Volcano Usliima. Id/u. X07 



There is reason for believing that the magma of Oshima 

 closely matches in density calcic plagioclase crystals (p. 109). 

 Olivine crystals are therefore of much higher specific gravity than 

 the magma from which they were separated ; consequently the 

 importance of their subsidence in the highly fluidal magma during 

 the course of crystallization may be reasonably assumed. 



Olivine forms a mix-crystal series belonging to Type I of 

 Koozeboom's classification^'^ composed of forsterite (Mg.SiO,, the 

 liigher-melting component with the melting point 1800°±20°C.'''') 

 and fayalite (Fe^iSiO^, the low^er- melting component). Following 

 the discussions given on pp. 105-106, the net result of the 

 separation and subsidence of olivine crystals, when the magma 

 cools uniformly throughout tlie whole mass, would be : (1) at a 

 given portion of the magma basin, the earlier the stage is, the 

 richer the olivine crystals there present are in forsterite, and (2) 

 at a given stage, the lower the portion of the magma basin is, the 

 more abundant the oUvine crystals are, and the richer they are in 

 forsterite. Thus, there must be a stage where the forsterite-rich 

 crystals are concentrated in the lower portion of the magma basin, 

 and where only a few olivine crystals less rich in forsterite are in 

 the upper portion. If the subsidence of the olivine crystals goes 

 on still further, the upper portion of the magma basin will at last 

 become entirely free from olivine. 



The olivine in the lavas rich in this mineral is rich in for- 

 sterite, (Fe. SiO,)i.._,,(Mg, SiO,),-_si ; while that in the olivine-poor 

 lavas is less rich in forsterite, (Fe, SiO,),g_3o(Mg, SiO,),,_„ d». 102). 

 This can be explained by supposing that the olivine-rich lavas were 



1) Harkee, The Nuiural History of Igneous Bocks, p. 372, 1909. 



2) BowEX and Axdeusex, "The System MgO-SiO.,," Aw. Jour. 'id. (4), Vol. XXXVII., 

 p. 487, 1914. 



