362 PHYSICAL PROPERTIES 



which is employed as solvent. Similarly, the value of a, in the 



above equation, is the same for "insoluble" serum globulin 



whether this protein be dissolved in A^/10 KOH or in iV/40 HCl. 



The influence which a protein exerts upon the refractive index of 



its solution is therefore independent of the nature or proportion of 



acid or base which is combined with it, differing in this respect very 



strikingly from the optical rotatory power of dissolved protein, 



which, as we have seen, is very intimately dependent upon the 



nature and proportion of combined inorganic acid or base. The 



reason for this fact is readily perceived when we reflect that the 



power of dissolved protein to refract light is a function of the space 



which is occupied by the protein molecules. Now the molecular 



volume, and, indeed, the molecular refractivity is an additive 



function of the atomic volumes (or refractivities) of the atoms 



which together make up the molecule. In a molecule which 



contains over a thousand atoms, as a protein molecule does, the 



substitution of even several H atoms by K atoms or the addition 



of a few H or CI atoms or OH groups might be expected not to 



exert an appreciable influence upon the volume or refractive power 



of the whole molecule. The refractive index of a protein solution 



also remains unaltered by hydrolysis, upon which fact I have 



based a method of determining the comparative activities of 



trypsin solutions (50). 



From the investigations of Gladstone and Dale (11) and of 



A^ — 1 

 Landolt (27) it appears that the expression — - — , where N is the 



refractive index of a substance and d its density, which is known 

 as the specific refractivity, is very constant, being only slightly 

 dependent upon the temperature. Moreover each particular 

 substance in a mixture preserves its own specific refractivity 

 nearly unchanged; hence the refractive index of a mixture or of 

 a solution can generally be readily calculated from the refractive 

 indices of its components. Now the density of a dilute protein 

 solution is always very nearly that of water, that is 1, so that the 

 specific refractivity of a dilute protein solution may, with a toler- 

 able approach to accuracy, be taken as n — 1, where n is the re- 

 fractive index of the solution. Suppose c per cent of protein were 

 to be dissolved in a solvent of specific refractivity A^i — 1 (for 

 instance, a dilute acid or alkali, which is nearly equal in density to 

 pure water), then, if Gladstone's law holds good, and the density 



