614 



Albert P. Mathews 



The following values in Table 3 were computed by formula 

 (10) from Schiff's surface-tension figures at the boiling point. 

 a is in dynes per gram mol. 



TABLE 3 



These results are uniformly somewhat lower than those 

 computed from Ramsay and Shields' figures by formula (7). 



Returning to the value C of Eotvos formula, the following 

 computation shows that it has in it the components ascribed 

 to it. 



It was believed, for the reasons given, that C of Eotvos must 

 be equal to K C R/3P C N 1/3 . The ratio of the internal to the ex- 

 ternal pressure is supposed to be, at the critical temperature, 

 very approximately equal to 7. Its real value is as follows: 

 If we assume that b c of van der Waals' equation is always 

 twice the volume at absolute zero, then b c = 2V . Since 

 V - V C /S, b c == 2V C /S. If this is so then the ratio between 

 K c and P c is equal to (S 2 S* + 2)/(S 2). S =- RT C /V C P C . 

 Hence C of Eotvos must be as follows : 



(i i) C = (S 2 S + 2)R/(S 2) 3 N 1 /a. 



Formula (n) can now be tested since C is known to lie 

 between 2.27 and 2.34 for several non-associating substances. 

 The results of a computation of C by this formula and the 

 values given by Eotvos are compared in Table 4. 



The results are evidently closely similar, but unfortunately 

 Eotvos did not give the value of C for many substances of 

 which S is known. 



