Mercurys Perihelion

Chris Pollock
March 31, 2003

Introduction In 1846, the French Astronomer Le Verriere, doing calculations based on Newtons theory of gravitation, pin-pointed the position of a
mass that was perturbing Uranus orbit. When fellow astronomers aimed their
telescopes at his location, they recognized, for the first time, the eighth planet.
Newtons theory had reached its zenith. Shortly after, however, it became clear
to Le Verriere that additional mass, nearer to the sun than Mercury, was needed
to explain the strange advance of Mercurys orbit. When no such mass was observed, astronomers began to doubt Newtons theory. Then, along came Albert
Einstein, whose theory nearly perfectly explained Mercurys erstwhile mysterious motion. This essay is a history of Newtons theory of gravity, the enigma
of Mercury, and Einsteins convincing solution. It will blend together mathematical and physical theories with a narrative about the brilliant scientists who
chose to tackle the problem of gravitation. In particular, this paper will show
the following computations:
1. A derivation of Keplers First Law concerning elliptical orbits from Newtons Law of Gravitation and Newtons Second Law.
2. An outline of a method to approximate non-relativistic perturbations on
Mercurys orbit by assuming external planets are heliocentric circles of
uniform linear mass density.
3. A calculation of relativistic perihelion shift using Einsteins theory of relativity and the Schwarzschild solution.
Pre-Newtonian Theories and Ideas The problem I set out to explain,
the advance of Mercurys perihelion, was of tantamount importance to the astronomical community. This seemingly miniscule enigma stood glaringly in the
way of humanitys understanding of the solar system. Lets begin by reviewing
the early evolution of celestial theories.
In the second century CE, the Greek astronomer Claudius Ptolemy hypothesized that the sun, moon, and planets orbited the earth along circles called
deferents. On a smaller scale, they travelled along smaller circles, called epicycles, whose centres moved along the deferents. This theory was sufficiently
accurate to explain observations of the time. It dominated until around 1843,
when Copernicus published his book, de Revolutionibus Orbium Caelestium.
His solar system was heliocentric but retained the orbital deferents and epicycles of Ptolemys theory. It is interesting to note that Copernicus had difficulty
explaining Mercurys motion and once commented that this planet has . .
. influenced many perplexities and labours on us in our investigation of its
wanderings. (from Baum, 11)
More exact naked-eye planetary observations were taken by Tyco Brahe,
who, I discovered while writing this paper, did not cut off his own nose. I had
always believed, courtesy of my eleventh grade science teacher, that the great
Danish observational astronomer removed his nose to allow his face a better
seal with his telescope. My teacher, it turns out, was doubly dishonest since

Brahe lost part of his nose in a duel with a student and telescopes had not been
invented in 1600, when Brahe made his observations.
A young mathematician who worked with Brahe examined the elders notes
and tried to calculate the orbit of Mars. Although an apparently simple task,
this was disconcertingly complex since planets appear to change speed and direction based on the motion of not only the planet but also Earth.
When Kepler finished his calculations, he determined that Mars moved in an
ellipse with the sun at one focus. Kepler became quite proficient at predicting
Mercurys passes between the Sun and Earth, called transits. By hand, he
calculated Mercurys 7 November 1631 transit time, accurate to within five hours
(Baum, 15), although he died before witnessing the event he had predicted. A
subsequent transit, on 23 October 1651, was predicted, using corrections of
Keplers calculations, with an accuracy of a few minutes.
Based on his planetary observations, Kepler made the following statements,
known collectively as Keplers Laws (Stewart, 897):
1. A planet revolves around the Sun in an elliptical orbit with the Sun at
one focus
2. The line joining the Sun to a planet sweeps out equal areas in equal times.
3. The square of the period of revolution of a planet is proportional to the
cube of the length of the major axis of its orbit.
And so, by the middle of the 17th century, kinematic data on planetary paths
were fairly accurately known. Hence, all of the pieces were on the table for a
brilliant theorist (read: Newton) to assemble, explaining the dynamics behind
planetar motion.
Newtons Law Issac Newton was born on Christmas Day, 1642. Since
he was a failure at farming, his mother sent him to university. In the middle
of his study at Cambridge, the plage broke out, and the school was closed
to students for 1665-1666. Newton returned home, and in this marvellously
creative period, wrote about both gravitation and calculus. On the former
question, he considered a rock twirling around on the end of a string. The
rock, he knew, tended to launch, but the string provided a counteracting force.
Newton wondered, then, what provided a counteracting force in the case of
planetary motion. Could it be gravity, the force that held people on the Earths
surface?
In his magnum opus, 1687s Philosophiae Naturalis Principia Mathematice,
Newton calculated that the path of planets would be elliptical if they were
subjected to a force of gravitation that varied with the inverse square of the
separation between the planets and the Sun. Starting from Newtons Second
Law and the Newtons Law of Gravitation, it is possible to prove each of Keplers
Laws. Since Keplers First Law deals with the elliptical shape of planetary
orbits, a derivation is included below. The discussion that follows is based on
Chapter 11.4 of Stewarts Calculus. This is the so-called one-body problem
since the system under consideration contains a test particle (the planet) moving

