What about Bordes and de Sonneville-Bordes' Azilianisation model after thirty years of archaeology? At the end of the 1970's F. Bordes and D. de Sonneville Bordes proposed the progressive Azilianisation model for Tardiglacial human populations in Dordogne Valley based on the sites of Morin and Gare-de-Couze. More recent excavations of key series such as Pont d'Ambon by Guy Cé lé rier or Bois Ragot by Andre Chollet contrasted with this paradigm by characterizing an internal Azilian phasing with no Magdalenian elements. As part of the ANR Magdatis project, we reassessed the emblematic series of Morin in order to test the model advanced by F. Bordes and D. de Sonneville-Bordes in 1979. Abri Morin (Gironde, France) was excavated by Deffarge at the end of the 1950s. The site stratigraphy proposed by Deffarge consists of two main units (A and B). The lower unit B is divided into two levels, BII and BI. Unit A is divided into four levels (AIV at the base to AI at the top). The site yielded a large collection of lithic objects attributed to the Upper and Final Magdalenian (Magdalenian V and VI), as well as many faunal remains (Delpech, 1983), a large quantity of worked bone and antler (Deffarge et al. 1974 a-b) and/or engraved objects (Deffarge et al. 1975; Guy 1993; Paillet and Man-Estier 2014). Prior to our work, only one radiocarbon date was available (Delibrias et al. 1974; Delibrias et Evin 1974). This date (Gif-2105 10 480 ± 200 years BP) appeared to be very recent in comparison to the chronology of archaeological complexes related to the end of the Magdalenian. The reassessment of the lithic material focuses on the armatures. More than 2,000 remains were studied but our analysis is limited to the distribution of five morphotypes in the stratigraphic sequence proposed by Deffarge (Figure 2). Unit B can be attributed to an early phase of the Upper Magdalenian (Figure 4), but unit A contains unprecedented morphotype associations, comprising typically Magdalenian elements (shouldered point, Laugerie Basse type point) with elements characterizing the early Azilian (backed bi-point), the recent Azilian (backed point with tapered base) and the Laborian (backed Malaurie type point). The faunal remains were studied by Delpech (1983) and are mainly composed of reindeer remains associated with Bovine (Bison and/or Aurochs), horse, red deer, and to a lesser extent roe deer and wild boar remains. Avian remains are abundant and diversified (Mourer-Chauvire 1975; Delpech 1983; Gourichon 1994), with a predominance of snowy owl remains. The reassessment of the faunal material resulted in the identification of more than 5,000 remains (Table 1). The proportion of the different taxa is not significantly different from previous publications. Three additional taxa were identified from several remains. These are the dog (Canis familiaris, Boudadi-Maligne et al. 2012), and in lesser proportions, the mammoth and the chamois. Refits were carried out on the faunal material. Most of the associations were from objects from the same level but some of them concern objects found in different levels and in both units (A and B, Figures 5 and 6). In the light of these results, twenty-two faunal remains of known stratigraphic provenance were radiocarbon dated (Table 2). These dates enabled us to test the stratigraphy proposed by Deffarge, but also to document the different problems linked to the appearance or disappearance of different taxa from the region (Szmidt et al. 2009; Boudadi-Maligne et al. 2012; Costamagno et al. 2016), or the dates of the different technocomplexes from the end of the Tardiglacial (Barshay-Szmidt et al. 2016), or the evolution of armature elements made of antler (Pe tillon 2016). The dating results were calibrated with the OxCal software, version 4.3.2 (Bronk Ramsey 2009) with the IntCal13 curve (Reimer et al. 2013) (Figure 7). The cross-analysis of the dates obtained in relation to the stratigraphy proposed by Deffarge reveals numerous discrepancies. The organization of the dates by chronological order is not consistent with the stratigraphic order of the Deffarge levels (figure 7). However, two units can be observed. The first comprises dates between 16 ka calBP and 14 ka calBP. The second comprises four dates obtained on a barbed Azilian type point ("flat harpoon": Deffarge et al., 1974b, Fig. 6, n° 8), a rabbit remain and two wild boar remains. In comparison with the corpus of dates from the different technocomplexes from the end of the Paleolithic (Barshay-Szmidt et al. 2016; Fat Cheung et al. 2014; Langlais et al. 2012, 2014), these four dates can be excluded from what is usually found during the Upper Magdalenian. They correspond more to the early Azilian (wild rabbit), the recent Azilian (wild boar) and the Laborian (barbed Aziloid type point). These results imply that the Morin stratigraphy was not limited to an occupation by the last Magdalenians and must have been much more complex than previously thought. It is very probable that the site was occupied not only by the last Magdalenians, but also during an early and recent phase of the Azilian, and again by Laborian groups. We thus refute the Azilianization model proposed by Bordes and de Sonneville-Bordes and we support hypotheses of gradual evolution brought to light at well studied sequences such as Pont d'Ambon du Bois Ragot, Murat or Rochereil.