Brovst, en kystboplads fra ældre stenalder

Resumé

Brovst

Settlements of the Kongemose Culture have previously only been recorded in Zealand. Sporadic single finds, for example the flint-edge dagger from Flynderhage (see KUML 1969), which was assigned on typological grounds to the Kongemose Culture, have been the only representatives of this culture in Jutland.

In the summer of 1969, three settlements which could be assigned to the Kongemose Culture were discovered in northern Jutland: at Brovst and on the island of Mors. Excavation of these sites is still in progress and the present article is a preliminary report from Brovst.

Topography

The settlement is situated west of a small morainic hill which was in Atlantic times an island in a broad reach of the Limfjord. This part of the fjord was sheltered from the Skagerrak by a late glacial beach ridge, fig. 1. Geological investigations have shown that the area was a freshwater lake before it was inundated by the sea in Atlantic times and transformed into a shallow fjord.

In order to gain a general impression of the nature, extent and stratigraphy of the settlement, a series of trial pits, an E-W section and a small area of the settlement were excavated, fig. 2.

Stratigraphy

The E-W section is representative of the stratigraphy of the settlement, fig. 3.

Immediately above bed rock, layer 1, lies a thin, black, marine clay containing a great deal of charcoal and numerous artefacts of flint, bone and antler, layer 2. To the west, this layer contains more sand, while the cultural material decreases. On account of this increasing sand, bone and antler were found only in the most easterly parts of the deposit. To the east, the flint had a white patina, but showed no signs of water wear. To the west, the degree of patination decreased, and in the most westerly parts of the excavation the flint had a completely unaltered surface.

Immediately above layer 2 is a strongly compressed layer of oyster, cockle, mussel and periwinkle shells containing charcoal and numerous artefacts of flint, bone and antler, layer 11.

Layer 11 is succeeded by a series of marine sand layers, 8, 10, 13 and 15, of different colours but similar composition and physical properties. At the bottom of sand layer 8, many artefacts and numerous millimeter-thick layers of shell litter occurred, presumably due to erosion from the top of layer 11. In the western parts of the excavation, the lowest parts of the sand complex are strongly rust coloured, layer 13. Throughout the sand, worked flint occurred, with a white patina and strong signs of water rolling. The sand layers are succeeded in the east by layer 4 and in the west by layer 12.

In layer 4, which has a matrix of marine sand, a series of small, separate heaps of shells was found to the east, resting directly on the surface of layer 8. These heaps consisted of shells of oysters, cockles, mussels and periwinkles. Mixed with these shells were numerous pieces of worked flint, bone, antler, charcoal and pottery. In the southwest, a fireplace was found in one of the shell heaps, showing that they had been deposited on dry land. The flint has a white patina and exhibits faint signs of water wear, but none of the shell heaps was redeposited or showed signs of erosion.

Layer 12 consists of coarse marine sand and gravel with a sparse occurrence of worked, strongly water-rolled flint.

To the east, layer 4 is covered by the marine sand layers 5, 6 and 7, containing much worked flint with white patina and strong signs of water wear.

The succession is completed by the ploughed top-soil, layer 9, containing numerous pieces of worked flint, especially in the most westerly parts.

Apart from layers 4 and 9, all deposits are of marine origin and formed in connection with a coast which has been variously influenced by vertical movements of the land. The stratigraphy is relatively simple and can in several respects indicate the circumstances of deposition.

Layer 2 was deposited in quiet, shallow water near the coast. Since it lies above the moraine, it must represent a rise in the level of the sea, which at some stage in the Mesolithic encroached on the former freshwater basin. Later, the sea level fell, as indicated by the rust-coloured sand of layer 13, and layer 11 was probably deposited on dry land. Later, the sea level rose again, layer 11 was inundated, and its upper parts eroded and redeposited at the bottom of layer 8, which registers a lengthy rise in the sea level. The shell heaps of layer 4, which were deposited on dry land, must stem from a period when the sea had once again withdrawn. Later, the shell heaps were inundated by the sea and layers 5, 6 and 7 deposited.

Occupation has apparently been continuous throughout the period of these variations in the sea level, moving up and down the shore accordingly. Apart from layer 8, where artefacts were confined to the deepest parts, cultural remains were found in all layers.

Scientific investigations in northern Jutland have shown that the Stone Age sea in these regions reached its maximum level during the middle or late Atlantic transgression. At Brovst, the maximum sea level is represented by the highest point of layer 6 in the eastern parts of the excavation.

