lOG Vol. XLni , Art. C. — S. Tsnboi : 



material may have been. With a rate of cooling, somewhat 

 quicker or slower than that which gives maximum zoning, the 

 range of zoning is not so great on account of either a moderate 

 degree of undercooling or a partial adjustment of composition 

 between various zones. 



(lY) Crystallization icith Local Collection of Crystals. — In 

 the foregoing discussion, sinking or floating of the crystals has 

 been left out of consideration. The results as outlined above are 

 to be obtained only wlien the relative movement of the crystals 

 in the melt from which tliey are separated is prevented, In the 

 case of slow cooling, if the separated crystals differ from the 

 melt in density, sinking or floating of the crystals would take 

 place whereupon a local collection of them w^ould be accom- 

 plished. In this case, in the part where many crystals have 

 accumulated, only a small amount of liquid is available for the 

 interchange of material between liquid and crystals, so that the 

 crystals would not be subject to a continual indefinite change in 

 their composition toward enrichment in B, but would remain rich 

 n A. 



Crystallization of the Component Minerals. 



The separation of the component minerals of our rocks took 

 place in two distnict generations— before and after the extrusion 

 of the magma. The phenocrysts of olivine, plagioclase, hypersthene 

 (and clino- hypersthene), and augite belong to the intratelluric 

 separation, while the groundmass constituents, plagioclase, augite, 

 and magnetite, are of the effusive stage. A discussion on the course 

 of the crystallization of these component minerals follows :— 



Olivine. — Olivine, which contains no other mineral as inclusion, 

 appears to be the earliest separated mineral. 



