During the summer of 2009, an important archaeological campaign was carried out in the heart of the downstream Seine region. It allowed us to locate a well-preserved occupation site attributed to the Upper Magdalenian. The operation concerns a surface area of 103 hectares located from the left bank of the Seine along the edge of the river and slightly downstream, to the most concave part of a meander which gives a strongly asymmetrical aspect to the bottom of the valley. The Magdalenian occupation was discovered in an alluvial sandy silt layer which provides a favourable context for the preservation of archaeological remains. The occupation is located on the edge of a former channel, probably restored during the Late Glacial. A gravel ridge has protected the deposits from the Holocene alluvial erosion. At present, only a surface of about 10 square meters in the main find concentration has been uncovered, but a quantity of test pits show that the archaeological layer extends over nearly 200 square meters. We provisionally suggest a possible relationship of contemporaneity between this occupation unit and two other sets of remains (a horse skull and a stony structure) discovered a few hundred metres away, in a similar stratigraphic position. With the exception of a few fragile remains that were removed during the field work, the other objects have been left in situ and the entire occupation site has been preserved for later excavations. The archaeological remains are mainly composed of reindeer bones and antlers, flint tools, backed bladelets and lithic remains, an antler point as well as a probable shell adornment. More than two hundred flint artefacts were removed, among which we found two burins. The raw materials exploited by the Magdalenians consisted mainly of a flint of the Late Cretaceous epoch, probably stemming from the Campanian and Santonian stages. Furthermore, the presence of a tertiary material has been detected. This raw material, strongly recalling the red flint from the Ludian stage of the Eocene epoch, has already been found in several other regional occupations (blades or other tools). Among the artefacts preserved in situ, we identified several other burins and, to a lesser extent, scrapers and piercers. A few backed bladelets were mixed with the debitage products close to the northwest edge of the hearth. A reindeer antler point was also removed from the rest of the archaeological floor before the excavated surface was refilled. This antler point was discovered less than 2 metres from the west edge of the hearth and a few centimetres away from a large worked reindeer antler fragment. It has a subrectangular section and presents a double bevel as well as a longitudinal groove. The morphology of this point is very similar to other examples known in the regional Magdalenian. All the rocks have been left in situ, and the information concerning this category of artefact is only based on field observations. They are characterized by their morphological and petrographical diversity, and include mainly sandstone, sandstone-shale and limestone. Apart from rocks contained in the hearth, probably reserved for its functioning, some other large blocks could have composed elements of the infrastructure in view of their size and weight. Such elements are known in the other Magdalenian occupations of the Paris Basin where they sometimes delimited a circle several metres in diameter. One of the main interests of this occupation lies in the presence of numerous bone remains. Almost all the determined elements relate to the reindeer. The discovery of reindeer tarsi in anatomical connection and the presence of skeleton elements particularly sensitive to taphonomical depletion, like antler fragments, ribs and vertebrae, obviously show exceptional preservation conditions and the absence of post-depositional disturbances. A first estimation based on teeth series and forelimbs (the radius fragments), gives a Minimum Number of Individuals of two reindeers brought back to the occupation site. At present, the absence of teeth series belonging to very young individuals prevents a more accurate seasonal determination of the Magdalenian visits. However, the presence of a large-sized antler from a hunted animal, belonging to an adult male reindeer, permitted the processing of this specimen to be estimated at between September and mid-December. The occupation would thus have taken place in the autumn or at the beginning of the winter, unless it is considered that the reindeer antler could have been transported for several months as a stock of raw material. Besides the reindeer remains, we observed a diaphysis fragment with dimensions suggesting an attribution to a horse or a bovine, and the horse skull discovered outside the occupation site. All the artefacts are located around the pit hearth, close to a large ochrestained area. The low density of remains suggests a short-term residence. A small pit was identified close to the hearth. Considering the bone remains, we note an asymmetrical distribution in relation to the hearth. There are a lot of bone fragments to the south of the fire area, while they are almost absent to the north. Such a pattern probably points to the repetitive disposal behaviour of the prehistoric people. Two debitage product concentrations, related to bladelet production, are located to the north-east of the hearth. Two other flint concentrations seem to be more probably toss zones. Finally, we observe a group of tertiary flint blades near a large sandstone slab. This group could be interpreted as a spare space of unused blades. With regard to the spatial organization, some levels of Pincevent (Seine-et-Marne) offer the most striking parallels. Two bone fragments were sent to a laboratory in order to obtain radiocarbon dates. The first one, a reindeer radius fragment which was collected close to the hearth, delivered a date of 12,000 ± 70 BP. The second one, a cheek tooth from the isolated horse skull, delivered a date of 12,230 ± 70 BP. These results are very similar to those obtained for the nine other Paris Basin Magdalenian sites. We hope that OSL (Optically Stimulated Luminescence) analysis (in progress) could provide this Upper Magdalenian occupation with an accurate chronological context through the Weichselian Late Glacial chronozones. The outstanding conservation of this occupation site already makes La Haye one of the most important site for regional Magdalenian studies. The analysis of this occupation will be completed by exhaustive excavations. Since the planned building project has been cancelled, the site area is no longer threatened. However, new excavations are scheduled for the coming years., Une opération d’archéologie préventive réalisée durant l’été 2009 dans le département des Yvelines, au coeur du territoire Seine-Aval, a permis de repérer une unité d’occupation très bien conservée qui est rapportée à la phase supérieure du Magdalénien. Elle a été découverte dans une couche d’alluvions limono-sableuses qui s’est révélée un contexte particulièrement favorable à la conservation des vestiges. Actuellement, seuls une dizaine de mètres carrés ont été fouillées, mais les sondages réalisés dans les environs immédiats indiquent qu’elle se développe sur une surface d’environ 200 mètres carrés. À l’exception des éléments les plus fragiles qui ont été prélevés, les vestiges ont été conservés in situ pour une fouille ultérieure. Ils se composent principalement d’ossements et de ramures de renne, d’outils et de déchets de taille en silex, d’une pointe de sagaie ainsi que d’un probable élément de parure. L’ensemble se répartit de manière structurée autour d’un foyer en cuvette aménagé au moyen de blocs rocheux et à proximité duquel s’étend une nappe d’ocre rouge. La faible densité des vestiges évoque un séjour plutôt bref que les premiers indices de saisonnalité tendent à placer à l’automne. L’équipement technique et la structuration de l’espace habité permettent d’établir de nombreuses comparaisons avec quelques autres gisements magdaléniens de plein air du Bassin parisien. Par ailleurs, nous suggérons provisoirement une possible relation de contemporanéité entre cette unité d’occupation et deux autres ensembles (un crâne de cheval et une structure pierreuse) découverts à quelques centaines de mètres de distance dans une position stratigraphique analogue. Les aspects chronologiques sont d’ores et déjà documentés au moyen de datations radiocarbone obtenues sur os. Nous fondons aussi de nombreux espoirs sur les résultats d’analyses OSL (Optically Stimulated Luminescence) en cours qui pourraient permettre un calage plus précis de l’occupation au sein de l’une des chronozones du Tardiglaciaire weichsélien., Debout Grégory, Le Jeune Yann, Djema Hélène, Bignon Olivier, Charier Marie-Aline, Peschaux Caroline. La découverte du gisement de la Haye aux Mureaux (Yvelines) et ses implications sur la connaissance du peuplement magdalénien d’Île-de-France. In: Bulletin de la Société préhistorique française, tome 108, n°2, 2011. pp. 221-246.