Jökull - 01.12.1986, Síða 40

above, the Maelifell picrite basalt has a whole-rock
composition rich in MgO (26 wt.%), whereas the
quenched glass composition has a low MgO content
(8.8 wt.%). Based on experimental petrology, field evi-
dence and the general primitive characteristics of
picrites, Clarke (1970), Reid & Woodes (1978),
Clarke & O’Hara (1979), Elthon (1979) and Maaloe
& Jakobsson (1980) have made a strong case for the
existence of a high-MgO basic liquid (18-30 wt.%).
Further confirmation of the existence of a high-MgO
liquid has been manifested by the primitive character-
istics of the Maelifell picrite basalt and by the calcu-
lation of the two end-members involved in the mix-
ing; one end-member is inferred to be primitive (MgO
approximately 18 wt.%). Flowever, the Maelifell
whole-rock composition has a much higher MgO
(approximately 26 wt.%) content than do these calcu-
lated end-members (18 and 13 wt.%). This can only
indicate that the magma has picked up olivine pheno-
crysts, as will be shown below. By combining whole-
rock composition, mineralogy, mineral chemistry and
textural evidence, the following appears to be the
most reasonable model to explain the above mixing.
The Maelifell volcano is probably underlain by a
complex plumbing system which allows magma from
the mantle to ascend to the surface. Magma must have
been caught in a trap (magma chambers) on the way
to the surface where cooling and crystallization caused
differentiation. This magma was relatively evolved
and crystallized olivine (Fo86_89) that accumulated at
the base. Plagioclase (An85_89, present as inclusions)
crystallized as well. This resulted in a more evolved
melt remaining in the upper part of the chamber.
Fresh pulses of upper-mantle-derived melt arrived, as
indicated by the primitive characteristics of one of the
ealculated end-members. This melt carried primitive
olivine (FOg9_91), Cr-spinel and Cr-Al-endiopside with
a high Cr content (averaging > 1 wt.% Cr703). Part of
the cumulate-olivine- plagioclase-bearing melt mixed
with this primitive melt, resulting in the occurrence
together of reversely zoned olivine (group I) and
normally zoned olivine (group II), and erupted as
subglacial picrite basalt. It has been suggested that hot
picritic parental liquids are too dense, upon injection
into a magma chamber, to mix directly with the
residual (or evolved) liquid of the chamber (Sparks et
al. 1980, Stolper & Walker 1980, Huppert & Sparks
1980). Several recent publications, including those of
O’Hara & Mathews (1981), Huppert et al. (1982),
Sparks (1983), Bell (1983) and Fitton et al. (1983),
show, however, that mixing must take place with time
and that commonly only part of the fluids mix. Grön-
vold & Óskarsson (pers. comm. 1983) do not find the
reversely zoned olivine in their samples of Maelifell
picrite basalt from a nearby area, which supports mix-
ing on a local scale at Maelifell.
FIELD WORK AND ANALYTICAL
TECHNIQUES
Detailed petrolocical study was carried out on a 4
kg sample. This involved four thin sections, four
microprobe sections and XRF analyses for major and
trace elements.
XRF ANALYSES
INSTRUMENT: Phillips PW1400 spectrometer with
Rh target.
Major elements analyzed using fused beads, dilution
ratio 1:5, diluted with lithium tetraborate. Fused
using Claisse Fluxer. Analysis procedure and correc-
tions outlined in the Geochemistry Laboratory, De-
partment of Geological Sciences, XRF Circular # 1.
ELECTRON MICROPROBE ANALYSES
The mineral and the glass analyses were performed
using the CAMEBAX electron microprobe at McGill
University. Stoichiometric silicates were employed as
standards and ZAF corrections were made on the
data. The accelerating voltage and beam current used
were 15 kV and 5 nÁ respectively. One to twenty
spots (an average of 5) in each crystal were analyzed
depending upon homogeneity. The same standards
used for the minerals were used for the glass analyses.
Following analyses were done on the microprobe; 196
olivine anal. (16 grains of group I, 15 of group II), 19
plagioclase anal., 31 pyroxene anal., 6 spinel anal.
and 6 glass anal.
ACKNO WLEDGEMENTS
This paper summarizes part of main results of the
authors M.Sc thesis. I would like to acknowledge the
help of my professors and advisors R.F. Martin and
D.M. Francis at McGill University Montreal Canada.
Kristján Saemundsson at the National Energy Auth.,
Sveinn P. Jakobsson at the Icelandic Museum of
Natural History, Karl Grönvold and Niels Óskarsson
at the Nordic Volcanological Institute are thanked for
reading an early draft and for many helpful sug-
gestions.
38