The Threebody Problem Valtonen M Karttunen H

download

https://ebookbell.com/product/the-threebody-problem-valtonen-m-
karttunen-h-2048332

Explore and download more ebooks at ebookbell.com

Here are some recommended products that we believe you will be
interested in. You can click the link to download.

The Threebody Problem Cixin Liu

https://ebookbell.com/product/the-threebody-problem-cixin-liu-56481884

The Threebody Problem Cixin Liu

https://ebookbell.com/product/the-threebody-problem-cixin-liu-56888538

The Threebody Problem And The Equations Of Dynamics Poincars

Foundational Work On Dynamical Systems Theory 1st Edition Henri
Poincar Auth

https://ebookbell.com/product/the-threebody-problem-and-the-equations-
of-dynamics-poincars-foundational-work-on-dynamical-systems-
theory-1st-edition-henri-poincar-auth-5882734

The Threebody Problem From Pythagoras To Hawking 1st Ed Mauri Valtonen

https://ebookbell.com/product/the-threebody-problem-from-pythagoras-
to-hawking-1st-ed-mauri-valtonen-6745112
The Threebody Problem Omnibus Cixin Liu

https://ebookbell.com/product/the-threebody-problem-omnibus-cixin-
liu-35112578

The Threebody Problem Trilogy Remembrance Of Earths Past Omnibus Cixin

Liu

https://ebookbell.com/product/the-threebody-problem-trilogy-
remembrance-of-earths-past-omnibus-cixin-liu-232240994

The Threebody Problem Remembrance Of Earths Past Cixin Liu

https://ebookbell.com/product/the-threebody-problem-remembrance-of-
earths-past-cixin-liu-35470160

The Threebody Problem Liu Cixin

https://ebookbell.com/product/the-threebody-problem-liu-cixin-11003028

The Dark Forest The Threebody Problem Liu Cixin

https://ebookbell.com/product/the-dark-forest-the-threebody-problem-
liu-cixin-35996694
 THE THREE-BODY PROBLEM

How do three celestial bodies move under their mutual gravitational attraction? This
problem was studied by Isaac Newton and other leading mathematicians over the
last two centuries. Poincaré’s conclusion, that the problem represents an example
of chaos in nature, opens the new possibility of using a statistical approach. For the
first time this book presents these methods in a systematic way, surveying statistical
as well as more traditional methods.

The book begins by providing an introduction to celestial mechanics, including

Lagrangian and Hamiltonian methods, and both the two- and restricted three-body
problems. It then surveys statistical and perturbation methods for the solution of
the general three-body problem, providing solutions based on combining orbit
calculations with semi-analytic methods for the first time. This book should be
essential reading for students in this rapidly expanding field and is suitable for
students of celestial mechanics at advanced undergraduate and graduate level.
THE THREE-BODY PROBLEM

MAURI VALTONEN AND HANNU KARTTUNEN

Väisälä Institute for Space Physics and Astronomy,
University of Turku
  
Cambridge, New York, Melbourne, Madrid, Cape Town, Singapore, São Paulo

Cambridge University Press

The Edinburgh Building, Cambridge  , UK
Published in the United States of America by Cambridge University Press, New York
www.cambridge.org
Information on this title: www.cambridge.org/9780521852241

© Cambridge University Press 2005

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

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

First published in print format 2005

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

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

- ---- hardback

- --- hardback

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

Preface page ix
1 Astrophysics and the three-body problem 1
1.1 About the three-body problem 1
1.2 The three-body problem in astrophysics 5
1.3 Short period comets 8
1.4 Binary stars 12
1.5 Groups of galaxies 15
1.6 Binary black holes 17
2 Newtonian mechanics 20
2.1 Newton’s laws 20
2.2 Inertial coordinate system 21
2.3 Equations of motion for N bodies 22
2.4 Gravitational potential 24
2.5 Constants of motion 25
2.6 The virial theorem 27
2.7 The Lagrange and Jacobi forms of the equations of motion 29
2.8 Constants of motion in the three-body problem 31
2.9 Moment of inertia 32
2.10 Scaling of the three-body problem 34
2.11 Integration of orbits 34
2.12 Dimensions and units of the three-body problem 38
2.13 Chaos in the three-body problem 39
2.14 Rotating coordinate system 43
Problems 45
3 The two-body problem 47
3.1 Equations of motion 47
3.2 Centre of mass coordinate system 48
3.3 Integrals of the equation of motion 49

v
vi Contents

3.4 Equation of the orbit and Kepler’s first law 52

3.5 Kepler’s second law 53
3.6 Orbital elements 54
3.7 Orbital velocity 57
3.8 True and eccentric anomalies 58
3.9 Mean anomaly and Kepler’s equation 60
3.10Solution of Kepler’s equation 61
3.11Kepler’s third law 63
3.12Position and speed as functions of eccentric anomaly 64
3.13Hyperbolic orbit 66
3.14Dynamical friction 68
3.15Series expansions 70
Problems 78
4 Hamiltonian mechanics 80
4.1 Generalised coordinates 80
4.2 Hamiltonian principle 81
4.3 Variational calculus 82
4.4 Lagrangian equations of motion 85
4.5 Hamiltonian equations of motion 87
4.6 Properties of the Hamiltonian 89
4.7 Canonical transformations 92
4.8 Examples of canonical transformations 95
4.9 The Hamilton–Jacobi equation 95
4.10 Two-body problem in Hamiltonian mechanics: two dimensions 97
4.11 Two-body problem in Hamiltonian mechanics: three dimensions 103
4.12 Delaunay’s elements 108
4.13 Hamiltonian formulation of the three-body problem 109
4.14 Elimination of nodes 111
4.15 Elimination of mean anomalies 113
Problems 113
5 The planar restricted circular three-body problem and other special cases 115
5.1 Coordinate frames 115
5.2 Equations of motion 116
5.3 Jacobian integral 119
5.4 Lagrangian points 123
5.5 Stability of the Lagrangian points 125
5.6 Satellite orbits 130
5.7 The Lagrangian equilateral triangle 133
5.8 One-dimensional three-body problem 136
Problems 139
 Contents vii

6 Three-body scattering 141

6.1 Scattering of small fast bodies from a binary 141
6.2 Evolution of the semi-major axis and eccentricity 148
6.3 Capture of small bodies by a circular binary 152
6.4 Orbital changes in encounters with planets 154
6.5 Inclination and perihelion distance 157
6.6 Large angle scattering 162
6.7 Changes in the orbital elements 165
6.8 Changes in the relative orbital energy of the binary 169
Problems 170
7 Escape in the general three-body problem 171
7.1 Escapes in a bound three-body system 171
7.2 A planar case 179
7.3 Escape velocity 180
7.4 Escaper mass 183
7.5 Angular momentum 184
7.6 Escape angle 188
Problems 195
8 Scattering and capture in the general problem 197
8.1 Three-body scattering 197
8.2 Capture 203
8.3 Ejections and lifetime 207
8.4 Exchange and flyby 211
8.5 Rates of change of the binding energy 214
8.6 Collisions 216
Problems 219
9 Perturbations in hierarchical systems 221
9.1 Osculating elements 221
9.2 Lagrangian planetary equations 222
9.3 Three-body perturbing function 225
9.4 Doubly orbit-averaged perturbing function 227
9.5 Motions in the hierarchical three-body problem 231
Problems 239
10 Perturbations in strong three-body encounters 240
10.1 Perturbations of the integrals k and e 240
10.2 Binary evolution with a constant perturbing force 243
10.3 Slow encounters 246
10.4 Inclination dependence 260
10.5 Change in eccentricity 264
10.6 Stability of triple systems 268
viii Contents

10.7 Fast encounters 274

10.8 Average energy exchange 281
Problems 285
11 Some astrophysical problems 288
11.1 Binary black holes in centres of galaxies 288
11.2 The problem of three black holes 296
11.3 Satellite black hole systems 310
11.4 Three galaxies 310
11.5 Binary stars in the Galaxy 313
11.6 Evolution of comet orbits 320
Problems 327
References 329
Author index 341
Subject index 343
 Preface

Classical orbit calculation in Newtonian mechanics has experienced a renaissance

in recent decades. With the beginning of space flights there was suddenly a great
practical need to calculate orbits with high accuracy. At the same time, advances in
computer technology have improved the speed of orbit calculations enormously.
These advances have also made it possible to study the gravitational three-body
problem with new rigour. The solutions of this problem go beyond the practicalities
of space flight into the area of modern astrophysics. They include problems in the
Solar System, in the stellar systems of our Galaxy as well as in other galaxies. The
present book has been written with the astrophysical applications in mind.
The book is based on two courses which have been taught by us: Celestial
Mechanics and Astrodynamics. The former course includes approximately Chapters
2–5 of the book, with some material from later chapters. It is a rather standard
introduction to the subject which forms the necessary background to modern topics.
The celestial mechanics course has been developed in the University of Helsinki
by one of us (H. K.) over about two decades. The remainder of the book is based
on the astrodynamics course which arose subsequently in the University of Turku.
Much of the material in the course is new in the sense that it has not been presented
at a textbook level previously.
In our experience there has been a continuous need for specialists in classical
orbit dynamics while at the same time this area of study has received less attention
than it used to in the standard astronomy curriculum. By writing this book we hope
to help the situation and to attract new students to the research area, which is still
modern after more than 300 years of studies.
We have been privileged to receive a great deal of help and encouragement
from many colleagues. Especially we would like to thank Douglas Heggie, Kimmo
Innanen and Bill Saslaw who have between them read nearly the whole manuscript
and suggested numerous improvements. We also appreciate the comments by Victor
Orlov, Harry Lehto and Tian-Yi Huang which have been most useful. Seppo Mikkola

ix
x Preface

has generously provided research tools for the calculations in this book. Other
members of the Turku research group have also helped us with illustrations. Most
of all, we would like to thank Sirpa Reinikainen for typing much of the final text,
including the great many mathematical formulae.
Financial support for this project has been provided by Finland’s Society for
Sciences and Letters and the Academy of Finland (project ‘Calculation of Orbits’),
which gave the opportunity for one of us (M. V.) to concentrate on writing the book
for a period of two years. The generous support by the Department of Computer
Science, Mathematics and Physics (in Barbados) and the Department of Physics
(in Trinidad) of the University of the West Indies made it possible to carry out
the writing in optimal surroundings. Parts of the text were originally published in
Finnish: H. Karttunen, Johdatus Taivaanmekaniikkaan, Helsinki: Ursa, 2001.
Finally, M. V. would like to express his appreciation to Sverre Aarseth who taught
him how to calculate orbits (and much else), and to his wife Kathleen, the Caribbean
link, whose encouragement was vital for the accomplishment of the book.
 1
Astrophysics and the three-body problem

1.1 About the three-body problem

The three-body problem arises in many different contexts in nature. This book deals
with the classical three-body problem, the problem of motion of three celestial
bodies under their mutual gravitational attraction. It is an old problem and logically
follows from the two-body problem which was solved by Newton in his Principia
in 1687. Newton also considered the three-body problem in connection with the
motion of the Moon under the influences of the Sun and the Earth, the consequences
of which included a headache.
There are good reasons to study the three-body gravitational problem. The motion
of the Earth and other planets around the Sun is not strictly a two-body problem.
The gravitational pull by another planet constitutes an extra force which tries to
steer the planet off its elliptical path. One may even worry, as scientists did in the
eighteenth century, whether the extra force might change the orbital course of the
Earth entirely and make it fall into the Sun or escape to cold outer space. This was a
legitimate worry at the time when the Earth was thought to be only a few thousand
years old, and all possible combinations of planetary influences on the orbit of the
Earth had not yet had time to occur.
Another serious question was the influence of the Moon on the motion of the
Earth. Would it have long term major effects? Is the Moon in a stable orbit about the
Earth or might it one day crash on us? The motion of the Moon was also a question
of major practical significance, since the Moon was used as a universal time keeping
device in the absence of clocks which were accurate over long periods of time. After
Newton, the lunar theory was studied in the eighteenth century using the restricted
problem of three bodies (Euler 1772). In the restricted problem, one of the bodies
is regarded as massless in comparison with the other two which are in a circular
orbit relative to each other. At about the same time, the first special solution of the
general three-body problem was discovered, the Lagrangian equilateral triangle

1
2 Astrophysics and the three-body problem

solution (Lagrange 1778). The theory of the restricted three-body problem was
further developed by Jacobi (1836), and it was used for the purpose of identifying
comets by Tisserand (1889, 1896) and reached its peak in the later nineteenth
century with the work of Hill (1878) and Delaunay (1860). The ‘classical’ period
reached its final phase with Poincaré (1892–1899).
In spite of these successes in special cases, the solution of the general three-body
problem remained elusive even after two hundred years following the publication
of Principia. In the general three-body problem all three masses are non-zero and
their initial positions and velocities are not arranged in any particular way. The
difficulty of the general three-body problem derives from the fact that there are no
coordinate transformations which would simplify the problem greatly. This is in
contrast to the two-body problem where the solutions are found most easily in the
centre of mass coordinate system. The mutual force between the two bodies points
towards the centre of mass, a stationary point in this coordinate system. Thus the
solution is derived from the motion in the inverse square force field. Similarly, in
the restricted three-body problem one may transfer to a coordinate system which
rotates at the same rate around the centre of mass as the two primary bodies. Then
the problem is reduced to the study of motions in two stationary inverse square
force fields. In the general problem, the lines of mutual forces do not pass through
the centre of mass of the system. The motion of each body has to be considered
in conjunction with the motions of the other two bodies, which made the problem
rather intractable analytically before the age of powerful computers.
At the suggestion of leading scientists, the King of Sweden Oscar II established
a prize for the solution of the general three-body problem. The solution was to
be in the form of a series expansion which describes the positions of the three
bodies at all future moments of time following an arbitrary starting configuration.
Nobody was able to claim the prize for many years and finally it was awarded in
1889 to Poincaré who was thought to have made the most progress in the subject
even though he had not solved the specific problem. It took more than twenty years
before Sundman completed the given task (Sundman 1912). Unfortunately, the
extremely poor convergence of the series expansion discovered by Sundman makes
this method useless for the purpose of calculating the orbits of the three bodies.
Now that the orbits can be calculated quickly by computer, it is quite obvious why
this line of research could not lead to a real solution of the three-body problem: the
orbits are good examples of chaos in nature, and deterministic series expansions
are utterly unsuitable for their description. Poincaré was on the right track in this
regard and with the current knowledge was thus a most reasonable recipient of the
prize.
At about the same time, a new approach began which has been so successful in
recent years: the integration of orbits step by step. In orbit integration, each body,
 1.1 About the three-body problem 3

P1 (1, 3)
m1 = 3

P2 (−2, −1) P3 (1, −1)

m2 = 4 m3 = 5

Figure 1.1 Initial configuration of the Pythagorean problem.

in turn, is moved forward by small steps. In the most basic scheme, the step is
calculated on the basis of the accelerations caused by the two other bodies during
that step while they are considered to remain fixed. There is an error involved when
only one body is moved at a time, and others move later, but this can be minimised
by taking short steps and by other less obvious means.
Burrau (1913) considered a well defined, but in no way special, initial configu-
ration of three bodies which has since become known as the Pythagorean problem
since the three bodies are initially at the corners of a Pythagorean right trian-
gle. The masses of the three bodies are 3, 4 and 5 units, and they are placed
at the corners which face the sides of the triangle of the corresponding length
(Fig. 1.1). In the beginning the bodies are at rest. Burrau’s calculation revealed
the typical behaviour of a three-body system: two bodies approach each other,
have a close encounter, and then recede again. Subsequently, other two-body en-
counters were calculated by Burrau until he came to the end of his calculating
capacity. Only after the introduction of modern computers and new orbit integra-
tion methods was the celestial dance in the Pythagorean problem followed to its
conclusion.
Later work has shown that the solution of the Pythagorean problem is quite typical
of initially bound three-body systems. After many close two-body approaches, a
configuration arises which leads to an escape of one body and the formation of
a binary by the other two bodies (see Fig. 1.2, Szebehely and Peters 1967). A
theoretical treatment of a three-body system of this kind is given in Chapter 7. In
the following chapter, situations are discussed where a third body comes from a
large distance, meets a binary, and perhaps takes the place of one of the binary
members which escapes. Such orbits were calculated already in 1920 (Becker
1920). Sometimes the third body is always well separated from the binary; then the
situation is best described by perturbations on the binary caused by the third body.
Some examples of these systems are discussed in Chapter 10.
4 Astrophysics and the three-body problem
6 6

4 4

2 2

0 0

-2 -2

-4 -4

-6 -6
-6 -4 -2 0 2 4 6 -6 -4 -2 0 2 4 6
t = 40 t = 60
6 6

4 4

2 2

0 0

-2 -2

-4 -4

-6 -6
-6 -4 -2 0 2 4 6 -6 -4 -2 0 2 4 6
t = 20 t = 70
6 6

4 4

2 2

0 0

-2 -2

-4 -4

-6 -6
-6 -4 -2 0 2 4 6 -6 -4 -2 0 2 4 6
t = 10 t = 80

Figure 1.2 Trajectories of the Pythagorean problem. The orbits up to time 10, 20,
40, 60, 70 and 80 units are shown. The last two panels are identical since the escape
has already happened and the bodies are outside the frame.

