MAGMATIC DIFFERENTIATION OF IGNEOUS ROCKS 249 



This review proves, in spite of the sources of error connected 

 with the computation, especially of the first two rows of figures, 

 that the "first" phenocrysts in the andesites, etc., here considered 

 never contain quite as much An as the theoretical calculation 

 requires. The phenocrysts indeed, in many cases, may be zonal; 

 they may consequently contain in the kernel a little more An than 

 the chemical analyses of the whole phenocrysts show. But even 

 if we take this into consideration, the kernel of the phenocrysts 

 will always or almost always contain not quite as much An as the 

 theoretical calculation requires. That is to say, at the beginning of 

 crystallization, as long as the phenocrysts were extremely small, 

 the equilibrium between the solid and the fluid Ab+An phase was 

 very incomplete, but not absolutely lacking. Under the microscope, 



we observe that the zonal structure of the plagioclases, as well as 

 of the pyroxenes, etc., is in general more pronounced in the periphery 

 than in the kernel of the individuals. 



Concerning the dikes and flows, we may thus maintain as the 

 crucial point that the equilibrium between the solid and the fluid 

 mix-crystal phase was lacking to a very great extent. At the 

 beginning of the crystallization, as long as the crystals were quite 

 small, and at a stage with a relatively high temperature and conse- 

 quently with a moderate viscosity of the magma, we may indeed 

 assume a limited equilibrium. But at the later stages of the 

 crystallization, viz., with larger crystals and at lower temperature 

 and a higher degree of viscosity of the residual magma, only 

 very incomplete equilibrium occurred. 



As well known, we occasionally observe in the plagioclases of 

 dikes and flows a recurrence of zonal structure. For example, 

 around a series of zones, showing decreasing An, there may suddenly 

 occur a zone with more An and less Ab. This probably depends 



