434 7 He LaVOGT 
We especially call attention to the following: In the magma, so 
extraordinarily rich in ferromagnesian metasilicate, only hyper- 
sthene crystallized at the beginning. Then the formation of this 
mineral stopped, as the ferromagnesian silicate still remaining in 
the magma entered into biotite. The change from hypersthene 
to biotite was probably caused by the quantity of H.O present in 
the magma, and this quantity had been relatively enriched by the 
separation of the large quantity of hypersthene. We shall return to 
this matter in a following chapter. 
Hynersthene 
Ih ITH WV V 
Fic. 9.—Diagram illustrating the different stages of the crystallization of the 
orbicular quartz-norite from Romsaas. 
We further emphasize that from the original magma, so rich in 
ferromagnesian metasilicate, a quartz-dioritic magma was sepa- 
rated at a far-advanced stage of the solidification and there resulted, 
by continuous solidification, at last even a magma for special 
‘“‘oligoclase (or andesine) granite dikes,” consisting of biotite, 
andesine (AbssAn,.), and quartz. We refer to a special chapter in 
Part IT. 
The normal quartz-norite (Fig. 10) from Romsaas, consisting 
of ca. 63 per cent hypersthene (included a little secondary horn- 
blende), 8 per cent biotite, 24 per cent plagioclase (Ab,An,), 4 per 
cent quartz, and, in addition, a little apatite, rutile, and pyrrhotite, 
in part shows accumulation (together-swimming or synneusis struc- 
ture) of hypersthene individuals which often have a well-developed 
idiomorphic contour on their boundary toward the plagioclase or 
quartz. The hypersthene, therefore, must have crystallized com- 
pletely or in a great measure before the plagioclase and the quartz. 
