Angular Momentum Conservation in Central Fields

When Halley's comet last swung past the Sun in 1986, it was moving at 54.6 km/s — fast enough to cross the continental United States in under a minute and a half. Out at aphelion, beyond Neptune's orbit, the same comet crawls along at 0.91 km/s, slower than a rifle bullet. Nobody steers it. No engine fires. Yet every return, for millennia, the speeds at those two points have stood in exactly the ratio of the distances, inverted: sixty times closer, sixty times faster. Johannes Kepler read this rhythm out of Tycho Brahe's Mars data in 1609 — equal areas swept in equal times — almost eighty years before anyone could say why the Solar System keeps such perfect books. The why is one line of vector algebra. A central force points along the line to the center, so it has no lever arm about that center: the torque $\vec{r}\times\vec{F}$ vanishes identically, and the angular momentum $\vec{L}=\vec{r}\times\vec{p}$ freezes — magnitude and direction alike. You already know what a central force is and how the two-body problem collapses to one body orbiting a fixed center; this article delivers the payoff of that setup. Mastering this one conservation law buys you three things at once: the orbit is trapped in a fixed plane, the sweep rate of the radius vector is a constant you can compute from any single snapshot of the motion, and the angular variable can be eliminated entirely — the gateway through which the whole of orbit theory, from the effective potential to the Kepler problem, passes.

Key takeaways

A central force points along the line to the center, so it exerts zero torque about that point: $\vec{N}=\vec{r}\times\vec{F}=\vec{r}\times f(r)\hat{r}=0$ .
Zero torque means the angular momentum $\vec{L}=\vec{r}\times\vec{p}$ is a constant vector — fixed in both magnitude and direction.
Because $\vec{r}$ must stay perpendicular to the fixed $\vec{L}$ , the entire orbit is confined to a single plane through the force center: a 3D problem collapses to 2D.
In that plane the magnitude $L=mr^2\dot\theta$ is constant, which is precisely Kepler's second law: the radius vector sweeps equal areas in equal times, $\mathrm{d}A/\mathrm{d}t = L/2m$ .
The conservation holds for any radial profile $f(r)$ — inverse-square, harmonic, anything — but breaks the moment the force acquires a tangential component.
Solving $L=mr^2\dot\theta$ for $\dot\theta$ and substituting into the energy eliminates the angle, reducing the planar motion to a one-dimensional radial problem with a centrifugal term $L^2/2mr^2$ .
Key idea: a central force is parallel to $\vec{r}$ , so its torque about the center vanishes; therefore $\vec{L}=\vec{r}\times\vec{p}$ is conserved — its fixed direction traps the orbit in a plane, and its fixed magnitude $L=mr^2\dot\theta$ is Kepler's equal-area law in disguise.

The core insight

A central force is parallel to $\vec{r}$ , so its torque about the center vanishes; therefore $\vec{L}=\vec{r}\times\vec{p}$ is conserved. The constant direction of $\vec{L}$ forces planar motion; the constant magnitude gives $L=mr^2\dot\theta=\text{const}$ — equivalently, a constant areal velocity.

Where this lives today

Every interplanetary mission is flown on equation (4). During a gravity-assist flyby, the spacecraft moves in the planet's (nearly) central field, so its angular momentum about the planet is conserved along the encounter hyperbola; together with energy conservation this fixes the entire flyby geometry, and mission designers chain such encounters years in advance — ESA's JUICE, currently en route to Jupiter for its 2031 arrival, threads four such assists. The same law runs the bookkeeping inside mission software closer to home: for any satellite, one radar snapshot of $(r, v)$ at a known flight angle fixes the areal rate, and with it the period — exactly the computation of Worked Example 2. At the other end of the scale, angular momentum conservation is the central obstacle of astrophysics. A collapsing interstellar cloud spins up as $r^2\dot\theta$ locks in place — shrink $r$ by a factor of a thousand and $\dot\theta$ grows a million-fold — so matter cannot simply fall onto a forming star; it must first shed angular momentum. The result is accretion disks, from protoplanetary disks imaged by ALMA to the glowing ring around M87's black hole photographed by the Event Horizon Telescope: their whole existence, and the slow viscous outward transport of $L$ that lets matter spiral in, is the "angular momentum problem" that this article's one-line conservation law sets up.

