Birefringent Action of Strained Glass. 325 



of the neutral band in the strained plate, with nearly half 

 an inch of screen between it and each of the glass margins. 

 The screw-press being clamped firmly to a heavy vertical 

 stand, the slit-screen is brought into the course of the two 

 pencils of the refractor, as shown in section at F in the 

 diagram of art. 2 : the refractor also is supposed to be so 

 arranged, that a rise of the fringes indicates a relative retar- 

 dation of the pencil V. Things being thus prepared, the 

 experimental proof of the proposition extends through the 

 four following cases, which are got by simple displacements 

 of the plate-stand. 



(1) V through compressed margin, H through slit ; 

 strain raises the fringes ; V is retarded relatively, and there- 

 fore absolutely. 



(2) V through extended margin, H through slit; strain 

 lowers the fringes ; V is accelerated relatively, and there- 

 fore absolutely. 



(3) H through compressed margin, V through slit ; 

 strain lowers the fringes ; V is relatively accelerated, and H 

 therefore is absolutely retarded. 



(4) H through extended margin, V through slit ; strain 

 raises the fringes ; V is relatively retarded, and H therefore 

 is absolutely accelerated. 



I have reasoned here as if the strains in the margins of the 

 plate were pure vertical compression and pure vertical tension 

 respectively, and also as if there were no strain behind the 

 slit, in the neutral part of the plate. It is certain that things 

 are not exactly thus ; and objection may therefore be made 

 to the proof as wanting in rigour. If any doubts remain 

 on this account as to the truth of the proposition, they will 

 disappear very soon. 



The results of these experiments are quite regular. When 

 a good plate is put in position unstrained, and the central 

 fringe is brought on the wire of reference, a quarter turn of 

 either screw of the press gives a clear displacement of the 

 fringes, always in the right direction of rise or fall according 

 to the proposition ; and as the strain is gradually increased, 

 the displacement of the fringes also increases regularly, 

 through one, two, three, or more fringe-widths, till the plate 

 breaks. The effects are equally clear with different kinds of 

 . and with different thicknesses. 



In the course of a few repetitions of the experiments, there 

 was another important result that presented itself : the four 

 preceding cases fell clearly into two sets — (1) and (2) as cases 

 of strong action, (3) and (4) of weak action ; but this de- 

 serves a detached and formal statement. 



