the Michelson Interference Experiment. 153- 



consequence of the motion of the earth) in the following* 

 way: — 



1°. The velocity oi:' propagation of light being c km. /sec- 

 in the direction of translation SA ; along SA' (fig. 1) form- 

 ing an angle <£ with SA, it is then 



C0 =c(l-i^sin^) ; C ;+(rsin4>)W. 



2°. The light has the same frequency in all directions- 

 from S. 



We note that v sin is the component of the velocity of 



translation perpendicular to the direction SA'. The velocity 



of light is according to this hypothesis not very different 



v 2 

 in different directions, the deviation being of the -^=10 8 



order of magnitude. It is therefore not in the least strange 



that this difference of the light's velocity of propagation has 



not been detected by direct measuring, the accuracy here 



v 

 obtained being no greater than that of - =10~ 4 order of 



c 



magnitude. The aether becomes symmetrical about an axis SA 



parallel to the direction of translation. The velocity of light 



is least perpendicular to this axis and equal to c^ = ell — ^ ^ 1 . 



It can easily be shown that the expected effect of the 

 Michelson experiment must fail to appear according to this 

 hypothesis. 



Imagine the arrangement of the experiment swung so 

 as to make the line SSi (fig. 1) form the angle (/> with the 

 direction of translation SA. During the time tf 1? which the 

 light consumes in going from S to the mirror Si, the latter 

 will have moved to S/ and the mirror S to S' and 



SS'=S! 81'=!*!. 

 The light covers the distance SSi'= Zj, where 

 y = Z 2 + tft 1 2 + 2vt 1 l cos 0. 

 According to our hypothesis the light's velocity along 

 SS X is c =c(l — f-gsin 2 ^)) and consequently* l^cU. 



* It will be shown later that taking the Telocity along SS 1 instead 

 of that along SS/ is without influence if we want an accuracy of 



the second order with respect to -. 



