( 898 ) 
; sae a d? 
unless in a point in which either gen 0, or EE. it is required 
dv? da? 
2 
for the possibility of the splitting up that the curve ==? lies 
le 
f d, 
entirely on the same side of (=) =O as the point C, which had 
az}, 
already been taken into account in fig. 25. It is true if the position 
de 
of 7 bi — 0 should be as drawn in fig. 44, a very complicated course 
av 
for the spinodal line might be imagined, in which this line passes 
; dp ' Op 
twice through de = 0, every time touching AS = 0, but besides 
Lt) y 2 ‘ 
with other difficulties, we meet then with the following objection. 
t 
baie dv 
In this case there must be a point in which | — ] =o on the right 
av spin 
d? d*v 
of En for the spinodal line, which would require é == 0. 
dax? da?) 
And such a point cannot be pointed out there. At all events we 
shall assume, if it were only for simplicity’s sake, that the position 
W 
da” 
this figure C represents the double point. At higher temperature the 
—=0 is as indicated by the small figure P in fig. 45 *). In 
two parts of the spinodal line are quite separated, and they behave 
independently of each other. On increase of 7’the spinodal line which 
dp : 
belongs to = —() moves to greater volumes, and the spinodal 
V Jr ; ’ 
1) To the solution of the question whether, if the spinodal line belonging to the 
equilibrium liquid-vapour, is found at a certain value of x, there can occur again 
a point of a spinodal line for smaller volume, the following equation may 
contribute 
d? 
inven 
En 
dp de) . 
¢ Se ER 
Following the line x= constant, { dp - | must be =0. If dp is always 
pT 
U 
2 
Gre - v : - : 
positive, the quantity =) must reverse its sign between two such points. 
a pr 
From this follows also that a detached closed longitudinal plait can only occur in 
2 
2 
v 
a region of fig. 1, where the circumstance occurs that for small volumes ) 
Fe PU, 
P 
is negative. 
