Archaeologia Islandica - 01.01.2007, Blaðsíða 66

Arne Espelund
Figure 8. The slag heap at Belgsá, indicating also the two pieces of the broken anvil stone. Draw-
ing by the author.
rial, leaving even less traces. Turf and
stones are readily available, while plastic
clay is not common in Iceland. The Ice-
landers have created their own word for
slag: gjall, after the sound that pieces of
slag gave when thrown onto the slag heap
(compare with German Tonerde for clay
applied for potterymaking).
Analyses of ore, iron and slag
In the year 1996 one to two pieces were
taken for analysis from each of the three
sites. Fresh samples of brown ochre were
also obtained in ditches less than 200 m
from the sites Lundur and Viðivellir. At
Belgsá no such ore was visible. The ochre
and also a sample of true ore - found later
- were roasted at 600 °C prior to analysis
in order to avoid ambiguity with respect
to free and combined water and also to
transfer iron to the highest oxide Fe203.
Due to a low ratio between the oxides of
iron and silica, it was soon realized that
the ochre could not have been used as raw
material. Iron in the slag is assumed to be
present as FeO. Among other oxides only
Si02, MnO and A1203, are significant.
The latter is regarded as an amphoteric
diluent. CaO is present around 1%, the
other oxides MgO, BaO, Ti02, P205 and
K20 being well below 1%. However, P205
is of importance even at low values
because it says something about the qual-
ity of the iron produced. In table 6 the
important elements are presented as
oxides. Notice satisfactory sums for the
samples 1, 2, 3 and 5. The slag analyses
in the first four rows are reasonable, with
the following values for R: Viðivellir (V)
3.22, Lundur (L) 2.49, Belgsá (Bl) 2.15
and (B2) 2.58. Ochre ore found at Viðivel-
lir represents a ratio Fe0/Si02 like that of
64