Bornholm i ældre stenalder – Status over kulturel udvikling og kontakter


  • Claudio Casati
  • Lasse Sørensen



Bornholm, ældre stenalder


The cultural development and contacts on Bornholm during the late Palaeolithic and Mesolithic

This paper presents the current status of research concerning the late Palaeolithic and Mesolithic cultures on the island of Bornholm in the Baltic Sea. Due to the island’s geographical position between Scandinavia and continental Europe, it can reveal important regional knowledge of all cultures throughout prehistoric times. The article also discusses subjects relating to island-archaeology, such as migration and possible isolation. Furthermore, the results show an updated picture of the settlement pattern on Bornholm during the Mesolithic. Finally, the archaeological finds reveal cultural as well as social contacts between hunter/gatherer societies in the Western Baltic and Bornholm. During the last few decades, the Mesolithic research on Bornholm has been focusing on the Maglemosian Culture. Systematic surveys in search of Maglemosian settlements were not conducted until the early 1980s. As an added bonus from these surveys, some late Palaeolithic and late Mesolithic sites were registered on Bornholm for the first time. Today, more than 125 Maglemosian settlements are recorded, but most of them are still a result of surface collections. Unfortunately, the preservation of organic material is not good on these sites, as they are found on Late Glacial shoreline deposits, i.e. on sandy soil. The Mesolithic habitants were forced to adapt and adjust to a very different raw material situation, and this gives the lithic industry during the Maglemosian Culture an extremely small and microlithic appearance. The lithic material from the Maglemosian Culture on Bornholm reveals some important regional aspects, which have similarities with Maglemosian settlements in Scania. These similarities are primarily caused by the same size of the raw nodules. The raw materials from Bornholm are all from the Maastrichtien and located in secondary deposits brought to the island with the different glaciers from the Quaternary Period. The most common flint type is »kugleflint« (nodular flint), often of good quality and not larger than four to six cm. The second most common material is the Kristianstad flint, which is dark grey and quite coarse and often up to 10-15 cm in size. The other two flint types are grey Danien, which is coarse to fine grained and tends to vary between four to six cm to around 10-15 cm in size. The use of local raw materials makes it easier to find the lithics from the Maglemosian Culture (Figs. 1 & 10). Many of the registered Mesolithic sites are a result of several survey projects.