under the attraction of a massive body (the Sun). The planets mass is assumed
to be so small relative to that of the Sun that the Sun remains fixed in space.
Also, the effects of all other planets are neglected.
Lets begin by stating Newtons Second Law:
F = ma

(1)

where F is the force experienced by the planet, m is its mass, and a is its
acceleration.
And Newtons Law of Gravitation:
GM m
GM m
F=
r=
u
(2)
3
r
r2
The coordinate system is radial. F is the force experienced by the planet, G
denotes the gravitational constant, r = r(t) is the planets position vector, u is
a unit vector in the direction of r, and M and m are the masses of the Sun and
planet, respectively. Further, r = |r|, v = r0 , and a = r00 .
First, we will show that the planet moves in a plane. Equating the Fs from
(1) and (2) gives
GM
a=
r
(3)
r3
hence a and r are parallel, so
ra=0
From the properties of cross products, we know that
d
(r v) =
dt
=
=
=

r0 v + r v0
vv+ra
0+0
0

So, we can conclude that r v is a constant vector, say h. We can assume

that r and v are not parallel. This implies that h 6= 0. We can now conclude
that r and h are perpendicular, so the planet always lies in a plane through the
origin perpendicular to h. Planetary motion, according to Newtons Laws, is
planar.
Now, we can prove examine the shape of the orbit within this plane. Lets
start by rewriting h:
h =
=
=
=
=
=

rv
r r0
ru (ru)0
ru (ru0 + r0 u)
r2 (u u0 ) + rr0 (u u)
r2 (u u0 )

So, combining the above result with (3) gives

GM
u r2 (u u0 )
r2
= GM u (u u0 )
= GM [(u u0 )u (u u)u0 ]

ah =

Since u is a unit vector, u u = |u|2 = 1, so

a h = GM [(u u0 )u u0 ]

(4)

Now, lets consider the (u u0 ) term in (4). Since u u is constant,

d
(u u) = 0
dt
u0 u + u u0 = 0
u u0 = 0

Substituting into (4) we get

a h = GM u0
But h is a constant vector, so
(v h)0

= v0 h
= ah
= GM u0

(v h)0
Integrating both sides yields

v h = GM u + c

(5)

where c is a constant vector.

Since the planets motion is confined to an arbitrary plane, we can, without
loss of generality, call this the xy-plane. Recalling that h is perpendicular to this
plane, lets define a standard basis vector k in the direction of h. Looking at (5)
and recalling that (v h) and u both lie in the xy-plane, we can conclude that
c is also in the xy-plane. Lets choose the x - and y-axes such that the standard
basis vector i is in the direction of c. There is really nothing special about
this choice. Any set of axes would do just fine. Lets go one step further and
call the angle between c and r. Hence (r, ) is the planets position in polar
coordinates. If, at this point, it is unclear where we are heading, recall that we
set out to show that the planets path is elliptical. So, ultimately, were seeking

an expression for r(). Lets look at an expression for r (v h), substitute in

(5), and turn the crank . . .
r (v h) =
=
=
=

r (GM u + c)
GM r u + r c
GM ru u + |r||c| cos
GM r + rc cos

where c = |c|.
Now, lets solve for r in the previous equation.
r=
Define e =

c
GM ,

r (v h)
GM + c cos

and substitute into the above equation:

r=

r (v h)
GM [1 + e cos ]