In recent years there has been increasing demand for solutions of the general
three-body problem in various astrophysical situations. For example, binaries and
their interactions with single stars play a major role in the evolution of star clusters
(Aarseth 1973). Triple stellar systems are another obvious astrophysical three-body
 1.2 The three-body problem in astrophysics 5

problem. Many other astrophysical bodies, ranging from compact bodies to galax-
ies, occur in triple systems. These will be discussed in Chapter 11.
Before then, we have to do some preparatory work which is partly found in
standard textbooks on mechanics and celestial mechanics. The proper discussion
of the three-body problem is thus delayed until Chapter 5 where special cases are
first introduced. This leads to consideration of the general three-body problem in
the later chapters, the main objective of this book.

1.2 The three-body problem in astrophysics

For more than 300 years there have been many different motivations to solve the
three-body problem and many different techniques have been applied to it. In this
book we have the rather limited purpose of looking at solutions of astrophysi-
cal significance. At the present time we can solve any given three-body problem,
starting from the known positions, velocities and masses of the three bodies, by
using a computer. There is of course the limitation of the accuracy of calculation
which may be quite significant in some cases. But notwithstanding the accuracy, the
solution of an astrophysical problem usually involves much more than a calculation
of a single orbit. Typically we have to sample three-body orbits in a phase space of
up to eleven dimensions. Then the calculation of orbits is only one tool; one has to
have a deeper understanding of the three-body process to make sense of the limited
amount of information that is derivable from orbit calculations.
Therefore we do not deal with the mathematical three-body problem. Fortunately,
there are excellent books by Marchal (1990) and Hénon (1997, 2001) which deal
with the mathematical aspects very thoroughly. As an example, periodic orbits of
the general three-body problem are of great mathematical interest, but there are
very few examples where they are important in astrophysics.
In problems of astrophysical importance, one may almost always identify a
binary and a third body. A binary can be treated as a single entity with certain
‘internal’ properties (like a molecule). It is described by its component masses,
by its energy and angular momentum, as well as by its orientation in space. This
binary entity interacts with a third body once, or more frequently, which changes the
internal properties of the binary. At the same time, the third body absorbs whatever
energy is given out from the binary, in order to conserve the total energy. Similarly,
conservation of angular momentum between the binary and the third body has to
be satisfied.
Before we can take up the discussion of the three-body problem, we have to
be familiar with binaries, i.e. the two-body problem. The two-body problem is
treated in basic courses of mechanics and celestial mechanics. Therefore the dis-
cussion of Chapter 3 may appear as unnecessary repetition to some readers, and they
6 Astrophysics and the three-body problem

may want to skip much of the first chapters. Also the Hamiltonian techniques of
Chapter 4 are commonly studied in courses of advanced mechanics. They form
such an essential part of the treatment of weak perturbations of binaries that it is
necessary to introduce the Hamiltonian concepts also in this book.
A planet going around the Sun is an example of a binary. The third body could then
be another planet, a moon, an asteroid or a comet. Because of the large differences
between the masses of the bodies, from the dominant Sun down to asteroids of
kilometre size or even smaller, Solar System dynamics is special in many ways. A
very up to date treatment exists in this area (Murray and Dermott 1999). We discuss
only a small class of Solar System problems which are related to stellar dynamics
and therefore form a suitable introduction to later studies of three-body systems
with more equal masses. Thus Chapter 5 contains many topics which readers may
have encountered earlier.
At the present time, three-body astrophysics is primarily motivated by the need
to understand the role of binaries in the evolution of stellar systems. For most of the
time, a binary acts just like a single star in a stellar system. The distances between
stars are large compared with the sizes of the stars and even compared with the
sizes of close binary orbits. For a relatively brief moment a binary and a third star
interact strongly, a ‘new’ binary forms, and a ‘new’ third star leaves the scene.
The importance of the process lies in its ability to redistribute energy and angular
momentum efficiently; the population of binaries may become more and more
tightly bound as time and three-body scattering go on, while the population of single
stars may gain speed and become ‘heated’. This will have profound consequences on
the structure and evolution of a star cluster; for example single stars and sometimes
also binaries escape from the cluster; binary orbits may shrink to form contact
binaries, and also triple stars may form where the third body remains bound to
the binary. The end products of the three-body process may appear as sources of
radio jets, X-rays, gamma rays or as other kinds of ‘exotic’ objects (Hut et al.
2003).
These various scenarios can be reproduced by numerical orbit calculations. The
orbits of thousands of stars can be calculated in a simulation of a star cluster.
Even though these simulations have now reached a great level of complication
and trustworthiness (e.g. Heggie and Hut 2003, Aarseth 2003a), it is still useful to
examine the three-body process to see how much is understood from elementary
principles. Together with the simulations of large numbers of bodies one may attain
a deeper understanding of the evolutionary process.
Chapter 7 starts with the discussion of initially strongly interacting three-body
systems. We will learn that such systems have a limited lifetime, and we do not
expect to find very many of them in nature. But they are important in the descrip-
tion of the intermediate state between the impact of a third body on a binary and
 1.2 The three-body problem in astrophysics 7

the departure of a ‘new’ third body from a ‘new’ binary. The techniques used in
Chapter 7, assuming a complete reshuffle of the positions and velocities of the three
bodies in the phase space, work surprisingly well.
By ‘work well’ we mean that a large body of numerical orbit calculations can
be described in a statistical sense by simple physical principles. That this should
be so is not at all obvious at the outset. Therefore we devote considerable space
to comparing numerical calculations with the theory. The beauty of it is that once
the theory is established for certain parts of the parameter space, we have good
reasons to expect that it applies more generally. Also we do not need to go back to
calculating millions of orbits when a slightly different astrophysical problem arises
but we can use the theory directly with a fair amount of confidence.
In addition to these practical considerations, it gives the reader a certain satisfac-
tion to learn that simple analytical solutions of the three-body problem exist, even
though only in a statistical sense and for a limited part of the parameter space. It
will also become clear that these are the only solutions of any significance in large
parts of the phase space due to the chaotic nature of the problem.
Chapters 9 and 10 try to cover the remaining parameter space, i.e. when the
binary and the third body are so well separated throughout the interaction that the
principle of complete chaos is not productive. At one extreme there is the very slow
and gentle perturbation of the binary known as the Kozai mechanism. Then only
the binary eccentricity and inclination change periodically while the orbital sizes
are unaffected. At the other extreme we have a high speed intruder which gives the
binary a ‘shock treatment’ during its brief encounter with the binary. In between, a
binary is strongly perturbed at close encounters but is able to maintain its identity
and not break up or exchange members. A stability boundary is derived which tells
us where the perturbation treatment ends and the chaos theory begins.
In this way we can give a rather complete coverage of the astrophysical
three-body problem. As in the case of the chaos theory of Chapters 7 and 8, also
in the perturbation theory of Chapters 9 and 10, a great deal of space is dedicated
to comparing numerical results with the theory. This is necessary since it is not
always obvious, in the absence of exact theory, what approximations should lead to
the best understanding of the experiments. Often we even find that in the final steps
we just have to accept the guidance of the numerical experiments without clear
justification of the theory. This is not because the theories could not be pushed any
further but more because we like to keep the theory at a rather simple level (and it
may appear quite complicated to some readers already as it is). But also we have to
remember that in the general three-body problem with strong interactions no exact
theory exists, and we should not spend too much effort towards this elusive goal.
Throughout the development of the theory we will look at some small astrophys-
ical problems which are easily solved at this stage. In the final chapter a couple of
8 Astrophysics and the three-body problem

larger issues are discussed which require solutions from different parts of the text
as well as other astrophysical information. To give the reader an idea of what sorts
of problems we are dealing with we outline a couple of astrophysical examples in
the next sections.

1.3 Short period comets

The origin of short period comets is one of the oldest three-body problems. Lexell
(1778, 1779) studied the motion of the comet found in 1770 by Messier, and sug-
gested that the orbit had become elliptical with a period of 5.6 years, when the comet
passed close to Jupiter. Later, Laplace (1799–1825, 1805) and Leverrier worked on
the capture hypothesis. Tisserand (1889) and H. A. Newton (1891), among others,
discussed details of the capture process. The idea, already put forward by Lexell,
was that short period comets are created from long period comets which pass near
Jupiter and lose energy during the encounter. Everhart (1969) carried out a major
survey of close encounter orbits between comets and planets using computers. In
spite of all these and later efforts, the origin of the short period comets is still an
open question to a large extent.
There are more than two hundred known short period comets, even though it is
well documented that comets fade away after 102 –103 revolutions around the Sun
and that they escape the solar system after 104 –105 revolutions (Fig. 1.3). There
must be a source of comets which constantly (or from time to time) replenishes the
population. Two such sources have been suggested: the Oort Cloud of comets and
the Kuiper Belt of asteroids and comets. Sometimes also other source regions, such
as interstellar comets, have been mentioned. The processes which may keep the
short period comet population intact are mostly related to the three-body problem.
The Oort Cloud (Oort 1950) is a collection of as many as 1012 comets loosely
bound to the Solar System. The orbits of the comets are such that they generally
do not enter the planetary region, and their orbits are mainly affected by the Sun,
passing stars, gas clouds and the tidal field of the Galaxy as a whole. As a result of
these influences, there is a more or less steady flux of ‘new’ comets which enter the
planetary system for the first time. The flux may be rather uniform per pericentre
interval (closest approach distance to the Sun, in AU) up to the distance of Jupiter;
beyond that the flux is expected to rise but the difficulty of observing distant comet
passages prevents observational confirmation of this expected trend.
Since we have some idea of the Oort Cloud flux of comets we may ask how much
of this flux is captured to short period orbits and what are the orbital properties of
the captured comets. An Oort Cloud comet comes in a highly eccentric orbit, and in
the absence of planets it would return back to the Oort Cloud in the same elliptical
orbit. But the orbit is influenced by one or more planets, and this influence is likely
 1.3 Short period comets 9

Figure 1.3 Halley’s comet, the most famous periodic comet, photographed in May
13, 1910. The big round object is Venus and the stripes are city lights of Flagstaff.
Image Lowell Observatory.

to change the orbit either to a hyperbolic escape orbit or to a more strongly bound
short period elliptical orbit. Basically we need to solve the three-body problem
consisting of the Sun, a planet and a comet. Since the comet has much smaller
mass than the other two bodies, the problem is restricted only to the question of
the motion of the comet. Generally we may assume that the planet goes around the
Sun in a circular orbit, and remains in this orbit independent of what happens to
the comet. This is an example of the restricted circular three-body problem which
will be discussed in Chapter 5.
It is a straightforward procedure to make use of an orbit integrator, a computer
code which calculates orbits, and to calculate the orbit of an Oort Cloud comet
through the Solar System, past various planets, perhaps through millions of orbital
cycles, until the comet escapes from the Solar System, or until it collides with the
Sun or one of the planets, or until the comet disintegrates. The calculation can be
long but it is possible with modern computers. However, questions remain: how
representative is this orbit and how accurate is the solution? Indeed, we do not know
the exact starting conditions for any Oort Cloud comets and neither can we carry
out the calculation over the orbital time of millions of years without significant loss
of accuracy.
To some extent these problems can be avoided by using the method of sampling.
We take a large sample of orbits with different initial conditions and calculate them
through the necessary period of time. We may then look at observational samples
10 Astrophysics and the three-body problem

and compare them with samples obtained by orbit calculation. In practice we need
millions of orbits in order to get satisfactory statistics of the captured short period
comets. It becomes a major computational challenge.
We can learn quite a lot by studying the interaction of a comet with only one
of the planets. A typical Oort Cloud comet has a high inclination relative to the
plane of the Solar System. It comes from one side of the plane (say, above), dives
through the plane, turns around below the plane, and crosses the plane again. A
strong three-body interaction takes place only if a planet happens to be close to
one of the crossing points at the right time. Most likely no planet is there at all, but
when a close encounter happens, almost always it is only with one planet only at
one of the two crossing points. Therefore we have a three-body problem.
But even then the exact nature of the three-body encounter is unclear. The best
we can do is to develop a statistical theory of how the comet is likely to react to
the presence of the planet close to the crossing point. Öpik (1951) developed such
a theory where the comet is assumed to follow an exact two-body orbit around the
Sun until it comes to the sphere of gravitational influence of the planet. At this
point the comet starts to follow an exact two-body orbit relative to the planet. After
leaving the sphere of influence of the planet, the orbit is again a (different) two-body
orbit around the Sun. A theory along these lines will be discussed in Chapter 6,
together with some more recent work. It is the most basic form of a solution to
the three-body problem. Notice that we will be discussing probabilities; this is the
recurring theme of the solutions of astrophysical three-body problems.
The statistical properties of comets which we will have to confront are primarily
the distribution of their orbital sizes (semi-major axes), their perihelion distances
and the distribution of orbital inclinations relative to the Solar System plane. In
terms of orbital sizes, the comets can be classified as being either Jupiter family
comets (orbital period below 20 years), Halley type comets (period above 20 but
below 200 years) or long period comets (longer periods than 200 years). Oort Cloud
comets have orbital periods in excess of a million years. Within these groups, the
inclination distributions vary (Fig. 1.4). One of the aims of a successful theory of
short period comets is to explain their observed numbers (28 Halley type, 183 Jupiter
family comets; Marsden and Williams 1999) in relation to the rate of comets coming
from the Oort Cloud. Also it should explain how the differences in the inclination
distributions originate.
If it turns out that a decrease in the semi-major axis of a comet is associated
with a decrease in its inclination (and this will be shown in Chapter 6) then we
might propose that comets evolve from Oort Cloud comets to Halley type comets
to Jupiter family comets by successively decreasing the semi-major axis and the
inclination of the orbit. How this works out in detail will be discussed in Chapter 11.
 1.3 Short period comets 11
Inclinations in comet families
45

40
Jupiter family
35

30

number
25 Oort Cloud

20

15

10

5 Halley types

0
0 20 40 60 80 100 120 140 160 180
inclination

Figure 1.4 The inclination distributions in three classes of comets: Oort Cloud
comets, the orbits of which agree with random orientations (number proportional
to sin i), Halley type comets which are more often in direct orbits than in retrograde
orbits, and Jupiter family comets which are in direct orbits close to the Solar System
plane.

