For many years, only few institutional researchers were interested in the countless engravings located in the sandstone shelters of the southern & Icirc;le-de-France. Discovered as early as the 1860s, especially in the Fontainebleau massif, these rock engravings form one of the largest concentrations of this type known in Europe. In 1975, the foundation of the "Groupe d'& Eacute;tudes de Recherches et de Sauvegarde de l'Art Rupestre" (GERSAR), a local dynamic association of volunteer researchers, has given decisive boost to researches concerning this very rich rock art heritage. During more than twenty years, its prospecting, inventory and study were almost completely carried out by GERSAR, with supervision from the French Ministry of Culture. From the 2010s onwards, the association was finally assisted by professional researchers. In this paper, we will focus in detail on the relationships developed over the last 9 years through a "Projet Collectif de Recherche" (PCR) concerning the alleged Paleolithic and Mesolithic engravings. We'll begin by describing the genesis of this collaborative project. Then, we'll take a very concrete look at the scientific cooperation between professional researchers and volunteers. From the very beginning of our common work, the latter have not only acted as guides and informants: the way in which each of them has introduced the participants into this vast heritage has greatly influenced the study strategies developed by the PCR. These strategies were intensive when the aim was to study in detail a few shelters which were considered to be emblematic, then they were extensive when we were interested in the broader distribution of engraved shelters across the whole region or in a few selected areas. As part of the PCR, some of the volunteer researchers have developed in-depth personal research projects, for instance on the distribution of sandstone quarries of historic age, as their extension constitutes a bias to be estimated before analyzing the spatial distribution of engraved shelters. Finally, we'll note that the PCR has benefited greatly from the pre-existing network of scientific relations forged by the GERSAR, and has since endeavored to develop it, which is the role of such inter-institutional teams. In addition to these research activities, we share a common responsibility for the preservation of these innumerable engravings, which are as fragile as the sandstone on which they are made. The shelters, which are impossible to guard, are highly threatened by the intense tourist frequentation of the forests in which they are found, which increases the risk of various kinds of irreparable deterioration (graffiti, destructive camping fires, abrasion of the walls by repeated visits, etc.). As a result, the collaboration between volunteers and professionals also involves a form of heritage protection, which consists of educating the general public, by all kinds of means, about the precautions to be taken to avoid damaging these fragile remains. For this purpose, PCR and GERSAR have developed partnerships with local institutions such as the "Office National des For & ecirc;ts" (ONF) and the "Parc Naturel R & eacute;gional du G & acirc;tinais fran & ccedil;ais" (PNRGF). Information for the general public is provided through scientific valorization, such as the co-creation of information boards, as well as educational activities for young people, which are partly carried out by another association, Ark & eacute;oM & eacute;dia. This association has already reached 7,500 children, most of whom live close to the engraved shelters, by organizing school workshops and visits to two shelters that can be entered without risk of damage. At the end of this paper, the question arises as to the nature of all this cross-fertilization (do they fall under the definition of "participatory/collaborative sciences" in the full sense of the term?). We'll also consider the perspectives they open up, as well as their limits in a context where discretion is sometimes the best way to protect sites that are obviously impossible to maintain under close surveillance.

An investigation of the Magdalenian occupation at Étiolles-Les Coudray (France) was conducted using geophysical methods. Based on ground-penetrating radar (GPR), electrical resistivity tomography (ERT), electromagnetic induction (EMI), and stratigraphic information, we present a reconstruction of the main sediment interfaces carrying the Magdalenian occupation. Étiolles-Les Coudray is one of the most important open-air campsites in the Paris Basin, where consecutive settlements distributed along the Hauldres stream were preserved by silts. The geoarchaeological goals were, in particular, the reconstruction of the ancient environment in which hunter–gatherers settled, providing spatialized known stratigraphies able to find an echo in the Seine Valley. Moreover, a focus on the capability of geophysical methods to detect archaeological features is also presented and discussed. We observed that the major reflections in the GPR records were generated from interfaces that have grain size variation: (1) the bottom of the Holocene colluvium and (2) the bottom of the upper Late Glacial silt. EMI and ERT show a very clear horizon associated with the upper Late Glacial silt, in some places even more clearly defined than with GPR. We confirmed the presence of a channel along the slope, placed under Locus 1, and a second channel of the same type globally following the paleotopography of Locus 2. We created a thickness map of the “beige sandy silt” and hypothesized a high probability of good preservation conditions of Magdalenian evidence. Finally, the detection of several localized diffraction hyperbolas in the GPR record offers the possibility to obtain the ground truth of the geophysical results in the near future and verify the nature (archaeological or geological) of these features.

