2Disregarding palimpsests enables us to observe periods of typological stability over four centuries, separated by breaks. Some of these breaks could correspond to a simple transformation of the technical system due to intrinsic factors, outside influences or responses to climatic and environmental fluctuations (Robinson et al., 2013). Others result from deep societal changes, or even immigration. It is still impossible to answer these questions due to a lack of data for each phase. The aim of this brief note is nonetheless to present an inventory of the differences observed between these different periods of the Mesolithic.

3The study of lithic productions enables us to distinguish between blade manufacturing by direct percussion with a soft stone from the manufacturing of regularized pieces by indirect percussion in the most recent assemblages with trapezes and blades with Montbani retouch (Rozoy, 1968, Coincy and then Montbani styles). During the full debitage phase, rather large pieces were detached, along with some smaller ones that might be qualified as bladelets according to morphometric criteria. There is no evidence yet for two distinct production sequences (chaînes opératoires) for the blades and possible bladelets. For the moment, in Picardie, technological analyses are not more precise than weapon element typology for chronological attributions (Paris et al., 2012), except for the Initial Mesolithic (work in progress).

4Much less frequent artifacts, such as “axes” are always present in the “Early Maglemosian”, as well as at La Chaussée-Tirancourt at the end of the local Beuronian and in the RMS A. They are absent in the Beuronian with segments, in which we find a few grooved sandstone pieces, and probably some prismatic tools of the Montmorencien type (Ducrocq et al., 2014).

5The bone and antler tools include points and bevels, though perforated antler sheaths are found only in the later phases, represented by an object in the upper level at La Chaussée-Tirancourt (contact peat / silt, between 7 500 et 6 900 BP, Ducrocq, 2001: 193), and isolated pieces with dotted decorations (Fagnart, 1991).

6Other than a few red deer teeth, the personal ornaments consist mostly of perforated Tertiary fossil shells (Ampullina in the Beuronian with segments at Warluis II and especially Bayana in the Late Beuronian at La Chaussée-Tirancourt). These objects are similar to those attributed to the Ahrensbourgian at Remouchamps (Lejeune, 1984), but they are still absent from the Initial Mesolithic and the “Early Maglemosian”. During the RMS A, a change appears to have occurred with the use of perforated Holocene cardiums (Pit 2 at La Chaussée-Tirancourt).

7Charred hazelnut shells are present in highly variable proportions in most of the sites. Their absence from the oldest sites can be linked to their chronological position before the rise of Corylus within the deciduous woodland.

8The animal species most consumed are wild boar, red deer, roe deer and aurochs (Bridault, 1997). We observe a specificity with the Beuronians with segments whose predation was focused on wild boar (Ducrocq, 2013).

9There is no clear evidence for fishing before the end of the Beuronian and during the RMS A, as shown by pike remains at La Chaussée-Tirancourt (Ducrocq, 2001).

10The Mesolithic sites of all periods are located in the same places, either on the alluvial plains, on the edges of plateaus, or on sandy mounds (Ducrocq, 2001).

11The filling of valley bottoms by peat gradually reduced the dry surfaces near streams and rivers, which were the preferred living spaces. The number of sites in the valleys thus decreased depending on their age. There was generally a much greater density of the most recent sites due to multiple diachronic returns to the same, much more restricted, locations. The high density and rather diverse toolkits of RMS A and Mesolithic with trapeze sites could therefore be the result of taphonomic factors. On the other hand, the abundance of endscrapers and burins in the Initial Mesolithic and “Early Maglemosian” contrasts greatly with the Beuronian with segments, in which the tools consist mostly of unretouched blanks. These latter sites are the most numerous (slightly less than 50% of the sites), but their abundance in the peaty valleys, such as Le Thérain (Coutard et al., 2010) or La Selle (Fagnart et al., 2008), could be due to sedimentation processes that favored the development and preservation of shallow levels. However, the greater representation of this phase is not limited to the peaty zones. It also exists in the Oise Valley (Paris et al., 2012), and on the sandy mounds and plateau edges (Ducrocq, 2001). This is probably due to an increased mobility and a resulting multiplication of campsites. In Flandre, the abundance of sites contemporary with this period is interpreted as the result of a high mobility that contrasts with that of the later Mesolithic populations (Crombé et al., 2011).

12Finally, the zones of distribution of the “Early Maglemosian”, the Beuronian and the RMS A are very different (Ducrocq, submitted). This could be linked to major the paleo-geographic modifications associated with the transgression (Sturt et al., 2013), which could explain the movements of populations, the modifications of resources and / or the establishment of new exchange networks.

13This brief inventory shows that the phasing that is gradually becoming apparent is based not only on typological data, but on a set of factors indicating a complex evolution of the Mesolithic in northern France.