The physical picture

Twirl a ball on a string and the string force runs straight along the string to your hand. Pull the string in or let it out and you change the ball's speed — but you can never give it a sideways twist about your hand, because a force pointing at a pivot has no lever arm there. With no twist, the ball's angular momentum about your hand cannot change. A central force is exactly this "string pointing to the center" situation, with one generalization: the pull can follow any radial law $f(r)$ whatsoever — inverse-square gravity, a linear spring, anything — and the no-twist argument never notices. The conservation has two visible signatures. The first is planarity: $\vec{L}=\vec{r}\times m\vec{v}$ is perpendicular to both $\vec{r}$ and $\vec{v}$ , and since $\vec{L}$ is a frozen arrow in space, the position vector must remain forever perpendicular to it. The orbit is trapped in the plane through the center normal to $\vec{L}$ , the way a wheel's spokes can never leave the wheel's plane while the axle stays fixed. Every planet, comet, and spacecraft coasting in a central field moves in a plane chosen once and for all by its initial position and velocity. The second signature is the sweep rate. In a time $\mathrm{d}t$ the radius vector sweeps out a thin triangle of area $\tfrac12 r^2\,\mathrm{d}\theta$ . Constant $L=mr^2\dot\theta$ says this area grows at a constant rate: close to the center the body races through a short, fat triangle; far away it crawls through a long, thin one — and the two areas match exactly. That is Kepler's second law, and it is nothing but angular momentum conservation wearing seventeenth-century clothes.

Figure

Equal areas in equal times — the conserved L sets a universal sweep rate

Before the algebra, fix the geometry of why the torque dies:

Figure

No lever arm — a force aimed at the center can never twist the motion about it

From zero torque to a trapped plane

Angular momentum earns its keep through one definition and one dynamical law. The definition, for a particle of mass $m$ at position $\vec{r}$ (measured from the force center — the choice of origin is part of the physics here) moving with momentum $\vec{p}=m\vec{v}$ :

\vec{L}=\vec{r}\times\vec{p}=m\,\vec{r}\times\vec{v} \qquad (1)

where: $\vec{L}$ — the angular momentum about the center; it is perpendicular to the instantaneous plane spanned by $\vec{r}$ and $\vec{v}$ . In the two-body problem the same formula governs the relative coordinate with $m$ replaced by the reduced mass $\mu$ . Note that $\vec{L}$ depends on the origin you measure $\vec{r}$ from — a particle moving in a straight line has zero $\vec{L}$ about points on the line and nonzero $\vec{L}$ about everything else. How does $\vec{L}$ change in time? Differentiate the definition with the product rule and use Newton's second law: the result is the rotational counterpart of $\dot{\vec{p}}=\vec{F}$ ,

\frac{\mathrm{d}\vec{L}}{\mathrm{d}t}=\vec{N}=\vec{r}\times\vec{F} \qquad (2)

where: $\vec{N}$ — the torque about the center. The term $\dot{\vec{r}}\times m\vec{v}=\vec{v}\times m\vec{v}$ that the product rule also produces vanishes identically, which is why only the force term survives. Equation (2) holds for any force whatsoever; the magic of central forces is what they do to its right-hand side. For a central force $\vec{F}=f(r)\hat{r}$ , the force is parallel (or anti-parallel) to $\vec{r}$ , and the cross product of parallel vectors is zero:

\vec{N}=\vec{r}\times f(r)\hat{r}=\frac{f(r)}{r}\,\big(\vec{r}\times\vec{r}\big)=0 \quad\Longrightarrow\quad \vec{L}=\text{const} \qquad (3)

where: the conclusion needs only $\vec{F}\parallel\vec{r}$ — the radial profile $f(r)$ never enters. Gravity's inverse square, the spring's linear pull, or any exotic $f(r)$ you invent all conserve $\vec{L}$ equally well. This is a symmetry statement, not a statement about gravity: a central force looks the same in every direction from the center, and rotational symmetry is what pays out a conserved angular momentum.