The rock art assemblage in the Parisian basin is well-known for its extensive number of engraved sites that reflects a long-duration tradition spanning from Upper Paleolithic to modern times. By contrast, only a handful of rock art sites with rock paintings are currently known in the area. In this paper we present the results of the first comprehensive study of one of these rock painting sites, Croc-Marin (Montigny-sur- Loing, Seine-et-Marne) applying state of the art methodologies to reassess their exceptionality in the framework of this prehistoric graphic assemblage. The presence of similar conventions in pictographs and engravings suggest both were part of a common symbolic system. Some of these conventions could be deeply rooted in the last stages of the Upper Palaeolithic that subsequently transformed into the Mesolithic graphic repertoire in the area. This painted rock art site is a key element in understanding the local evolution of the graphic productions since the PleistoceneHolocene transition and during the Mesolithic, not only in the region, but at an European scale, reflecting the changes and transformation in the rock art traditions and the arousal of local particularities progressively becoming more and more different across Europe.

Anthropogenic engravings were made within more than 2000 small natural cavities found in the sandstone boulder fields of the central Paris Basin, in an area of 1800 km2 to west of the town of Fontainebleau. These engravings consist mainly of clustered rectilinear grooves arranged in vertical parallel series or organized in grids. These dominant classes of motifs form a unique type of rock art attributed to Early Mesolithic hunter-gatherers (-9500-6300 BCE in northern France) on a growing body of evidences. Here, we approach the so-cial context in which these abundant symbolic manifestations were produced by developing an experimental study guided by two main questions: what efforts (i.e difficulty of work and working time) and know-how were required for these engravings? On what occasions were they made? After quantifying the highly variable hardness of the Fontainebleau sandstone with a Schmidt Hammer, we selected natural blocks and blank cavities in recent quarries to conduct our experiments. These show that the engravable part of the sandstone corresponds to a friable cortex often less than 1 cm thick, which covers a core of hard sandstone that cannot be engraved. In the sandstone boulder fields, this friable cortex is only preserved on surfaces protected from weathering. Our experiments also reveal that curved grooves can be made without difficulty and that the omnipresence of rectilinear grooves is a cultural choice. Finally, our experiments show that these rectilinear grooves are very easy to engrave and make it possible to propose equations for estimating the time required to make archaeological motifs. The average engraving time of a groove has been estimated at less than one minute and less than 15 min for the most common grids. These results, placed in their archaeological context, allow us to formulate the following proposals: (1) the engravings attributed to the Early Mesolithic period were technically easy and could be made in rather short periods of time; (2) the cramped configuration of the engraved places and the absence of apparent distribution pattern of motifs on the walls points to the hypothesis of an accumulation of individual engraving episodes, possibly by several engravers; (3) the engraved cavities are abundant, easily accessible and connected to daily-life territories. We can deduce that the Fontainebleau rock engraving phenomenon is the result of numerous small, ordinary and stereotyped engraving rites, likely carried out by a large number of engravers.

This article begins by considering some very common notions in prehistoric archaeology. In this respect, it questions what current approaches and sources allow to describe and what their limits designate as inexpressible, momentary or definitive. The challenge is to establish an adequate, operative vocabulary. On the other hand, insisting at all costs on using sociological approaches can weaken this ambition, which requires highly meticulous prior palaeoethnographical studies. The fading of the meaning is, however, inevitable when discussing prehistoric 'cultures', and therefore the notion of 'traditions' is to be preferred to refer to the fragments or remains of these cultures, for which precise distribution ranges are difficult to identify. At the same time, it seems that the notion of 'cultural currents' is very well adapted to expressing the fluidity with which practices spread among mobile hunter-gatherers, who, according to contemporary analogues, have a reputation of social flexibility. Some reflections follow on what the impressive transmission of ideas in Europe during the late Palaeolithic and Mesolithic may reveal, focusing in particular on the fluctuating intensity of contact between the Magdalenian and late Azilian.

