Between around 29.000 and 21.000 BP, Europe was occupied by human societies whose archaeological traces are gathered within the Gravettian entity. However, several authors are now openly questioning the unity of the Gravettian, and one of the key aspects of these debates is the supposed formation period of the Gravettian, that is to say the period covering the disappearance and replacement of the traditions related to the Aurignacian around 29.000-28.000 BP. Dated to around 28.000 BP (De Heinzelin, 1973a, p. 45 ; Haesaerts and Damblon, 2004 ; Jacobi et al., 2010), the main occupation of Maisières-Canal (Belgium) is clearly relevant to the period. For a long time, the site’s lithic industry has been recognized for its peculiarities, most notably numerous points with flat retouch and tanged tools. The industry was therefore the key factor in the definition of the Maisierian, which is sometimes seen as a “ tradition” (Campbell, 1980) or a “culture” (Dewez, 1989) that has no link with the Gravettian itself. On the other hand, some have postulated that the Maisierian simply represents a northern variation of Early Gravettian industries, possibly as the origin of the industries with Font-Robert points from South-West France (former Perigordian Va ; De Heinzelin, 1973a, p. 54 ; Otte, 1976 and 1979b, p. 632 ; Desbrosse and Kozłowski, 1988, p. 47). The difficulties in the classification of Maisières-Canal illustrate its importance for understanding the period covering the development of the Gravettian. However, although previous research on Maisières-Canal’s lithic industry has focused on typological, technological and functional aspects, as well as raw materials provisioning, little has been done to characterize the lithic technical system and the chaînes opératoires followed to produce lithic tools. Maisières-Canal was discovered in 1966 during enlargement work of the “ canal du Centre” near the village of Maisières and the city of Mons, and was excavated between 1966 and 1967. Two distinct areas were identified. The most important in terms of the quantity of remains is known as the “ Champ de fouille”, while the other is much more limited and was interpreted as a knapping area. This article presents study of the lithic material coming from the Champ de fouille, with specific focus on both the production of points and the laminar reduction sequence. The industry includes 945 retouched tools, mostly made on large blades. Retouched tools on bladelets are extremely rare. The retouched assemblage is dominated by several types of burins and points. The point category comprises several morphological groups. Among these, Maisières points and tanged points are the most important, but a limited number of shouldered points are also present, as well as several distal fragments that cannot be associated with any of these groups. Despite their morphological differences, these point type display a similar shaping method, using direct flat retouch prepared with an organic hammer. A specific succession of actions is identified that ultimately allows detachment of removals overshooting ridges of the blank’s dorsal face. These so-called “ ridge-overshot” removals offer the possibility to reduce the blank’s thickness. This characteristic of the shaping method can potentially lead to a complete modification of both the morphological and volumetric properties of the blank. However, the extent to which the method was applied was determined by characteristics of the chosen blank and their proximity to the functional parameters of the desired tool. Consequently, only some points display a completely invasive flat retouch, while most are retouched in a more limited way. This observation illustrates the flexibility of the method which allows various degrees of investment for the knapper, ranging from a very limited correction of the blank’s initial properties to its complete shaping. Blade production is made on several flint types all coming from the Mons region (Miller, 2001 ; Moreau et al., 2013 ; Moreau et al., 2016). The maximum distance between the site and exploited flint sources is therefore around 10 km. Blades are generally produced on regular blocks that have both narrow and wide surfaces. Initialization of the reduction sequence is made on one of the narrow sides after a limited configuration of the volume that usually includes preparation of a two-or one-sided crest. The reduction sequence then becomes dissymmetrical (Valentin et al., 2014) as it progresses gradually toward one of the adjacent wide sides of the core. The bipolar exploitation of this wide débitage surface leads to production of straight, large and moderately thick blades. Maintenance of the convexities needed is made with the detachment of blades at the limit between the débitage surface and its flanks, as well as by the alternate use of both striking platforms. Percussion techniques include soft and hard stone, which clearly demonstrate a change of hammers between the laminar production and the transformation of blades into points. Outside the few, and often isolated discoveries of Maisierian elements across North-Western Europe, these two components of the lithic technical system appear unique with regards to the Early Gravettian. No similarity is evident with industries uncovered at other sub-contemporaneous sites in France (Digan, 2008 ; Pesesse, 2008a ; Floss and Taller, 2011), Belgium (Touzé, Flas et al., 2016) and Central Europe (Moreau, 2010 and 2012b). The characterization of blade and point productions at Maisières-Canal thus questions the history of lithic technical systems between the end of the Aurignacian and the (presumably) full development of the Gravettian. It seems that this history is complex and followed different trajectories across Europe.

International audienceLocated at the interface between the Paris and the Rhine Basins and the more northern territories of Europe, Belgium contains several Gravettian occupations, both open-air and in caves. The available documentation is unfortunately limited by the earliness of the excavations conducted at most sites. Stratified records in karstic context, primarily excavated in the 19th century, are particularly affected by this situation. The analysis of Gravettian lithic technological behaviours can, however, rely on two open air sites excavated in the late 20th century which provide more rigorous data: Maisieres-Canal and Station de l'Hermitage. These two sites have been the focus of comparisons that highlight their similarities, both in terms of the lithic industry and location. In fact, the former stands out for the presence of tanged tools, a typological marker which is usually associated with the Early Gravettian of Western Europe. In this article, we present a new study that evidences the differences in the lithic technical systems represented at these sites. After a presentation of the most recent data, we consider the causes that may be responsible for such diversity, highlighting the possible role of the chronological factor, but also of the existence of an original technical tradition in north-western Europe during the appearance and development of the Gravettian. (C) 2015 Elsevier Ltd and INQUA. All rights reserved

Le Gravettien en Belgique est documenté dans un nombre limité de gisements. Il s’agit en général de gisements de grottes ou d’abris sous roche situés dans le bassin mosan, qui ont souvent fait l’objet de fouilles dès le xixe siècle, notamment les plus importants d’entre eux. La connaissance que nous avons du Gravettien belge doit énormément aux recherches de M. Otte. Dans les années 1970, cet auteur a compilé, synthétisé et structuré une documentation éparse et bien souvent imprécise et, sur cette base, formulé une proposition de séquence régionale. Les connaissances accumulées au cours des dernières décennies permettent d’aborder certains aspects de ce modèle sous un angle différent, tandis que de nouvelles questions sont également apparues. Cet article a ainsi pour objectif de dresser un bilan critique du phasage de la séquence gravettienne belge, à partir des données chronologiques et lithiques de quatre sites majeurs : Maisières-Canal, les grottes et l’Abri supérieur de Goyet, et la Station de l’Hermitage. Au terme de ce bilan, trois entités sont identifiées. Le Maisièrien et le Gravettien ancien correspondent à des traditions lithiques, dont les aires de répartition s’étendent sur une partie de l’Europe nord-occidentale, et qui sont bien documentées en Belgique grâce à deux sites de référence. Un Gravettien « post-phase ancienne » est aussi détecté, mais uniquement dans des ensembles dont l’intégrité pose question, ou qui sont limités quantitativement. Seule la découverte de nouveaux sites, fouillés avec des moyens modernes, permettra d’en acquérir une connaissance plus précise.