Birefringent Action of Strained Glass. 329 



art. 4, and had set so well into its bearings by usage, that 

 any moderate strain gave equal optical effects in its total rise 

 and total foil. Beginning with V through either margin and 



t5 O O ill 



H through the slit, the compensator at zero, and the central 

 fringe on the wire of reference, the plate was then moderately 

 strained, and the displaced fringe was brought back to its 

 first position by the compensator, which was then read. The 

 places of V and H in margin and slit were then interchanged 

 by a simple displacement of the plate-stand, the compensator 

 was again brought to zero, and the central fringe to the wire 

 of reference ; the plate was then carefully unstrained, and 

 the displaced fringe brought back as formerly to its first 

 position by the compensator, which was again read. The 

 two readings were taken as measures of the two strain- 

 generated retardations. 



Of a number of measurements taken in this way, I give 

 here the last eight, the numbers in successive columns being 

 the pairs of readings for the successive strains, left as they 

 were found, in vernier-units (about 84 to the wave-length). 

 The strain was tension in the first four columns, compression 

 in the others. 



103 



69 



39 



74 



75 



92 



55 



85 



49 



34 



21 



37 



38 



45 



27 



43 



Approximately, therefore, the greater retardation is twice the 

 less. 



9. Measurements of a more accurate kind, but through a 

 rather small range, were taken next with the winged plate 

 and refractor, checked by a J amines quartz compensator. 

 The plate was tested beforehand by the quartz compensator, 

 and was found to be so well fitted to its bearings, that a 

 given strain of moderate intensity could be recovered per- 

 fectly, a good number of times in succession : the plate was 

 then placed on a suitable stand, in the course of the pencils 

 of the refractor. 



(1) V passing through the middle of the central pillar, 

 and H through a wing, the given strain (vertical compression) 

 was applied, and its effect was neutralized by the glass com- 

 pensator, which was then read. The plate was then un- 

 strained, and the reverse effect was compensated and the 

 compensator again read. If the two readings differed in 

 arithmetical value, their average was taken as the measure 

 of V, the strain-generated retardation of the pencil V. 



(2) The plate-stand was moved across the pencils, till H 

 took the former place of V in the winged plate ; the opera- 

 tions (1) were repeated, and the average of the two new 