Derivation AZero torque, a frozen vector, and the trapped plane

We are deriving the two headline consequences of a central force in one sweep: the angular momentum vector is a constant of the motion, and the orbit is confined to a fixed plane through the force center. The strategy is to cross Newton's second law with $\vec{r}$ from the left, recognize the result as a total time derivative, and then convert the constancy of a vector into a geometric prison for the position. The step to watch is the last one — a scalar triple product identity does the geometric work, and it is easy to miss that the argument needs $\vec{L}\neq 0$ : the degenerate case $\vec{L}=0$ confines the motion to a line, not a plane.

Step 1. Write the equation of motion for a particle in a central field: $m\ddot{\vec{r}}=f(r)\,\hat{r}$ . Why: this is the most general central force — $f(r)$ is any function of the distance alone, attractive where $f<0$ .

Step 2. Cross both sides with $\vec{r}$ from the left:

\vec{r}\times m\ddot{\vec{r}}=\vec{r}\times f(r)\hat{r}=\frac{f(r)}{r}\,(\vec{r}\times\vec{r})=0

Why: $\vec{r}\times\vec{r}=0$ for any vector — the right side dies for every radial profile $f(r)$ , which is exactly why the conclusion is so general.

Step 3. Recognize the left side as a total derivative:

\vec{r}\times m\ddot{\vec{r}}=\frac{\mathrm{d}}{\mathrm{d}t}\big(\vec{r}\times m\dot{\vec{r}}\big)=\frac{\mathrm{d}\vec{L}}{\mathrm{d}t}

Why the extra product-rule term vanishes: differentiating $\vec{r}\times m\dot{\vec{r}}$ also produces $\dot{\vec{r}}\times m\dot{\vec{r}}$ , which is a vector crossed with itself — zero.

Step 4. Conclude the conservation law: $\mathrm{d}\vec{L}/\mathrm{d}t=0$ , so $\vec{L}=\vec{L}_0$ is fixed in magnitude and direction for the entire history of the motion.

Step 5. Convert the frozen vector into geometry. For any instant,

\vec{r}\cdot\vec{L}=\vec{r}\cdot(\vec{r}\times m\vec{v})=0

Why: a scalar triple product with a repeated vector vanishes identically. Since $\vec{L}=\vec{L}_0$ never moves, $\vec{r}(t)$ is perpetually perpendicular to one fixed direction — the position is trapped in the plane through the force center normal to $\vec{L}_0$ . The three-dimensional problem has collapsed to two dimensions, at zero cost in approximation.

Figure

The frozen L pins the orbital plane — r and v may wander, but only inside it

The scalar workhorse: L in polar coordinates

With the motion flattened into a plane, set up polar coordinates $(r,\theta)$ in that plane, centered on the force center. What does the conserved magnitude $L$ look like in these natural variables? The velocity splits into a radial and a tangential part, $\vec{v}=\dot{r}\,\hat{r}+r\dot\theta\,\hat{\theta}$ , and only the tangential part survives the cross product:

L=m r^2\dot\theta=\text{const} \qquad (4)

where: $r$ — the distance from the center, $\dot\theta$ — the angular velocity in the orbital plane. Read it as a trade law: whenever $r$ halves, $\dot\theta$ must quadruple. A comet plunging toward the Sun must whip around faster — not because anything pushes it sideways, but because nothing does.

Go deeper · Derivation B · 1 minFrom the vector L to the polar scalar and the equal-area law

Derivation BFrom the vector L to the polar scalar and the equal-area law

The goal is the two working formulas of orbit theory: the polar expression $L=mr^2\dot\theta$ , and its geometric alter ego $\mathrm{d}A/\mathrm{d}t=L/2m$ — Kepler's second law. The strategy is a straight computation: decompose the velocity in polar unit vectors, push it through the cross product, then connect $r^2\dot\theta$ to the area swept by the radius vector. Watch how the radial part of the velocity drops out of the cross product — it is the reason approaching or receding motion carries no angular momentum — and note that the final area relation, integrated over a full orbit, already contains the seed of Kepler's third law.

Step 1. In the orbital plane write the velocity in polar form: $\vec{v}=\dot{r}\,\hat{r}+r\dot\theta\,\hat{\theta}$ . Why: $\hat{r}$ stretches the radius, $\hat{\theta}$ swings it — every planar velocity is this sum.

