TE BASIC: MA SSIVIS ROCKS, 2 TC. 7 
the pyroxene in these rocks are thus very similar in their - 
development to the cellular cordierite, andalusite and orthoclase 
grains described by Salomon’ in contact rocks. The cellular 
structure is thought by this writer to be so characteristic of min- 
erals formed by contact action that he has called it the contact 
structure. In contact rocks it is probably due to the selection 
of material needed in building up the contact minerals, which 
would grow around the substances not needed in their construc- 
tion and thus include them. In the present case the cellular 
structure of the hypersthene must be an original one, and the direct 
consequence of the conditions under which the minerals separated 
from the rock magma. The plagioclase, though included within 
the hypersthene, is probably the younger of the two minerals, 
just as the ice, in a frozen sponge saturated with water, is younger 
than the mass of the sponge which encloses it. 
A third form of structure is produced by the aggregation of 
the hypersthene plates into groups or masses that are imbedded 
in the granulitic matrix from which hypersthene is wanting. The 
masses or groups are composed of large irregular grains of the 
orthorhombic pyroxene, sometimes cellular, sometimes compact, 
with which are intermingled many grains of diallage and an occa- 
sional one of plagioclase. Figure 2 shows a portion of one of 
these aggregations in the upper part of the picture. In its lower 
portion is the granulitic mosaic of plagioclase and diallage. 
_ The diallage in all these rocks is in the rounded granulitic 
forms, except in the rare instances when it is associated with the 
hypersthene in the aggregates, where it is in large irregular grains. 
In certain of the non-hypersthenic varieties it is sometimes in 
cellular plates, but in those containing the orthorhombic pyrox- 
ene the diallage, when it occurs, is nearly always granulitic. In 
the figure last referred to the small, highly refractive, rounded 
grains are of this mineral. In nearly all cases the diallage is 
light green in color, is fresh and is without pleochroism. Its 
inclusions are magnetite grains, small flakes of brown mica and 
tiny particles of limonite. 
* Zeits, d. d. geol. Ges., XLII., 1890, pp. 487, 511, e seg. 
