Quina or not? Techno-economic review of a Mousterian charentien site in eastern Languedoc: the Roquette cave in Conqueyrac (Gard, France)

For years, the Quina Charentian industries from the Languedoc region, and more widely from South-Eastern France, have been the subject of long discussions due to their typological particularities which set them apart from the series of the classical Charente region. Technological reappraisal of this cultural facies allowed the identification of a distinct technical entity in South-Western France. Now there is a consensus to reserve the term 'Quina Mousterian' for lithic industries which combine a specific core reduction system (called the Quina method) with particular processes of tool retouching, resharpening and recycling. The presence of this techno-complex must now be tested in other regions by a reassessment of assemblages formerly attributed to the Quina Charentian facies. The analysis presented here of the industries from Grotte de la Roquette is part of this reappraisal project, partly because it is the only recently excavated 'Quina' site in the Languedoc region. Our technological and techno-economical analysis of the lithic assemblages has led us to refute its initial attribution to the Quina Mousterian. Despite a low density of artefacts and little evidence of in situ knapping, the core reduction strategies were identified. First of all, these lithic series provide no characteristic core or any diagnostic products of the Quina flaking system such as thick large and often cortical blanks with an asymmetrical cross-section, or flakes with a dihedral asymmetric butt or with the typical 'lisse a pans' (smooth faceted) butt. The lithic industries are on the contrary mainly composed of relatively thin and elongated flakes resulting from recurrent centripetal production systems, although with differences according to the raw material. Quartz pebbles collected in the river just below the cave show the use of a discoidal flaking method, which was relatively commonly used to exploit such materials. Constituting the major part of the assemblage, chert and flint have on the other hand an undeniable Levallois component evidenced by the joint presence of exhausted Levallois cores, technical products like "debordant" flakes and some characteristic end-products. However, a lack of Levallois blanks, especially typical flakes, is clearly highlighted by the technological counts. Their proportions in the lithic assemblages vary from 8 to 15%, according to the layer and type of raw material. At the same time, however, characteristic products of the discoidal method such as pseudo-Levallois points, flakes with a 'broken profile', thick quadrangular flakes or short flakes with large inclined butts, are also not well represented. It seems therefore that the partial deficit of Levallois products is rather the result of high flexibility in the knapping processes and of economic parameters related to blank import and export strategies. Indeed our techno-economic study has shown that interestingly these lithic industries are predominantly made on semi-local flints collected for the most part in the Vidourle Valley and introduced into the cave as tools, cores and end-or by-products. As suggested by the presence of retouching flakes and the weak evidence of blank production, in situ activities are mainly focused on the manufacture and the maintenance of this diversified imported equipment. The retouched tool kits, which compose a large part of the assemblage, are highly dominated by sidescrapers, some of them manufactured by scaled stepped retouches. However they are mostly of the 'half-Quina' type and affect different kinds of blanks mainly selected for their large size (thus offering a greater potential for reworking). Sometimes thick and asymmetrical, most often with peripheral cutting edges, these tool blanks also include exhausted cores. Although undoubtedly present in the lithic series, these stepped retouched sidescrapers, almost exclusively made on good quality non-local flint, are in much lower proportion than in typical Quina assemblages such as Marillac-Les Pradelles, for instance, especially for imported raw material. The ratios of Quina sidescrapers at La Roquette are on the contrary close to those observed within the exotic flint series from Les Canalettes, a clearly Levallois assemblage from Languedoc. The coexistence of Levallois core reduction with Quina tools in a same assemblage is unusual and when such a situation is identified, the homogeneity of the assemblage needs to be checked. In the case of La Roquette assemblages, the stepped retouched tools do not seem intrusive. They are homogeneously distributed throughout the layers; no spatially or vertically specific distribution reflecting possible visits of Quina groups was observed. Another argument may support this assertion. Raw materials for the Quina tools are exactly the same as for the rest of the assemblage (especially Levallois cores and flakes), even if it cannot be absolutely excluded that different groups roamed the same territories, thus maybe exploiting the same outcrops. Whatever the case, the composition of the retouched tool kits also differs from the typical Quina assemblages from South-Western France. Some tool types frequently represented in these industries, such as transverse scrapers with stepped retouches, 'limaces' and bifacially retouched scrapers ('tranchoir' type), are very rare here. Typical resharpening and recycling flakes (types III and IV) resulting from tool curation in the Quina system are also very scarce. However, most of the half-Quina and Quina retouched tools show abrupt lateral cutting edges, sometimes close to the percussion bulbs, indicating a sharp decrease in the blank dimensions due to several successive retouching phases. To sum up, the repetitive resharpening sequences, probably intended to extend the use-life of the imported lithic pieces, combined with the high proportion of retouched tools and the presence of retouching flakes, characterize the large imported component of these assemblages. It seems that the techno-typological characteristics of the La Roquette assemblages, which clearly individualize them from the classical Quina complex, could be the result of economic strategies related to the lithic technological organization, the site function and the settlement mobility pattern of the Mousterian groups living there. All the criteria taken into account indicate that La Roquette cave served for short-term occupations, with the introduction and maintenance of personal equipment which composed a large part of the lithic assemblage.