Step 2. Cross with the position $\vec{r}=r\hat{r}$ :

\vec{L}=m\,\vec{r}\times\vec{v}=m r\dot{r}\,(\hat{r}\times\hat{r})+m r^2\dot\theta\,(\hat{r}\times\hat{\theta})=m r^2\dot\theta\,\hat{z}

Why the first term dies: $\hat{r}\times\hat{r}=0$ — radial motion has no lever arm. The second uses $\hat{r}\times\hat{\theta}=\hat{z}$ , the normal to the orbital plane, so the vector result also re-confirms which plane the motion lives in.

Step 3. Read off the conserved magnitude: $L=mr^2\dot\theta$ . Why this is the workhorse: it converts a vector conservation law into one scalar first integral relating $r$ and $\dot\theta$ at every instant.

Step 4. Now the geometry. In time $\mathrm{d}t$ the radius vector and the displacement $\mathrm{d}\vec{r}$ span a triangle of area $\mathrm{d}A=\tfrac12|\vec{r}\times\mathrm{d}\vec{r}|$ . Divide by $\mathrm{d}t$ :

\frac{\mathrm{d}A}{\mathrm{d}t}=\frac12\,\big|\vec{r}\times\dot{\vec{r}}\big|=\frac{1}{2m}\,\big|\vec{r}\times m\dot{\vec{r}}\big|=\frac{L}{2m}=\frac12 r^2\dot\theta

Why this is Kepler's second law: the sweep rate is a constant of the motion — equal areas in equal times, for ANY central force, not just gravity.

Step 5. Integrate over one full period of a closed orbit of semi-axes $a,b$ :

A_{\text{ellipse}}=\pi ab=\frac{L}{2m}\,T \quad\Longrightarrow\quad T=\frac{2\pi m ab}{L}

Why keep this: it converts a single measured snapshot ( $L$ from one $r$ and $v$ ) into the full orbital period — and combined with inverse-square dynamics it grows into Kepler's third law. *Note: inserting $b=a\sqrt{1-e^2}$ and the inverse-square force into $\pi ab=(L/2m)T$ reproduces Kepler's third law.*

The areal velocity deserves its own display, because it is the form in which the conservation law was first discovered — read off planetary tables by Kepler four decades before Newton was born:

\frac{\mathrm{d}A}{\mathrm{d}t}=\frac12\,r^2\dot\theta=\frac{L}{2m}=\text{const} \qquad (5)

where: $A$ — the area swept by the radius vector. For Earth ( $e\approx 0.0167$ ) the orbit is nearly circular, yet the law is still visible in the calendar: we are about 3.4% closer to the Sun in early January than in early July, and orbital speed is higher by the same factor (30.29 versus 29.29 km/s) — one reason the seasons have slightly unequal lengths. The full development of this law and its consequences lives in Kepler's second law and the Kepler problem.

The gateway: eliminating the angle

Here is why orbit theory treats equation (4) as its master key. Solve it for the angular velocity:

\dot\theta=\frac{L}{mr^2} \qquad (6)

where: the right side contains no trace of $\theta$ itself — the angle's rate of change is slaved entirely to the current radius and the conserved constant $L$ . Any expression containing $\dot\theta$ can now be purged of angular variables. The most consequential place to do it is the total energy.

Go deeper · Derivation C · 1 minPurging the angle from the energy

Derivation CPurging the angle from the energy

We are deriving the one-dimensional energy equation (7) — the statement that a planar central-force orbit is energetically equivalent to a single coordinate $r(t)$ moving in a modified potential. The strategy is minimal: write the kinetic energy in polar coordinates, then use the conserved $L$ from equation (6) to eliminate $\dot\theta$ wherever it appears. The point to watch is what the substitution does to the angular kinetic energy: it does not disappear — it reappears as a repulsive function of $r$ alone, the centrifugal barrier, and misreading its sign (it enters the effective potential as $+L^2/2mr^2$ , a positive, repulsive term) is the standard error downstream.

Step 1. Write the total energy in the orbital plane using $\vec{v}=\dot{r}\,\hat{r}+r\dot\theta\,\hat{\theta}$ :

E=\frac12\,m\big(\dot{r}^2+r^2\dot\theta^2\big)+U(r)

Why: the two velocity components are orthogonal, so the kinetic energy splits cleanly into a radial and an angular piece.

