1910.] TROWBRIDGE— THE ETHER DRIFT. 55 



through space of the point of the earth's surface where the appa- 

 ratus is set up, and suppose the relative positions of the three points 

 be such that each thermometer receives the same amount of radia- 

 tion from the source of hght which hes between them. Now, if 

 the ether be drifting across the apparatus in the direction joining 

 the three points as the theory of Lorentz would have it drift, and 

 if thereby the dimensions of the apparatus in this direction be 

 shortened as Lorentz and Fitzgerald have supposed it to be short- 

 ened, tlicfi if the whole apparatus be rotated i8o° about a vertical 

 axis the distance between the three points cannot have altered so 

 that if any change in the amount of radiation received by the two 

 thermometers were to be noticed on rotating the apparatus through 

 1 80° it might be taken as proof positive that the ether was in motion 

 relative to the apparatus. On the other hand, if the apparatus were 

 sufficiently sensitive to detect a second order optical effect and no 

 change took place on rotation, the objection could not be made that 

 the eft"ect was there but had been masked by shortening, since no 

 hypothetical shortening could be conceived to alter the relative 

 positions of three points on a straight line. 



Thus the problem which confronted us was to devise apparatus 

 capable of detecting a change in radiation received by its two parts 

 so small as one one-hundred millionth part of this radiation — 

 expressed differently it was to devise a pair of electrical ther- 

 mometers capable of detecting a dift'erence of temperature of less 

 than the ten-millionth part of a degree and yet to so protect them 

 that in spite of the fact that they must stand several feet apart 

 that they should be subject to no irregular fluctuation in temperature 

 of this order of magnitude. Also they must be mounted on a support 

 so rigid that it may be rotated without introducing irregularities 

 due to change of shape. After considerable trouble we have suc- 

 ceeded in satisfying both these conditions, but there remains a 

 third more difficult condition which we have as yet not been able 

 to wholly satisfy — this is that the two conditions just mentioned 

 must remain satisfied when the thermometers are subjected to 

 radiation from a light source standing between them which is at a 



