MAGMATIC DIFFERENTIATION OF IGNEOUS ROCKS 247 



I.I kg., on the other hand, there was found to be (by more parallel 

 measurements) an undercooling of 30°-4o'^. In very thin silicate 

 melts of sufficiently large quantity and with sufficiently slow cooling 

 so that the crystallization lasts at least one or several hours, the 

 crystallization, no doubt throughout, takes place with only an ex- 

 tremely small degree of undercooling. In very viscous melts, on the 

 other hand, the undercooling may play an extraordinarily important 

 part, even when very large molten masses are involved. 



In normal striated deep-seated rocks I do not know of any struc- 

 tural element which indicates any traceable supersaturation. This 

 must depend on the fact that they were cooled so slowly that the crys- 

 tallization took place with quite insignificant supersaturation. To a 

 certain degree it is otherwise with orbicular granite, diorite, etc., 

 having concentric orbs. Here the successive shells from the center 

 to the periphery (as I have shown in Tscherm. Min. Petrogr. Mitt., 

 XXV (1906), 396-403) may be explained under the supposition that 

 the crystallization first began after some undercoohng, especially 

 along the eutectic boundary lines. This orbicular structure of the 

 deep-seated rocks occurs as a rule only as a border phenomenon 

 near adjacent older rocks, and this must be due to the fact that the 

 cooUng here took place somewhat quicker than in the central 

 parts of the eruptive fields. 



In the still more quickly cooled dikes and effusives, the super- 

 saturation generally has played a certain role, as is indicated by 

 partial resortion on account of supersaturation (see a following 

 chapter). In very viscous dikes and effusives rich in silica, the 

 supersaturation may have been of very considerable importance. 



ON THE EQUILIBRIUM BETWEEN THE SOLID AND THE LIQUID PHASE OF 

 MIX-CRYSTAL COMPONENTS 



In dikes and effusives, the plagioclases, pyroxenes, etc., may show, 

 as is well known, a zonal structure, and the ''second generation" of the 

 plagioclase has as a rule a different composition (a higher percentage 

 of Ab) from the "first generation." This comes from an absolutely 

 lacking or a more or less incomplete equilibrium between the segre- 

 gated mix-crystals (or the first "embryonal" mix-crystals together 



