626 Albert P. Mathews 



and fi, the heat rendered latent by an increase in the intra- 

 molecular energy, is 14.80 cals. While the figure 14.80 is 

 undoubtedly a little too low it appears that approximately 

 one-fifth of the total internal latent heat goes within the mole- 

 cules at the fraction 0.313^. Furthermore, the internal 

 intramolecular latent heat does not increase much, if at all, at 

 temperatures lower than this. The theoretical latent heat at 

 absolute zero was calculated by Mills' formula, and d was 

 taken as $d c . 



5. Computation of "a" from the Number of Valences and 

 Molecular Weight 



Finally I have calculated "a" for a gram mol by the formula : L 



(34) a = C(M X Val) 2 / 3 . 



C is taken arbitrarily as equal to 1.259 X io n . M is the molecu- 

 lar weight, and Val the numbers of valences per molecule. 



The values computed in these different ways are given in 

 Table 8. For the surface tension computations I have taken 

 the data from Ramsay and Shields, Ramsay and Ashton and 

 Renard and Guye. In all cases when computing "a" by the 

 last formula the valence of carbon has been taken as 4 ; oxygen 

 as 2, except in oxygen gas, where it is unity; nitrogen as 3, 

 except in nitrogen gas which has been taken as monovalent; 

 and hydrogen as i . The critical data of oxygen are accurately 

 determined; for hydrogen, the critical density being uncertain, 

 I have assumed S to be 3.4, the same as oxygen ; in computing 

 nitrogen, the critical density being uncertain, I took S as 3.5 

 and the density 0.33, which is between the values given by 

 Sarrau and Hautefeuille and Cailletet. For all substances 

 included in his list Young's critical data have been used. 

 The critical temperatures and pressures of the other substances 

 have been taken from the Landolt-Meyerhoffer tables. Where 

 the critical density was unknown I have been unable to com- 

 pute by formulas which involve that factor. In other cases 

 the surface-tension data or the vapor densities could not be 

 found. Hence there are many gaps in the table. 



1 Mathews: Jour. Phys. Chem., 17, 181 (1913). 



