Why does Fizeau's wheel go dark at a certain speed?

Light passes one gap, travels 8.63 km to a mirror and back in about 58 µs. If the wheel turns exactly half a tooth pitch in that time, the returning light hits the next tooth and is blocked. From 2L/c = 1/(2Nf) you get c = 4NLf; a 720-tooth wheel first goes dark near 12 rev/s, and again at three times that speed.

How did watching Jupiter's moon measure the speed of light?

Io's orbit is a very steady clock and it is eclipsed once per orbit. As Earth moves away from Jupiter, light from each eclipse has a little further to travel, so eclipses arrive later and later; as Earth approaches they come earlier. The accumulated delay over half a year is the time light needs to cross Earth's orbit, about 16.6 minutes today. Rømer's estimate was about 22 minutes, so his value of c came out low.

Is the speed of light still measured today?

No. Since 1983 c is defined as exactly 299 792 458 m/s and the metre is the distance light travels in 1/299 792 458 of a second. Modern "speed of light" experiments really calibrate lengths using that fixed c and atomic clocks.

What did the Michelson–Morley null result show?

If light moved at c relative to an ether, turning the interferometer should have shifted the fringes by about 0.4 fringe. The observed shift was under 0.01, meaning any ether wind was below roughly 5 km/s, far less than Earth's 30 km/s orbital speed. Special relativity later explained it: the speed of light in vacuum is the same in every inertial frame, and there is no detectable ether.

Why is the 1935 value low and outside its own error bar?

That experiment bounced light many times inside a 1.6 km vacuum pipe and reported 299 774 ± 11 km/s, about 18 km/s too low. Unaccounted systematic errors such as ground vibration and length measurement are usually blamed. An error bar only covers the uncertainties the authors knew about; systematic errors tend to show up only when different methods are compared.

💨

💨Speed of Light History - Fizeau Wheel, Rømer, Michelson–Morley Simulator

From Galileo flashing lanterns between hills to fixing c as an exact number in 1983, every classic experiment here is a simulation you can tune yourself. Spin Fizeau's wheel until the light first goes dark, watch Io's eclipses drift 16.6 minutes late as Earth orbits, and compute c from your own reading. A timeline with error bars then shows how the measurements closed in on 299 792 458 m/s.

Free · No sign-up · Runs in your browserLast updated: 2026-10-29

Loading tool...

How to use Speed of Light History

1Pick an experiment at the top, for example "Fizeau's wheel".
2Drag the sliders or type the teeth, distance or speed; the animation and chart update live.
3Find the key state (darkness, a steady image and so on) and read the speed of light your measurement implies.
4Open the Michelson–Morley tab and compare the fringe shift the ether predicts with the null result.
5Use the timeline to see historic error bars, including the 1935 value that missed its own.

FAQ

Why does Fizeau's wheel go dark at a certain speed?: Light passes one gap, travels 8.63 km to a mirror and back in about 58 µs. If the wheel turns exactly half a tooth pitch in that time, the returning light hits the next tooth and is blocked. From 2L/c = 1/(2Nf) you get c = 4NLf; a 720-tooth wheel first goes dark near 12 rev/s, and again at three times that speed.
How did watching Jupiter's moon measure the speed of light?: Io's orbit is a very steady clock and it is eclipsed once per orbit. As Earth moves away from Jupiter, light from each eclipse has a little further to travel, so eclipses arrive later and later; as Earth approaches they come earlier. The accumulated delay over half a year is the time light needs to cross Earth's orbit, about 16.6 minutes today. Rømer's estimate was about 22 minutes, so his value of c came out low.
Is the speed of light still measured today?: No. Since 1983 c is defined as exactly 299 792 458 m/s and the metre is the distance light travels in 1/299 792 458 of a second. Modern "speed of light" experiments really calibrate lengths using that fixed c and atomic clocks.
What did the Michelson–Morley null result show?: If light moved at c relative to an ether, turning the interferometer should have shifted the fringes by about 0.4 fringe. The observed shift was under 0.01, meaning any ether wind was below roughly 5 km/s, far less than Earth's 30 km/s orbital speed. Special relativity later explained it: the speed of light in vacuum is the same in every inertial frame, and there is no detectable ether.
Why is the 1935 value low and outside its own error bar?: That experiment bounced light many times inside a 1.6 km vacuum pipe and reported 299 774 ± 11 km/s, about 18 km/s too low. Unaccounted systematic errors such as ground vibration and length measurement are usually blamed. An error bar only covers the uncertainties the authors knew about; systematic errors tend to show up only when different methods are compared.

Features

Nine tabs: Galileo, Rømer, Bradley, Fizeau, Foucault, Michelson, modern methods, Michelson–Morley, timeline
Fizeau animation keeps the true phase relation, so light pulses pass gaps or hit teeth, with a brightness vs speed curve
Rømer tab computes the Earth–Jupiter distance and eclipse delay from real orbits, and turns your delay into c
Foucault image shift, Michelson's steady-image speed, cavity and laser methods all work forwards and backwards
Michelson–Morley uses the exact round-trip formula for the expected fringe shift and gives an ether wind limit
Historic values with their published uncertainties, plus a zoomed view of the difference from the defined value

Related tools

View all 196 Physics Tools →