658 FRANK F. GROUT 
red rock. It seems doubtful if any other process than the sepa- 
ration of immiscible fractions could operate with the separation of 
crystals to give such a distinct differentiate. 
SUMMARY 
The rocks of the Duluth gabbro lopolith are found to fall 
naturally into two series, one related to the gabbro family, the 
other more closely to the granites. Intermediate types are rare, 
though in the field the gradation between types is visibly complete. 
The abruptness of the separation in the field as well as in the 
arrangements on the basis of laboratory data indicates that the 
process involved in the separation of granite and gabbro was of a 
different nature from the more easily understood process by which 
a variety of phases developed in the gabbro. ‘The several modifi-. 
cations of the gabbro probably resulted from a differentiation by 
crystallization during convection, aided by a slight amount of 
settling of crystals. Before the solidification was complete the 
granitic magma must have separated from the gabbro, probably by 
some other process, for the process giving variety to the gabbro 
seemed to produce no modification in it approaching the granite. 
The evidence is strong that differentiation of two sorts may occur 
in a single magma chamber. If crystallization and settling and 
convection are involved in the main process, it is interesting to 
suggest other possibilities for the other process. The difference 
may be due largely to a difference in the concentration of water, 
but the field and laboratory studies both strongly suggest an 
immiscible separation of the red rock from the gabbro. 
November, 1918 
