( 140 ) 



region where — is positive (These Proc, April 26, 1907, p. 833) 



d.v 

 — either for a very great difference in the size of the molecules of 

 the components, or for a small difference. In the latter case the 



highest and the lowest points of = are to be found at almost 



d.v 1 * 



the same value of x. But this is one of the many particularities 



which is to be left to a later investigation. 



Particularly the last described way of splitting up of the spinodal 



dp 



curve takes place far to the left of the point where 



dx 



has mi- 



nimum volume, and so at a value of .i\ not very different from 

 that for which ,i\ = ,i\ on the vapour binodal curve, and maximum 

 pressure exists; and so this leads to the opinion that this detaching 

 of a longitudinal plait is to be found for mixtures with minimum 

 pressure and very different size of the molecules; but also this sup- 

 position must be further defined by a fuller investigation. 



The following remarks may serve for a full characterization of 

 the course of the spinodal line before and after the splitting. Before 



the splitting the curves - =0 and 



must be thought as 



dx' dv 



intersecting, as in fig. 8 (These Proc. March 30, 1907), but the line 

 d'tp 



d.r* 



as having moved to smaller volumes. This figure holds indeed 



for a left-hand region of the />figure, but this figure would change little 



in its essential features if we also insert the line 



dp 

 dx 



in it, but 



place it on the right so that — = is no longer intersected by it. 



dx' 



For a region of the left-hand side extended towards the right is the same 



as a region of the right-hand side extended towards the left. If = 



dx* 



and — = intersect there is a complicated plait, with the hidden 

 dv' 



plaitpoint on the right side. If now with rise of temperature the two 



curves get further apart, because they both contract, splitting up of 



the spinodal curve does not always immediately follow. For this to 



be brought about the curves must be pretty far apart, and intersection 



d'v d'v 



of 



dx' 



— = and = must take place between the two curves, 



(/./■ 



and the temperature must be reached at which this point of inter- 



