Pierre M. Vermeersch and Philip Van Peer (eds.), 1990. Contributions to the Mesolithic in Europe, Leuven University Press, p. 81 _ p. 86

Studies in Settlement Patterns and Submarine Bogs: Results and Strategy for Further Research

Ole GR0N

Introduction

At the previous conference I outlined the results of my analysis of the artefact distributions in the dwellings of tbe Maglemose culture (Gnm 1989). Tbe central assumption is - on the basis of the available etbnograpbic information: that tbe location of individuals, and thus of the activities, within dwelling-areas (central living-areas protected by dwellings, encircled by wind screens or just by symbolic markers) in hunter-gatherer societies are always organized spatially in accordance with a set of culture - specific rules tbat reflect their relative social relations and status (Gmn 1987b: 64-65). The existence of such rules of spatial behaviour appears to be a so-called "cultural universal", like tbe existence of initiation rites, incest taboos etc. wbich are found in all societies. Two patterns (fig. 1) of

® Ilrrtllith toll(tnlntion

@ hearth

,

~ A

rtille slde ,/ llIe side

B

rllily I rllily 2

___________________ IMn ____________________ _

lu.

Figure I: A model for the organization of the dwelling space on the basis of the locations of microlith concentrations, hearths, and dwelling traces. and the orientation of the pattern relative to the shore.

spatial organization were distinguished. Pattern A consisting of relatively small concentrations of material witb one concentration of microliths, and one fireplace. Pattern B consisting of somewhat larger concentrations of material (nearly the double size) with two concentrations of microlitbs and in a number of the cases also two fireplaces symmetrically arranged around a line perpendicular to the shore. The sites are often located only a few meters from the shores of lakes or streams.

Of 21 sites analyzed at the moment, only 3 do not fit into one of tbe two categories. Among tbe remaining ones tbe two patterns are represented by equal numbers. Tbere is a marked tendency tbat the units following tbe A-pattern - tbe smaller ones­chronologically are Preboreal, wbereas tbe larger units - of B-type - are Boreal. If the interpretation is correct, tbat we have to do witb one - respectively two family dwelling units, this does not necessarily imply tbat tbe settlements became larger, population density increased etc. We need more information on tbe settlement size and organization to draw such conclusions. Their number of dwelling units may for instance have decreased.

The dwelling units distinguisbed are signified by a number of features: I. In nearly all cases tbey appear as marked

concentrations of worked flints , less tban 10m in diameter. Only concentrations of an apparently typologically pure cbaracter bave been taken into consideration.

2. In a number of cases floors of bark and brancbes are preserved, and in a few cases wbat appears to be stakes from the dwelling structure itself bas been found.

81

3. With sites on sand it is normal to see a lenticular thickening of the "dwelling-area", consisting of homogeneous light grey sand. This seems to reflect that it was normal to dig a shallow pit for a dwelling area. Probably the homogeneous sandy lenses consist of a mixture of charcoal, organic waste, and sand that has fallen through a floor of boughs and/or bark. Where more of these features appear in connection with one dwelling unit, they coincide.

The following is a discussion of the latest results and the further aspects of the approach to settlement structure chosen for analysis of the Maglemosian material. First I shall discuss a possible new dwelling structure. Second I shall discuss research strategies for providing new and better data for further analysis of settlement structure. More specific a recently started research project on reconstruction of cultural landscape and settlement patterns from the stone age in a submerged bog basin measuring approximately 9 by 14 km will be discussed.

Flaadet - a Winterhouse?

In 1987 I suggested that the late Preboreal Flaadet site should represent a large winter dwelling of the Maglemose culture (Gr0n 1987a: 311-315). The arguments for this were the following: a. The unit was located 20 - 25 meters from the

prehistoric shore. 3 - 5 meters is normal for dwellings on the shores of lakes. The proposed explanation is that this withdrawal from open water was made to avoid the layer of cold air that in winters will form just above the ice of frozen lakes. The location chosen for settlement was elevated approximately 1.5 m over the water level at that time (Skaarup 1979: plan I). Comparable behaviour has been registered in several ethnographic cases (Gr0n 1987a: 311-312).

b. The site was located on the northern part of a little island in the prehistoric lake. According to Knud Frydendahl's reconstruction of the climatic situation, northerlies were the least common winds during winters (Gr0n 1987a: 317-318). Thus a position oriented to the north would be profitable in winter. Ethnographically this phenomenon is quite well known from colder regions.

