184 VICTOR ZIEGLER 
For a different case let us assume we have three metals fused 
into an alloy. Provided they are completely soluble in each other, 
the eutectic ratio is the controlling factor, the metal in greatest 
excess crystallizing first, until the second metal reaches its satura- 
tion point, when the two come down together, until finally the 
ternary eutectic point is reached and crystallization of all goes on 
simultaneously. In case the metals interact, chemical compounds 
may be formed and complicate the system. Thus in a combination 
$35°C 
N 
a 
—w 
) 
ir) 
utelctic 
lecth ce 
FAS 
feS 
2 
| 
Ca,§ —_¢-—_—_ 
Fic. 2.—Order of crystallization in alloy of PbhS=7o per cent, FeS=20 per cent, 
Cu.S = Io per cent. 
a a p74 
: 
@® 4% SO) b%0 x 
x 
! XA 
, 7ce GS 
Fic. 3.—Crystallization of K and Mg sulphate solution, when K,SO, is in excess 
and saturates solution between 92° and —3° C. Schénite (2) is stable only between 
92° and —3°C. See van’t Hoff’s Theoretical and Physical Chemistry, p. 86. 
of bismuth, lead, and tin, the order of beginning of crystallization 
will depend on relative masses and will end with a ternary eutectic. 
In the case of a copper, tin, and antimony mixture, we get chemical 
reactions during cooling which may result in the formation of three 
pure metals and seven chemical compounds. There is no recur- 
rence of the substance first crystallized in either case. If substances 
are capable of forming solid solutions they do not affect the above 
in any way. 
For sake of simplicity let us assume we have a magma of SiO,, 
KAISi,Os, and NaAlSi,Os, an example worked out by Harker. 
t Harker, Natural History of Igenous Rocks, p. 251. 
