296 SCIENCE PROGRESS 



his values for the dispersion of oxygen and nitrogen are both 

 greater than that of air, so that one of them must be wrong. 

 Dufet, in his collection of optical data, records no attempt to 

 measure the dispersion of chlorine or bromine and only one 

 rough observation on iodine. It is very much to be regretted 

 that these comparatively easy experiments should not be 

 performed. The dispersion of mercury, sulphur, and phos- 

 phorus has been measured recently, and in these results we 

 find the first indications of what may, if confirmed, turn out 

 to be a simple relation. The dispersive power of sulphur is 

 found to be about four times that of oxygen, and the dispersive 

 power of phosphorus almost exactly twice that of nitrogen. 

 But the most promising results have been obtained from the 

 inert gases. It is found that, in the four cases in which the 

 dispersions have been measured, if the refractivity for infinite 

 waves is plotted against the value of b in Cauchy's formula the 

 result is a straight line. If this rule should be confirmed in the 

 case of the fifth gas, neon, it may perhaps assist mathematicians 

 in the selection of the hypotheses connected with their theories 

 of dispersion. 



There is one further result which is worthy of the attention 

 of mathematical physicists, and is, I believe, the only one of its 

 class at present. The dispersion of S0 2 has been measured by 

 Ketteler, and can be compared with that of sulphur and oxygen. 

 If Ketteler's figures are reduced to the standard density, the 

 three formulae would then stand thus : 



For one atom of sulphur — 



u - i = "000523 ( 1 + r — — ) Cuthbertson and Metcalfe. 



For two atoms of oxygen — 



H - 1 = "0002663 ( 1 + ^ 0/ ll j Scheel. 



For one molecule of S0 2 — 



it — 1 = '000642 (1 + .., , ) Ketteler corrected. 



The figure g'6 is the mean of the two values 91 and io"i 

 obtained from Ketteler's three values, and is consequently not to 

 be trusted to 10 per cent. Roughly speaking the dispersive 

 power of S0 2 is half that of sulphur and double that of oxygen. 



It is much to be desired that similar figures should be 

 obtained for HC1, HBr, HI, FLS, and many other compounds. 



