1 i BEDINGS OP THK A.MEWCAN A.< A.DEMT. 



'I In- constancy of the quantity, li I , or a. we have <>iilv proved 



1 • ,1 p • l 



in the cases of th< which were experimented upon l>y Joule and 



Thomson, and through limited range of pressure. However, Bince the 

 constancy of this term u iblished for nil the gases tried, we may 



assume that the same result would have been obtained with any other 

 b through the same limits of pressure. It cannol be taken for 

 ated that this constancy will hold at all pressures. In fact, the inter- 

 pretation given by van der Wan)-- for the quantity ". as the iutermoleo- 

 ular attraction, would lead to the idea that it must be a function of the 

 volume. Otherwise the attraction between two molecules would lie 

 independent of their distance apart. The variation of a with the vol- 

 ume will be mentioned later. 



Van der Waals considered apriori that a would not depend upon the 

 temperature. But it is evident from page II that it i- a constant onfy if 



dc 



— — - = 0. Although the change <>f specific heat with moderate change of 

 dv & i & 



volume is negligible, the change through wide limits of volume is proba- 

 bly always a measurable quantity, as will be shown later. 



dJS, 

 Since the quantity a has been defined as equal to — II T , we 



should he able to determine its value from the data of Joule and 

 Thomson. A calculation which is in effect the same as this has already 

 been made by van der Waals,* who calculated from the values of a and 

 b, given by experiment, the cooling effect that should be observed in free 

 expansion. This cooling effect, in Centigrade degrees, for a change of 

 one atmosphere in pressure is given below, as calculated by van der 

 Waal-- and as determined by the averages of Joule and Thomson's 

 results : — 



The agreement, although not close, is very satisfactory considering the 

 Uncertainty of the experimental results. 



2. Liquids. — The experimental proof of the continuity between the 

 liquid and gaS60U8 state's BhoWS that we must regard them as identical, 



* Kontinuit&t der F. und <;. Zustand, p. L16. 



