A new "very large site" of the fourth millennium in the South of France: the contribution of the "Cazan l'Heritiere 2008" project in Verneques (Bouches-du-Rhone)

Few well-studied Neolithic open-air sites are known in the Provence region, in contrast with the Languedoc, middle Rhone and Garonne valleys regions, where many large Chassey culture settlements from the 5th and 4th millennia BC have been studied during the last decades. In the municipality of Vernegues, repeated preventive archaeology operations (diagnostics and excavations) in the last twenty years have allowed the partial exploration of a vast Middle and Final Neolithic settlement. We propose naming this settlement `Cazan', after the nearby hamlet. With a surface area of at least 160 hectares, the site is located in a little sheltered valley in the hills to the south of the Durance River valley. Its geographical position, between the Alpilles hills to the west, the Crau plain to the south-west, the Durance valley to the north and the Aix-en-Provence basin to the south-east, as well as its large surface area, suggests that this site could have played a central role in the organization of the Chassey culture territories in western Provence. The Cazan site is currently documented by 23 discovery points, including four planimetrical archaeological excavations, representing almost 2.7 hectares of Neolithic occupations brought to light. The density of remains is variable, hut the small surface actually excavated (less than 2% of the site's estimated total area of 160 hectares) does not allow a comprehensive approach to this vast occupation. At the most, all that can be said is that this type of extensive site is rare in southern France. In this paper, we present data from an archaeological excavation that took place in 2008, in a locality called `L'Heritiere'. During the four weeks of this preventive archaeological operation, 5600 m(2) were excavated. The data collected allowed several questions related to these megasites to be answered: timespan and nature of occupation, detailed chronology and site functions. The geomorphological study allowed the insertion of the archaeological remains in a complex stratigraphic sequence deposited in a slight depression inherited from the Ice Ages, in which the highest density of remains were found, to be understood. The current gently sloping topography and the lateral variations in the sedimentary sequence demonstrate that the natural and archaeological layers are not evenly preserved, some of the areas having suffered more than others from erosion. In contrast, some areas show more important sedimentary deposits, where archaeological remains have been preferentially trapped and well-preserved. Among the 35 actual features discovered, 15 storage pits, 2 or 3 water wells, one artefact deposit, 3 heated stone features, 8 to 10 fire pits, 2 hearths and what is probably wedging for a post can be distinguished. The pits are of average volume, and, on the whole, have a quick and non-anthropogenic filling. The two wells have been used as disposal places for artefacts and various organic remains, and provided two stratigraphic sequences rich in information on the nature of human activities. The fire-pits display evidence of burning on their bottom and walls, but some of them have probably been disturbed, as the stones partially filling the pits are not always in primary position. The presence of only two hearths and one post wedging suggests activities which left little evidence on the ground. The ceramic production study outlined four morphological groups, interpreted as as many occupation phases. Technological data demonstrate, for each of the groups, the use of local raw materials but with some variability in the fabrication processes (`chains operatoires'). The latter could point to the existence of several domestic production units using different methods. The technical and morphological specificities observed could be interpreted as a local style variation but overall, the four phases identified on the site can be placed in the existing regional seriation of ceramic production. The lithic collection, a series of 745 elements, demonstrates that the site was part of several exogenous raw material exchange and distribution networks (Sardinian obsidian, Bedoulian flint from Vaucluse, Oligocene flint such as the Apt-Forcalquier type). The typo-technological analysis indicates two successive Chassey chronological phases. The oldest phase is only represented by scattered artefacts while the recent phase is more largely represented, but it was not possible to distinguish any stages within it. Cross-examining the data provided by the stratigraphy, ceramic seriation and the seven radiocarbon dates suggests four Chasse", culture phases during the occupation of the THeritiere' site: 1) an older phase before 4000 BC, attested by scattered artefacts and a radiocarbon date conflicting with the related artefacts; 2) a following phase with both 'Terms Longues' and `Rocalibert' type Chassey culture stages (C and DI stages of the regional ceramic seriation) during which the wells were dug and then filled up again, one stage also being associated with pits; the dating ranges between 3950 and 3790 BC, that is to say a century and a half at most; 3) a final phase, related to the Pertus' type Chassey culture (stage D2), associated with the fire-pits, which took place between 3790 and 3690 BC, lasting a century at the most. The spatial distribution of artefacts does not reveal any specific polarity that could be related to one of the occupation phases. In other words, on the 2008 archaeological excavation scale, we are dealing with repeated and extensive occupations over three centuries, each occupation having left only very few remains. This result suggests a non-intensive occupation model, but a real appropriation of the location. The high chronometric resolution achieved in the `L'Heritiere' site study demonstrates the usefulness of applying an integrative approach to resolve spatial temporal discontinuity and continuity issues on archaeological sites. Future research directions include the confrontation of our interpretation with other sectors excavated on this megasite and further analyses on archaeological material to define subsistence patterns, raw material provenance, and their evolution during the Chassey period (translation Alexis Taylor and Dorcas Vannieuwenhuyse).