general form of p for an N -particle system in an arbitrary. then the transformation of p to P follows from comparing eq. (3. t) ˙i ∂Q 55 . with a new Hamiltonian K(Q. t).6. P . pθ ) = H(x. θ. eq. Q.1 Introduction A canonical transformation is a transformation of the phase-space coordinates q and p to new coordinates Q and P that obey Hamilton’s equations of motion. In this case. y. t) as discussed in the context of the general Lagrangian formalism in Sec. q = g(Q.2. y in a twodimensional system to polar coordinates r. pr . g(Q. if we change from one conﬁguration-space coordinate system. The discussion in Sec. to another. (2. Applying eq. t).68) and using eq.Chapter 4 Canonical transformations 4. For instance. θ. 3. 2. py ) .

6. ˙ Lph ≡ p·q − H(q. Although it clearly has no eﬀect on Hamilton’s equations of motion. we can ﬁnd a physically equivalent Lagrangian (1) by adding to Lph the total time derivative of an arbitrary function. p) for time-independent transformations as in eq. p.4) However.3 where we compared the Hamiltonian for the harmonic oscillator derived from the natural form for the Lagrangian with that from a gauge-transformed version. (4. The discussion of the time dependent case will be left to the next section.1). ∂Qi j=1 (4.j ≡ ∂gj /∂Qi . p] = t1 ˙ ˙ dt Lph (q. p. (4.2. 2. in the case of the particle in an electromagnetic ﬁeld treated in Sec. q. (1) (4. CANONICAL TRANSFORMATIONS = p· ˙ ∂g ˙ ∂ Qi n ∂gj = pj . and thus does not generate a canonical transformation.2) Thus by inverting the matrix Ti. This suggests that gauge transformations may hold the clue to constructing a general theory of canonical transformations and we show in the next section that this is indeed the case. t2 Sph [q. gauge-transformed phase-space Lagrangian ˙ Lph ≡ −q·p − H(q. M (q. This gives the alternative. t) .3) which we showed to be stationary on physical paths with the form of phasespace Lagrangian given in eq. 4.31). We have also encountered another special case of a canonical transformation (for which p was changed but q remained the same) in Sec. we shall have use of the particular choice M = −q·p. t) . (3.5) . p. (2) (4.2. P ) = H(q. We can also show by direct calculation that K(Q. t) . we complete the construction of the canonical transformation by ﬁnding p = T−1 ·P . Likewise.56 CHAPTER 4. t) say. (3.2 Generating functions To discuss gauge transformations in the Hamiltonian context we use the modiﬁed Hamilton’s Principle based on the phase-space action integral eq.32). p.2.5 an electromagnetic gauge transformation eﬀects a change in the momenta but leaves the conﬁguration-space coordinates unaﬀected. 3. 3. p. as in Sec.

t) = Lph (Q. This condition rules out. P . t) + . (4. we ensure simultaneous stationarity of the two functionals by requiring that the old and new phasespace Lagrangians diﬀer only by the total time derivative of some function. P . We require the two sets of coordinates used to be functionally independent.6) that is stationary under variations vanishing at the endpoints for the same paths (represented in the new coordinates) as Sph is. t). t). P . where Lph is a phasespace Lagrangian in either of the forms given in eqs. For example. Otherwise the choice of generating function is somewhat arbitrary and we choose whichever is convenient for the problem. F3 (p. q.4. so that F really is a function of 2n independent phase-space coordinates. P ] = t1 ˙ ˙ dt Lph (Q. P ) where the mapping between the two coordinates is invertible and suﬃciently smooth that we can evaluate the required derivatives. F2 (q. p. not only the physical ones. the use of the type 1 generating function for point transformations.9) These functions are known as the generating functions of types 1 to 4. t). (4. Q. ˙ ˙ Lph (Q.7) As for the gauge transformation discussed above. p) → (Q. p. P . Q. F4 (p. P and H replaced by K.10) (4.4) or (4.8) This equality is to hold on all phase-space paths. t) . there must also exist an action integral Sph of the form t2 Sph [Q. The total time derivatives in the four cases are given by dF1 ∂F1 ˙ ∂F1 + ∂F1 ˙ = q· + Q· dt ∂q ∂Q ∂t dF2 ∂F2 ∂F2 ∂F2 ˙ ˙ = q· + P· + dt ∂q ∂P ∂t (4. t) (4. giving four possibilities F1 (q. which we denote by F .11) . P . Q. P . P . P . Do we take F to be a function of the old or of the new variables? The traditional answer is to take F to be a mixed function of n of the old and n of the new (since there are only 2n independent variables). Since we are seeking a canonical transformation. including the identity transformation. p replaced by Q.5) with q. dt (4. for example. dF ˙ ˙ ˙ ˙ Lph (q. Q. GENERATING FUNCTIONS 57 Now assume we have a phase-space transformation (q. (1) (4. Q. t) ≡ P ·Q − K(Q. t) .2.

P ) ˙ + F4 (p. P .13) The four types of generating function correspond to choosing the two alternative forms. p) = (2) ˙ Lph (Q.12) (4. (4.4) and (4. p. p1 z(t2) z(t) z(t1) q1 q2 . ˙ where Fi denotes dFi / dt and we have suppressed explicit time dependences for brevity.58 CHAPTER 4. p(t)) in phase space. (2) ˙ Lph (q. eq. p. (4. as indicated in the following table: (1) (1) ˙ ˙ Lph (q.8). since both the Lagrangians . ˙ each passing through the same point at time t but with diﬀerent z. p) = (1) ˙ P ) + F1 (q. CANONICAL TRANSFORMATIONS ∂F3 dF3 ˙ ∂F3 + ∂F3 ˙ = p· + Q· dt ∂p ∂Q ∂t ∂F4 dF4 ˙ ∂F4 + ∂F4 .5). p) = Lph (q. ph (4.. P ) . ˙ + F3 (p. eqs. (4. on either side of the gauge transformation. As indicated in eq. P . q. L (2) (Q. pn Figure 4.14) (2) ˙ Lph (q.1: Some possible paths z(t) ≡ (q(t). Type 1 generating functions combine most (1) (1) naturally with the choice Lph = Lph . P ) .14). P ) + F2 (q. Q) . p) = (1) ˙ P) Lph (Q.. Lph = Lph . q. Q. ˙ = p· + P· dt ∂p ∂P ∂t (4. Q. ˙ ˙ ˙ Lph (Q. Q) .

4. p.18) ∂t These equations deﬁne. Since we ˙ require q and Q to be functionally independent. 4.15) Since eq.20) (4. and thus eq. As ˙ ˙ illustrated in Fig. t) + ∂q ∂t ∂F1 ˙ P+ ·Q − K(Q. a canonical transformation. Substituting eq. (4. ∂Q (4.16) (4. (4. (1) (2) For type 2 generating functions we choose Lph = Lph . it must in particular hold at any given phase-space point (q. p.8) with these choices and collecting terms in q and Q we get p− = ∂F1 ∂F1 ˙ ·q − H(q.21) (4. (4. (4.8) is to hold on all paths. albeit implicitly. but not p or P . (4.22) .8) becomes K=H+ ∂F2 ∂F2 ˙ ·q − H(q. t) .19) ˙ ˙ Equating the coeﬃcients of q and P to zero we get the implicit equations deﬁning type 2 canonical transformations ∂F2 ∂q ∂F2 Q = ∂P p = and K=H+ ∂F2 . ∂Q p = ﬁnally leaving ∂F1 .1 the possible total time derivatives (q. p) span the full 2n-dimensional space of tangent directions through the given point. t) . ∂P p− (4. the possible choices of q and ˙ also span a 2n-dimensional space. P .15) can hold identically Q ˙ ˙ only if the coeﬃcients of q and Q vanish separately ∂F1 ∂q ∂F1 P = − . p) for all possible paths passing through this point.17) (4. GENERATING FUNCTIONS 59 ˙ ˙ ˙ ˙ ˙ and F1 involve only q and Q. Lph = Lph so that eq.2. t) + ∂q ∂t ∂F2 ˙ =− Q− ·P − K(Q. P . ∂t (4.10) in ˙ ˙ eq. (4.

P p = (2mω0 P )1/2 cos φ . Lph = Lph and ﬁnd ∂F3 ∂p ∂F3 P = − ∂Q q = − (4. (2. to ﬁnd ∂F4 ∂p ∂F4 Q = ∂P q = − and K=H+ ∂F4 .23) (4. 2 (4. H = p2 /2m + mω0 x2 /2 and try the type 1 generating function F1 = 1 mω0 x2 cot φ . ∂t (4. p and φ.28) 4.29) 2 with x playing the role of the old generalized coordinate. Using eq.61) and Hamiltonian eq. ∂t (2) (4.30) (4.29) in eqs.26) (4. Q.) 2 Now consider the Hamiltonian.16).2.24) (2) (1) and K=H+ (2) ∂F3 . (3. (Introducing is a useful device for the formal asymptotic theory. q. ω0 = ω0 ( t) = 0.25) For type 4 we choose Lph = Lph .27) (4. (4.31) Eliminating x between these equations we ﬁnd the canonical transformation between x.17) we have p = mω0 x cot φ P = 1 mω0 x2 cosec2 φ . Lph = Lph . CANONICAL TRANSFORMATIONS For type 3 we choose Lph = Lph .1 Example 1: Adiabatic Oscillator Consider the harmonic oscillator Lagrangian eq.32) . where is small parameter. and φ playing the role of the new generalized coordinate. in which we consider the scaling of error terms as → 0.60 CHAPTER 4. taking ω0 to be a slowly varying function of time which never vanishes. (4. (4.16) and (4. (4.

t ln P (t) − ln P0 ≈ t0 dt d ln ω0 cos 2φ . Thus we should be able to show its adiabatic invariance from the Hamiltonian equations of motion. with no secular increase. Integrating the equation to get the phase as a function of time we have t φ(t) ≈ t0 ω0 ( t) .34) ˙ ∂K = ω0 + 1 d ln ω0 sin 2φ φ= ∂P 2 dt and (4. we see that the integrand averages to zero over times long compared with 1/ω0 and thus the integral will also be oscillatory.33) with eq. Comparing with eq.37) where t0 is an arbitrary starting time. which from eq. Thus to leading order we may drop the sin 2φ term in eq. Observe that d ln ω0 / dt = ω0 ( t)/ω0 is small. so that we are making the same approximation. Functions with this property are known as secular terms.35) d ln ω0 ∂K ˙ P =− = P cos 2φ . (4.38) Using eq. First we need the new Hamiltonian. (4.18) is given by 2 p2 mω0 x2 mω0 x2 cot φ ˙ K = + + 2m 2 2 1 d ln ω0 = ω0 P + P sin 2φ . (4. (4. O( ).36) is O( ) so we see that P is approximately constant over short times. (2. Thus the canonical transformation method conﬁrms the adiabatic invariance of the action. . 2 dt The equations of motion are thus (4.4. (2.62) we identify P = mω0 A2 /2. (4. However we can say more than that since we can estimate the long-term change by integrating the equation.36) ∂φ dt So far the results are exact. ω ≡ φ ≈ ω0 .37). dt (4. as we were led to in the averaged Lagrangian approach.2.33) Comparing eq. the action. Thus the phase increases with time without bound. (4. (4.35) because it is small in comparison ˙ with ω0 .65) we see that P is in fact J. but to proceed further (unless we solve these equations numerically) we need to make use of the expansion parameter . (4. GENERATING FUNCTIONS x= 2P mω0 1/2 61 sin φ . The right-hand side of eq.