Another idea which is as old as the Oort Cloud theory is the concept of a disk
of comets left over from the formation of the Solar System. This is usually called
the Kuiper Belt or Edgeworth–Kuiper Belt (Edgeworth 1949, Kuiper 1951). These
comets reside mostly by the orbit of Neptune or beyond (Transneptuneans). Comets
do not develop a tail at this distance from the Sun and they can be observed only if
the nucleus of the comet (the solid body from which the tail originates) is quite large,
greater than about 100 km in diameter. Therefore our knowledge of the comets at the
Kuiper Belt is limited to the very largest bodies among them. By extrapolation it has
been deduced that there may exist as many as 107 ‘ordinary’ comets in the Kuiper
Belt and in its vicinity (Levison and Duncan 1997), even though only somewhat
over 300 very bright ones are known. Some of the Transneptunean comets certainly
approach the Sun at some point in their orbital evolution and become visible by
their bright tails; however, it is difficult to know exactly which comets have this
origin.
The Oort Cloud comets are very loosely bound to the Solar System and therefore
the Oort Cloud may require replenishment. A more tightly bound and dynamically
more robust cloud of comets has been suggested to lie inside the Oort Cloud, the
so-called Inner Oort Cloud (Hills 1981). Numerical simulations of the origin of
the Oort Cloud by ejection of cometary bodies from the region of the outer planets
also suggest the existence of the Inner Oort Cloud (Duncan et al. 1987). Passing
stars and molecular clouds perturb the orbits of the Inner Oort Cloud comets and
12 Astrophysics and the three-body problem

Figure 1.5 In the course of years the components of the binary star Krüger 60 are
seen to orbit their common centre of mass. Image Yerkes Observatory

some of them end up in the Oort Cloud proper. The Inner Oort Cloud is yet another
potential source region for short period comets.

1.4 Binary stars

About half of the stars in our Galactic neighbourhood are binary stars (Fig. 1.5).
Binary stars have been studied since the eighteenth century when William Herschel
discovered that stars may appear in physical pairs. Combined with Newton’s laws
of gravity and dynamics, binary stars have been used in the studies of stellar masses,
radii and other properties. Yet even today, we do not have a full understanding of
the origin of even the most basic properties of the binaries.
The study of binary stars is an interesting field in its own right but it also has
wider implications. Binaries appear as mass points when seen from far away, but
in close encounters with single stars or other binaries their internal motions may
release or absorb energy. Therefore they behave somewhat similarly to molecules
in gas dynamics. The role of binaries is very crucial in understanding the dynamics
of stellar systems.
Binaries may have very different semi-major axes and orbital eccentricities,
ranging from two stars in contact with each other and in a circular orbit, to a
wide and typically eccentric binary with a semi-major axis of the order of 0.1 pc.
Corresponding orbital periods range from a fraction of a day to millions of years.
One of the basic questions in binary star studies is to understand these distributions:
the distributions of binary periods, of binary orbit eccentricities and of mass ratios
of the binary components.
 1.4 Binary stars 13
Binary star periods per log interval
1.4
1.2

log[dN/d(logP)]
1
0.8
0.6
0.4
0.2
0
-2 -1 0 1 2 3 4 5 6
log(P/yr)
Figure 1.6 The distribution of binary star periods P, plotted per logarithmic inter-
val d log P (Abt 1977). The horizontal dashed line represents the so-called Öpik
law.

Because of the wide range of physical sizes of binary systems, studies of their
distributions require different methods on different scales. Various selection effects
also complicate these studies. When all the studies on different scales and with
corrections for selection effects have been combined, the following results have
been obtained. The period distribution is flat over five orders of magnitude when
plotted per logarithmic interval of the period P (Fig. 1.6; Abt 1977). A distribution of
this type is sometimes called ‘scale free’. Written in the usual way the distribution is

f (P) dP ∝ P −1 dP (1.1)

which corresponds to

f (|E B |) d|E B | ∝ |E B |−1 d|E B | (1.2)

in terms of the binary orbital energy E B . This form of binary energy distribution
is known as Öpik’s law (Öpik 1924). One of our aims is to explain the origin of
this law; we will come back to it in Chapter 11.
The distribution of orbital eccentricities e depends on the period: short period
orbits have low eccentricities while long period orbits are more eccentric with the
weight of the distribution at the high end of the range of e (Fig. 1.7; Duquennoy
and Mayor 1991). In the latter case the data are consistent with

f (e) de = 2e de, (1.3)

an equilibrium distribution in stellar dynamics (Jeans 1919, Ambartsumian 1937).

Short period orbits are of low eccentricity (Campbell 1910). For contact binaries it
is easy to understand why the orbit is circular: tidal friction between the two stars
14 Astrophysics and the three-body problem
Binary eccentricities
12
P > 3 yr
8
Number

4
P < 3 yr
0
0 0.3 0.6 0.9
e
Figure 1.7 The observed distribution of eccentricities in a sample of binary stars.
The data for binaries of longer orbital periods (over 3 years) are shown by points
with error bars while the data for binaries of short orbital periods (less than 3 years)
are indicated by the broken line (Duquennoy and Mayor 1991). The straight dotted
line represents a possible interpretation of the long orbital period data.

operates most strongly during the pericentre of the orbit, when the two stars are
closest to each other. The resulting loss of orbital speed at the pericentre leads to
a rather circular orbit. But the low eccentricity orbits are dominant even among
binaries whose orbital size is as large as one astronomical unit (AU), the distance
between the Earth and the Sun. Tidal friction does not play any role at these sepa-
rations since the radius of the Sun (and a typical main sequence star) is only about
0.005 AU. We have to look for other explanations for the different eccentricity
distributions.
The mass ratio distribution depends on the type of the primary star (Abt 1983).
The most massive stars (primaries of spectral type O) tend to be in rather equal
mass pairs, m 2 /m 1 >
∼ 0.3 while the reverse is true for less massive upper main
sequence stars. For the low mass, solar type primaries the mass ratio peaks at about
m 2 /m 1 = 0.25 (Duquennoy and Mayor 1991).
To some extent the differences can be understood by the form of the mass dis-
tribution function for single stars:

f (m) ∝ m −α (1.4)

where α = 2.35 over most of the range (Salpeter 1955). If the more massive com-
ponent is taken to be m 1 = 1, then the distribution for m 2 should follow Eq. (1.4),
and the same should also apply to the mass ratio m = m 2 /m 1 . However, obser-
vations give a less steep function. For example, Kuiper (1935a, b) found that
f (m) = 2(1 + m)−2 gives a good fit to data on close binaries while Trimble (1990)
prefers f (m) ∝ m −1 . Thus the masses of the binary components are more equal
 1.5 Groups of galaxies 15
Mass ratios in binary stars
1.8
1.6
1.4
log(number) 1.2
1
0.8
0.6
0.4
0.2
0
0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
m_2/m_1
Figure 1.8 The distribution of the mass ratio m 2 /m 1 (m 2 < m 1 ) in a B spectral
type binary star sample (points with error bars; Evans 1995). In comparison, the
dashed line gives the expected distribution if the companion of the B-type star has
been picked at random among ordinary stars lighter than the primary.

than we would predict by picking a companion for the primary at random from the
single star mass function (Fig. 1.8; Evans 1995).
On the other hand, the mass function for low mass stars (below the mass of the
Sun) has a lower power-law index α  1.25. Therefore the mass ratio distribution is
also expected to be less steep, f (m) ∝ m −1.25 . This agrees with the observed flatter
distribution of m for solar type primaries. Quantitative comparisons are difficult
because the single star mass function below about 0.5 solar mass is poorly known.
At the other end, the mass function for O-type stars is steeper than Eq. (1.4)
with α = 2.35. A power-law index α = 3.2 may be used at the high mass end,
above m  10M (Mihalas and Binney 1981). This leads to an even steeper mass
ratio distribution, in contrast with observations which go in the opposite direction.
Thus there is much more to the mass ratio distribution than simply picking a pair at
random. Even if the pairs might have formed like this initially, there has obviously
been strong evolution which has modified the mass ratios. We will discuss the
three-body process in Chapter 11 as a possible cause of this evolution.

1.5 Groups of galaxies

Just like stars, galaxies also appear in binaries, triples, quadruples and other small
groups, as well as in large groups and clusters (Fig. 1.9). Binaries and their
interactions with third galaxies may have consequences for the evolution of these
systems. There is also important information about galaxies themselves which
16 Astrophysics and the three-body problem

Figure 1.9 Coma Cluster of galaxies. Picture taken with the KPNO 4-m Mayall
Telescope. Image NOAO/AURA/NSF.

requires an understanding of the multiple systems. First and foremost, the total
masses of galaxies, including their dark halos, are best derived from the study of
multiple systems.
The method which is commonly used is to assume that the multiple system
is in a bound state, i.e. that none of the members is currently escaping from the
grouping. In addition, some assumptions are made about the nature of the orbits
(e.g. circular, radial or intermediate) as well as about projections onto the plane of
the sky. With these assumptions it is possible to connect observed properties, i.e.
projected separations and radial velocities with inferred quantities such as the total
mass of the system.
The total amount of light emitted by all the member galaxies is also measurable.
Then the calculated mass (in solar mass) may be divided by the measured light
output (in solar luminosity), and the resulting number is called the mass-to-light
ratio. It is usually assumed that it is rather narrowly confined to a value around
30–50, at least for spiral galaxies. It is still an unsolved problem in astronomy as to
how frequently galaxies in a complete volume sample fall into this range of mass-to-
light ratio, and to what extent there is variation from one class of galaxies to another.
 1.6 Binary black holes 17

The existence of binary galaxies complicates matters. Their interaction with

third galaxies may lead to escape orbits. It is very difficult to infer observationally
whether a particular galaxy is in an escape orbit. The natural assumption is that
any galaxy associated with a group is bound to it; this assumption overestimates
the true mass by a factor of two or so in cases where there are escapers among the
galaxies. This will be discussed in the final chapter. There we will also derive the
maximum speed of escape based on the three-body theory.

1.6 Binary black holes

Black holes represent a form of matter which has collapsed on itself, in principle
into a single point. In practice, the event horizon at the Schwarzschild radius Rg ,
defined as
2G M
Rg = , (1.5)
c2
where M is the mass of the body and c the speed of light, represents the radius of
the body. In a non-rotating black hole, this radius separates the internal space of the
black hole from the outside world; no communication is possible from the inside
of the Schwarzschild radius to the outside.
For many purposes we may treat black holes as mass points, and calculate their
motions relative to each other as we do for ordinary bodies in Newtonian physics.
Close to the Schwarzschild radius the dynamics differs very much from Newtonian
dynamics, but far away, say at 100R g , only small corrections to Newtonian theory
are required. In this chapter and in the following we will only consider the latter
situations.
There are black holes of stellar masses, and supermassive black holes of millions
of solar masses. These we are fairly certain of. There may be a whole mass range
of black holes, but for various reasons we tend to observe only these two extremes.
In this book we will only discuss the supermassive variety.
Supermassive black holes have been identified in the centres of many galaxies.
Possibly every galaxy of sufficient size has or has had a supermassive black hole
in its centre. Black holes are surrounded by disks of gas, and associated with them
are phenomena which sometimes produce extremely bright radiation. For example,
jets of relativistic particles are thought to flow out in two opposite directions, along
the rotation axis of the disk. These jets may appear very bright, especially when
seen end-on.
When galaxies merge, their central supermassive black holes also approach each
other, but generally do not merge. They form a binary system of supermassive black
holes. Eventually there may also be a merger of the binary, but typically the time
18 Astrophysics and the three-body problem

Figure 1.10 Radio galaxy Cygnus A, as seen in radio waves. The two bright
emission regions straddle the galaxy which is centred on the spot of radio emission
at the centre of the picture. Image courtesy NRAO/AUI.

scale is very long, of the order of the current age of the Universe (Hubble time).
Occasionally the binary merger may happen in a shorter time scale through an
increase of eccentricity in the binary, and a collision of black holes at the pericentre
of the binary orbit (Aarseth 2003b). The problem of calculating this time scale will
be discussed in the final chapter using three-body theory.
Multiple mergers of galaxies also take place, and these may produce triple or
quadruple supermassive black hole systems. The evolution of the triple systems
is also discussed in the last chapter. Then we will find that supermassive black
holes should occasionally fly out of their parent galaxies, or sometimes remain as
oscillators in the galaxy. These are processes not well understood at present since
no unique identification of a supermassive black hole outside a galaxy has yet been
made. However, the currently popular CDM cosmology leads us to expect that
the black hole escape process was extremely common in the early universe when
large galaxies were assembled from their small progenitors.
Supermassive black holes are most likely associated with strong radio sources,
radio galaxies and quasars. Radio galaxies typically have a double lobe structure
where the regions of strongest radio emission lie on diametrically opposite sides of
the galaxy (Fig. 1.10). Two categories of theories are usually advanced to explain
the double radio source phenomenon: beam theory (Blandford and Rees 1974) and
slingshot theory (Saslaw et al. 1974). According to the former theory, supermassive
black holes remain stationary in the centres of galaxies, or at most move around
each other in binary orbits, and send out beams of relativistic plasma in two oppo-
sitely directed beams. These beams supposedly reach outside the galaxy and cause
radio emission there. In the slingshot model supermassive black holes themselves
 1.6 Binary black holes 19

are thrown out of the centre of the galaxy via the three-body process. Due to con-
servation of linear momentum, a single black hole escapes in one direction and the
binary in the opposite direction from the galaxy, and each produces radio emission
along its escape path.
As long as no supermassive black holes have been directly observed or their
existence in the radio lobes has not been disproven, it is difficult to verify either
theory. However, there are indirect ways to do so. The most clear-cut test is related
to the escape direction of radio lobes. In the beam theory the direction is along the
rotational axis of the disk of gas (the accretion disk) which surrounds the black hole.
Even though the disk usually cannot be imaged directly, it is a fair assumption that
it arises from a more extended disk of gas in the galaxy via accretion. Therefore the
observed orientation of the extended gas in the galaxy should also give information
about the accretion disk of the central black hole. The projection of a circular disk
in the sky is an ellipse; the direction of the minor axis of the ellipse shows the
projection of the axis of symmetry of the disk.
Another way to find the orientation of the accretion disk is to observe small
scale jets which emanate at right angles to the disk. Again the jet line shows the
projection of the axis of symmetry in the sky. In the beam theory the radio lobes
should lie further along the jet line, and the projected angle in the sky between the
jet line and the double lobe radio axis should be zero. Thus the radio axis should
be in the direction of the minor axis of the image of the extended gas.
It is also possible that the distribution of stars in the galaxy is spheroidal and
flattened along the same axis as the gas disk. This is certainly true of spiral galaxies,
but may also pertain to elliptical galaxies which are hosts of double radio sources.
The alignment of the gas disk and the stellar disk could have its origin in mergers of
galaxies: the angular momentum of both components should be strongly influenced
by the orbital angular momentum of the binary galaxy which has merged.
In contrast to the beam theory, the slingshot theory predicts the escape of radio
lobes along or close to the central plane of the galaxy, and perpendicular to the
minor axis of the gas distribution and perpendicular also to the jet directions. In
Chapter 7 we will discuss the escape directions in the three-body problem in more
detail, and also compare them with observations.
 2
Newtonian mechanics

This chapter introduces the basic concepts of Newtonian mechanics. We will empha-
size the areas which are most useful in the three-body problem, and also familiarise
ourselves with a system of units and scaling laws. The calculation of orbits using
Newton’s laws is a central theme of this book, and therefore a brief introduction to
the methods follows. It is not the purpose of this work to teach the latest orbit cal-
culation techniques; therefore only a brief introduction is given. Finally, we discuss
the connection of Newtonian mechanics to chaos. It may come as a surprise that
the introduction of just one more body to the well behaved two-body system brings
about a chaotic, unpredictable dynamical system. This was realised by Poincaré
well before the concept of deterministic chaos became a popular topic.