c. The internal structure of the site fits into the B-pattern. The only difference is the size: where this pattern in the other cases is connected to structures measuring approximately 4 by 6 meters, it takes up an area which in this case according to the analysis of the distributions measures something like 7 by 8 meters. The two concentrations of microliths normally take up one or two adjacent square meters each. In this case, they are considerably larger, and of the shape of

82

two crescents symmetrical around the axis perpendicular to the shore. Their concave sides are directed from each other (fig. 2B). Likely they represent waste accumulated around two persons (Gnm 1987a: 313-314). A variation in the dwelling space available per person in summer and winter is often found with hunter-gatherers (Gr0n 1987a: 314). The smaller B-units, are in a number of cases known to have been in use between April and October.

If the interpretation of the Flaadet site as a large winter dwelling is correct, one should expect a construction of that size to have left some structural traces such as for instance postholes. One of the weaker methodological parts of traditional excavation in squares is the difficulty in observing and registering faint structures and colourings showing in the surface. Since the special geological conditions on this site furthermore made it likely that traces of structural remains could have escaped observation during the 1973 excavation, a "reexcavation" was carried out in August 1989 for Langelands Museum by the author.

The plough soil was removed mechanically down to untouched underground. Immediately four grey oval to circular structures, approximately 50 cm in diameter, appeared on the central line of the rectangle postulated in 1987 (fig. 2) (Gnm 1987a: 313). The northernmost one was located 10 cm south-east of the middle of the postulated north­western gable, the southernmost one did exactly touch the south-eastern gable-line from the inside (fig. 2).

, , ' .. --'"

Figure 2: Dwelling outline postulated in 1987: dotted line. Border of 1973 excavation: line. Square size - I m2

Figure 2a: A) Post holes(?) excavated in 1989: black areas. Border of 1989 excavation: dash line.

Figure 2b: Microlith concentrations: hatched areas. Concentrations of burnt nints (hearths?): black areas. Concentration of scrapers: contour lines (equidistance - I). Border of interpolation: dash line.

To explain the structures, it is necessary to describe the original geological features of the locality (fig. 3). I. The underground of the little island consisted of

light brown and stiff sandy moraine clay. 2. [n 1973 a compact layer of 10- 20 cm large

stones with sand/gravel between them was observed directly above the underground. This is typically a layer of stones and sand/gravel originating from the top of the little hill , from which the clay has been washed away by wave activity, at a relatively high water level. The stone-layer had a thickness of 15 - 20 cm within the excavation area.

3. Above the stone layer were registered remains of an approximately 5 cm thick layer of white stiff lime-gyttja, of late Preboreal origin (Skaarup 1979: 31).

4. The stratigraphical position of the culture layer was not registered exactly, since it was the destruction of this layer by ploughing, that led to the discovery and excavation of the site. But it was located just a few centimetres above the lime­gyttja.

5. According to observations made in other parts of the little island/hill, the peat in which the Maglemose material was found , was of an Atlantic origin, and thus has covered it as a protective layer, until it was destructed by ploughing (Skaarup 1979: 17-34).

® ERTEB0LLE LOCALITIES

o NEOLITHIC LOCALITIES

Figure 3: The Stryne basin. Land: black areas. Depths of 0-2 meters: white. Depths of 2 - 4 meters: light hatched. Depths of 4 - 6 meters: dark hatched.

[n 1989 the stone layer had been removed by heavy ploughing. Excavation showed the circular/ oval structures observed in the surface to be 12 - 15 cm deep regular pits in the underground clay, with evenly rounded bottoms, and a content of sandi gravel. Diameters varied between 40 and 50 cm. [n the top of this sand were traces of 10 - 20 cm large stones removed recently by ploughing. These and the sand/gravel probably have fallen into the pits from the stone layer.

The four pits have been covered by a culture layer containing 500 - 1000 pieces of worked flint per m2 in their locations. They did not contain even small pieces of worked flints . This strongly indicates that the pits were created before the culture layer was accumulated.

This leaves us with three possibilities. The structures observed may be periglacial phenomena. This possibility is rejected by all geologists [ have discussed the matter with. The general geological situation on the location, the regular and uniform shape of the structures and the fact that they are found on a line but do not have "symmetrical" counterparts within a distance of a few meters, seems to exclude this possibility.

