342 Dr. J. Kerr on a new Relation between 



or raised so that the light, after traversing the first piece, is trans- 

 mitted to the analyzer through the ripper or lower parts of the 

 second piece. Through the extended parts the light is weak- 

 ened, and, with a right degree of tension, extinguished perfectly, 

 and with a greater tension restored again ; through the com- 

 pressed parts it is always strengthened. 



Generally, when the directions of stress in the two pieces 

 are parallel, compression and tension counteract each other; 

 two compressions or two tensions reinforce each other. When 

 the lines of stress are at right angles, two compressions or 

 two tensions counteract each other ; compression and tension 

 strengthen each other. 



The action of a strained piece in the polariscope is most 

 distinct when the direction of the stress lies, as it does here, 

 midway between the principal sections of the two crossed Nicols. 

 There is no action when the direction of stress lies in either of 

 these planes. 



There are several variations of the illustrative experiment that 

 ought to be noticed here for their bearing on what follows. 

 The first piece may be stationed on the other side of the di- 

 electric ; it may even be removed altogether, and the com- 

 pression or tension applied to the dielectric plate itself in a hori- 

 zontal direction perpendicular to the beam ; the results are then 

 the same as formerly. It appears thus, without reference to 

 theory, that horizontal compression of the first piece has always 

 to be compensated in the same way (that is, by horizontal ten- 

 sion of the second piece), whether the first action is applied 

 before the dielectric, or behind it, or in it. 



The illustrative experiment takes its simplest form when di- 

 electric and neutralizing plate are both removed. It appears 

 thus that horizontal compression of a first piece has to be 

 compensated in the same way (that is, by horizontal tension of 

 a second piece), whether the other two mutually balanced pieces, 

 the unexcited dielectric and the neutralizing plate, are present 

 or absent. Accordingly, and without reference to theory, in 

 proceeding to characterize the birefringent action of the excited 

 dielectric by means of the compensating slip, I assume that the 

 mutually balanced actions of unexcited dielectric and neutralizing 

 plate are without effect, and therefore to be left out of account. 

 13. The compensating slips used in the following experi- 

 ments were of different sizes. But there was one which I 

 came at last to employ almost exclusively, as 1 found it well 

 adapted to the whole range of effects examined. It was a rect- 

 angular piece of very good plate, -fir inch thick, 2 inches wide, 

 and 10 long. It had no sensible action in the polariscope while 

 unstrained. 