2.1 Newton’s laws

We begin with the fundamental laws of mechanics, as given by Newton in his
Principia in 1687, although in a more modern form.

First law If there are no external forces, an object will maintain its state of motion,
i.e. it will stay at rest or continue rectilinear motion at constant velocity.

Second law The rate of change of the momentum of an object is proportional to

the applied force. If the force is F, momentum p, mass m and the radius vector r ,
then
d
F = ṗ = (m ṙ ). (2.1)
dt
Third law If a body A exerts a force F on a body B, the body B will exert a force
−F on body A.
Following these fundamental laws Newton presents a corollary, which essentially
states that the forces acting on a body can be evaluated independently of each other.

20
 2.2 Inertial coordinate system 21

In more modern language this means that in the presence of several forces F 1 , F 2 ,
. . . , F n the body will behave as if there were only a single force F, which is the
vector sum of the individual forces: F = F 1 + F 2 + · · · + Fn .
These laws determine how objects will move in the presence of forces. Another
law was discovered by Newton to provide the force:
Law of gravity If the bodies A and B have masses m A and m B , respectively, and
if their mutual distance is r , A will act on B with a force that is directed towards A
and has a magnitude
Gm A m B /r 2 , (2.2)
where G is a constant, the constant of gravity, the value of which depends on the
units chosen.
When we calculate the gravitational force or acceleration due to an object, or
the effect of gravity on a body, we assume that all the mass is concentrated in the
centre of mass, i.e. the objects are pointlike. This is valid also for real bodies if they
are spherically symmetric, which is true for many celestial objects.
This is almost all one needs to know about physics; from here on, mainly math-
ematical methods are required.

2.2 Inertial coordinate system

Vectors and rectilinear motions appearing in Newton’s laws are geometric entities,
independent of any coordinate frames. In practice, however, our calculations are
carried out in some coordinate frame. If an object is at rest, then none of its co-
ordinates changes with time. Since things look different in different frames, the
definition of the coordinate frame is fundamental.
An inertial frame is a coordinate frame in which Newton’s laws hold true. If
we had an absolute frame of the whole universe, at least all the frames moving at
constant velocity with respect to this absolute frame would be inertial. The concept
of such an absolute frame has, however, turned out to be rather problematic. Yet
we can define that a frame is inertial if experiments show that Newton’s laws are
valid in that frame.
Often the situation depends on the required accuracy. An observer on the surface
of the Earth is in circular motion around the Earth’s axis, the Earth orbits the Sun,
and the Sun orbits the centre of the Milky Way galaxy. Accelerations are involved
in all these motions, and thus the frames moving with the Earth or the Sun are not
inertial. Yet we can use them as if they were. The deviation from the inertial frame
has to be taken into account only if (1) we are interested in such a long period of
time that the curvature of the trajectory becomes evident, or (2) we require such a
high accuracy that the accelerations affect the results.
22 Newtonian mechanics

2.3 Equations of motion for N bodies

Assume we have n point masses with radius vectors r i and masses m i . The total
mass of the system is


n
M= mi (2.3)
i=1

and the radius vector R of the centre of mass

 
1
n
R= mi r i . (2.4)
M i=1

The gravitational force affecting the object i is

n
1 ri − r j
Fi = −G mi m j
j=1, j=i
(r i − r j )2 |r i − r j |
(2.5)

n
ri − r j
= −G mi m j .
j=1, j=i
|r i − r j|
3

From Newton’s second law

d
Fi = (m i r˙i ) = m i r¨i . (2.6)
dt
Equating these two forces and dividing by m i the equation of motion of the object
i becomes:


n
rij
r¨i = −G mj , (2.7)
j=1, j=i ri3j

where
rij = ri − r j,
(2.8)
ri j = |r i − r j |.

In terms of rectangular x yz coordinates and basis vectors ê x , ê y and êz , the

position of m i is

r i = xi ê x + yi ê y + z i ê z . (2.9)

Defining the gradient operator ∇i as:

∂ ∂ ∂
∇i = ê x + ê y + ê z , (2.10)
∂ xi ∂ yi ∂z i
 2.3 Equations of motion for N bodies 23

the potential energy V of our system is

n
n
m j mk
V =−G
j=1 k= j+1
r jk
(2.11)
1
n
n
m j mk
=− G ,
2 j=1 k=1,k= j r jk

and
 
1
n
n
1
∇i V = − G m j m k ∇i .
2 j=1 k=1,k= j r jk

Since r jk depends only on the coordinates of the bodies j and k, only the terms
with j = i or k = i are non-zero:

n  
n  
1 1 1 1
∇i V = − G m i m k ∇i − G m j m i ∇i
2 k=1,k=i rik 2 j=1, j=i r ji

n  
1
=−G m j m i ∇i ,
j=1, j=i
ri j

where
 
1
∇i = ∇i ((x i − x j )2 + (yi − y j )2 + (z i − z j )2 )−1/2
ri j
(xi − x j )ê x + (yi − y j )ê y + (zi − z j )êz
=−
ri3j
rij
=− 3.
ri j

Then immediately


n
rij
∇i V = G mi m j . (2.12)
j=1, j=i ri3j

Comparing this with the equation of motion (2.7) one may write

m i r¨i = −∇i V, i = 1, . . . , n, (2.13)

where
1
n
n
m j mk
V =− G . (2.14)
2 j=1 k=1,k= j r jk
24 Newtonian mechanics

2.4 Gravitational potential

When studying the effect of a system on an external test body, the potential instead
of the potential energy is often used. The potential

n
mj
U (r ) = −G   (2.15)
r − r j 
j=1

has the dimension of energy per unit mass, and the potential energy of a mass m at
r is U (r )m. The equation of motion of this object is

r̈ = −∇U. (2.16)

A point mass can be replaced by a continuous mass distribution, the density of

which at r is denoted by ρ(r ). The potential at r 0 due to this mass distribution is

ρ(r ) dV
U (r 0 ) = −G , (2.17)
|r − r 0 |

where dV is the volume element and the integration is extended over the whole
mass distribution. Evaluation of the potential of an arbitrary object often leads to
complicated integrals.

Example 2.1 Find the potential of a homogeneous disk with radius R at a distance
z along the axis of the disk.
Let the surface density of the disk be ρ (kg/m2 ). Then its total mass is ρπ R 2 . In
cylindrical coordinates (r, φ) the potential at a distance z is
 
R 2π
ρr dφ
U (z) = −G dr √
0 0 z2 + r 2
 R
r dr
= −2π Gρ √
0 z2 + r 2
R 

= −2π Gρ  z2 + r 2
0
 √
= −2π Gρ z2 + R2 − z2
R2
= −2π Gρ √
z2 + R2 + z
GM 2
=−  .
z 1 + (R/z)2 + 1
 2.5 Constants of motion 25

If the distance z is much greater than the diameter of the disk, we can replace the
square root with the two first terms of its Taylor expansion:

GM 1
U (z) ≈ −
z 1 + 4 (R/z)2
1

 
GM 1 R 2
≈− 1− .
z 4 z

The potential of a point mass would be −G M/z. It is the shape of the mass
distribution that gives rise to the second term. Thus we cannot replace such an
object with a point mass.
The gradient of the potential gives the force (per unit mass). In this case only the
component parallel to the z axis remains. Using the previous series approximation,

 2
dU GM 3 R
Fz = − ≈ 2 1− .
dz z 4 z

Example 2.2 Potential of an infinite plane.

In the previous example we let the radius R of the disk grow without limit. The
potential obviously approaches infinity. However, this does not necessarily mean
that the force will also become infinite. When R → ∞ in the expression of the
potential U (z), U (z) → −2π Gρ(R − z) and we get F = −2π Gρ. This does not
depend on the distance at all. The gravitational attraction of an infinite plane is
everywhere constant and proportional to the density of the disk. For example, when
calculating stellar motions with respect to the galactic plane we can approximate
the galaxy as an infinite plane to get very simple equations of motion.

2.5 Constants of motion

The equation of motion of a body is a second order, vector differential equation. Thus
the total solution involves six integration constants, which in celestial mechanics
are usually called the integrals of the equation of motion. The general solution of
n bodies thus requires 6n integrals. Some of them can be easily evaluated.
Summing all of the equations of motion gives:

n

n

n
rij
m i r¨i = −G mi m j .
i=1 i=1 j=1, j=i ri3j
26 Newtonian mechanics

The sum on the right hand side consists of pairs (r i − r j )/ri3j and (r j − r i )/r 3ji .
These terms cancel each other, and hence


n
m i r¨i = 0.
i=1

We integrate this twice with respect to time and get

n
m i r i = at + b, (2.18)
i=1

where a and b are constant vectors. Comparison with the definition of the centre
of mass (2.4) shows that the sum on the left hand side is M R. Thus we have
at + b
R= . (2.19)
M
Thus the centre of mass or barycentre moves at constant velocity along a straight
line. The obvious consequence of this is that the barycentric coordinate frame is an
inertial frame.
The total angular momentum of the system is


n
L= m i r i × r˙i . (2.20)
i=1

The time derivative of this is

 
d

n
L̇ = m i r i × r˙i
dt i=1


n

n
= (m i r˙i × r˙i ) + (m i r i × r¨i )
i=1 i=1


n

n
r i × (r i − r j )
= −G mi m j
i=1 j=1, j=i
ri3j


n

n
ri × r j
=G mi m j .
i=1 j=1, j=i ri3j

Again the terms on the right hand side cancel each other pairwise, because the
vector product is anticommutative. Thus

L̇ = 0, (2.21)

i.e. the total angular momentum is constant.

 2.6 The virial theorem 27

Finally the total energy of the system is derived. Taking the scalar product of the
equation of motion of the ith body and r˙i , and adding all equations, one obtains:


n

n
m i r¨i · r˙i = − r˙i · ∇i V
i=1 i=1

n  
∂V ∂V ∂V
=− x˙i + ẏi + z˙i
i=1
∂ xi ∂ yi ∂z i
dV
=− .
dt
Integration with respect to time gives

1
n
m i r˙i · r˙i = −V + E, (2.22)
2 i=1

where E is a constant of integration. The left hand side of the equation gives the
kinetic energy of the system

1
n
T = m i vi2 , (2.23)
2 i=1

where vi is the speed of object i. Thus Eq. (2.22) can be written as

E = T + V, (2.24)

which is the familiar law of total energy conservation.

Thus far we have found ten integrals for the equations of motion: vectors a and
b describing the trajectory of the centre of mass, the total angular momentum L,
and the total energy E. For the complete solution, 6n − 10 additional constants
are required. The missing constants are easily found when n = 2. When n > 2,
no additional independent constants are known. In fact, it has been shown that
there are no additional independent constants that are algebraic functions of time,
coordinates and velocities.

2.6 The virial theorem

In addition to constants of motion, statistical properties can be derived. The ‘virial’
of the system is defined as


n
A= m i r i · ṙ i . (2.25)
i=1
28 Newtonian mechanics

The time derivative of this is

n
Ȧ = (m i ṙ i · ṙ i + m i r̈ i · r i )
i=1


n
= 2T + Fi · r i .
i=1

Consider the time average in the interval t ∈ [0, τ ]. Denoting this average value
by ,
  
 1 τ
n
Ȧ = Ȧ dt = 2T  + Fi · r i . (2.26)
τ 0 i=1

If the system remains bounded, its virial cannot increase without limit. Thus the
integral in (2.26) remains finite. When the length of the interval increases, Ȧ
approaches zero, and thus
 

n
2 T + Fi · r i = 0. (2.27)
i=1

This is the virial theorem in its general form. If the forces Fi are due to mutual
gravitational forces only, the sum becomes


n

n

n
ri − r j
Fi · r i = −G mi m j · ri.
i=1 i=1 j=1, j=i
ri3j

It is possible to write the double sum using only the vector differences r i − r j . We
start by writing the sum twice but by changing the order of the summation indices
i and j. We add the two identical sums and divide by two:


n
Fi · r i
i=1
 
1
n
n
ri − r j r j − ri
=− G mi m j · r i + m j mi ·rj
2 i=1 j=1, j=i ri3j r 3ji
1
n
n
ri − r j
=− G mi m j · (r i − r j )
2 i=1 j=1, j=i ri3j
1
n
n
mi m j
=− G .
2 i=1 j=1, j=i ri j
n  n
The double sum i=1 j=1 , where i = j, gives every term twice, due to the
symmetry relative to interchange between i and j. In order to have every term only
 2.7 The Lagrange and Jacobi forms of the equations of motion 29

once, the latter sum should start from j = i + 1. We make this change and multiply
the above expression by two:


n
n
n
mi m j
F i · r i = −G = V.
i=1 i=1 j=i+1
ri j

Thus
1
T = − V. (2.28)
2
The virial theorem (2.28) can be used to study, for example, the stability of
clusters of stars. The kinetic energy can be determined from the velocity dispersion
and the potential energy from the mass distribution derived from the brightness of
the objects. If the kinetic energy is much higher than the average value given by the
virial theorem, the system is unstable. The virial theorem indicates that clusters of
galaxies should be unstable, which contradicts observational evidence about their
stability. This is one reason to believe that such systems must contain large amounts
of dark matter.

2.7 The Lagrange and Jacobi forms of the equations of motion

For three bodies, Eq. (2.7) gives us
 
r 12 r 13
r̈ 1 = −G m 2 3 + m 3 3 (2.29)
r12 r13
 
r 23 r 21
r̈ 2 = −G m 3 3 + m 1 3 . (2.30)
r23 r21
In the Lagrangian formulation of the equations of motion the coordinate dif-
ferences are used instead of the coordinates themselves. The acceleration between
bodies 1 and 2 is obtained by subtracting (2.30) from (2.29):
  
r 12 r 13 r 32
r̈ 12 = −G (m 2 + m 1 ) 3 + m 3 3
+ 3
r12 r13 r32
  
r 12 r 13 r 32 r 21
= −G M 3 + m 3 3
+ 3 + 3 (2.31)
r12 r13 r32 r21
 
r 12
= G m3W − M 3 ,
r12
where
r 12 r 23 r 31
W= 3
+ 3 + 3 . (2.32)
r12 r23 r31
30 Newtonian mechanics
3
2 R3

r r31
1 r3
rB r1

Figure 2.1 Coordinates of the Jacobi system: the bodies 1 and 2 form a binary
while the third body (3) is more distant.

