156 
IDDINGS. 
accord with the behavior of molten magmas in the process 
of differentiation, as well as in that of crystallization. Their 
behavior as eutectic substances, or those in which the point 
of solidification is below that of their components, has been 
pointed out elsewhere. 
The differentiation of molten magmas is not according to 
stoichiometric proportions and is therefore not a mineral- 
ogical differentiation. Other arguments which might be 
brought against the idea that the differentiation was a pro¬ 
cess of mineral separation are: the wholly uncrystallized 
condition of many lavas at the time of their eruption, and 
of most coarsely granular rocks previous to their arrival in 
the conduit in which they have solidified, and the consequent 
crystallization of the minerals in situ, and the comparatively 
rapid growth of all phenocrysts in porphyritic rocks. 
The character of the differentiation of molten magmas is 
therefore that of a chemical separation, affecting the ele¬ 
mentary oxides of the constituent elements. The question 
naturally following this conclusion is, How has such a differ¬ 
entiation been brought about? 
CAUSE OP THE DIFFERENTIATION OF IGNEOUS MAGMAS. 
Molten magmas as solutions .—It was pointed out in 1861 
by Bunsen* that the order of crystallization of minerals in 
igneous rocks indicated that molten rock magmas are solu¬ 
tions of silicate compounds which are liquid at high tem¬ 
peratures. This fact has been enlarged upon by Lagorio,f 
who considered molten magmas as saturated solutions of 
silicates, which are capable of carrying different amounts of 
different oxides. The result of his investigations led him to 
consider the solvent as a silicate with the composition of 
(K, Na) 2 0, 2 Si0 2 . 
The character and probable behavior of molten magmas 
* Bunsen (R.) Zeitschr. deutsch. Geol. Gesell. 8°. Berlin, 1861, vol. 13, 
p. 62. 
fLagorio (A.) Tschermak’s min. petr. Mitth. 8°. Vienna, 1887, vol. 8, 
pp. 437 and 508. 