The Ségognole 3 carved shelter in Noisy-sur-Ecole (France) is known for its Palaeolithic-style panel engraved with two horses. They are arranged in file on either side of three slots, initially assumed to be natural, evoking a pelvic triangle. A thorough re-examination revealed the artificial character of the slots as well as numerous anthropogenic interventions to modify the hydrology of the shelter to drain water to the slot representing the vulva.The shelter has two galleries offset in height and separated by a thin wall. The upper gallery passes behind the engraved panel. It encompasses a basin that overhangs the pelvic triangle and often fills with rainwater. This disposition is the key part of the hydraulic system. Two particular interventions were made: (1) the slots at the front of the pelvic panel had been widened by percussion and grooving; and (2) the basin at the rear of the panel had been deepened by percussion, thus opening fractures on its floor. As a result, infiltration of meteoric water into the sandstone was encouraged and directed to the base of the pelvic triangle.An experiment demonstrated that filling the basin with 50 L of water caused the vulvar slot to drain after two days. The Palaeolithic people could potentially have operated the ‘vulvar outflow’ on demand and without weather constraints. It is likely that this setup had symbolic and ritual purposes and formed with the parietal arrangement an integral part of a complex and dynamic installation linked to the feminine sex symbol.

Le site de la Haute-Île à Neuilly-sur-Marne (Seine-Saint-Denis) se trouve sur la rive droite de la Marne, au sein d’un méandre situé à 15 kilomètres en amont de Paris (fig. 1). L’important potentiel archéologique a été mis en évidence à partir de 1999 lors des opérations préventives liées à l’aménagement de ce terrain en parc urbain. Ces premiers travaux ont révélé que le méandre correspond à une ancienne île conservant un paléosol holocène au niveau de sa paléoberge-est (fig. 1). Partiellement fouillé dans un cadre préventif (2000-2004) puis programmé (2011-2015), ce paléosol, dont l’extension est estimée à 3 400 m2 (fig. 3), contient les trois espaces typiques de bord de rivière : berge, rive (en partie organique) et chenal (fig. 2). Sur la berge, seul espace pour l’instant fouillé, les occupations sont regroupées au sein d’une même unité de sol complexe correspondant à des «sols cumulés » . Ce niveau est très riche en matériel archéologique accumulé avec peu d''apport sédimentaire et d’importants phénomènes de pédogénèse. Des études géostatistiques (par projection altitudinale de l’ensemble des artefacts et écofacts) ont néanmoins permis de repérer que cette couche conserve une microstratigraphie composée d’au moins deux niveaux «fantômes » (fig. 4). Bien que livrant des témoins fugaces d’occupations néolithiques et protohistoriques, ces «sols cumulés » sont surtout riches en vestiges mésolithiques avec des témoins d’activités funéraires et domestiques. Les sépultures mésolithiques de la Haute-Île sont pour l’instant au nombre de quatre (fig. 6, 7, 8 et 9). Réparties sur une surface minimale de 120 m2 (fig. 5), elles consistent en des inhumations primaires et individuelles réalisées dans de très petites fosses (aux contours ovalaires de moins d’un mètre de diamètre et aux profils en U à fonds plats). Les corps sont tous fortement contractés (tête inclinée avec les membres inférieurs et supérieurs repliés sur le torse) mais ils ont été déposés de façon peu normalisée (fig. 11) : en position redressée avec le buste vertical (sépultures 2 et 4), ou bien sur le dos (sépulture 1) ou encore sur le ventre (sépulture 3). La décomposition en espace vide ainsi que des effets de paroi ou de contrainte conduisent à formuler l’hypothèse de contenants funéraires souples et périssables, autrement dit d’une préparation des corps avant inhumation. À ce stade, la présence d’objets de parure n’est attestée que dans la sépulture 3 et ils se composent de coquillages fossiles et subactuels découverts en petite quantité à proximité du crâne, du bassin et du coude gauche (fig. 10). Les datations radiocarbone (AMS) obtenues placent les sépultures de la Haute-Île durant le 7e millénaire avant notre ère, soit à la transition entre le premier et le second Mésolithique (tabl. 4 ; fig. 18). Avec les découvertes de La Vergne en Charente-Maritime et celles des îles bretonnes d’Hoëdic et Téviec, le groupement de sépultures de la Haute-Île constitue la quatrième nécropole mésolithique connue sur le territoire français. Les pratiques funéraires particulières décrites à la Haute-Île se retrouvent sur plusieurs autres sépultures isolées du Bassin parisien, principalement