( •'^O? ) 



is so slight that if the latter lias been discussed, also the i)articidar- 

 ities of the particular case are easily seen. 



In the adjoined tigure I have represented the position of the 

 spinodal and binodal curves at a temperature somewhat above 7^^ . 

 AB and CD denote two portions of the binodal curve of the trans- 

 verse plait, for so far it is to be realized. EBP^CF represents the 

 binodal curve of that part of the longitudinal plait that has split off 



and moves towards smaller volumes at rising temperature. P^ is the 

 plaitpoint moving towards smaller volumes. By P^ the plaitpoint 

 moving to greater volumes is represented. At Tg^j P^ and P^ coin- 

 cide. But now we know from former considerations and I refer 

 among others to these Proc. VIII p. 184 that there must also still 

 exist a closed binodal curve, of which P^ is a plaitpoint, and that 

 there must be found one more (third) plaitpoint to close this binodal 

 curve. The points P^ and Pj together form a pair of heterogeneous 

 plaitpoints. At Tgp this hidden plaitpoint already exists, but P^ and 

 Pj dg not coincide until at a temperature that lies higher (but how 

 much higher cannot be indicated here any more than in the cited 

 paper). 



At a temperature above that for which the above figure holds, 

 the longitudinal plait moving to smaller volumes has the plaitpoint 

 Pj on the binodal curve of the transverse plait, which can then be 

 realised over its full width, and at still higher temperature it has 

 no longer any points in common with this binodal curve. 



If figure 47 of Contribution XV (These Proc. XI p. 904) is 

 compared with the result of this discussion, we see that the tem- 

 perature of the point C of figure 47 is Tg^, , and that of the [)oint 

 D the temperature at which the pair of heterogeneous plaitpoints 

 Pj and Pj coincide. Already in contribution XV I remarked about 

 fig. 47 that the righthand branch betw^een E and F may be omitted. 

 Then, however the point E must be thought to lie at p^co, and 

 the point P at P^O. Since then I have not been confirmed in the 

 opinion that this branch should be omitted. 



In fig. 47, however, a mistake has slipped in, with regard to the 

 closed curve which represents the concentration of the two liquid 



