346 



Ml;. GLIVK CUTHBEBT8ON ON THK 



TABLE XL 



The most obvious fact shown by this table is oue which the study of the refraction 

 equivalents had long ago made familiar, i.e., that in each family of elements the 

 refractivity increases with the atomic weight, though not in simple proportion to it. 

 It is true that the existing figures for the refraction equivalents show many excep- 

 tions to this rule ; but the figures themselves are so untrustworthy that anomalies do 

 not destroy the general effect of the evidence. 



The second, and perhaps more interesting, point is that the refractivities of the 

 different members of the same chemical family ;ire, within the limits of experimental 

 error, related in the ratios of small integers. This fact, observed first in the family 

 of the inert gases, is equally true in that of the halogens, and, if the present results 

 are approximately correct, is now shown to exist between the two first members of 

 the families of O and N respectively. 



The case then stands thus : The indices of fourteen elements have been measured 

 in the gaseous state.* Of these, two, mercury and hydrogen, are not related to any 

 others whose indices have been determined. The remaining twelve all conform to 

 the rule.f The probability that these coincidences are fortuitous is small. But if 

 they are not, the extreme simplicity of the relations raises the hope that some 

 fundamental property of matter may be involved. 



If we attempt to go beyond the elements whose indices have been determined in 

 the free state and enquire how far the refraction equivalents lend support to the 



* I do not reckon arsenic, since M. LE Eoux himself did not place much reliance on his result, and the 

 present enquiry renders it still more dubious. 



t It is a strange fact that the index of IL> is practically identical with that of Ne-, but whether this is 

 more than a pure chance it would l>e hazardous to conjecture. 



