MAGMATIC DIFFERENTIATION OF IGNEOUS ROCKS 523 
gations of zonal olivine apply to the common rock-forming olivine, 
with at most 35-40 per cent Fe,SiO, (and rest Mg,SiO,). 
c) The earlier approximate determinations, especially by 
Deelter, show for olivine with increasing percentages of Fe.SiO,, 
continuously decreasing ‘‘melting-points” (o:melting-point in- 
terval). 
d) The later investigations prove inter alia that even so small a 
percentage of Fe,SiO, as 10-15 per cent lowers the melting-point 
interval of the olivine considerably below the melting-point of 
Mg,siO,. A maximum of the melting-curve, which in this case 
must have occurred at predominant Mg,SiO, and little Fe,SiO,, is 
therefore out of the question. 
In conformity with all the above observations I believe myself 
justified in illustrating the binary system Mg,SiO,:Fe.SiO, by 
the foregoing sketch (Fig. 22); it must be emphasized, however, 
that the course of the curve is only sketched.’ As the zonal 
structure of the olivine is far less prominent and there is less differ- 
ence between the two components than in the plagioclase in rocks 
with about the same cooling-rate, the difference between the 
liquidus and solidus curves must be less for Mg,SiO,: Fe,SiO, than 
for An: Ab. 
OLIVINE AND FPLAGIOCLASE 
As explained in my earlier treatise ‘‘Die Silikatschmelzlés- 
ungen,”’ I and II (1903-4), olivine and anorthite do not crystallize 
at the pressure of one atmosphere in molten masses of certain 
intermediate mixtures of Mg,SiO, and CaAl,Si,O¢, for here new 
minerals are formed, chiefly melilite and spinel. In conformity 
with this, in the treatise above cited, I discussed for basic silicate 
melts with anorthite and olivine as the two extremes, not the 
individualization boundary between olivine and anorthite, but 
between (a) olivine and melilite, and (6) melilite and anorthite. 
O. Andersen (at that time with the Geophysical Laboratory, 
Washington), in his treatise ‘‘The System Anorthite: Forster- 
ite: Silica, ’”? arrived at the same results, for by intermediate mixtures 
of An and Mg,SiO,—viz., between 90 An:10 Mg,SiO, and 54 
An:46 Mg,SiO,—he obtained crystallized spinel. At high pressure, 
t A minimum in the neighborhood of Fe,SiO, is not excluded but very improbable. 
eRIFOG ICI. 