(6)

So, we now have the desired equation for r. Lets simplify the numerator of
(6).
r (v h) = (r v) h
= hh
= h2

where h = |h|. So, (6) becomes

r

=
=

h2
GM [1 + e cos ]
eh2
c[1 + e cos ]

And, for one final simplification, lets rename

r=

h2
c

ed
1 + e cos

as d. So
(7)

This is the equation of a conic section in polar coordinates. The value of e

in the denominator determines the type of conic section (7) represents. Since
planets have closed orbits, 0 < e < 1 and r represents an ellipse. Further, e is
the eccentricity of Mercurys orbit, and one of the ellipses foci is located at the
origin. So, beginning with only Newtons Second Law and Law of Gravitation,

we have verified Keplers First Law, that planets moves in an elliptical orbit
with the sun at one of the foci.
Of course, the actual solar system is far more complex than this model.
Newtons central idea, that planetary motions are governed by an inverse square
force of gravity, gave mathematicians and astronomers the tools to explain the
clockwork motion of the planets, as long as precise masses and positions were
known for all planets. Shortly after Principias publication, scientists began to
use these laws to make successively more accurate predictions.
Applying Newtons Amazing Law of Gravitation Comet Halley had
been predicted to return, by Halley himself, in winter 1758-59. In 1757, AlexisClaude Clairault, of France, and two assistants rushed to find a more precise
return date. Using Newtons Law of Gravitation and considering perturbations
caused by Jupiter and Saturn, they arrived at a date of mid-April 1759. Comet
Halleys perihelion, the point in its orbit nearest to the sun, was only 33 days
earlier than they had predicted. This error was due mainly to computational
shortcuts they made to ensure that they finished their calculations before Comet
Halley arrived.
William Hanover, who was born in Hanover in 1738 but moved to England in
1757, made a shocking discovery in 1781, making use of the telescope. He found
a seventh planet orbiting the Sun. Uranus became the first planet discovered
since antiquity. As predicted by Newtons theory, Uranus followed an elliptical
path. The astronomers of the world excitedly turned their attention to this new
planet.
There is one figure who played so important a role in verifying and employing
Newtons amazing theory that it is necessary to look at his life in more detail.
That man is Urbain Jean Joseph Le Verriere, born n Normandy in 1811. He
attended Ecole Polytechnique and graduated with a degree in chemistry. He
received an appealing job offer that required him to move away from Paris and
his girlfriend. He declined the job, chose marriage, gave up chemistry, and ended
up in astronomy through a fortuitous job offer.
In 1839, he began calculating the stability of the solar system and planetary
orbits. In 1841, then, he applied his unique analytic skills and almost superhuman endurance for calculation to master the motion of Mercury. (Baum, 71)
This planet orbits nearest the sun and has a large eccentricity and short period.
A detailed description of its motion had, so far, eluded theorists and stood out
as a critical test for Newtons Gravitation. In 1845, Le Verriere gained reknown
for calculating Mercurys 1845 transit of the Sun to within 16 seconds (Baum,
73). Unsatisfied by this error, however, he did not publish his tables of Mercurys motion. At this point, he left Mercury (!) temporarily and turned to
another problem.
Since the discovery of Uranus, its orbit had been lagging behind predictions.
Le Verriere, always sure of Newton, set to calculating the position and mass of
an object that could bring about the observed lag in Uranus orbit. He began
with an assumption that the unknown mass must be exterior to Uranuss orbit,
since a perturbation had not been observed in Saturns orbit. He then tried to
calculate the location, mass, and orbital parameters of such an object.
6

