348 Mi; <'1JYF. CUTHBEBT8ON OX TIIF. 



These two values are widely different, but lx>th are Broiler than the highest 

 refractivity which it is ever neceasary to attribute to carlxm. The inference is, either 

 that the index of boron is actually higher than that of carbon, or, that the presence 

 of boron has the effect of increasing the retarding power of OH. and Br in an abnormal 

 degree. It is difficult to say which of these alternatives is the more interesting ; 

 and the determination of the indices of B01 :t , BBr :t , and BH 3 , in the gaseous state, 

 ought to yield results of the highest importance. 



This rule, that increased valency means increased refractivity, is no doubt closely 

 connected with GLADSTONE'S observation that, when an element, such as iron, forms 

 two compounds, that in which it has the highest valency has the highest refracting 

 power, and the allied fact that a double or treble Ixmd is accompanied by increased 

 refracting power in the compound in which it occurs. The problem is one of great 

 interest and is yet unsolved. 



Beyond the bare fact that the refractivity does increase with the valency it is 

 difficult to go at present. I have tried, but in vain, to guess the law which connects 

 the refractivities of A, 01, S, and P (139, 192, 270, 297). It is easy to find specious 

 relations between two or three members of the series, but, until we know the indices 

 of the alkali elements, and can make a nearer approximation to the true values for 

 carbon, boron, and silicon, the data are insufficient, and speculation is useless. 



Nor have I been more fortunate in my endeavours to find a meaning for the 

 numbers, 1, 2, 8, 12, and 20, which express the ratios between the refractivities of 

 the different members of the inert family of gases. 



CLAUSIUS has shown that, on the assumptions of the electromagnetic theory of 

 light, if the molecules of gaseous matter are perfect spheres, conducting electricity, in 

 a non-conducting medium, the refractivity of the gas is proportional to the fraction of 

 space occupied by the molecules. And if, for molecules, we put electrons, it would 

 follow that the sum of the volumes of the electrons (or possibly their number), in 

 similar elements, are in the ratios of small integers. But this theory throws no light 

 on the peculiar numbers actually observed. 



Indices of Compound?.. 



From the tables given above, it will l>e seen that the determination of the indices 

 of S and P enables us to add 15 compounds to the slender list of those in which the 

 refractive indices of all the components in the free state are known. Of these, 1 1 , 

 via., PBr 3 , P 4 S, PA, POC1 3 , PSC1 S , SOC1.,, SCI* S 2 C1 2 , S(XC1 2 , H 2 S0 4 , SO.HC1, have 

 had their indices measured in the liquid state ; and 4, viz., SO 2 , H 2 S, PC1 3 , and PH ;! , 

 in the state of gas. It might have been expected that this addition to the available 

 d-ata would have afforded some clue to the relation of the indices of compounds to 

 those of their constituents, but I have been unable to find any ; and I doubt if the 



. T , 



