324 IEE. VOGE 
feldspar and quartz, as is well known, sometimes produces 
micropegmatitic or granophyric structure, as in graphic granite. 
Corresponding structure, which gives an evidence of a crystallization 
along a eutectic boundary curve, or exceptionally by a binary (or 
ternary or still more complex) eutectic, is also sometimes found 
with other minerals, for instance, between olivine and magnetite 
(see Fig. 28). 
In many, possibly in most, cases the crystallization of the 
minerals A, B, C, etc., takes place in the following manner: 
Each mineral begins crystallizing at its proper stage, and continues 
to grow until the entire magma has solidified. As an extreme 
example we may choose apatite. This phosphate is only slightly 
soluble in silicate magmas at temperatures just above that at 
which the silicates commence to crystallize. If there is 0.20 per 
cent apatite present, the essential portion, perhaps 0.18, 0.10, 
or 0.195 per cent, has already crystallized before the silicate 
minerals have commenced segregating. But we may be pretty 
certain that a trifle phosphate, 0.02, 0.01, or perhaps only 0.005 
per cent, still exists in solution at this stage and little by little 
solidifies later. It has, however, not been possible for me to 
substantiate this by observation with respect to the apatite; but 
I have been able to establish that spinel, when present only as 
©.o1 or at most 0.02 per cent, only commenced crystallizing after 
a great part of the silicate mineral A had solidified. (See Fig. 33 
and the chapter on spinel.) 
Fe,O, and the different ferromagnesian silicates are only slightly 
soluble in acid—or granitic—magmas, and therefore commenced 
crystallizing at an early stage. We find, however, as is discussed 
below, a small remnant of magnetite and ferromagnesian silicate 
in the final product of the solidification. We may consequently 
draw the conclusion that the essential part of the magnetite and 
the ferromagnesian silicate was certainly solidified during the 
first stage of the crystallization, but that a little remnant stayed in 
the solution and was solidified later. 
In a binary system, type IV, of two discontinuous mix-crys- 
tals—A, melting at relatively high temperature (for example, FeS,, 
