336 Dr. J. Kerr's Experiments on the 



vertical ; then, in the notation of art. 12, the strain-generated 

 relative retardation of the component ray H 



= \e8m+ {m-l)Se\ - \e8m + (m-l)8e\ 



= —e8m. 



Thus, in glass vertically compressed, the relative acceleration 

 of the component ray H is produced exclusively by a decrease 

 of velocity of the component ray V; and similarly, with 

 change of sign, in the case of vertical tension. This com- 

 pletes the solution of prop. III. 



Proposition X. 



In the glass plate of prop. VII., strained homogeneously in 

 one direction parallel to its faces, the strain-generated variations 

 of the thickness (e), and of the index (m) for light incident 

 normally, and polarized in a plane parallel to the line of strain, 

 are such as to satisfy approximately the equation 



Be Sm 



m 



r 



For the equality of eSm and (m — l)he has been proved by 

 experiment. 



Proposition XI. 



To find the effect of directional compression or tension in 

 glass, upon light passing along the line of strain. 



15. I have made several attempts on this part of the sub- 

 ject at different times, and I begin here with the latest. The 

 particular plate employed in the experiment is a rectangular 

 parallelopiped of light optical flint (of Chance's), \ inch thick, 

 6 inches long, and 2% wide, its two ends (2-J by £) very well 

 planed and polished, so as to give good vision through the 

 glass, along the length of the plate. This piece is fixed to a 

 suitable stand by one of its edges, with its faces horizontal, 

 so that the two pencils of the refractor (from a vertical and 

 very narrow lamp-slit) traverse the two ends at right angles, 

 one of the pencils passing as close as possible to the outer 

 lateral surface (6 by J) through the whole of its length. In 

 these trying circumstances, the fringes are obtained almost as 

 well through the glass as through air, the plate being a singu- 

 larly good one, the glass homogeneous and isotropic, and the 

 surfaces of incidence and emergence accurately plane and 

 parallel. 



To strain the glass moderately, and in a manner sufficiently 

 regular for a qualitative experiment, an assistant lays his bare 



