408 FRANK F. GROUT 
the result is, at least temporarily, a heterogeneous condition of the 
magma. Similar heterogeneity may result from crystals settling 
from an upper cooling zone into a deeper superheated zone. How- 
ever, even more important variations in the magma are thought to 
result from the removal from the cooling zone of crystals of an 
early period of formation. Thus if a uniform magma had cooled 
on the outer edges of a laccolith until olivine crystallized, not only 
would there be convection due to the increased density of the 
olivine substance, but as the current moved along the floor some 
olivine crystals would settle out, leaving the liquid to rise in the 
central part of the mass with a different composition from the 
average. 
It is important to consider in detail the result of this variation 
in the circulating magma. The material supplied by such a magma 
to the cooling border zone will differ from time to time and is 
almost certain to show some alternation because of a lack of rapid 
diffusion. As different material passed the cooling zone different 
crystals would be likely to develop, and a layer of different rock 
would be deposited on the walls and floor, giving rise to bands 
parallel to the current of magma and parallel to the walls of the 
chamber. 
A further possible cause of alternation of materials deposited 
may be found in rhythmical activity of the cooling, or intrusion, 
or gas supply of the magma. Cooling is affected by annual and 
longer rhythms, but the depth of most of the igneous masses makes 
it unlikely that the rhythm from the surface would have notable 
effects. A rhythm in extrusive action is well recognized,’ and the 
causes usually assigned to it would apply equally well to intrusive 
action. The supply of heat, gases, and lava may be distinctly 
periodic and may be responsible for the alternation of material 
crystallizing in the cooling zone of the magma. 
Banding developed in this way is not likely to be perfectly 
regular, because of the irregularities of the current, its rise in the 
center, and its possible tendency to corrode or resorb some bands 
already deposited. Settling crystals on a large scale would tend 
tJ. D. Dana, Volcanoes (Dodd, Mead & Co., 1891), p. 124; Bonney, Volcanoes, 
pp. 274 ff. 
