Cet article est consacré à l’étude de la conservation des ossements animaux dans un amas coquillier mis en place sur une période longue, de l’Âge du Bronze à la période romaine, au centre de l’île d’Ouessant. Le terrain acide (granit) ne permet pas la conservation des ossements animaux, et c’est grâce à la présence du calcaire de nombreuses coquilles de berniques qu’ils ont été préservés. Une approche cartographique de l’ensemble du dépôt permet d’en restituer les modalités de mise en place pendant plus d’un millénaire et de rendre compte des phénomènes de conservation différentielle. C’est ainsi que l’on montre que la préservation est meilleure au centre de l’amas, où se situe l’essentiel des restes de poissons, que sur les bordures, où subsistent essentiellement des restes de bœufs.

This article addresses the impacts of differential preservation in an alkaline shell midden overlying an acid soil which does not allow bone to survive. This midden accumulated over an extended period and displays unusual traits attributable to practices of a ritual nature (Le Bihan and Méniel, 2002). Excavation extended to approximately 1000m². Bone was collected on a quarter metre grid by individual context. Some 35,000 animal and bird bones were recovered; fish bones, which were very numerous, are presently still under study. Consideration of the entire bone assemblage indicates that the principal deposition occurred along the SW/NE axis of the midden, with a lesser dump on the NW (zone XLI). The sector with the greatest concentration of bones covered some 50m²; there, fragile remains, notably those of birds and fish, were at their most frequent. The distribution of ovicaprid bones is more uniform. Cattle and pigs are more frequent on the edges of the midden. Overall, there was a noticeable gradient in the quality of bone preservation between the core and the periphery of the deposit. The principal dump of bones displays two remarkable characteristics: a great quantity of shoulder blades and a preponderance of right-side remains over their left-side equivalents. The chronology of the accumulation of the midden shows four major phases: Bronze Age, Hallstatt Iron Age, La Tène Iron Age and Gallo-Roman. Plotting the bones by phase shows that the prime area of the deposit moves from NE to SW over time. It is thus possible to track the evolving rules of selection of species, individuals or animal parts from their distributions on the ground. This is particularly true for domestic pig, which increases in importance over time, and is encountered especially towards the south of the midden. Some bones show clear signs of alteration attributable to their exposure to the weather, making it plain that the midden was not covered with soil during its accumulation. The impact of such degradation on the bones can be very visible. The fact that some bones display signs of weather-related damage on more than one facet indicates that there was considerable secondary manipulation of this material. This study demonstrates the value of coupling a standardized but rapid method of data collection (by quarter metre squares) with detailed cartographic analysis of the material recovered. Both the taphonomic effects on a deposit built up over a very long time scale and continuity in certain practices (preferential selection of right shoulder blades) are clear from an initial consideration of the data.