Perihelion Precession in Schwarzschild Geometry

A planet on a Newtonian orbit traces the same ellipse forever: its perihelion, the point of closest approach, is nailed to a fixed direction in space. This is not a generic fact about central forces — it is a fragile coincidence of the inverse-square law, protected by Bertrand's theorem, and the moment gravity departs from a pure $1/r$ potential the ellipse begins to turn. General relativity supplies exactly such a departure. In the Schwarzschild geometry a massive test particle still moves on a plane, still conserves an energy $E$ and an angular momentum $L$ from the metric's two Killing vectors, and its radial motion still reads as a one-dimensional problem in an effective potential. But that potential carries one extra term, attractive and going as $1/r^3$ , absent from Newton. That single term unlocks the perihelion: the radial oscillation from perihelion out to aphelion and back now takes slightly more than one full turn of azimuth, so each orbit the perihelion creeps forward. This article is the flagship of the Schwarzschild-orbits head. It takes the metric, the conserved quantities, the geodesic equations, and the massive-particle effective potential — each owned by a preceding article — and assembles from them the most famous number in the early history of the theory: the $42.98''$ per century by which Mercury's perihelion advances, unexplained by Newtonian celestial mechanics and reproduced by Einstein in November 1915 with no free parameters. The same effective potential, pushed harder, governs the ISCO and the photon sphere; perihelion precession is its gentlest, weak-field face, and the cleanest place to see relativistic orbital dynamics at work.

Key takeaways

Bound timelike orbits in the Schwarzschild geometry are not closed ellipses: the perihelion advances by a small angle every revolution, tracing a slowly turning rosette.
The advance comes from one extra term in the relativistic effective potential, an attractive $-GML^2/(c^2r^3)$ absent from Newton, which makes the radial period slightly longer in azimuth than $2\pi$ .
To leading post-Newtonian order the advance per orbit is $\Delta\varpi = 6\pi GM/[c^2 a(1-e^2)]$ , with $a$ the semi-major axis and $e$ the eccentricity.
For Mercury this is $42.98''$ per century — matching the observed residual after Newtonian planetary perturbations are removed, one of the three classical tests of general relativity.
The clean route is the relativistic Binet equation $u''+u = GM/L^2 + (3GM/c^2)u^2$ for $u=1/r$ ; its resonant $u^2$ term drives a secular perihelion shift.
The same physics, scaled up, gives $4.23^\circ$ per year for the Hulse–Taylor binary pulsar and a measured $\sim12'$ per orbit for the star S2 around Sgr A*.
Key idea: The relativistic effective potential adds an attractive $1/r^3$ term to the Newtonian $-GM/r + L^2/2r^2$ ; this makes the radial oscillation period slightly longer than the angular one, so the perihelion advances by $6\pi GM/[c^2 a(1-e^2)]$ per orbit.

Where this lives today

Perihelion precession became a general-purpose probe of gravity. In the Solar System the measured advance constrains the post-Newtonian combination $(2 + 2\gamma - \beta)/3$ , with planetary radar and the BepiColombo mission tightening the bounds. In binary pulsars the periastron advance is a post-Keplerian parameter that, combined with the Shapiro delay and orbital decay, over-determines the two stellar masses — the double pulsar PSR J0737−3039 does this to better than a percent. At the galactic centre the GRAVITY interferometer tracked the star S2 through its $16$ -year orbit around the $4\times10^6\,M_\odot$ black hole Sgr A* and detected its Schwarzschild precession of about $12'$ per orbit, confirming the same $6\pi GM/[c^2a(1-e^2)]$ scaling around a supermassive black hole. And for nearly circular orbits the precession reappears as the split between the orbital frequency $\Omega$ and the radial epicyclic frequency $\kappa$ , $\kappa^2 = \Omega^2(1 - 6GM/c^2r)$ , invoked to model the quasi-periodic oscillations seen in the X-ray light of accreting compact objects.

The physical picture

Picture the radial coordinate as a bead rolling in a one-dimensional well. One radial cycle — perihelion out to aphelion and back — takes a definite lapse of proper time, during which the planet also sweeps some azimuth $\phi$ . In Newtonian $1/r$ gravity these two clocks are locked in an exact $1:1$ resonance: the bead returns to perihelion after precisely $2\pi$ of azimuth, so the orbit closes into a fixed ellipse. The Schwarzschild $1/r^3$ term deepens the well on its inner flank, stiffening the radial spring just enough that the radial cycle now needs a touch more than $2\pi$ of azimuth. That excess, accumulated turn after turn, is the precession — the perihelion sliding forward in the same sense as the orbit. The first figure shows the resulting rosette.