The first survey project concentrated on inland sites located abnormally far from any water resources. We retrieved information from old maps about the old bogs and lakes, which had been drained through the last 100 years. By raising the ground water level on modern maps using GIS, it was possible to recreate the size of the former inland lakes and hereby reconstruct the landscape and the lake. The majority of the Maglemosian sites turned out to be located exactly on the edge of these former lakes. The second survey project involved sites lying near the different creeks on Bornholm. These sites were repeatedly visited during all the phases of the Maglemosian Culture. What caused the habitants to return to a certain area over a 1000 years period? One of the reasons could be the access to migrating salmon and trout. Recent biological research of the trout population from the Kobbe and Bagå creeks indicates that a large number of migrating trout still migrates up the major streams, which may coincide with some of the Maglemosian hotspots. The geographical development in the Baltic region can be divided into some main stages, the Baltic Ice Lake stage, 12000-9300 cal BC, the Yoldia Sea stage, 9300-8500 cal BC, and the Ancylus Lake stage, 8500-7000 cal BC, (Figs. 2 A-D). From the Baltic Ice Lake stage until the beginning of the Ancylus Lake stage, approx. 8200 cal BC, the island was either the northern part of a peninsula covering an area from Rügen in Germany to Bornholm, or an island with a substantial land bridge towards Rügen. In the following phases, from 8200-7200 cal BC, the sea level of the Ancylus Lake was low, and due to continued transgressions, the land bridge was flooded and several smaller islands were created between Pomerania, Rügen, and Bornholm. The size and geographical spreading of these smaller islands is still heavily debated. However, it is clear that Bornholm became an island some time during the Boreal period. The late Palaeolithic finds occur randomly on some Maglemosian sites on the island (Fig.3). A particularly interesting find is the elk antler harpoon found in Vallensgård Mose (Fig. 6). This double rowed harpoon is of a type known from both the Ahrensburg and the Sviderian Cultures. So far, the only excavated Palaeolithic site on Bornholm is in Vallensgård Mose (“mose” is a peat bog). The assemblage from Vallensgård Mose consists of lithic material, found in a layer, which was cryoturbated during permafrost in the Dryas III. The lithic material consists of blade and flake cores, flakes, blades, one end scraper and one tanged point. The raw material is high quality Senonian flint (Figs. 4-5). The typological dating and technological observations from the blade core suggests that the Vallensgård Mose material belongs to the Bromme, the Swiderian, or the Ahrensburgian Period. The question as to which technological complex the Vallensgård material belongs remains open and can only be answered through future excavations of the site. The key to an understanding of the settlement pattern during the late Palaeolithic cultures is the fact that reindeer tend to migrate along fixed routes. However, is it possible to relate an actual reindeer migration route to Bornholm? The datings prove that reindeer were present throughout the entire Late Glacial. The reindeer remains found on Bornholm display no certain signs of human working (cut marks and marrow fracturing). If we consider the datings of reindeer remains from Bornholm, an actual reindeer migration route to or through Bornholm can neither be established nor excluded. The absence of the Ahrensburgian Culture on Bornholm could be caused by the smaller size of the raw materials on Bornholm. This could have forced the Ahrensburgian lithic production to adapt a new technology, which had a Mesolithic character. A climatic explanation for the absence of Ahrensburg material could also be made – for instance that the harsh climate around the Baltic Ice Lake frightened off the Ahrensburg humans, as argued by Svante Björck. According to him, we should expect no scenarios with human activity near the shores of the Baltic Ice Lake. However, this is contradictive to the find of a longitudinally split metatarsal from elk from the bay of Køge Bugt. The faunal picture from the Preboreal containing reindeer elks and beavers proves that Bornholm had a complete package of migrating animals during this period. The reindeers and elks became extinct in the early Boreal when Bornholm became an island. The isolation did not have any effect on other larger mammals, such as red deer, roe deer, and wild boar, which migrated to Bornholm during the late Preboreal and the early Boreal. These animals had the ability to reproduce and adapt to a warmer climate and a denser forest during the Boreal and the Atlantic period. These faunal changes had an impact on the hunter-gatherers who migrated to the island during the Preboreal. A limited number of finds have been registered from the Preboreal phase (9.500-8.000 Cal. BC.) on Bornholm. At Lundebro, a few microliths and blanks were found, which show similarities with the early Mesolithic Barmose Phase, (Figs. 7-8). The pieces are up to three times broader on the average than the typical blades from the Middle and Later Maglemosian phases on Bornholm. This proves that the later habitants on Bornholm were forced to adapt and adjust to a very different raw material situation. From the later Boreal phases, a large number of Maglemosian settlements with different topographic characteristics have been registered. Two different types of coastal sites have been observed on Bornholm, with major difference in accumulation, use, and exploitation of the settlement area (Figs. 7, 20-21). Kobbebro was settled repeatedly, which resulted in a 70 cm thick cultural layer. Less than 500 m from Kobbebro, two other sites – Melsted and Nr. Sandegård – have been excavated. At these sites, a different picture of a coastal site type with separate flint concentrations emerges. All the sites are dominated by microliths, which indicates that hunting played an important role. The coastal sites have been located 100 to 200 metres from the Ancylus Lake, which played an important economic part, as indicated by the repeated habitation. Seal hunting could have taken place in the Ancylus Lake during the Maglemosian Culture, as the ringed seal migrated into the Gulf of Bothnia during the late glacial period. The inland settlements on Bornholm are located on higher elevated grounds near a lake, a stream, or a forced passage. So far, they are only known from surface finds. Another type of settlements is the observation site, which revealed differences in size and duration. The larger site Loklippen is located on elevated ground approx. 115 m above the sea, with a broad outlook over the Vallensgård Mose and near a forced passage, where hunters shoot their pray even today (Figs. 12-14). A big surprise was the fact that the inhabitants of the Loklippen site used quartz as raw material. This rather untraditional, yet systematic, flake production indicates that they were forced to use unconventional materials such as quartz. Quartz assemblages dominate the early Mesolithic settlements in Central and Northern Sweden, with the settlement of Hjälmsjön as the southernmost site dominated by a quartz material. This could prove a direct contact between hunter-gatherers in Bornholm and in Scania, as revealed by the systematic production of quartz flakes. A smaller type of observation site was excavated at Smedegade in Klemensker. This site had a more typical appearance compared to other observation sites and covered an area of four to eight square metres. The site had a limited lithic assemblage and so it was interpreted as a short-term hunting station. However, the situation at Loklippen proves that certain observation sites were more frequently used. The last type of site is the transit camp, which lies on a sandy plateau near a spring and a creek, where the conditions for water