• -10 000
  • -9 000
  • -8 000
  • -7 000
  • -6 000
  • -5 000
  • -4 000
  • -3 000
  • -2 000
  • -1 000
  • 0
  • 1 000
  • 2 000

Docteur Jean-Georges Rozoy

Résumé des abréviations utilisées dans les articles : consulter la liste.


Dr J.-G. Rozoy



The structures of 50 Mesolithic camps which were dug by the authors on 9 open air sites on siliceous grounds (19 among them were utilizable thoroughly) are compared to the structures of 4 Paris Basin Magdalenian sites, made up of several tens of recognizable camps. The Mesolithic camps were single, according to the Crombé's model, the Magdalenian camps were complex, according to the Leroi-Gourhan's model, and much larger (70 to 300 m2, versus 20 to 60 m2 for the Mesolithic ones), they bore heaps of raw knapping waste, the flints were distributed unevenly, they were distributed evenly on the Mesolithic camps. On the Magdalenian sites, the hearths were in the middle of the camps; they were outside and leeward on the Mesolithic camps. The Magdalenian working places were very close, the tool distribution was quite uneven, the Mesolithic tools were distributed at random all through the inner spaces of the camps. The Magdalenians used only a small part of the knapping products ("always less than 10 %"), 30 to 50 % of the Mesolithic knapping products more than 2 cm long have conspicuous using wears, but 80 % are found broken, then we can presume that they were used. The Magdalenian knapping way was very stiff, "made thoroughly without even stirring", the Mesolithic one was looser, often stopped and resumed by different persons in different parts of a camp. The Magdalenian tool ratio was unsteady, according to the logistic exploitation of the land; the Mesolithic one was steadier, because of residential mobility.


Les structures de 47 camps mésolithiques fouillés par les auteurs dans 8 sites de plein-air sur terrains siliceux (dont 16 camps totalement utilisables) sont comparées à celles de 4 sites magdaléniens du Bassin parisien comportant plusieurs dizaines de camps identifiables. Les camps magdaléniens sont beaucoup plus grands (70 à 300 m2 contre 20 à 60 m2 pour le Mésolithique), ils comportent des amas de débitage brut, la dispersion des silex y est très inégale, elle est uniforme dans les camps mésolithiques. Les foyers sont centraux au Magdalénien, extérieurs et sous le vent au Mésolithique. Les lieux d'activité sont très concentrés au Magdalénien, la répartition des outils y est très inégale, au Mésolithique les outils sont distribués aléatoirement sur tout l'espace interne du camp. Les Magdaléniens n'utilisent qu'une faible proportion du débitage ("toujours inférieure à 10 %"). Au Mésolithique 30 à 50 % des produits de plus de 2 cm portent des traces d'usage visibles, mais 80 % sont cassés et ont donc été employés. Le débitage magdalénien est très rigide, fait "de bout en bout sans bouger", celui du Mésolithique est plus souple, souvent arrêté et repris par diverses personnes en divers endroits du camp. La proportion d'outils est au Magdalénien variable, en fonction de l'exploitation logistique du terrain, elle est au Mésolithique plus constante du fait de la mobilité résidentielle.


