280 



DR. L. N. G. FILON ON THE 



of the glasses, the latter were analysed for me by Mr. W. J. REES, on the staff of 

 Messrs. CHANCE BEOS. To Mr. REES' skill I am indebted for the following results : 



The majority of these glasses belong to the borosilicate variety, excepting 2783. 

 2783 is a flint glass, and was stated by the makers to be identical in composition 

 with another glass, O 154, the composition of which (see ' Camb. Phil. Soc. Proc.,' 

 vol. XH., Part V., p. 314) was stated by Messrs. ZEISS to include Na 2 O and BaO. 

 It seems probable that the composition of the later glass is a little different to that 

 of O 154. 



15. Linear Law connecting X and the Stress. 



Since it was known beforehand that corrections to W of the type W + KW 2 would 

 have to be applied, W being due to the imperfect annealing, and KW 2 to relative 

 and absolute rise and fall (see 6, 13), instead of calculating the stress-optical 

 coefficient C directly, as was done in previous experiments, the relation between W 

 and X was first studied, with a view to disengaging the corrections. 



In practice, readings for W and X were taken for both first and second orders 

 of the band, and even, where possible, for third orders for both tension and pressure. 

 The tension and pressure observations were obtained by altering the relative heights 

 of the two slabs by interchanging two steel slips which raised the supports of one 

 of the slabs. The bending moments were not altered. 



A typical set of results is embodied in Table I. below. 