transportation, fishing, hunting, and gathering are favourable. Ålyst and Hullegård are two such sites that were visited repeatedly during the Maglemosian Culture, and so contain a complex of smaller or larger settlements. The main lithic production is blades for the production of microliths (Figs. 19-20, 24). The microliths at Ålyst can be dated typologically from the beginning of the Boreal phase (8000 cal BC) until the end of the Boreal period (7000 cal BC) (Fig. 22). Finds from Ålyst indicate that there was also a more permanent settlement that lasted for up to one or two months. It became clear when visible structural evidence representing two oval-shaped huts was found (Figs. 16-18). The two huts show remarkable similarities as to orientation, size, entrance area, fireplaces, and pits, as well as to the combination of lithic tool types. However, differences are seen with respect to the microliths. Lanceolates with lateral retouch and triangular microliths dominate in Hut I, while the microlith inventory of Hut II was confined to lanceolates with lateral retouch. The two huts are probably not contemporary, which can be established by future C-14 dating of the different features from the huts. There are also problems concerning the flint concentrations in the two huts: are they altogether contemporary with the huts? To prove this hypothesis, it is necessary to do extensive refitting between the flint concentrations, the postholes, and the pits both inside and outside the huts. The huts are 7 x 4 m, which gives them an inner area of 30 square metres. The Ålyst huts are thus of a middle size hut structure, compared to the rest of the Maglemosian huts from Northern Europe. Could the oval shaped huts be a normal hut type connected to the Maglemosian Culture in the Western Baltic? If we look at some of the huts found in the Western Baltic, there are indeed parallels to the oval shaped hut at Tingby and Årup in Scania and at Wierzhowo 6 in the north eastern part of Poland, (Figs. 9 & 19). The similarities are particularly remarkable between the Ålyst huts and the Årup hut. These hut structures show similarities in orientation, dimensions, position of the postholes, lithic material, typological dating, etc. (Fig. 18). The striking parallel gives rise to interesting questions about cultural influences and regionalism during the Early Mesolithic in the Western Baltic. The connection between Årup in North-Eastern Scania and Ålyst on Bornholm should be considered as more than chance. However, at the present it is still unclear whether the oval-shaped hut type has a special geographical or topographical distribution in southern Scandinavia. The settlement pattern on Bornholm is influenced by a regional mobility strategy connected to the special geographical conditions on the peninsula. These observations also illustrate the regional differences between the Maglemosian societies and their ability to change mobility patterns and to adapt to the local situation. This opens up for the discussions regarding the coastal vs. inland problem in Southern Scandinavia. The discussion should consider the geographical differences, which could lead to a different mobility strategy in each region of the Maglemosian Culture. Especially the sites from Holmegård, Sværdborg, and Lundby are located closer to the Ancylus Lake than to the Kattegat coast during the Maglemosian Culture. This location could lead to a commuting strategy between the inland lakes and the Ancylus Lake, with sporadic contacts to the marine areas as indicated by several objects of marine origin and the imported flint. This hypothesis – which is supported by the Carbon-13 values from Zealand and from the Barum Woman in Scania – indicates a mobility strategy orientated towards the Ancylus Lake rather than towards the marine coast along Kattegat. All the areas surrounding the Ancylus Lake, including Bornholm, probably had similar mobility and settlement patterns orientated towards the freshwater lake. The settlement pattern around the inland lakes changed drastically in the late Maglemosian Period when Bornholm had become an island. The main difference on Bornholm between the settlement pattern of the island and the settlement patterns from the earlier periods is the apparently deliberate rejection of the inland as a habitation zone and the concentration of settlements in the coastal zone. A possible explanation for this major change could be that some important resources were lost as the larger inland lakes became overgrown and filled with sediments. A similar pattern and decline in site number has been observed around some of the flat-bottomed lakes on Zealand, such as Barmose, Lundby, Sværdborg, and Holmegård. Part of the repeated settlement pattern on Bornholm proves that some of the creeks were used through more than a thousand years during the Maglemosian Culture. This may have had both historical and ecological causes. Some locations may be recognized as specifically orientated towards a certain gender, or a specific season. Furthermore, some sites could be devoted to the exploitation of specific resources of primarily symbolic or mythological rather than economical causes. However, it is clear that the island was occupied and used by hunter-gatherers who shared a landscape with territorial and ideological components. The repeated use of certain hotspots in the landscape could indicate bordered territories determined by the creeks. One family group would hunt and fish in one particular creek, whereas others were connected to another creek. Unfortunately, it is impossible at the present to locate these territorial borders due to the incomplete picture of the site distribution. The settlement patterns presented in this paper must be regarded as preliminary and subject to later modifications, mainly because the dating base for the sites is their content of microliths. As for Bornholm, a preliminary regional microlithic typology with four phases has been suggested (Fig. 22). However, we face serious problems in fine-tuning the typo-chronology of the Maglemosian Culture. If this phase could be split into minor segments, the sites would appear much more sporadic compared to the current picture of the habitation. The repeated settlement pattern and the fact that the typological and functional expressions are unchanged during the Maglemosian Culture indicate that the societies on Bornholm had a continuing social and cultural contact with other groups or tribes within the Maglemosian Culture. Towards the end of the Maglemosian Culture, the habitation became sporadic, and the possibilities of creating contact with other cultural groups became limited and difficult because Bornholm was an island. It is however important to keep in mind that Bornholm seems to have never been completely isolated, and that it had a continuous social and cultural contact with the later Kongemose and Ertebølle Cultures. This is currently supported by the fact that the first Kongemose site (Sandemandsgård) has been registered at Bornholm (Figs. 23-24). Furthermore, a submerged site was located on Southern Bornholm at Boderne at a depth of four to five metres (Figs. 23 & 25). This indicates a now submerged landscape around Bornholm, which was settled in the Mesolithic. The use of this submerged landscape and its impact on the settlement pattern is currently uncertain. These arguments demonstrate that Bornholm was never out of sight or out of mind for the hunter-gatherers of the Kongemose and early Ertebølle Cultures. During the following late Ertebølle Period, a large habitation along the Littorina coast is registered on the island. One of these sites is Troldskoven, which is of particular interest as it is the only settled cave site in Denmark (Figs. 23 & 26). It was found by a coincidence when a German tourist excavated the site in 1939 and collected a large lithic material. The material was subsequently lost under the World War II. During the following years, the site was more or less forgotten until we took an interest in the cave.

