TWO-PHASE CONVECTION IN IGNEOUS MAGMAS 485 
to separate if intruded in a region of less pressure, even if it does 
not reach the surface. The solubility of a gas also varies with the 
temperature and with certain changes in the composition of the 
magma. For example, the assimilation of other rocks or liquids 
may so change the composition as to make the gas less soluble. 
It becomes evident that gases may separate at considerable depth 
as well as at the surface. 
Crystal phase.—It is a commonplace that a magma on cooling 
and under various modifying influences will crystallize in a series of 
mineral compounds. In the closed system, with some of the 
mother-liquor, each constitutes a phase. 
New liquid phases.—The question of immiscibility in magmas 
may be left open. While it is not well to use the term as if the 
process were known, the discussion of probable cases would not be 
complete without the consideration of new liquid phases. Changes 
in temperature or composition are the explanations given for their 
separation. 
FACTORS £N MAGMA MOVEMENTS 
The points to consider in a discussion of convection are the 
forces applied and the viscosity and related effects tending to oppose 
or retard movement. We are not now concerned with the forces 
leading to the intrusion of magmas. The estimate of forces is 
wholly dependent on specific gravities and on the volume of the 
portions of different specific gravities. These must be estimated. 
The viscosity of magmas is known to be variable with composition, 
pressure, and temperature; and in the present case we must esti- 
mate the added effect of the presence of a second phase. 
Viscosity.—The factor of viscosity is so great that by many the 
possibility of active convection in a crystallizing magma is dismissed 
as an absurdity; but the field evidence is strong enough to warrant 
a quantitative estimate. Furthermore there may be in the minds 
of many a misconception of the true nature of increased viscosity. 
A truly viscous fluid, as distinct from a weak solid, will yield to any 
small force if given time. Viscosity cannot inhibit the movement, 
but can only retard it. In a pitch a million of million times as 
viscous as water, stones will sink and cork will rise in a few weeks. 
