496 FRANK F. GROUT 
The probable size of crystals in convection.—It is suggested’ that 
if convection kept a supply of material at hand for the growth of a 
crystalline border, the largest crystals would be at the borders. 
This is actually the case in some pegmatitic differentiated dikes 
and indicates some movement in a thinly fluid magma. However, 
the mechanics of the process as outlined is not such as to furnish 
growing crystals a large supply of molecules from the central 
mother-liquor, though some such action may occur. The crystal- 
line-border phase grows as a result of small crystals becoming 
caught in a more viscous, less rapidly moving wall. Here the prog- 
ress of cooling is so advanced that, with the high viscosity, it is not 
to be expected that large crystals will develop. 
The accumulation of crystals—Most of the crystals, being 
heavier than the magma, would tend to settle; and though formed 
during the cooling along the top and sides of the chamber they 
would probably lodge at the sides and especially along the bottom 
during the forced circulation inward toward the rising current. 
Settling would here be entirely sufficient to remove crystals from 
the current. Similarly the crystals lighter than the magma would 
have a tendency to lodge along the roof and outer corners. Thus 
the segregation of minerals depends on gravity and is assisted by 
convection. It is largely independent of the place where cooling 
and crystal growth occur. 
Orientation of crystals —When a crystal once lodges in a viscous 
wall, too viscous to be again involved in general circulation, there 
may still be sufficient fluidity to allow orientation. The viscous 
matrix crystallizes, while the magma near it is still in motion, and 
the crystals would probably be oriented in the direction of the 
current—in most cases parallel to the walls of the chamber. 
Gravity differentiation.—As most of the crystals of an igneous 
rock are heavier than the magma from which they grow, it will be 
expected that whichever forms first will segregate toward the 
bottom. It is only the coincidence of high gravity and early 
crystallization that results in a strict gravitative arrangement in 
the resulting rock. However, the general order of crystallization 
is, as a matter of fact, roughly the order of decreasing specific 
tN. L. Bowen, ‘““‘The Later Stages of the Evolution of Igneous Rocks,” Journal 
of Geology, Supplement to Vol. XXIII (1915), p. 12. 
