Jökull - 01.12.1985, Qupperneq 49

and readvance stages, from a central ice mass and from
valleys as large as Fljótsdalur and Skriðdalur, should
have left much more obvious traces in the form of
terminal moraines than exist anywhere in the district
now, even considering the effects of weathering and
erosion in post-Glacial time. Some possible reasons for
their absence are examined below.
the evidence of the giljahólar
These hills and ridges of deposited material are well
described by Hjartarson et al. (1981, p. 78). They lie on
the E side of Jökuldalur, (Fig. 1), in a fairly narrow part
of the valley where the Jökulsá á Brú is well entrenched,
(Fig. 14). A little upstream lie the Hauksstaðamelar, a
rather similar group of hills but on the other side of the
river. It is possible that the two groups of hills were once
one system, now divided and broken up by the river,
but this does not seem to be the case since the Giljahól-
ar climb up the E valleyside, while the Hauksstaðamelar
lie in the middle or W of the valley and are too far
upstream to be justifiably considered as part of the
other group. One of the hills of the Hauksstaðamelar
has been excavated for road material and the exposed
section seems to be that of an esker (Fig. 5).
In overall appearance the Giljahólar have the
morphology of moraine but in detail there are major
discrepancies. For instance, although there are a shall-
ow pond and several hollows similar to kettle holes, the
hills are rounded and smooth in appearance and all
loose material is rounded pebbles and cobbles. When
viewed in plan, the hills tend to form sinuous ridges
running up or down the valleyside, one or two forming
transverse lines for several hundreds of metres. The
suggestion óf Hjartarson et al. (1981) that these are
eskers is almost certainly correct.
A close examination of the western part of the feature
showed only one section of the deposits, a sandpit by
the road opposite the homefield of the Gil farm.
Although this exposure shows some layered sands and
loose rounded stones, it is part of the bank of a small
stream and may not be representative of the material
higher up. The amplitude of these hills is about 100 m
from base by the river to top at 207 m, but this partly
represents the slope of the hillside on which they rest.
However the highest of the Hauksstaðamelar rises to a
similar height of 204 m.
The Giljahólar appear to have been derived from the
hillside above. In this hillside lies a rock-cut trench
which has been mentioned above, in the edge of which
may be traces of a small glacier tongue. At its other end
the trench is connected with the basin of Sandvatn stóra
on the Fellaheiði plateau top, (Fig. 1), by a system of
dry channels over the ridge in the middle of the heath
with sandy deltas on the W side and it is possible that
the Giljahólar were built up by material carried in
subglacial flow across Fellaheiði. As supporting evi-
dence there is the ending of some of the channels on the
E side of the plateau against its upper edge, (Fig. 2),
and some evidence from deposits on the E plateau
surface. These were found near the first bridge on the
new Snaefell road from Fljótsdalur over Fljótsdalsheiði,
(Fig. 1), and show a line of deposits running from
Garðavatn N to NE over Bessastaðaá for many
kilometres, (Fig. 16a), with samples showing layered
sands and rounded stones, and giving a well-sorted
analysis, (No. 80/4 in Fig. 6). Excavations show these
rounded hills, up to 6 m high, to be eskers, (Fig. 16b).
Many of the other features in lower Jökuldalur are
attributed by Hjartarson et al. (1981) to water activity
caused by jökulhlaups from lakes dammed up by local
glacier readvance. Whilst, clearly, Jökuldalur and its
environs have been affected by large-scale regional
glaciation, there are no definite signs of readvance
glaciers and the phenomena in the valley can be
explained in other ways. The valley itself is a complex
one, and the upper part, near and above the junction
with Gilsá, is a typical river valley, while the present
Gilsá valley may have at one time carried the main
stream. The enormous terraces of Jökulsá á Brú near
Brúarjökull and map evidence of a high-level valley
with long lakes, which near the ring road have impress-
ive terrace systems themselves, suggest that Jökulsá, at
an early stage, flowed into Vopnafjörður, perhaps sub-
glacially, in a route further to the W than at present.
Similarly, there is evidence of considerable water flow
in channels above the W edge of Jökuldalur from Hof-
teigur northeastwards, some of the channels cutting
through spurs, and a group above Fossvellir in Jökuls-
árhlíð containing very large eskers, (Fig. 17). Varying
amounts of rounded material, some quite large, havé
been deposited in Jökuldalur by apparent flow from S
or SW tributary valleys.
IMPLICATIONS
It is suggested in this paper that there are no definite
signs, attributable to any specific date, of stadial glacial
readvance in Fljótsdalur or lower Jökuldalur, nor are
there any clear signs of marine incursion into Fljóts-
dalur above Lagarfoss. The broad range of altitudes in
which deposits occur seems to eliminate any concept of
precise base-level. A recent study of the broad Skaga-
fjörður valley, in central N Iceland, with open access to
the central plateau, (Víkingsson, 1978), indicated that
there were no signs of large end moraines but numbers
JÖKULL 35. ÁR 47