( 653 ) 
I showed already, that this correction-term is small in the case of 
normal substances, about of order » — 5. 
As the second member of the expression (6%) is always positive, 
even when a,, should be CVa, a,, the longitudinal plait on the 
y-surface (for it is obvious, that in the neighbourhood of v=d 
the spinodal curve belongs to the longitudinal plait, which can 
be regarded as a prominence of the transversal plait) will always 
close itself above a definite temperature at the side of the small 
volumes. 
This temperature 7, is the plaitpoint-temperature, corresponding to 
ak 
Ax 
(6%); it is given by (6%, in connexion with the expression for ==0); 
deduced from it, yielding for the plaitpoint after elimination of 7 
the value 
CI 
1 eneen ees 
[on verter et |, 
b,—b 
where r= ———. (compare vaN per Waars, Cont. IL, and also my 
1 
preceding communication, p. 579). Only when 5, = 4, (r = 0), ze will 
be ='/,. In each other case x, will be removed to the side of the 
smallest molecular volume. 
Just at 7, the closing will take place at the limit of volume 
v=b(«=a-); for values of 7’< 7, the longitudinal plait will 
remain wiclosed up to the smallest volumes. For in that case (compare 
the representation in space) a section 7=const. will cut the boundary- 
curve (6%), lying in the boundary-plane v = %, in a straight line. 
This temperature 7’, may consequently be regarded in any respect 
as a third critical temperature. For above that temperature a for- 
mation of two liquid layers will never present itself at values of v in 
the neighbourhood of 5, that is to say at very high pressures; just 
in the same manner as above the ordinary critical temperatures of 
the single substances can never appear a liquid phase in presence 
of a gaseous one. 
nd Is v=o, then for each value of x, 7’ will be =O, that 
is to say, the equation (6) cannot be satisfied in that case. The plait 
will consequently never extend to v=o. 
3. Is «=O or 1, then (6) passes into the two boundary-curves, 
lying in the two limiting 7’ v-planes, viz. 
Lae SAG . 2a, i 
R1 re he ban RE 3 (v — b,)?. 
With »>— 3+, (resp. 3b,) these two curves yield duly: 
45* 
