68 THE TRUE VALUE OF a OF VAN DER WAALS' 



tion on the carbon atom either in CS 2 or in CCl 4 , where it might 

 perhaps have been anticipated. 



Figure 2. Representing the molecular structure of 



2 , CO, CO%, S0% and X. 2 <> with the number of valence electrons 



called for l>v the cohesion. 



OO CO oco 



ee se G0G00 



2 CO CO» 



(2 valences) (2 valences) (5 valences) 



OSO NON 



0000 000 • © © © © 



so., N 2 



(S valences) (5 valences) 



IV. THE RELATION OF a TO THE OTHER CONSTANT b. 



The rather simple and interesting relationship appears to hold, 

 if we compare the values of a computed from the gravitation and 

 valences with the value V c — b,., that Log a = K' {{V c — b c )/b c ). 

 That is: the ratio of the space between the molecules to the space 

 within the molecules at the critical temperature is proportional to 

 the logarithm of a. For the constant K' I have taken the value 

 of 13.3 when using logarithms to the base 10. This is the value 

 of Log a when b r = V c j2. With this formula and the values of a 

 given in column 8 of table 2 we find the following values of à,.-. 



K 



Hydrogen r,./1.8G 



Nitrogen Vjl . 92 



Benzene. V c /2 . 00 



Carbon tetrachloride . . . / " c /2 . 00 



Methyl butyrate FJ2.01 



Octane V c /2 . 02 



Diisoamyl V r /2.03 



The other substances of the table will come in between the extreme 

 values. It will be seen that, if this formula is correct, b c is, indeed, 

 for nearly all substances quite close to the value of / 7 c /2. The for- 

 mula is, however, purely empirical. It seems a not improbable relationship 

 and is certainly very nearly true, I think, if it is not exactly true. 



There is one point in this relationship, however, which 'deserves 



