of the double refraction in strained glass. 



491 



If in the table last given we take the mean of all the values 

 of 1/Cp corresponding to wave-lengths lying inside intervals of 

 500 tenth-metres, we find 



\ 



from to 



Mean value 

 of (Cp)' 1 



Corrg. value of C in 

 sq. cms. per kg. wt. 



6750 



6250 



1060 



2-77 x 10- 6 



6250 



5750 



1063 



2-76 x 10- 6 



5750 



5250 



1056 



2-78 x 10- 6 



5250 



4750 



1048 



2-80 x 10- 6 



4750 



4250 



1027 



2-86 x 10- 6 



4250 





997 



2-94 x 10- 6 





It seems therefore probable that, although some hidden source 

 of error (possibly a difference in the behaviour of the glass under 

 tension and under pressure or, again, imperfect homogeneity of the 

 glass) may have partly masked this effect in the observations on 

 the lower fringes, and exaggerated it in the observations of the 

 upper ones, yet the differences obtained are significant. 



Wertheim's law is therefore not accurately true and the arti- 

 ficial double refraction due to strain does exhibit dispersion, the 

 difference in the refractive indices being smaller in the red than 

 in the violet by about 6 per cent, (taking mean values). 



7. With regard to the law of variation of the coefficient G, 

 the present observations are too rough to allow it to be deter- 

 mined : the table of mean values given above, however, would 

 seem to indicate that G varies far more rapidly towards the violet 

 than towards the red. But, having regard to the probable error 

 of the observations, no definite conclusion can be drawn. 



Since the above work was done, Herr F. Pock els has published 

 an account (see Wied. Ann. 1902, Ser. iv. Vol. vn. p. 745, " Ueber 

 die Aenderung des optischen Verhaltens verschiedener Glaser 

 durch elastische Deformation") of some extremely interesting ex- 

 periments on various glasses under compression. He has used 

 light of three kinds, namely the Bunsen name coloured with Na, 

 Li and Tl-salt. 



For heavy glasses Pockels finds a deviation from Wertheim's 

 law of the same kind as the one indicated here, i.e. the coefficient 

 C is greater (numerically) for the green and yellow rays that) for 

 the red. For lighter glasses this dispersion of double refraction 

 appears to be insensible. 



