874 Mr. S. D. Wicksell 



on 



The convergence of this series I have not ascertained other- 

 wise than numerically. It is clear, however, that the series 

 cannot converge for a value of P or y greater than the interval 

 of convergence of (3) when = O . 



For not too high pressures we can as before neglect all but 

 the first power in P and get 



F=l/^ [l + p(^±i 2 a -5U ... (7) 



That is, if a and b do not vary with P we have by unvarying 

 temperature: 



A -jfe?*0 o _l = P,' w 



where kg is the expansion-coefficient at constant pressure 

 by a temperature equal to 6 and a pressure equal to P. 



This formula is similar to (6). 



I have tested this formula numerically, and the results are 

 given by the following table. 



I have used values of h given in the tables of Landolt- 

 Bornstein, Those values are not, however, referred to a 

 constant temperature. They denote the mean expansion 

 for an interval of temperature : — 





For Carbon dioxide. 



X 







Temp. -interval 



1. 



(0°-20°C.). 



Pressure 



about 1 



-2 atm. 



k. 



P. 





A. 



Pi 



P 2 ' 



(Case 1) 0-0037128 



518 mm. Hg. 



case 1 

 case 2 



0-517 



0519 



( „ 2) 37602 



998 



case 2 

 case 3 



0-376 



0-376 



( „ 3) 37972 



1377 



II. 



case 1 

 case 3 



0-727 



0-725 



Temp.-interval (64° -100°). 



Pressure about 17- 



40 atm. 



k. 



P. 





A. 



P 2* 



(Casel) 0-004747 



12988 mm. Hg. 



case 1 

 case 2 



0-61 



0-68 



( „ 2) 5435 



18856 



case 2 

 case 3 



0-37 



0-49 



( „ 3) 6574 



26212 



case 1 



0-60 



0-71 



In this case the pressure is too high to allow us to neglect 

 the second power of P. 