In mid-1845, a young Cambridge mathematician, John Couch Adams, was

also considering the same problem. Adams was, in fact, the first to theorize the
new planets position, but his results were not well disseminated. By 1846, the
famous Le Verriere too issued a prediction as to the new planets position, and a
team at the Berlin Observatory found it almost immediately, only 55 arc minutes
from Le Verrieres calculation, and 2.5 degrees from Adams. The director of the
observatory exclaimed that this was the most outstanding conceivable proof
of the validity of universal gravitation (Baum, 117). Indeed, Newtons theory
had reached its peak. Its power, in the hands of a talented mathematician or
astronomer like Le Verriere of Adams, seemed without bounds. With another
success behind him, Le Verriere then returned to the problem of Mercury.
The Hunt for an Intermercurial Planet In 1854, Le Verriere was appointed director of the Paris Observatory. He undertook, with the assistance of
many human computers, a study of the whole solar system. His dream was to
explain every minute movement of the jewelled clockwork that seemed, for an
elusive moment, to be stable, self-adjusting, and eternal (Baum, 2).
In 1859, he began examining observations of Mercurys motion in detail,
relying mainly on 14 very accurate solar transit times, recorded between 1697
and 1848. He concluded that the ellipse of Mercurys orbit was precessing slowly,
which was expected, since the other planets of the solar system were expected
to Mercurys perihelion to precess. The observed precession amounted to a
perihelion advance of approximately 565 seconds of arc per Earth century. Le
Verriere then calculated an expected precession by considering the effects force
each planet exerted on Mercury. This was an accurate, though computationally
demanding technique. The outer planets, Le Verrier predicted, should cause an
advance of 527 seconds of arc per century, leaving a residual 38 seconds that
he was not able to explain using Newtons theory (Baum, 136). In the next
section, we will calculate an approximation of this Newtonian perturbation.
An Approximation of the Newtonian Precession of Mercurys Orbit It is possible to model, very precisely, the Newtonian force each planet was
expected to exert on Mercury. An elegant approximation, however, was described by Price and Rush in their paper entitled Non-relativistic contribution
to Mercurys perihelion precession. They replaced each of the outer planets
by a ring of uniform linear mass density. Since Mercurys precession is slow
compared to the ortits of even Uranus, their approach yields a fairly accurate
time-averaged effect of the moving planets effects. An sketch of their calculations follows.
First, lets replace each planet by a ring with linear mass density given by
the follwing equation:
Mi
i =
2Ri
where i is the linear mass density along the orbit of the ith planet from the
sun, Mi is the mass of the ith planet, and Ri is the radius of the ith planets
orbit, which is assumed to be circular.
We can calculate the gravitational pull exerted on Mercury by each of the
exterior planets by considering the radial components of the force exerted by
7

differential angular elements on Mercury using Newtons Law of Gravitation

and integrating over the circle. Please see page 532 of Price and Rushs paper
for further details on this integration.
This yields
Gi ma
Fi = [ 2
]u
(8)
Ri a2
where Fi is the radial force exerted on Mercury by the ith planet, a is Mercurys
distance from the Sun, and u is a unit vector for Mercurys position. Note that
since Ri > a, Fi in (8) is positive, so the force exerted on Mercury by each
external planet is directed outward, opposite to the force exerted by the Sun.
Now, we will refer to some results from Classical Mechanics. From the requirements that planetary orbits are stable and closed, the following differential
equation relates gravitational force and planetary motion:
(r) = m(
r 2 r)
where (r) is the total applied force, and represents differentiation with
respect to time.
is a constant, say h. SubstiAlso, the planets angular momentum, mr2 ,
tuting this value into the above equation yields
(r) = m(
r

h2
)
m2 r 3

Now, lets consider an orbit for Mercury that oscillates about a circular
orbit of radius a. Using a Taylor series approximation and solving the resulting
differrential equation gives an equation for apsidal angle, which is defined as the
angle between the perihelion and aphelion:

= q
0 (a)
3 + a[ (a)
]

(9)

where is the apsidal angle, (a) is the net central force, and a is the radius
of the circular orbit around which we are perturbing.
Now, lets use (8) to find the sum of the forces of all of the external planetary
rings on Mercury:
F (a) = Gm

9
X

i a
= 7.587 1015 N
a2

Ri2
i=2

(10)

The numerical value was obtained by substituting masses and orbital radii for
all external planets into (10).
The Gravitational force exerted by the Sun on Mercury is F = 1.318
1022 N , and the net force that Mercury experiences is given by (a) = F +F (a).
Lets differentiate this expression and multiply by a.
a0 (a) = aF0 + aF 0 (a)
8

(11)

