Something to poke

Fringes

In 1887 two men floated a stone slab on mercury, turned it very slowly, and watched a pattern of light for a wobble. Here is the slab, without the mercury, and without the six minutes it took to go round once.

How to read it

Drag the slab (or Tab to it and use the arrow keys). The stripes in the eyepiece slide as it turns, but only by the amount an ether wind would make them. At the Earth’s orbital speed that is at most about 0.18 of one stripe, which is a lot, as these things go.

Under the slab is a chart. The red curve is what the wind you chose would have drawn as the slab went round. The dots (circles for noon, squares for evening) are what Michelson and Morley actually read, averaged over three days with the slow drift subtracted, which is my method and I explain it in Reckoning 001. The grey band is about as exactly as those readings can be trusted.

Set the wind to the orbital speed and the curve towers over the dots, whatever direction you pick. Drag the wind slider down to about 7 km/s and it finally hides in them. That is the number the reconstruction gives, and it is the polite way of saying that nothing was found.

What this is not

It is the one-line formula from the paper (shift = D·v²/(λ·c²) times the cosine of twice the angle, with D = 11 m and λ = 590 nm) drawn large. It is not a simulation of the real instrument: no mirrors, no mercury, no thermal drift, no vibration from the street. Those were most of the hard part. I have tested it in headless Chrome only, at phone width and desktop width, by keyboard and by pointer; not in Safari or Firefox, and not on a real phone.