( 828 ) 
component at the same temperature. Observations, however, have 
repeatedly been made which would on the contrary give a greater 
value for the three-phase-pressure. If this result is real, and not the 
consequence of errors of observation, the circumstances must be 
different from those supposed above. Accordingly I have tried if it 
could be possible to account for such a phenomenon. If we think 
three-phase-pressure also possible for a system with minimum value 
of Ti, pios must really always be larger than the vapour tension of 
each of the components at the same temperature. And though objections 
might be advanced against this possibility which I do not consider 
as entirely devoid of importance, I have set these objections aside, 
and examined what would be then the further circumstances of the 
course of the three-phase-pressure. 
The first representation of the y-surface, which I gave already in 
my Théorie moléculaire ete., supposed this possibility, and in figure 11 
of these contributions I have expressly pointed it out. 1 will, how- 
ever, immediately state that then the highest temperature at which 
three-phase-pressure is still possible, must lie below 7%, . 
At lower temperature the p,2-figure consists of two branches which 
start from p, and point upward. In the same way of two branches 
starting from p,, in which p, must be thought much smaller than p,. 
At the point where the two vapour branches intersect, the coexisting 
vapour phase is found. Let us again call the concentration of this 
phase «,. An horizontal line drawn through this point of intersec- 
tion, contains the points which represent the two coexisting liquid 
phases with concentrations 2, and 2,. Then we have z, < #,<4,. 
Let us complete this figure by tracing the rapidly ascending line for 
V za 
Uv 
“7: . ‘ 1 
the equilibrium 2,3. We always have then —— 
With rise of temperature the branches starting from p, contract, 
and close to 7), they must have almost entirely retracted into the 
axis «=O, and so have got clear from the other part of the p,a- 
figure. So there is a temperature and that below 7, at which the 
three-phase-pressure has vanished. What has been left on the: right, 
is a continuous curve for the equilibrium 3,2. If we examine the 
circumstances occurring when the branches get detached more in 
details, we observe that if the line for the equilibrium 2,3 is vertical, 
; d, 
and ‘so v,,=— 0, is positive, and hence & is positive. 
12 
a, at Vs Us 
Where the curve retracting into the z-axis, is vertical, the value of 
d 
= is negative for the equilibrium 2,3. And when the two points 2 
Hi 
