•336 Mr Filon, On the variation with the Wave-length, etc. 



In the above, to get C in sq. cms. per kilogramme weight, 

 multiply by 10 7 . 



These results, with the exception of O 428 and 658, are 



shown in Fig. 7. There are one or two points which are irregular, 



for example those due to O 192 and O 2154. These indicate an 



AC 

 irregularity both in the curve of C and in that of — , so that 



A A/ 



they are probably explained by the presence of some disturbing 

 constituent. 



In the case of 2154 this may be due to the fact that this 

 glass contains a small percentage (1*5) of B. 2 3 , and the presence 

 of this substance appears to have a very decided effect upon both 

 the stress-optical coefficient and its variation .with the wave- 

 length. Thus both 428 and O 658, which are borate glasses, 

 show a comparatively high C. A glass, S 205, containing 69 per 

 cent, of B„0 3 , examined by Pockels, gives C = 4*33 for Na-light, 



and -— =--000084. The other two borate glasses 428 

 AA, J Tl 



AC 

 and 658 show a positive -^r with a high positive C, quite an 



exceptional result. 



The present series of experiments, however, was intended 

 solely to deal with lead glasses and the consideration of borate 

 glasses may here be postponed. 



In conclusion, these experiments confirm, in every way, the 

 results obtained by Pockels for the effect of lead on artificial 

 double-refraction produced by stress for a given wave-length. 

 They indicate further that the difference of the refractive indices 

 (to which C is proportional) is greater in the violet than in the 

 red for those lighter glasses which are not borates. That as the 

 percentage of lead increases, this variation in the difference of 

 the refractive indices decreases : it is probably zero, or Wertheim's 

 law is obeyed, for a glass containing about 66 per cent, of PbO. 

 After that, the sign of the variation is altered and it begins to 

 increase rapidly numerically. For glasses nearly on the limit of 

 high density C and its rate of change with the wave-length 

 increase with great rapidity (the initial sign of both, however, 

 being now changed). For such glasses the variation of C is 

 parabolic, the difference of the refractive indices increasing much 

 more rapidly towards the violet than it decreases towards the red. 

 There are indications of a similar phenomenon in some of the 

 other glasses, but in many cases, the variation appears, at all 

 events to the first approximation, to follow a linear law. 



I wish to renew my thanks to Professor Trouton, F.R.S., and 

 to Assistant-Professor Porter, of University College, London, for 

 their continued courtesy and kindness during the progress of these 

 experiments, which were carried out in the Physical Laboratory of 

 University College, London. 



