September 28, 2012

16:18

BC:8312 Chaos in Nature - 9in x 6in

main-ChaosNat

Chapter 1

The Laws of Dynamics

1.1

Keplers Empirical Laws

In order to show the importance of the three-body problem for the emergence of the idea of chaos, we begin by looking at the set of conceptual problems surrounding celestial mechanics at the time when Nicholas Kopernik
(1473-1543) published his De Revolutionibus in 1543. He had the magnicent idea to place the sun near the center of the planetary orbits, following
in the footsteps of his Greek predecessors Philolaus the Pythagorian (5th
century BC), who argued that the earth revolved around a central re, or
Aristarchus of Samos (-310: -230). In spite of this Kopernik assumed that
the only role that the Sun played was to provide light for the Earth. As did
his predecessors, Kopernik was content simply to describe the motions of
the planets, without any real understanding of the causes. Johannes Kepler
(1571-1630) was one of the rst to claim that the Sun was responsible for
the movement of the planets. In fact, in his Mysterium Cosmographicum
of 1596, he attributed a real dynamical role to the sun by arguing that
it was the source of an attracting force on all planets, and that this force
descreased with increasing distance of the planets from the sun.
In spite of this insight, Keplers name is forever associated with the
description of planetary motion in terms of elliptical orbits. Up to then,
planetary orbits had always been described by uniform circular motions or
combinations of uniform circular motions as had been taught by Aristotle
(-384: -322), following Platos lead (-428: -348). The idea that circular
motion was fundamental to the description of celestial motion goes back
at least as far as Platos Timeus. Aristotle justied this by stating that
circular motion was the only perfect type of motion that stars could deign
to follow. It was only after failing to describe planetary motion in this

September 28, 2012

16:18

BC:8312 Chaos in Nature - 9in x 6in

main-ChaosNat

Chaos in Nature

way over a period of ve long years that Kepler became convinced that
planetary motions could only be described by ellipses. He once wrote 1 :
My rst mistake was to assume that the planetary orbits
were perfect circles. This error cost me so much time because
it had been propagated by the authority of all the philosophers,
and further was metaphysically very reasonable.

Aware of the criticism that his contradiction of Aristotelian physics would

incite, Kepler divided his Astronomia Nova (1609) into seventy chapters
where he attempted to show that any other description of planetary motion
contradicted the very precise measurements of Tycho Brahe (1546-1601),
in which he had absolute condence.
Already persuaded of the dynamical role played by the sun, Kepler
naturally placed the sun at one of the foci of an ellipse (Figure 1.1). For
Kepler, the elliptic shape of planetary orbits was explained by two dierent
contributions. The rst one corresponds to a force emanating from the sun
and inuencing each planet. Propagated within the ecliptic plane the
mean plane in which planets orbits were inscribed this force depends on
the inverse of the distance between the sun and each planet. For Kepler,
this force shall induce a circular motion around the sun. A second
magnetic force emanating from each planet was responsible for varying
the distance between each planet and the sun. In other words, the magnetic
force was responsible for the ellipticity and only emanates from each planet,
just because ellipticity was planet dependent. Despite the imperfections of
his system, this was a description of Nature that reduced the role of God
to that of a creator: having given us the laws according to which the world
evolves, He no longer needs to intervene. This view is presented in a letter
dated 10 February 1605 2 :
My aim is to show that the heavenly machine is not a kind
of divine, living being, but a kind of clockwork [...], insofar as
nearly all the manifold motions are caused by a most simple,
magnetic, and material force, just as all motions of the clock
are caused by a simple weight. And I show how these physical
causes are to be given numerical and geometrical expression.
1 J.

Kepler, Astronomia Nova (1609), Chap. xl.

Kepler, Letter dated on February 10, 1605, quoted in I. Peterson, Newtons clock:
Chaos in the Solar system, Freeman, 1993.
2 J.