2) with the roles of the old and new variables interchanged.2. whereas the type 1 class does not [though a type 1 transformation may come arbitrarily close to the identity—see Problem 4. However. E ≡ T + V . P . (4.2 Example 2: Point transformations F2 (q.7. (4. .2(c) and consider the limit ∆t → 0]. Note that. 2. substituting eq. (4. then E = N ω0 (ignoring the zero-point energy 2 ω0 ). t) The second equation shows that eq.20) and (4.5.62 CHAPTER 4.39) is the generating function for a point transformation. (4. t) + ·P . because of the possibility of asymptotically small corrections “beyond all orders”). t) = G(q. (4. Thus the adiabatic invariance of P = N can be interpreted as saying the number of quanta is conserved during adiabatic changes in the system.40) Q = G(q. and the ﬁrst equation is the same as eq. P . The quantum mechanical interpretation of this was pointed out around 1911 by Einstein: If there are N >> 1 quanta of energy in the os1 cillator.22) gives ∂G(q. take where G(q. t) is a function such that the matrix ∂Gj /∂qi is nonsingular. (4. p. (4.21) give n pi = j=1 ∂Gj Pj . CANONICAL TRANSFORMATIONS Remark 4. (4. K is still approximately the total energy of the oscillator. Thus. ∂qi (4. apart from the small correction term ∂F1 /∂t. t) K(Q. in the case of a time-dependent transformation.1 As implied by the remark at the end of Sec. the class of canonical transformations that can be generated by a type 2 generating function includes the identity. the generating function method now provides us with the transformation of the Hamiltonian. Then eqs.34) gives P ≈ E/ω0 . we could go on to construct a formal asymptotic procedure in which the canonical transformation was successively reﬁned to remove the oscillatory terms in P at each order in to produce an adiabatic invariant that is conserved to arbitrarily high order (but not exactly. so eq. That is. t)·P .39) in eq. t) = H(q. 4.39) As an example of a type 2 generating function.41) ∂t Note that included among the point transformations is the identity transformation. a useful result that we did not have easily before.

we consider the type 2 generating functions F2 (q.20) and (4. t) ∂H ∂σ(q. P .3.43) Due to the nature of the generating function method these equations are mixed between the perturbed and unperturbed canonical variables.21) we have Q= ∂σ(q.45) ∂t This is again a mixed expression in the perturbed and unperturbed variables so we use eq. (4. ) = p − + O( 2 ) . P .44) . t) + + O( 2 ) .20) and (4. P . P .21) that F2 = q·P gives p = ∂F2 /∂q = P and Q = ∂F2 /∂P = q. P . (4. p. (4. continuous symmetry operations and canonical perturbation theory. p.44) to get an explicit expression K(Q.46) ∂t (4. t) ∂H · − · ∂q ∂p ∂p ∂q ∂σ(q. ) = H(q. t) P (q. q·P generates the identity transformation. t) p= = P+ + O( 2 ) . t) + K(q.4. ) = H(q. INFINITESIMAL CANONICAL TRANSFORMATIONS 63 4. p.3 Inﬁnitesimal canonical transformations The class of canonical transformations continuously connected to the identity transformation is important for several reasons—it includes time evolution. Since type 1 transformations do not include the identity. That is.22). t. p. The generating function of such a family can be written in the form F2 = q·P + σ(q. ∂q Q(q. t) + O( 2 ) . t) ∂F2 = q+ + O( 2 ) . ∂p ∂σ(q. p. p. From eqs.42) where the notation O( 2 ) means a term scaling like 2 in the limit → 0. (4. ∂σ(q. p. t) + O( 2 ) .42) in eq. t) and note from eqs. ∂P ∂P ∂F2 ∂σ(q. t) + O( 2 ) . p. up to ﬁrst order in . inserting eq. (4. (4. such that = 0 is the identity. and thus negligibly small compared with the O( ) term. ) = q + Also. The term σ is called the inﬁnitesimal generator of the family of transformations. t) + ∂σ(q. t. p. t. We now move continuously away from the identity among a family of canonical transformations parameterized by . p. (4. ∂q ∂q (4. but they can easily be solved by iteration to give. say. (4. p. t. ∂σ(q.

we can integrate it to obtain a statement valid over ﬁnite time intervals. we have shown that the Hamiltonian is the inﬁnitesimal generator for dynamical evolution in time. periodically forced system. (This is e most useful when n = 1.s (q s .1 Time evolution Normally is a parameter that we imagine varies at ﬁxed time. that ˙ Q(q. but it is also of great interest to consider the case where the parameter is time itself.12). (4.) For this reason systems where the Hamiltonian has an explicit periodicity in time are sometimes called (2n + 1)-dimensional Hamiltonian systems. ps ) at time s to the corresponding point on the physical path at time t: ϕt. δt) = q(t) + q(t)δt + O((δt)2 ) = q(t + δt) + O((δt)2 ) . ˙ ˙ with q = ∂H/∂p and p = −∂H/∂q. In this case the map is called a period map. we may think of the period map as giving the intersections of orbits with the Poincar´ surface of section. pn ). t. ϕtn+1 . t) = H(q.tn (q n . Another example is the construction of a numerical integrator of the ODEs making up Hamilton’s equations of motion. (3. ps ) evolving any phase space point (q s . which evolves a phase-space point at time tn = n∆t to the corresponding point at time tn+1 = (n + 1)∆t as if the state of the system were captured periodically in the ﬂashes of a stroboscope. Since the result is true at all times t.47) . eq. ˙ P (q. make the space topologically toroidal).e. p.3. p. which is the function ϕt. p(t)). ps ) = (q(t). Of necessity we replace the inﬁnite number of points in time of the exact problem with a ﬁnite number of points in a numerical approximation. t. we consider transformations at times t + δt where t is ﬁxed but the time variation comes from taking = δt. That is. Then we can state the integrated result in the following form: Hamiltonian ﬂows are canonical transformations. for then the surface of section is two-dimensional and easy to visualize. The main use of of such a stroboscopic map is the analysis of the dynamics of a nonautonomous. p. A commonly encountered example of such a ﬂow is what we shall call a stroboscopic map. in which one takes ∆t to be the period of the forcing.s (q s .44) with Hamilton’s equations of motion. p. and evaluate t modulo ∆t (i. δt) = p(t) + p(t)δt + O((δt)2 ) = p(t + δt) + O((δt)2 ) (4. Now consider the choice σ(q. CANONICAL TRANSFORMATIONS 4. Then we see by comparing eq.64 CHAPTER 4. t). Although we do not know the ﬂow exactly (otherwise there would be no point in integrating the equations . If we visualize the motion in the (2n + 1)-dimensional space formed by adding the time dimension to phase space. First we introduce some dynamical systems jargon by deﬁning the Hamiltonian ﬂow . That is. t = 0.

g} = −{g. p) is the 2n-dimensional phase-space position deﬁned by concatenating q and p. eq. w}} + {v. bop ] with i {a. {u. v}} ≡ 0 . can be written in the symmetric fashion ˙ q = {q. A property shared with the commutator is the Jacobi identity {u.) In this case the numerical algorithm is called a symplectic integrator.4. We follow the sign convention used by Goldstein (see Sec. ∂q ∂p ∂p ∂q (4. b}. (4. then the classical limit may be obtained by replacing [aop .4 Poisson brackets We deﬁne the canonical Poisson bracket {f.2(c).49) The utility of the Poisson bracket notation lies in the fact that Hamilton’s equations of motion. so that {f. (4. {w. 6) whereas Scheck deﬁnes the Poisson bracket with the opposite sign. (The full signiﬁcance of the fact that a map is canonical will become apparent when we have established some more properties of canonical transformations.48) Note that the bracket is an antisymmetric bilinear form in f and g: {f. H} . g} ≡ ∂f ∂g ∂f ∂g · − · . (3. u}} + {w. 1 .4. We observe that this algebraic property is analogous to the properties of the commutator of two operators. Indeed there is a deep physical connection between the commutator and the Poisson bracket through the quantummechanical correspondence principle: if aop and bop are quantum-mechanical observables. A and B say: [A. g} between any two phase-space functions f and g by1 {f. f } ≡ 0. f }. Do not confuse these curly brackets with the similar notation we have been using for sets of variables. B] ≡ AB − BA.12). {v. which allows the more compact notation ˙ z = {z. where a and b are the classical counterparts of the respective quantum-mechanical operators.7.50) where z ≡ (q.51) (4. POISSON BRACKETS 65 numerically) it is sometimes very important to require the approximate stroboscopic map to be exactly canonical. H} . H} ˙ p = {p. An example is the discrete-time dynamical system derived in Problem 4. 4.

t) + {σ(z.44) and (4. H} + . ) = z + {z. Note that. the angle measured about the z-axis. CANONICAL TRANSFORMATIONS We can generalize this result to the time derivative of any function. take σ = pφ . t). (4.) 4. the inﬁnitesimal generator of a symmetry of the Hamiltonian is an integral of the motion. (3.55) (4. That is.53) which agrees with eq. Then eq. since the absence of φ from the Hamiltonian immediately gives pφ = 0. (Of course the result is obvious if we use φ as a generalized coordinate. t)} + O( 2 ) . t. (4. t) ∂t + O( 2 ) . (4. in an autonomous system.4. (4. (4. f is an integral of the motion if it does not depend explicitly on time and its Poisson bracket with the Hamiltonian vanishes. σ(z. df ∂f ∂f ∂f ˙ ≡ z· + = {f. where the case = 0 is exactly solvable and H1 . t.66 CHAPTER 4.52) shows that this implies that σ is an integral of the motion.46) in the more compact forms Z(z. and K(z. where Z ≡ (Q. We can also write eqs.1 Symmetries and integrals of motion In this section we consider only autonomous systems. the momentum canonically conjugate to q = φ.55) the Hamiltonian is invariant under the canonical transformations generated by σ if and only if its Poisson bracket with σ(z) vanishes. if we take f = H we immediately get dH ∂H = . ˙ but we now do not need to use such coordinates—all we need to do is evaluate the Poisson bracket of H with pφ . dt ∂t (4. This is a Hamiltonian form of Noether’s theorem. Thus we see that pφ is the inﬁnitesimal generator for rotations about the z-axis.52) dt ∂z ∂t ∂t For example. Then the above result shows that pφ is an integral of the motion.30). If the system is rotationally symmetric then the Hamiltonian is invariant under increments in φ.54) 4. (4. H(z. ) = H(z.4. For example. t).2 Perturbation theory Often we are given a system for which the Hamiltonian may be resolved into the form H0 + H1 . From eq. P ). But eq. t)} + ∂σ(z.44) implies that φ is mapped to φ + . f (z.

