MAGMATIC DIFFERENTIATION OF IGNEOUS ROCKS 347 
while the feldspar first commenced crystallizing at a somewhat 
later stage. The sequence of the commencement of the crystalliza- 
tion in the granites is in most cases (1) pyrite, zirkon, apatite, etc.; 
(2) iron ore and ferromagnesian silicate; (3) feldspar; (4) quartz. 
But it appears from the structure that the ferromagnesian silicate, 
especially biotite, continued crystallizing after the commencement 
of the solidification of both feldspar and quartz, and that the 
feldspar continued crystallizing also during the segregation of the 
quartz. In most of the granites, however, we are unable to deter- 
mine with accuracy from the structure, the quantitative propor- 
tions of feldspar and quartz during the intermediate and later 
stages of the crystallization. 
The case is complicated by the fact that the granite 
magma contains, besides the usual ferromagnesian silicate, Or, 
Ab, An, and Qu components, some H.,O, probably partly entering 
into a SiO, combination, for example, as H,SiO, (?), and the 
latter was not split up until a later stage of the crystallization 
period. If this supposition is correct, the consequence will be a 
somewhat reduced quantity of the independent quartz component 
during the first part of the crystallization period—that is to say, 
during the first part of the crystallization the feldspar was rela- 
tively more abundant than that corresponding to the proportion 
calculated from the relation between feldspar and quartz in the 
resulting solid rock. 
We have an instructive orientation on the composition of granite 
magmas at a late stage of the solidification in the composition of 
the intervening masses between basic concretions, or orbicules, 
in granites, which show these structural elements. (See analyses 
Nos. 23¢-29¢.) These intervening masses prove throughout that 
during the crystallization a displacement of the composition of 
the magma remainder took place in the direction of the—in other 
ways determined—‘‘granite eutectic,’ and we especially empha- 
size that the analyses of the intervening masses Nos. 2 BG, QEGoy Biol 
26¢c almost exactly correspond with the “‘granitic eutectic.” 
Above we have only considered granites with relatively basic 
concretions, or orbicules. But from strongly acid granites, with 
about 78-80 per cent SiO, in the whole rock, we know a couple of 