Lets now examine the terms on the right-hand side of (11). Differentiating
(10) yields:
9
X
i (Ri2 + a2 )
0
F (a) = Gm
(12)
Ri2 a2
i=2
And F , the magnitude of the force exerted on Mercury by the Sun, is given
by Newtons Law of Gravity
F =

GMs m
a2

where Ms is the mass of the Sun. Differentiating,

F0 =

2GMs
2
=
F
a3
a

aF0 = 2F

(13)

Substituting (12) and (13) into (11) gives the following for Mercurys apsidal
equation:

= s
(14)
2F +Gma

3+

P9

i=2

R2 +a2
i
(R2 a2 )2
i

F +F (a)

We now apply a binomial expansion on both the numerator and denominator

of (14) and neglect all terms higher than first power of FF(a)
:

F (a) Gma
= (1

P9

i=2

R2 +a2

i (R2ia2 )2

2F

(15)

Now, compute the numerical value for by substituting the value of F (a)
calculated in (10), F , and mass and orbital radius for the exterior planets into
(15):
= (1 + 9.884 107 )
where is the angle between perihelion and aphelion. Please see p533 of Price
and Rushs article for the numerical valuse used in calculating .
The precession of Mercurys orbit per revolution, then, is
2 2
T
where T is the period of Mercurys orbit, 87.969 days. So, in more conventional
units,
531.9arcsec
precession =
(16)
century
Le Verriere predicted a perihelion advance of 527 arcseconds per century,
and this approximation yields 531.9 arcseconds. Still, Newtons theory was not
able to explain the observed perihelion precession using known masses in the
9

solar system. Le Verrier, naturally, tried to apply the same technique that had
led him to the discovery of Neptune and began looking for a missing body in
the solar system.
The Hunt for an Intermercurial Planet An increase of approximately
10 percent in Venus mass would explain Mercurys perihelion advance, but it
would also affect Earths orbit in a way that had not been observed. Since the
missing mass must not affect Earth, Le Verriere decided that it must be nearer
to the Sun than Mercurys orbit. And so began the hunt for Le Verrieres ghost
planet, or rather planets. He quickly realized that a single planet so near the
Sun would have been bright and, hence, visible during solar eclipses. Since no
such planet had been observed during past eclipses, Le Verriere hypothesized,
instead, that the mass was in the form of many small bodies.
In 1860, he visited a small-town French doctor and part-time astronomer
who had recorded observations during what he believed to be a transit of an
intermercurial planet. Le Verriere, convinced by the mans story, released the
news of the new planet, which was quickly dubbed Vulcan. He had, again,
delighted the French scientific community. The team of Newton and Le Verriere
had again triumphed. Or so it seemed.
Based on the doctors observations, Le Verriere calculated the planets distance from the sun, 0.147AU, and period, 19 days 17 hours (Baum, 156). The
astronomical community tried again and again to observe the elusive planet
Vulcan, but, as time passed with no further sightings, doubts as to the planets
existence began to mount. With so much attention directed at observing the
area around the sun during subsequent eclipses, and with no, or at least very
few, credible sightings, most astronomers lost faith in Vulcan. Until his death
in 1877, Le Verriere remained utterly convinced that the missing mass existed
and would eventually be found, verifying the supremacy of Newtons Law of
Gravitation.
By 1890, however, almost nobody believed a sufficient amount of matter
would be found inside Mercurys orbit to explain the perihelion advance. In
1895, Newcomb corrected some inconsistencies in planetary mass and repeated
Le Verrieres calculations. He found an extra perihelion shift of 43 per century,
slightly larger than Le Verrieres result. Since hidden mass was, by this time,
out of the question, perhaps, he thought, the problem lay with Newtons Law of
Gravitation. If, for example, the exponent in the denominator were 2.00000016
(Baum, 233), then the motion of Mercury could be more accurately explained.
This represented a true paradigm shift. While formerly observations were
questioned and Newtons theory was unassailable, now credible scientists began
to question the foundations of Newton Law of Gravitation. The observational
problem was, essentially, closed. The theory problem, in contrast, was wide
open.
Einsteins General Theory of Relativity In November 1915, Albert
Einstein, sitting at a desk in Berlin, wrote his General Theory of Relativity. It
is ironic to note that Newtons theory had peaked in the same town, only 70
years earlier, when Neptune was discovered at the Berlin Observatory.
Einstein did not set out to solve the problem of Mercurys perihelion shift.
10

