( 623 ) 
which converts this branch plait into a main one. On the other 
hand that part of the plait, which at 7, was situated in the neigh- 
bourhood of the existing plaitpoint lying below 7, and which was 
then a main plait, must have been reduced to a branch plait for 
values of 7’ slightly below 7. 
That the distinction between a main plait and a branch plait 
is not arbitrary, but essential, appears when we determine which 
of the two tops which occur between 7, and 7, belongs to the 
base of the plait, and when this is ascertained, examine in what 
way the binodal curve of the other top must be completed. 
So the question is, when the bi-tangent plane is rolled over the 
binodal curve from the base part of the plait, which of the two 
occurring tops will be reached by continued rolling. 
If we consult fig. 1, it is easily seen that a rolling tangent plane 
which comes from the right side, and which has reached the two 
points of contact A’ and A”, has obtained a new point of contact in 
A, lying on the same isobar and in this way has become a plane 
touching in three points. At the assumed temperature we have there- 
fore a three-phase-pressure. In this case there are two tops of a 
plait viz. P and Q. But there cannot be any doubt as to which 
of these two tops belongs to the base part lying right of A’ A", 
If viz. we continue to roll the tangent plane when it has the 
line A A' as nodal line, the binodal line on the side of the small 
volumes between the points A” and A is completed by the curve 
A" BCA, the configuration A'B'CA' giving on the other hand the 
completion on the side of the larger volumes. This harmonizes 
with the diagram in my Théorie Moléculaire. (Cont. II p. 23). 
So when continuing to roll we reach P as top of the plait. We 
are therefore justified in considering the part of plait A'PA as 
belonging to the main plait. There lies, however, on and by the 
side of the main plait, a second configuration, of which AQA" is a 
part. If a rolling tangent plane is moved over it, starting from Q, 
the binodal curve described in this way does not end in the points 
A and A", but if the plane has reached those points and has there- 
fore again assumed the position of the three-phase-triangle, we may 
roll it continuously further till it has reached a point of the spinodal 
curve. This curve is denoted by D in fig. 1. The binodal curve 
under consideration has then obtained a minimum pressure; the 
conjugate point D' is then a cusp‘). 
1) For a proof of these and similar properties consult Cont. Il, fig. 3. Further 
the very important papers of Korrewee on the theory of plaits. 
