2i8 SCIENCE AND METHOD. 



would be a systematic error; for, since light takes a 

 certain time, /, to travel from B to A, A's watch would 

 always be slower than B's to the extent of t. This 

 error is easily corrected, for it is sufficient to inter- 

 change the signals. A in his turn must send signals 

 to B, and after this new setting it will be B's watch 

 that will be slower than A's to the extent of t. Then 

 it will only be necessary to take the arithmetic mean 

 between the two settings. 



But this method of operating assumes that light 

 takes the same time to travel from A to B and to 

 return from B to A. This is true if the observers are 

 motionless, but it is no longer true if they are involved 

 in a common transposition, because in that case A, for 

 instance, will be meeting the light that comes from B, 

 while B is retreating from the light that comes from 

 A. Accordingly, if the observers are involved in a 

 common transposition without suspecting it, their set- 

 ting will be defective ; their watches will not show the 

 same time, but each of them will mark the local time 

 proper to the place where it is. 



The two observers will have no means of detecting 

 this, if the motionless ether can only transmit luminous 

 signals all travelling at the same velocity, and if the 

 other signals they can send are transmitted to them 

 by mediums involved with them in their transposition. 

 The phenomenon each of them observes will be either 

 early or late — it will not occur at the moment it would 

 have if there were no transposition ; but since their 

 observations are made with a watch defectively set, 

 they will not detect it, and the appearances will not 

 be altered. 



It follows from this that the compensation is easy to 



