Volcano Oshima, Idzu. 113 



its crystallization was restricted only to the intratelluric stage. 



Hypersthene phenocrysts are often surrounded by augite 

 crystals in parallel intergrowth. From this it is inferred that 

 augite is a later crystallization product than hypersthene. 



Augite is found both as phenocrysts and as a constituent of 

 the groundmass, showing that its separation took place both in the 

 intratelluric and effusive stages. 



Summarizing what has been stated above the following con- 

 clusion is deduced as the most probable course of crystallization 

 of pyroxenes : — 



IVie hypersthene crystals began to separate from the magma after 

 the crystallization of olivine teas completed and in the middle of the 

 crystallization of j^^f^gioclase. The duration of the separation of 

 hypersthene loas comparatively short, and after the cessation {?) of 

 its crystallization, augite began to separate and continued to do sa 

 even in the effusive stage. 



Pyroxenes form mix-crystal series, and the change in their 

 composition during the course of crystallization is exj)ected to have 

 been accompHshed as in the cases of olivine and of plagioclase. 

 The data at hand, however, are insufficient for a discussion on 

 this subject. Tlie only point bearing on it is that in one of the 

 somma lavas, two-pyroxene-basaltic bandaite exposed at Gyôja'^ 

 (3c; p. 74), the phenocrystic augite has the refractive indices: 

 «< 1.6869 1.6869</3< 1.6906 1.7014<r, while the mineral consti- 

 tuting the groundmass has the refractive indices: «< 1.6836 <i? 

 < 1 .6869 < r (p. 76). From this, it is imagined that the augite crystals 

 whose refractive indices are «< 1.6869 1.6869</?< 1.6906 1.7014<r 

 would, by a change in chemical composition, approach those whose 

 refractive indices are «< 1.6836 </3< 1.6869 <^, if the crystallization 