Besides Le Tillet (Rozoy 1996, 2000 c), which is in the Somme Culture (fig. 1), our team would dig and analyse topographically several Mesolithic open air sites : Roc-la-Tour-II, Montbani-II and Sablonnière-II (Coincy) (Rozoy 1978) and Tigny (Rozoy 1998 c) in the Tardenoisian early/middle stage; l'Allée Tortue (Parent 1967, Rozoy 1978, Rozoy and Slachmuylder 1990, Rozoy 2000 a) in Fère-en-Tardenois in the Tardenoisian late stage; Marlemont (the dig was impeded by an American military camp, Rozoy 1978) and Roche-à-Fépin (Rozoy 2000 d) in the Ardennian middle stage. On these 8 sites, we could identify 47 distinct camps, 16 of which are utilizable entirely (9 in Le Tillet, 2 in Tigny, 2 in La Roche-à-Fépin and those of Roc-la-Tour II, Montbani-II and Sablonnière-II, with two occupation periods at least for each of these last two). We can also refer to different colleagues' works (to a certain extent depending on the details given by the authors), more particularly J. Hinout (1991) who published total plans with most of the tool types in the Tardenoisian, Les Mazures (Pirnay & Straet 1978, Rozoy 1982) and in Hergenrath-Flönnes-2 (Hubert 1967, Rozoy 1978) in the Ardennian; and in the Ardenne Tardenoisian l’Ourlaine (Lausberg-Miny, Lausberg & Pirnay 1982, Pirnay 1982, Rozoy 1982). Some other contemporaneous sites in other more or less similar ones can also be used partially for topography. That corpus is fairly rich and allows to draw tendencies common to those various Mesolithic open-air sites in Northern France and to try and compare them to the corresponding features which were established admiringly by A. Leroi-Gourhan and his scholars in four well-known Paris Basin Magdalenian open-air sites : Pincevent, Etiolles, Verberie and Marsangy. Our team dug also the Roc-la-Tour-I Magdalenian site (Rozoy 1988) but it is a palimpsest and it does not seem possible to distinguish successive camps there, because there were not enough deposits of sediment; moreover a part of the ground had been washed off by solifluxion. Of course the Mesolithic camps on sand cannot provide so precise nor so detailed data as do the Seine, Yonne and Oise lower terraces. Besides the lack of preserved bones (which allow to obtain quite remarkable plans on the Magdalenian sites concerned, except in Etiolles), two extra features are unfavourable in the Mesolithic : the very light soil leads to the scattering of the lithic material (probably partly as soon as the beginning, increased with later taphonomy), the only one that have been preserved, in a layer 20 to 30 cm thick (Rozoy 2000 c, "dig analysis methods"). The tiny flints, usually all of them from the same local material, make it illusive any hope to reset a sufficient number of chipping-proceedings, and then do not allow to restore the successive operations on a same nucleus nor even to identify correctly the chipping places (Pirnay 1982, Walczak 1997). Consequently the topographical analysis is reduced to far more simple features. Though, it is possible to compare them with those known from Magdalenian.

The Magadalenian camps are much larger

All the camps that we could identify in the Mesolithic early and middle stages would spread on rather small a space, between 20 and 40 m2, the largest ones being Tillet-2-NW (we know that it was occupied at least twice, Rozoy 1999 b and fig.2) and Fépin-1, with a poor 20 m2 annex to the latter. Montbani-II was more than 60 m2 but it may have been occupied twice on close areas. Sonchamp-III (dug by Hinout, in Bailloud 1967, p. 304) is as much as 70 m2. For the final stage, two camps among the Allée Tortue X ones were 60 m2 each (those with Vielle trapezes without flat inverse retouch) (Rozoy 1999 a and fig. 3). We have been able to demonstrate that none of these camps was strictly contemporaneous with another on the same site, and that they were occupied by different persons with nearly similar working though separate habits. So it was with the units determined in Bergumermeer (final Mesolithic) by R.R. Newell (1978) varying from 40 to 60 m2.

In the Magdalenian, M. Olive (1997, p. 88) mentions 5 units in Etiolles between 60-70 and 180 m2, the unit plans established by M. Jullien and coll. (1988) for Etiolles and Pincevent vary between 50 and 120 m2 and we can find 130 m2 for "the richest area in the U 5 occupation unit (NE area)" in Etiolles (Pigeot 1987 and fig. 4). The plans established by Y. Taborin, M. Olive and N. Pigeot (1979) for the W 11 and U 5 units in Etiolles cover 180 and 225 m2, and none of both concerns the whole corresponding camp, P. Coudret and coll. (1994) explaining (p. 135) that the P 15 unit was an annex to U 5. So it was for the Pincevent Dwelling place I (Leroi-Gourhan and Brézillon 1966) where the sheltered  area was 36 m2 large, the camp itself (destroyed deeply by gravel quarrying) was much larger. M. Olive (1992, p.88) refers to 6 small units in Etiolles varying between 10-12 and 35 m2, but those were secondary structures, the annexes to broader elements forming clearly larger complex camps with them. For example in Verberie F. Audouze (1987) points out butchering places as complements to the dwelling areas and M. Jullien, Cl. Karlin and P. Bodu (1987), for the IV-2 level in Pincevent, points out ten contemporaneous dwelling units forming one camp, among which 1 600 m2 are known, an unestablished part having been destroyed by dredging before the sand-quarry stopped being worked. In Marsangy (Schmider 1992) the n° II area lay through 155 m2. A first opposition takes place then : the smallest units in Magdalenian camps (which did not necessarily include the whole camp) had quite the same surfaces as the largest entire Mesolithic camps we know. It is a first indication about an amount of people much larger in a Magdalenian camp than in a Mesolithic camp. On the other hand, there were much more of the latter, and they were scattered more widely on the land, which was occupied entirely, giving a total population four times more important than in the Magdalenian (Rozoy 1992).

