Mr. W. Sutherland on Molecular Refraction. 153 



modifying constant c is specially large. The general con- 

 clusion, then, is that the modified Gladstone equation is capable 

 of representing the relation of index of refraction and density 

 under all circumstances within the Kmits of experimental 

 error. We see how the values of m{n — l)/d, tabulated by 

 various experimenters, for various liquids ought to be reduced 

 to the values corresponding to zero density ; or in all cases 

 the observations ought to be made on bodies in the state of 

 vapour, if the results are to be as trustworthy as possible to 

 the chemist. 



The advantage that the Lorenz formula possesses is that 

 it happens to give the same value for a substance in the 

 liquid and vapour states ; and since for vapours 



{n'-l)/{n^ + 2)d = ^{n-l)ld, 



we see that the Lorenz formula gives approximately, when 

 applied to a liquid, two thirds of the desideratum, namely the 

 value of (n — l)/d determined on the vapour of the liquid. It 

 is on this account I believe that the Lorenz expression has 

 been found by Briihl to prove more accurate than Gladstone's 

 in its chemical applications ; in an indirect manner the 

 Lorenz formula applied to liquids gives a measure of the 

 constant term in the modified Gladstone equation. 



The most important fact that the study of refraction from 

 a chemical point of view has brought out is that the refrac- 

 tion-equivalent of an atom of an element is not constant but 

 depends on its method of binding with others. Now, accord- 

 ing to the theory of our first approximation, the refraction- 

 equivalent of an atom ought not to vary unless an actual 

 change in the physical structure of the atom has been pro- 

 duced ; the second approximation introduces a constant b 

 which may depend on structure, but there is evidence which 

 would seem to indicate that b is zero. Thus the following 

 table contains the values of the Lorenz molecular refraction, 

 which I have calculated from the data given by Bartoli and 

 Stracciati (Ann. de Ch. et de Ph. 6 ser. vii. 1886) in their 

 valuable table of physical constants for the liquid paraffins ; I 

 employ the Lorenz expression for the reason given above. I 

 have added the value for CH4 calculated from Mascart's data 

 for gases {Comptes Rendus , Ixxxvi.). The density at 0°, given 

 by Bartoli and Stracciati for pentane (iso), appears to be a 

 misprint, and I have taken '6368 as the density at 16°. 



