X 



( 423 ) 



the phenomenon of double retrograde condensation predicted by 

 VAN DER Waai-s bj a different method ^). Tlic plienomenon occurs 

 when the vapour is about to form a critical {)oint with one of the 

 liquids while its representative point is still on the vapour side of 

 the plait, (which involves that it has to turn the corner of the plait) 

 and this condition is certainly fulfilled, as van der Waals pointed 

 out, if there are two liquid layers below the critical temperature of 

 the lower component and this is nearly always the case. But the 

 phenomenon is not limited as might be supposed from his paper ^) 



to the case that the three phase 

 pressure is lower than the vapour 

 pressure of the lower component. 

 In order to show this I will con- 

 sider the case not discussed by van 

 DER Waals where the three phase 

 pressure is above the vapour pressure 

 of both substances, as with ether 

 and water (Fig. A). The vapour 

 1 must come together with the 

 liquid 2 : therefore either 1 must 

 move round the extreme point of 

 the plait towards 2 or 2 comes 

 round to 1. In the first supposition (Fig. BCD) the maximum in the 

 vapour pressure which lies hidden inside the triangle appears outside, 

 at the moment when 1 and 2 have the same composition x{B), 



/I 



\r 



gradually moves off the plait and disappears ; at this moment the 

 plaitpoint P belonging to the equilibrium 1,2 passes over to the 

 liquid side, and the case has now become identical with that considered 



1) 1, c. p 816, sqq. 



2) Ph. Kohnstamm and J. Ghr. Reeders. Proc. XI p. 917. 1909. Professor 

 KoHNSTAMM draws my attention to a note in a recent paper by the same authors 

 n the September Meeting (p. 272) where the limitation is withdrawn. 



