404 N. L. BOWEN 



present, and the crystalline product cannot fail to show the blotchy 

 condition that this predicates. The evidence of immiscibility 

 would not be confined to rapidly chilled flow and dike rocks alone, 

 though it would presumably be especially clear in them. Further 

 cooling would result simply in the separation of more crystals of A 

 with a consequent change in the composition of the liquid from D 

 to C, where eutectic crystallization of both A and B would occur. 

 It will be noted too that the liquid C, which is the last material to 

 crystallize, is closely related in composition to the liquids K-D, 

 the first material to separate as a liquid. This is important in 

 connection with the well-recognized paralleHsm between ''Differ- 

 entiationsfolge " and " Kristallizationsfolge. " If liquid immisci- 

 bility were a prominent factor, or even a subsidiary factor, in the 

 differentiation of igneous rocks no such parallelism would exist. 



If we return now to those discontinuous changes of composition 

 in associated igneous rocks that have led to the suggestion of 

 immiscibility, we may make the discussion more concrete by refer- 

 ring to a specific case for which immiscibility has been invoked, 

 viz., the case of the Duluth laccolith. Grout has concluded that 

 the material of this great laccolith or lopoKth, as it has been called, 

 was intruded as a homogeneous magma, that crystallization- 

 differentiation controlled by convection currents and gravity 

 ensued, whereby the peridotite and banded gabbro were produced, 

 and that when a certain stage of crystallization had been reached 

 the material of the red rock separated from the gabbro as an immis- 

 cible liquid.^ The first stages of this process, with modifications 

 that are perhaps not fundamental, the writer is able to agree with, 

 but he can see nothing in the described Duluth rocks to warrant the 

 assumption of immiscibility, nor indeed in any other igneous series 

 that shows this association of gabbro and granophyre without 

 intermediate types. The kind of immiscibility assumed by Grout, 

 namely that which ensues only after a certain amount of crystalli- 

 zation has occured, is in no sense different from that which we have 

 discussed. Indeed some of the liquids of the system described show 

 this behavior, viz., those having a composition lying between A 

 and E (Fig. i). This liquid, which formed only after crystallization 



^Jour. Geol., XXVI (1918), 658. 



