350 PROCeeDINGS : PHILOSOPHICAIv SOCIETY 



cylinder with polished ends at a constant rate between crossed nicols 

 and noting the temperature at which the double refraction began 

 sensibly to diminish and also the temperature at which it vanished 

 rapidly. It was found by experiment that the lower temperature gave 

 good results as an annealing temperature and that the glass could be 

 held at the higher temperature for a reasonable time without serious 

 deformation. 



Maxwell's "relaxation time" for the internal stresses in unannealed 

 glass was determined at various temperatures by measuring the con- 

 stants of the elliptical polarization by means of a modified Stokes 

 eUiptic analyzer, 



. In measurements of the second type, the stretching and bending of 

 glass strips under load was employed, following Twyman and others. 



The results indicated that the relaxation time not only depends on 

 the temperature and the kind of glass, but also that larger stresses will 

 relax more quickly. Stretching and bending gave practically the same 

 results but the optical method gave considerably larger values, the dis- 

 crepancy varying with the kind of glass. For large and equal stresses 

 there was a better agreement. 



Twyman's empirical exponential law for the change in the relaxation 

 time with temperature holds quite closely for small temperature ranges. 

 The constant in the exponential term appears to increase with the 

 temperature and vary considerably with the glass. The most efficient 

 method of cooling seems to be such that the rate increases exponentially 

 according to the same law, as long as the stresses in the glass remain 

 at all times below the breaking stress. 



A phenomenon which may be of considerable importance in annealing 

 and other processes in glass manufacture as well, was an apparent endo- 

 thermic transformation found when careful heating curves were taken 

 by means of differential thermocouple. 



Mr. C. G. Peters has shown that this effect is accompanied by a very 

 marked increase in the thermal expansion coefficient. This would ap- 

 pear to make it advisable to carry out the annealing below the region 

 of this effect. It occurs in general, immediately following the tempera- 

 ture at which double refraction begins sensibly to diminish. It was 

 found in all the optical glasses tested as well as many others. 



Lately Mr. M. So has pubhshed similar results for some Japanese and 

 other glasses. He failed however to observe a corresponding effect 

 on cooling which had been the first indication of it observed in this 

 laboratory. 



A very plausible explanation would appear to be that the rapid heating 

 disturbs the equilibrium between the components existing in the glass, 

 which, in connection with the rapid decrease of viscosity, tends to re- 

 establish itself with corresponding rapidity causing the thermal effects 

 observed. 



Discussion: Major Wright stated that the temperature ranges cited 

 by the authors agreed very closely with those found by him. Mr. 