The presence or absence of chipping heaps

The four analysed Magdalenian open-air sites in the Paris Basin (and also Tarterets II, Brézillon 1971) contained huge heaps of chipped flint, made up of several thousands of pieces, all of them being stored one above the other, with practically no sediment having percolated through them, liable to be more than 60 cm high and one or several square metres broad : 4 000 pieces in only one heap in Verberie (Audouze 1994), 30 000 flints in W1 unit in Etiolles (Audouze 1988). Usually there are several heaps in one unit, they are separated more or less clearly and between these heaps and the structures (hearths, heaps of stones, ... ) there are empty areas, with sometimes only two or three flints in one square metre. Those empty areas were not only the sleeping areas protected by the animal furs used for the beds, they exist also outside the shelters, even in the central areas of the camps. That corresponds to the existence of well defined chipping pauses; in all the pauses they would handle (usually "from end to end, without even stirring", Olive 1992, p. 115) several flint-nodules, the products (still raw) of almost the whole of which have been stacked in a heap, only a part of them (blades and /or bladelets) have been chosen to make tools. That corresponds to "a very clear taste for focussing their activities in a reduced space" (Y. Taborin 1987). Was that "taste" referring to the climate, considering that fine activities were easier inside, with less stiff fingers? If that was so, we could find summer activities more scattered around, but we don't. Sometimes (Olive 1997) the peeling off was made a little further, in the outer part of the camp, and the chipping itself was achieved rather in the middle, and always produced heaps. Cases of full chipping begun somewhere and achieved elsewhere did exist but they were scarce. We can also perceive bare areas in sites where bone was preserved, as in Pincevent (Leroi-Gourhan and Brezillon 1966, 1972). We could perceive areas where flint waste was brushed off or ashes pushed off hearths.

It was exactly the opposite in the Mesolithic : the distribution of flint chipping is much less clear, there are no bare areas in the concentration inner spaces and there the poorly covered parts are scarce (fig. 2), nearly always in Le Tillet there are more than 150 flints on a quarter of a square metre, in Fépin or Tigny there are more than 25 or 30, i.e. for each site more than the usual quantity found in the outer spaces, in which we could perceive very scarcely and doubtfully areas where pieces were thrown out; the only pieces clearly thrown out were cores, but we cannot state that they were thrown out together. In some Mesolithic sites we could hardly perceive a few pauses for elementary peeling off, but the largest crust-flakes were split off everywhere and so it was for the processing of full chipping, when most part of the crust was being on peeled off. That corresponds to a kind of chipping which was often interrupted then resumed (possibly by other people and on another place). That was demonstrated with "mental reconstructing" by J. Walczak (1997); it is also determined partly by the much smaller sizes of the products, which were not bulky and could easily become parts of the sandy soil.

The place of the hearth in a camp

In the Mesolithic sites on sand, it is not always easy to determine where the hearth(s) would be placed, because of the permanent washing off, draining the coal particles downwards deeply. That could be photographed by René Parent in Sablonnière-II when he directed our dig there (Parent 1973, Rozoy 1978, pl. 129). Moreover, the sites have remained uncovered for various spans of time and they are liable to have received further deposits. In Le Tillet we could gather coally sand from 9 hearths which had been set up against sandstone blocks and also from the supposed remains of the cleaning of 9 other hearths, but we had to get rid of those which gave coal incompatible with Boreal (fig.5). In the Allée Tortue, we know a hearth from the early stage, out of the middle, fit up with stones. In a few words, most hearths (all the well established ones) were on the borders of the concentrations of flints, and usually in such places as to avoid to fill the working areas with smoke (Rozoy 2000 c). That is quite the opposite of the central position which is the rule (not a fast rule, though) in the Paris Basin Magdalenian camps (fig.4). Similarly the hearths would structure camps, but in quite different ways. The fact that a hearth structures a setting up is so elementary, so common an ethnographic fact, that it demonstrates nothing, what is important is the structuring way, which in the Mesolithic was quite different from that in the Magdalenian. Of course the climate (cold or not) could have an influence. And though Magdalenian camps are supposed to have been summer camps, we know some units there with no marks of shelters, which have also central hearths. Moreover those hearths were often set up with stones. At last another main difference in the Mesolithic is in the excluding ratio between hearths and chipping areas (fig. 2) or tool areas (fig. 7). Should either the domestic tools or the armatures be concerned, in all cases they would keep apart from the verge of the hearth. The same exclusion can be found in Tigny (Rozoy 1998 b). That is quite on the opposite of the Magdalenian custom.

