258 FRANK F. GROUT 



CRYSTAL SETTLING 



The reality of the separation of crystals from liquid magma 

 by gravity has been clearly shown/ In crucibles a few centimeters 

 deep, crystals formed and settled and grew to considerable size. 



In order to determine equilibrium relations, efforts were made in 

 the laboratory work to maintain uniform conditions of heat through 

 the whole melt. Contrasted with these conditions are the large 

 size and irregular coohng of intruded magmas. A temperature 

 variation of 100° C. is not unlikely in the viscous liquid beginning to 

 crystallize. If crystals form in a cooling top layer and settle to a 

 hot interior layer they remelt and make the magma heterogeneous. 

 Eventually the continued settling of crystals may so cool the central 

 part of the magma that the crystals formed near the top settle 

 through the central part without either growth or solution. It 

 would take so long for this balance to be attained, however, that 

 it seems likely that the roof phase would grow as a solid before 

 many crystals settled. Meanwhile in any large mass the cooling 

 sides of the chamber would start convection and entirely modify 

 the situation. The probability seems to be that chambers a few 

 feet thick are chilled; up to 100 feet crystal settling may be the 

 dominant effect, but in larger chambers convection stirs things up 

 too much. Settling a few feet removes crystals from the circulating 

 magma. 



The condition of the gabbro at Duluth was taken as a sign that 

 no great amount of settHng occurred,^ and Bowen agrees that simple 

 settling does not adequately explain the series of rocks developed. 

 He finds it possible still to refer to the peridotites as a result of 

 settling. They are bands, like other bands in the gabbro, and if 

 most of the bands are not the result of settHng it seems inadvisable 

 to select the peridotites to support the original idea. 



In a remark on the banded structure at Duluth, Bowen says 

 that if crystallization is rhythmic, crystals brought down to the 

 bottom by settling would result in precisely the same alternation as 



^ N. L. Bowen, "Crystallization Differentiation in Silicate Melts," Amer. Jour. 

 ScL, XXXIX (1915), p. 186. 



' Frank F. Grout, "A Type of Igneous Differentiation," Jour. GeoL, XXVI (1918), 

 p. 639. 



