Aggersund. An Ertebølle settlement on the Limfjord

Abstract

Aggersund. An Ertebølle settlement on the Limfjord

The Ertebølle culture in Denmark is probably one of the best studied mesolithic cultures in northern Europe. Archaeological investigations have, however, concentrated almost exclusively on the large settlements, particularly their middens, while excavations of the smaller settlements of the culture have only rarely been undertaken or published.

Large sites, representing lengthy occupation, are not suited to the elucidation of conditions in the short term. Small settlements on the other hand open the possibility of erecting a finer chronological system, which can be used in the analysis of the larger sites, as well as the possibility of studying more short term, seasonal subsistence activities. Finally the small sites make it possible to check the degree to which subsistence affects the tool kit present.

This situation can be shown by means of an example from Aggersund in north Jutland. The area round Aggersund is now some 6 m higher in relation to the sea than it was during the stone age (2), during which period it was broken up into a series of larger or smaller islands separated by narrow straits. or broader shallower stretches of water, fig. 1. Traces of the action of the stone age sea are much in evidence in the form of old seacut cliffs, raised beach systems and areas of raised Littorina seabed (3). Because of the particular topographic and geological situation Aggersund was the narrowest point of the fjord in the prehistoric period as well. West and northwest of Aggersund are areas of fen -which in the stone age were wide areas of shallows- where large quantities of waterbirds collect while migrating through in the spring and autumn. In this part of the Limfjord it should therefore be possible to find settlements which exploited these waterbird migrations to some considerable degree.

The excavation of a small Ertebølle settlement on the north side of the fjord at the farm »Bundgård« near Aggersund in the central part of the Limfjord was therefore awaited with particular interest.

The site is today on a south facing, evenly sloping field, 6.50 m above and ca. 850 m away from the water, but originally lay on the beach itself.

Fig. 2 shows two sections, running respectively N-S and E-W through the settlement. Bedrock consists of chalk, overlayn by a sterile, yellow layer of marine sand. In this sand is an oval pit measuring 12 X 6 m (N-S and E-W respectively) with a maximum depth of 30-40 cm.

At the bottom of the pit the upper 2-3 cm of the sand are grey or browny black, with clay in a few places and small areas with burnt crushed masses of shell. This layer contains a great number of animal bones, much charcoal, and fragments of flint waste and tools. The flint is bluewhite, sharp edged and not rolled. This layer apparently represents an old surface, into which these remains have been trampled down into the loose surface of the sand.

Above this is an even transition to a shell layer with cultural debris (a kitchenmidden or »køkkenmødding«) which is up to 40 cm thick and fills the whole of the pit. The midden, a primary deposit, measures about 12 X 6.5 m (N-S and E-W) and lies between the 5.80 and 6.20 m contour lines.

The shell layer consists almost entirely of oysters (93%) (13), mixed with flint, animal bones, antler, charcoal and a little pottery. The flint is white-blue and not rolled.

The structure and consistency of the layer is very homogenous; the general impression is that it was deposited over a very short time and seems to represent a single occupation.

Above the shell layer is a black sandy layer about 10 cm thick, containing charcoal, and many flint tools, animal bones and potsherds. The stratigraphic sequence finishes with the ploughsoil, which contains rolled flints.

Summary of the stratigraphy

The sequence shows that the sea covered the area and deposited the yellow sand before the site was occupied. This sea level, reaching at least to the 6 m level and probably nearer the 7-8 m level, was probably attained during the High Atlantic Transgression. After the sea receded, human occupation followed and the depression became filled with cultural remains as a result.

The degree to which the area remained above sea level or was again flooded immediately after the occupation cannot be determined.

After some time, however, the black layer was deposited, and at that time at any rate the site must have been dry. The finds show that this occupation occurred during the transition from early to middle neolithic (the sub-boreal era). It is possible that this black level could be of marine origin; in which case it would signify another marine transgression. It must however be mentioned that no signs of marine erosion were observed in the midden, which is an entire primary deposit. On the other hand it must be admitted that a possible subsequent transgression could have washed away layers lying higher up and therefore more exposed to erosion than the midden, which was protected in the abovementioned depression.

The fact that the ploughsoil contained many rolled flints shows that after this occupation the settlement was covered by a transgression. The archaeological finds show that this transgression dates to the sub-boreal, and reached a height of at least 7.00 m. After this the sea once again receded and the site has since been dry.

As the purpose of this article is principally to sketch the archaeological background to U. Møhl's work on the fauna) material from the midden, emphasis will be laid on the finds from the midden and the level immediatly below it, while the remaining layers will be dealt with rather more briefly.

