I'm sorry if this is a dumb question. Recently I was learning that it's impossible for us to measure the speed of light in one direction. We can only measure it in two directions and we assume that it's speed is the same in both directions by convention.


But we have been able to directly observe the effects of the finite speed of light since the 1600s, when astronomers tried to use Io as a way to measure time without a clock and found that it appeared to orbit slightly faster while the Earth was moving towards it and slightly slower while moving away from it, for a total maximum discrepancy of 16 minutes. This delay is not from light bouncing off of Io's surface, since it does so in both situations. The delay is only from the light needing to travel an extra 186 million miles between the two measurements.

So my question is, is this second example an actual measurement of the speed of light in one direction? I don't see why one source says we can't measure light in one direction while the second source suggests that we can. Please help me clear up this contradiction.

  • $\begingroup$ 1. First set up a method to determine if the speed differs in two directions. e.g. use a resonant cavity. 2. Then measure the there-and-back speed. $\endgroup$ Jun 21, 2021 at 10:34

5 Answers 5


Yes, it is often assumed that Rømer measured the speed of light in one direction. It may seems trange, but Rømer velocity is also the velocity obtained under the tacit assumption of the equality of the speeds of light in opposite directions. The fact of the matter is that Rømer and Cassini were speculating about the movement of Jupiter’s satellites, automatically assuming that the observers’ space was isotropic.

The Estonian - Australian physicist Leo Karlov showed that Rømer actually measured the speed of light by implicitly making the assumption of the equality of the speeds of light back and forth.

L. Karlov, “Does Roemer's method yield a unidirectional speed of light?” Australian Journal of Physics 23, 243-258 (1970)


L. Karlov “Fact and Illusion in the speed of light determination of the Roemer type” American Journal of Physics, 49, 64-66 (1981)

Some reflections on the one-way speed of light are here.

Another interesting method to measure one - way speed of light that you may discover soon or later was so - called Double Fizeau Toothed wheel. That is two toothed wheel attached to opposite sides of long rotating shaft and a beam of light between the teeth. This method was employed (probably without proper due - diligence) by S. Marinov and M. D. Farid Ahmet.

However, Herbert Ives in his 1939 article "Theory of Double Fizeau toothed wheel" predicted that outcome of the measurement will be exactly c due to relativistic twist of the rotating shaft.


With two modern atomic clocks on well separated mountains, one observer could record the time when a short pulse passed his clock, and the second observer could record the time when it arrived at his. You might want to synchronize the clocks at a mid point, and position them at the same altitude.

  • 1
    $\begingroup$ When you move your synchronized clocks apart, they no longer keep their synchronization if the speed of light is different in one way and the other. Time dilation will affect them differently. $\endgroup$ Feb 24, 2021 at 21:58

I had a thought for measuring light speed in one direction that wasn't affected by the reasons presented in Veritasiums video. Set up a string of light sensitive materials, equal distances apart, along the path of the laser. When those materials are hit by the laser, it triggers 2 other lasers to fire, 1 laser aimed at the next light sensitive material in the string, the other a millisecond burst to an observation post. (For more reliable measures, the observation post should be at the start or end point of the experiment so all the light is travelling along the same path.)

Now all you do is time the delay between the 2nd lasers pulses. Now, repeat the experiment by firing a laser in the opposite direction, and if there is a difference in the speed of light, in either direction, the 2nd laser will show it by either arriving closer together or farther apart.

To make sure the test is as accurate as possible, use clocks at each laser station just for measuring the time from when the laser from the previous station was received, to when the lasers from that station was fired. And to make sure both lasers on the station fired at the exact same time. Adjust calculations if that was indeed the case.

This would remove the need for 2 clocks (one at each end), mirrors, or anything else that vast distances can distort.

  • $\begingroup$ It doesn’t make any sense. If the light sensitive matters fire the second light burst towards the end point both light beam are reaching at the same end how is the opposite direction considered? Its never possible to measure at multiple points across the path without using multiple clocks at the same time without sync problem. $\endgroup$ Nov 24, 2021 at 18:42

The speed of light has never been measured in one direction but always using a reflected media. Thus c is constant in that method.

In ref to this I present a video by YouTuber Veritasium https://youtu.be/pTn6Ewhb27k

I also suggest thae following experiment, Using a photon emitter with a clock, target a second photon emitter with a clock.

The emitter pulses photons at 1 ms bursts of 1ms and each records the recieved photon. Have the test run for 10 minutes several times, one at dawn, one at noon and one at dusk, this will give data corresponding to the location of the test, it's direction and that on the sun.

I also submit that additional tests be done at the earth apogee and at the moons apogee to add that data. This will also give additional information with regards to the gravitational pull of those bodys that effect us most.

If light is constant then the data will prove this, if not then we will have mpre questions.


We need to redefine simultaneity. My proposition is as follows: if a rigid body AB of length l is moving without any acceleration parallel to X axis and at time t0 its point A is at location x, then simultaneously its point B is at location x+l

Let's now design the experiment to synchronize distant clocks and/or measure one way speed of light:

Imagine four spaceships flying as perfect square EFGH towards (or away from) not moving (at least relative to each other) points ABCD, where AD is parallel to EF and distance EF equals AD. Points EG should be collinear with points AB and points FH collinear with CD. Making sure that ABCD (and EFGH) is a square is relatively easy, since 2-way speed of light is constant:

we can measure (and correct, if necessary) distances BD and CA by sending light signals from B to D (and from C to A) and back

Inline image

Now at certain time (clocks at A and D can be pre-synchronized using Einstein convention, but it is not absolutely necessary) we can measure distance from A to G (L) and from D to H (L’) using light (laser) signal send from A to G (and reflected back to A) as well as distance from D to H. If the distance AG (L) equals DH (L’) signals from A and D had been sent simultaneously; if not, it would be easy to adjust the clocks so they are synchronized.

We can also measure one way speed of light directly, without synchronizing the clocks: Let's send the light signal from A to G (and reflect it back to A) and simultaneously from A to a reflector at D and reflect the light from D to H (and from H back to D). If we measure the time for the light to travel from A to G and subtract it from it the time the light takes to travel from D to H', we will have the time the light travels from A to D. Please let me know if I made any wrong assumption

Of course, theoretically it would be sufficient to have only the lines AD parallel to GH, but practically it could be difficult to make sure they are parallel to each other


Not the answer you're looking for? Browse other questions tagged or ask your own question.