MAGMATIC DIFFERENTIATION .OF IGNEOUS ROCKS 535 
of MgO in the olivine in some cases, and in the orthopyroxene in 
others—approximately balance, and probably depend chiefly on 
the source of errors connected with the determinations. 
If we pass on to the gabbroidic rocks, we find that hypersthene, 
in the common anchi-eutectic norites, usually shows—as well on the 
basis of the analysis of isolated hypersthene, as by the determina- 
_ tions of the axial angle undertaken by earlier investigators and by 
myself—a composition between about 30 and 38-4o per cent 
FeSiO, (stoechiometric). And the olivine shows, as well on the 
basis of the analyses of isolated material, as on my own determina- 
tions of the axial angle, about 32-35 per cent Fe,SiQ,. 
‘In the hypersthenite-norites (with only relatively little plagio- 
clase) we usually find, however, a relatively lower percentage of 
iron, as well in the rock as in the separated silicate minerals. ‘This 
is discussed more elaborately in Part II. We shall include a couple 
of separate determinations: The thin secretion of a hypersthenite- 
norite, above mentioned (Fig. 25), consisting chiefly of hyper- 
sthene, augite, hornblende, and olivine shows: 
Hypersthene, optically negative, 2V=ca. 80°, gives 25-30 (about 27) 
per cent FeSiO;; 
Olivine, optically negative, 2V=ca. 83°, gives about 30 per cent Fe,SiO,. 
Olivine-carrying norite with only about 4o per cent labradorite 
from Skjeekerdalen (Figs. 20-21): 
Hypersthene, optically negative, 2V=ca. 80-85°, gives about 25 per cent 
FeSi0,; 
Olivine, optically negative, 2V=ca. 85-88°, gives 20-25 per cent Fe,SiO,. 
Also in the igneous rocks containing feldspar, we find approximately 
the same MgO:FeO proportion in both minerals. Any relative 
enrichment of MgO in the olivine is usually not to be found. 
ORTHORHOMBIC AND MONOCLINIC PYROXENE 
As special Fe,O, determinations are lacking in several cases and 
in others are little instructive on account of a later oxidation, 
we in both minerals originate from the entire percentage of iron, 
this giving a quite true image of the relative proportions of MgO 
and FeO in the two minerals. On account of the small percent- 
age of Fe,O, the statements for FeO ought, however, for the 