Figure

A bound Schwarzschild orbit does not close: each revolution the perihelion advances by Δϖ = 6πGM/(c²a(1−e²)), tracing a slowly turning rosette. The precession here is exaggerated to 58.6° per orbit for clarity; for Mercury the real advance is only 5.0×10⁻⁷ rad per orbit, which accumulates to 42.98″ per century

The extra term is relativistic: read it as the gravitational field energy itself gravitating, or as the centrifugal barrier $L^2/2r^2$ undercut at small $r$ by the coupling to curvature. Either way the effect scales as the compactness $GM/(c^2r)$ times an orbital factor — about $2.5\times10^{-8}$ at Mercury's orbit, giving a per-orbit shift of a few tenths of a microradian: invisible in one pass, relentless over the $415$ orbits of a century. The engine lives in the shape of the effective potential, which the second figure lays bare against its Newtonian counterpart.

Figure

The Schwarzschild effective potential V_eff = c²/2 − GM/r + L²/2r² − GML²/(c²r³) against the Newtonian one, in units GM = c = 1 with L² = 20. Far out the two agree, but the extra −1/r³ term turns the Newtonian centrifugal wall (green, rising to +∞) into a finite barrier (blue) that plunges to −∞. A bound orbit oscillates in radius between perihelion and aphelion about the stable minimum, and the azimuth swept in one such radial cycle exceeds 2π by the precession angle

The core equations

Everything begins with the geometry. The Schwarzschild line element is the unique static, spherically symmetric vacuum solution of Einstein's equations; its coefficients are what the geodesics feel.

\mathrm{d}s^2 = -\left(1-\frac{2GM}{c^2 r}\right)c^2\,\mathrm{d}t^2 + \left(1-\frac{2GM}{c^2 r}\right)^{-1}\mathrm{d}r^2 + r^2\,\mathrm{d}\Omega^2 \qquad (1)

where: $M$ — mass of the central body; $r,\,t$ — Schwarzschild radial and time coordinates; $\mathrm{d}\Omega^2 = \mathrm{d}\theta^2+\sin^2\theta\,\mathrm{d}\phi^2$ — the unit-sphere metric. Reading: the two factors $1-2GM/c^2r$ carry all the gravitation; $c\to\infty$ flattens them to unity and returns flat spacetime. Because the metric depends on neither $t$ nor $\phi$ , the timelike and rotational Killing vectors hand us two constants of the motion — the specific energy and specific angular momentum recalled by name from the Killing-vector article.

E = \left(1-\frac{2GM}{c^2 r}\right)c^2\frac{\mathrm{d}t}{\mathrm{d}\tau}, \qquad L = r^2\frac{\mathrm{d}\phi}{\mathrm{d}\tau} \qquad (2)

where: $E$ — specific energy (per unit rest mass, carrying a factor $c^2$ ); $L$ — specific angular momentum; $\tau$ — proper time. Reading: with $E$ and $L$ fixed by the initial conditions, the four-velocity has only $\mathrm{d}r/\mathrm{d}\tau$ left as a genuine unknown. Feeding the two constants into the four-velocity normalization collapses the radial motion to a first integral with the structure of one-dimensional energy conservation.

\frac{1}{2}\left(\frac{\mathrm{d}r}{\mathrm{d}\tau}\right)^2 + V_{\rm eff}(r) = \frac{1}{2}\frac{E^2}{c^2} \qquad (3)

where: $V_{\rm eff}$ — the effective potential of the next card; $E^2/c^2$ — the conserved parameter playing the role of total energy. Reading: turning points are where $\mathrm{d}r/\mathrm{d}\tau=0$ , i.e. $V_{\rm eff}(r)=E^2/2c^2$ — the perihelion and aphelion radii. The effective potential is where relativity leaves its fingerprint: the Newtonian gravity-plus-centrifugal-barrier, plus one uninvited term.