There is reason to believe that layer 6 represents the late Atlantic transgression. If this is correct, the shell heaps in layer 4 must have been deposited during the regression between the middle and late Atlantic transgressions. Layer 8 must represent the close of the middle Atlantic transgression, whilst layers 2 and 11 belong to its opening phases.

These conclusions can of course only serve as a working hypothesis until more comprehensive investigations into the stratigraphy of the area have been carried out.

Artefacts

In the trial excavation, a relatively large artefact material of flint, bone, antler and pottery was recovered. The concentration of artefacts was greatest in the eastern parts of the excavation, and decreased steadily to the west. The vertical distribution of flint waste is recorded in fig. 4 in diagramatic form, based on the amounts recovered from 10 cm. thick layers in squares A 10 and A 12. Artefacts of bone and antler were found everywhere in layers 4 and 11, but only in the most easterly parts of layer 2.

The list p. 77 indicates the distribution of flint tools in the various layers. The raw material was flat blocks of cretaceous flint. The blade blocks are mono- or bipolar, in the latter case always with parallel striking surfaces. Cylindrical or prismatic blocks have not been found, which is unusual. The flake material is dominated by a large number of regular flakes, particularly in the deepest layers.

The stratigraphically and archaeologically oldest deposit is layer 2. Its artefacts are mainly of flint with a predominance of regular blades, which have been the raw material for tools of the common forms, fig. 6. A few flake axes have been recovered, but core axes are dominant. The layer is characterized by many rhombic points, here for the first time found in a limited cultural layer in Jutland, fig. 9. Transverse arrow-heads also occur, however, but in considerably smaller numbers, fig. 8. A little bead of red deer tooth has been recovered, fig. 7.

An artefact inventory in all respects identical with that of layer 2, is found in layer 11, figs. 6, 8 and 9. In any consideration of layer 11, it is necessary to establish whether it should be considered a true midden or whether it has been formed by the washing together of shells and artefacts. The preliminary investigations indicate that the layer has been deposited on dry land and constitutes a true kitchen midden. If future investigations confirm this, then Brovst exhibits the oldest kitchen midden found in Denmark. All previously investigated shell heaps seem to be later than the middle Atlantic transgression.

The inventory from the sand layer 8 closely resembles that of layers 2 and 11. It can be remarked, however, that rhombic points do not occur in this layer, only transverse arrow-heads being present, fig. 8. In layer 8 there are also more flake axes than occur in the deeper layers.

The artefacts from layer 4 differ radically from those of the layers described above. The considerable number of surface flaked symmetrical flake axes, numerous transverse arrow-heads of the type with strongly splayed side, fig. 8, and occurrence of thick-walled, coiled pottery, which on this site occur only in layer 4 and above, are characteristic. Of bone artefacts, a very regular needle, fig. 7, may be mentioned.

A rich artefact material stems from the layers above the shell heaps, in the main resembling that of layer 4. There is a sporadic occurrence of Neolithic objects in these layers, which must be intrusive.

Dating

The material from Brovst can be compared with that of the settlement at Norslund (KUML 1965). Since we know from coastal settlements in eastern Denmark that rhombic points belong to a different milieu than the transverse arrow-heads, and since the oldest deposits at Norslund contain only the latter, layers 2 and 11 at Brovst must be older than Norslund layer 3, which is radiocarbon dated to c. 3755 B. C. Norslund layer 3 is on geological grounds contemporary with or older than the middle of the middle Atlantic transgression, which agrees well with the preliminary results from Brovst, where layers 2 and 11 must be assigned to an early stage of the middle Atlantic transgression.

The sand layer 8 can be compared on a typological basis with Norslund layer 3 and assigned the same date.

The material from the shell heaps in layer 4 can be compared with that of Norslund layer 2 and Dyrholm II and must be dated accordingly to c. 3200 B. C. Geologically, Dyrholm II is dated to the regression between the middle and the late Atlantic transgressions, which is in complete agreement with the observations at Brovst, where the shell heaps can also be assigned to this period.

The artefact inventory in layers 5 and 6 is identical with that of layer 4, but stratigraphic conditions indicate that the former strata are more recent. In content, they are comparable with Dyrholm II, i. e. can be assigned to the period around 3200 B. C. Geologically, layers 5 and 6 are placed in the late Atlantic transgression, which does not conflict with the archaeological dating.

The sporadic Neolithic artefacts occurring in layers 5 and 6 must stem from later occupation of the site.

The preliminary investigations at Brovst have established for the first time the presence of the Kongemose Culture in Jutland. If layer 11 really represents the earliest kitchen midden in Denmark, the settlement has also contributed decisive information concerning the means of subsistence in the Kongemose Culture.

Søren H. Andersen