DACHY Tiphaine, GUÉRET Colas, CAMPMAS Émilie, SIMONNET Robert, BON François, PERRIN Thomas (2018) ‒ Saint-Trivier/Chabet el Houidga (Mascara, Algérie) : nouvel éclairage sur un faciès méconnu du début de l’Holocène : le Columnatien, Bulletin de la Société préhistorique française, 115, 2, p. 215‑251.
Résumé : C’est à la fin des années 1950, durant la guerre d’indépendance algérienne, que Georges Simonnet entreprit plusieurs sondages, sous forme de tranchées, sur le gisement de Saint-Trivier, à Mascara (Algérie). Le site se présente sous la forme d’une accumulation de vestiges d’occupations préhistoriques plus ou moins cendreux et de niveaux coquillers, en bordure d’un ravin abrupt l’ayant partiellement entaillé. Les sondages réalisés ont livré de nombreux vestiges de silex taillés, d’ossements animaux et de gastéropodes, ainsi que plusieurs aménagements, fosses et foyers. Issues d’une zone géographique à l’écart de l’essentiel des gisements algériens connus pour le début de l’Holocène, ces données s’avèrent particulièrement précieuses. La restitution du déroulé des opérations de terrain, des choix méthodologiques et l’examen des documents photographiques permettent de replacer les vestiges archéologiques au sein d’une stratigraphie de plus d’un mètre, scandée de plusieurs occupations distinctes. Adossée à plusieurs datations radiocarbone, l’analyse de l’évolution des industries lithiques et osseuses suggère que l’occupation principale se rapporte à un faciès particulier de l’Épipaléolithique méditerranéen d’Afrique du Nord : le Columnatien. Individualisé dans les années 1970, celui-ci n’était pour l’heure documenté que par deux gisements : Columnata et le site du Cubitus. Ces nouvelles données permettent d’apporter un éclairage inédit sur ce faciès particulier, daté du 8e millénaire avant notre ère, en Algérie occidentale.
Mots-clés : Columnatien, Ibéromaurusien, Épipaléolithique méditerranéen, Élassolithique, Algérie, industrie lithique, tracéologie, malacologie, datations radiocarbone, modélisation bayésienne, stratigraphie.
Abstract: At the end of the 1950s, during the Algerian War of Independence, Georges Simonnet, a french military officer and archaeologist, excavated several test pits at the site of Saint-Trivier, at Mascara (Algeria). The site was discovered accidentally during the development of a military camp. It consists of an accumulation of remains of more or less ashy prehistoric occupations and shell levels, located at the edge of a steep ravine that partially carves into it. Three main trenches were dug out by G. Simonnet, resulting in the discovery of abundant knapped flints, vertebrate remains and mollusc shells, as well as several pits and hearths.
This site is located in a geographic zone away from most of the early Holocene Algerian sites, and therefore the data are particularly valuable. After the war, the excavator returned to France with his collection, and the latter was left in the family cellar for several decades, before being temporarily handed over to the TRACES laboratory of Toulouse. A first examination of a sample of these remains was carried out as part of the collective programme funded by the IDEX Toulouse project ?€?MeNeMOIA – Du Mésolithique au Néolithique en Méditerranée occidentale : l’impact africain – From the Mesolithic to the Neolithic in the Western Mediterranean: the African impact?€?.
In an initial phase, the examination of the photographs, the slides, the test-pits plans, the two section drawings and rare written notes enabled us to roughly reproduce the field operations. Rigorous methods were applied to the excavations and the sediment was finely sieved. The archaeological remains, mainly made up of flint, bone and shells, were practically all stored in small boxes indicating the excavation units they came from. Based on these data, we could propose a spatial distribution of the diverse categories of remains for each excavation unit. In this way, two, and in some places three large levels of accumulations were identified in a stratigraphy with a thickness of about one metre. They are demarcated by what the excavator referred to as ?€?paving?€?; sub-horizontal piles of (seemingly unburnt) rough blocks. The identification of several clearly different levels, as regards the sediments and the density of the remains, points to the presence of several successive occupations under a superficial reworked level (layer 1).
The examination of the remains confirms this impression. The knapped lithic industry was studied over a test area of one square metre. It was mainly made on small flint pebbles, and shows a clear evolution in the operative debitage schemes and the typological range of tools. In this way, we can distinguish the first occupations at the very base of layer 6, characterized by a debitage oriented towards the production of arched backed bladelets with truncated bases. Above this, a second occupation phase (upper part of layer 6) is rich in trapezes or isosceles triangles, as well as in sharp bladelets with straight retouched backs and retouched bases, which are also present at the base of layer 5. However, this latter layer differs in that it contains a higher number of backed bladelet fragments and microburins. The following levels, from layer 5 to the lower half of the paving of layer 3, make up a coherent complex. They contain more or less elongated scalene triangles, sharp bladelets with rectilinear backs and rounded bases and blades with partially backed edges. Microburins, backed bladelet fragments and bladelets with arched retouched backs are still numerous. Above this complex (the upper half of the paving of layer 3 and layer 2), the levels are very different and are characterized in particular by bladelets with arched tips. Segments of circles are present in all the layers. They comprise hyper-microlithic specimens but also large objects with traces of impacts showing that they were used as cutting armatures. In addition, the taphonomic conditions allowed for the preservation of the original functional marks on blades, resulting in particular from plant working.
The bone industry is represented by pointed tools and tranchets with oblique bevels, although the latter are only present in the main occupation (layers 3 and 4).
Among numerous vertebrate remains, the few that have been identified for now include gazelle, Barbary sheep and hartebeest. The continental malacofauna is abundant and contains various species. A few marine mollusc shells are also present, including at least one shell probably used as an ornament.
Several radiocarbon dates were carried out, mainly on bone apatite, as collagen was not preserved. The Bayesian modelling of the results obtained in the stratigraphic sequence places the first occupations towards 11 ky BC cal., whereas the central occupations date to the second half of the 8th millennium cal. BC. The age of the upper levels is less clear; they probably date to the 6th millennium cal. BC.
In the context of our knowledge of the recent prehistory of North Africa, these diverse observations suggest that the lowest occupations, which have not yet been extensively studied in detail, are probably related to a final Iberomaurusian complex. The main occupations can be compared to a specific Mediterranean Epipalaeolithic facies in North Africa; the Columnatian. This was identified in the 1970s, and up until now has only been recorded at two sites, Columnata and le Cubitus. These new data thus provide an unprecedented insight into this particular complex, dating from the 8th millennium cal. BC, in Western Algeria.
Keywords: Columnatian, Iberomaurusian, Mediterranean Elassolithic, Algeria, lithic industries, use wear analysis, malacology, radiocarbon dates, Bayesian models, stratigraphy.