attribuées au premier Mésolithique (tabl. 5), ce qui inscrit la nécropole de la Haute-Île au sein d’un possible modèle régional. Parmi les vestiges mis au jour dans les «sols cumulés » , une grande partie peut être rattachée au Mésolithique. Sur les 7 000 éléments d’industrie lithique, la très grande majorité des pièces attribuables se rapporte à cette période (tabl. 1). Seulement une trentaine d’éléments appartient au premier Mésolithique (composante A parmi les armatures) tandis que le reste est attribué au second Mésolithique (composantes B, C et D ; fig. 12 et 13). Les composantes B et C peuvent être rapprochées du Mésolithique récent. Elles se distinguent l’une de l’autre par des critères typologiques (respectivement trapèzes rectangles à base souvent concave versus trapèzes plus longs à base décalée) ainsi que métriques (lamelles étroites et minces versus lamelles plus larges et plus épaisses) ou lithologiques (usage principal de silex d’âge Mésozoïque versus Cénozoïque) et même spatiaux (répartition relativement dissociée : fig. 14 et 15). La composante D, peu nombreuse, se rapporte plutôt au Mésolithique final et se compose de pointes de Sonchamp, de trapèzes asymétriques et de triangles scalènes. Parmi les 17 000 restes de faune, les espèces sauvages dominent largement (tabl. 2 ; fig. 16) avec du sanglier (34,3 % du NRD), du cerf (22,8 % du NRD), du chevreuil (15,1 % du NRD) et de l’aurochs (5,5 % du NRD). Les taxons domestiques (boeuf, caprinés et équidés) sont en revanche très peu représentés (1,3 % du NRD) et rappellent la présence d’occupations discrètes postérieures au Mésolithique. Les sept mesures radiocarbone réalisées sur la faune sauvage se situent toutes dans la même fourchette chronologique (tabl. 4 ; fig. 18) : durant le 6e millénaire avant notre ère, soit au cours du second Mésolithique. Les stratégies d’acquisition et d’exploitation identifiées correspondent à une chasse majoritaire aux ongulés terrestres sauvages avec traitement des carcasses et consommation in situ (à l’exception peut-être de l’aurochs dont certaines parties squelettiques ont pu être exportées ailleurs). La présence du castor, de la loutre et de poissons témoigne de l’exploitation du milieu aquatique mais celle-ci paraît être minoritaire. Des fragments de poinçons en os, des canines de suidés refendues et des métapodes de cerf et de chevreuil travaillés signalent la présence d’une industrie en matières dures animales dont une partie au moins pourrait être prudemment rattachée au Mésolithique (tabl. 3 ; fig. 17). L’attribution à cette période de l’essentiel sinon de la totalité de la faune sauvage ainsi que de l’industrie lithique abondante et variée indique une occupation fréquente et répétée de la paléoberge-est de la Haute-Île durant le second Mésolithique. La nature du mobilier découvert (outils et armatures lithiques, restes de chasse et de boucherie, possibles pièces d’industrie en matières dures animales) indique sans conteste des occupations à caractère domestique. À ce titre, le site de la Haute-Île constitue un nouveau gisement de référence, car c’est le seul qui soit multiphasé en Île-de-France pour le second Mésolithique, les témoins de cette période étant de surcroît encore rares à plus large échelle dans le Bassin parisien. Les datations radiocarbone laissent penser que la nécropole pourrait être plus ancienne que les occupations domestiques, ce qui conduit à formuler l’hypothèse que l’occupation humaine mésolithique de la Haute-Île s’est déroulée en plusieurs moments avec, tout d’abord, une occupation à fonction principalement funéraire durant le 7e millénaire avant notre ère, suivie de plusieurs installations successives à vocation domestique durant le 6e millénaire. À ce stade de la fouille, les sépultures ne paraissent pas être en lien avec un habitat. Si elle se confirme, une telle implantation des structures funéraires à l’écart des lieux d’habitation correspond à une situation courante dans le Mésolithique régional, ce qui renvoie peut-être à une conception particulière de l’espace à cette période. Cette succession de fonctions à la Haute-Île pourrait indiquer que la nature funéraire du lieu a pu être rapidement oubliée ou négligée, peut-être à l’occasion des bouleversements de diverses natures qui ont marqué le passage du premier au second Mésolithique, mais il est également possible que les populations du second Mésolithique se soient volontairement installées sur un lieu ayant gardé une forte charge symbolique. Il faut noter par ailleurs que la paléoberge-est fut un lieu particulièrement attractif durant tout le Mésolithique, en étant à l’abri des crues, avec une bonne visibilité et un accès facile à la rivière.