Step 2. Eliminate $\dot\theta$ with equation (6): $r^2\dot\theta^2 = r^2\,(L/mr^2)^2 = L^2/m^2r^2$ . Why this is legitimate: $L$ is a constant of the motion, so the substitution trades a dynamical variable for a fixed number — no information is lost, and $E$ remains exactly conserved.

Step 3. Assemble the one-dimensional energy equation:

E=\frac12\,m\dot{r}^2+\frac{L^2}{2mr^2}+U(r) \qquad (7)

Why this is a reduction: everything on the right depends on $r$ and $\dot{r}$ alone. The angular motion survives only as the positive term $L^2/2mr^2$ — a "centrifugal barrier" that blows up at small $r$ and guards the center against any orbit with $L\neq 0$ .

Equation (7) says the two-dimensional orbit has been reduced to a bead sliding in the combined potential $U_{\text{eff}}(r)=U(r)+L^2/2mr^2$ , where the barrier term is the ghost of the eliminated angular motion. How this reduction works in full, and what the barrier means physically, are the subjects of the next two articles in the chain: reduction to the radial equation and the effective potential. One step further, trading $t$ for $\theta$ via (6) yields Binet's orbit equation for the orbit's shape.

Which equation answers which question

you want	use	notes
is $\vec{L}$ conserved here?	check $\vec{F}\parallel\vec{r}$ , eq. (3)	any $f(r)$ qualifies; tangential components disqualify
the orbital plane	normal to $\vec{L}_0=m\,\vec{r}_0\times\vec{v}_0$	set once by initial conditions
relate $r$ and $\dot\theta$ along the orbit	$L=mr^2\dot\theta$ , eq. (4)	the scalar workhorse
speeds at two apsides	$r_p v_p=r_a v_a$	velocity purely tangential at apsides only
sweep rate / Kepler II	$\mathrm{d}A/\mathrm{d}t=L/2m$ , eq. (5)	any central force, not just gravity
period of a closed orbit	$T=2\pi m ab/L$	from integrating the sweep rate
eliminate $\theta$ from dynamics	$\dot\theta=L/mr^2$ , eq. (6) → eq. (7)	gateway to the radial equation

Limiting cases

Case	Result	Meaning
inverse-square force, $f(r)=-k/r^2$	$\vec{L}$ conserved, plus a bonus conserved vector (Laplace–Runge–Lenz)	closed, non-precessing ellipses — the Kepler problem
isotropic oscillator, $f(r)=-kr$	$\vec{L}$ conserved just the same	closed ellipses again, but centered on the force center, not a focus
tangential perturbation, $\vec{F}=f(r)\hat{r}+g(r)\hat{\theta}$	torque $N_z=r\,g(r)\neq 0$	$L$ drifts; the orbit precesses or spirals secularly
linear drag, $\vec{F}_{\text{drag}}=-b\vec{v}$	$\mathrm{d}\vec{L}/\mathrm{d}t=-(b/m)\,\vec{L}$ , so $L(t)=L_0e^{-bt/m}$	direction fixed, magnitude decays — a clean inspiral
$\vec{L}_0=0$	$\vec{r}\parallel\vec{v}$ at all times	motion on a straight line through the center — the "plane" degenerates to a line

Common trapAngular momentum is conserved because gravity is inverse-square — no, and the converse trap too

The conservation needs only $\vec{F}\parallel\vec{r}$ . The radial profile is irrelevant: the isotropic harmonic oscillator $f=-kr$ conserves $\vec{L}$ exactly as well as gravity does. What the inverse-square law adds is a second, separate conserved quantity (the Laplace–Runge–Lenz vector) that closes the ellipse — do not credit $\vec{L}$ for that. The converse trap: a force whose magnitude is isotropic but which carries any tangential component $g(r)\hat{\theta}$ breaks the conservation immediately, because the torque $rg(r)$ no longer vanishes. And a bookkeeping trap on top: it is the torque about the force center that vanishes. About any other origin, $\vec{r}\times\vec{F}\neq 0$ and $\vec{L}$ about that point is not conserved — always state your origin before invoking equation (3).

Worked examples

1 · Halley's comet, fast and slow (confidence builder)

Halley's comet has perihelion $r_p=0.586\ \mathrm{AU}$ and aphelion $r_a=35.1\ \mathrm{AU}$ . Its perihelion speed is $v_p=54.6\ \mathrm{km/s}$ . Find its aphelion speed.

