A TYPE OF IGNEOUS DIFFERENTIATION 639 
Nearly all the feldspar of the gabbro, through a thickness of 
three miles, is basic labradorite or acid bytownite, with surprisingly 
little variation and no apparent relation between position and the 
slight variations found. Near the top for a few feet the feldspars 
are zoned, and directly above is the red rock in small amount. Can 
15,000 feet of bytownite, Ab,An., settle out of a magma of more 
acid composition without changing the composition of the residual 
liquor, so as to produce a notable change in the crystals forming ? 
And, if so, can 15,000 feet of bytownite, Ab,An., settle from a 
mother liquor that amounts to less than 300 feet of acid andesine ? 
It is evidently absurd to think that the main gabbro settled, leaving 
a mother liquor of red rock in such small amount. The gabbro 
feldspars are too uniform. The early crystals, which were very 
basic according to theory, forming from a labradorite melt, must 
have remained in contact with the mother liquor until equilibrium 
was established and they became average in composition. The 
crystals may have settled a little, but the viscous magma more than 
likely moved with them in convection most of the way. The end 
of crystallization came when the crystals lodged in the more viscous 
wall or floor, and there slowly, maintaining equilibrium, the crystals 
adjusted their composition to that of the magma around them, 
some bytownite, some labradorite." 
In connection with crystal settling the gravitative position of 
differentiates is cited as strong confirmation. The Duluth gabbro 
is supposed to be one of the best illustrations, since it is commonly 
thought that magnetite separated at the base. As a fact, the seg- 
regated ores are far from the base; the best concentrations are 
bands centrally placed in banded gabbro. Ores near the base are 
contact ores or xenoliths, and the gabbro at the base shows very 
«The mechanics of the convection has been outlined elsewhere. See Frank F. 
Grout, “Two-Phase Convection in Igneous Magmas,” Jour. Geol., XXVI (1918), 481. 
Another more general criticism of crystal settling may be noted at this time. 
Bowen records in the Am. Jour. Sci.. XX XIX, 175, that crystals grow during set- 
tling in a crucible from an infinitesimal start to one-tenth of a millimeter, in settling 
15 millimeters. How far-fetched it is then to think of the grains of common igneous 
rocks as having settled thousands of feet in a laccolith or batholith! Crystal settling 
is an idea to think of in terms of a few feet rather than in hundreds of feet. 