A more modern approach. (4.4. . is known as the Lie transform. pi . and put the small gravitational interactions between the planets into H1 . but they all make use of the idea that we can represent the Poisson bracket with an inﬁnitesimal generator W as an operator LW such that its action on any phase-space function f is given by LW f ≡ {f. This is known as a problem in celestial mechanics.5 Action-Angle Variables It is clear from the equations of motion that if the Hamiltonian does not depend on one of the generalized coordinates qi then the canonical momentum conjugate to it. (We can as before refer to qi as ignorable. amenable to computerized algebraic manipulation.5. This is an alternative to the traditional generating function method that can be motivated by noting that time evolution generates a canonical transformation continously connected to the identity. to generate a canonical transformation that is an explicit function of only the old variables. and shown to give rise to Kepler’s Laws. then we can use a pseudo-Hamiltonian. though the terminology cyclic coordinate is also used in this context. the canonical transformation is eﬀected by exponentiating LW exp( LW ) = 1 + LW + 1 2 2 2 LW + . W } . (4. W say. This is known as canonical perturbation theory. and this consists of extending eq. (4. (An example is the planets moving around the sun—the motion of a single planet was exactly solved by Newton.56) In the case that W does not depend on . ACTION-ANGLE VARIABLES 67 represents the eﬀect of a small perturbation. P such that the new Hamiltonian is cyclic in all the new generalized coordinates . is conserved. The classical method of generating the canonical transformations is called Von Zeipel’s method. There are various variants of the Lie transform method.57) 4.) One strategy for solving the perturbed problem is to successively transform the new Hamiltonian into an exactly integrable one order by order in a power series in .42) to higher order in .. The problem with this method is the awkward mixture of old and new variables that has to be unscrambled.) This suggests a general strategy for solving Hamiltonian dynamics problems— make a canonical transformation to new canonical coordinates Q. Thus we can take H0 to be the Hamiltonian for the planets with only the gravitational ﬁeld of the sun included. If we regard as pseudo-time..

but time does not permit the development of explicit action-angle transformations for these systems. (For an autonomous system in fact the integrals of motion have no explicit dependence on t at all. We have already come across a transformation to action-angle coordinates.58) A system for which this strategy can. Clearly. they are invariant sets of the dynamical system (just as a ﬁxed point is an invariant set consisting of just one point) and are often called invariant tori. In this. rather than increasing secularly. Then the dynamics is very simple since all the new momenta are constants of the motion and thus the generalized velocities ∂K/∂Pi are also constant. Other well known integrable systems are the particle moving under the inﬂuence of an inverse-square law force. and thus such a set of canonical coordinates is called action-angle variables.) If the problem is such that the orbits remain in a bounded region of phase space. and the new coordinates Qi can be taken to be angles deﬁned on the tori. Although much of the classical mechanics literature is concerned with integrable problems it must be emphasized that integrability is not generic— given an arbitrarily chosen Hamiltonian. since there are n of the Pi . . Of particular interest are inﬁnite-dimensional integrable Hamiltonian systems. so all the new generalized coordinates evolve linearly in time Qi = Qi0 + ∂K t. ∂Pi (4.29). The dimensions of the Pi are then those of action. Thus the study of integrable systems remains an important one. such as the Korteweg– de Vries equation and nonlinear Schr¨dinger equation that give rise to soliton o phenomena. then the coordinates have to oscillate in a quasiperiodic fashion. be carried out is called integrable. including the limits t → ±∞. case the orbits are conﬁned to topological tori labelled by the new momenta. (4. 4. in principle at least. rather general. where H0 is integrable.2). and the physical pendulum.68 CHAPTER 4. a necessary condition for integrability is that there must exist n independent integrals of the motion that are well behaved as functions of the old canonical coordinates for all time. Nevertheless we are very often interested in cases which are not far integrability.4. in the sense that the Hamiltonian can be decomposed into the form H0 + H1 required to apply perturbation theory (see Sec. CANONICAL TRANSFORMATIONS K = K(P ) (we consider the autonomous case only here). from the set of all possible ones. for the harmonic oscillator problem (in the case when ω0 = const). Since the Hamiltonian ﬂow maps the Pi = const tori onto themselves. eq. it will almost always not be integrable.

. The essential idea in this theory is to ﬁnd an invariant torus of the unperturbed system which is as far from resonance as possible and to keep changing parameters in such as way as is increased away from zero that the torus remains nonresonant. 4. However by choosing the region of phase space carefully one may be able to construct an invariant torus by perturbation theory. To do this we calculate the divergence of ˙ z = (∂H/∂p. not only to Hamiltonian 2 Unless we also include time reversal. This is most easily seen (in the case of transformations continuously connected with the identity at least) by showing that Hamiltonian ﬂows are incompressible (see Fig. That is. as shown by Kolmogorov. This result obviously applies to any ﬂows built up from inﬁnitesimal canonical transformations. x → x.e. This blowup is an example of a resonance between the frequency of rotation and the frequency of the perturbation (in this case zero frequency). For H 2mgl the potential energy term is small compared with the kinetic energy and we can use H0 = p2 /2m and take g as the expansion parameter. as the rotation frequency approaches zero and we approach the separatrix.6.1 Properties of canonical transformations Preservation of phase-space volume One of the most important properties of canonical transformations is that their Jacobian determinants det(∂Zi /∂zj ) are always unity2 . perturbation theory will in general give results useful over only a ﬁnite time interval as the long-time behaviour may be chaotic and impossible to describe in detail. ∂q ∂p ∂p ∂q (4.6. a region of phase space has the same volume whether represented in the old or new canonical coordinates. PROPERTIES OF CANONICAL TRANSFORMATIONS 69 If the problem is not integrable. −∂H/∂q): ˙ div z = ∂ ∂H ∂ ∂H · − · ≡0. p → −p.2).4.6 4. 4. i.59) Then the conservation of the phase-space volume follows from the application ˙ of Gauss’ theorem since the surface integral of z is the rate of change of the volume within the surface. This will give θ a perturbation theory involving inverse powers of pθ . even in a nonintegrable system. Although we shall not go into this further. which will obviously blow up as |pθ | → 0. Arnol’d and Moser (the KAM theorem). some idea of the meaning of resonance can be had by considering the pendulum problem. where the Jacobian is −1.

60) . CANONICAL TRANSFORMATIONS p1 z(t) z(t1) q1 q2 . which is the phase-space density of a statistical ensemble of systems normalized so that the probability of ﬁnding a system in a small region of phase space of volume d2n z is f (z. f obeys the continuity equation ∂f ˙ + div(zf ) = 0 .70 CHAPTER 4.) This result has profound implications both in statistical mechanics and in dynamical systems theory. t) d2n z. f (q. each particle of which represents a system of the ensemble. Liouville’s Theorem In statistical mechanics we exploit the ﬂuid dynamical analogy implied by the above “proof” by deﬁning a distribution function.) Then. p. (In fact the conservation of phase-space volume holds for all canonical transformations.. t). (Think of phase space as being ﬁlled with a dust cloud. pn Figure 4. since the number of systems in the ensemble is conserved.2: Incompressible motion of a volume in phase space. ∂t (4. time evolution..

A general closed N -body system in statistical mechanics is nearly always assumed to be completely chaotic (the ergodic hypothesis). is obtained when {f. merely that for any attracting direction there must be a repelling direction to balance it. Hamiltonian mechanics as a dynamical system The conservation of phase-space volume provides a strong constraint on the long-time behaviour of a system in that it rules out the possibility of purely attracting or repelling sets.59). Thus the period map is area preserving. A standard example is the “kicked rotor”—a pendulum (see Fig. which can be achieved by taking f to be a function of the integrals of the motion. if the basin of attraction of an attracting ﬁxed point had the same dimensions as phase space we could draw a surface entirely enclosing the ﬁxed point which would contract with time. 2. This is Liouville’s theorem of statistical mechanics. thus contradicting the conservation of phase-space volume.61) That is. ∂f /∂t = 0. H} = 0. though in a “collisionless” plasma it can be appropriate to consider other possibilities]. ˙ Using z = (∂H/∂p. proportional to exp(−H/kT ) where k is Boltzmann’s constant and T is the temperature. The simplest system in which to see some of the consequences of the conservation of phase-space volume is provided by the period map of a periodically forced one-degree-of-freedom system (sometimes called a 1 1 degree 2 of freedom Hamiltonian system). For instance. In this case the equilibrium distribution function is a function of H only [usually the Maxwellian distribution. using the incompressibility result eq.4. PROPERTIES OF CANONICAL TRANSFORMATIONS or.6. so that the only integral of motion in a closed system is H itself.62) ∂t As a consequence we see that statistical equilibrium. This does not mean that ﬁxed points that are attracting in some directions cannot exist in Hamiltonian systems. ∂t ∂z 71 (4. (4.2) in which gravity is varied periodically (say by oscillating the support vertically . −∂H/∂q) we can write the result in terms of Poisson brackets ∂f + {f. Then the phase-space is only two dimensional and the “volume” is in fact an area. H} = 0 (4. ∂f ∂f ˙ + z· =0. the distribution function is constant on a physical path in phase space.

