98 
IDDINGS. 
Application of the Foregoing Conclusions to Rock Magmas. 
Saturation of Rock Magmas. —From the foregoing observa¬ 
tions we may reasonably conclude that the degree of satura¬ 
tion of natural molten glasses or rock magmas will depend 
on the temperature and pressure at which they exist. In a 
perfectly molten state the relation between the temperature 
and pressure is such that all of the compounds remain in 
solution. The character of the saturation will depend pri¬ 
marily on the relative amounts of these chemical compounds, 
which are present in all rock-magmas in variable propor¬ 
tions. 
In a cooling magma which originally was perfectly molten 
we may imagine the rate of cooling such that crystallization 
is possible; then as the temperature decreases a point will 
be reached where some one or more of the chemical com¬ 
pounds supersaturates the solution and begins to crystallize 
out. At a lower temperature a point is reached where another 
combination must crystallize out, and so on until the whole 
mixture has solidified—it may be to a complete mass of crys¬ 
tallized compounds or to a mixture of crystals and amor¬ 
phous, uncrystallized glass. 
(1.) The nature of the compound that crystallizes first will 
depend primarily on the chemical make-up of the magma, 
and will vary accordingly. From the order already given 
of the ability possessed by the chemical compounds in ques¬ 
tion to saturate such a solution, it appears that it requires a 
smaller amount of the iron oxides and magnesia than of 
lime, and a still smaller amount of these than of the alkalies 
and alumina to saturate the magma. Hence, if they occurred 
in nearly equal quantities, the first compounds to separate 
out would be the ferro-magnesian silicates, followed by those 
bearing lime and alumina, with more or less of the alkalies, 
while the last to separate would be the alkali-aluminous 
silicates and free silica. This order would be maintained if 
the proportion of the -first-named compounds increased, but 
