Glenlair, Autumn 1865

The velocity came out of the equations at Glenlair in the autumn of 1865, and I confess I checked it three times hoping it would not. Take the ratio of electromagnetic to electrostatic units - Weber’s measurements give 310,740 kilometres per second - and the waves in the medium must propagate at exactly that speed. Fizeau, measuring light through air between Suresnes and Montmartre, had found 314,858. The agreement sits within the error of both determinations, and no amount of re-derivation moves it.

I had wanted the equations to describe a mechanical medium; I built the model with vortices and idle wheels precisely so that the displacement current would have something visible to displace. What I did not bargain for was that the same mathematics, once the displacement term was admitted, would not stay in the machine. It demanded waves. Transverse waves, at the speed of light, in the very medium that carries induction between two charged bodies. I set out to account for Faraday’s lines and found myself accounting for colour.

The coupling is the whole argument. Charge moves; the medium between takes up the strain; the strain, changing, acts as a current; that current regenerates the changing field; and the disturbance propagates outward at a speed fixed by the ratio of the two electrical constants, not by anything I choose. The phase relation is exact - no lag permitted, no damping in the ideal case - so that a disturbance begun at one wire arrives at another with its sign preserved across any distance the medium spans. Two oscillators, coupled through the field, ring together whether they know of each other or not.

I had expected the theory to close a ledger on electricity. Instead it opened a door into optics, and I am uneasy in the way a man is uneasy when an argument proves too much. Still, the number stands.

A correction, then, to my own earlier position: I believed the medium was a hypothesis the equations tolerated. I now believe the equations are the hypothesis, and the medium - whatever it proves to be - must conform to them.