In fact, the answer fell out as a neat consequence of his theory. In Newtons
theory, as demonstrated previously, the motion of planets are ellipses with the
Sun at one focus. Here, I will show that Einsteins Theory of General Relativity
changes this conclusion, although this change is minor in most cases.
The general relativistic calculation that follows will consider the motion of a
test particle in the gravitational field of a massive body. The test particles mass
is assumed to be so small that it has no effect on the massive body. Fortunately,
the Schwarzschild solution to Einsteins field equations describes precisely this
case.
Here, I will assume that the reader if familiar with the Schwarzschild solution. A good derivation and discussion of the Schwarzschild solution can be
found in Hans Stephanis General Relativity. Schwarzschilds spherically symmetric vacuum solution has a line element
2m 2
1
)dt
ds2 = (1
dr2 r2 (d2 + sin2 d2 )
(17)
r
1 2m
r
where m, in our case, is the mass of the sun in relativistic units.
Starting from the Schwarzschild line element, it is possible to deduce the
motion of a test mass (a planet). Please refer to section 15.3 of Ray DInvernos
Introducing Einsteins Relativity for further details. Before we begin, though, it
is interesting to note that Einstein reached the same conclusion about Mercurys
perihelion precession in 1915, without the use of Schwarzschilds solution.
Since the test mass (Mercury) moves along a timelike geodesic, the Lagrangian is identical to kinetic energy, and g x x = 1, where , = 0, 1, 2, 3.
So, the Lagrangian, L is as follows:
m 2
v
L =
2
m
dx dx
=
g
2
d d
m
=
g x x
2
where x = dx
d and is proper time.
So, from (17), the Lagrangian for Mercurys force-free motion is given by
the following equation:
L=

m
m
2m 2
1
r 2 r2 2 r2 sin2 2 ] =
[(1
)t
2
2
r
1 2m
r

(18)

Let x0 = t, x1 = r, x2 = , and x3 = . Recall the Euler-Lagrange equation:

L
d L

(
)=0
xa
d x a

(19)

Lets now apply the Euler-Lagrange equation to (18). For the a = 0 case:
d
2m
[(1
)t] = 0
d
r
11

(20)

For the a = 2 case:

d 2
(r ) r2 sincos 2 = 0
d

(21)

And for the a = 3 case:

d 2 2
(r sin ) = 0
d

(22)

It is not necessary to take the a = 1 case, which corresponds to r. Since

Mercurys motion involves four equations, t = t( ), r = r( ), = ( ), and
= ( ), equations (18), (20), (21), and (22) provide sufficient information.
We have chosen to omit the a = 1 case since differentiating (18) with respect to
x1 = r will clearly make the biggest mess.
As with the Newtonian case, we first must establish whether planar motion
is possible. Lets consider motion in the equatorial plane and check that it is
stable. So, at = 0, lets assume that = 2 and = 0. Since = 2 , (21) gives
d 2
(r ) = 0
d
r2 = constant
At = 0, r2 = 0, so r2 = 0 for all . If we assume that r 6= 0, then = 0
for all . Hence = 2 for all , and planar motion is possible. Lets consider
motion in the xy-plane.
Integrating (22) gives
r2 = h
(23)
where h = constant. Hence, h represents angular momentum. Next, integrate
(20) to give
2m
(1
)t = k
(24)
r
where k = constant.
Now, substitute (24) and = 2 into (18):
r 2
k2
2 2

2m r = 1
1 2m
1

r
r
Let u = 1r . So
r

dr
d
d 1
=
( )
d u
1 du d
= 2 ( )( )
u d d
1 du
= 2 ( )hu2
u d

12

(25)

 r = h(

du
)
d

(26)

Pretty soon, well be at a tidy differential equation for Mercurys motion.

