( 233 j 
pass into the vapour line gp in p, and in consequence of this the solu¬ 
bility line will have to bend to the left, which gives rise to the 
phenomenon of retrograde solubility. 
As is evident from fig. 1 we get the retrograde phenomenon at 
q on the vapour line for the same reason. 
Another remarkable phenomenon which fig. 1 also reveals, is that 
of retrograde condensation and of retrograde solidification. 
It is clear that if starting from the point a we roll a vertical 
tangent plane algng the spacial figure in the direction of the axis of 
temperature, this plane touches the vapour sheet of the continuous 
liquid-vapour surface first for some time, but a change has come 
before the temperature of the point p is reached. The before men¬ 
tioned vapour sheet is no longer touched, and contact now takes 
place with the vapour sheet of solid B. The change takes place at 
the moment that the two vapour surfaces are touched, which reveals 
a double retrograde phenomenon. 
This behaviour now proves that we have first only the phenomenon 
of retrograde condensation, which at a certain temperature is suc¬ 
ceeded by the phenomenon of retrograde condensation and solidification, 
and afterwards only by the phenomenon of retrograde solidification. 
What has been discussed here holds also for the neighbourhood 
of q , but here the phenomenon takes place in reversed order. Starting 
from q we get first retrograde solidification and then retrograde con¬ 
densation with rise of temperature. 
The phenomenon of retrograde solidification continues to exist 
between p and q (this was found for ether-anthraquinone, but it has 
proved not necessary theoretically), and this is the reason why the 
surface for the equilibrium between solid B and fluid has such a 
peculiar shape. 
Now it is to be seen in fig. 1 that at the temperature of p the 
solubility isotherm has a somewhat different shape from that of q; 
this is in perfect harmony with the experiment. It has viz. appeared 
that the part, of line pe running to the right is much larger than 
that of line qd , and in consequence of this the melting-point line 
will have a shape as indicated by the line cdef, which is a melting- 
point line with two vertical tangents at a pressure higher than that 
corresponding with the highest three-phase pressure 1 ). 
In conclusion I may call attention to fig. 3, which shows the 
backside of fig. 2, or rather the spacial figure on the .B-side seen 
slightly from above, so that the melting-point line cdef is clearly visible. 
Amsterdam , Sept. 1909. 
Anorg. Chem. Laboratory 
of the University. 
b This 
will be examined in detail. 