The corresponding equations for the acceleration between bodies 2 and 3 as well
between bodies 3 and 1 are obtained from (2.31) by changing the subscripts ac-
cordingly.
In the Jacobi system of coordinates, the three-body system is considered to be
composed of two parts: a clearly defined binary, and a somewhat distant third body
(Fig. 2.1). The centre of mass of the binary is used as a reference point whose
position is r B . The relative position vector of the binary is r 21 = r . By definition
of the centre of mass (Eq. (2.4))

1
rB = (m 1r 1 + m 2r 2 )
m1 + m2
m1 m2 m2 m2
= r1 + r2 + r1 − r1 (2.33)
m1 + m2 m1 + m2 m1 + m2 m1 + m2
m2
= r1 + r
m1 + m2

or, alternatively,

m1 m2 m1 m1
rB = r1 + r2 + r2 − r2
m1 + m2 m1 + m2 m1 + m2 m1 + m2
m1 (2.34)
= r2 − r.
m1 + m2

Let us call the position vector of body 3 relative to the centre of mass R3 = r 3 − r B .
Then

m2
r 31 = r 3 − r 1 = R3 + r B − r 1 = R3 + r,
m1 + m2
m1 (2.35)
r 32 = r 3 − r 2 = R3 + r B − r 2 = R3 − r.
m1 + m2
 2.8 Constants of motion in the three-body problem 31

Also
m1 m2
r 31 + r 32
m1 + m2 m1 + m2
   
m1 m2 m 1m 2 m 1m2
= + R3 + − r
m1 + m2 m1 + m2 (m 1 + m 2 )2 (m 1 + m 2 )2
= R3 . (2.36)
Using Eqs. (2.31) and (2.36), the equations of motion are:
  
r r 31 r 32
r̈ = −G (m 1 + m 2 ) 3 + m 3 3
− 3 (2.37)
r r31 r32
and
m1 m2
R̈3 = r̈ 31 + r̈ 32
m1 + m2 m1 + m2

m1 m 1 r 31
= G m2 W−M
m1 + m2 m 1 + m 2 r313
 (2.38)
m2 m 2 r 32
+ m1 (−W) − M
m1 + m2 m 1 + m 2 r32
3
 
m 1 r 31 m 2 r 32
= − GM + .
m 1 + m 2 r31
3 m 1 + m 2 r323

If r  R3 , r 31 ≈ r 32 ≈ R3 . Then the equations of motion (2.37) and (2.38) are

simply
r
r̈ = −G(m 1 + m 2 ) , (2.39)
r3
and
R3
R̈3 = −G M . (2.40)
R33
These are two separate two-body equations of motion the solutions of which will
be discussed in the next chapter. The solutions are Keplerian orbits, an inner orbit
(Eq. (2.39)) and an outer orbit (Eq. (2.40)). As Eqs. (2.39) and (2.40) are not exact,
one is really dealing with two perturbed Keplerian orbits.

2.8 Constants of motion in the three-body problem

The angular momentum for a three-body system is (Eq. (2.20))

L = m 1r 1 × ṙ 1 + m 2r 2 × ṙ 2 + m 3r 3 × ṙ 3 . (2.41)
32 Newtonian mechanics

In the Lagrangian form the angular momentum in the centre of mass coordinate
system reads
 
m 1 m 2 m 3 r 12 × ṙ 12 r 23 × ṙ 23 r 31 × ṙ 31
L= + + (2.42)
M m3 m1 m2
and in the Jacobi form

L = M(r × ṙ ) + m(R3 × Ṙ3 ) (2.43)

where M and m are the reduced masses:

m 1m 2
M= ,
m1 + m2
(2.44)
m 3 (m 1 + m 2 )
m=
M
(Problem 2.5).
The plane through the centre of mass and perpendicular to L is called the invari-
able plane. When L = 0, Eq. (2.43) tells us that the motion is confined to a plane
which contains all four vectors r , ṙ , R3 and Ṙ3 .
The energy integral for the three-body problem is
1 
E= m 1 v12 + m 2 v22 + m 3 v32
2   (2.45)
m 1m2 m 1m3 m 2m3
−G + + .
r21 r13 r23
The corresponding equations in the Lagrangian and Jacobi forms are
  2 2 
m 1 m 2 m 3 1 v23 v31
2
v12
E= + +
M 2 m1 m2 m3
  (2.46)
1 1 1
− GM + +
m 3r21 m 2r13 m 1r23
and
 
1  m 1m 2 m 1m 3 m 2m 3
E= Mv 2 + mV32 − G + + , (2.47)
2 r r31 r23
respectively, where v = |ṙ | and V3 = | Ṙ3 | (Problem 2.6).

2.9 Moment of inertia

In the centre-of-mass coordinate system, the moment of inertia of a three-body
system is

I = m 1 r12 + m 2r22 + m 3r32 . (2.48)

 2.9 Moment of inertia 33

In the Lagrangian form it becomes

 2 
m1m2m3 r21 r2 r2
I = + 13 + 32 , (2.49)
M m3 m2 m1
and in the Jacobi form
I = Mr 2 + m R32 (2.50)
(Problem 2.7).
Differentiating the moment of inertia twice with respect to time,
I˙ = 2m 1r 1 · ṙ 1 + 2m 2r 2 · ṙ 2 + 2m 3r 3 · ṙ 3 ,
     
I¨ = 2m 1 v12 + r 1 · r̈ 1 + 2m 2 v22 + r 2 · r̈ 2 + 2m 3 v32 + r 3 · r̈ 3
from which it follows that
1¨
I = 2T + V = 2E − V (2.51)
2
(Problem 2.8). Equation (2.51) is known as the Lagrange–Jacobi identity.
The moment of inertia is a measure of compactness of the three-body system, in
the sense that the more compact the system, the smaller is the moment of inertia.
When the moment of inertia increases, a greater and greater sphere is required to
surround the system. When in addition I¨ > 0, the bounding sphere increases at an
accelerated pace. In practice, I¨ > 0 often means the escape of one body from the
other two.
In the evolution of a three-body system there are periods when I¨ ≈ 0. During
those periods the system is bounded in a relatively small volume, and the virial
theorem (Eq. (2.28)) is approximately satisfied. From numerical orbit calculations
(see Fig. 1.2) the potential energy V is known to fluctuate a great deal, and since
E is a constant in Eq. (2.51), I¨ must fluctuate on either side of zero. Sooner or
later I goes through a deep minimum after which I¨ remains positive over a period
which is sufficiently long to permit an escape of one of the bodies. Over that period,
the time averages of T and V satisfy
2T  > −V  . (2.52)
Later, when the escape velocity of the third body Ṙ3 asymptotically approaches
a constant value, Eq. (2.28) is satisfied again. Therefore the general behaviour
of the moment of inertia I in the three-body systems, as revealed by numerical
orbit calculations, together with the Lagrange–Jacobi identity, clearly leaves open
the possibility of an escape in the three-body problem. It does not prove that an
escape must happen, and indeed not every minimum of I leads to an escape, but
when the possibility of escape is offered often enough, it is easy to believe that the
evolution of a three-body system ends in an escape of one of the bodies, just as in
the Pythagorean problem (Fig. 1.2).
34 Newtonian mechanics

2.10 Scaling of the three-body problem

One of the nice features of the solutions of the three-body problem, as well as the
N -body problem, is that the solutions can be freely scaled from one physical system
to another. For example, the Pythagorean problem could represent the motions of
three stars, of 3, 4 and 5 solar masses, starting from mutual separations of 3, 4 and 5
parsecs from each other, or three isolated planets of 3, 4 and 5 Earth masses, starting
from 3, 4 and 5 AU (astronomical units) apart. It can be easily shown (Problem 2.9)
that if r 1 (t), r 2 (t) and r 3 (t) represent the solution (orbits) of a three-body problem,
and k is a real number, then also k 2r 1 (t), k 2r 2 (t) and k 2r 3 (t) is a solution of the
same problem, as long as the time t is scaled as k 3 . In the new solution the velocity
v scales as k 2 /k 3 = k −1 , the energy E scales as v 2 , i.e. as k −2 , and the angular
momentum L scales as r × v, i.e. as k 2 k −1 = k. The product E L 2 is scale free, a
fact which will be used in the following to normalise the angular momenta.

2.11 Integration of orbits

It is generally possible to calculate the orbits in N -body systems by using a com-
puter. Here, a brief outline of the principles commonly used is given (Aarseth 1971).
There are numerous methods in contemporary use, but the details are beyond the
scope of this book. Good expositions of the latest methods are available for example
in Aarseth (2003a). Usually the orbit calculation procedure is referred to as integra-
tion of orbits. To calculate the orbits, Eq. (2.7) is to be solved for all bodies in the
system. As long as the current positions of the bodies are known, the accelerations
at that moment are given by the solutions of Eq. (2.7). If the current velocities are
also known, the motions at that moment are fully specified. The new position r 1 is
obtained from the current position r after a small time step t by the Taylor series
1 1 1
r 1 = r + ṙ t + Ft 2 + Ḟt 3 + F̈t 4 + · · ·. (2.53)
2 6 24
Here we have written F = r̈ . The derivatives Ḟ and F̈ are obtained from Eq. (2.7)
by differentiation:

n
ṙ i j r i j (r i j · ṙ i j )
Ḟ = −G mj 3
−3 ,
j=1, j=i ri j ri5j


n
r̈ i j ṙ i j (r i j · ṙ i j )
F̈ = −G mj 3
−6 (2.54)
j=1, j=i
ri j ri5j
 
r i j (r i j · ṙ i j )2
+3 5 5 − ṙ i j · ṙ i j − r i j · r̈ i j .
ri j ri2j
 2.11 Integration of orbits 35

Similarly, the new value of the velocity ṙ 1 is obtained from

1 1
ṙ 1 = ṙ + F t + Ḟt 2 + F̈t 3 + · · ·. (2.55)
2 6
By repeating the short time step for all of the bodies we may hope to calculate the
time evolution of the whole system.
There are obvious problems with this scheme. Complicated calculations are car-
ried out at every time step which consume computer time and which cause round-off
errors. Thus, in order to finish the calculation in a reasonable amount of computer
time, the time step should not be very small. But then the convergence of the Taylor
series becomes poorer. We realise that during a relatively long time step the acceler-
ation can change considerably from what it was at the beginning of the step, because
in reality all bodies move simultaneously. Some form of prediction of motions of
all bodies has to be introduced prior to taking the time step. The prediction can be
based on a low order Taylor series which does not require complicated calculations:

1 1
r 1 = r + ṙ t + F t 2 + F̂ t 3 . (2.56)
2 6

We define F̂ as the backward difference

F − F −1
F̂ = (2.57)
t−1

where t−1 is the length of the preceding time step, and F −1 is the value of accel-
eration at the beginning of that step. In general, the computer algorithm becomes
faster when the derivatives (2.54) are constructed from the corresponding back-
ward differences. The values of F only need to be stored for the several previous
steps, and the derivatives of F are then calculated by simple operations from these
quantities. In particular, the prediction makes use of the values of the acceleration
at two successive points where the acceleration has to be calculated in any case,
and thus the prediction requires only modest extra calculation effort.
Other features which save computer time are the use of variable time step and
individual time step. At some points in the orbit a body may be advanced by
long time steps, at other more critical times a small step is necessary to preserve
accuracy. Every body has a different situation in this regard; therefore it is useful to
assign a different, individual, time step to each of them. The prediction of the orbits
of all bodies up to a common moment of time then becomes crucial. When the
acceleration and its derivatives change rapidly, it is necessary to go to a short time
step. Therefore one may calculate the length of the step by requiring that the ratio
of successive terms in the Taylor series is of the order of a small number η  1.
36 Newtonian mechanics

For example,
1
| Ḟ|t 3
6
= η,
1
2
|F|t 2

from which it follows that

3|F|
t = η . (2.58)
| Ḟ|
Usually a somewhat more complicated rule for t is used which takes account of
the higher derivatives of F but the general principle is the same. One may also set
other simple demands on t based on practical experience with computations.
A special situation arises when two bodies come very close to each other. Then
a rule like Eq. (2.58) leads to such a small step that the computation may virtually
terminate before the two bodies pass each other. The source of the trouble is the
1/r 3 factor in Eq. (2.7). Let us write this equation again just for two bodies, with
accelerations caused by other bodies lumped together as a perturbing acceleration
f:
r
r̈ = −µ + f. (2.59)
r3
Here µ = G(m a + m b ); m a and m b are the masses of the two bodies. One attempts
to transform this to another form where there are no singular terms which approach
infinity when r → 0. This kind of transformation is called regularisation of the
equation of motion. The common first step is to introduce an auxiliary variable τ
by
dτ 1
= . (2.60)
dt r
If we represent differentiation relative to τ by primes Eq. (2.59) becomes
r  r
r  = r − µ + r2 f (2.61)
r r
(Sundman 1912; Problem 2.10). This already helps somewhat since now the 1/r 3
factor is replaced by 1/r . In order to get rid of the remaining 1/r we introduce two
auxiliary variables h and e:

1 r 2 µ
h= − ,
2 r2 r (2.62)
r 1   
e = − − 2 r × r × r .
r µr
 2.11 Integration of orbits 37

It is then easy to see that Eq. (2.61) transforms to

r  = 2hr − µe + r 2 f . (2.63)
When f = 0, the quantities h and e are constants. We see this by evaluating
 
2
d 1 r µ r  r  2 r  · r  µ
h = 2
− = − 3
+ 2
+ 2 r ,
dτ 2 r r r r r

where
  
r  · r  r r  r
= 2· r −µ +r f
2
r2 r r r
r r  2 µ
= 3 − 2 r + r · f .
r r
Thus
h = r  · f . (2.64)
After some calculations we find also
1 
e = − r  (r · f ) + f (r · r  ) − 2r h  (2.65)
µ
(Problem 2.11). Therefore a regularized equation of motion has been derived which
has no singular terms when r → 0 as long as f → 0. Quantities h and e arise again
in the next chapter where their physical significance is clarified (Eqs. (3.11) and
(3.13)).
The above regularisation scheme (Burdet 1967, Heggie 1973) is not the only
possibility. Even though the Burdet–Heggie regularisation is known to function
perfectly well in numerical solutions of the N -body problem (Saslaw et al. 1974,
Heggie 1975, Valtonen and Heggie 1979), actually another transformation called
the K-S regularisation is more widely used (Kustaanheimo 1964, Kustaanheimo
and Stiefel 1965). However, the important point to us is that there are methods to
handle close two-body encounters without slowing down the orbit calculation and
without loss of numerical accuracy.
How does one then measure the accuracy of the orbit calculation? One may
test the method with a problem where the analytic solution is known, such as
the two-body problem (Chapter 3, Problem 3.8). Alternatively, one may fol-
low the conservation of quantities which are known to be strictly constant in
the N -body problem, i.e. the centre of mass and momentum, the total energy
and angular momentum. Perhaps the most stringent test is based on the time-
reversability of the orbits. If at any point along the orbit the velocities of all bodies
are reversed, the bodies should retrace their trajectories exactly. Computational
38 Newtonian mechanics

errors make this unlikely to happen in numerical orbit integration. The ability
to retrace the orbit backward is a very demanding test for the accuracy of the
method.