V_{\rm eff}(r) = \frac{c^2}{2} - \frac{GM}{r} + \frac{L^2}{2r^2} - \frac{GM L^2}{c^2 r^3} \qquad (4)

where: $-GM/r$ — Newtonian gravity; $L^2/2r^2$ — centrifugal barrier (also Newtonian); $-GML^2/(c^2 r^3)$ — the relativistic correction. Reading: the last term is attractive and steepens as $1/r^3$ , so it dominates at small $r$ , converting the Newtonian centrifugal wall into a finite barrier — the blue-versus-green contrast in the second figure. Changing the independent variable from proper time to azimuth, with $u=1/r$ , turns the radial equation into the relativistic Binet equation — a driven harmonic oscillator in $u(\phi)$ .

\frac{\mathrm{d}^2 u}{\mathrm{d}\phi^2} + u = \frac{GM}{L^2} + \frac{3GM}{c^2}u^2 \qquad (5)

where: $u=1/r$ — inverse radius; $3GM u^2/c^2$ — the relativistic term, absent in Newton where the equation is linear. Reading: without the $u^2$ term the solution is a fixed conic; with it, the closed conic is lost. Solving (5) perturbatively gives a conic whose argument advances slightly slower than $\phi$ — a conic that precesses.

u(\phi) = u_0\left[1 + e\cos(k\phi)\right], \qquad k = \sqrt{1 - \frac{6GM\,u_0}{c^2}} \approx 1 - \frac{3GM}{c^2}u_0 \qquad (6)

where: $u_0 \approx GM/L^2$ — inverse semi-latus rectum; $e$ — eccentricity; $k<1$ — ratio of radial to angular frequency. Reading: perihelion recurs when $k\phi = 2\pi$ , i.e. at $\phi = 2\pi/k > 2\pi$ — the planet oversteps a full turn, so the perihelion has moved forward. Reading off the overshoot gives the headline result, the advance per orbit, converted to the arcsecond-per-century rate astronomers report by dividing by the period.

\Delta\varpi = 2\pi\left(\frac{1}{k}-1\right) \approx \frac{6\pi GM}{c^2}\,u_0 = \frac{6\pi GM}{c^2\,a(1-e^2)}, \qquad \dot\varpi = \frac{\Delta\varpi}{T} \qquad (7)

where: $\Delta\varpi$ — advance per revolution (radians); $a(1-e^2)$ — semi-latus rectum $p=1/u_0$ ; $T=2\pi\sqrt{a^3/GM}$ — Kepler period; $\dot\varpi$ — the observable rate. Reading: the advance grows with mass, shrinks with orbit size, and blows up as $e\to1$ where the perihelion grazes the strong field.

Derivation AThe relativistic effective potential from the geodesic norm

The goal is to show that a massive test particle's radial motion in Schwarzschild spacetime reduces exactly to one-dimensional energy conservation, and to expose the single extra term that distinguishes the relativistic effective potential from the Newtonian one. We start from the constraint that the four-velocity has constant norm $-c^2$ , insert the two conserved quantities to eliminate the time and azimuthal velocities, and rearrange into the form of equation (3), whereupon the $1/r^3$ coefficient appears on its own and cannot be transformed away.

Step 1. Restrict to the equatorial plane $\theta=\pi/2$ (any orbit can be rotated into a plane) and write the norm $g_{\mu\nu}u^\mu u^\nu=-c^2$ using the metric of equation (1), keeping the terms in $\dot t$ , $\dot r$ , and $\dot\phi$ with $\dot{}=\mathrm{d}/\mathrm{d}\tau$ . Why: the norm is the one relation tying the four-velocity components together — the natural place to impose "massive particle".

Step 2. Substitute the conserved quantities of equation (2), $\dot t = (E/c^2)(1-2GM/c^2r)^{-1}$ and $\dot\phi = L/r^2$ , to remove $\dot t$ and $\dot\phi$ , leaving $-\,(E^2/c^2)(1-2GM/c^2r)^{-1} + (1-2GM/c^2r)^{-1}\dot r^2 + L^2/r^2 = -c^2$ . Why: the Killing constants reduce a three-component problem to a single radial equation in $\dot r$ .

Step 3. Multiply through by $(1-2GM/c^2r)$ and isolate $\dot r^2$ to get $\dot r^2 = E^2/c^2 - (1-2GM/c^2r)(c^2 + L^2/r^2)$ . Why: clearing the inverse metric factor readies the equation for comparison with a one-dimensional energy law.