The structures may be cultural features originating from an earlier settlement on the site. Meanwhile such a hypothetical Palaeolithic/late­Palaeolithic one has left no cultural items. One very little piece of lime (only a few miHimetres large) was found near the top of the sand/gravel in one of the pits, in a position with no worm traces (these were easily observable as dark channels in the light sandi gravel). This piece must originate from the lime­gyttja and points to an origin of the structures later than the formation of the lime-gyttja in late Preboreal (Skaarup 1979: 31). Thus it seems that the pits must have been created at some time after the beginning of the Maglemose phase - but on the other hand they cannot be later than the formation of the culture layer that belongs to an early phase of the Maglemose Culture (Preboreal/ early Boreal) (Skaarup 1979: 10 I-I 04). Thus they have likely been created in a narrow time interval before the Maglemosian settlement.

The dating of the pits in relation to the settlement and their positions in relation to the hypothetical dwelling outline based on the artefact distributions on the site, strongly indicates that the four structures represent some kind of structural elements related to the postulated dwelling structure. Any use as storage pits or pits for hearths should have shown as charcoal or organic colouring of the sand/gravel. The well preserved phosphate in the top of the underground clay, proves that the chemical traces are rather well preserved, and have not been washed out.

83

The features have originally been at least 20 cm deeper than when observed in 1989, and thus must have measured 40 - 50 cm in diameter, and 32 - 35 cm in depth. Immediately one would think of holes from a central row of posts supporting the ridge of a roof. Since the stone-layer has been destructed, structures indicating the courses of walls can easily have been removed. Postholes of similar dimensions are found as a part of the inner construction of the large Russian late Palaeolithic houses of Kostienki (Jefimenko 1958: fig. 5; Rogatchev 1955; 1957).

If the observed structures do represent postholes, it seems strange that no worked flints have found their way into the pits when the posts turned over at some time after the abandonment of the site. The material found in the pits meanwhile was sand/gravel of the type found in the "stone­layer". Higher there had been stones of the size found in the stone-layer - according to their negative imprints. This loose material may well have found its way into the bottom of the pit first. The little piece of lime lay centrally, close to the top of the sand in one pit. The stiff lime-gyttja may have withheld the culture layer that rested on it from falling down, so that cultural items were deposited only in the upper parts of the pits, removed by ploughing today.

It is interesting to note a preliminary publication of a very similar construction, with clear holes from posts with a diameter of 15 cm, and wall lines marked by stakes with a diameter of 5 cm, excavated by Ebbe Westergren in south-western Sweden. The structure is rectangular, 3.5 by 8.8 m­half the size of the postulated Flaadet-house. There is one post in the middle of each gable, and one centrally in the room. These three likely supported a saddle roof. There are also posts in each of the four corners, and one in the middle of one long side and two - marking a proposed entrance - centrally in the other. The culture layer appeared as a 12 cm thick layer inside the structure, whereas only thin blotches were preserved outside. The concentration of finds was clearly restricted to the outline of the structure. This structure was located 2.5 m above the water, and like the FIAdet site, 25 m from it (Westergren 1988: 5).

The results - not least the structural similarity between the two cases - seem to support the idea put forward by Ebbe Westergren and the author, that larger and more solid dwelling types than the hut/ tent structures hitherto found , were in use in the Maglemose culture (Gmn 1987a: 311-315; Westergren 1988: 5-6).

In the discussion of the seasonal rotation of the Magjemose Culture, it has been postulated that the winter quarters were on the coasts of the North sea and the Baltic. Meanwhile here exist some indications that this was not the case. Skottemarke and Favrbo were clear inland locations in early

84

Magjemose. At the former were found at least six and at the latter two slaughtered carcasses of elks. The 14C datings point to 7500 BC (uncalibrated) for both. According to the seasonal indicators they were killed in the winter (M0hl 1908: 5-15). According to the excavation of one of the few known Magjemosian coastal sites, Bua VastergArd in Sweden, the diet estimated in kg consisted of approximately 50 % from the sea, and 50 % from the forest. The site was located on an island, and within a reasonable hunting range surrounded by considerably more water than land (Wigforss el al. 1983: 8-13, 140, 196). This does not seem to support the picture of the Magjemose hunters as stubborn sea-hunters that would travel far to exploit the sea­resources each winter.