72 CHAPTER 4. so that g0 is the time average of g(t). with n an integer. p) = + mg(t)l(1 − cos θ) . 2ml (4. and pn immediately Figure 4. −N < n < N .32) over the interval t−N + ε ≤ t ≤ tN + ε.64) 2ml2 Between the kicks the motion is p(t) one of uniform rotation with conpn stant angular velocity p/ml.65) for t in each range tn ≡ n∆t < t < tn+1 ≡ (n + 1)∆t. t = tn as yet unknown. (4. ∆t p(t) = pn+1 θ(t) = (4.67) .20) p2 H(θ. (3. ε → 0+.3. N ≥ 1. Furthermore we assume the gravity to be applied purely impulsively ∞ g(t) = g0 ∆t n=−∞ δ(t − tn ) (4. piecewise linear function of t and p a piecewise tn−1 tn tn+1 t constant step function with only the values θn at. As the kicked rotor. CANONICAL TRANSFORMATIONS and transforming to the accelerating frame in which the support is stationary). (3. illustrated in Fig. Inserting this trial function in the phase-space action integral eq. we get N −1 Sph = n =−N p2 +1 ∆t pn +1 (θn +1 − θn ) − n 2 − mg0 l∆t (1 − cos θn +1 ) . and δ(t) is the Dirac δ function. Thus θ(t) we can give a potentially exact repθn ε resentation of the phase-space path g(t) ∆t with θ a continuous. (noting that θn occurs in the terms of the sum for both n = n and n = n − 1) are ∂Sph pn+1 ∆t = θn+1 − θn − =0. 4. The form of the equations of motion is unaﬀected by the fact that g depends on time. so the Hamiltonian is the same as in eq.3: Trial function for before. ∂pn+1 ml2 (4.66) The conditions for Sph to be stationary under variations of θn and pn+1 .63) where tn ≡ n∆t. 1 [(tn+1 − t)θn + (t − tn )θn+1 ] .

(4. which is the value at which the last KAM invariant curves with the topology of a rotating pendulum orbit become unstable and break up into invariant Cantor sets. . Two of these KAM curves are shown.4 shows some typical iterated orbits of the map. showing both chaotic and regular regions. each with chaotic separatrices.69) shows that the width of the physical pendulum’s separatrix shrinks to zero as k → 0 when represented in terms of y]. 0) and (0. 3.70) where the chaos parameter k is deﬁned by k≡ g0 (∆t)2 . . while for k greater than this value a phase-space point can diﬀuse without bound in the positive or negative y-direction. l (4. which is p expressed in units of 2πml2 /∆t. being the only curves completely crossing the ﬁgure from left boundary to right. (4. It is clearly seen that they approximate the Hamiltonian equations of motion for the physical pendulum as ∆t → 0.71) Figure 4. 0) and (1. 1) [also (1. (4. PROPERTIES OF CANONICAL TRANSFORMATIONS and 73 ∂Sph = pn − pn+1 − mg0 l∆t sin θn = 0 .6..1 are seen as large “islands” around (x. There are also new islands due to resonances not present in the physical pendulum problem (in fact an inﬁnite number of them). From eq. 2πml2 p = y. . y) = (0.1 is highly chaotic. The remnants of the librating pendulum orbits around the O points in Fig.71) we see that the integrable limit ∆t → 0 corresponds to k → 0 [noting that eq.4.4 is kc = 0. θ = 2πx . (4. The value of k used in Fig. 3. Commonly the map is re-expressed in terms of an angle coordinate x with period normalized to unity and a nondimensionalized momentum y.69) (4. 4. 1) because the phase space is periodic in y as well as x]. For k < kc the motion is bounded in y. ∆t Then the map becomes yn+1 = yn − xn+1 k sin 2πxn .971635406 .68) ∂θn These equations deﬁne an area-preserving map known as the Standard Map. but the region corresponding to the X point in Fig. 2π = xn + yn+1 .

respectively. φ . θ and φ . ∂q ∂p ∂p ∂q ∂Q ∂P ∂P ∂Q (4. K= p2 p2 φ θ + + V (θ. t) .4: Some representative orbit segments for the Standard Map showing both chaotic and regular motion.72) 4.2 Transformation of Poisson brackets Finally. we note the important result the Poisson bracket is invariant under canonical transformation {f. CANONICAL TRANSFORMATIONS y 0 0 x 1 Figure 4.7.74 1 CHAPTER 4.7 4. g} = ∂f ∂g ∂f ∂g ∂f ∂g ∂f ∂g · − · = · − · . 4.1 Problems Coriolis yet again Given the Hamiltonian for a particle on a sphere in nonrotating latitude and longitude coordinates.6. 2mR2 2mR2 cos2 θ .

(b) Show that the approximate action integral from t−N to tN (N ≥ 1 being an arbitrary integer) is stationary for −N < n < N if xn obeys the secondorder diﬀerence equation 2 ω0 (∆t)2 (xn−1 + 4xn + xn+1 ) 6 xn−1 − 2xn + xn+1 = − ¯ (c) Using the approximate action integral S(xn . Use units such that m = ω0 = 1 and use three timesteps: ∆t = 0. Find the Hamiltonian H in the rotating coordinates. use Maple or Mathematica.) This provides a discrete-time dynamical system approximation to the true continuous-time system for the harmonic oscillator. F1 (q. p0 = 0 and plot the result. where φ is the usual latitude measured in the rotating frame deﬁned in problem 2. (Again. φ. Represent this linear transformation as a matrix and show that the determinant is unity.7. (d) Iterate the map obtained in (c) 100 times taking the initial point as x0 = 1. Q) = −S(xn . For a bonus mark verify one of the solutions by showing that the secondorder diﬀerence equation derived in (b) above (when arranged to give xn+1 in terms of xn and xn−1 ) gives the same sequence of x-values as the discrete-time dynamical system in (c) if two successive values of x given by the dynamical system are used to start the second-order diﬀerence equation. φ = Ωt + φ. pn+1 ≡ P ). (It would be advisable to use Maple or Mathematica for the algebra.1.7. That is.4. I. pn ≡ p) to (xn+1 . ﬁnd the linear canonical transformation from (xn .) . as a type 1 generating function.e. xn+1 ). pθ .47. 4. For another bonus mark calculate the eigenvalues for the three values of ∆t given above. the discrete time “Lagrangian” and “Hamiltonian” discriptions are dynamically equivalent. with q = xn and Q = xn+1 . pφ canonical coordinates. ∆t = 3.7. PROBLEMS 75 ﬁnd the transformation to the θ.1. xn+1 ) evaluated over the ¯ range tn < t < tn+1 .45 and ∆t = 3.2 Diﬀerence approximations 2 1 Using the harmonic oscillator Lagrangian L = 2 m(x2 − ω0 x2 ) ˙ (a) calculate an approximate action integral S using the piecewise-linear trial function 1 x(t) = [(tn+1 − t)xn + (t − tn )xn+1 ] ∆t for t in each range tn ≡ n∆t < t < tn+1 ≡ (n + 1)∆t.

CANONICAL TRANSFORMATIONS .76 CHAPTER 4.

m1 + m2 (5.Chapter 5 Answers to Problems 5. Answer Let the position of particle 1 with respect to a stationary Cartesian frame be {x1 . However. The rigid rod constraint equation is then (x1 − x2 )2 + (y1 − y2 )2 + (z1 − z2 )2 = l2 .1(a)]. y1 .6. z1 } and that of particle 2 be {x2 .5. (1.6.1) This reduces the number of degrees of freedom from 6 to 5. whose position vector is rCM = m2 r1 + m1 r2 . and the spherical polar angles θ and φ depicted in Fig. (5. z2 }. y2 . 1.1 to specify the orientation of the rod. {x1 .2) and ﬁnd suitable generalized coordinates for the system incorporating this holonomic constraint. which we could take to be position of particle 1. a better choice for the three position coordinates would be the coordinates of the centre of mass [see Problem 3.2) Problem 1. y1 . Write down a constraint equation of the form eq.1 Chapter 1 Problems Problem 1.1 Two particles are connected by a rigid rod so they are constrained to move a ﬁxed distance apart. z1 }.2 ˙ Suppose we know that the angular momentum vectors rk ×mk rk of a system of particles are all nonzero and parallel to the z-axis in a particular Cartesian 77 .

78 CHAPTER 5. . so the assumption that zk = 0 must be wrong. 2. and show that they lead to a set of holonomic constraints. . .3) we might ask whether this is the only solution.3) by zk xk .3) which form a set of holonomic constraints. Thus we can divide the ﬁrst eq. (5. Answer The diﬀerential constraints are provided by the vanishing of the x and y components of the angular momentum vectors (multiplied by dt) yk dzk − zk dyk = 0 zk dxk − xk dzk = 0 . Dividing by xk yk we clearly have a contradiction. yk = rk sin θk . r2 .4) (5.1 This problem has been called “Ecliptic” because this is the astronomical name of the plane to which the planets are approximately “constrained”. xN . and add to show (since the terms dzk /zk cancel) that dxk /xk − dyk /yk = 0. Thus the motions are constrained to lie in the x. (5. We can assume that. for k = 1. θN }. where θ and ζ are arbitrary curvilinear coordinates on S. . except for isolated instants. N . y2 .6. However we are told that the z components of the angular momenta are nonzero. y plane and we can either take q = {x1 . . so that q = {r1 . . y1 . θ1 . Write down the diﬀerential constraints implied by this fact. Note: Although eq. (5. . . . so that xk dyk − yk dxk = 0. suppose that zk is not equal to zero. It suggests that the reason for this constraint is because the planets were all created by condensation from the same spinning disk of dust and gas and thus have parallel angular momenta. Remark 5. .4) clearly solves eq.3 Show that any geodesic r = x(τ ) on a two-dimensional manifold S : r = X(θ. (5. the second by yk zk . These are satisﬁed by taking zk ≡ 0 . yN } or transform to polar coordinates using xk = rk cos θk . . . ANSWERS TO PROBLEMS coordinate system. Thus. is such that the curvature vector κ(τ ) is everywhere normal to S (or zero). x2 . Hence write down suitable generalized coordinates for the system. 2 Problem 1. rN . . θ2 . xk and yk are nonzero also. ζ) embedded in ordinary Euclidean 3-space.