Step 4. Expand the product on the right: it yields $c^2 - 2GM/r + L^2/r^2 - 2GML^2/(c^2r^3)$ . Divide by two and move the constant to define $V_{\rm eff}(r) = c^2/2 - GM/r + L^2/2r^2 - GML^2/(c^2r^3)$ , so that $\tfrac12\dot r^2 + V_{\rm eff} = E^2/2c^2$ . Why: the cross term between the centrifugal $L^2/r^2$ and the metric factor $2GM/c^2r$ is precisely the $1/r^3$ piece — a product of gravity and angular momentum with no Newtonian counterpart, and it is what precesses the orbit.

Go deeper · Derivation B · 2 minFrom the orbit equation to the perihelion advance

Derivation BFrom the orbit equation to the perihelion advance

The goal is to convert the radial energy equation into the Binet equation for $u(\phi)=1/r$ , solve it by perturbation theory treating the relativistic term as small, and extract the closed-form advance $\Delta\varpi = 6\pi GM/[c^2 a(1-e^2)]$ . The strategy: change the independent variable from proper time to azimuth using $L=r^2\dot\phi$ , differentiate to remove the square root, then feed the Newtonian ellipse into the small nonlinear term and watch the resonant piece build a secular drift that resums into a precessing conic.

Step 1. Use $\dot r = (\mathrm{d}r/\mathrm{d}\phi)\dot\phi = -L\,\mathrm{d}u/\mathrm{d}\phi$ , since $u=1/r$ gives $\mathrm{d}r/\mathrm{d}\phi = -r^2\,\mathrm{d}u/\mathrm{d}\phi$ . Substitute into the radial equation to get $L^2(\mathrm{d}u/\mathrm{d}\phi)^2 = E^2/c^2 - c^2 + 2GM u - L^2 u^2 + (2GML^2/c^2)u^3$ . Why: trading $\tau$ for $\phi$ turns "when the planet reaches a radius" into "what shape the orbit is" — and the shape is what precesses.

Step 2. Differentiate both sides with respect to $\phi$ and divide out the common factor $2L^2\,\mathrm{d}u/\mathrm{d}\phi$ to reach the Binet equation $\mathrm{d}^2u/\mathrm{d}\phi^2 + u = GM/L^2 + (3GM/c^2)u^2$ . Why: differentiating removes $E$ and the square root, leaving a clean oscillator whose only non-Newtonian feature is the quadratic driver.

Step 3. Solve at zeroth order by dropping the $u^2$ term: $u_0(\phi) = (GM/L^2)(1 + e\cos\phi)$ , a closed ellipse with apsidal angle exactly $\pi$ . Insert this into the small term to get a forcing $(3GM/c^2)u_0^2 = (3G^3M^3/c^2L^4)(1+e\cos\phi)^2$ . Why: perturbation feeds the Newtonian orbit into the correction; the constant and $\cos^2\phi$ pieces only shift and rescale, but the $2e\cos\phi$ term is dangerous.

Step 4. Keep only the resonant $\cos\phi$ forcing, $\mathrm{d}^2u_1/\mathrm{d}\phi^2 + u_1 = (6G^3M^3e/c^2L^4)\cos\phi$ , whose particular solution grows without bound as $u_1 = (3G^3M^3e/c^2L^4)\,\phi\sin\phi$ . Why: forcing an oscillator at its own natural frequency produces a secularly growing $\phi\sin\phi$ response — the mathematical signature of a slow, cumulative drift rather than a one-time shift.

Step 5. Recognize $u \approx (GM/L^2)[1 + e\cos\phi + \delta\,e\,\phi\sin\phi]$ as the first two terms of $(GM/L^2)[1 + e\cos((1-\delta)\phi)]$ with $\delta = 3G^2M^2/c^2L^2$ , so the apsidal argument is $(1-\delta)\phi$ , not $\phi$ . Perihelion recurs when $(1-\delta)\phi = 2\pi$ , i.e. at $\phi = 2\pi(1+\delta+\cdots)$ . Why: resumming the runaway term into a frequency shift is legitimate because $\delta\ll1$ , converting a divergent-looking correction into a finite precession.