2.12 Dimensions and units of the three-body problem

In the general three-body problem, 21 independent variables are to be specified:
three position and three velocity components plus the mass value for each body.
The position and the motion of the centre of mass, as well as the total mass and
the linear scale (scale factor k above) of the system are not of much interest, so
that the number of ‘interesting’ variables is reduced to 13. The same can be said
about the spatial orientation of the system, described by two angles specifying the
direction of the total angular momentum. Therefore the number of dimensions of
the three-body problem is 11.
The magnitude of the total angular momentum L 0 and the value of the total
energy E 0 are important parameters. The combination

L 20 |E 0 |/G 2 m 50 (2.66)

is a dimensionless quantity. Here m 0 signifies the average mass of the three bodies,
to be defined more specifically in Chapter 7 (Eq. (7.28)). If this quantity is given
the value 6.25,

L 0 ≡ L max = 2.5Gm 2.5

0 /|E 0 | .
0.5
(2.67)

L max is an important standard of reference for the angular momenta of three-body

systems (Chapter 7).
The reason why the total mass M and the linear scale factor r are quantities of
little interest is that the three-body problem is scale free as mentioned above. If one
three-body problem with a given mass scale (total mass) M, distance scale r and
time scale t has been solved, then it is obvious (Eq. (2.7) and Section 2.10) that
this solution also applies to any other mass scale M and distance scale r . The time
scale t ∼ M −1/2r 3/2 and the velocity scale v ∼ M 1/2r 1/2 transformations ensure
this. Therefore M and r are ‘uninteresting’ parameters.
In numerical orbit calculations it is usual to fix the system of units by putting the
gravitational constant G = 1. Writing this constant in the common astronomical
units of solar mass (M = 1.99 × 1030 kg), parsec (1 pc = 3.086 × 1016 m) and
km/s, we get

G = 4.3 × 10−3 (km/s)2 pc/M . (2.68)

 2.13 Chaos in the three-body problem 39

G = 1 ties the units such that corresponding to pc and km/s the mass unit is
1/(4.3 × 10−3 M ) = 233M . Since
1 km/s ≈ 1 pc/Myr (2.69)
the corresponding unit of time is very nearly one megayear (1 Myr = 106 yr).
Obviously the calculation can be carried out in other units but this does not bring
a new dimension to the three-body problem. In practice, the time is frequently
given in the natural time unit of the three-body system which is the crossing time
Tcr = G M 5/2 /(2|E 0 |)3/2 (see Eq. (8.26) and the discussion there). This is roughly
the time that it takes a body to travel through the three-body system.
In principle, the three-body problem may be solved by going through the 11-
dimensional phase space with a fine tooth comb, and by calculating the orbital
evolution at each phase space point. Let us briefly consider how practical this
proposal is. Construct a grid in the phase space composed of 100 values in each
coordinate and calculate the orbits at each grid point. It is in total 10011 = 1022
orbits. If every orbit takes 1 second to calculate in a fast machine, the project would
last 1022 seconds, i.e. about 3 × 1014 yr. Even so, the solution would be rather rough
since a mesh of 100 grid points is far too coarse to cover the phase space adequately.
Therefore it is not possible to find a complete solution of the three-body problem by
‘brute force’, even if we have perfect integration methods and powerful computers
in use.
The alternative, an analytical solution for example in the form of a series has
been proposed and searched for. It has been estimated that 108 000 000 terms are
required in the Sundman series to represent a three-body solution with the accuracy
commonly used in the ephemeris calculations. The computer time requirement
would be impossible, and certainly the representation could not be said to be simple.
It is still possible that, in the future, computers that use mathematical manipulators
(e.g. Maple, Mathematica) may be used to identify important terms in long series
expansions.
In recent years it has become clear that solutions of the three-body problem can
be found and represented in simple form by using a combination of analytical and
computational methods. Analytical methods give the basic functional forms which
can be tested and improved by computer experiments. They also provide useful
physical insights into the problem. We will describe this avenue of research mainly
in Chapters 7 to 10. Before that, some computer experiments will be described
which are useful to develop the analytical theory.

2.13 Chaos in the three-body problem

A two-dimensional description of the state of the system is usually called a map.
In the three-body problem the free dimensions require reduction in order to make
Other documents randomly have
different content
 CHAPTER III.

My sleep was sound and sweet; nevertheless, when the morning

began to dawn, I was awakened by its first glimmerings, and found
that my thoughts became at once too busy to admit of a return to
slumber. I therefore arose, and went to walk in an open gallery, with
which my chamber was connected. This gallery commanded a
prospect of a great part of the city, which at that hour appeared no
less tranquil than stately, nothing being in motion except a few small
boats gliding here and there upon the river. Neither as yet had any
smoke begun to darken the atmosphere; so that all things were seen
in a serene and steady light, the shadows falling broadly westward
over streets and squares—but pillars, and obelisks, and arches,
rising up every where with unsullied magnificence into the bright air
of the morning. The numerous poplars and other lofty trees of the
gardens, also, seemed to be rejoicing in the hour of dew and
silence; so fresh and cheerful was the intermixture of their branches
among the piles of white and yellow marble. Near at hand, over the
groves of the Philoclean Mansion, I could see the dome of the
Pantheon, all burnished with living gold, and the proud colonnades
of the Flaminian Circus, [pg 21]loaded with armies of brazen statues.
Between these and the river, the theatres of Pompey and Marcellus,
and I know not how many temples, were visible. Across a more
crowded region, to the westward, my eye ascended to the cliffs and
towers of the Capitol; while, still farther removed from me,
(although less elevated in natural situation,) the gorgeous mansion
of the Emperor was seen, lifted up, like some new and separate city,
upon its enormous fabric of arcades. Behind me, the Flavian
Amphitheatre, the newest and the most majestic of all Roman
edifices, detained the eye for a space from all that lay beyond it—the
splendid mass of the Esquiline—and those innumerable aqueducts
which lie stretched out, arch after arch, and pillar after pillar, across
the surrounding plain.

As I stood upon a projecting balcony, I heard some person stepping

softly along the floor, and, being screened by some pillars, looked
back into the gallery without subjecting myself to observation in
return. The noise, I found, was occasioned by one of the slaves of
Licinius, (the same I had remarked over night,) who had an air of
anxious vigilance on this occasion, looking about from side to side as
if afraid of being detected in some impropriety. I heard him tap at
one of the apartments adjoining my own, and young Sextus,
opening the door, eagerly asked, “Well, Dromo, good Dromo, what
news?—Have you seen or heard any thing of her?—Speak low, I
beseech you, and remember that my preceptor is near.” “Which
preceptor?” replied Dromo; “count me your best, and I will teach
you how to manage all besides.”—“Hush!” whispered the young
man; “he may be astir with these eternal parchments.”[pg 22]—“Be
easy,” returned the slave; “I have found out facts which will serve to
bridle that tongue at any time.”—“Dromo,” said Sextus, “have a care;
remember the thong of sleek leather which hangs at the foot of the
stair-case; and many is the time I have saved you from it; for which
you may, perhaps, have to thank the beauty of her who has
rendered you necessary to me, as much as my own good nature.
But no more idle words at present—what have you to tell me?”

“I have just been down,” answered he, “to the herb-market. I had
made my bargain, and was coming away, when I met one of old
Capito’s men, driving an ass laden with articles from the country. So
I asked if he was carrying a present to his master’s brother. He said
he had brought nothing for Lucius but a letter; and that he believed
its purport was to invite the two young ladies, to come out to-day
and enjoy the beauty of the season. I no sooner got this
information, than I ran hither as swiftly as my legs would carry me.
You can easily go out, as if by chance, to pay your respects to the
Patrician.”
“Ah, Sempronia!” sighed Sextus, “shall I approach you at last?—
What will she think when she sees me there?—Oh, how will she
speak to me?”

While he was uttering these words, Dromo suddenly started, and

came peeping on tiptoe towards the place where I stood. I stepped
from behind my pillar, and said to the astonished youth, “Fear not,
Sextus, that I shall intermeddle with your secrets, or make any use
of what I have accidentally overheard. But I wish you would satisfy
my curiosity, and inform me who is this [pg 23]lady, and what may
be the meaning of all this concealment?”

Here Dromo, perceiving that his young master was a good deal
confused, came forward and said, “From observing your looks last
night, when I was making a handle of yon barbarian to torture our
friend of the porch, I think you are a good-natured person, who
would not willingly bring any of us into trouble. The truth is, that
Licinius wishes my young master here to marry a certain lady, who
has already had wet eyes over the ashes of a first husband; but who
is of noble birth, and very rich. Now Sextus, being only eighteen,
does not like this great lady so well as she likes him; and has, in
fact, lost his heart elsewhere.”—“Dromo,” answered I, taking young
Sextus by the hand as I spoke, “this is a pretty common sort of
story; but I shall take no side till I have seen both of the ladies; and
the sooner your ingenuity can bring that about, the more shall I be
beholden to you.”—“We shall try,” replied the slave, observing that I
had overcome the reluctance of the lover; “but in the meantime I
observe that the clients are beginning to assemble in the porch, to
await the forthcoming of Licinius. Go, therefore, and get some
breakfast, for, by and by, you will both be expected to accompany
the Senator to the Forum, to hear him plead; which, between
ourselves, will be a six hours’ job for you, unless you manage
matters dexterously.”

This hint produced a visible effect on Sextus; but we went down

together immediately to an apartment, where some bread and
grapes were prepared for us; and there, with much ingenuousness,
he opened his [pg 24]heart to me. But what surprised me most of
all, was to hear, that although he had been enamoured of Sempronia
for several months, and was well acquainted with several of her
relations, he had never yet seen her, except at certain places of
public resort, nor enjoyed any opportunity of making known his
passion.

While I was expressing my astonishment at this circumstance, we

were interrupted by Xerophrastes, who came to inform us that
Licinius, having already descended into the hall, was about to issue
forth, and desirous of my company, if no other occupation detained
me. We accordingly followed the philosopher, and found his patron
where he had indicated, pacing to and fro, in the highest state of
excitation, like a generous steed about to scour the field of battle.
The waxen effigies of his ancestors stood at one end of the hall,
some of them defaced with age; and upon these he frequently fixed
his ardent eyes. Seeing me enter, he immediately cried out, “Come
hither, young friend, and I shall presently conduct you to a scene
worthy, above all others, of the curiosity of a stranger.”

With this, arranging his gown, and putting himself into a dignified
attitude, he ordered the porter, who stood chained by the door, to
throw wide its massy valves; which being done, the litigants and
consulters, who were without, received the orator with acclamations,
and surrounded him on all sides. Some of the poorer ones, I
observed kissing the hem of his garment, and dodging wistfully at
his elbows, without ever attracting a word or look from him; while
those of a higher class came forward more familiarly, seeking to
impress particular circumstances upon his memory, and [pg
25]paying him compliments on the appearance he had made the day
before in the Centumviral Court. Encircled by this motley group, he
walked towards the great Forum, followed at a little distance by
Sextus, the preceptor, myself, and some freedmen of his household.
In moving on, we passed, by accident, the door of another great
pleader, by name Bruttianus, who stood there attended in a similar
manner. When he perceived Licinius, this man took from his door-
post a green palm-branch, and waved it towards us in a vaunting
manner; but our friend, saluting him courteously, cried out, with his
sharp and cutting voice, “We shall try it again.” Whereon
Xerophrastes, immediately stepping up to his patron, began thus,
“How this vain-glorious person exposes himself!—he is certainly a
weak man; and his tones, by Hermes, are more detestable than
those of an African fowl.”—At which words, Sextus tipped me the
wink; but I did not observe that Licinius was at all displeased with
them. Yet, soon after, Bruttianus having overtaken us, the
processions were joined, and the two pleaders walked the rest of the
way together in a loving manner, exchanging complimentary
speeches; to which Xerophrastes listened with edifying gravity of
visage.

At length we entered that venerable space, every yard of whose

surface is consecrated to the peculiar memory of some great
incident in the history of Rome. Young Sextus allowed me to
contemplate for some time, with silent wonder, the memorable
objects which conspired to the decoration of this remarkable place;
but after the first gaze of astonishment was satisfied, proceeded to
point out, in order, the names and uses of [pg 26]the principal
structures which rose on every side over its porticos—above all, of
its sublime temples—into whose cool and shady recesses the eye
could here and there penetrate through the open valves. Nor did the
ancient rostrum from which Tully had declaimed, escape our
observation—nor within its guarding rail of silver, the rising shoots of
the old mysterious fig-tree of Romulus—nor the rich tesselated
pavement which covered the spot that had once yawned an abyss
before the steady eye of Curtius—nor the resplendent Milliary pillar
which marked the centre of the place. In a word, had the gathering
crowds permitted, I could have willingly spent many hours in
listening to the explanation of such magnificent objects; but these,
and the elevated voice of Licinius, who was just beginning his
harangue, soon compelled me to attend to things of another
description.
Within one of the proud ranges of arcade, on the side nearest to the
Capitoline stairs, a majestic Patrician had already taken his seat on
an elevated tribunal—his assessors being arranged on a lower bench
by his side, and the orators and clients congregated beneath. When
I heard the clear and harmonious periods of my kinsman; when I
observed with what apparent simplicity he laid his foundations in a
few plain facts and propositions; with what admirable art he
upreared from these a superstructure of conclusions, equally easy as
unexpected; when he had conducted us to the end of his argument,
and closed with a burst of passionate eloquence, in which he
seemed to leave even himself behind him, I could not but feel as if I
had now for the first time contemplated the practised strength of
[pg 27]intellect. Yet I have lived to discover that the talent which so
greatly excited my wonder is often possessed from nature, or
acquired through practice, in a measure which at that time would
have afforded me scarcely inferior delight, by men of no
extraordinary rank.

The keen and lively gestures of the fervid Licinius, whose soul
seemed to speak out of every finger he moved, and who appeared
to be altogether immersed in the cause he pleaded, were succeeded
by the solemn and somewhat pompous stateliness of Bruttianus,
who made a brief pause between every two sentences, as if he were
apprehensive that the mind of the judge could not keep pace with
the stream of his illustrations, and looked round ever and anon upon
the spectators with a placid and assured smile, rather, as it seemed
to me, to signify his approbation of their taste in applauding him,
than his own pleasure in their applauses. Nevertheless, he also was
a splendid speaker, and his affectation displeased the more, because
it was evidently unworthy of his understanding. While he was
speaking, I observed that the Stoic preceptor was frequently shifting
his place among the crowd, and muttering every where expressions
of high contempt. But this did not disgust me so much as the fixed
attitude of ecstasy in which he listened to the discourse of his own
patron, and the pretended involuntary exclamations of his delight.
“Oh, admirable cadence!” he would say, “I feel as if I were draining a
honey-comb. Oh, harmonious man, where have I, or any other
person here, sucked in such sweetness!” These absurd phrases,
however, were caught up forthwith, and repeated by the numerous
young men who hung upon the skirts of [pg 28]the orator, and
seemed, indeed, to be drinking in nectar from the speech, if one
might judge from their countenances. From their taking notes in
their tablets from time to time, and from the knowing looks they
assumed at the commencement of every new chain of argument, I
guessed that these might be embryo jurisconsults, preparing
themselves by their attendance for future exertions of the same
species; and, indeed, when I listened to their conversation at the
close of every speech, I thought I could perceive in their tones and
accents, studied mimicry of the natural peculiarities of Licinius,
Bruttianus, and the other orators. Altogether, the scene was as full of
amusement as of novelty, and I could willingly have remained to the
end of the discussion. But my eyes chanced to fall upon young
Sextus, and I could not but see that his mind was occupied in
matters remote from the business of the Forum. He stood with his
arms folded in his gown, and his eyes fixed upon the ground, only
lifting them up from time to time with an impatient air towards a
side entrance, or to observe by the shadows on the porticos what
progress the sun was making.

Perceiving, at length, that Xerophrastes had his back turned to us,

and that his father was engaged with his tablets, he plucked me by
the sleeve. I understood his meaning, and followed him quickly
through the crowd; nor did we look back till we had left the noise of
the Forensic assembly entirely behind us. “I am depriving you,” he
said, “of no great gratification, for that old creature is, indeed,
possessed of much natural shrewdness, but he is bitter from
observing that his reputation is rather eclipsed by younger people,
and looks like [pg 29]some worn-out and discarded cat, grinning
from the top of the wall at the dalliance of some sleeker rival. You
could find no delight in the angry sneerings of such an envious
person; and his age would prevent you, at the same time, from
willingly giving way to contemptuous emotions. I will be your guide
to the villa. But if any questions be asked on our return, you can say
I was anxious to shew you something of the other regions of the
city.”

He hurried me through noble streets, and past innumerable edifices,

before each of which I would gladly have paused. Nevertheless,
seeing him wrapped up in anxious thoughts, I did not oppose myself
to his inclinations; and ere long, having passed the Hill of Gardens, I
found that we had gained the eastern limit of the city.

[pg 30]
 CHAPTER IV.

A sharp walk of about an hour and a half on the Salarean Way,

brought us within sight of the Suburban of Capito.

A lofty wall protected the fields of this retirement from the intrusive
eyes of passengers. We entered by a small side-door, and found
ourselves, as if by some magical delusion, transported from the glare
of a Roman highway, into the depth and silence of some primeval
forest. No nicely trimmed path conducted our feet. Every thing had
at least the appearance of being left as nature had formed it. The
fern rustled beneath us as we moved; the ivy was seen spreading its
careless tresses from tree to tree; the fawn bounded from the
thicket. By degrees, however, the gloom lessened, till at length, over
an open space of lawn, we perceived the porch of entrance, and a
long line of colonnade. We passed under the porch, and across a
paved court where a fountain was playing, into the great hall, which
commanded all the other side of the place—a noble prospect of
elaborate gardens gradually rising into shady hills, and lost in a
distance of impenetrable wood.

