The reason for the failure of any such experiment anywhere conducted can best be made plain by a crude paraphrase of a classic proposition from Relativity.  Suppose it is required to determine the same moment of time at two different places on the earth’s surface, as must be attempted in finding their difference in longitude.  Take the Observatory at Greenwich for one place, and the observatory at Washington for the other.  At the moment the sun is on the meridian of Greenwich, the exact time of crossing is noted and cabled to Washington.  The chronometer at Washington is set accordingly, and the time checked back to Greenwich.  This message arrives two seconds, say, after the original message was sent.  Washington is at once notified of this double transmission interval.  On the assumption that HALF of it represents the time the message took to travel from east to west, and the other half the time from west to east again, the Washington chronometer is set one second ahead of the signalled time, to compensate for its part of the loss.  When the sun has reached the meridian of Washington, the whole process is repeated, and again as before, half of the time the message has taken to cross and recross the Atlantic is added to the Greenwich record of noon at Washington.  The number of hours, minutes, seconds, and fractions of a second between these two corrected records represents the difference in solar time between the two places, and incidentally the same moment of time has been established for both—­at least, so it would appear.

But is it established?  That each message took an equal time to travel each way is pure assumption, and happens to be a false one.  The accuracy of the result is vitiated by a condition of things to which the Relativists have called attention.  Our determination might be defended if Washington and Greenwich could be assumed to remain at rest during the experiments, and some argument might even be made in its favor if we could secure any cosmic assurance that the resultant motion of the earth should be the same when Greenwich signalled its noon to Washington and Washington its noon to Greenwich.

Our present discussion is merely illustrative, or diagrammatic; so we will neglect the velocity of the earth in its orbit round the sun, some forty times greater than that of a cannon ball, and the more uncertain and more vertiginous speed of the whole solar system towards its unknown goal.  Let us consider only the rotation of the earth on its axis, the tide-speed of day and night.  To fix our idea, this may be taken, in our latitudes, at eighteen thousand miles per day, or perhaps half the speed of a Mauser rifle bullet.

So fast, then, will Washington have been moving to meet the message from Greenwich.  So fast will Greenwich have been retreating from Washington’s message.