Jökull - 01.12.2006, Qupperneq 66

David W. McGarvie et al.
879 m and rises from a base level at 580 m. Sam-
ple TJ97-22 was collected from the base of the lava
cap, and is a porphyritic obsidian containing 7% phe-
nocrysts. Phenocrysts are dominated by euhedral and
unzoned feldspars (anorthoclase) up to 4 mm in size
but mostly around 1–2 mm, often forming clusters up
to 7 mm in size; other mineral phases are ferroaugite,
fayalitic olivine (Fo1), and ilmenite that occur in mi-
nor amounts and do not exceed 0.5 mm in size, all
set in a flow-banded and microlite-rich glass (A. G.
Tindle, unpublished data). This sample is a mildly-
peralkaline rhyolite (agpaitic index of 1.2) and is a
comendite according to the classification scheme of
Macdonald (1974). The base-summit height differ-
ence of 300 m suggests that the ice sheet was at least
this thick at the time of eruption. This sample has an
Ar-Ar age of 278±18 ka.
Older peralkaline rhyolites
Two samples were collected from the eastern rim of
the suspected caldera (Figure 2). Sample TJ97-14 was
collected from the eruptive unit that forms the high-
est ground above the eastern caldera rim (the moun-
tain Hábarmur, 1199 m). This unit is around c. 40 m
thick, comprises a number of lava bodies up to 200
m in size, and drapes older and more-altered rhyo-
lites of unknown provenance. The presence of perva-
sive columnar jointing in the lava bodies with promi-
nent obsidian margins suggests interaction with water
(Tuffen et al., 2001; Stevenson et al., 2006), which
at this elevation is most likely to have been derived
from melting of snow or thin ice (Lescinsky and Fink,
2000). Sample TJ97-14 is an aphyric black obsidian,
which is flow-banded and contains scattered micro-
lites that are mostly aligned parallel to flow bands.
This sample is a peralkaline rhyolite (agpaitic index
of 1.3), is a pantellerite according to the classification
scheme of Macdonald (1974), and has an Ar-Ar age
of 384±20 ka.
Sample TJ97-9 was collected c. 1.5 km to the
north of the Hábarmur summit, from a separate (c. 30
m thick) eruptive unit that forms a short (c. 300 m)
WNW-ESE oriented ridge which overlies older, al-
tered rhyolite of unknown provenance (Figure 2). The
base of the unit is composed largely of microcrys-
talline grey-yellow rhyolite whilst its top is domi-
nated by strongly flow-banded, green-black obsid-
ian and obsidian-dominated breccias. Whilst this unit
may be an eroded subaerial lava flow that has lost
its pumiceous carapace, the abundance of obsidian
together with the presence of columnar-jointed lava
bodies suggests interaction with water (cf. Lescin-
sky and Fink, 2000; Edwards et al., 2002: Steven-
son et al., 2006), in an environment similar to that of
sample TJ97-14, where erupting lava may have inter-
acted with snow or thin ice. Sample TJ97-9 is green
and black, aphyric obsidian which is strongly flow
banded and contains sparse microlites which are gen-
erally aligned parallel to flow bands. This sample is a
mildly-peralkaline rhyolite (agpaitic index of 1.1), is
a comendite according to the classification scheme of
Macdonald (1974), and has an Ar-Ar age of 83±6 ka.
DISCUSSION
Ring fracture rhyolites
In the absence of absolute ages, a tentative eruption
age for the ring fracture rhyolites of c. 80 ka (Mc-
Garvie, 1984; McGarvie et al., 1990) was based on
two assumptions: firstly that good preservation of the
glaciovolcanic edifices implied eruption during the
most recent (Weichselian) glacial period; and sec-
ondly that the mild erosion experienced by the edi-
fices (such as smoothed surfaces on some summit lava
caps) was caused by the later and thicker ice of the last
glacial maximum (c. 21–25 ka).
Samples from two widely-separated tuyas were
dated (Table 1): Illihnúkur in the east (72±7 ka), and
SW Rauðfossafjöll in the west (67±9 ka). There is
substantial overlap between the two ages, but the ap-
parent 5 ka age difference is less than either of the
two uncertainties. Consequently, it is reasonable to
suggest that these two tuyas were erupted simultane-
ously. However as the Ar-Ar ages for these two sam-
ples were determined on feldspar (anorthoclase) sep-
arates, these two ages only represent eruption ages if
these feldspars crystallized immediately prior to erup-
tion.
Petrographic evidence such as euhedral shapes
and lack of zoning (McGarvie, 1985) suggests that
the feldspars have not experienced variable thermal
64 JÖKULL No. 56