290 



SCIENCE PROGRESS 



It will be noticed that the list is extraordinarily meagre con- 

 sidering the volume of work which has been done on kindred 

 subjects. It is true that in addition a very long list could be 

 made of the carbon compounds whose gaseous index has been 

 determined. Most of these are, however, comparatively useless 

 for the present purposes, since they consist of large complicated 

 molecules. But the following, which are among the simplest, 

 and involve only one unknown refractivity, that of carbon, may 

 be selected : 



Table IV 



Refractivities of some carbo?i compounds 



A study of these figures confirms quite definitely the con- 

 clusions of Briihl that the law, whatever it be, is not additive. 

 In not a single instance is the refractivity of the compound 

 equal to the sum of its constituents, and the divergence is in 

 most cases far greater than can be explained by errors in 

 experiment. Thus, for example, the refractivities of oxygen 

 and hydrogen are certainly known to less than one part in a 

 thousand, but the refractivity of water vapour differs from the 

 additive value by about 5 per cent. 1 Nor can it be affirmed 

 that an element has one refractive power when free and another 

 which, though different, is constant in combination. For if we 



1 It is much to be regretted that the experimental values for water differ so 

 widely ; a new determination is urgently required. 



