28 
MESSRS. C. T. HEYCOCK AND E. H. NEVILLE OiS" 
more C is dissociated the more rounded will be the point P, and the less conspicuous 
the middle ” curve, of which it is the summit. Attention has been drawn to this 
by Lorentz and Stortenbecker (‘Zeitsch, Phys, Chem.,’ vol 10, p. 194). If C 
is wholly dissociated by fusion, we should not expect its existence to be indicated by 
the freezing-point curve ; and as Le Chatelier has pointed out, if its melting 
point ■“ lies below the curve of fig. 1, the existence of the compound may equally fail 
to appear in the freezing-point curve. We believe that such cases of partial dis¬ 
sociation on fusion ai’e probable in the case of alloys (see ‘Jour. Chem. Soc.,’ 1891, 
p. 936, and 1894, p. 65). 
Eig. 4. 
/I 
Ik 
! 
_ 
' 
Q 
\ 
L 
\ 
r 
1 
. 
] 
L ' 
\/ 
A- 
h 
— 
) 
rt 1 
/ 
/ 
/ 
w 
\ 
? 
\fr' 
■es^/ 
do/' 
-Zo, 
4 C 
'/fie- 
im 
s 
/O/'icy/'i 
Si' 
ina 
■rn 
m 
%4 
p 
KCi 
■514 
In fig. 3 we have a case of probably frequent occurrence, where P, the melting 
point of C, lies below the curve B, but a portion of the system CA is possible, as 
shown by the curve QE. Both Le Chatelier and Boozeboom give examples of 
this, and some of the curves occurring in this paper probably contain examples 
of the same phenomenon. 
Thus an intermediate maximum indicates a compound and gives its composition, 
while an angle such as Q may indieoJe a compound, but does Hot necessarily give its 
composition. 
If, instead of one of the pure bodies A or B, or a compound, the substance formed 
is a “ solid solution,” in Van’t Hoff’s sense of the word, we should expect the line 
of freezing points starting from the point A to rise, or at all events the fall in the 
* In such a. case the compound decompo,ses into a solid and a liquid part; as when the deca-hydrate 
of sodium sulphate “ melts ” in the presence of a saturated solution of the salt. 