Step 6. Read off the advance $\Delta\varpi = 2\pi\delta = 6\pi G^2M^2/c^2L^2$ , and use the Newtonian identity $L^2 = GM\,a(1-e^2)$ to write it as $\Delta\varpi = 6\pi GM/[c^2 a(1-e^2)]$ . Why: expressing the answer through the observable orbital elements $a$ and $e$ makes it directly testable against a planet's catalogued orbit.

Limiting cases

Case	Result	Meaning
Weak field, bound orbit	$\Delta\varpi \approx \dfrac{6\pi GM}{c^2 a(1-e^2)}$	Leading post-Newtonian result, valid when $GM/(c^2 r)\ll1$ ; the Mercury and binary-pulsar regime.
Nearly circular orbit	$\Delta\varpi \approx 6\pi\,\dfrac{GM}{c^2 r}$	Set $a\to r$ , $e\to0$ : the radial epicyclic frequency $\kappa$ falls just below the orbital frequency $\Omega$ , with $\kappa^2 = \Omega^2(1 - 6GM/c^2r)$ .
Strong field, near ISCO	$\kappa^2 = \Omega^2\left(1 - \dfrac{6GM}{c^2 r}\right)\to0$	At $r = 6GM/c^2$ the epicyclic frequency vanishes; the perturbative formula breaks down and the advance per orbit diverges — this radius is the ISCO.
Newtonian limit	$\Delta\varpi \to 0 \quad (c\to\infty)$	The $1/r^3$ term vanishes, orbits close into fixed Keplerian ellipses, and Bertrand's theorem is recovered.

Worked examples

1 · Mercury's forty-three arcseconds (fully worked)

Compute the general-relativistic perihelion advance of Mercury in arcseconds per century, using $GM_\odot/c^2 = 1.4766$ km, $a = 5.7909\times10^{10}$ m, $e = 0.2056$ , and orbital period $T = 87.969$ days.

Step 1. Semi-latus rectum. The formula needs $p = a(1-e^2)$ . With $e^2 = 0.04227$ , $1-e^2 = 0.95773$ , so $p = 5.7909\times10^{10}\times0.95773 = \mathbf{5.5461\times10^{10}\ m}$ . This is the length scale that sets the advance — the smaller it is, the deeper into the field the perihelion dips.

Step 2. Advance per orbit. Substitute into equation (7): $\Delta\varpi = 6\pi\,(GM_\odot/c^2)/p = 6\pi\,(1.4766\times10^3\ \mathrm{m})/(5.5461\times10^{10}\ \mathrm{m}) = \mathbf{5.0185\times10^{-7}\ rad}$ . Barely half a microradian — the angle Mercury's perihelion gains each Mercurian year.

Step 3. Orbits per century. A Julian century is $100\times365.25 = 36525$ days, so the number of orbits is $N = 36525/87.969 = \mathbf{415.20}$ . The tiny per-orbit shift only becomes measurable because it repeats this many times, all in the same prograde sense.

Step 4. Accumulate and convert. The total angle is $N\,\Delta\varpi = 415.20\times5.0185\times10^{-7} = 2.0837\times10^{-4}$ rad. Converting with $1\ \mathrm{rad} = 206265''$ gives $2.0837\times10^{-4}\times206265'' = \mathbf{42.98''\ per\ century}$ .

Step 5. Verdict. This matches the observed relativistic residual (after equinox precession and planetary perturbations are removed); modern radar and spacecraft ranging confirm it to better than $0.1\%$ . Mercury precesses at 42.98″ per century from general relativity alone.

2 · The Hulse–Taylor binary pulsar (fully worked)

The binary pulsar PSR B1913+16 has orbital period $P_b = 7.752$ h, eccentricity $e = 0.617$ , and total mass $M = 2.828\,M_\odot$ . Estimate its periastron advance in degrees per year, treating it with the same one-body formula using the total mass and the relative orbit.

Step 1. Relative semi-major axis. From Kepler's third law $a^3 = GM(P_b/2\pi)^2$ . With $GM = 2.828\times1.327\times10^{20} = 3.75\times10^{20}\ \mathrm{m^3\,s^{-2}}$ and $P_b = 2.791\times10^4$ s, this gives $a = \mathbf{1.95\times10^{9}\ m}$ — about $6.5$ light-seconds, a pair of neutron stars in an orbit smaller than the Sun.

