Archaeologia Islandica - 01.01.2002, Qupperneq 83

Brattahlíð reconsidered
guish from the later ones.
Keeping these insecurities in mind, it
is still believed, that the otherwise quite
uniform and intact farm sites can be com-
pared with relatively reliable results,
even though some "noise" must be
expected. At least, this approach by far
outweighs the alternative, namely
Roussell's outdated classifications.
As noted, the total survey material
includes almost 450 sites, or ruin groups,
in the Eastern Settlement. Only about
half of these are actual farms, the rest
represents small sites and saeters (shiel-
ing). The distribution of these is shown in
figure 2.
The demarcation of the individual
farms is simple, since most sites are situ-
ated two to six kilometres apart. In some
cases, more than one site is found within
a smaller area. In this test, these are
regarded as being holdings of the same
farm (see Keller 1989, 113).
The calculations of the total floor area
are based on extemal measurements,
since this is how the ruins are described
in the survey material. Dikes and pens
are discarded, as they are roofless struc-
tures. The floor areas of the churches are
not included either, which makes all the
farm sites more equally comparable.
However, this has no significant efifect on
the calculations.
When studying the survey material, it
appeared that there was no clear demar-
cation between larger and smaller farm
sites, as already noted by Roussell (1941,
30). A dividing number had to be chosen
to limit the number of farms in the test
and a 2000 square metre limit appeared
to be a good choice.
A total of 37 farm sites qualified for
the test, having either a church or a total
floor area larger than 2000 square metres.
The farm sites, arranged according to
size, are shown in fígure 3. Due to the
insecurities of the test, the succession of
the sites should not be taken too serious-
ly. On the other hand, the test shows
some obvious, yet also unexpected
results:
1. The farms with parish churches are
generally not among the largest in the
settlement. A number of other farms
without churches are considerably larger.
These farms seem to be better situated
too, when studying the sites in detail.
Furthermore, in most cases byres and
dwellings are larger on these sites as
well.
2. Roussell’s model of the social struc-
ture, placing Garðar (0 47) and
Brattahlíð (0 29/29a) at the top closely
followed by the other church farms,
should no longer be trusted. These farms
do not form a seperate class of major
farms. The social structure must be more
complex than this.
3. The largest of all the Norse farm sites
was not any of the classical sites, but a
rather anonymous site (number 0 39) sit-
uated in Qinngua in the Tunulliarfik fjord
- the former Eiríksfjörður. Could it be
possible, that Jónsson and Norlund were
mistaken, when placing Brattahlíð at
Qassiarsuk?
81