September 28, 2012

16:18

BC:8312 Chaos in Nature - 9in x 6in

The Laws of Dynamics

main-ChaosNat

P1
111111111
000000000
Earth
111111111
000000000
P4
11111111111111111111111
00000000000000000000000
111111111
000000000
11111111111111111111111
00000000000000000000000
111111111
000000000
11111111111111111111111
00000000000000000000000
111111111
000000000
perihelion 111111111
aphelion
11111111111111111111111
00000000000000000000000
000000000
11111111111111111111111P 3
00000000000000000000000
111111111
000000000
Sun
111111111
000000000
111111111
000000000
111111111
000000000
P2

Figure 1.1 Elliptical trajectory of the Earth around the Sun, which is placed at one
of the foci of the ellipse. The point on the trajectory closest to the sun is called the
perihelion and the furthest point is called the aphelion. The variation in the distance
between the Earth and the Sun was driven by a magnetic force emanating from the
earth.

Kepler believed that the motions of the planets were completely determined, governed by unchanging laws and evolving like clockwork. He succeeded in describing planetary motion, within the limits of Tycho Brahes
observations, by three laws that were stated as follows:
the path of the planets about the sun are elliptical in shape, with
the center of the sun being located at one focus;
an imaginary line drawn from the center of the sun to the center of the planet will sweep out equal areas in equal intervals;
[
areas delimited by P[
1 P2 and P3 P4 are described in equal intervals; thus, the planet moves at the perihelion faster than at the
aphelion (Figure 1.1);
the ratio of the squares of the mean motions 3 , n, is equal to the
inverse of the ratio of the cubes of the semi-major axis a of their
ellipses, that is, the product n2 a3 is constant. This law expresses
the fact that, the more distant from the Sun the planet is, the
slower its motion is. It was published in Keplers book Harmonice
Mundi (1619).
Despite this, if Kepler obtained very good agreement with the observations
of planetary motion, and if his Tables Rudolphines were incontestably the
best of the time, he had more diculty with the problem of the moons
motion, which is particularly sensitive to the motions of the other planets 4 :
how adding or substracting the small lunar orbit, which
could not extend very much beyond the thinness of the Earths
3 The mean motion is dened as the mean speed the planet would have if its motion
was uniform circular.
4 J. Kepler, Le Secret du Monde (1596), Transl. A. Segonds, Gallimard, 1993, p. 150.

October 2, 2012

11:19

BC:8312 Chaos in Nature - 9in x 6in

main-ChaosNat

Chaos in Nature

orbit by very much, could inuence the increase or decrease of

all the other spheres.

He stated, as had Kopernik before him, that he was often faced with large
deviations caused by small errors in either the planetary masses or the
diameters of their orbits 5 :
we are constrained to determine large orbital changes from
very small and almost imperceptible [elements], and from
which, due to a departure by a few minutes, this is 5 or 6 degrees which are lost, and thus a small error can be immensely
propagated.

This rapid growth of small errors in the description of the trajectories of

heavenly bodies is a phenomenon that we run into throughout this work.

1.2

The Law of Gravitation

Keplers elliptical trajectories involve only two bodies, the more passive
being placed at a focus of the ellipse. Combinations of the ellipse with
additional motions are used to take into account the inuence of other
bodies. Before dening the three-body problem we should describe the
interactions among several bodies. The law of gravitation is expressed in
this way.
Towards the middle of the 17th century several scientists attempted to
nd an expression according to which the force exerted by the sun decreases
as a function of the distance r which separates the interacting bodies. Kepler imagined a law depending on the inverse square of the distance as he
got for the propagation of light in his Optica (1604), but he abandoned this
idea when he realized the force seemed to only propagate within the ecliptic
plane, that is, according to a circle perimeter. The force should therefore
depend on the inverse law of the distance since it is varying according to
a circle and not to a sphere. Jeremiah Horrocks (1617-1641) was one of
the earliest to take up Keplers works. While he was a student at Cambridge University he did this himself because, at that time, there was no
course in astronomy, much less courses describing the works of Kopernik,
Tycho Brahe, or Kepler. Horrocks discovered Keplers works by reading
5 N.