Substitute (23) and (26) and u = 1r into (25)
2 du

h d
k2

h2 u 2 = 1
1 2mu 1 2mu
Multiply through by

12mu
h2

k2
du
1 2mu
( )2 u2 (1 2mu) =
2
h
d
h2
And, rearranging this expression will give a first order differential equation
for Mercurys motion:
(

du 2
k 2 1 2m
) + u2 =
+ 2 u + 2mu3
d
h2
h

(27)

By differentiating (27) and dividing through by 2, we get a second order

differential equation for Mercurys motion:
d2 u
m
+ u = 2 + 3mu2
d2
h

(28)

This term very closely resembles the differential encountered in Newtonian

motion, except there is an additional 3mu2 term in (28). This term, however,
is very small. In the case of Mercury, for example, the ratio of the terms on the
right-hand side of (28) is on the order of 107 . We can, then, solve (28) using a
2
perturbation method. Lets introduce a parameter  = 3m
h2 . Also, lets express
differentiation with respect to with a prime and rewrite (28):
u00 + u =

m
h2 u 2
+
(
)
h2
m

(29)

All that remains is solving this differential equation. Lets assume the solution has a form u = u + u1 + O(2 ). Now, we will differentiate this solution
twice, substitute it into (29), and rearrange:
u00 + u

m
h2 u2
00
+
(u
+
u

) + O(2 ) = 0
1
1
h2
m

(30)

For the first approximation of a solution, lets equate the coefficients of ,

2 , . . . to zero. Then u = hm2 (1 + ecos) is the zeroth order solution to
(30). (Actually, the solution is u = hm2 (1 + ecos( )), but we can set
to zero for simplicity) This solution can be verified by differentiating twice and
substituting into (30). Remembering that u = 1r , this solution agrees with the
Newtonian solution.
13

Now, lets examine the coefficient of  in (30):

u001 + u1

=
=
=

h2 u2
m
m
(1 + ecos)2
h2
m
(1 + 2ecos + e2 cos2 )
h2

m
(1 +
h2
Lets try a general solution u1
coefficients.
u001 + u1 =

u001

1 2
2me
me2
e ) + 2 cos +
cos2
(31)
2
h
2h2
= A + Bsin + Ccos2 and solve for its

u01 = Bsin + Bcos 2Csin2

= 2Bcos2 Bsin 4Ccos2

So,
u001 + u1 = (A) + (2B)cos + (3C)cos2

(32)

Comparing (32) to (31), we arrive at the following values for A, B, and C:

A=

1
m
(1 + e2 )
h2
2
me
B= 2
h
me2
C= 2
6h

Hence,
m
1 2
me
me2
(1
+
e
)
+
sin

cos2
h2
2
h2
6h2
And, at last, the general solution to first order is
u1 =

u u + u1
m
1 1
[1 + esin + e2 ( cos2)]
(33)
h2
2 6
Examining the correction term in (33), we see that the esin term increases
after each revolution, and hence becomes dominant. Lets substitute our solution for u , neglect the other terms in the correction, and obtain a simplified
version of (33):
m
u 2 (1 + ecos + esin)
h
u u +

14

m
[1 + ecos[(1 )]]
(34)
h2
It is straightforward to check that (34) satisfies (30) to first order by differentiating and substituting. Examining (34) we see that Mercurys orbit is no
longer an ellipse. It is still periodic, but the period is now given by the following
equation:
2
period =
2(1 + )
(35)
1
And Mercurys perihelion precession per orbit, in relativistic units, is given
by subtracting 2 from its period
u

precession 2 =

6m2r
h2r

(36)

where  is dimensionless, mr is the mass of the sun in relativistic units, and hr

is Mercurys angular momentum in relativistic units. converting mr and hr into
non-relativistic units gives a more useful form of the perihelion shift equation:
precession

6G2 m2
c2 h2

(37)

where G is the Gravitational constant, m is the suns mass, c is the speed of

light, and h is Mercurys angular momentum per unit mass.
Lastly, lets use Keplers second and third laws to rewrite an approximation
of Mercurys perihelion precession in a more convenient form.
From Keplers second law,
dA
L
=
dt
2
where A is the area swept out by the orbit, L is angular momentum, and is
reduced mass. Integrating this equation over one elliptical orbit,
ab

L
T
2
hT
2

For an ellipse, b2 = a2 (1 e2 ), so
T2 =

4 2 (1 e2 )a4
h2

(38)