Here a freedman attended us, who informed us that Capito had

retired into a sequestered part of the grounds [pg 31]with some
friends from the city; but that if we chose we could easily join him
there. We assented, and, following his guidance, ere long traversed
no narrow space of luxuriant cultivation. From one perfumed terrace
we descended to another; till, having reached a certain green and
mossy walk, darkened by a natural arching of vines and mulberries,
the freedman pointed to a statue at the farther end, and told us it
stood over against the entrance of his master’s summer-house.
When we reached the statue, however, we could perceive no
building. The shaded avenue terminated in face of a precipitous
rock, from which there fell a small stream that was received in a
massive basin, where the waters foamed into spray without
transgressing the margin. A thousand delicious plants and far-sought
flowers clustered around the base of the rock and the brink of the
fountain, and the humming of innumerable bees mingled with the
whispers of the stream. We stood for a moment uncertain whether
we should move on or retire, when we heard one calling to us from
beyond; and passing to the other side of the basin, descried,
between the rock and the falling water, a low entrance into what
seemed to be a natural cave or grotto. We stooped, and found
ourselves within one of the most luxurious retirements ever haunted
by the foot of Dryad. A sparry roof hung like a canopy of gems and
crystals over a group of sculptured Nymphs and Fawns, which were
placed on a rustic pedestal within a circular bath shaped out of the
living stone. Around the edge of the waveless waters that slumbered
in this green recess, were spread carpets rich with the dyes of Tyrian
art, whereon Capito was reposing with [pg 32]his friends. He
received Sextus with kindness, and me with politeness, introducing
us both to his companions, who were three in number—all of them,
like himself, advanced in years, and two of them wearing long
beards, though their demeanour was destitute of any thing like the
affected stateliness of our friend Xerophrastes. These two, as our
host informed us, were Greeks and Rhetoricians—the third, a
Patrician of the house of the Pontii, devoted, like himself, to the
pursuits of philosophy and the pleasures of a literary retirement.

They were engaged, when we joined them, in a conversation which

had sprung from the perusal of some new metaphysical treatise.
One of the Greeks, the more serene-looking of the pair, was
defending its doctrines with earnestness of manner, although in a
low and measured cadence of voice; the other espoused the
opposite side, with quickness of utterance and severe animation of
look; while the two lordly Romans seemed to be contenting
themselves, for the most part, with listening, although it was not
difficult to perceive from their countenances, that the one sided in
opinion with the Stoic, and Capito himself with the Epicurean.

They all arose presently, and proceeded to walk together, without

interrupting the conversation, along the same shaded avenue which
Sextus and myself had already traversed. He and I moved along with
them, but a little in their rear—my companion being still too much
abstracted to bestow his attention on what they were saying; while I
myself, being but little an adept in such mysteries, amused myself
rather with the exterior and manners of the men, than with the
merits of the opinions they were severally defending. The [pg
33]Greeks were attired in the graceful costume of their country,
which was worn, however, far more gracefully by the Epicurean than
his brother,—the materials of his robe being delicate, and its folds
arranged with studied elegance, whereas the coarse garment of the
Stoic had apparently engaged less attention. Nevertheless, there
was a more marked difference between the attire of Capito and that
of Pontius Mamurra; for the former was arrayed in a tunic of the
whitest cloth, beneath which appeared fine linen rollers, swathing
his thighs and legs, to protect them, as I supposed, from the heat
and the insects, and a pair of slippers, of dark violet-coloured cloth,
embroidered with silver flowers; while the other held his arms folded
in the drapery of an old but genuine toga, which left his yet strong
and sinewy nether limbs exposed to the weather, all except what
was covered by his tall black sandals and their senatorian crescents.

As we passed on, our host from time to time directed the attention
of his visiters, more particularly of the two Greeks, to the statues of
bronze and marble, which were placed at convenient intervals along
the terraces of his garden. The symmetry of these figures, and the
graceful simplicity of their attitudes, inspired me with I know not
what of calm and soothing pleasure such as I had never before
tasted, so that I thought I could have lingered for ever amidst these
haunts of philosophic luxury. The images were, for the most part,
portraits of illustrious men—Greeks, Romans—sages and heroes; but
beautiful female forms were not wanting, nor majestic
representations of gods and demi-gods, and all the ethereal
imaginations of the Grecian poets. Seeing the [pg 34]name of
Jupiter inscribed upon one of the pedestals, I paused for a moment
to contemplate the glorious personification of might and wisdom,
depositing, at the same time, a handful of roses at the feet of the
statue—upon which I could observe that my behaviour furnished
some mirth to the Epicurean Demochares; while, on the contrary,
Euphranor, the disciple of the Porch, approved of what I did, and
rebuked his companion for saying any thing that might even by
possibility disturb the natural piety of an innocent youth. But the
Roman Stoic stood by with a smile of stately scorn; and utter
indifference was painted on the countenance of Capito. At another
time, Sextus having staid behind to examine the beauties of a
certain statue of Diana, which represented the goddess stretched
out in careless slumber on the turf, with a slender grayhound at her
feet, the Epicurean began to rally me on having a taste inferior to
that of my friend, whose devotion, he said, could not be blamed,
being paid to an exquisite imitation of what the great Nature of
things had decreed should ever be the most agreeable of all objects
in the eyes of a person of his age.—“Whereas you,” continued he,
“appear to be more occupied with deep-hung eye-brows, ambrosial
beards, and fantastic thunderbolts, and the other exuberances of
Homeric imagination.”

To this reproach I made no reply, but Capito immediately began to

recite some verses of a Hymn of Calimachus, in which both the
Greeks joined him; nor could any thing be more delightful than the
harmonious numbers. A sudden exclamation of my friend, however,
interrupted them, and Capito, looking up a long [pg 35]straight
pathway, said, “Come, Valerius, we shall see whether you or Sextus
is the more gallant to living beauties, for here come my nieces. I
assure you, I know not of which of them I am the more proud; but
Sempronia has more of the Diana about her, so it is probable she
may find a ready slave in our Sextus.”
We advanced, and the uncle, having tenderly saluted them, soon
presented us to their notice. Sextus blushed deeply when he found
himself introduced to Sempronia, while in her smile, although she
looked at him as if to say she had never seen him before, I thought I
could detect a certain half-suppressed something of half-disdainful
archness—the colour in her cheeks, at the same time, being not
entirely unmoved. She was, indeed, a very lovely girl, and in looking
on her light dancing play of features, I could easily sympathize with
the young raptures of my friend. Her dress was such as to set off
her charms to the utmost advantage, for the bright green of her
Byssine robe, although it would have been a severe trial to any
ordinary complexion, served only to heighten the delicious brilliancy
of hers. A veil, of the same substance and colour, richly embroidered
with flowers of silver tissue, fell in flowing drapery well-nigh to her
knees. Her hair was almost entirely concealed by this part of her
dress, but a single braid of the brightest nut-brown was visible low
down on her polished forehead. Her eyes were black as jet, and full
of a nymph-like vivacity.

The other, Athanasia, was not a dazzling beauty. Taller than her
cousin, and darker-haired, but with eyes rather light than otherwise,
of a clear, somewhat melancholy gray—with a complexion paler than
is usual in Italy, [pg 36]a demeanour hovering between cheerfulness
and innocent gravity, and attired with a vestal simplicity in the old
Roman tunic, and cloak of white cloth—it is possible that most men
might have regarded her less than the other. A single star of
diamonds, planted high up among her black hair, was the only
ornament she wore.

At the request of the younger lady, we all returned to the grotto, in

the neighbourhood of which, as I have already mentioned, our
tasteful host had placed the rarest of his exotic plants, some of
which Sempronia was now desirous of inspecting. As we paced again
slowly over those smooth-shaven alleys of turf, and between those
rows of yews and box, clipped into regular shapes, which abounded
in this more artificial region, the conversation, which the appearance
of the cousins had disturbed, was resumed; although, as out of
regard to their presence, the voices of the disputants preserved a
lower and milder tone than before. I must confess, however, that
mild as was the manner of the discourse, I could not help being
somewhat astonished, that a polite Roman could permit such topics
to be discussed in the hearing of females; above all, that he did not
interpose to prevent Demochares from throwing out so many
sarcastic reflections concerning the deities whose statues decorated
the garden. A beautiful Mercury, in particular, which we all paused to
admire, elicited many observations, that I could easily see were far
from being agreeable to the fair cousins. But greatest of all was my
wonder at the behaviour of Capito himself, who, after we had again
entered that delightful grotto, turned himself to me as if peculiarly,
and began [pg 37]a deliberate and ingenious piece of declamation
concerning the tenets of his favourite philosophy;—such as the
fortuitous concourse of atoms, the transitory and fluctuating nature
of all things, and the necessity of snatching present enjoyments, as
nothing permanent can be discovered whereon to repose the mind.
With great elegance, indeed, did he enlarge on these golden
theories, nor did he fail to intersperse his discourse with many
exquisite verses from Lucretius and other poetical followers of his
sect. Such, however, was the earnestness of his declamation, that I
could not help believing him to be quite sincere to what he said, and
asked him, not without anxiety, whether he had all his life been an
Epicurean, or whether it was only of late that he had espoused that
discipline.

“Valerius,” said he, “the question is not discreditable to your tender

age and provincial education. To be born wise, Fate or Heaven has
denied to the human race. It is their privilege to win wisdom for
themselves; the fault is their own, if they do not die wise. When the
stripling enters upon the theatre of the world, bright hopes are
around him, and he moves onward in the buoyancy of conscious
power. The pride of young existence is the essence and extract of all
his innumerable sensations. Rejoicing in the feeling of the real might
that is, it is his delight to think—to dream—of might existing and
exerted as for ever. New to the material, but still more to the moral
world, he believes in the stability of all things whose transitory
nature has not been exhibited before him. New to the tricks of
mankind, he believes that to be said truly, which, why it should be
said falsely, he is unable to conjecture. For him, superstition has [pg
38]equal potency to darken the past, and illuminate the future. At
that early period, when ignorance is of itself sufficient to produce a
certain happiness, the ambition is too high to admit such doctrines
as I have no shame in avowing. But time moves on, and every hour
some tender plant is crushed beneath his tread. The spirit clings
long to its delusions. The promise that is destroyed to-day springs
into life to-morrow in some new shape; and Hope, like some warring
deity of your poets, bleeds and sickens only to revive again. But
disappointment at length gathers to itself the vigour of an enduring
form. The horizon becomes colder around us—the soul waxes faint
and more faint within. It is then that man begins to recognize the
true state, not of his own nature alone, but of all things that
surround him—that having tasted much of evil, he is taught to feel
the value of good—and weaning himself from vain-glorious dreams,
learns the great lesson of wisdom, to enjoy the moments as they
pass—to snatch some solid pleasure, at least, amidst a world of
vision and imagination; so, in a word, as the poet has expressed it,
he may not have reason to complain in the hour of death that he
has never lived.

“In me,” he continued, “you behold one that has gone through the
experience necessary to produce an entire acquiescence in these
doctrines. I am one of those, Valerius, who have resolved to
concentrate, after this fashion, the whole of my dreams upon the
hour that is. There are not wanting, indeed, here and elsewhere,
persons who profess the same theories, only in the view of finding
excuse and shelter for the practice of vice. But till it be proved that
the practice of vice [pg 39]is the best means of enjoyment, in vain
shall it be asserted that our doctrine is essentially adverse to virtue.
The mistakes or the misdeeds of individuals must be estimated for
nothing; for where is the doctrine that may not be shewn to have
been defended by impure livers? The founder of our sect is
acknowledged, by its most virulent enemies, to have been the most
blameless of men, and they, I must take leave to believe, can never
be sincere friends of virtue, who doubt, that he who is a true
worshipper of pleasure, may also be the worshipper of virtue.”

There was a certain something, as I thought, more like suppressed

melancholy than genuine hilarity, in the expression of the old man’s
face, as well as in the tone of his voice, while he gave utterance to
these sentiments; nor did any of those present appear desirous of
protracting the argument; although I did not imagine from their
looks that any of them had altered their opinion. What, however, I
could not help remarking in a particular manner, was the gentle
regret painted in the countenance of the elder niece, while Capito
was speaking. The maiden sate over against him all the while, her
cheek supported on her left hand, with an expression of tender
affection. From time to time, indeed, she cast her eye upward with a
calm smile, but immediately resumed her attitude of pensive
abstraction. Her uncle took her hand in his when he had done
speaking, and kissed it gently, as if to apologize for having said any
thing disagreeable to her. She smiled again upon the sceptic, and
walked by herself, (for I could not help following her with my eye,)
down into a dark walk of pines that branched off [pg 40]at the right
hand from the entrance into the grotto. There I saw her stoop and
pluck a pale flower. This she placed in her bosom, and then rejoined
us with a more cheerful aspect; after which, we all walked towards
the villa. Nor did it escape my notice, that, although Sempronia
appeared willing to avoid Sextus as we went, it always happened, by
some accident or other, that he was nearer to her than any other
person of the company.

They were both at a little distance behind the rest of the party, when
Euphranor addressed himself to me, saying, “Is not this youth, your
companion, the same that is under the guidance of a certain
Xerophrastes?”—“The same,” said I, “and a wary, sage-looking
Athenian is his tutor. I believe he also is of the Porch.”—“No doubt,”
interrupted Demochares; “he has a beard that Zeno might have
been proud of, and walks as if he conceived himself to be the chief
pillar of the Porch, if not the Porch itself.”—“Who shall prevent
Demochares from having his jest?” replied Euphranor. “The man is
by birth a Thessalian, and his gutturals still remind one strongly of
his native hills.”—I would gladly have heard more of it, but he was
interrupted by the nearer approach of the rest.

[pg 41]
 CHAPTER V.

Before the hour of taking the bath, we exercised ourselves for some
time in the tennis-court, where I could not but admire, especially
after having heard Capito philosophise, the vigour and agility
displayed by him as well as his companions. I was then conducted
into the baths, where, after being washed and perfumed in the most
luxurious manner, I was arrayed in an elegant supper-garment by
one of the slaves of our host. At table we were joined again by the
ladies, who both reclined on the same couch with their uncle. Three
comely youths attended us, in short tunics, and girt with napkins of
fine linen; but, during the repast, an ancient female slave stood in
silence behind the couch of the young ladies. A small fountain of
alabaster played between two tall candelabra of the same material,
at the farther end of the apartment; and a young damsel stood
beside them, swinging slowly from time to time a silver censer, from
which clouds of delicate odour rolled up to the mirrored roof.

In all things the feast was splendid; but there was no appearance of
useless or vain ostentation. Every thing was conducted in a style of
great calmness and order, without the least formality. The repast
inter[pg 42]rupted not the conversation, which went on in a manner
to me equally instructive as entertaining; although I must confess
the presence of Athanasia sometimes rendered me inattentive to
what was spoken. I could not divest myself of the idea, that some
unknown circumstance was pressing on the mind of the fair
creature, and that when she smiled upon those who addressed her,
it was sometimes to conceal her ignorance of that which had been
said.
Being asked by Capito, I endeavoured, among other things, to
inform him and his friends, as far as I could, concerning the then
condition of this island, which, more particularly after the exploits of
Agricola, had come to be a subject of some interest. In return, the
chief topics which then occupied the capital were discussed by them,
as I perceived, in a great measure on my account; and I listened
with delight to the praises, which they all agreed on bestowing on
the new Emperor. Many anecdotes were narrated, which tended to
strengthen the feelings of admiration, with which I had already been
accustomed to contemplate his character. But others were told, as
the conversation went on, which I could not so easily reconcile with
the idea I had formed of him.