5. we have. ζ) dτ with f≡ and dl ˙ ˙˙ ˙ = X2 θ2 + 2Xθ ·Xζ θζ + X2 ζ 2 θ ζ dτ 1/2 (5. ∂θ X and ∂ζ X. ∂θ f = 1 ˙2 ˙˙ ˙ θ ∂θ X2 + 2θζ∂θ (Xθ ·Xζ ) + ζ 2 ∂θ X2 θ ζ 2f (5. (5.7) Similarly. where Xθ ≡ ∂θ X ≡ ∂X/∂θ ˙ ˙ and Xζ ≡ ∂ζ X ≡ ∂X/∂ζ. Also.6) (5.5).8) . θ. where e (τ ) ≡ dx/ dl is the unit tangent vector at each point along the path r = x(τ ). by interchanging θ and ζ we have ∂ζ˙ f = e ·Xθ . Answer The vector length element is dl = dθXθ + dζXζ . Thus l = f (θ. and the normal vectors n and κ.ζ) S Figure 5.1. for any path e ≡ ∂θ˙ f = 1 ˙ ˙ X2 θ + Xθ ·Xζ ζ θ f 1 ˙ ˙ = Xθ θ + Xζ ζ ·Xθ f ≡ e ·Xθ (5.1: Surface S with curvilinear coordinate system. The curvature vector is deﬁned by κ ≡ de / dl. the three tangent-plane vectors e . CHAPTER 1 PROBLEMS 79 ζ = const ∂ζX(θ. a geodesic.ζ) n e || κ θ = const ∂θX(θ. f dτ f From eq. ζ.5) ˙ ˙ 1 dl θXθ + ζXζ = .

1(a) As a model of the motion of a ﬂuid element or dust particle in a planetary (e. ≡ e· dτ (5. n.7) and eq. Thus either κ = 0 or it is parallel to the unit normal. (5. dτ d e =0. dτ (5. expanding out and cancelling. ζ) (see Fig. Xθ · d e =0.80 = CHAPTER 5.13) and (5. dτ θ Using the identities eq. dτ dτ Or. interchanging θ and ζ. we have ∂ζ f ≡ e · dXζ . (5. dτ (5.11) Similarly. Earth’s) atmosphere. (5.15) Since the vectors Xθ and Xζ span the tangent plane at the point r = X(θ.13) (5. (5.2 Chapter 2 Problems Problem 2.1).9) we can write this as d dXθ e ·Xθ − e · =0. the Euler–Lagrange equation for ζ variations becomes Xζ · (5.g. 2 5.12) (5. we have shown that the curvature vector κ has no component in the tangent plane.9) Similarly. consider the motion of particle of unit mass constrained to move on the surface of a perfectly smooth sphere of radius R . 5.10) Now consider the Euler–Lagrange equation for a geodesic d ∂ ˙ f − ∂θ f = 0 .14) as Xζ ·κ = Xθ ·κ = 0 .14) Recognizing that κ ≡ ( d/f dτ )e we can write eqs. ANSWERS TO PROBLEMS 1 ˙2 ˙˙ ˙˙ ˙ θ Xθ ·Xθθ + θζXθθ ·Xζ + θζXθ ·Xθζ + ζ 2 Xζ ·Xθζ f 1 ˙ ˙ ˙ ˙ = θXθ + ζXζ · θXθθ + ζXθζ f dXθ .8.

(a) Write down the Lagrangian in a frame rotating with the planet. 2.17) and the Lagrangian. where x.6).5. φ) (see Sec.20) . Answer ˙ ˙ x = −Rθ sin θ cos(φ + Ωt) − R(Ω + φ) cos θ sin(φ + Ωt) ˙ ˙ ˙ y = −Rθ sin θ sin(φ + Ωt) + R(Ω + φ) cos θ cos(φ + Ωt) ˙ ˙ z = Rθ cos θ ˙ Thus the kinetic energy is given by T = = 1 2 (5. Write down the equations of motion.18) (5. 2 Thus ∂L ˙ = R2 θ ˙ ∂θ ∂L ˙ = R2 (Ω + φ) cos2 θ ˙ ∂φ (5.4. L = T − V . Among this class of potentials ﬁnd the special case V = V0 (θ) required to make equilibrium possible (i. φ) . x = R cos θ cos(φ+Ωt).19) and (5. y = R cos θ sin(φ+Ωt). taking the generalized coordinates to be the latitude and longitude so that z = R sin θ.16) x2 + y 2 + z 2 ˙ ˙ ˙ ˙ θ2 sin2 θ cos2 (φ + Ωt) + sin2 (φ + Ωt) ˙ + (Ω + φ)2 cos2 θ sin2 (φ + Ωt) + cos2 (φ + Ωt) ˙ + θ2 cos2 θ 1 2 2R = 1 2 2R ˙ ˙ θ2 + (Ω + φ)2 cos2 θ . y. Suppose the force on the particle is given by an eﬀective potential V (θ.e. z is a non-rotating Cartesian frame. and ﬁnd a ﬁrst integral (i. so that the equations of motion admit the solution ˙ ˙ θ = φ = 0 at each latitude). constant of the motion) in the case where V is independent of longitude.e. where θ and φ are the latitude and longitude respectively. CHAPTER 2 PROBLEMS 81 rotating with angular velocity Ω about the z-axis. (5.2. by ˙ ˙ L = 1 R2 θ2 + (Ω + φ)2 cos2 θ − V (θ.

Assuming the winds are slow. Hence show that motion in a depression or low is cyclonic.82 CHAPTER 5.21) ˙ In the case where V = V0 (θ). 2. (5. φ) = V0 (θ) + h(θ.23) where the second equality follows from the identity cos 2θ ≡ 2 cos2 θ − 1 and the choice const = 1 R2 .20). φ is an ignorable coordinate and thus ∂L/∂ φ.8. given by eq.22) (5. counter-clockwise in the Northern Hemisphere for Ω > 0). and use the identity sin 2θ ≡ 2 sin θ cos θ. where h is the enthalpy (proportional to p(γ−1)/γ ) deﬁned in Sec. ˙ ˙ If in this latter case we further demand that V0 be such that θ ≡ φ = 0 for all θ and φ.e. then eq. Draw a sketch of a typical weather-map “low” (i. 4 2 (5.6. 4 Problem 2. φ). ∂θ or.4.1(b) Now assume that the general form of the potential is V (θ.e. clockwise in the Southern Hemisphere.21) gives ∂V0 1 = − 2 R2 Ω2 sin 2θ . so that second order time derivatives of θ and φ (and products of ﬁrst order time derivatives) can be neglected. the direction of the force on a ﬂuid element and the direction of motion.24) (5. (5. . show that the velocity of the particle in the rotating frame is at right angles to the pressure gradient − p. ANSWERS TO PROBLEMS ˙ and thus the Lagrange equations of motion ( d/ dt)∂L/∂ θ = ∂L/∂θ and ˙ ( d/ dt)∂L/∂ φ = ∂L/∂φ become ∂V ¨ ˙ R2 θ = −R2 (Ω + φ)2 sin θ cos θ − ∂θ and ∂V ¨ ˙ ˙ R2 φ cos2 θ − 2R2 θ(Ω + φ) sin θ cos θ = − . is an integral of the motion. ∂φ (5. integrating. where “cyclonic” in geophysical ﬂuid dynamics means “in the direction of the planet’s rotation” (i. V0 (θ) = 1 R2 Ω2 cos 2θ + const = 1 R2 Ω2 cos2 θ . a localized depression in p) showing the pressure contours.

and so. and the slow approximation will be best satisﬁed when h is small. rather. (5. It is for this reason that the vortices caused by highs and lows of pressure are so long lived—the air does not rush in to equalize the pressure diﬀerences but.25) and (5.25) ∂h ˙ R2 Ωθ sin 2θ ≈ . CHAPTER 2 PROBLEMS 83 Answer Making the slow approximation.e. since γ > 1. eqs.8.26).5. in the θ-φ plane) and use eqs.27) from eqs.21) and (5. . (5.1(c) Consider a charged particle constrained to move on a non-rotating smooth insulating sphere. Write down the Lagrangian in the same generalized coordinates as above and show it is the same as that for the particle on the rotating planet with appropriate identiﬁcations of Ω and V . and taking into account eq. (5. Thus the eﬀect of the Coriolis forces is to make the air move along the pressure contours.26) ∂φ Thus the pressure gradient causes a wind proportional to spatial derivatives of h. These two equations give the wind speed in angular velocities.25) and use (5. even though the forces acting on a ﬂuid element act at right angles to the contours.23) the equations of motion. (5. immersed in a uniform magnetic ﬁeld B = Bez . Problem 2. We now calculate the time rate of change of the enthalpy along a stream line ˙ h ≡ w· h ˙ ˙ ∂h + φ ∂h ≡ θ ∂θ ∂φ 1 ∂h ∂h ∂h ∂h ≈ − 2 Ω sin 2θ R ∂φ ∂θ ∂θ ∂φ ≡0 (5. (5.49) that p/ρ = constp(γ−1)/γ . the enthalpy is an increasing function of the pressure.22). Note: It is easy to show from eq. become ∂h ˙ R2 Ωφ sin 2θ ≈ − ∂θ and (5. on which the electrostatic potential is a function of latitude and longitude. rotates around the maxima and minima of pressure.26) to show the wind direction at a few representative points on a typical pressure contour.2. (2. In order to establish the direction of rotation it is suﬃcient to draw a sketch in a Mercator’s projection (i.

2 Consider the following potential V . φ) . φ) . 2 2 Completing the square we can rewrite the Lagrangian in the form L= 2 1 2 mR (5.32) 8m Note that eq.28) Thus the Lagrangian is ˙ ˙ ˙ L = 1 mR2 θ2 + φ2 cos2 θ + 1 eBR2 φ cos2 θ − eΦ(θ. Ay = 1 Bx. (5. 2 2 e˙ ·A = r = 1 ˙ 2 eB(xy 2 1 2 eBR − y x) ˙ ˙ ˙ cos θ cos φ(−θ sin θ sin φ + φ cos θ cos φ) ˙ ˙ + sin φ(θ sin θ cos φ + φ cos θ sin φ) = 2 ˙ 2 1 2 eBR φ cos θ. where the constant ω0 is the angular frequency of oscillations having amplitude small compared with the characteristic length l0 . (5. from eq. Az = 0.31) can also be written Ω = 1 ωc . Thus.8. φ) . where ωc ≡ eB/m is 2 the cyclotron frequency. (5.17) with Ω set r to zero. (5. We now need a vector potential A such that Bz ≡ ∂x Ay − ∂y Ax = B. . one with a nonlinear restoring force).30) 8m This is the same as eq.e.31) e2 B 2 R2 cos2 θ + eΦ(θ. corresponding to a particle of mass m oscillating along the x-axis under the inﬂuence of a nonideal spring (i. φ) = Problem 2. V (x) = 2 mω0 2 x2 + σ x4 2 l0 . where T is given by eq. (5. (5. and σ = ±1 depends on whether the spring is “soft” (σ = −1) or “hard” (σ = +1). ANSWERS TO PROBLEMS Answer The Lagrangian is L = T +e˙ ·A−eΦ.16).18) with the identiﬁcations Ω= and eB 2m (5.29) ˙ ˙ eB θ2 + φ + 2m 2 cos2 θ − e2 B 2 R2 cos2 θ − eΦ(θ. V (θ.84 CHAPTER 5. A suitable choice is Ax = − 1 By. (5.

respectively. Show that C ≡ 0. CHAPTER 2 PROBLEMS Consider the trial function x = l0 [A cos ωt + B cos 3ωt + C sin 3ωt] . but plot ω/ω0 and B vs.5. ﬁnd implicit relations giving approximate expressions for ω. . 85 where A. B and C as functions of A. The trial function is strictly appropriate only to the case A 1. By using this trial function in the time-averaged Hamilton’s Principle.2. B. and ω is the nonlinearly shifted frequency. Answer See the following Maple notebook. A from 0 to 1 in the case of both a hard and a soft spring. C are the nondimensionalized amplitudes of the fundamental and third harmonic.

