Perihelion Precession of Mercury

Equivalent forms

\Delta\phi = \frac{24\pi^3 a^2}{T^2 c^2 (1-e^2)}

\dot{\varpi}_{\text{GR}} = 43''\,\text{per century (Mercury)}

A pure prediction with zero adjustable knobs reproduced a 56-year-old anomaly to the arcsecond — theory beating epicycle-style fixes.

Symbol	Name	SI unit	Dimension	Range
$Δφ$	Precession per orbitoutput Forward shift of perihelion each revolution	rad	1	0 – 0.000001
$M$	Central mass Mass of the Sun	kg	M	1e+30 – 3e+30
$a$	Semi-major axis Orbit semi-major axis	m	L	50000000000 – 60000000000
$e$	Eccentricity Orbital eccentricity	dimensionless	1	0 – 0.9

Unit systems

Symbol	SI	CGS	Imperial
$M$	kg	g	lb
$a$	m	cm	ft

Where it holds

Weak-field, slow-motion (post-Newtonian) limit of general relativity,

GM/(ac^{2})

≪ 1. Near a black hole the full Schwarzschild geodesic is required and the advance can reach order unity per orbit.

Dimensional analysis

[1] (radians)

[M^{-1}L^{3}T^{-2}][M] / ([L][L^{2}T^{-2}]) = [L^{3}T^{-2}]/[L^{3}T^{-2}] = [1]

GM/(ac^{2})

is dimensionless,

so \Delta \varphi is

a pure angle.

\checkmark

Discovery

Urbain Le Verrier (anomaly); Albert Einstein (explanation) · 1915

Le Verrier reported Mercury's anomalous precession in 1859 and even proposed a planet 'Vulcan' to explain it. None existed. In November 1915 Einstein computed the general-relativistic correction and got 43"/century with no free parameters — the first triumph of GR. He wrote that the result gave him palpitations.

Try this

Mercury's orbit doesn't quite close — its closest point creeps forward by 43 arcseconds a century that Newton simply cannot explain.

Newtonian gravity plus the other planets accounts for most of the drift, but leaves 43"/century unexplained. General relativity's correction to the orbit supplies exactly that, turning the ellipse into a slow rosette.

Research status: stable

Real-world applications

First and still-classic test of general relativity
Periastron-advance tests in binary-pulsar timing
High-precision Solar-System ephemerides and spacecraft navigation
Probing strong gravity around the galactic-center black hole (star S2)

Common misconceptions

Newton predicts zero precession for Mercury — he predicts $\sim 531$ "/century from other planets; only the extra 43" is anomalous
The precession is huge — it is 43 arcseconds per CENTURY, about $0.012^{\circ}$
A hidden planet (Vulcan) causes it — searches found none; spacetime curvature does it

Experimental verification

Mercury's residual 43"/century, unexplained by Newtonian planetary perturbations

(\sim 531

"/century), matches GR exactly; radar ranging and the MESSENGER mission refined it

to \sim 0.1

% agreement; binary pulsars like PSR B1913+16 show the same periastron advance at thousands of times the rate.

Derivation

Solve the Schwarzschild geodesic for

u = 1/r

as a function of angle

\varphi

:

d^{2}u/d\varphi ^{2} + u = GM/h^{2} + 3GM u^{2}/c^{2}

.

The first term gives the Newtonian ellipse; the small

3GMu^{2}/c^{2}

term shifts the angular period of the radial oscillation.

Perturbation theory yields a perihelion advance per orbit

of \Delta \varphi = 6\pi GM/[

a

(1-e^{2})c^{2}]

.

Plugging in Mercury's a, e, and the Sun's mass gives

5.0\times 10^{-7} rad

/orbit

\times (415

orbits/century

) \approx 43

arcseconds per century.

Limiting cases

a \to

large (outer planets)⟶

\Delta \varphi \to 0

Far from the Sun spacetime is nearly flat; Earth's GR advance is only

\sim 3.8

"/century.

c \to \infty

⟶

\Delta \varphi \to 0

In the Newtonian limit the orbit is a closed ellipse with no precession.

e \to 1

(near-radial)⟶

\Delta \varphi

enhanced by

1/(1-e^{2})

Highly eccentric orbits dive deep into the Sun's potential, amplifying the relativistic kick.

What if…

What if the Sun were more compact (same mass)?

No change — the exterior Schwarzschild field depends only on M, so the advance is identical.

What if we orbited a black hole instead?

Deep in strong gravity $GM/(ac^{2})$ approaches 1 and the orbit can precess by large fractions of a turn each lap.

What if light, not a planet, passed the Sun?

The same metric bends starlight by 1.75" — Eddington's 1919 eclipse test, GR's second classic confirmation.

1

Mercury per orbit

Given ·

M:: 1.989e+30
a:: 57900000000
e:: 0.2056

Find ·

\Delta \varphi (rad

/orbit)

Steps

Numerator $6\pi GM = 2.50e21$
Denominator a $(1-e^{2})c^{2} = 5.79e10\times 0.9577\times 8.99e16 \approx 4.99e27$
$\Delta \varphi \approx 5.0e-7\,\mathrm{rad}$ /orbit

Result ·

\Delta \varphi = 6\pi \times 6.674e-11\times 1.989e30 / (5.79e10\times 0.9577\times (3e_{8})^{2}) \approx 5.0\times 10^{-7} rad

2

Scale up to a century

Given ·

Δφ:: 5e-7

Find · arcsec/century

Steps

Mercury makes $\sim 415$ orbits per century
Total $= 415\times 5.0e-7 = 2.08e-4\,\mathrm{rad}$
$\times 206265 \approx 43$ arcsec/century

Result · 415 orbits/century

\times 5.0e-7\,\mathrm{rad} \times (206265

"/

rad) \approx 43

"/century

Kepler's Third Law→Newton's Law of Universal Gravitation→Tidal Force→Escape Velocity→Gyroscopic Precession→