256 . ■ FRANK F. GROUT 



CONVECTION 



In making an estimate of the rapidity of convection^ an argu- • 

 ment was built up by starting with a variety of assumptions. This 

 was done because several writers had assumed, without any sug- 

 gestion of calculation, that the magma was too viscous to circulate. 

 The value of assumptions not based on quantitative data may be 

 illustrated by references to the effects of viscosity. Harker 

 approves diffusion, but considers the magma too viscous to allow 

 settling of crystals. Daly approves settling, but considers the 

 magma too viscous to allow convection. Pirsson approves con- 

 vection, but considers the magma too viscous to allow diffusion. 

 Thus we get around the circle and all processes are still worthy of 

 consideration. 



The results of calculation, in actual figures, indicate great 

 probability of some circulation. The calculated rates are not to 

 be taken as at all accurate. Neither is it important to distinguish 

 whether cooHng, gas phase, or crystal phase is the dominant cause.^ 

 The evidences are sufficient to indicate some active circulation. 

 Bowen says now that he has never doubted the reahty of convection 

 in magmas.^ 



The contrast presented in a paragraph headed "Crystal Setthng 

 vs. Convection" is apparently misleading. The present writer 

 had no thought of differentiation except through the agency of 

 crystallization. Convection seemed a necessary factor in explain- 

 ing the combination of banded and fluxion structures. It is argued 

 that the rhythm of setthng would give the same banding as rhythm 



'Frank F. Grout, "Two-Phase Convection in Igneous Magmas," Jour. GeoL, 

 XXVI (1918), pp. 481-89. 



2 Bowen in a later paper questions whether the two-phase idea can be extended 

 to aggregates of crystal and liquid. Let him try it. Simply suspend crystals in side 

 of dish and release. His own reference to heavy liquid separations indicates that he 

 has had experience with an operation in which two-phase convection is very 

 common. 



3N. L. Bowen, "Crystallization-Differentiation in Igneous Magmas," Jour. 

 GeoL, XXVII, p. 412. One would certainly infer a doubt from the statement in the 

 earlier paper, p. 12: "The same objections apply to the supposed maintenance of 



approximate uniformity .... through the agency of convection currents " 



Certainly convection would tend to keep the composition uniform, and one would 

 infer that since the magma does not remain uniform he saw no evidence of convection. 