Step 1. At the apsides — and only there — the velocity is purely tangential ( $\dot{r}=0$ ), so the angular momentum takes the simple form $L=mrv$ at both points.

Step 2. Conservation between the two apsides: $m r_p v_p = m r_a v_a$ , so $v_a = v_p\,(r_p/r_a)$ — the mass never mattered.

Step 3. Insert numbers: $v_a = 54.6 \times (0.586/35.1) = 54.6 \times 0.01670 = \boxed{0.91\ \mathrm{km/s}}$ .

Step 4. Sanity check: the distance ratio is $35.1/0.586 = 59.9$ , so the speed ratio must also be $59.9$ — sixty times closer, sixty times faster. Note what we did not need: the comet's mass, the Sun's mass, or even the fact that gravity is inverse-square. One conservation law, one line.

2 · A satellite's period from a single snapshot (exam level)

A satellite passes perigee at $r_p=7000\ \mathrm{km}$ from Earth's center with speed $v_p=8.5\ \mathrm{km/s}$ . Using $\mu_\oplus \equiv GM_\oplus=3.986\times10^{5}\ \mathrm{km^3/s^2}$ , find the apogee speed and the orbital period — using the areal law for the period, then cross-checking with Kepler's third law.

Step 1. The areal velocity is fixed for the whole orbit by this one snapshot (velocity is tangential at perigee): $\mathrm{d}A/\mathrm{d}t = \tfrac12 r_p v_p = \tfrac12 \times 7000 \times 8.5 = 29{,}750\ \mathrm{km^2/s}$ .

Step 2. Find the orbit's size from the energy (per unit mass): $\varepsilon = \tfrac12 v_p^2 - \mu_\oplus/r_p = \tfrac12(8.5)^2 - 3.986\times10^5/7000 = 36.13 - 56.94 = -20.82\ \mathrm{km^2/s^2}$ . The semi-major axis follows from $\varepsilon = -\mu_\oplus/2a$ : $a = 3.986\times10^5/(2\times20.82) = 9573\ \mathrm{km}$ .

Step 3. Apogee distance and speed: $r_a = 2a - r_p = 19{,}147 - 7000 = 12{,}147\ \mathrm{km}$ , and by angular momentum conservation between apsides $v_a = v_p\,(r_p/r_a) = 8.5 \times 7000/12{,}147 = \boxed{4.90\ \mathrm{km/s}}$ .

Step 4. Total area of the ellipse: eccentricity $e = 1 - r_p/a = 1 - 7000/9573 = 0.269$ , semi-minor axis $b = a\sqrt{1-e^2} = 9573\times0.9632 = 9221\ \mathrm{km}$ , so $A = \pi ab = \pi \times 9573 \times 9221 = 2.773\times10^{8}\ \mathrm{km^2}$ .

Step 5. The period is total area over sweep rate: $T = A\,/\,(\mathrm{d}A/\mathrm{d}t) = 2.773\times10^{8}/29{,}750 = \boxed{9323\ \mathrm{s} \approx 2\ \mathrm{h}\ 35\ \mathrm{min}}$ .

Step 6. Cross-check with Kepler's third law: $T = 2\pi\sqrt{a^3/\mu_\oplus} = 2\pi\sqrt{(9573)^3/3.986\times10^5} = 2\pi \times 1484 = 9322\ \mathrm{s}$ — agreement to rounding. The areal law turned one instant of tracking data into the whole clock of the orbit.

Origins · 1 minJohannes Kepler and how this was discovered

Johannes Kepler

1571–1630 · Imperial mathematician, Prague

Kepler found the equal-area law the hard way: buried in Tycho Brahe's Mars observations, the most precise astronomical data then in existence, which he inherited — after considerable intrigue — on Tycho's death in 1601. His "war with Mars" lasted years; the second law (found before the first, around 1602) emerged from brutal arithmetic long before calculus existed to make it natural. Kepler himself had no concept of angular momentum: he pictured a magnetic "motive soul" streaming from the rotating Sun, sweeping the planets around like spokes of a broom. That the equal-area law is really a statement about any force aimed at the Sun — true far beyond gravity — was Newton's insight, proved geometrically as the very first proposition of the Principia (1687). Kepler's working life was repeatedly upended: religious expulsions, the Thirty Years' War, and a years-long legal defense of his mother against witchcraft charges — through which he still produced the Rudolphine Tables, the best planetary tables of the century.

