( 228 > 
> X L and x g denote the concentration of the coexisting solid, 
liquid, and vapour phase, expressed in gr. mol. of the component 
B per one gr. mol. total, and V s , V L , and V g indicate the mol. 
volumes of these phases. 
If we now Confine ourselves to the system H,0 — Na,S0 4 , and if 
we apply the above formula successively to the melting-point lines 
of Na, S0 4 10 aq, of rhombic Na, S0 4 , and of monoclinic Na, S0 4 , 
we arrive at remarkable results. 
Starting with the melting-point line of Na, S0 4 10aq, we must 
remark beforehand that for the case that no maximum or minimum 
occurs in the p-^-figure, which is here excluded, (— ) is always 
\opJr 
negative. 
As now the denominator is already positive some degrees below 
the melting point in consequence of the predominating value of 
(a-S — xl)V 3 , the sign of- will be further governed by the nume¬ 
rator, and so we see that it follows from the, positive values of 
for the melting-point line of Na, S0 4 .10aq that the numerator, 
so the sum of the caloric quantities, is positive there. 
This is the usual case, and does not call for further remarks. 
From the figure p. 229, which represents the projection of the 
three-phase regions aSb + Z-j- G on the T,X-plane we see, however, 
that the melting-point line of the rhombic Na 2 S0 4 , which becomes 
stable in E presents the particularity of at first retrograding, and 
now the question suggests itself, how this behaviour can be accounted for. 
When in a solution the phenomenon of hydration takes place, the 
differential heat of mixing, which is on the whole thermodynamically 
positive, Can become negative, because the formation of hydrate is 
usually attended with generation of heat. 
Now it is clear from the above equation that the melting-point 
line can hecome a retrogressive one only when the numerator wi 
the caloric quantities has become negative , which will be the case 
when Q m assumes a predominantly large negative value. ^ 
If we now consider the transition point E, we know that e 
solution, which coexists there with solid Na,S0 4 .10 aq., is a S0 ^ 
equilibrium with rhombic, anhydrous Na 2 S0 4 . Now if this so ntion 
is hydrated, as is made probable by various phenomena, it is c ^ 
that the differential heat of mixing of anhydrous Na,S0 4 
much greater than that of Na,S0 4 .10 aq., as the molecules, 