From Keplers third law,

T2 =

4 2 a3
G(m + mm )

where mm is Mercurys mass. Since mm is very small compared to m, it can be

neglected:
4 2 a3
T2
(39)
Gm
15

From (39), we can solve for G2 m2

G 2 m2 =

16 4 a6 1
( 2)
T2
T

(40)

And now, substituting (38) into (40) gives

G2 m2 =

4 2 a2 h2
T 2 (1 e2 )

(41)

Finally, combining (38) and (37) gives an equation for Mercurys relativistic
perihelion precession per orbit:
precession =

24 3 a2
cT 2 (1 e2 )

(42)

where a is the semimajor axis of Mercurys orbit, c is the speed of light, T is

the period of Mercurys orbit, and e is the eccentricity of Mercurys orbit.
Using values from NASAs Solar System Bodies: Mercury website, this gives
a perihelion shift of 42.9 seconds of arc per century. That Einsteins theory
of General Relativity so elegantly explained Mercurys perihelion shift served
as an early verification of the theory. Shortly after, during the eclipse of 21
September 1922, light deflection was observed, verifying another prediction of
Einsteins theory.
A more striking example of perihelion shift than Mercurys has been observed. In 1975, binary pulsar PSR 1913+16 was discovered. The two massive
sources, in very tight orbit, display a perihelion advance of 4.23 degrees per
year!
Conclusions
The perihelion shift was initially observed in only Mercurys orbit since it
is nearest to the sun, and hence experiences the greatest perihelion drift. Since
Mercurys orbit is farily eccentric, this perihelion shift is easily detectable (Freundlich, 40). If it were more circular, on the other hand, it would have been
much more difficult to observe.
More recently, significant perihelion drift has been observed in other planets.
Einsteins calculations predict a shift for Venus of 8.6 seconds of arc per century,
and 8.4 seconds has been observed. Similarly, relativity predicts a shift for Earth
of 3.8 seconds of arc per century, while 5.0 seconds has been osberved (DInverno,
198).
Einsteins paper elegantly and convincingly explained the observed drift in
perihelion without the need for ghost planets or asteroid belts. From Einsteins
original 1916 paper, entitled The Foundations of the General Theory of Relativity: Calculation gives for the planet Mercury a rotation of the orbit of 43 per
century, corresponding exactly to astronomical observation (Le Verriere); for
the astronomers have discovered in the motion of the perihelion of this planet,
after allowing for all disturbances by other planets, an inexplicable remainder
of this magnitude. (Lorentz et al., 164)
16

References
Baum, Richard and Sheehan, William. In Search of Planet Vulcan: The
Ghost in Newtons Clockwork Universe. Plenum Trade, New York. 1997.
Callahan, James J. The Geometry of Spacetime: An Introduction to Special
and General Relativity. Springer, New York. 1991.
Davis, Phil (Web curator). Solar System Bodies: Mercury. http://solarsystem.nasa.gov/
features/planets/mercury/mercury.html. Updated 18 March 2003.
DInverno, Ray. Introducing Einsteins Relativity. Oxford University Press,
Oxford. 1993.
Freundlich, Erwin. The Foundations of Einsteins Theory of Gravitation.
Translated from German by Henry L. Brose. Cambridge University Press, Cambridge. 1920.
Price, Michael P., and Rush, William F. Non relativistic contribution to
Mercurys perihelion. in American Journal of Physics 47(6). 531-534. June
1979.
Lorents, H. A., Einstein, A., Minkowski, H., and Weyl, H. The principle
of relativity: a collection of original memoirs on the special and general theory
of relativity. contained The foundation of general relativity, by A. Einstein.
Dover, New York. 1952.
Roseveare, N. T. Mercurys Perihelion from Leverriere to Einstein. Caledon
Press, Oxford. 1982.
Stephani, Hans. General Relativity: An introduction to the theory of the
gravitational field. Cambridge University Press, Cambridge. 1996.
Stewart, John. Calculus: Fourth Edition. Brooks/Cole, Pacific Grove. 1999.
Torretti, Roberto. Relativity and Geometry. Permagon Press, Oxford. 1983.

17