MAGMATIC DIFFERENTIATION OF IGNEOUS ROCKS 239 



were present in smaller quantities. In addition to this, as the 

 crucial point is the fact that basic magmas with relatively small 

 or no content of Hght volatile compounds are considerably more 

 fluid than are the acid with the same amounts. Thus the basic 

 magmas, even if they were much less extended in the deep-seated 

 basins than were the acid, must reach the surface in much greater 

 quantity and form much larger flows. 



The anorthosites and the almost pure olivine rocks (dunites), 

 according to my opinion, must be explained by the extremely high 

 temperature of crystallization of these rocks. The magmas would 

 consequently be cooled at once to quite advanced crystallization 

 as they were pressed as dikes into the surrounding colder rocks. 

 Thus the magmas of these rocks could not get to the surface and 

 could consequently not form flows. 



As treated in my work Die Sulfid-Silikatschmelzldsungen (191 7), 

 molten FeS, alone or with some NiS, PbS, CU2S, etc., is characterized 

 by being considerably thinner than even the thinnest silicate melts. 

 As to the geologic consequences of this, I refer to my earlier publica- 

 tions on the genesis of the nickel-pyrrhotite deposits. 



ON THE SIZE OF CRYSTALS AND THE RATE OF CRYSTALLIZATION 



Many investigators in their synthetic experiments have confined 

 themselves to melts in a platinum crucible with a comparatively 

 small quantity of siKcate as, e.g., 10-20 gm. (or about 4-8 cm.^). 

 In consequence only quite small individuals have resulted, thereby 

 producing a wrong conception of a slow rate of crystallization for 

 the rock-forming minerals.^ 



If we wish to obtain from experiments a correct idea as to the 

 size of the minerals and the rate of crystallization, we must operate 

 with sufficiently large molten masses. Even with a duration of 

 crystallization of only half an hour we may get very large crystals, 



' For example we may quote C. Doelter, Phys. chem. Mineralogie (1905), p. 108: 

 "Unter normalen Verbal tnissen wurde ein etwa 10 bis 20 mm. langer Augitkrystall 

 unter der Voraussetzung dass die Krystallizationsgeschwindigkeit pro i Minute o.ooi- 

 0.002 mm., wie es die meisten Versuche ergeben, betragt, 200 Stunden brauchen." 

 In fact, the rate of crystallization for augite (diopside) from a pure melt, as will be 

 discussed below, is about i mm. (along the c axis) per minute, thus 500-1,000 as much 

 as indicated by Doelter. 



