318 Mr. W. Sutherland on the 



actual value of H may be determined ; but for the present, 

 as a general result, we regard H as negligible. 



If we regard the number n as measuring the actual volume 

 of the molecule, and if we consider molecules to have similar 

 shapes, then n^ measures the surface of the molecule; and we 

 can say that the parameter A varies inversely as the surface 

 of the molecule. This interpretation is partially justified by 

 the consideration of certain results in molecular refraction. 

 Let i be the index of refraction of a substance whose molecule 

 occupies a domain u, but actually fills only a volume U, and 

 let I be the index of refraction of the matter of the molecule, 

 then I have shown in my paper on Molecular Refraction that 

 (2 — l)i< = (I — 1)U, which is Grladstone's law. Now Landolt, 

 Gladstone, and Briihl have shown that in general {i — l)u 

 for a molecule is the sum of certain quantities known as the 

 refraction-equivalents characteristic of the different atoms, 

 but that there are pronounced exceptions : thus, when an atom 

 possesses two valencies, it has two corresponding refraction- 

 equivalents. The same results hold therefore for (I — 1)U; 

 thus, if an atom has different effective valencies, it contributes 

 different amounts to the value of (I — 1)U in a molecule ac- 

 cording to its valency; thus either I or U, or both I and U, 

 alter with the effective valency. Now we have considered n to 

 be proportional to U; and we have found n in the case of 

 oxygen to vary with valency, so that the optical argument for 

 a possible variation of U with valency agrees with the varia- 

 tion in the value of n which we have found. It is to be 

 observed here that this effect of change of valency in changing 

 n may tend to obscure the value of H as compared with C ; for 

 in most cases a reduction in the number of H atoms is accom- 

 panied by a change of valency of the C atoms. In the study 

 of molecular refraction we find in a large number of cases 

 that the reduction of molecular refraction caused by the ex- 

 pulsion of H atoms for a molecule is almost exactly counter- 

 balanced by the increased refraction-equivalent of C due to 

 changed valency. 



In the previous table the value obtainable from Schiff^s 

 data for certain organic bromides and iodides have been 

 omitted, because of their exceptional character. The following 

 table contains the values of kA. for various substances. 



Values of Mx 1000. 



Brs C.HsBr CsH^Br CsH.BrCIso) CsH^Br C^HgBr 

 104 259 269 270 269 271 



CfiHnBr CeH^Br C.H^Br C.H.Brg CsHeBr^ 

 270 247 247 149 149 



