342 MR, CLIVE CUTHRKRTSON ON TTIE 



Table IX. gives the details of the experiments. Having regard to the smallness of 

 their number, I consider that the error in this determination may equal, though it 

 probably does not exceed, l per cent. 



If, as in the case of P and Hg, we try to verify this value by a comparison with 

 the indices of compounds of sulphur and of its refraction equivalent, we are met by 

 anomalies which show that no single value of the refractivity of sulphur will satisfy 

 all cases. 



Table X. gives the indices and refractivities of the compounds of sulphur, the 

 refractivities of all of whose constituents are known. It will be seen that the 

 refractivities which must be assigned to S, in order to satisfy the additive law, 

 vary widely from the value of 551 here given for S,. In SO 2 , H 2 S, SO 2 CL, and 

 notably in H 2 S0 4 and S0 3 HC1, a smaller value would best accord with the facts. In 

 SOCL, PSC1 3 , P 4 S, the value 551 would fit in excellently. In SC1 2 , S 2 C1, a value 

 slightly larger, and in CS 2 , whose refractivity is 1476, a much larger value is 

 required ; for the highest contribution we can expect from C. is 250, which would 

 leave 610 to be accounted for by each of the sulphur atoms. Finally, the smaller of 

 the two refraction equivalents calculated by GLADSTONE for S (14'1), when compared 

 with that of O (2 '8), would point to a value as high as 680. 



All that can safely be concluded from these highly interesting anomalies is that 

 the refractive power of sulphur varies greatly, according to the nature of the 

 compound which it forms, and that the value now found is not far from the mean 

 round which the discrepant values are grouped. 



In the periodic table of the elements, oxygen occupies, with regard to sulphur, the 

 same relative position that nitrogen does with regard to phosphorus, and argon to 

 neon. And, if we compare the refractive powers of the first pair, it is astonishing to 

 find that their ratio is, within the limits of experimental error, the same as that found 

 for the other two. 



The most recent determinations of the index of oxygen for sodium light are 

 1-002702 (MASCART), 1 '0002716 (LOBENZ), T0002702 (RAMSAY and TRAVERS). 



If we accept the value in which two of these researches concur, and compare it 

 with the value 1 '00 11 01 now given for S 2 , it appears that the ratio of their refrac- 

 tivities is 4'08 against 4 '03 found for N and P and 4'13 for Ne and A. The 

 agreement is 1 per cent, worse in the present case than in that of P, perhaps owing 

 to the fewness of the experiments, or possibly to the wide dispersive power of 

 sulphur. But the concordance is too close to be accidental ; and it may therefore be 

 said that, at least to a first approximation, an atom of sulphur retards light four 

 times as much as an atom of oxygen. 



