(441 ) 



And according to the formula Vol. IX p. IV,), Contribution 11) 



'd'v 





ifjtn 



„.,,, m 



— has the same sign as f — , and is therefore negative. In the 

 aa/spin \dx/ a 



point of contact the two lines p and g do no! intersect. The //-line 



lies in the point of contact on the same side of the 7-linc — e. g. 



on the lower side. But in lig. 12 contact of a //- and a q line has 



f /'if' 



again been drawn on the left side of - =0. But there the «-line 



il.i' 1 



remains above the 7-line all through. So there must be a point of 

 contact somewhere between, where there is a transition between 

 these two cases, and where the contact is at the same time inter- 

 mi 1 "''' dv (I' 2 r d*V , di> 



section. I hen not only = , but = and so also =0. 



dasp dtcq d.v*p dx* q dx 



Then we are in a plaitpoint. If taking due account of the course 



of the />- and 7-lines, we seek this plaitpoint it appears that this 



point does not lie on that particular 7-line that passes through the 



, • , d *ty 



highest point ot the curve =0, and has there a direction parallel 



to the ,c-axis, and also possesses there a point of inflection. Hut 

 it lies on a q-line lying left of the former, where /> has a greater 

 value; while this plaitpoint must lie below the point of inflection of 



d'v 



I he 7-line, because is always positive. 



dx i p 



Of course, bul this is not necessary -for our argument, if for points 



of the spinodal line with very small x, the contact of the q- and 



//-lines is to take place again in such a way that the //-line remains 



again throughout on the same side of the //-line, which we may also 



call the lower side, there must exist another plaitpoint also on the 



d*y 

 lett hand of — =0. So in this second plaitpoint the o-line, comine 

 dx" 111 



from the right, must first run above the 7-line, which it will touch, 

 and will be below it from the point of contact. What is indeed 

 essential for our argument, is the circumstance thai the first-men- 

 tioned plaitpoint, the upper of the pair of heterogeneous plaitpoinls, 

 which I called the realizable one in a previous Contribution, though 

 it only fully deserves this name when it also lies above the binodal 

 curve, lies on an isobar of higher value of the pressure than the 