History

1609

— Kepler's Astronomia Nova publishes the equal-area law, distilled from Tycho Brahe's Mars data — a conservation law recorded eight decades before anyone could derive it.

1687

— Newton's Principia, Book I, Proposition 1: a purely geometric proof that equal areas follow from any force directed at a fixed center — Kepler's law becomes a theorem about central forces in general.

1746

— Patrick d'Arcy states the "principle of areas" for general systems; with Euler and Daniel Bernoulli, angular momentum begins its life as an independent mechanical quantity.

1834

— Hamilton's dynamics absorbs angular momentum into the general machinery of conserved quantities and symmetries.

1918

— Emmy Noether proves that every continuous symmetry yields a conservation law: conservation of

\vec{L}

is the fingerprint of rotational invariance.

Going further

Texts. Morin, Introduction to Classical Mechanics, Ch. 7 (Central Forces) · Goldstein, Poole & Safko, Classical Mechanics, 3rd ed., Ch. 3 §3.1–3.2 · Landau & Lifshitz, Mechanics, 3rd ed., §14 (Motion in a central field) · Taylor, Classical Mechanics, Ch. 8. The key results on this page are: $\mathrm{d}\vec L/\mathrm{d}t=\vec r\times\vec F$ via the product rule with $\vec v\times\vec v=0$ ; $L=mr^2\dot\theta$ from $\hat r\times\hat\theta=\hat z$ ; $\mathrm{d}A/\mathrm{d}t=L/2m$ ; the apsidal speed ratio $r_pv_p=r_av_a$ ; and Earth's eccentricity $e\approx0.0167$ . No nonstandard numbers are used. Worked-example values (Halley $v_a=0.91$ km/s; satellite $a=9573$ km, $v_a=4.90$ km/s, $T\approx9320$ s, cross-checked two ways) are arithmetic results from the boxed formulas.

Builds on: Central force fundamentals · The two-body problem and reduced mass. Leads to: Kepler's second law: equal areas · Reduction to the radial equation · The effective potential · The orbit equation and Binet's equation.

Does angular momentum conservation require the inverse-square law?

No — this is the single most common confusion in the topic. The derivation needs only $\vec{F}=f(r)\hat{r}$ , i.e. a force parallel to the position vector; then $\vec{r}\times\vec{F}=0$ no matter what $f(r)$ is. The isotropic harmonic oscillator $f=-kr$ conserves $\vec{L}$ exactly as well as gravity. What is special about the inverse-square law is a different, additional conserved quantity — the Laplace–Runge–Lenz vector — which is what makes Kepler ellipses close without precessing. Keep the two separate: $\vec{L}$ conservation is generic central-force physics; orbit closure is inverse-square physics.

Why exactly is the motion confined to a plane, and what happens if L = 0?

$\vec{L}=\vec{r}\times m\vec{v}$ is perpendicular to $\vec{r}$ by construction — that is just the scalar triple product $\vec{r}\cdot(\vec{r}\times m\vec{v})=0$ . Ordinarily that means nothing special, because $\vec{L}$ could change direction and let the "perpendicular plane" tilt around. The central-force miracle is that $\vec{L}$ is frozen: one fixed direction, forever. Then $\vec{r}\cdot\vec{L}_0=0$ pins the position into the single plane through the center normal to $\vec{L}_0$ . If $\vec{L}_0=0$ , the position and velocity are parallel at every instant, and the "plane" degenerates further — the particle moves on a straight line through the force center: a radial plunge or escape, the extreme case of an orbit with no sideways motion at all.

Is the ice skater pulling in her arms the same physics as the comet speeding up at perihelion?

The conservation law is the same; the mechanism is instructively different. The comet's angular momentum is conserved because the Sun's pull is central — an external force with zero torque. The skater's is conserved because the forces that pull her arms inward are internal to her body, and internal forces cancel in pairs when you compute the total torque. In both cases $L$ stays fixed while the mass distribution tightens, so the rotation rate must rise: the comet obeys $r^2\dot\theta=\text{const}$ , the skater $I\omega=\text{const}$ . The comet version is the purer one — a single particle, a single conserved product $mr^2\dot\theta$ — and it is the version that powers all of orbit theory.