36) . with the coordinates of the two particles being given by x1 y1 z1 x2 y2 z2 = = = = = = x − α1 l sin θ cos φ y − α1 l sin θ sin φ z − α1 l cos θ x + α2 l sin θ cos φ y + α2 l sin θ sin φ z + α2 l cos θ . Take the generalized coordinates of the system to be q = {x. z.86 CHAPTER 5. 2 2(m1 + m2 ) and 1 ˙2 2 m1 z1 + 1 m2 z2 = ˙2 2 (5. 2 2(m1 + m2 ) (5. θ. 2 2(m1 + m2 ) (5. y. y. ANSWERS TO PROBLEMS 5.35) m1 + m2 2 m1 m2 l 2 ˙ 2 2 z + ˙ θ sin θ . 1 ˙2 2 m1 x1 + 1 m2 x2 = ˙2 2 m1 + m2 2 m1 m2 l 2 ˙ ˙ x + ˙ (φ sin θ sin φ − θ cos θ cos φ)2 .1(a) Write down a Lagrangian for the problem of two particles of mass m1 and m2 connected by a light rigid rod of length l in a gravitational ﬁeld g.33) Because m1 α1 = m2 α2 the cross terms cancel when we expand the x-contribution to the kinetic energy. φ}.34) Similarly 1 ˙2 2 m1 y1 + 1 ˙2 2 m2 y2 m1 + m2 2 m1 m2 l 2 ˙ ˙ = y + ˙ (φ sin θ cos φ + θ cos θ sin φ)2 . Answer ˙ ˙ x1 = x + α1 lφ sin θ sin φ − α1 lθ cos θ cos φ ˙ ˙ ˙ ˙ x2 = x − α2 lφ sin θ sin φ + α2 lθ cos θ cos φ ˙ ˙ (5. where α1 ≡ m2 /(m1 + m2 ) and α2 ≡ m1 /(m1 + m2 ) (so that (x. z) is the centre of mass).3 Chapter 3 Problems Problem 3.5.

eq. CHAPTER 3 PROBLEMS Thus. in contrast to the latitude angle whose origin was at the equator—hence sin θ and cos θ are interchanged. 2 2(m1 + m2 ) 87 (5. From eq. We see from eqs. (5.] This is a consequence of the well-known result that the kinetic energy of a system of particles is the kinetic energy of the centre of mass plus the kinetic energy in the centre-of-mass frame.37) and (5.37) the mass matrix is µ= M 0 0 0 0 0 M 0 0 0 0 0 M 0 0 0 0 0 µl2 0 2 0 0 0 0 µl sin2 θ . (5. y. (5. adding.38) that.37) This is the kinetic energy of a free particle of mass (m1 +m2 ) plus that of a particle with the reduced mass µ ≡ m1 m2 /(m1 + m2 ) constrained to move on the surface of a sphere of radius l. (5.2. the kinetic energy is T = m1 + m2 2 m1 m2 l 2 ˙ ˙ (x + y 2 + z 2 ) + ˙ ˙ ˙ (θ2 + sin2 θ φ2 ) . and recalling that we are here using spherical polar coordinates where the origin of θ is at the pole.38) and the Lagrangian is L = T −V .5.3. Is it of the form T + V ? Answer Answering the second part of the question ﬁrst.17) of the Coriolis problem with Ω set to zero. (5.6. the Hamiltonian must be of the form T + V because the constraint |r2 − r1 | = l does not depend explicitly on time and thus the system comes under the general case treated in Sec. y and φ and that the centre-of-mass coordinates x.5.1(b) Use this Lagrangian to construct the Hamiltonian for the system. The potential energy is V = (m1 z1 + m2 z2 )g = (m1 + m2 )zg . z form a subsystem completely independent from that of the angular coordinates. 3. by using the centre of mass as the origin for our polar coordinates. [Cf.39) . Problem 3. we have made L ignorable in x.

px . eq. 2 Is it of the form T + V ? Answer From the above. z θ φ The z-dynamics is just uniformly accelerated motion.6. This is trivially inverted to give. Thus θ is an oscillatory function of time.2. φ) . py and pφ are constants of the motion. cos2 θ Also (5.0 − mgt. p2 p2 p2 p2 p2 y φ z x θ H= + + + + + M gz . ˙ ∂φ ˙ φ= R2 pφ −Ω. ˙ ˙ L = 1 R2 θ2 + cos2 θ (Ω + φ)2 − V (θ. ﬁxed direction. 3.5.8.1.44) . with solution z = z0 + (pz.41) (5. from Sec. R pφ ≡ ∂L ˙ = R2 cos2 θ (Ω + φ) . 2 2 sin2 θ 2M 2M 2M 2µl 2µl (5. (In fact we know from elementary physics that the motion can be integrated in closed form because the rod spins at right angles to a constant angular momentum vector with arbitrary. The motion of the rod can thus be found by studying two independent. with Hamiltonians p2 + M gz and p2 /2µl2 + p2 /2µl2 sin2 θ respectively.0 /m)t − 1 gt2 .88 CHAPTER 5.1(c) Find the Hamiltonian corresponding to the Coriolis Lagrangian in Problem 2. ANSWERS TO PROBLEMS where M ≡ m1 + m2 is the total mass and µ is the reduced mass deﬁned earlier. y and φ are cyclic (ignorable) coordinates. A qualitative understanding of the θ2 dynamics can be had by considering that the “pseudo-potential” p2 /2µl2 sin2 θ φ forms a well about θ = π/2 (assuming pφ = 0) with sides becoming inﬁnitely high at θ = 0 and π.) Problem 3.42) pθ ˙ θ= 2 . pθ ≡ so ∂L ˙ = R2 θ ˙ ∂θ (5. one-dimensional dynamical problems in z and θ. (3.27). pz = pz.43) so (5.40) Note: Because x.

Problem 3. Indicate representative trajectories of the system in phase-space.3. which is in the standard form treated 2 in Sec.45) 2 2 cos2 θ 2R 2R Given that T = = 1 2 2R ˙ ˙ θ2 + cos2 θ (Ω + φ)2 (5.8. (From Hamilton’s equations of motion we see that such points correspond to x = ˙ p = 0 and are thus equilibrium points.2 Find the Hamiltonian corresponding to the Lagrangian of the anharmonic oscillator in Problem 2.2.5 -0. p) in the 2 2 soft potential case for (a) H < mω0 l0 /8 and 2 2 (b) H > mω0 l0 /8.e.3. H 0.46) p2 p2 φ θ + 2R2 2R2 cos2 θ it is clear that the Hamiltonian diﬀers from T + V by the term −Ωpφ . ˙ The Lagrangian is L = 1 mx2 + V (x).5 1 x Answer Figure 5.2. CHAPTER 3 PROBLEMS Thus 89 ˙ ˙ H ≡ pθ θ + pφ φ − L p2 p2 p2 p2 φ φ θ = + 2 − Ωpφ − θ 2 − + V (θ.2: H(x. φ) R2 R cos2 θ 2R 2R2 cos2 θ p2 p2 φ θ = + − Ωpφ + V (θ. (5.47) Before sketching the phase space we ﬁrst get a qualitative feel for its salient features by looking at the behaviour of H as a function of x and p and ﬁnding its stationary points.5.05 -1 -0. Sketch contours of H(x. Thus x = p/m and the Hamiltonian is ˙ H= 2 p2 mω0 + 2m 2 x2 − x4 2 l0 . φ) .5. where ∂H/∂x = ∂H/∂p = 0.1 0. 3. 0). Try to include representatives of each qualitatively distinct orbit. i.05 -0.1 0. where the system does not move—ﬁxed ˙ . (5.

3: Phase 2 2 mω0 l0 /8. respectively. H = mω0 l0 /8. Thus H approaches +∞ as |p| → ∞. with its minimum at x = 0. It is seen that the contour H = mω0 l0 /8 forms the separatrix between ﬁve classes of qualitatively diﬀerent orbit: 0. using units such that l0 = ω0 = m = 1. On the other hand. There are 0.3.5 1 are unstable towards small perturbations. ANSWERS TO PROBLEMS points of the dynamical system. x = 0.6 (±l0 / 2. Points where the Hamiltonian is a minimum are stable.5 0 0.2. For |x| l0 the term quartic in x is very small compared with the quadratic term so the graph of H is a parabola. At x 2 2 both these saddle points.4 maxima x = ±l0 / 2.) For any ﬁxed x the graph of H is a parabola.6 • bounded oscillatory (librating) orbits contained within the separatrix. As in Fig. with its minimum at p = 0. concave upwards. (5. Directions of motion (to the right in the upper half plane because x > 0 and to the left in the lower ˙ 2 2 half plane) are marked by arrows. whereas saddle points correspond to unstable equilibria. 0). It is seen that there are maxima of H at ﬁnite values of |x|. corresponding to a particle trapped in the potential well. and have in2 2 dicated the contour H = mω0 l0 /8 by a dashed line. 3. so that they are linked by the contour H = const = Figure 5.1 tor. 5. . Since the origin is the intersection of the lines of minima. Thus the points (x. we have marked the stable and unstable ﬁxed points by O and X. for |x| l0 the negative quartic term dominates over the quadratic term and H approaches −∞ as |x| → ∞. p) = (0.4 equilibrium point at (x.2 of the line of minima p = 0 and the lines of √ -0. p = 0. We now examine the behaviour of H in the limits of large and small x. The maxima and minima of H as a function of x are the roots of ∂H 2x2 2 = mω0 x 1 − 2 ∂x l0 .2 mimimax or saddle points at the intersections p 0 X O X -0. but -1 -0.90 CHAPTER 5. The behaviour of H on the x-axis is graphed in Fig. 0) are also equilibrium points. space of the anContours of H in units such that l0 = ω0 = harmonic oscillam = 1 are plotted in Fig. p) = √ -0. there is a stable 0. 2 2 so that H < mω0 l0 /8. 5.48) √ which are at x = 0 (the minimum) and x = ±l0 / 2 (the maxima). concave upwards.