One possible explanation of the fact that so few winter sites from the Magjemose Culture have been observed, is that the sites were slightly withdrawn from the shore. If their waste has not been deposited in a nearby water logged waste layer and preserved, no organic material and thus no winter indicators will be preserved. At Flaadet only a few bone fragments and one elk (?) tooth were preserved (Skaarup 1979: 136-137).

A Submarine Bog - Research Strategy

Research in the organization of dwellings and settlements in the Magjemose Culture systematically depends on casualities: The majority of the sites analyzed, have been discovered by accident, and therefore were often partly damaged when excavated. With such a procedure it is not possible to pick out sites with optimal conditions for preservation of organic matter or to focus on sites related to a specific prehistoric resource area, as for instance a large bog basin.

For studies of settlement patterns, the ideal study area would be a large bog basin that in prehistoric time slowly grew into a bog, so that the different phases of settlement on its shores were separated horizontally. This situation is known from Amosen on Sealand (Andersen 1983: 189). Furthermore the conditions for preservation of bone/antler and wood should be of a high quality and the basin should not have been subject to peat digging or drainage connected to modern agriculture. In Denmark this rules out all the relevant land-based basins and leaves as the only possibility submerged bogs at reasonable working depths.

Ideally one should be able to locate sites before they are revealed by human or natural destruction. I have been involved in quite a number of speculations of how one could detect Mesolithic sites in bogs, before they were revealed by

destruction. The geo-radar is one possibility. Meanwhile it is sensitive to very small concentrations of salt in the ground water in moist areas. Very few p.p.m. (parts per million) incapacitate it totally. This means that the equipment will be impossible to use in most moist areas in Denmark. Sonar technique is impossible to use on land with a realistic solution. Since soil with its numerous small air filled cavities is an extremely bad sound transmitter, penetration will here be linked to waves too long for registration of 10- 20 cm thick anomalities such as culture layers. In submarine situations the situation is different, because the soft sediments will be saturated by water which is a very good sound transmitter. In 1985 I became aware of Tjeerd van Andels work with sediment echo sounder for reconstruction of the now submerged resource area of the Franchthi cave (Andel and Lianos 1984). Possibly one could use such equipment to locate settlements in submerged peat. Subsidiary one could at least spot the topographical situations where settlement is likely to be found , such as little islands covered by peat in the basin, positions where water courses meet the basin etc. One problem is layers of sand, which are unpenetrable even at a rather restricted thickness. Another is soft sediments which, due to their production and content of numerous little bubbles of gas, are unpenetrable. With a sediment echo sounder, one will have to chose a research area where none of the problems are present.

Through the last years Langelands Museum systematically has surveyed some the most promising Danish locations, and in collaboration with professor Jens Tyge Maller, Dept. of Geology, Aarhus University, have carried out experimental work with a sediment echo sounder with 20 and 300 kHz transducers. One preliminary result is that structures embedded in soft sediments with seismic characteristics and dimensions similar to those of Mesolithic culture layers, can be distinguished. Another is that we, after we had concluded that such an ideal basin probably did not exist, found what appears to be one less than 1 km from the Museum.

The Stryna basin (fig. 3) measures approximately 9 by 14 km, and surrounds the little island Stryna. The top of the peat is 4 - 5 m deep. According to dated sites in the vicinity the sea-level should have risen this high between 5000 and 4000 BC The peat is exposed, but apparently not eroded. Branches and trunks of typical bog-vegetation such as pine, birch, aspen etc. are found in its top. Around the peat surface is a submarine landscape with numerous oak trunks turned over, and in many cases with the stumps still standing in the bottom. Outside but adjacent to the bog are late Erteballe sites with organic matter preserved (Skaarup 1983). More distant from it is a series of Funnel beaker sites, and, amongst other things, a submerged longbarrow (Skaarup 1985: 138) is found. On

Stryna, centrally in the prehistoric lake, three Maglemose sites are known. The peat in the basin seems to be without development of gas or sandlayers embedded.

The strategy for 1990 is: I. To proceed to collection of samples of oak trunks

for dendro analysis. The aim is to produce a local main curve back to 5000 BC The dating of trunks and stumps will give an extremely detailed insight into the development of the water level locally (the trunks can only be preserved because they have fallen into the water of the rising sea, and likely most of them have fallen because their roots were overswum). This will be an important contribution to the discussion of the development of the water level and the relevance of the tilting line model. A local dendrochronological curve also will make exact dating of later Mesolithic sites possible, if such are found.

