DOI Kapitel:
Grenzgänger, traders and the last hunter-gatherers of the North European Plain
DOI Kapitel:Anscher, Theo J. ten; Knippenberg, Sebastiaan: Unexpected dimensions of a Swifterbant settlement at Medel-De Roeskamp (the Netherlands)
DOI Seite / Zitierlink:https://doi.org/10.11588/diglit.66745#0173
172
Unexpected dimensions of a Swifterbant settlement at Medel-De Roeskamp (the Netherlands)
or were slightly moved, supporting continuous use of
the southern site. The larger number of structures in its
northern part indicates that this part must have been
occupied the longest. The slightly different orientation
of some southern plans, including the outlier house 1
with a perpendicular orientation, may point to a slight
shifting of the settlement to the south. The overlap in
house plans and differences in orientation indicate
that the house plans represent several phases, each
accounting for perhaps three to five houses. Getting
to grips with the phasing will be a challenging task,
considering the problems connected with a palimpsest
situation and a notorious plateau in the 14C calibra-
tion curve.
The other types of features at Medel are much
less frequent. They include shallow pits and hearths.
Pits have been identified on the basis of their larger
diameter and flat to slightly bowl-shaped bottom. They
often had not been dug in much deeper than the top-
soil layer, so originally they must have had a depth
of around 0.3-0.4 m. Pits are both situated inside as
well as outside the house plans. Within house 2 or 3
a pit was found containing a large part of a ‘Vorrats-
gefaR’ (Fig. 4:16151). It is hard to prove whether these
are the remains of an intentional house offering (as
found in for instance Funnelbeaker [TRB] houses).
They might relate to some functional use as well. It
is noteworthy that the overall number of pits at the
site is small and that deep ones (> 0.6 m) are absent,
indicating that pits were not used for food storage.
A conspicious element is formed by a row of very
large postholes centrally located in the settlement
perpendicular to the gully. The postholes were around
0.5-0.6 m wide and up to 2.0 m deep. They contained
the remaining lower parts of seven massive oak posts
and one made of alder, all showing cutting marks of
stone axes. In alignment with these posts the upright
trunks of two alder trees were uncovered. The fact that
these trees were rooted in a layer below the levee signi-
fies that they already existed before the formation of
the levee, and their alignment with the oak posts sug-
gests that they were still visible when the latter were
erected. It must have been an impressive structure,
presumably with a substantial height. The oak posts
did produce tree ring series of considerable length
(about 120 years), but they could not be matched to
the regional dendrochronological curve. Another row
had been placed around 50 m to the south. Its length
and orientation, as well as its distance to the gully, are
very similar to that of the northern one, though the
southern row consists of paired postholes, less deep
and wide, and not deep enough for the posts to be
preserved.
How to interpret these rows? They do not sur-
round the site, their length is too short and the post
intervals are too wide for a protective fence or wall.
So a defensive function seems unlikely. It couldn’t
have served as a protection against the wind, con-
sidering that west/northwest is the prevailing wind
direction. Since the northern row seems to have in-
corporated trees that had survived the formation of
the levee, it is likely that it belonged to the earliest
habitation. Unsuitable as a defensive structure, the
exceptional post sizes suggest a ritual role rather than
a practical one, perhaps associated with the founding
or structuring of the initial settlement. The parallel
orientation to the houses would be in agreement
with this idea. Whether the southern row of paired
postholes is associated with the assumed younger
southern shift of the settlement mentioned before,
is a tempting thought but again difficult to prove.
The northern site is much smaller and does not
have the high density of posthole features. Their pres-
ence, however, attests to the erection of one or more
structures there too. Unfortunately this cannot be
supported by the identification of a clear house plan.
The preservation in this site is a lot better, probably
due to less intensive and shorter use, resulting in less
trampling. This is exemplified by the presence of two
shallow hearths yielding a wealth of macrobotanical
remains and other finds.
Settlement structure and houseplans
in context
Though lacking the defensive palisades, Medel-De
Roeskamp has a village structure that recalls the ar-
rangements encountered a few centuries earlier in
settlements of the Rossen culture of the Niederrhe-
inische Bucht.28 This resemblance could very well be
the accidental outcome of repeatedly rebuilding and
moving of the house location within a confined area.
A difference, however, is seen in the choice of a new
location when re-building a house. Within Rossen sites,
houses were almost always moved to another location,
whereas houses at Medel were sometimes re-build on
the same spot.
