LATER STAGES OF EVOLUTION OF IGNEOUS ROCKS 13 



If either free diffusion or convection toward the cool boundary 

 were the effective agent in producing the concentration in that 

 region of the minerals of early crystallization, we should expect 

 large crystals of these minerals attached directly to the wall rock, or 

 if the crystals were of small size they should be attached to each 

 other and ultimately to the wall rock. In nearly all cases these 

 early minerals occur, on the contrary, in small discrete crystals 

 separated from each other by crystals of later formation which 

 must have been a part of the liquid magma at the time the supposed 

 diffusive or convective processes were taking place. At this time, 

 then, they could not have been attached either directly or indirectly 

 to the wall rock and there is no reason why they should have 

 remained in the border position. 



For the reasons above outlined, the processes of diffusion or 

 convection during crystallization cannot in general be considered 

 satisfactory explanations of "basic border phases." Diffusion 

 toward the cold boundary prior to crystallization, the Soret action, 

 has also been found wanting. Daly^ has offered an explanation 

 of basic border phases which seems to be entirely satisfactory in 

 most cases. He pictures the border phase as a quickly chilled 

 portion having the composition of the original magma and the 

 more saKc phase adjacent to it as the lighter differentiate formed 

 during the slow cooling of the rest of the magma.^ For stocks, 

 batholiths, and large laccoKths, this process has distinct advantages 

 over a process involving diffusion toward a chilled border. Daly 

 has in this manner successfully explained that classic example of a 

 body exhibiting a basic border, the Shonkin Sag laccolith. 



In the case of small bodies like most dykes, it seems to the 

 writer to be more commonly true, for those having basic border 

 phases, that the dyke fissure serves as a channel for the passage of 

 magma during a considerable period. As the chilled layer forming 

 on the cold walls gradually increases in thickness, the composition 

 of the magma passing through gradually becomes more acid on 

 account of a like change in the composition of the magma in the 



^ Daly, Igneous Rocks and Their Origin, p. 244, Fig. 123, and p. 246, Fig. 126. 

 ^ In a later part of the paper the manner in which quick cooling prevents differen- 

 tiation and slow coohng promotes it will be shown in detail. 



