DISPERSION IN ARTIFICIAL DOUBLE REFRACTION. 



285 



TABLE II. Observations of S 57 (re-reduced). 



Physik,' 1902, p. 745) that for a glass containing l>etween 60 and 70 per cent, of 

 Pb( ) the stress-optical coefficient changes sign, and an experiment made by him with 

 such a glass pointed to the fact that the stress- optical coefficient did not vanish 

 simultaneously for all colours, a result which has been independently confirmed by 

 the present author from considerations of curves showing C and AC'/ AX plotted to 

 percentage of lead (see ' Camb. Phil. Soc. Proc.,' vol. XXI., p. 335). Now if the law 

 C = C /(l X,,/X) held universally, the vanishing of C u would imply the vanishing of C 

 for every wave-length. 



Moreover, it seems impossible to find theoretical justification for such a formula. 

 It is well known (see DRUDE'S ' Theory of Optics,' cap. V., pp. 388, 389)* that the 

 index of refraction p. is given by the formula p. 1 = l + SA^/jl (X^/X)*}, where 

 X,, = wave-length in. racuo of light belonging to one of the natural periods of the glass. 



[*Nolt atkM July 3rd, 1907. Throughout the paper I have followed DRUDK. But if we adopt 

 LORENTZ'S formula, viz. : 



or similar formula 1 , the most essential part of the reasoning remains in most cases practically unaltered.] 



