Birefringent Action of Strained Glass. 337 



arm gently along the outer lateral surface of the plate for five 

 seconds, and then withdraws it. Moderate though the heating 

 is, the fringes are largely affected by it, being rapidly displaced 

 and inclined, and before the end of the five seconds deformed 

 even to extinction. In 1 to 1J seconds they reappear dis- 

 tinctly, but so placed as to indicate a large relative retardation 

 of the pencil which is next the heated surface, and so inclined 

 as to indicate an extremely rapid decrease of the optically 

 effective strain from the heated surface inwards. As the 

 observation is continued, the fringes fall back to their initial 

 positions and directions, quickly at first, then more and more 

 slowly, for 4 or 4^ minutes. When the experiment is repeated 

 with a time of heating as short as 2 seconds, or even 1 second, 

 the effects are distinct enough though less intense, the fringes 

 being well displaced and inclined, without extinction or great 

 deformation. This experiment was often repeated, and the 

 effects were perfectly regular. 



When the plate, as it stands, is examined between crossed 

 Kicols, the effect of the regular heating is very faint, being 

 easily neutralized by a hand-compensator of thin glass. Direc- 

 tional strain across the rays is therefore out of account in the 

 principal experiment, which is found indeed to give sensibly 

 the same quantity of effect whether V or H is next the heated 

 surface. 



The birefringent action of a glass plate, strained by change 

 of temperature along one edge, and traversed by light per- 

 pendicular to the faces, was examined carefully by Sir David 

 Brewster, and the phenomena were fully discussed afterwards 

 by Sir John Herschel*. From their results it is evident, that 

 the predominant strain in our heated plate is compression 

 along the length at the heated edge and at the opposite edge, 

 with a diffuse tension along the length in the intermediate 

 parts of the plate : and this is proved easily as follows. 



When the plate is placed between crossed Nicols, with its 

 length vertical, and its faces perpendicular to the rays, the 

 effect of the regular heating indicates pure vertical compres- 

 sion at the heated surface, falling off rapidly inwards, and 

 disappearing at J to ^ inch from the surface. Inside of this, 

 about J inch from the surface, which is the place of the inner 

 pencil in the principal experiment, there is generally an effect 

 of the contrary kind, but so faint as to be clearly out of 

 account, in comparison with that close to the heated surface. 

 When the plate is in the same position, and the refractor is 

 employed instead of the polariscope, the effect of the regular 



* " Treatise on Light,"' Encvcl. Metrop., art. 1097. 

 PUL Mag. S. 5. Vol. 26. No. 161. Oct. 1888. Z 



