( 91 ) 



we get : 



. , 1 A « / 1 1 



n lo(j 



n'l\^ n^ ) RT^ \ni>.^ h 



"l — nbA 1 



a —Lb / uh;\ _ a { — Lhf 



RT, bT.lib., l ~ b7j~ RT, IJTah, 



from which 



a 



Ahv r h' 1 



Of course this expression also holds for Ab positive, if only a 

 triple point occurs, and this lies far enough from any critical [)oint 

 to justify the just mentioned suppositions. The value of 7\ to be 

 calculated from (31) only holds for the case Lb negative, for only 

 then there is a pressure of coexistence />„ at 7'=0. Moreover T„ 

 cannot be explicitly solved from (31). (.)n the other hand in (32) the 

 quantity if occurs, of which we only know that it will l)e near (). 

 But as we shall see, all the same some inferences may be made 

 concerning 7\ or rather concerning the relation ^'"/^t- 



If we suppose that at the critical temperature (vapour-liquid) the 

 molecules have become single for the greater part, 7^c can be cal- 

 culated from : 



8 a„ 8 a, : n" 8 a 



' 27 b, 27 b.^ 27n nb, ' ^ 



27 n a 



Hence RJc can be written for — , in consequence of which 



(32) becomes : 



2\ 27 /A/zV / ^i' 1 A 



— = — : /or/ ( -i- ._.... (32^) 



Formula (32«) differs in this from (27") on p. 461 loc. cit. that 



apart from the substitution of {nh^Y for (2/;.J-, the numerator 2ti' 



has now changed into ni^ . This is very essential, and brings the 



value of TJ7\ into the neighbourhood of the experimental value \/^, 



without such a large value of Lb being required for this. We saw 



Lb 

 in V p. 461 that [i' would slill have lo be =0,37 for =— y^, 



to bring the ratio TjT,. to 7^ for /^ = 2. Only for still greatei- values 

 of Lh, {i' might have beeji slightly smaller. This is no longer the 

 case now. 



8 a 8 CI. 



1) Also from (1 -f (yi—l)[3) yi2' = - _ y- , whicli passes into nRT,- = - ~- with 



^ 27 b 27 Jib^ 



(3 = 1. (a is viz. independcaL ol' p and -^ Ui lur an «told molecular quautilyj. 