(5.49) we have B· ψ = ∂y ψ ∂x ψ − ∂x ψ ∂y ψ ≡ 0. y)]2 p2 y x + + . corresponding to particles approaching from the left or right.2. By = ∂z Ax − ∂x Az = −∂x ψ . Problem 3.3 Consider the motion of a particle of charge e and mass m in a straight. Answer The given expression for H follows directly by writing eq. By = −∂ψ/∂x.3. Answer Given Ax = Ay = 0. y). respectively. which cross the p-axis (x = 0) at a value x < mω0 l0 /2 or x > mω0 l0 /2. respectively) without being reﬂected. (3. H= 2m 2m 2m write down the Hamiltonian equations of motion and give two integrals of the motion. y)ez .) From eq. corresponding to particles with suﬃcient energy to cross the potential barriers (from right to left. CHAPTER 3 PROBLEMS 91 • unbounded orbits with H <√ 0 l0 /8 which cross the x-axis (p = 0) at a mω 2 2 √ value x < l0 / 2 or x > l0 / 2. Az = ψ(x. The Hamiltonian .5. Az = ψ(x.5) in Cartesian components and using Ax = Ay = 0. Thus ψ = const on a magnetic ﬁeld line and so plotting contours of ψ maps out the ﬁeld.49) (Alternatively. y) we calculate B ≡ Bx = ∂y Az − ∂z Ay = ∂y ψ . 2 2 • unbounded orbits with H > mω0 l0 /8.5. inﬁnitely long magnetic conﬁnement system with vector potential A = ψ(x. or left to right. (b) Show that the Hamiltonian is p2 [pz − eψ(x. we can observe that B = ψ×ez and take the x and y components. Bz = 0 and thus that contours of ψ(x. (a) Show that Bx = ∂ψ/∂y. y) deﬁne magnetic ﬁeld lines. and being reﬂected by a potential barrier. ×A: (5.

2 ψ = 0. Also.50) implies that ζ ≡ x + iy = (px + ipy )/m ≡ pζ /m. 1 . since H does not depend explicitly on time. 3. pz is clearly an integral of the motion. Thus. ψ(x. ∂y m ∂y ∂H = − =0. m where the prime on Ψ means derivative with respect to its argument. ANSWERS TO PROBLEMS x = ˙ ∂H px = . and ∗ means complex conjugate. This is because Amp`re’s law × B = µ0 j = 0 leads to Laplace’s equation. (c) It can be shown1 that in any region with no electric current.3. ∂x m ∂x e(pz − eψ) ∂ψ ∂H = − = . Answer ˙ Equation (5. y) can be represented as the real part of an analytic function of the complex variable ζ ≡ x + iy.51) From the last equation above. y) = Re Ψ(ζ) and show that the equations of motion for ζ and pζ ≡ px + ipy are ˙ ζ = pζ ˙ 1 pζ .92 equations of motion are CHAPTER 5. H itself constitutes a second integral of motion as shown in Sec. ∂z (5. write ψ(x.50) and p˙x = − p˙y p˙z ∂H e(pz − eψ) ∂ψ = . e and it is a standard result of complex analysis that the real and imaginary parts of analytic functions obey the two-dimensional Laplace’s equation. ∂px m py ∂H y = ˙ = . ∂py m (pz − eψ) ∂H z = ˙ = ∂pz m (5. m e = [pz − eRe Ψ(ζ)] [Ψ (ζ)]∗ . as was to ˙ ˙ be shown.

Answer We treat this as a “point transformation”. φ = Ωt + φ.4. pφ canonical coordinates. from eq. with q = {θ.56) (5.5. (5.53) as was to be shown. That is. CHAPTER 4 PROBLEMS We are given that ψ(x. t) = θ pθ + φ pφ . pφ .1 Given the Hamiltonian for a particle of unit mass on a sphere in nonrotating latitude and longitude coordinates.51). pζ = ˙ e (pz − eψ)(∂x ψ + i∂y ψ) m e = (pz − eRe Ψ)[Re Ψ (ζ) − i Im Ψ (ζ)] m e ≡ (pz − eRe Ψ)[Ψ (ζ)]∗ m 93 (5. t) = θpθ + (Ωt + φ)pφ . φ . where φ is the usual latitude measured in the rotating frame deﬁned in Problem 2. Thus.55) we have committed the usual physicist’s “abuse of notation” in taking θ and φ to denote functions of θ and φ rather than . φ. pφ . Therefore ∂x ψ = Re [∂x ζ Ψ (ζ)] = Re Ψ (ζ) . pθ . as in Sec. φ}. φ. y) = Re Ψ(ζ). Q = {θ . K= p2 p2 φ θ + + V (θ.55) [Note that in eq. pθ . φ }.4 Chapter 4 Problems Problem 4. θ and φ .52) (5.8. where θ ≡ θ and φ ≡ Ωt + φ.7.2. F2 (θ. p = {pθ .1. t) . We thus have the generating function F2 (θ. That is. pθ . 2R2 2R2 cos2 θ (5. (5.54) ﬁnd the transformation to the θ. (5. Find the Hamiltonian H in the rotating coordinates.2. pφ } and P = {pθ . pφ }. ∂y ψ = Re [∂y ζ Ψ (ζ)] = Re [iΨ (ζ)] ≡ −Im Ψ (ζ) . 5. respectively. φ. 4.

rather than generate a new notation such as G. (5. We avoided this abuse during the formal development in Sec. (5.2. H = K − Ωpφ = p2 p2 φ θ + − Ωpφ + V (θ. t) = V (θ. Ωt + φ. but one needs to be aware of what one is doing.1(c). 4. 2 2 cos2 θ 2R 2R (5. In practice it is convenient to take it as understood from the context that Q is a function of q. from Sec.2 by denoting the corresponding function by G rather than Q. ANSWERS TO PROBLEMS independent variables.2 2 Using the harmonic oscillator Lagrangian L = 1 m(x2 − ω0 x2 ) ˙ 2 (a) calculate an approximate action integral S by evaluating the action integral exactly using the piecewise-linear trial function x(t) = 1 [(tn+1 − t)xn + (t − tn )xn+1 ] ∆t for t in each range tn ≡ n∆t < t < tn+1 ≡ (n + 1)∆t.7.] The transformation equations are.57) With the identiﬁcation V (θ.59) as the same Hamiltonian derived in Problem 3. Ωt + φ. This derivation makes it clear that the reason H is not of the form T +V is because the transformation to the rotating frame is time dependent. (5. 4.2.58) ∂F2 ∂θ ∂F2 = ∂φ ∂F2 = ∂pθ ∂F2 = ∂pφ = pθ = pφ =θ = Ωt + φ . pθ = pφ θ φ and K=H+ That is. t) . Problem 4.5. . This is convenient.94 CHAPTER 5. ∂t (5. from eq.59) ∂F2 = H + Ωpφ . φ) we recognize eq.54).

. = n+1 n 2 ∆t 3 (b) Show that the approximate action integral from t−N to tN (N > 1 being an arbitrary integer) is stationary for −N < n < N if xn obeys the secondorder diﬀerence equation xn−1 − 2xn + xn+1 = − 2 ω0 (∆t)2 (xn−1 + 4xn + xn+1 ) 6 Answer ¯ ¯ The total action S is a sum. CHAPTER 4 PROBLEMS 95 Answer In the interval tn ≤ t ≤ tn+1 the velocity is constant. in which xn occurs in two terms. xn+1 ) ≡ = tn+1 dt L tn xn+1 (xn+1 − xn )2 m∆t 2 dx − ω0 x2 2(xn+1 − xn ) xn (∆t)2 3 − x3 ) (xn+1 − xn )3 m∆t n 2 (x − ω0 n+1 = 2 2(xn+1 − xn ) (∆t) 3 2 2 ω ∆t m (xn+1 − xn ) − 0 (x2 + xn+1 xn + x2 ) (5. (5. with q = xn and Q = xn+1 .4. . after multiplying by −∆t/m. + S(xn−1 . xn ) ∂ S(xn . From eq. xn+1 ) + . ﬁnd the linear canonical transformation from (xn .60). pn ≡ p) to (xn+1 . xn ) + S(xn .61) ( 6 ¯ (c) Using the approximate action integral S(xn .5.. The condition for S to be stationary with respect to variations in xn is ∂S/∂xn = 0. xn+1 ) evaluated over the ¯ range tn < t < tn+1 . x = (xn+1 − xn )/∆t. This provides a discrete-time dynamical system approximation to the true continuous-time system for the harmonic oscillator. xn+1 ). Show that the determinant of the map is unity. xn+1 ) − + m ∂xn ∂xn 2 (ω0 ∆t) = −(xn − xn−1 ) + (2xn + xn−1 ) 6 (ω0 ∆t)2 + (xn+1 − xn ) + (xn+1 + 2xn ) 6 (ω0 ∆t)2 = xn−1 − 2xn + xn+1 + (xn−1 + 4xn + xn+1 ) = 0 2 . . as a type 1 generating function. 5. . pn+1 ≡ P ).60) . ˙ Thus the contribution to the action from this interval is given by ¯ S(xn . F1 (q. . this may be written ¯ ¯ ∆t ∂ S(xn−1 . Q) = −S(xn . .