Kopernik, Des R
evolutions des orbes c
elestes, Book iii, Chap. 20 quoted by J.
Kepler, Le secret du Monde (1596), Ibid.

September 28, 2012

16:18

BC:8312 Chaos in Nature - 9in x 6in

The Laws of Dynamics

main-ChaosNat

the work of Philips van Lansberge 6 in which this Belgian astronomer defended Kopernik against Tycho Brahe but argued against Keplers ellipses.
Horrocks convinced himself very quickly that Keplers ellipses were more
accurate than uniform circular motion. Horrocks applied Keplers ideas to
the interactions between the sun and the planets 7 :
I, on the contrary, make the planet naturally to be averse
from the Sun and desirous to rest in his own place, caused by
a material dullness naturally opposite to motion and averse to
the Sun without either power or the will to move to the Sun
of itself. But then, the Sun by its own rays attracts and by its
circumferential revolution carries about the unwilling planet,
conquering that natural self-rest that is in it; yet not so far,
but that the planet doth much abate and weaken this force of
the Sun, as is largely disputed afore.

Reviewing Keplers ideas that the Sun is the source of planetary motion,
he proposed a mathematical law describing the decrease in the force with
distance between the sun and planets. In his theory of the Moon, Horrocks
observed that although the Moon is between the Earth and the Sun, the
ellipse described by the Moon with the Earth at one of its foci is
elongated: this is because the eccentricity and apogee vary in time. He
also studied cometary motion which he found to be almost elliptical
and the tides. All these ideas found their way into Newtons Principia.
In fact, among Horrocks contemporaries at Cambridge was John Wallis
(1616-1703), whose works were read attentively by Newton, and Ralph
Cadworth, who inuenced Newtons mechanistic philosophy. We point out
that Wallis, who directed Newton to the calculus of changes, was charged
by the Royal Society of London with the publication 8 of Horrocks works
in 1666, the year during which Newton declared that he began his studies
of planetary motions. Newton held a very high opinion of Horrocks 9 :
the hypothesis of Horrox which is the most ingenious, and,
if I do not deceive myself, the most accurate of all
6 P. van Lansberge, Tabulae motuum caelestium perpetuae: ex omnium temporarum
observationibus consentientes, 1632.
7 J. Horrocks, Philosophical exercices, Part 1, para 26, (1661), quoted by Robert Brickel,
A Chapter of Romance in Science 1639-1874: In Memoriam Horroccii and as mentioned
by V. Barocas, A country Curate, Quaterly Journal of the Royal Astronomical Society,
12, 179-182, 1971.
8 Horrocks Opera were published in 1672.
9 I. Newton, The System of the world.

September 28, 2012

16:18

BC:8312 Chaos in Nature - 9in x 6in

main-ChaosNat

Chaos in Nature

and in the rst edition of Principia

10

Our Countryman Horrox [who] was the rst who advanced

the theory of the moons moving in an ellipse about the Earth
placed at its lower focus.

Horrocks was an important bridge between the works of Kepler and those
of Newton.
Christiaan Huygens (1629-1695) had studied uniform circular motion
and established that the centrifugal force followed a 1/r2 law. In addition,
Robert Hooke (1635-1703) and Christopher Wren (1632-1723) tried to establish a relation between the 1/r2 law and the elliptical trajectories of
celestial bodies. Specically, Hooke developed a qualitative model based
on the principle of inertia, correctly stated by Descartes 11 :
Each particular part of matter continues always to be in the
same state unless collision with others constrains it to change
that state.