The finds: The catalogue, p. 46-49, shows which tool types were found in the various layers.

The ploughsoil: This layer contained the most finds; but because of its »non-sealed« nature there is no certainty that all the artifacts are contemporary.

The inventory from the ploughsoil includes many Ertebølle types together with a few from the middle neolithic Funnel-Beaker and Pitted Ware cultures; this can be seen in more detail from the list of finds, p. 46-49.

The finds from the black layer are few, but from the potsherds can be dated to the transition from early to middle neolithic. The finds belong to a variant of the Funnel Beaker culture local to North Jutland (15) fig. 8-9.

The animal bones from this layer include wild and domestic pig, domestic cattle, red deer, sheep and/or goat, dog, swan and cod (16).

The site location as well as the animal bones both show that this layer represents a small, neolithic hunting station (17).

The shell layer: Finds come from both the shell layer itself and from the very top of the underlying sand.

No fireplaces were found in the pit, but blackening of parts of the top of the sand as well as the presence of burnt bones and flint show that fires did burn in the hollow, fig. 3.

The horizontal distribution of the finds was very uneven and several local concentrations could be seen, figs. 4-7. Activity areas are defined as areas where the number of artifacts of a particular type is above the average number per square metre of the type. Flint waste is particularly concentrated in three areas, fig. 4, and as it has proved possible to reconstruct larger parts of cores from many flint fragments within these concentrations, one can conclude that flint working took place in these areas.

The tools are found with a few exceptions evenly spread, figs. 5-7. The exceptions are axes and blade borers, which are found in smaller, clearly delimited concentrations, figs. 6-7.

As the axes are frequent in the same areas that much flint was found, one can conclude that it was mainly axes that were worked on the locations.

The animal bones were also found in smaller, clearly delimited concentrations which apparently represent butchering waste.

These observations show that the depression functioned as a work- and activity area before it became filled with shells.

That the flint and bones lie in such clearly delimited groupings must mean both that the settlement was in use only for a very short period, and that the infilling of the depression occurred quickly after the activities had taken place. Had this not been the case the concentrations would surely have been spread over a larger area and have been mixed with other waste.

Technology

The raw material is predominantly cores of danien-flint from the area of the settlement, where the chalky flint from the bedrock must have been easily accessible. The cores are low rectangular pieces of flint with fresh cortex on their sides, figs. 10-11. Natural frostshattered flakes of flint were also used, in particular for the larger tools, such as axes and core borers, fig. 14d. As these »natural flakes« are very common and appear in large numbers in the tight-packed basal sand, it is possible that they were dug up or collected from this layer, which might account for the presence of the depression in the sand layer.

The material shows the usual Ertebølle technology, i.e. many fine blades, which are however on this site generally thinner and narrower than is usual, cf. fig. 7. Many microblades are also present, which is unusual for an Ertebølle context. It cannot be determined whether these differences are due to the raw material or reflect a definite intention in the blade production.

Many of the core trimming flakes show scars of thin, narrow blades or microblades, which shows that the rather unusually large number of microblades is not due to chance.

Hammerstones: fig. 12. These occur mainly in the southwest part of the settlement -or in the same area where the largest of the concentrations of flint waste have been demonstrated.

Cores, fig. 10-11. Conical cores with one flaking surface are the most usual. They are found evenly distributed throughout the central area of excavation, and there is no clear parallel between their distribution and the concentrations of flint waste, fig. 5.

The cores most often have two flat areas, opposite each other, from which cortex has been flaked. Blade removal is almost always at the »narrow end«. Only one core is round and has had blades removed from all over its surface, fig. 10a.

Scrapers, fig. 13: This group, forming 8.3% of the implements, consists predominantly of blade scrapers; fine, regular pieces similar to those known from other Ertebølle sites in Jutland.

Burins are also common at 10.4%, fig. 15. Angled burins are the most common, and of these as many are on snapped blades as are on pieces with a transvers preparatory blow. Angled burins on transversal retouch are on the other hand rare. The burins were found evenly spread over the whole settlement area and do not seen to cluster in »special activity areas«, fig. 6.

Pieces with transverse retouch, fig. 16, make up 9% of the material. Blades with concave retouch predominate; three are made on thick »natural flakes«, fig. 16e, and are clearly not similar to the remainder -they are presumably rough-outs for flake axes; compare e.g. fig. 16e with fig. 22.

The transverse-retouched pieces are more or less evenly distributed over the whole settlement, fig. 5.

One partly-backed blade must be described as a knife.

