Jökull - 01.12.1984, Page 7

minerals. This reflects the state of the magma at
the time of the eruption with the glass having the
same composition as the liquid except the vola-
tiles that have escaped. In the present study only
the glass phase was analyzed using an
ARL-SEMQ microprobe. Only clear siderome-
lane was analyzed, about ten spots 5-10 microns
in diameter, in each thin section. These analyses
are therefore not the same as conventional whole
rock analyses but as the mineral content is low in
the samples, less than 5%, the difference will be
fairly small. The advantage of the glass analyses
is that it represents the liquid composition and is
not affected by differential movement of liquid
and minerals during the eruption. In black
tachylitic glass, oxides have crystallized making
them unsuitable for microprobe analyses but in
most of the samples clear sideromelane is the
dominating glass phase.
DESCRIPTION OF
SAMPLES ANALYZED
The samples analyzed were collected by
various people during an expedition to Gríms-
vötn during the middle of June 1983. They are
listed in an appendix. Most of the samples were
collected from the northern tephra fan in differ-
ent places. Continuous snow cover of the caldera
wall above the lake and the presence of volcanic
bombs but absence of hyaloclastite fragments
makes it unlikely that any of the material in this
fan originates from the caldera cliffs. One sample
comes from the southern air-borne tephra fan
and two from the eastern fan.
For comparison samples available from the
1934, 1922 and 1903 eruptions were also in-
cluded. The eruption in 1934 was a sizable one
with the same main eruption site as in 1983.
Tephra fell in eastern and north-eastern Iceland
(Thórarinsson 1974) and samples collected at six
distant weather stations on Easter Day, April lst
are included. Samples were also collected close to
the eruption site and are included but the exact
date is uncertain.
One dated and one undated samples from the
1922 eruption were also available. In 1903 erup-
tions took place near Thórdarhyrna (Thórarins-
son 1974) and in Grímsvötn (Jóhannesson 1983).
The one sample available from this is most likely
from Grímsvötn. The samples from the 1983
eruption contain only minor amounts of mine-
rals, less than 5%. These are small grains of
mainly plagioclase with additional olivine and
pyroxene. The one exception from this is sample
0013 which contains a noticeably greater amount
of these minerals but this is not reflected in the
composition of the enclosing glass.
The samples from the 1934 eruption are simi-
lar, mainly glass with small amount of the same
minerals. The samples 0024—0029, collected at
the distant weather stations, are all identical with
very fine grained glass and small amount of small
plagioclase but little or no olivine and pyroxene,
possibly due to eolian differentiation. Samples
0018 and 0019 contain mainly black tachylitic
grains with very subordinate amounts of clear
sideromelane so the homogeneity of these sampl-
es is less certain than in the other samples.
In the 1922 samples the same minerals are
found but sample 0007 has markedly higher mine-
ral content which again is not reflected by any
difference in the enclosing glass.
The 1903 sample contains small amounts of
crystals, mainly plagioclase.
CHEMICAL COMPOSITION
OF THE GLASS
The 1983 eruption. All the analyses are listed in
Table 1 and it should be emphasized that these
are analyses of the glass phase but not the whole
rock. All the analyses show very similar chemical
composition. The rock type is an evolved quartz
normative tholeiite and similar tholeiites
characterize many central volcanoes in Iceland.
The chemical composition of the different sam-
ples is very similar but small scale differences
seem to be present. The results for one oxide can
by considered at the time and the variation is no
greater than could be expected if the same sam-
ple was analyzed repeatedly at different times. If,
however, it is considered that all the samples
were analyzed during the same period under very
similar conditions then it becomes more likely
that minor compositional differences are actually
present. If the variation of all the oxides is con-
sidered together certain correlations would be
expected. In other basalt suites certain trends are
generally present and explained by mineral-liquid
control likely to operate during the evolution of
JÖKULL 34. ÁR 5