Angular momentum about which point? Does the choice of origin matter?

It matters completely. $\vec{L}=\vec{r}\times\vec{p}$ depends on where you measure $\vec{r}$ from, and the conservation argument works only about the force center, because only there is $\vec{F}$ parallel to $\vec{r}$ . Compute the torque of the Sun's pull about any other point and you get something nonzero — the planet's angular momentum about that point genuinely changes in time. This is why the two-body problem is always reduced first: after the reduction, the relative coordinate feels a central force about the center of mass frame's origin, and the conserved $\vec{L}$ is that of the relative motion with the reduced mass, $\vec{L}=\mu\,\vec{r}\times\dot{\vec{r}}$ .

If L is always conserved for gravity, why do real satellite orbits precess and decay?

Because the real Earth is not a point mass, and low orbits are not in vacuum. The oblate Earth's equatorial bulge makes the gravitational field slightly non-central — the force no longer points exactly at the center of mass — so a small torque acts and the orbital plane precesses (the $J_2$ effect, several degrees per day for low orbits; sun-synchronous satellites exploit it deliberately). Atmospheric drag is a velocity-dependent force, $-b\vec{v}$ -like, whose torque $\vec{r}\times(-b\vec{v})=-(b/m)\vec{L}$ shrinks $L$ exponentially and produces the slow inspiral. Conservation laws are exact only as long as their symmetry is: every observed violation is a measurement of the perturbation that broke the symmetry.

Kepler's second law was discovered before Newton's laws. How can you find a conservation law without knowing the dynamics?

Empirically — the way Kepler did it. Tycho Brahe's Mars data gave positions on the sky over decades, accurate to about two arcminutes. Kepler reconstructed the orbit geometrically and noticed the sweep-rate regularity around 1602, publishing it in 1609. He had no dynamics to explain it; his own theory involved a "motive soul" streaming from the Sun. The law sat as pure phenomenology for nearly eighty years until Newton's Principia opened with precisely this theorem: equal areas hold if and only if the force always points at the center. That equivalence is what made Kepler's second law the evidence that the Sun's force is central — a beautiful case of a conservation law observed first and understood later.

Earth's orbital speed varies by only 3% — does the equal-area law have any visible effect on daily life?

Yes: it is written into the calendar and the sundial. Because Earth sweeps area at a constant rate but sits at varying distance, its angular speed around the Sun is larger in January (perihelion) than in July — so the seasons are unequal: northern-hemisphere winter (solstice to equinox) is about 89 days while summer is about 93.7. The same varying angular rate makes the solar day (noon to noon) drift up to ±16 minutes relative to a uniform clock across the year — half of the famous "equation of time" that makes sundials run fast or slow, and the reason the figure-eight analemma printed on globes is lopsided. All of that is $mr^2\dot\theta=\text{const}$ acting on a 1.67% eccentricity.

How does this connect to Noether's theorem — is the conservation "because of" symmetry or "because of" zero torque?

They are the same statement at two levels of sophistication. The elementary argument — $\vec{F}\parallel\vec{r}$ makes the torque vanish — is a computation you can do in one line from Newton's laws. Noether's 1918 theorem explains why such a computation had to work: the potential $U(r)$ depends only on distance, so the system is invariant under rotations about the center, and every continuous symmetry generates a conserved quantity — rotational invariance generates exactly $\vec{L}$ . The Noether view pays off when systems get complicated: even where torque bookkeeping is a mess, symmetry inspection tells you instantly which components of $\vec{L}$ survive. A potential symmetric only about the $z$ -axis (a spinning oblate planet, say) conserves only $L_z$ — and that is precisely the component satellite orbits keep while their planes precess.

Angular Momentum Conservation in Central Fields

Where this lives today

The physical picture

From zero torque to a trapped plane

The scalar workhorse: L in polar coordinates

The gateway: eliminating the angle

Which equation answers which question

Limiting cases

Worked examples

1 · Halley's comet, fast and slow (confidence builder)

2 · A satellite's period from a single snapshot (exam level)

Going further

Before this

Where this leads