THE ORIGIN OF IGNEOUS ROCKS. 
157 
when considered as saturated solutions or alloys of silicate 
molecules has been discussed by the writer in a paper “ On 
the crystallization of igneous rocks/’* when the influence 
of physical conditions on the saturation and crystallization 
of the compounds was especially dwelt upon. 
The analogy between alloys and solutions is so close that 
there seems to be no essential difference between them, and 
they may be considered as synonymous. 
That one particular compound in the mixture of a magma 
should be considered the solvent for the others in all cases 
seems to the writer to be contrary to the evidence furnished 
by the results of crystallization w T hich are exhibited by the 
microstructure of rocks; for, if it is necessary to consider 
in each instance that a certain portion of the magma acts as 
the solvent for those compounds which separate first from 
the magma when it crystallizes into rock, then a study of the 
order of crystallization of the minerals in rocks proves that 
different compounds have played the part of solvents in dif¬ 
ferent magmas; and since the general character of these com¬ 
pounds is the same in all rocks, it is evident that the behavior 
of any one of these compounds as solvent for the others de¬ 
pends on the conditions of their association—that is, chiefly 
on their relative amounts, or on their mass. A few illustra¬ 
tions of extreme cases will be sufficient. 
In the highly silicious and feldspathic rocks, when the 
crystallization is coarse enough to permit all of the minerals 
to be identified, it is found that the last to solidify was a 
mixture of alkalies, alumina, and silica in such proportions 
that it separated on crystallizing into alkali feldspar and 
quartz. Since the crystallization of these two minerals is 
in many cases synchronous, the solvent of the other com¬ 
pounds of the rock must have been a mixture of the con¬ 
stituents of these two. In certain phases of the gabbro 
occurring in the granular core in the Crandall region, the 
last mineral to crystallize was orthoclase, which formed a 
biddings (J. P.) Bull. Phil. Soc. Washington. 8°. Washington, 1889, 
yoL 11, pp. 65-113. 