Transverse arrowheads, fig. 17, are rare on this site (11%) compared with other Ertebølle sites. It is however not certain whether this is the real state of affairs or whether the low representation could be due to aspects of excavation technique.

Toothed and notched pieces, fig. 19, are rare; the same is true of pieces with continuous edge­retouch, fig. 20.

Axes, fig. 21-25, are common; they include both core and flake axes. There are a few more core than flake axes. Both types are characterized by many irregular and atypical examples.

Predominant among the core axes are the asymmetrical and atypical types, although symmetrical axes are also well represented, figs. 22-23. On the other hand no symmetrical axes with radial edge-flaking were found. The flake axes are mainly asymmetrical flat flaked and atypical specimens, fig. 24, while the symmetrical flat flaked type, one of the most characteristic of Ertebølle tools, is rare and also irregularly flaked.

Pointed core tools (picks), fig. 22, are also present in the material.

Bone and antler artifacts are rare, figs. 25-28. Antler axes of the type to be hafted as an axe rather than as an adze and with the shafthole near the base are present in two specimens, fig. 25. There are also pressureflakers, fig. 26, and sawn off tines of red deer an tier, fig. 27. Among the last-named the piece illustrated in fig. 27b may have been used as a »retouchoir« (31).

Finally the material includes a number of bone points, fig. 28, made on splinters of birdbone (swan). One fragment of a fish-hook, fig. 28e, was also found.

Pottery, fig. 29, is rare. Only sherds of thick-walled, coarse, coil-built Ertebølle ware have been found.

Dating

In the thickest part of the shell layer three shell samples (Ostrea edulis) were taken out in stratigraphic sequence. The lowest is sample no. K-2638, which gave 3470±95 b.c. K-2639 is central in the layer, and gave 3460±95 b.c. and finally sample K-2640 is from the top of the layer; 3510±95 b.c.

The average is 3480±95 b.c.

The radiocarbon dates show that the shell layer was deposited over a very short period of time -and apparently represents a single shortlived (seasonal?) occupation- a result in accordance with the archaeologicalstratigraphic observations.

The radiocarbon dates also show that, whether the site is a small settlement or just a part of a larger one, the find may be regarded as a short chronological unit.

The archaeological material confirms the scientific datings; on a typological basis the settlement is relatively dated to the period between Norslund levels 3 and 2 (43).

Conclusion

The Aggersund site represents a short occupation, which on the basis of radiocarbon dates is to be placed at about 3,500 B.C. in uncalibrated C-14 years. In terms of relative chronology on is dealing with a short phase of the Ertebølle culture of Jutland, which has not previously been isolated on a settlement of its own. Because of this the site aquires significance with regard to future analyses of the large »classic« Limfjord settlements of this culture.

The subsistence base was hunting, fishing (?) and collecting.

The surrounding woodlands provided wild pig, red deer and roe deer, the fjord an addition in the form of gray seal and oysters. A few furproviding animals, fox and pine marten, provided pelts for clothing and dwelling structure.

That fishing must also have played a part in the economy of the settlement is shown by the presence of a fish-hook. The absence of fish bones is probably not of major significance -factors of preservation and excavation technique can be decisive in such cases.

It is on the other hand swans which characterize the settlement the numbers of these birds being unique among other small Ertebølle settlements.

To this must be added the fact that the settlement is seasonal, occupied in late autumn and winter -a seasonal exploitation which, according to U. Møhl must be connected with the specialized hunting of the migrating swans.

That this bird exploitation really was specialized is shown by the fact that only bones of swan have been identified, despite the fact that there must have been large numbers of ducks and geese present as well at this time of year.

The composition of the fauna thus confirms the initial hypothesis, in which it was suggested that in precisely this part of the Limfjord Ertebølle settlements would be expected, the economy of which was based on the large-scale migration of birds.

On this basis one could perhaps expect to find an artifactual material which differed either in choice of types or in its proportions from what is known from other contemporary Ertebølle sites, but this does not seem to be the case to any larger degree. It is noteworthy that all types are present in the finds. As there is for the present a lack of relevant comparative material in the form of correspondingly small settlements of the same age from this part of the country, it is difficult to decide which aspects of the assemblage and its composition are determined by function and which by chronological factors.

The Aggersund settlement in the meantime helps give a more varied picture of the Ertebølle culture's economy; it is a seasonal site -used during autumn and winter for the specialized hunting of swans, a new economic facet not previously documented, but which could well have been a part of the large settlements in the region, for example Ertebølle itself; due to the nature of the site (a large settlement) and the excavation technique, this cannot be established with any degree of exactitude in relation to any particular toolkit or point in time.

Søren H. Andersen