CLASSICAL AND NEW MECHANICS 177 



bodies and is subject to experimental verification; sec- 

 ondly, when it passes through space absolutely de- 

 prived of material bodies and so is not subject to 

 experimental evidence. 



In the first case, we know that there is a true path 

 and that light moves very approximately in straight 

 lines, and we have quite accurately measured the time 

 light requires to move from one place to another 

 through various substances. V is here the length of 

 path divided by the time, a true velocity. We know 

 that this V is a variable ; it is less in water than it is in 

 air, and still less in glass. It is also less in dense air 

 than in rarefied air. Not only does V vary with the 

 kind of matter through which light passes, but it also 

 depends on the motion of the medium as shown in the 

 experiments on the velocity of light passing through 

 columns of moving water made by Fizeau and repeated 

 by Professor Michelson. The velocity of light in 

 material media is therefore subject to all the variations 

 which influence the velocity of sound and other types 

 of motion and evidently cannot be the V assumed by 

 Professor Einstein to be an absolute constant. 



He must then limit his postulate strictly to what is 

 called the velocity of light in absolutely immaterial 

 space. There are two methods of obtaining this value. 

 We observe the difference in time between the cal- 

 culated eclipse of some satellite and the recorded ob- 

 servation of the event or we use other stellar phe- 



