The late Middle Palaeolithic in Southwest France: new TL dates for the sequence of Pech de l'Azé IV
The relative and numerical chronological position of the technological and typological variants of the Mousterian in southwest France has been the subject of debate for over fifty years. Since the advent of both ESR and TL dating methods in the 1980s, a database of chronometric dates for a growing number of sites has been steadily accumulating. A recent summary by Guibert et al. appears to show a complex pattern of broadly overlapping Mousterian variants in the late Middle Palaeolithic and has led some to conclude that Bordes' initial interpretation of these variants, as distinct cultural groups, was essentially correct. This paper adds to this database with new thermoluminescence dates from Pech de l'Azé IV. This site, originally excavated by Bordes in the 1970s, contains a deep sequence of Mousterian assemblages, which express considerable technological and typological variability and which are associated with well-preserved faunal remains. Excavations were undertaken from 2000 to 2003, in part to gain a better understanding of the geological context of the assemblages and also to obtain chronometric data. In this paper, new TL dates for the levels 3B, 4C, 5A and 8 are presented. The deepest of these, the Typical Mousterian of Layer 8, is placed in Marine Isotope Stage (MIS) 5c. Another Typical Mousterian industry from level 5A dates to the transition from MIS 5 to MIS 4. Level 4C, which is rich in scrapers, is placed in MIS 4, while the Mousterian of Acheulian Tradition is attributed to MIS 3. At a local scale, these new dates allow for the integration of the Pech de l'Azé IV sequence with chronometric dates available for the nearby sites of Pech de l'Azé I and II, and at a more regional scale they provide additional data points from the more poorly dated late interglacial and early MIS 4 period. These dates are in line with the emerging pattern suggesting that the Mousterian variants, as they are defined, overlap considerably in time, and call into question their interpretation as a succession of chronological units, while a correlation with climate change of the technological units is not clear either.