He coupled this principle with the equilibrium between the centrifugal force
and the force due to the Sun. If he found motion qualitatively in accord
with observations, he was not able to obtain results quantitatively in accord
with Keplers Laws. Nevertheless, he claimed that the force was not only
between the sun and the planets, as it was in Keplers Astronomia, but
between any massive bodies 12 :
all Celestial Bodies whatsoever, have an attraction or graviting power towards their own Centers, whereby they attract
or not only their own parts, and keep them from ying from
them, as we may observe the Earth to do, but that they do
also attract all the other Celestial Bodies that are within the
sphere of their activity; and consequently that not only the Sun
and Moon have an inuence upon the body and motion of the
Earth, and the Earth upon them, but that also, and by their attractive powers, have a considerable inuence upon its motion
10 I.

Newton, Principia Mathematica, Book iii, Scholium 475, 1687.

Descartes, Le Monde: ou Trait
e de la lumi`
ere, in Oeuvres de Descartes, ed. AMTannery (Paris: L. Cerf, 1897-1913), 11, p. 435 (1677 pagination). English translations
are in Ren
e Descartes, Le Monde, trans. M. S. Mahoney (New York: Abaris, 1979) p.
61.
12 R. Hooke, An attempt to prove the Motion of the Earth from Observations, 11674.
11 R.

September 28, 2012

16:18

BC:8312 Chaos in Nature - 9in x 6in

The Laws of Dynamics

main-ChaosNat

as in the same manner the corresponding attractive power of

the Earth has a considerable inuence upon every one of their
motions also.

It was Isaac Newton (1642-1727) who reaped the honor of showing that
a 1/r2 force leads directly to Keplers laws. To do this he resorted to
very subtle geometric arguments that related motion to innitely small
displacements. He studied dierent laws of the form 1/rn and stated, in
Book iii of his Principia Mathematica (1687) that gravity follows a law in
which it decreases like the inverse square of the distances 9 :
Proposition ii: That the forces by which the primary planets are continually drawn o from rectilinear motions, and retained in their proper orbits, tend to the sun; and are reciprocally as the squares of the distances of the places of those
planets from the suns centre.

and is universal:
Proposition vii: That there is a power of gravity tending
to all bodies proportional to the several quantities of matter
which they contain.

With these two Propositions, Newton unied celestial mechanics with

the empirical laws of Kepler and the terrestial mechanics of falling bodies
studied by Galilei (1564-1642): he showed that a single force was responsible
for all these phenomena. To show that Keplers Laws result from a 1/r2
force law Newton used, among other arguments, the idea of an accelerating
force introduced in Book i of the Principia:
Lemma x: Spaces which a body describes by any nite force
urging it, whether that force is determined and immuable, or
is continually augmented or continually diminished, are in the
very beginning of the motion one to the other in the duplicate
ratio of the times.

It was in this form that the fundamental principle of dynamics was presented by Newton. We had to await the development of dierential calculus, independently discovered by Newton and the school of Gottfried
Leibniz (1646-1716) for a modern formulation of this principle, stated by

September 28, 2012

BC:8312 Chaos in Nature - 9in x 6in

16:18

main-ChaosNat

Chaos in Nature

10

Leonard Euler in 1747 13 : the product of the mass and the acceleration of
a body is equal to the sum of the forces on the body.
Using this fundamental dynamical principle, the geometries, as it was
called, were able to show the relation between changes in motion and the
forces that caused them. Any system in which motion results from applied
forces is a dynamical system, in the sense that its behavior evolves in time.
Since this fundamental principle of dynamics establishes a relation between
an acceleration and the applied forces, it is expressed by a dierential equation which involves the second derivative of the position the acceleration
and the position evolving under the inuence of the forces on which it
depends 14 .
In the meantime, the dierential equations describing the problem of
two massive bodies under gravitational interaction were rst written in 1710
by Jacob Hermann (1678-1733) 15 , one pupil of Johann Bernoulli (16671748). Both of them opened the door for Eulers principle. They solved
the dierential equations and conrmed Newtons result that the laws of
motion are, in this case, exactly Keplers Laws. Thus, when a single planet
orbits the sun, it moves exactly according to the laws found by Kepler. The
planet follows an elliptical trajectory that repeats itself forever: this is a
periodic solution. Keplers Third Law gives us the period of this solution
as a function of the distance from the sun as measured by the semi-major
axis a of the ellipse (in astronomical units):

