Archaeologia Islandica - 01.01.2013, Blaðsíða 60

MAGNUS HELLQVIST
organic content. The organic content may
also be a consequence of, and increased
by, the indoor activities in the house and
the generally good preservation. For
example; it was possible to visually
identify dung remains during sampling.
The floor deposit exposed during
excavation provided few visible or
sedimentological indications about its
suitability for sampling, so the strategy
during 2002-2003 was to test different
ways of sampling to guide later work.
These strategies varied between floors in
different rooms and several methods were
used, including: i) whole extant floor
sections; ii) ‘chess-board’ arranged
pattems of extant floor sections; and iii)
specified sampling of visually distinctive
parts of the house remnants such as the
floor-wall borders, thresholds etc., both
from a spatial point of view and also
connected to different types of
constmctions in and parts of the former
buildings. When sampling sections of
extant floor remnants, the method was to
collect “lumps” of soil from for squares of
a gridded house floor, to avoid damaging
subfossil remains of plants
(Guðmundsson, 2009). The sampling was
focused when there was any sign or
indication of room usage (e.g. storage or
kitchen), or special constraction details
(e.g. wood constmction or walls).
The method of sampling contexts like
floors in their entirety is also described
and discussed by Guðmundsson (2009) in
investigations at Flofstaðir in north-eastem
Iceland. Guðmundsson justifies this as the
best way to extract large and satisfactory
numbers of plant remains even though this
means that more time must be used for
processing samples. Sampling extant
sections or ‘chess-board’ pattem sampling
of floors in Hólar gave similar results, in
that it generated extremely large volumes
of sediment that were time-consuming to
process.
Samples for several analytical
purposes, primarily for insects and plant
macrofossils, were collected and stored in
5 to 10 litre buckets, where each bucket
(or, in the case of extant floor sampling,
buckets) contained samples from each
square of a grid. The volume (>5 litre) of
soil samples was to ensure that high
enough numbers of insect and plant
remains were, as discussed by
Konradsdottir (2009) who noticed a
relationship between sample size and
number of insect remains found, even
though this connection was not defínite.
Large sample volumes are especially
important when sampling floor deposits,
but there was no clear connection between
sample volume and number of fínds in the
Hólar material.
The samples were processed at the
same time for macrofossil plant and insect
remains. The used method of water
sieving follows Watson (1976), Wagner
(1982), and Gumerman and Umemoto
(1987). The paraffin method described by
Coope (1986) was not used, both because
the water sieving was effective and also to
avoid any chemicals. Samples of insect
and plant remains were prepared from the
same sediment samples and this collection
technique was one of the more important
strategies in the work at Hólar.
Macroscopic remains of insect and plant
remains were sorted from the sieve residue
under a stereomicroscope at low
58