Acta naturalia Islandica - 01.07.1964, Blaðsíða 20

18
SIGURDUR STEINTHORSSON
closes many entire plagioclase and pyroxene grains, or it may occur as discrete
crystals.
Even on a microscopic scale the pyroxene grains are fairly evenly distribu-
ted over the groundmass. The pyroxene phenocrysts, however, are sometimes
surrounded by a narrow zone rich in plagioclase and free of pyroxene grains.
(See e. g. Pl. II b). The plagioclases are then orientated parallel to the
edges of the phenocryst. It appears that in this zone the liquid was empover-
ished in pyroxene, which was being concentrated to precipitate on the already
formed phenocryst nucleus. This effect is hardly seen at all around the olivine
phenocrysts, nor would it be expected there anyway, because olivine is much
sparser in the groundmass, and crystallizes quickly to form discrete grains
of fair size.
In some sections the plagioclase laths are aligned in parallelism. The
granular texture may then be due to movement: the mush of already crystal-
lized olivines and plagioclases was still flowing whilst the pyroxene was crystal-
lizing. Against this cataclastic hypothesis appears to be the fact that the granular
groundmass is sometimes seen sitting in a cavity of phenocrysts (Pl. III b).
It is hard to visualize how the granular mass came to be inside the pheno-
cryst, for there is no reason why the crystals forming in the cavity should
not be precipitated from the liquid already there, instead of being brought
in from the turmoil outside. (It is assumed that the cavity was present in the
phenocryst when the groundmass began to crystallize).
Macdonald (1944) has described picritic basalt from Hawaii with horn-
fels-like groundmass of augite, hypersthene, labradorite and iron ore. The
texture is attributed to recrystallization of the groundmass in response to
thermal metamorphism. This explanation is hardly applicable for the Hvamms-
múli rocks, because the two textures, granular and ophitic, are frequently seen
associated in the same thin section, and showing no intrusive relationships
(Pl. V b). It is conceivable in such cases that the two types represent some
sort of banding governed by the movement of the crystallizing magma during
emplacement. As shown elsewhere, (next section), the flow was laminar, and
flow units may have formed where relatively unaffected units Were bounded
by planes of movement (the granular bands). Such relationship would be ex-
pected to be confused: an irregular pressure from behind would soon break
up the banded relationship, fold the mass and chew it about causing partial
remixing of the two units.
Composition seems to play a very important role as well in this connection,
because the granularity increases with pyroxene content (F. Walker 1957).
The sub-ophitic texture is characterized by optically continous pyroxene