Rensdyrjægere på farten – Opholdssteder fra senistid i Østjylland

Resumé

Reindeer hunters on the move
Small Lateglacial sites in eastern Jutland

The Søvind site is an example of a Late­glacial site discovered during the excavation of much later remains. Given its limited size and depth beneath the soil surface, it is a type of locality that is unlikely ever to have been found by “usual” archaeological methods.

Only scattered Lateglacial finds have been recorded from the central part of eastern Jutland to date (fig. 22). These probably do not provide a true picture of the settlement of the time, but reflect chance situations. The two new sites presented here, Søvind and Elhøj, are therefore very significant.

The Søvind site is situated on a “land bridge” between high ground extending south towards Horsens Fjord, a large east-west oriented valley in Lateglacial times, and higher hills to the north (fig. 1). There are no watercourses or lakes in the vicinity, and its location sets the site apart from the majority of known Lateglacial sites, which are frequently situated by fresh water.

Stratigraphy. Beneath the plough soil was a c. 40-50 cm layer of sand, which became increasingly brownish-yellow and reddish-brown with depth (fig. 2). The flint artefacts were found within this sand layer and in two pits (fig. 3). There were no secondary glacial changes, for example solifluction etc., of the archaeological deposits. The only cultural remains were of flint. The find-bearing deposits were sieved through a fine sieve and delimited in all directions by find-free areas (fig. 3). The flint was found in particular in two pits in the middle of the flint scatter (fig. 3), and outside these the artefact concentration was only c. 1-2 pieces/m². The flint covered an area of c. 9.5 x 4 m (c. 38 m²) (fig. 3). The area of the Søvind site is less than that of other Lateglacial sites, for example Bro I in northwest Funen, which covered an area of c. 50 m², but corresponds to that of the Lateglacial site of Langå I.

Settlement features. In the middle of the flint scatter were two pits, which measured c. 2.5 x 1.3 m and c. 1.5 x 0.8 m, and had a depth of c. 25-35 cm. They lay parallel, c. 1 m apart (fig. 3), and all the indications suggested they were coeval. In addition to their content of flint, the fill in the pits also differed from the surrounding sand layer in that it was more brownish-grey in colour. The pits could either be the result of human activity or have a geological origin, for example as a consequence of windthrows. Their number, their uniform size and form, and the fact that they lie parallel in the middle of the artefact concentration, as well as their high content of worked flint relative to the surrounding surface, are unique in the context of Lateglacial localities in Denmark. There is, therefore, much to suggest that they result from human activity and that they could have functioned as for example caches. No hearths were found, but a little fire-brittled flint shows that there had been fire at the site.

The finds. The finds assemblage from the Søvind site is small and consists of c. 550 pieces of worked flint, of which c. 8.4% comprises tools. None of the pieces shows evidence of frost damage. The flint is of Senon type and it was worked using hard, direct technique.

Cores are represented by five examples, of which four are conical, with a single platform (unipolar), and one is cylindrical/sub-cylindrical with two platforms (bipolar) (figs. 4-6). Many flakes have either reverse traces or two opposing platforms, which suggests that bipolar cores with two opposing platforms were more common than suggested by the composition of the recovered assemblage. One core can be refitted with three successive platform flakes, and a further core has been refitted with a “surplus” flake (figs. 5-6). The assemblage includes core flakes (figs. 8a, f and 9c). Simple flakes number c. 373 examples (fig. 7). They have thick percussion bulbs with a cone, a clear point of percussion, large flat platform remnants and trimming dorsally at the proximal end (fig. 10). Many flakes terminate in a hinge fracture. At Søvind there are 95 blades (16.1% of the assemblage), of which 71 are (large) blades, 24 are micro-blades (or preparation flakes) (fig. 8d-e). Only five flakes are regular in form, while the remainder (66 examples) are very irregular with a fluid transition between the categories (fig. 10). There are no backed blades. The blades show great variation with respect to both width and thickness (fig. 10), and several are “Grossklingen”. Søvind was clearly a locality where production of blades/elongate flakes was important.