2. To analyze the topography of the basin by means of sediment echo sounder, and a number of borings into the layers.

3. To try to locate Mesolithic culture layers in the peat by means of sediment echo sounder.

The basi n likely contains well preserved Mesolithic sites, with good preservation of hut floors and other structural elements. Our problem is to find them. If the seismic method proves successful, it will be possible to localize sites very fast and efficient within large areas. By means of relative dendrochronology it should be possible to determine whether spatially separate units are exactly contemporaneous, and thus on an exact basis to study the organization of possible larger settlements with several dwelling units (eg. Gran 1978: 303-310). Possible winter dwellings of the postulated Flaadet type will be difficult to distinguish by means of seismic methods, because they should be located directly on the prehistoric underground (clay or sand) and not on top of some peaty bank.

Note

The present research is supported by the G.E.C Gads Foundation, and in its initial phase was supported by the Augustinus Foundation. The Danish Research Council for the Humanities has granted a sediment echo sounder for interdisciplinary geological-humanistic research.

85

References

van Andel T. and Lianos N., 1984. High-Resolution Seismic Reflection Profiles for the Reconstruction of Postglacial Transgressive Shorelines: An Example from Greece. Quarternary Research, 22, 31-45

Andersen K. , 1983. Stenalder bebyggelsen den vestsjadlandske Amose. Copenhagen.

Gren 0 ., 1987a. Seasonal Variation in Maglemosian Group Size and structure. A New Model. Current Anthropology, 28, 303-327

Gr0n 0 ., 1987b. Dwelling organization - A key to the understanding if social structure in old Stone age societies? An example from the Maglemose Culture. In: New in Stone Age Archaeology. Archaeologia Interregionalis. (eds.) J .K. Kozlowski and S.K. Kozlowski. Varia, vol. CCXXX. Warszaw, 63-83

Gren 0 ., 1989. General Spatial Behaviour in Small Dwellings: a Preliminary Study in Ethnoarchaeology and Social Psychology. In: The Mesolithic in Europe. Papers presented at the third International Symposium. (ed.) C. Bonsall. Edinburgh , 99-105.

Jefimenko P.P. , 1958. Kostienki I. Moscow-Leningrad.

M0hl U., 1980. Elsdyrskeletterne fra Skottemarke og Favrbo. Skik og brug ed borealtidens jagter. Aarbeger for Nordisk Oldkyndighed og Historie 1978, 5-32.

Rogatchev A.N., 1955. Aleksandrovskoe poselenie drevnekamennogo veka u sela Kostienki na Donu. (Mater. iss. arh . S.S.S.R.). Moscow-Leningrad.

Rogatchev A.N., 1957. Mnogoslojnye stoanki Kostenkovsko-Borchevskogo rajona na Donu. In Paleolit i neolit. (ed. ) P.P. Jefimenko. (Mater. iss. arh. S.S.S.R. 59), Moscow, 9-118.

Rank G., 1951. Das System der Raumeinteilung in den Behausungen der nordeuroasischen Volker, voUI. Stockholm.

Skaarup J ., 1979. Flaadet. En tidlig Maglemoseboplads po Langeland. Rudk0bing.

Skaarup J ., 1983. Submarine stenalderbopladser i Det sydfynske 0hav. Antikvariske Studier 6. Copenhagen.

Skaarup J., 1984. Yngre stenaider pa @erne syd for Fyn. Rudk0bing.

Tanner A., 1979. Bringing Home Animals. Religious Ideology and Mode of Production of the Mistassini Cree Hunters. London.

Westergren E., 1988. Det lildsta kanda huset i Norden. Popular Arkeologi. arg.6,4, 4-6.

Wigforss J ., Lepiksaar J ., Olsson I.U . and Passe T., 1983. Bua Vastergard -en 8000 ar gammal kustboplads. Goteborg.

86

U.I.S.P.P. Mesolithic Commission

CONTRIBUTIONS TO THE MESOLITHIC IN EUROPE

PAPERS PRESENTED AT THE FOURTH INTERNATIONAL SYMPOSIUM 'THE MESOLITHIC IN EUROPE', LEUVEN 1990

Edited by Pierre M. VERMEERSCH & Philip VAN PEER

Leuven University Press 1990