For example, I was somewhat disturbed with what they told me

concerning his treatment of the Christians, who, as we understood in
Britain, had been suffered to live in tranquillity ever since Nerva
acceded to the empire. But now, from the circumstances related, it
appeared that the mild and humane Trajan had taken up, in regard
to this sect, the whole aversion of Domitian; every day some cruel
catastrophe was made known [pg 43]of some person who had
adopted their tenets. Being ignorant of the nature of those tenets,
and having heard only in general terms that they were of Jewish
origin, dark, and mystical, I was at a loss to account for the extreme
hatred of the Prince, or rather for his condescending to give himself
so much trouble concerning a matter so obscure and seemingly
trifling.

Capito, however, assured me, that although I might have good

occasion to wonder at the steps taken by the Emperor, it would no
longer be said by any one, that the progress of the Christian sect
deserved to be considered as a matter either of obscurity, or of no
consequence. “On the contrary,” said he, “from what you have just
heard of the numbers and quality of those that have lately suffered
various punishments, you cannot hesitate to admit that the head of
the empire has been justified in considering it as a subject worthy
his attention.”
“We have adopted the gods of many nations,” said Pontius Mamurra,
“nor do I see why, because the Jews have been unfortunate in a
contest with Rome, we should take it for granted that theirs are
unworthy of respect. If, however, as we have heard asserted, he
who embraces this creed becomes an infidel in regard to the deities
of Rome, I say Cæsar does well in refusing toleration to the
intolerant superstition. Domitian was a tyrant, and a monster of
humanity; the late prince was wise and good; and yet it may be,
that, in regard to these Christians, the principle of Domitian’s
conduct was right in the main, and that of Nerva’s wrong. But you,
Capito, regard both sides of the question, I have no doubt, with the
same indifference.”

[pg 44]
“I hope,” replied Capito, “I shall never regard with indifference any
question, in which the interest of the empire and the honour of
Trajan are concerned. But if you mean only that I am indifferent
about the nature of this Syrian superstition, you are in the right. I
have no knowledge of its dogmas, nor desire to have. I presume
they have their share of that old eastern barbarity, in the shady
places of which the elder Greeks used to think they could discover
the outlines of something really grand and majestic.”

“It may be so,” said Mamurra; “but if the superstition be found

dangerous to the state, the Prince does well in repressing its
progress. That is the only question of which I spake.”—“There is,
indeed, no other,” said Capito; “I thought of none.”—“And how do
you answer it, dear uncle?” said Athanasia, (lifting herself up, for the
first time, to take part in the conversation.)—“Nay, my love,” said the
old man, “to answer that is the business of Cæsar, and of the
Senate. I only regret, that blood should be shed, and citizens exiled;
above all, in the reign of a just and merciful Prince.—Sempronia,”
continued he, “what is that strange story your father was telling
about one of the daughters of Serennius?”
“They only allege,” replied Sempronia, with a smile, “that Tertulla
had a flirtation with a handsome young Greek, and the Greek
happened to be a Christian,—and she was converted by the Greek,—
and she was found out in going with him to some secret assembly of
these people, in a vault by the Vatican Hill,—and her father has been
glad to send her to Corsica, partly to escape the lawyers, and partly,
I suppose, in hopes [pg 45]that the quietness of the island, and the
absence of handsome young Christians, may perhaps, in time,
restore poor Tertulla to her right mind—This is all. Do you think that
a strange story, uncle?” “Not, if it be exactly as you have told it,
Sempronia. What says Athanasia?” Athanasia answered gravely, that
she was sorry for Tertulla, and had never heard any thing of the
young Greek before.

By this time, the increasing darkness of the chamber had warned me

that we ought to be thinking of our return. I had more than once
looked towards Sextus, but he refused to meet my eye. When I was
on the point of speaking, Sempronia, starting from her couch,
exclaimed, that she was sure there was thunder in the skies; and
presently flash after flash gleamed along the horizon. All sat silent,
as if awe-struck; but Sempronia was the only one that seemed to be
in terror from the tempest. Nevertheless, my eyes rested more on
Athanasia, who looked paler than she had done, although her
countenance preserved its serenity. “How awful,” said I, “is the voice
of Jupiter!” Athanasia folded her arms upon her bosom, and lifting
her eyes to heaven, said in a whisper,—“How awful is the voice of
God!” She then dropt her hand on the end of her couch, and half
unconsciously taking hold of it in mine, I asked her if she was afraid.
“No,” said she, “I am not afraid, but the heaviness of the air makes
me faint, and I never can listen to thunder without feeling something
extraordinary within me.”

Capito said, he could not think of our going into the city that
evening, and that we must all make up our minds to remain in the
villa. The countenance of [pg 46]Sextus brightened up, and he
looked to me as if to ask my assent. I was easily persuaded, and our
host despatched a messenger to inform Licinius of the cause of our
absence. The old man then led us into another apartment, which
was richly furnished with books and paintings. Here he read for
some time out of one of the poets, to a party, none of whom, I am
afraid, were very attentive in listening to him, till, the hour of rest
being come, we were conducted to our several apartments, Sextus
and myself, indeed, being lodged in the same chamber.

We were no sooner left alone than I began to rally my friend on the

beauty of his mistress, and the earnest court he had been paying
her. The youth listened with blushes of delight to her praises, but
seemed not to have the least idea that he had been so fortunate as
to make any impression on her mind. On the contrary, he scarcely
appeared to be aware of having done any thing to attract attention
from her, and expressed astonishment when I assured him, that his
behaviour had been such as could not possibly admit of more than
one explanation in the eyes of a person so quick and vivacious as
the lovely Sempronia.

After we had both retired to our beds, and the lights were
extinguished, we still continued for some time to talk over the
incidents of our visit, and the future prospects of Sextus and his
love; until at length sleep overpowered us in easy bonds, and
agreeable dreams followed, I doubt not, in the hearts of us both, the
thoughts and sights of a delightful day. Mine surely were delightful,
for they were all of Athanasia. Yet, even in the visions of the night, I
could never gaze [pg 47]on her face without some strange
impression of mystery. I saw her placid smile—I heard the sweet low
cadence of her voice—but I felt, and I could not feel it without a
certain indescribable anxiety, that her deep thoughts were far away.
[pg 48]
 CHAPTER VI.

I awoke early, and drew near to the bed of Sextus; but seeing that
he was fast asleep, and that a quiet smile was on his lips, I could not
think of awakening him. The sun shone bright into the apartment,
and I resolved to walk forth and breathe the balmy air of the
garden.

The moisture was still heavy on the green paths, and the birds were
singing among glittering leaves; the god-like statues stood
unscathed in their silent beauty. I walked to and fro, enjoying the
enchantment of the scene;—a new feeling of the beauty of all things
seemed to have been breathed into my soul; and the pensive grace
of Athanasia hovered over my imagination, like some presiding
genius of the groves.

I found myself near the favourite grotto, and had stood over against
its entrance for some space, contemplating the augmented stream
as it fell from the superincumbent rock, and regretting the ravage
which the nightly tempest had made among the delicate flowers
round its basin. Twice I thought I heard the murmurs of a voice, and
twice I persuaded myself that it was only the rippling of the waters;
but the third time I was satisfied that some person must be near. I
passed [pg 49]between the water and the rock, and beheld the fair
creature that had been occupying so many of my thoughts, kneeling
far within the grotto, as it seemed, in supplication. To disturb her by
advancing farther, would have been impious; to retire, without the
risk of disturbing her, almost impossible; but I remained there fixed
to the spot, without perhaps considering all these things as I should
have done. The virgin modesty of her attitude was holy in my eyes,
and the thought never occurred to me, that I might be doing wrong
in permitting myself to witness the simple devotions of Athanasia.
“Great God, listen to my prayers,” was all I understood of what she
said; but she whispered for some moments in a lowly and fervent
tone, and I saw that she kissed something with her lips ere she
arose from her knees. She then plunged her hands into the well, by
whose brink she had knelt, and turned round to the light.
“Athanasia, forgive me,” was already on my lips; but on seeing me,
she uttered a faint cry and fell prostrate upon the marble. I rushed
forward, lifted up her head, and laved water from the fountain, till I
saw her lips tremble. At last she opened her eyes, and after gazing
on me wildly for a moment, she gathered her strength, and stood
quite upright, supporting herself against the wall of the grotto.
“Great heavens!” cried I, “in what have I offended, that I should be
rendered the cause of affliction to Athanasia? Speak, lady, and say
that you forgive me.”

“I thought,” said she, with a proud calmness, “that Valerius was of

Roman—of Patrician blood. What brings him to be a spy upon the
secret moments of a Patrician maiden?”—Then bursting into a tone
of [pg 50]unutterable fervour, “Speak,” said she, “young man, what
have you heard? How long have you stood here? Am I betrayed?”

“Witness, heaven and earth!” cried I, kneeling, “and witness every

god, that I have heard nothing, except to know that you were
praying. I have only seen you kneeling, and been guilty of gazing on
your beauty.” “You heard not the words of my prayer?” said she.
“No, not its words, Athanasia, nor any thing of its purpose.” “Do you
swear this to me, young man?” “Yes, I swear by Jupiter and by
Rome—as I am a man and a Roman, I know not, neither do I desire
to know, any thing of what you said. Forgive me for the fault of my
indiscretion—you have no other to forgive.”

Athanasia paused for a moment, and then resuming more of her

usual tone of voice, (although its accents were still somewhat
disturbed and faltering,) said to me, “Valerius, since the thing is so, I
have nothing to forgive. It is you that must pardon me for my
suspicion.” “Distress me not, Athanasia,” said I, “by speaking such
words.” “From this hour, then,” said she, “what has passed here is
forgotten. We blot it from our memories;”—and with that, as if in
token of the paction, she extended to me her hand. I kissed it as I
knelt, and swore that all things were safe with me; but added, as I
arose, “that I was afraid I should be promising more than I should
be able to perform,—did I say I should be able to forget any hour, or
any place, where I had seen Athanasia.” “Nay,” said she, “no
compliment, or I shall begin to suspect you of insincerity.” I was then
about to withdraw from the [pg 51]grotto; but seeing a scroll of
parchment lying at the feet of Athanasia, I stooped, and presented it
to her, saying, “I was afraid she might forget it.” She took it eagerly,
and saying, “Of that there was no danger,” placed it in her bosom,
within the folds of her tunic. She was then gathering up her black
tresses, and fastening them hastily on the back part of her head,
when we heard the sound of footsteps not far off, and beckoning to
me to remain where I was, she darted from me, and in a moment
vanished among the trees. I waited for a few minutes, and then
stepping forth, beheld her walking at a distance, beside her sister, in
the direction of the villa. They were soon lost among the paths, and
I returned alone into the grotto.

I sat down beside the dark well, wherein she had dipt her hands,
and mused in a most disturbed mood on all the particulars of this
strange and unexpected interview. Every motion of her features—
every modulation of her voice, was present with me; I had gathered
them all into my heart, and I felt that I must cherish them there for
ever. From the first moment I saw her, my eyes had been
constrained to gaze upon her with an interest quite novel to me; but
now I knew that she could not smile, without making my heart faint
within me, and that the least whisper of her voice was able to bring
tears into mine eyes. Now I thought of my own unworthiness, and
could not help saying to myself, “Why should a poor ignorant
provincial, such as I am, be torturing myself with the thoughts of
such a creature as this?” Then, again, some benign glance of hers
would return before me, and I could not help having some faint
hopes, that her [pg 52]innocent heart might be won to me by
faithful unwearied love. But what always threw me back into despair,
was the recollection of the mystery that I knew hung over her mind,
although what it was I could not know. That she had been saying
something in her prayers which could not be overheard without
betraying her, she had herself confessed. What could be this secret,
so cherished in dread, and in darkness?—A crime?—No crime could
sully the clear bosom of her innocence. No consciousness of guilt
could be concealed beneath that heavenly visage. But perhaps she
had been made the confidante of some erring,—some unhappy
friend. Perhaps, in her prayer, she had made mention of another’s
name, and implored the pardon of another’s guilt. Last of all, why
might it not be so, that the maiden loved, and was beloved again;
that she might have some reason to regard any casual betrayal of
her affection as a calamity; and that, having uttered the name of her
lover in her secret supplications, her terrors might all have been
occasioned by her apprehensions of my having overheard it? And yet
there was something in the demeanour of Athanasia, that I could
not bring myself to reconcile entirely with any one of these
suppositions. Had she feared that I had overheard any confession of
guilt,—even of the guilt of another,—surely some semblance of
shame would have been mingled with her looks of terror. Had she
apprehended only the discovery of an innocent love, surely her
blushes would have been deeper, and her boldness less. Yet the last
solution of the difficulty was that which haunted me the most
powerfully.

When I came forth into the open air, I perceived [pg 53]that the sun
was already high in heaven, and proceeded in haste towards the
villa, not doubting that Sextus and Capito would be astonished by
the length of my absence. I found them and the ladies walking
under the northern colonnade, having returned, as they told me,
from a fruitless search after me through almost the whole of the
garden. I looked to Athanasia, as if to signify that she well knew
where I might have been found; but, although I saw that she
understood my meaning, she said nothing in explanation. Sextus
drew me aside shortly after, and told me, that his father had sent to
inform him, that our presence was necessary in the city before
supper-time, to attend a great entertainment which was to be given
that evening by the lady whose cause he had successfully pleaded in
the Forum on the preceding day; which lady, I now for the first time
learned, was no other than the same Marcia Rubellia, to whom his
father was very anxious the youth should be married. The success of
this pleading had increased very much the wealth of the lady, and,
of course, as Sextus well knew, the anxiety of Licinius for the
proposed union; and to remain at the villa any longer, was, he said,
entirely impossible, since he already suspected his father had not
been quite pleased with him for leaving the Forum the day before,
without staying to hear out a cause in which his duty, if not his
inclination, ought to have made him feel so greatly interested.

We bade adieu, therefore, to our kind host and the young ladies, not
without more reluctance than either of us durst express, and ready
promises to return soon again to the villa. We found Dromo and
Boto waiting [pg 54]for us at the gate, the former of whom looked
unutterable things, while the latter appeared to be as joyful in
seeing me again, as if we had been parted for a twelvemonth. The
two slaves were mounted on asses, but they led horses for our
conveyance; so we mounted with all speed, and were soon beyond
the beautiful enclosures of Capito. As soon as we were fairly out of
sight of the house, Dromo began to ply Sextus with innumerable
questions about the result of the visit, all of them in bad Greek; that,
as he said, there might be no chance of what passed being
understood by the Druid; for by that venerable designation, he
informed us, the primitive Boto had already come to be best known
in the vestibule of Licinius. “Ah!” quoth he, “there is no need for
many words; I am sure my young master has not been behindhand
with himself. If he has, it is no fault of mine, however. I put
Opportunity into his hands, and she, you know, as the poets say, has
only one lock of hair, and that is in front.”
Welcome to our website – the perfect destination for book lovers and
knowledge seekers. We believe that every book holds a new world,
offering opportunities for learning, discovery, and personal growth.
That’s why we are dedicated to bringing you a diverse collection of
books, ranging from classic literature and specialized publications to
self-development guides and children's books.

More than just a book-buying platform, we strive to be a bridge

connecting you with timeless cultural and intellectual values. With an
elegant, user-friendly interface and a smart search system, you can
quickly find the books that best suit your interests. Additionally,
our special promotions and home delivery services help you save time
and fully enjoy the joy of reading.

Join us on a journey of knowledge exploration, passion nurturing, and

personal growth every day!

ebookbell.com