On the other hand, the aviator who flies eastward will never reach his destination, because in each second of time he is driven westward to precisely the same extent as he is travelling eastward. He will remain stationary. To accomplish his journey he would need to cover in the wind an infinite distance.
If, instead of imagining a wind equal in velocity to the aviator (an extreme supposition in order to make the demonstration clearer), I had thought of it as less rapid, we should again find, by a very simple calculation, that the man who flies north and south has less distance to cover in the wind than the man who flies east and west.
Now take rays of light instead of aviators, the ether instead of the wind, and we have very nearly the conditions of the Michelson experiment. A current or wind of ether—since the ether has been already shown to be stationary in relation to the earth’s movement—proceeds from one to the other of our east-west mirrors. Therefore the ray of light which travels between these two mirrors, forth and back, must cover a longer distance in ether than the ray which goes from the south mirror to the north and back. But how are we to detect this difference? It is certainly very minute, because the speed of the earth is ten thousand times less than the velocity of light.
There is a very simple means of doing this: one of those ingenious devices which physicists love, a differential device so elegant and precise that we have entire confidence in the result.
Let us suppose that our four mirrors are fixed rigidly in a sort of square frame, something like those “wheels of fortune” with numbers on them that one sees in country fairs. Let us suppose that we can turn this frame round as we wish, without jerking or displacing it, which is not difficult if it floats in a bath of mercury. I then take a lens and observe the permanent interference-fringes which define the difference between the paths traversed by my two rays of light, north-south and east-west. Then, without losing sight of the bands or fringes, I turn the frame round a quarter of a circle. Owing to this rotation the mirrors which were east-west now become north-south, and vice versa. The double journey made by the north-south ray of light has now taken the direction east-west, and has therefore suddenly been lengthened; the double journey of the east-west ray has become north-south, and has been suddenly shortened. The interference-fringes, which indicate the difference in length between the two paths, which has suddenly changed, must necessarily be displaced, and that, as we can calculate, to no slight extent.
Well, we find no change whatever! The fringes remain unaltered. They are as stationary as stumps of trees. It is bewildering, one would almost say revolting, because the delicacy of the apparatus is such that, even if the earth moved through the ether at a rate of only three kilometres a second (or ten times less than its actual velocity), the displacement of the fringes would be sufficient to indicate the speed.
When the negative result of this experiment was announced, there was something like consternation amongst the physicists of the world. Since the ether was not borne along by the earth, as observation had established, how could it possibly behave as if it did share the earth’s motion? It was a Chinese puzzle. More than one venerable grey head was in despair over it.
It was absolutely necessary to find a way out of this inexplicable contradiction, to end this paradoxical mockery which the facts seemed to oppose to the most rigorous results of calculation. This the men of science succeeded in doing. How? By the method which is generally used in such circumstances—by means of supplementary hypotheses. Hypotheses in science are a kind of soft cement which hardens rapidly in the open air, thus enabling us to join together the separate blocks of the structure, and to fill up the breaches made in the wall by projectiles, with artificial stuff which the superficial observer presently mistakes for stone. It is because hypotheses are something like that in science that the best scientific theories are those which include least hypotheses.
But I am wrong in using the plural in this connection. In the end it was found that one single hypothesis conveniently explained the negative result of the Michelson experiment. That is, by the way, a rare and remarkable experience. Hypotheses usually spring up like mushrooms in every dark corner of science. You get a score of them to explain the slightest obscurity.