508 



UNIVERSITY OF VIRGINIA PUBLICATIONS 



future to find by experiment whether or not such a maximum really exists 

 at about this concentration. Also it would be interesting to see if the 

 diffusion curve falls, as does the theoretical curve, at concentrations higher 

 than 4.5 normal. All that we can say at present is that the empirical 

 equation given above in all probability gives a fair approximation to the 

 truth. 



The following empirical formula also gives a fair approximation to the 

 experimental value for diffusion constants, 



where A is a constant (about 6). This formula seems to have a somewhat 

 more rational basis than the one previously given, but it does not agree 

 quite so well with the experimental facts. Table XI gives the values so 

 calculated, and also those obtained by the former method, in comparison 

 with the experimental data. 



In any case it is necessary to assume that the density is an important 

 and direct factor in the rate of diffusion. (Of course we can not distin- 

 guish between density and concentration.) 



TABLE XI. 



A comparison of diffusion as calculated from the formula 



(I) K = 2F' + 0.4p - 1.42 

 and the formula, 



(II) K = F Vp + A (Where A = 6) 



and the experimental values. 