The village arrangement at Medel is probably not
unique for Swifterbant. Whilst at Medel many more
houses could be identified than the two or four houses
at Schokland-P14, large tracts of Schokland-P14 that
28 Rossen sites near Hambach, Inden and Aldenhoven, see for
instance Dohrn-Ihmig 1983.
Unexpected dimensions of a Swifterbant settlement at Medel-De Roeskamp (the Netherlands)
or were slightly moved, supporting continuous use of
the southern site. The larger number of structures in its
northern part indicates that this part must have been
occupied the longest. The slightly different orientation
of some southern plans, including the outlier house 1
with a perpendicular orientation, may point to a slight
shifting of the settlement to the south. The overlap in
house plans and differences in orientation indicate
that the house plans represent several phases, each
accounting for perhaps three to five houses. Getting
to grips with the phasing will be a challenging task,
considering the problems connected with a palimpsest
situation and a notorious plateau in the 14C calibra-
tion curve.
The other types of features at Medel are much
less frequent. They include shallow pits and hearths.
Pits have been identified on the basis of their larger
diameter and flat to slightly bowl-shaped bottom. They
often had not been dug in much deeper than the top-
soil layer, so originally they must have had a depth
of around 0.3-0.4 m. Pits are both situated inside as
well as outside the house plans. Within house 2 or 3
a pit was found containing a large part of a ‘Vorrats-
gefaR’ (Fig. 4:16151). It is hard to prove whether these
are the remains of an intentional house offering (as
found in for instance Funnelbeaker [TRB] houses).
They might relate to some functional use as well. It
is noteworthy that the overall number of pits at the
site is small and that deep ones (> 0.6 m) are absent,
indicating that pits were not used for food storage.
A conspicious element is formed by a row of very
large postholes centrally located in the settlement
perpendicular to the gully. The postholes were around
0.5-0.6 m wide and up to 2.0 m deep. They contained
the remaining lower parts of seven massive oak posts
and one made of alder, all showing cutting marks of
stone axes. In alignment with these posts the upright
trunks of two alder trees were uncovered. The fact that
these trees were rooted in a layer below the levee signi-
fies that they already existed before the formation of
the levee, and their alignment with the oak posts sug-
gests that they were still visible when the latter were
erected. It must have been an impressive structure,
presumably with a substantial height. The oak posts
did produce tree ring series of considerable length
(about 120 years), but they could not be matched to
the regional dendrochronological curve. Another row
had been placed around 50 m to the south. Its length
and orientation, as well as its distance to the gully, are
very similar to that of the northern one, though the
southern row consists of paired postholes, less deep
and wide, and not deep enough for the posts to be
preserved.
How to interpret these rows? They do not sur-
round the site, their length is too short and the post
intervals are too wide for a protective fence or wall.
So a defensive function seems unlikely. It couldn’t
have served as a protection against the wind, con-
sidering that west/northwest is the prevailing wind
direction. Since the northern row seems to have in-
corporated trees that had survived the formation of
the levee, it is likely that it belonged to the earliest
habitation. Unsuitable as a defensive structure, the
exceptional post sizes suggest a ritual role rather than
a practical one, perhaps associated with the founding
or structuring of the initial settlement. The parallel
orientation to the houses would be in agreement
with this idea. Whether the southern row of paired
postholes is associated with the assumed younger
southern shift of the settlement mentioned before,
is a tempting thought but again difficult to prove.
The northern site is much smaller and does not
have the high density of posthole features. Their pres-
ence, however, attests to the erection of one or more
structures there too. Unfortunately this cannot be
supported by the identification of a clear house plan.
The preservation in this site is a lot better, probably
due to less intensive and shorter use, resulting in less
trampling. This is exemplified by the presence of two
shallow hearths yielding a wealth of macrobotanical
remains and other finds.
Settlement structure and houseplans
in context
Though lacking the defensive palisades, Medel-De
Roeskamp has a village structure that recalls the ar-
rangements encountered a few centuries earlier in
settlements of the Rossen culture of the Niederrhe-
inische Bucht.28 This resemblance could very well be
the accidental outcome of repeatedly rebuilding and
moving of the house location within a confined area.
A difference, however, is seen in the choice of a new
location when re-building a house. Within Rossen sites,
houses were almost always moved to another location,
whereas houses at Medel were sometimes re-build on
the same spot.
The village arrangement at Medel is probably not
unique for Swifterbant. Whilst at Medel many more
houses could be identified than the two or four houses
at Schokland-P14, large tracts of Schokland-P14 that
28 Rossen sites near Hambach, Inden and Aldenhoven, see for
instance Dohrn-Ihmig 1983.