Again. For another bonus mark calculate the eigenvalues for the three values of ∆t given above. p0 = 0 and plot the result.62) n 2 ∆t 3 Thus. Thus we have constructed a symplectic integrator —a discrete-time dynamical system that preserves some of the Hamiltonian character of the continuous-time dynamical system. which is unity and conserved under canonical transformation. 2 ∆t 3 ∂F1 = − ∂xn+1 2 m 2(xn+1 − xn ) ω0 ∆t = − (2xn+1 + xn ) . For ∆t = 0.e.47.17) the transformation is given by pn = ∂F1 ∂xn 2 m 2(xn+1 − xn ) ω0 ∆t = + (2xn + xn+1 ) . the generating function is F1 (xn .45 and ∆t = 3. (4. That this must be the case follows both from the general principal that canonical transformations preserve phase-space volume (or. ∆t = 3. 2 ∆t 3 pn+1 (5. area) and from the fact that the determinant in this case is also the Poisson bracket between q and p.60). I.63) For further analysis see the following Maple printout where it is veriﬁed that the transformation matrix has unit determinant.96 CHAPTER 5. For a bonus mark verify one of the solutions by showing that the secondorder diﬀerence equation derived in (b) above (when arranged to give xn+1 in terms of xn and xn−1 ) gives the same sequence of x-values as the discrete-time dynamical system in (c) if two successive values of x given by the dynamical system are used to start the second-order diﬀerence equation. see the following Maple printout. ANSWERS TO PROBLEMS Answer From eq. (5. xn+1 ) = − 2 m (xn+1 − xn )2 ω0 ∆t 2 − (xn+1 + xn+1 xn + x2 ) .1. (d) Iterate the map obtained in (c) 100 times taking the initial point as x0 = 1. in this case. (5.16) and (4. from eqs.1 the diﬀerence approximation gives results very close to the true continuous-time harmonic . Use units such that m = ω0 = 1 and use three timesteps: ∆t = 0. the discrete time “Lagrangian” and “Hamiltonian” discriptions are dynamically equivalent.

which is a circle in phase space. mutually reciprocal numbers.5. which means they are either complex-conjugate pairs on the unit circle in the complex plane (the stable case).47. which occurs somewhere between ∆t = 3. For large enough ∆t the iteration goes unstable. . This can be predicted by evaluating the eigenvalues of the transformation matrix—because the determinant is unity they must satisfy λ1 λ2 = 1. one of which must be greater than unity.45 and ∆t = 3. CHAPTER 4 PROBLEMS 97 oscillator solution. leading to instability (exponential growth).4. or they are real.

98 CHAPTER 5. ANSWERS TO PROBLEMS .

Bell. Sykes and J. 1982). 1980) by Herbert Goldstein. 3. 1990) by Florian Scheck.Chapter 6 References and Index SOME TEXTS on classical dynamics are listed below. The call number is TA350.L283 in the Hancock.G6 in the Hancock and Earth Sciences libraries.S9 in the Hancock. 5. Introduction to Dynamics (Cambridge University Press. Another classic. M. The most referred to text is undoubtedly Classical Mechanics (AddisonWesley. A good introductory book is I. 4. Chemistry and Physical Sciences libraries. The call number is QA805. tr. 1969) by L.R35 in the Hancock library. 1. text is Dynamics (Wiley. The call number is QA807. 1983) by S.A. more modern text is Mechanics : from Newton’s Laws to Deterministic Chaos (Springer. the authors have space to devote to the qualitative analysis of nonlinear dynamical systems. The call number is QA805. 2. Stromlo. 1959). Another interesting. Percival and D. New York. A good. 2nd ed. B. By limiting attention to systems with one degree of freedom. 1953.L. including 99 . Richards. 1960. The call number is QA845. Neil Rasband. Landau and E. although it does not have much on modern developments in nonlinear Hamiltonian dynamics. D. Synge and B. A clearly written text of similar vintage is Principles of Mechanics by J. Chemistry and Earth Sciences libraries. Griﬃths (McGraw-Hill. S. Lifshitz. though not as carefully written. concise text is Mechanics (Pergamon 1st ed. 1st ed. Stromlo. 2nd ed. 6. from the Russian by J.S35313 in the Physical Sciences library.

Vogtmann and A. 11. 1st ed. Lichtenberg and Michael A. It includes the Lagrangian approach and Hamilton’s Principle. The call number is QA845. translated from the Russian by K. including treatment of electrical circuits using Lagrangian methods. Classical Dynamics: a Contemporary Approach Jorge V.T. Wiley. 1983 called Regular and stochastic motion) by Allan J. Despite the use of the word “applied” in the title. See Physics Today May 1999. for a favourable review.J73 in the Hancock library. it is actually a very scholarly work and contains an unusual amount of thoughtful historical discussion.W62 Hancock. c1996) is a large text used in the ANU Engineering Dynamics course. Williams (New York. The call number is QA1.W68 in the Hancock library. REFERENCES AND INDEX such modern topics as chaos. Weinstein (Springer–Verlag. P. 1992. Arnol’d. but not Hamilton’s equations. . Lieberman. QA845. The call number is QC133. Jos´ and Eue gene J. 7. see Regular and Chaotic Dynamics (Springer. Whittaker (Cambridge University Press. p..P46 in the Hancock library. Cambridge UK. 1998) is a large modern text. An older mathematically oriented classic is A Treatise on the Analytical Dynamics of Particles and Rigid Bodies by E. A more mathematically oriented book by one of the modern grand masters in the ﬁeld is Mathematical Methods of Classical Mechanics by V. For a great deal of detail on nonlinear dynamics. 10. It is written in an interesting manner with many physical examples having a modern ﬂavour. 2nd ed. 9. Saletan (Cambridge U.100 CHAPTER 6.I. 66. 1927). written in a style easily accessible by physicists. 8. New York. The call number is QA805. 1980). Fundamentals of Applied Dynamics by James H. New York.A647 in the Hancock and Stromlo libraries.

46 canonical system. 73 commutator. 71 dynamics. 27 autonomous system. 44 harmonic oscillator. 75. 45. 31 angular momentum. 3. 48. 41 discrete time. 41 Atwood’s machine. 63 time evolution. 81 constraint. 30 equation of motion. 38. 37. 41 Hamiltonian. 37. 65 conﬁguration space. 26. 30 billiards. 79 cycle. 77 energy integral. 92 . 46 particle in compressible ﬂuid. 23. 70 dynamical system. 53. 37. 82 discretization. 15 ecliptic. 83 d’Alembert equations. 23. 89 asymptotic expansion. 16.Index N -body problem. 64. 51 action-angle coordinates. 6. 55 inﬁnitesimal. 29 particle in e.m. 26 coordinate-free formulation. 46. 24. 32. 84. 30. 3. ﬁeld. 37. 43. 93 curvature vector. 74. 3 canonical coordinates. 21 abuse of notation. 82. 37 anharmonic oscillator. 44 canonical transformation. 50 energy methods. 20. 20. 45 enthalpy. 25. 27. 36 constant of motion. 15. 93 action. 42 holonomic. 33 Coriolis force. 88. 50. 13. 41 cyclic coordinate. 66 auxiliary condition. phase space. 43. 24 action integral. 87. 94 distribution function. 62 attractor. 77. 67 chaos parameter. 68 adiabatic invariance. 44 canonical momentum. 4 time-varying. 33 diﬀerence equation. 64 celestial mechanics. 17. 31 action integral. 35. 21 degrees of freedom. 13. 25 Coriolis. 10 averaged Lagrangian. 80. 2 diﬀeomorphism. 41 continuous time. 86 101 nonholonomic. 28. 38. 67 cyclotron motion. 2 conservation equation. 5. 73 Hamiltonian.

47 Lie transform. 8 INDEX Hamilton’s equations of motion. 6. 50. 26. 8 ﬁeld. 41. 5 expontially small term. 23. 5 ﬁxed point. 1. ﬁeld. 48. 69 KAM theorem. 16 Fr´chet derivative. 45 particle in e. 2 generalized force. ﬁeld. 69. 71 kinematic momentum. 81 integrator. 7 e functional derivative. 66. 66. 8 inﬁnite-dimensional. 29 test-particle Lagrangian. 45 kinematics. 62 extremum condition. 25. Hamiltonian. 7 gauge potential. 1. 35.m. 64 integrator. 21. 47 stable. 8. 26. 10. 23 Lagrangian. 67 . 38. 51 Eulerian description. 45 libration. 44 Hamilton’s Principle. 18 constraint. 41 natural form. 29. 65 interchange of delta and dot. 65 radial. 57 geodesic. 26 ergodic hypothesis. 63 inner product. 44 particle in e. 24. 47 ﬂow. 16 applied. 29 force. 31. 18 explicit. 35 gauge transformation. 47 elliptic. 5. 49 ignorable coordinate. 35. 18. 29 Lagrangian description. 36. 73 kicked rotor. 15 Korteweg–de Vries equation. 91 Hamiltonian system (2n + 1)-dimensional. 31. 56 generalized coordinates. 30 integrability. 19 implicit. 6 ideal. 34. 1. 43. 64 harmonic oscillator. 27. 28 particle in ideal ﬂuid. 80 ideal equation of state. symplectic. 13. 35. 47 unstable. 68 Jacobian. 34. 51 Hamiltonian. 43. 5. 32. 5 Legendre transformation. 64 ﬂuid Euler. 43. 37. numerical. 56. 67 inﬁnitesimal generator.m. 46 Hessian matrix. 7 instantaneous frequency. 6. 9 curvature. 68 integral of motion. 23 friction. 17 generating function.102 Poisson-bracket form. 68 Lagrange multiplier. 42. 71 Euler angles. 47 hyperbolic. 50 Hamilton’s principle modiﬁed. 33. 15 Euler–Lagrange equation. 78 gradient deﬁnition. 46. 12 Lagrange’s equations. 28. 23. 16. 7 invariant tori.

62 Poisson bracket. 43. 30 103 reduced mass. 32. 51 variational derivative. 64 mass conservation. 44 phase-space Lagrangian. 45 generalized. 73 statistical mechanics. 33 resonance. 16. 50. 86 rotation. 41 virtual displacement. 42 objective functional. 69 rigid rod. 16 path. 48. 29 texts on classical dynamics.INDEX Liouville’s Theorem. 77. 47 secular behaviour. 10 particle. 69 total energy. 8 symmetry. 9 natural form. 18. 21 quantum mechanics. 25. 64 e point transformations. 57. 24 perturbation theory. 42 spurious solution. 21 two-body. 53. 64. 47 Poincar´ section. 66. canonical. 8 variational calculus. 85 variation constrained. 6. 46. 67 phase space. 49 potential N -body. 58 physical. 66 nonlinear Schr¨dinger equation. 9. 4. 65 polar coordinates. 6. 15. 41 period. 78 map area preserving. 17 virtual work. 46. 6. 16. 24 simple. 35. 45. 71 stroboscopic. 10. 68 separatrix. 7 vector ﬁeld. 47. 47 solubility condition. 87 trial function method. 35. 21 Newton’s third law. 61. 41. 20. 99 time reversal. 21. 3. 57 physical pendulum. 48 Maxwellian distribution. 62 quasiperiodic motion. 7 second. 50. 70 subsystem. 70 Lorentz force. 13. 48 Newton’s second law. 13. 25 natural boundary conditions. 87 support of a function. 71 iterated. 6 mass matrix. 10 ﬁrst. 30. 71 metric tensor. 68 o nullspace. 10 Standard Map. 13. 66 test particle. 24. 19 Noether’s theorem. 37. 28 manifold. 17 . 49 motion in a central potential. 55. 3. 39. 2. 24 pendulum physical. 87 reduction to ﬁrst-order ODEs. 42 representation. 3.