Step 2. Advance per orbit. Here $GM/c^2 = 4.18\times10^3$ m and $p = a(1-e^2) = 1.95\times10^9\times0.619 = 1.21\times10^9$ m, so $\Delta\varpi = 6\pi\,(4.18\times10^3)/(1.21\times10^9) = 6.5\times10^{-5}$ rad $= \mathbf{0.00374^\circ}$ per orbit — over a hundred times Mercury's per-orbit shift.

Step 3. Rate per year. The system completes $N = (365.25\times86400)/2.791\times10^4 = 1131$ orbits per year, so $\dot\varpi = 1131\times0.00374^\circ = \mathbf{4.23^\circ\ per\ year}$ .

Step 4. Verdict. The measured periastron advance is $4.2266^\circ$ per year — roughly $35{,}000$ times Mercury's rate, reproduced here to three figures. The same $6\pi GM/[c^2a(1-e^2)]$ that gives Mercury 43″/century gives the Hulse–Taylor pulsar 4.23°/year, because the neutron-star pair is far more compact — a pillar of the system's Nobel-winning confirmation of general relativity.

Physics of the portrait

The perihelion of Mercury is where a two-century-old anomaly of celestial mechanics met a brand-new theory of gravity and matched it to the arcsecond, with nothing adjusted.

Origins · 1 minAlbert Einstein and how this was discovered

Albert Einstein

1879–1955 · general relativity and the perihelion of Mercury

By the 1910s the residual advance of Mercury's perihelion had been a puzzle for half a century. Urbain Le Verrier — who had predicted Neptune from Uranus's perturbations — announced in 1859 that after accounting for every known planet, Mercury's perihelion still crept forward by an amount he could not explain, later refined to about $43''$ per century. He and others postulated an unseen intramercurial planet, "Vulcan", searched for and never found. Bertrand's 1873 theorem, meanwhile, worked against any easy fix — only the $1/r$ and harmonic potentials give closed bound orbits, so a fixed perihelion was a privilege any modification would forfeit. Einstein, completing general relativity in November 1915 before Schwarzschild had even found the exact metric, computed the geodesic precession in the field of the Sun and obtained the missing $43''$ — a moment he later said left him with palpitations. It was the theory's first quantitative triumph: not a new prediction awaiting a telescope, but an old, precise anomaly explained at a stroke with no free parameters, taking its place beside light bending and gravitational redshift as one of the three classical tests.

History

1859

— Le Verrier reports Mercury's anomalous perihelion advance and proposes the planet Vulcan

1915

— Einstein computes 43″ per century from general relativity, matching the residual

1916

— Schwarzschild publishes the exact vacuum solution underlying the calculation

1974

— Hulse and Taylor discover the binary pulsar PSR B1913+16 with a huge periastron advance

2020

— the GRAVITY collaboration measures the Schwarzschild precession of the star S2 at Sgr A*

Going further

References. Misner, Thorne & Wheeler, Gravitation (1973), §25.5–25.6 and Box 25.3 · Wald, General Relativity (1984), §6.3 · Carroll, Spacetime and Geometry (2004), §5.5 · Hartle, Gravity (2003), §9.4 · Schutz, A First Course in General Relativity (2009), §11.1–11.2 · Weinberg, Gravitation and Cosmology (1972), §8.6. The Schwarzschild line element (1), the Killing constants (2), the geodesic equations behind (3), and the massive-particle effective potential (4) are combined here to yield the Binet equation $u''+u = GM/L^2 + (3GM/c^2)u^2$ from the four-velocity norm, and the perturbation gives $\Delta\varpi = 6\pi G^2M^2/(c^2L^2) = 6\pi GM/[c^2a(1-e^2)]$ . Mercury's advance follows from $GM_\odot/c^2 = 1.4766$ km, $a = 5.7909\times10^{10}$ m, $e = 0.2056$ , $T = 87.969$ d, yielding $42.98''$ /century; the Hulse–Taylor rate $4.23^\circ$ /yr follows from $P_b = 7.752$ h, $e = 0.617$ , $M = 2.828\,M_\odot$ . The Newtonian identity $L^2 = GM\,a(1-e^2)$ connects the two forms of the formula, and the limit $\Delta\varpi\to0$ as $c\to\infty$ recovers closed Keplerian ellipses.