With regard to refitting, it has proved possible to match up two core flakes (fig. 9c) and two burins with their burin spalls (fig. 16a-b). The approach employed in working the two cores is also described (fig. 5). The most important discovery is, however, that the pits are associated with the surrounding surface (fig. 3). It has not proved possible to refit flint tools and debitage – perhaps because the tools were not made at Søvind but brought to the site from another locality?

The Søvind assemblage includes 15 scrapers (fig. 12a-f). Seven of these are triangular with a scraper edge at the broadest end (fig. 11c, e), a feature that is not common in the Bromme culture but is frequently seen in assemblages from the Federmesser culture. There are 18 burins, which thereby constitute the commonest tool type and display great variation in type. The majority (16 examples) are made on very thick flakes. Angle burins dominate with nine examples (fig. 15a-c), while dihedral burins are represented by five examples (fig. 13a-c). Further to these are two transverse burins (fig. 13d) and there are two combinations of angle and dihedral burins (fig. 14b-c). Five angle burins are on a break, three are on concave truncation and one example has four corners (fig. 14a). This group includes the example shown in figure 14c, which could be refitted with the percussion-bulb end of the piece it was derived from, thereby giving a good impression of the manufacturing process. Two of the dihedral burins are on a concave truncation, and two are on a burin facet. Three dihedral burins have been produced with two blows, and one example is a combination of a retouched edge and a blow (fig. 13c). Two combinations of angle and dihedral burins are of a type that has not previously been described from a Danish Lateglacial site (fig. 13b-c). Twenty burin spalls show that burins have been heavily used; two of these spalls could be refitted with a burin (fig. 16a-b). A blade point with dorsal retouch (fig. 16d) with a curved back and straight truncation at the percussion-bulb end is a characteristic element of the assemblage. Similar pieces are known from Federmesser settlements. Three notched pieces and two denticulate pieces are also present (fig. 17c). Pieces with a coarse, denticulate edge (fig. 17a-b) (three examples) are unusual types at Danish Lateglacial sites. A curved knife also forms part of the assemblage. Hammer- or crushing stones of stratified porphyry with crush marks at one end are represented by a single example (fig. 7).

The flint-working at Søvind is characterised by hard, direct technique, and the cores are conical or cylindrical with a smooth platform. Blades and flakes have a large, broad, thick percussion bulb and also frequently a hinge. There are no backed blades. The number of tools is small and these are dominated by burins and scrapers; angle burins are the most frequent but dihedral burins are also common. The burin group contains several different types, for example combinations of dihedral and angle burins as well as burins on concave truncation. Further to these is a dorsally-worked blade point, some denticulate pieces and a hammer-/crushing stone.

Dating. The site must be dated typologically, which is difficult as there are no other well-dated assemblages from this part of eastern Jutland. Furthermore, the Søvind assemblage is so small that it cannot be determined whether the absence of types etc. reflects a cultural or a chronological difference, or is simply a consequence of the small number of finds. The stratigraphy suggests a date in the Lateglacial period, and this is supported by the flint technique employed, the tool types, their morphology and the dorsally-worked blade point present in the assemblage. Collectively, the flint technique of the Søvind assemblage points indicates the Bromme culture, while elements of the tool inventory correspond to the ­Federmesser culture. The assemblage can therefore perhaps tentatively be perceived as a mixture of elements from these two cultures which, according to the most recent research, means a date of around 10,900 BC.

Further conclusions. The Søvind site demonstrates several new and important aspects of a Lateglacial site in Denmark. It is small both with respect to area and the amount of worked flint; the composition of the tool inventory is also narrow, being dominated by burins and scrapers. There are no tanged points with a dorsally-worked blade tip. The situation of the site in the landscape is also unusual; and then there are the two kidney-shaped pits. There is no well-defined hearth. All the evidence suggests that the site must represent the result of a very short stay by a small number of people. The settlement activities comprised working of blades and flakes, butchering of prey, scraping of skins/hides, splitting antler and sawing through bone and antler. Then there is the use of fire. Hunting, on the other hand, did not play a prominent role. Perhaps the tools were brought to Søvind, where they were used and then discarded.