In 2004, we conducted a small survey and dry sieving of the surface of the cave and found lithic material. This could indicate that the cave was inhabited during the Early Ertebølle Culture. However, it is not possible to conclude any final dating of the assemblage until more investigations of the cave have been conducted. Another important late Ertebølle site on Bornholm is Grisby. This site demonstrates fishing and hunting – in particular on marine mammals. The artefacts from Grisby include imported lithic artefacts, Limhamn axes, and groove-decorated ceramics with an elongated cylinder-shaped base (Fig.27). All these artefacts are characteristic of the East Scandinavian Ertebølle sites, which were part of an established network across the Baltic Sea. These factors could be the basis of a swift transition to the Neolithic.

To sum up, Bornholm in the late Palaeolithic was the northern part of a peninsula or an island with a substantial land bridge, which covered an area from Rügen and Pomerania to Bornholm. The settlement in the late Palaeolithic is sporadic, although the excavated site in the bog Vallensgård Mose indicates possible contacts with Rügen. This material presumably belongs to either the Bromme or the Ahrensburgian Culture in the Allerød or Dryas III. During the Dryas III and Preboreal, faunal remains of reindeer and elks have been registered on Bornholm, but there is no evidence of settlements in the Ahrensburg Culture and only little evidence from the earliest Maglemosian Culture. In the following Boreal phases, a large migration to the island along with a warmer climate and a changed fauna has been registered. During the late Maglemosian Culture, Bornholm became an isolated island and the settlement pattern changed. The number of inhabited sites was reduced and the settlements concentrated near the coast. In the Maglemosian Culture, it has been possible from the archaeological material to observe continuous social and cultural contacts with other Maglemosian societies in the Western Baltic, as exemplified by the changes in flint technology and the similarities concerning the hut structures in this region. These facts illustrate how geographic developments challenge a hunter/gatherer group exceedingly and prove their ability to adapt to changed conditions, as seen during the Maglemosian Culture in the Baltic region. During the following Kongemose and early Ertebølle cultures, the island had a sporadic habitation, but new finds, especially underwater sites, could change the impression of the settlement distribution. In the late Ertebølle Culture, an increasing number of coastal sites have been registered on the island, and this clearly proves cultural contacts with Scania. Finally, the consistent contacts between Bornholm, Scania, Rügen, and Pomerania could be one of the main reasons why the process of neolithisation seems to have been swift compared to other parts of Southern Scandinavia. The prehistory on Bornholm also has its peculiarities with an exotic quartz production during the Maglemosian Culture and the first cave site attached to the Ertebølle Culture observed in Denmark. In this article, we have described how the geographical changes have challenged the Mesolithic hunter-gatherers to the extreme. One of their most important faculties was the ability to exploit and maintain cultural as well as social contacts with other Late Palaeolithic and Mesolithic societies in the Baltic region.

Claudio Casati og Lasse Sørensen
Afdeling for Arkæologi
Københavns Universitet





Casati, C., & Sørensen, L. (2006). Bornholm i ældre stenalder – Status over kulturel udvikling og kontakter. Kuml, 55(55), 9–58.