Prerequisites. The Schwarzschild Metric · Killing Vectors and Conserved Quantities in Schwarzschild · Geodesic Equations in Schwarzschild Spacetime · Effective Potential for Massive Particle Orbits. Leads to. Circular Orbits and Their Stability · Innermost Stable Circular Orbit (ISCO) · Photon Orbits and the Photon Sphere.

Ask the tutor: Ask me to walk through the perturbative Binet solution line by line, or to recompute the advance for any planet, pulsar, or S-star from its orbital elements.

Is the perihelion precession a real physical effect or an artifact of coordinates?

It is entirely physical. The advance $\Delta\varpi$ is the angle between two successive perihelia, and both the perihelion (a turning point, $\mathrm{d}r/\mathrm{d}\tau=0$ ) and the azimuth swept between them are coordinate-independent statements about the geodesic — you get the same number in any coordinate system. What is frame-dependent is the much larger $5025''$ per century from the precession of the equinoxes, which is why that piece must be subtracted before comparing with the $43''$ of relativity.

Where does the factor of 6 in 6πGM/c²a(1−e²) actually come from?

It factors as $2\times3$ . The $3$ is the coefficient of the $u^2$ term in the Binet equation, $3GM/c^2$ , which traces to the cross term between the centrifugal barrier $L^2/r^2$ and the metric factor $2GM/c^2r$ in the effective potential. The remaining $2\pi$ comes from accumulating the small frequency shift $\delta$ over one revolution: the apsidal argument is $(1-\delta)\phi$ , perihelion recurs at $\phi = 2\pi/(1-\delta)\approx 2\pi(1+\delta)$ , so the overshoot is $2\pi\delta$ with $\delta = 3G^2M^2/c^2L^2$ . Substitute $L^2 = GM\,a(1-e^2)$ and the $6\pi$ appears.

If a circular orbit has no perihelion, why does the formula stay finite as e → 0?

A perfectly circular orbit has no distinguished point of closest approach, so "perihelion advance" seems meaningless there. But the formula tends smoothly to $6\pi GM/c^2r$ as $e\to0$ , and that limit has a clear meaning: it is the precession of the slightly eccentric "wobble" of a nearly circular orbit. In frequency language it is the difference between the orbital frequency $\Omega$ and the radial epicyclic frequency $\kappa$ , since $\kappa^2 = \Omega^2(1-6GM/c^2r)$ . So the effect does not disappear at low eccentricity; it becomes the precession of an almost-circle rather than of an obvious ellipse.

Why does the extra term go as 1/r³ specifically, and not some other power?

The $1/r^3$ arises from the product of two Newtonian ingredients with the one relativistic metric factor. When you expand the four-velocity norm, the centrifugal term $L^2/r^2$ gets multiplied by $1-2GM/c^2r$ , producing $L^2/r^2$ (the ordinary barrier) minus $2GML^2/(c^2r^3)$ (the new piece). So the power is fixed: it is one power of $1/r$ steeper than the centrifugal barrier because it is that barrier weighted by $GM/c^2r$ . Such a $1/r^3$ correction is exactly the kind of deviation from inverse-square that Bertrand's theorem says must open the orbit.

How can the same effective potential describe precession, the ISCO, and photon orbits?

Because they are all features of one curve, read at different energies and angular momenta. Perihelion precession is the weak-field, bound-orbit regime: the orbit oscillates in the shallow outer well and its slight anharmonicity makes it precess. Push the angular momentum down and the outer minimum and inner maximum approach each other; when they merge, at $r = 6GM/c^2$ , the stable circular orbit disappears — the ISCO. For massless particles the potential has the same structure with its maximum at $r = 3GM/c^2$ , the photon sphere. Precession is simply the gentlest thing this potential does.

Did Einstein really need the full Schwarzschild solution to get the 43 arcseconds?

No — and that is a nice historical point. Einstein derived the Mercury result in November 1915, before Schwarzschild found the exact metric in early 1916. He worked to the necessary order in a weak-field expansion of his field equations, which is all the $6\pi GM/[c^2a(1-e^2)]$ result requires — it is a leading post-Newtonian effect, first order in $GM/c^2r$ . Schwarzschild's exact solution later confirmed the calculation and is the cleaner modern route, but the historic prediction preceded it. This is why the perihelion advance is called the first test of general relativity even though the "Schwarzschild geometry" was not yet in hand.