A multi-method luminescence dating of the Palaeolithic sequence of La Ferrassie based on new excavations adjacent to the La Ferrassie 1 and 2 skeletons

Abstract

a b s t r a c tA new interdisciplinary project was initiated to excavate a portion of the Palaeolithic site of La Ferrassie left intact by earlier excavations. One of the aims of this project was to provide chronological information on the succession of Middle and Upper Palaeolithic layers, as well as on the skeletons unearthed by Capitan and Peyrony in the early 1900's. We report here preliminary results on the lithics, faunal remains, site formation processes, and on the stratigraphic context of the La Ferrassie 1 and 2 skeletons that were found adjacent to our excavations. Finally, results from luminescence dating of the sediments and a preliminary set of radiocarbon ages are presented. Quartz OSL, both at the multi-grain and single-grain levels of analysis, and post-IR IRSL of feldspar at various stimulation temperatures are compared. The quartz/feldspar comparison revealed a bleaching problem for the quartz OSL (and the feldspar pIRIR signals) from Layer 2; as a consequence, the age of this Layer was determined using a minimum age model.A Mousterian industry with bifaces, at the base of the sequence, has been dated between 91 ± 9 and 44 ± 3 ka. The Ferrassie Mousterian layers are attributed to MIS 3, between 54 ± 3 and 40 ± 2 ka, and thus appear very late in the final Middle Palaeolithic of the region; furthermore, these ages constrain the chronology of the La Ferrassie 1 and 2 skeletons, which have been attributed to one of these Ferrassie Mousterian layers. The Châtelperronian layer is dated to 42 ± 3 ka and the Aurignacian to 37 ± 2 ka. Implications of the ages for the La Ferrassie 1 and 2 skeletons, and for the variability of late Mousterian, are discussed.

Testing the accuracy of a Bayesian central-dose model for single-grain OSL, using known-age samples

“…Samples 092201, -02, -03 and -04 come from Middle and Late Palaeolithic layers from the Bordes-Fittes rock shelter (France), and the expected ages range from 23 to 45 ka ( 14 C ages; Aubry et al, 2012;. Finally sample FER 3 is a colluvial sample from the Late and Middle Palaeolithic La Ferrassie rock shelter (France), with an expected age of 46 ka ( 14 C age; Guérin et al, 2015b).…”

The complementarity of luminescence dating methods illustrated on the Mousterian sequence of the Roc de Marsal: A series of reindeer-dominated, Quina Mousterian layers dated to MIS 3

“…The Single Aliquot Regenerative dose protocol (SAR protocol; Murray and Wintle, 2000) employed in these single grain measurements included a post-IR green signal (1 s stimulation with green laser at 125 °C following a 40 s stimulation with IR diodes at 125 °C) and is similar as that used by Guérin et al (2015a). A preheat at 260 °C for ten seconds and cutheat at 220 °C were used before, respectively, regenerated (and natural) and test dose signals.…”

“…Thomsen et al, 2003), some authors argue for strict rejection criteria (e.g., Yoshida et al, 2000;Jacobs et al, 2006, and references therein) based essentially on the response of grains to a number of quality-insurance tests of the SAR protocol, such as recycling, recuperation and IR depletion ratios (Murray and Wintle, 2000;2003;Wintle and Murray, 2006;Duller, 2003). Others have reported that the main effect of such rejection criteria is to reduce the number of selected grains with negligible effects on the estimated D e and overdispersion (OD) parameters, which might thus lead to a loss of robustness in the final results (e.g., Thomsen et al, 2012;Guérin et al, 2015a;Geach et al, 2015;Hansen et al, 2015;Kristensen et al, 2015;Zhao et al, 2015;Thomsen et al, 2016).…”

“…However, several studies have recently argued against this model, based on different strands of evidence. Thomsen et al (2016) and Guérin et al (2015a) presented single grain OSL ages in better agreement with independent age control when calculating Unweighted Arithmetic Means (UAM) of individual equivalent doses, compared to Central Dose Model (CDM)-based ages following Bailiff et al, 2013, we prefer to use the terminology Central Dose Model, rather than Central Age Model as this model only concerns dose estimation -not age). While the UAM approach has the drawback of ignoring analytical sources of uncertainty, Combès et al (2015) proposed a Bayesian model for the statistical analysis of SAR OSL data (hereafter, this model will be referred to as the BaSar model).…”

“…overdispersion), despite a significant reduction in the number of selected grains. Thus, we decided, as did Guérin et al (2015a), not to apply these criteria in the analysis of our samples.…”