a
1/n
n2 a3

Mercury

Venus

Earth

Mars

Jupiter

Saturn

0.38
88
7.09

0.72
224
7.44

1.00
365
7.50

1.52
687
7.44

5.2
4307
7.58

9.5
10767
7.39

where 1/n corresponds to the inverse of the average motion, that is, to the
period of revolution, given in Earth days. Only the planets known to Kepler
13 L. Euler, D
ecouverte dun nouveau principe de m
ecanique, M
emoires de lAcad
emie
des Sciences de Berlin (1750), 6, 185-217,
1752. The modern writing of this principle
where m is the mass of the body, a the
has the form of a vectorial
F
equation ma =
acceleration vector and
F is the sum of the forces applied on the body.
14 Writing this in a dierential form came after the contribution by Pierre Varignon
(1654-1722) who associated the concept of acceleration with the rst time derivative of

the velocity, a =

d
v
,
dt

or to the second time derivative of the position, a =

d2 OM
. The
dt2
2
d x
= F (x)
dt2

fundamental principle of dynamics is written as the dierential equation

which is a second-order dierential equation since it involves a second time derivative.
15 J. Hermann, Letter to J. Bernoulli, July 12, 1710. in J. Bernoulli, Opera, 85.

September 28, 2012

16:18

BC:8312 Chaos in Nature - 9in x 6in

The Laws of Dynamics

main-ChaosNat

11

are presented in this Table but the near constant value of the product n2 a3
is satised for all the planets (major and minor) known today. As a result,
we can consider the two body problem as solved, and that Keplers Laws
describe their solutions.
1.3

Theory of the Moon

Newton had already been interested in the problem where three bodies interacted gravitationally. This conguration occurs when the action of a
second body on a third cannot reasonably be neglected compared to the
action of the rst body. In our solar system there are two systems where
this type of interaction is particularly important. These are the Sun-EarthMoon system, because of the close proximity of the Moon to the Earth, and
the Sun-Jupiter-Saturn system, because of Jupiters large mass. Nevertheless, Newton treated the Sun-Earth-Moon system as a perturbation of the
two-body Earth-Moon system 16 :
The area which the moon describes by a radius drawn to
the earth is proportional to the time of description, excepting
in so far as the moons motion is disturbed by the action of
the sun, and here we propose to investigate the inequality of
the moment, or horary increment of that area or motion so
disturbed.

He specically stated that the plane of the orbit oscillates and that the
eccentricity also varies. Even so, the theory of the Moon that he published
in his Principia was very imperfect, as he himself stated in the preface to
this work. In order to produce a better theory, Newton undertook a more
general study of the three body problem but immediately ran into major
diculties. As a result he adopted a simplied form of this problem by
either by choosing a 1/r form of the gravitational law in order to simplify
the computations, or else by assuming a very massive central body whose
motion was almost unaected by the motion of the two other bodies, or else
taking one of the three bodies very far away from the ohter two. The general
problem was treated in a not very clear way in 22 corollaries in which he
proposed a semi-qualitative evolution of the motion. Even though Newtons
mechanics had already explained tidal phenomena and the attening of the
16 I.

Newton, Principes Math

ematiques de la philosophie naturelle, Livre iii, Proposition
xxvi, 1687.

September 28, 2012

12

16:18

BC:8312 Chaos in Nature - 9in x 6in

main-ChaosNat

Chaos in Nature

Earth at its poles, it was unable to solve the problem of describing the
motion of three bodies under their mutual gravitational interaction.
Because it is very simple to state but also because it is at the heart of
the theory of the Moon, the three body problem has been at the center of
mathematical and mechanical studies for close to two centuries.