CRYSTALLIZATION-DIFFERENTIATION IN MAGMAS 411 



It is suggested that the various kinds of filter-press action that 

 have been discussed hitherto probably cannot take place before the 

 mass is 80 per cent crystalline. In all cases we should have a marked 

 hiatus in the crystallization series that could have formed under 

 conditions favoring continuous differentiation, that is, conditions 

 favoring crystal settling. Even with the slow cooling and the quiet 

 cooling requisite for crystal settling a certain amount of the filter- 

 press action we have discussed may occur when crystal accu- 

 mulation has proceeded to the appropriate degree.^ No general 

 discussion of the combined effect will be attempted. Instead an 

 explanation will be offered for the phenomena exhibited by the 

 special example, the Duluth lopolith. 



DIFFERENTIATION IN THE DULUTH LOPOLITH 



On this igneous mass Grout has recently published an interesting 

 series of papers. It is a large, sheetlike body mainly of gabbroid 

 composition but showing differentiation, with masses of peridotite 

 at the bottom, a great thickness of olivine gabbro with striking 

 banded effects and anorthositic facies, and at or near the top a 

 so-called "red rock" or granophyre. Grout thinks that crystalli- 

 zation was a prominent factor in the production of differentiation, 

 and that convection was the dominant agent in bringing about the 

 observed distribution of crystals. There was also a wholly differ- 

 ent type of differentiation, according to Grout, which involved 

 the separation of the red-rock magma from the partly crystallized 

 gabbro magma as an immiscible liquid. To this latter phase of 

 his conclusions certain objections have already been offered that 

 later will be referred to again. Any type of differentiation in which 

 crystallization is the fundamental factor is, in the writer's opinion, 

 much more likely to occur, and any promising process that may 

 be put forward as conducive to the localized accumulation of 

 crystals is unlikely to meet with opposition from the writer. 

 Convection does not, however, seem to offer any promise. Grout 

 introduces a type of convection called two-phase convection, in 

 which a mass of liquid containing suspended crystals is considered 



' A combination of the two actions is probably the most promising explanation 

 of many monomineralic masses. 



