260 HENRY S. WASHINGTON 
each other should determine the minerals actually formed. In 
accordance with the general facts of chemistry these relative affinities 
would vary more or less with differences in the physical conditions, 
such as temperature and pressure, and their influence may conceiv- 
ably be altered by such factors as mass-action or the presence of 
catalyzers, bringing about different reactions between the constit- 
uents of the magma and the consequent formation of different min- 
erals. But for igneous magmas in general, and within the com- 
paratively narrow range of temperatures between the points where 
crystallization begins and the mass becomes solid, it will be per- 
missible to assume certain general orders of affinity which should 
control in the great majority of cases, at least, and which a study 
of the minerals formed in holocrystalline igneous rocks would reveal 
to us. 
Leaving out of account all of the (usually) minor chemical con- 
stituents, for the sake of simplifying the discussion, we may assume 
that rock magmas are composed of the oxides: silica, alumina, ferric 
oxide, ferrous oxide, magnesia, lime, soda, and potash. Whether 
these exist as such in the magma before solidification, or whether 
they are combined into more complex molecules, is an unsettled 
question, and one which need not concern us here. It is evident that 
a study of the relative affinities of each of these for the others would 
be very complex, but our knowledge of igneous rocks and of their 
component minerals clearly indicates that certain ones are of especial 
importance. 
Of these the most important, and the one which most interests 
us, is that of the bases for silica, as silica is always present in the 
largest amount, and the great majority and the most abundant of 
the igneous rock minerals are silicates. While no universally appli- 
cable order of affinity can be established, for reasons given above, 
yet the well-known facts of mineralogy and the experimental re- 
searches of Lagorio, Morozewicz, and others, indicate the following 
as being the one which controls in the great majority of cases,’ and 
the one which we may assume as fundamental in the present investi- 
gation. Beginning with the oxide which has the greatest affinity for 
t Cf. Cross, Iddings, Pirsson, and Washington, Quantitative Classification of 
Igneous Rocks (Chicago, 1903), p. 192. 
