268 DR. L. N. G. FILON ON THE 



experiments showed conclusively that for slabs of the same material Cj and C a 

 remained very approximately proportional one to the other for all the values of X 

 examined. In this way condition (6) is satisfied independently of the wave-length. 

 It follows from (5), using (6), that the relative retardation is given by 



(7). 



Accordingly the relative retardation, after the light from such a point-source has 

 traversed the two slabs, is the same for all the rays from S. Two such slabs are 

 therefore optically equivalent to a single slab which would be under an accurately 

 uniform tension. 



By properly adjusting the differences of height, z l z 2 , z 1 h, the amount of relative 

 retardation may be adjusted within certain limits. 



In general, I will be chosen small with respect to d. Thus a- is a proper fraction, 

 say of order y^. Equation (6) thus shows that M] and M 2 are to be chosen of 

 opposite signs, and approximately equal in magnitude. 



This gives at once the physical explanation of the result (6). The rays pass 

 through approximately horizontally. If we compare two rays passing through at 

 different levels, the ray which passes through the regions of greater tension in N 

 passes through the regions of lesser tension (or greater pressure) in F, and the two 

 variations balance one another. 



Further, since the amount of relative retardation as given by (7) involves only the 

 relative heights of the axes of the slabs and the source of light, the latter may be 

 moved parallel to the axes of the slabs without affecting the relative retardation. 

 Hence a horizontal line-source, parallel to the axes of the slabs, may be used instead 

 of a point-source. This was, in fact, indispensable in order to obtain the required 

 intensity. 



3. Description of the Apparatus, 



Light from an arc lamp L was passed through a condensing lens C and through a 

 thin horizontal slit T (fig. 2), which was placed from 1\ to 3 metres away from the 



Fig. 2. 



glasses and straining apparatus. It was polarized by a Nicol P, whose polarizing 

 plane was roughly at 45 to the horizontal, and then passed through the two slabs N 

 and F. These were adjusted so that their levels differed very nearly by \ centim. 



