The use of river valleys by prehistoric hunter-gatherers: a study in the eastern Flemish Valley.

River valleys and basins have played a primordial role in the ecology of prehistoric huntergatherers, from their earliest Palaeolithic appearance to the last nomadic groups of the Holocene. Regional distribution maps unvariably show the close association of human occupations and hydrographic features. Obviously, the latter were important in huntergatherer subsistence but they may also have functioned as demographic corridors or sink/source refuges (Finlayson 2013). The aim of the project, essentially designed as a fieldwork project in the Flemish Valley, is twofold. First, and most of all, it intends to contribute to a better general understanding of the role of wetlands in the settlement systems of hunter-gatherers in northwestern Europe. Second, it wants to start building the empirical base to eventually test the far reaching hypothesis that the Flemish Valley of coastal Belgium was a demographic refuge for Neanderthal groups during the contact period with modern humans. While the present evidence to even remotely support this idea is extremely thin indeed, there are some observations to suggest that as a research hypothesis, it is not premature. Given the completely random character of archaeological surveys for Pleistocene sites so far, the number of Middle Palaeolithic findspots here is surprisingly high. Most of these locations are situated along the southern fringe of the Valley, adjacent to the limestone regions of the Meuse basin and northern France (Locht and Depaepe 2011) with their wellestablished Middle Palaeolithic occupations. Stratigraphic observations are almost nonexistent but, very prudently, it would seem that they mostly belong to the Hengelo interstadial. Thinking further in terms of demographic corridors on a latitudinal gradient, according to the rhythm of Pleistocene climatic variation, the Flemish Valley with its very distinct morphology may have been an attractive area for settlement. The Flemish Valley (Tavernier and De Moor 1974) is a ~2000 km² depression in northern Belgium that was eroded during the Saale glaciation. It has a large eastern tributary valley (the Eastern Branch) in which the present rivers Zenne and Demer flow. The Valley contains a deep sequence of Upper Pleistocene sediments, with basal peats of Eemian age around its fringes. The few recorded sections that exist indicate that this is a high resolution geo-archive at least for certain episodes during the Upper Pleistocene. Yet, there has never been an attempt at systematic survey for sites in stratigraphic context even if stone tools and probably associated faunal remains have shown up whenever quarry activities were carried out. Only in very recent times, on the occasion of the extension of the Antwerp port, primary context Mesolithic site clusters in the central part of the Valley have begun to be investigated (Crombé 2005). As far as knowledge of the Valley’s Pleistocene occupation history is concerned, the only evidence so far available are small out-of-context lithic assemblages randomly collected by amateur archaeologists. The project proposes to set up a program of systematic field research in the Eastern Branch. Two portions of the present Demer basin are selected, in function of the two research questions mentioned above. The first is located upstream in the large plain of Halen-Schulen, the southeastern entrance to the Valley. At the site of Stevoort-Kanenveld, my research group has carried out a preliminary heritage project. It appears that this is a stratified Holocene site with several occupation levels ranging from Mesolithic into the Bronze Age, the first of its kind in the Eastern Branch. Pollen records from deep peat deposits that can be tied into the site stratigraphy show the effects of the 9,3 and 8,2 ka climatic downturns. Large scale excavations are imperative. The high resolution record preserved here will inform us on the function of alluvial plains in the settlement system of postglacial hunter-gatherers and their possible responses to such climatic events. The second research area, the Lower Demer valley between the city of Aarschot and the confluence with the river Dyle, is chosen in function of the Flemish Valley refuge hypothesis. In its present lower course the Demer is an incising meandering river due to the upstream existence of the Halen-Schulen plain mentioned above (De Smedt 1973). In that large floodplain the Demer deposits most of its sediment load allowing it to incise downstream. Whether there is a causal relationship to this specific geomorphology is at present unclear, but the Lower Demer valley is especially rich in prehistoric sites, including a number of Middle Palaeolithic findspots. Therefore, it is ideally suited to start a systematic survey of Middle Palaeolithic regional site distributions. Our aim is to locate primary context locales by means of an extensive and systematic deep sounding survey with a mechanical auger. Making use of the detailed geomorphological mapping of the Lower Demer Valley (De Smedt 1973) we shall single out zones where older Pleistocene terraces are preserved. There are indications that sites are associated with these terraces while Upper Pleistocene sediment covers are comparatively thin here. Thus, they should be within survey reach. Obviously, a full test of the refuge hypothesis would require an effort at a scale that largely transgresses the capacity of a 4yr project. Within the latter, our aim is to provide the first elements of solid empirical evidence, both in terms of data density and chronological resolution, that will ultimately allow to resolve a historic question of large importance for our understanding of recent human evolution.