LAW 1The orbit of a planet is an ellipse with the Sun at one of the two foci.
Johannes Kepler · 1609–1619
Kepler found the rules by fitting Tycho Brahe's observations of Mars. Each law below is a working simulation you can bend with your own hands — change the orbit, and watch what stays true.
LAW 1The orbit of a planet is an ellipse with the Sun at one of the two foci.
LAW 2A line joining a planet and the Sun sweeps out equal areas during equal intervals of time.
LAW 3The square of a planet's orbital period is proportional to the cube of its orbit's semi-major axis: T² ∝ a³.
Take the planet's velocity arrow at every instant, slide them all so their tails sit on one common origin, and watch where the tip travels. It does not wander, spiral or smear — it draws a perfect circle. That circle is called the hodograph, and it is the largest chapter on this page.
Every motion here is a true Kepler solver — mean anomaly M advances uniformly, then Kepler's equation M = E − e·sin E is solved by Newton iteration each frame. No uniform circular shortcuts.
“The orbit of every planet is an ellipse with the Sun at one of the two foci.”
Not a circle, and the Sun is not at the centre — it sits at one focus, F₁. The second focus, F₂, is empty space, yet it is geometrically real: for every point on an ellipse, the distances to the two foci add to the same constant, r₁ + r₂ = 2a. Watch the live check under the simulation — it holds at every instant, for every eccentricity.
Drag e toward 0.9 and the orbit stretches into a comet-like sliver; set e = 0 and the two foci merge into the centre of a circle. The shape is set by just two numbers: the semi-major axis a and the eccentricity e.
“A line joining a planet and the Sun sweeps out equal areas during equal intervals of time.”
The simulation below slices one orbit into 12 equal time steps of T/12 each and shades the sector the Sun–planet line sweeps in each step. Near the Sun the sector is short and fat; far away it is long and thin — but every sector's area, computed numerically from the simulated positions, comes out the same. That is why the planet visibly hurries near perihelion and dawdles near aphelion: the area rate dA/dt = h/2 never changes, so a shorter radius must be swept faster.
In modern language this is conservation of angular momentum: h = r²·dν/dt is constant because gravity pulls along the radius and exerts no torque.
“The square of a planet's orbital period is proportional to the cube of the length of the semi-major axis of its orbit.”
Kepler published this in Harmonices Mundi (1619), a decade after the first two laws. It is a single rule for the whole system: measure any planet's distance and period, and a³/T² comes out the same — about 7.496 × 10⁻⁶ AU³/day² — because the constant is set by the Sun's mass alone. Tap any planet in the table to load its a and e into all the simulators on this page.
| Planet | a (AU) | T (days) | T (years) | e | a³/T² (10⁻⁶ AU³/day²) |
|---|
On logarithmic axes, T = a^1.5 becomes a straight line of slope 3/2. All eight planets sit on it — Mercury at the fast, close-in end, Neptune taking 165 years at the far end. A slope of 1 would mean T ∝ a; nature chose 3/2, and that exponent is what Newton later traced to the inverse-square law.
Here is the construction, running live and synchronised. On the left, ordinary position space: the planet rides its ellipse, the blue arrow is its position vector r, the teal arrow at the planet is its velocity v, tangent to the orbit. On the right, velocity space: the same velocity arrow, picked up and re-drawn with its tail pinned to a fixed origin O. Only the arrow's tip is free to move — and as the planet completes its ellipse, the tip completes a circle.
Read the right panel carefully: its axes are velocities (km/s), not positions. The circle is not the orbit — the planet never travels around it in space. It is the orbit's shadow in velocity space, and its geometry encodes the whole motion: the circle's radius R = μ/h is set by angular momentum, and its centre is displaced from the origin by exactly e·R. The fading ribbon on the right panel is the velocity history: every velocity tip from the last full orbit, each coloured by its own speed, with a dot every T/24 — the dots bunch where the tip lingers and spread out where it rushes. Drag the eccentricity slider above the panels to stretch the orbit from a circle into a comet-like ellipse and watch the history redraw itself on the new circle.
The one reading to take away: on the hodograph, distance from the origin is speed. So the tip's farthest point from O is perihelion — the planet at its fastest — and its nearest point is aphelion, the slowest. For a near-circular orbit like Earth's, the origin sits almost at the circle's centre and the speed barely changes (30.29 vs 29.29 km/s). Load Halley's comet (e = 0.967) and the origin nearly touches the circle: the tip swings from 54.5 km/s at perihelion down to under 1 km/s at aphelion.
And once more, because it is the whole point: the axes on the right are velocities in km/s, not positions. The circle is not the orbit. The orbit is the ellipse on the left; the circle is what the velocity vector draws while the planet walks the ellipse.
Nothing about an ellipse suggests a circle hiding inside it, yet the proof is short. Work in the perifocal frame: x-axis toward perihelion, y-axis 90° ahead in the direction of motion, ν the true anomaly, μ = GM☉ and h the (constant) specific angular momentum.
Gravity is the only force, and it points at the Sun: a = dv/dt = −(μ/r²)·r̂. By Law 2, h = r²·dν/dt is constant, so the clock can be re-read as an angle: d/dt = (h/r²)·d/dν.
The r² cancels exactly — this cancellation is the inverse-square law doing its one trick. Integrating the rotating unit vector r̂ = (cos ν, sin ν) then gives v as plain sines and cosines, with one constant of integration: the eccentricity e.
Set R = μ/h. Then vx = −R·sin ν and vy = eR + R·cos ν, so the tip satisfies a circle equation identically, for every ν: radius R, centre displaced e·R along the vy axis. At e = 0 the centre lands on the origin and the speed never changes.
Why doesn't the circle's centre drift as the planet moves? The offset is locked to the eccentricity vector — the Laplace–Runge–Lenz vector — a conserved quantity that points toward perihelion and keeps the ellipse from precessing. It exists only for an inverse-square force. Perturb the force law even slightly and the LRL vector slowly rotates: the hodograph stops being an exact circle and the orbit stops closing. The perfect circle is the fingerprint of 1/r².
It is also why the result feels miraculous: energy and angular momentum alone don't fix the orbit's orientation. The LRL vector is an extra, "hidden" symmetry of the Kepler problem — in quantum mechanics the same symmetry explains the degeneracy of the hydrogen spectrum.
William Rowan Hamilton presented the hodograph to the Royal Irish Academy in 1846 and named it from the Greek hodos (path) and graphein (to write): the "path-writer" of the velocity vector. Möbius arrived at the same construction independently in the same decade, and Richard Feynman famously re-derived the circle from scratch in his 1964 lecture on planetary motion — the "lost lecture," reconstructed and published decades later.
Only bound (elliptical) orbits are simulated on this page, but the geometry extends: a parabolic orbit (e = 1) still gives a circle — one that passes exactly through the origin, the tip reaching zero velocity at infinity — while a hyperbolic orbit (e > 1) traces only an arc of its circle, because ν never sweeps the full turn.
Answered 0 of 5 · 0 correct