( 44fi ) 

 this value would be 0,0055 c. Thus we are led to the conclusion 



(C 



that — cannot be much different from 11, so that the least admis- 



sible value of the condensation is nearly n = e^^. 



Calculations which we shall omit here may serve to estimate the 

 error that has been committed in simplifying- the equation (2). It 

 is found that far away from the earth the error may become rather 

 large, but that nearer the surface, precisely on account ofthesmall- 

 ness of the velocity in these parts, we need not trouble ourselves 

 about it. Thus, what has been said about the condensation may 

 be true, even though the state of n.otion in the rarefied aether, at 

 great distances, depart widely fi'om the equation (4). 



Strictly speaking, the condensation must be still more considerable 

 than the value we have found to be necessary. If the aether be 

 attracted by the eaith, it is natural to suppose that it is acted on 

 likewise by the sun ; thus, the earth will describe its orbit in a 

 space in which the aether is already condensed. In this dense aether 

 the earth must produce a new condensation. 



Of course it is not necessary that the attraction follow precisely 

 the law of inverse squares; any law which leads to a sufficient 

 condensation will suffice for our purpose. To understand the con- 

 nexion between the condensation and the velocity of sliding, we 

 may consider a simple case. Let the aether have a constant small 

 density k outside a certain sphere, concentric with the earth, and 

 within this sphere a constant density k' > k. 



If now the earth were at rest, and the aether flowed along it, 

 a diametral plane of the sphere, perpendicular to the mean direction 

 of flow, would be traversed by a quantity of aether, equal to that 

 which enters the sphere on one side and leaves it on the other 

 side. If this shall be the case, the velocities inside the sphere must 



k 



be of the order — c , if outside the surface they are of the order c 



If we wish to maintain the theory of Prof. Stokes by the sup- 

 position of a condensation in the neighbourhood of the earth, it 

 will be necessary to add a second hypothesis, namely that the velo- 

 city of light be the same in the highly condensed and in the not 

 condensed aether. This is the theory that may be opposed to that 

 of Feesnel, according to which the aether has no motion at all. 

 In comparing the two we should, I believe, pay attention to the 

 following points. 



1. The latter theory can only serve its purpose if we introduce 

 the well known coefficient of Fresnel, concerning the propagation 