In order to gain an impression of how long people stayed at the site, an experimental reproduction of the Søvind flint was undertaken. This showed that it took only 2.5 hours of flint-working to “manufacture” the entire assemblage (fig. 18). This supports the idea of a very brief occupation, but it should not be understood as meaning that Søvind represents precisely 2.5 hours of flint-working and occupation.

If Søvind is compared with the Bromme culture site of Bro I (Andersen 1972), where the remains suggested brief (seasonal) occupation by a single, small social unit, Søvind’s special character becomes clearly apparent. In addition to an apparent chronological difference between the two sites, there are also many structural differences, with Bro I having a circular, slightly sunken feature, numerous finds and a distribution of artefacts in characteristic “activity areas” around a centrally-positioned hearth. There is also a difference in the number of tools and in the form and extent of the finds distribution. In flint-technical terms, the two sites are similar, but the tools at Søvind are thicker (and heavier) than those from Bro I. Within the various tool categories, the form of the scrapers differs, and angle and transverse burins dominate at Bro I, while a combination of angle and dihedral burins, in addition to burins on truncation, characterises Søvind. Then there is a dorsally-worked blade point found at Søvind – a type that is not known from Bro I. Finally, the absence of tanged points from Søvind, in contrast to Bro I, where these form an integrated part of the tool inventory, should be noted, too.

These variations could be due to a chronological difference, but they are much more likely an expression of different social units, subsistence activities and probably also the length of the occupation of the two sites. Bro I must be termed a settlement, while Søvind is a stopping-off place.

Elhøj. About 33 km north of Aarhus, on one of the highest points in eastern Jutland, is a small area which contains blade cores, flint debitage from flint-working in hard, direct technique and a few tanged points and other tools and implements (figs. 19-20). During periods of low, scattered vegetation, the site off­ered a 7-15 km wide vista across this part of Jutland. There are two small “patches” (c. 5 x 5 m and 10 x 20 m) of flint with a slightly higher concentration of flint debitage and tools (fig. 20), but neither of these can be characterised as a “settlement” (fig. 20). No excavations have been undertaken, but close by the flint scatters a minor, unproductive investigation was undertaken of several small bogs containing deposits of Lateglacial character. The artefacts are few in number and include characteristic, coarse tanged points and a dorsally-worked point (fig. 21). Further to these are some blade cores and large, thick flakes which, judging by the working technique employed, must date from the Lateglacial period. The location of the flint scatters only makes sense in an open, sparsely vegetated landscape. The site must therefore represent traces of hunter groups who occasionally took a break here and exploited the view to follow the movements of large wild animals in the landscape. During these stays, new tools and weapons were made and others were repaired or replaced. Elhøj must be interpreted as a lookout point or a hunting station. The flint-working technique and the tanged points indicate a date in the Bromme culture, but if the theoretical standpoint is adopted that all the artefacts derive from the same visit, then the site must rather, due to the dorsally-worked point, be assigned to the Federmesser culture.

Further conclusions. Both Søvind and Elhøj show new aspects of Danish Late­glacial cultures. They are examples of the difficulty and randomness of demonstrating Lateglacial localities. The Søvind site is not a settlement, but represents a short stop in the landscape, during which scrapers and burins were used. Elhøj tells us that single finds of tanged points need not always be lost hunting weapons or evidence of a settlement, but can also represent remains of the contemporaneous hunter population’s exploitation of the open landscape for hunting and lookout positions. If an excavation of the Early Neolithic remains at Søvind had not been undertaken, and an intensive field-walking programme not carried out in the hilly landscape north of Aarhus, it is very unlikely that these two small localities would ever have been found. Current reconnaissance work in the landscape of eastern Jutland is revealing new finds of types from both the Federmesser and Bromme cultures (fig. 22). These discoveries help to underline the fact that the sites found to date probably only reflect a small fraction of the original settlement, which has consisted of a continuum extending from single finds to small “patches” in the style of Søvind and Elhøj, then small settlements such as Bro I and, finally, large settlements such as Bromme.

Søren H. Andersen
Moesgaard Museum