The focussing or scattering of the activity areas

In the Magdalenian camps (thanks to flint-core rebuilding), we could perceive pauses for peeling (and preparation of cores) and pauses for full-chipping. There were often several of these pauses next to one another (Olive 1997), with a functional differentiation of the space on both sides of the hearth (Olive 1997, p.92-93). Apart from those pauses (and the spaces used to keep them clean), the ground was almost bare. The tools in Etiolles P15 Unit and even more in K12 were concentrated in one of those chipping areas, the core-preparation areas in K12 have none. Usually, in the Magdalenian, the tools (including retouched flakes and blades) were gathered very tightly on narrow areas, usually near the hearths. For example, most part of the inner space in P15 had none (Olive 1997, p. 92, and fig.5) : 70 tools on less than 5 square metres, and 24 on the 23 extra square metres. In Pincevent, Dwelling 1 included 332 retouched tools and 62 "cutting blades" with using wears, 394 genuine tools on the whole, for 31kg of flint on 30 square metres, i.e. numbers hardly lower than those in Tillet-2-SE, -3, -4 or -5, but their weights are four times higher. 48 of the 130 burins were gathered tightly around the two seating-blocks next to the hearths, each group on one square metre, 3 groups of 8 were near the 3 hearths (opposite the seating-blocks), the others were scattered along the sides of the tents or pushed off outside. Big borers and truncated blades were gathered even more closely next to the seating-blocks (19 out of 41), backed bladelets were also near the hearths (62 out of 64, on 6 of the covered 30 m2, amongst which one metre with 21 pieces). So it was on Pincevent section 36 (Leroi-Gourhan & Brézillon 1972 and fig.6 : on the plan we counted 569 pieces on 123 m2, 62 squares of which were without even one backed bladelet and some others with 60 or 80. The shortened table that we had at disposal did not allow us to calculate the square-Chi but it might be amazing!). In the Dwelling 1, there was the same quite uneven distribution for all the tools, only the 35 scrapers were scattered everywhere (Leroi-Gourhan and Brézillon 1966, p. 348. But one square with 5 scrapers biases the distribution, the other 30 are in accordance with the "loi de Poisson"). But in Pincevent section 36 in T 112 and V 105, the 118 scrapers on 121 m2 were scattered unequally with 70 squares with no scraper at all, the square-Chi is 136 (the limit for 1/1 000 is 22), the bias does not depend only on the two squares with 6 pieces, but on the whole distribution with too many empty squares and too few with one piece only. The case of the scrapers in the Dwelling 1 is then an exception, even for the scrapers. In Verberie (Audouze and coll. 1981, p. 111), the distribution of cores and borers is also quite uneven, more particularly with a group of 5 cores in C 17, far from the middle. 5 of the 9 borers were found together. In Marsangy, though most of the tools were scattered more evenly such a concentration was remarked for the axial big borers : 47 ones on 53 m2, there were 33 squares with no big borer at all, 3 with 6 ones in each, the square-Chi is 617, whereas the probability limit at 1/1 000 is at 20 (Schmider 1992, p. 169, Schmider 1994).

Such concentrations are quite the opposite of the situation in the Mesolithic, when we could prove (Rozoy 2000 c) that the dispersion was almost quite at random on the inner spaces of the concentrations, whereas the (very poor) anomalies regarded only a low minority of the pieces, as the case of the scrapers in the Pincevent Dwelling 1. In the Mesolithic that exception became an absolute rule : 3/4 of the distributions were perfectly at random, likely for more than 5% (more often 50%). The others belonged to one digging-square only (1/4 of one metre) or two, with a few more tools of the type concerned, the others being scattered according to the "loi de Poisson", the square-Chis were very little higher than the 1/1 000 limit. The Magdalenian working areas were specialized and accurate, some of them being liable to have been left on the outer spaces, but we are not supposed to know expressly that those outer spaces did exist. In the Mesolithic their scattering on the whole inner spaces made those last ones homogeneous for all the categories of tools and pieces. In Tillet-2 SE and in Tillet-3 the global localization took place in the southern parts of those inner spaces (fig.7), and it was common to all the types of the tools belonging to the common kit, without any possibility to identify precisely the places where they were made nor how these or those types were utilized, whereas the elements of the chipping and the sharp armatures had the same lots as the chipping waste and they were left all over the inner spaces. That means that for both those units the spaces where the domestic tools were utilized were smaller than those used for chipping, though they were common to all those tools. In Tillet-2-NW the tool area was larger than the chipping one (fig. 7), and it seems diffused, too. Those observations show that freedom in working was far greater : when chipping or making and using tools, the craftsmen would settle down either here or there, without they feel themselves bound to any working place they might have occupied previously. So both they changed pauses more often, making or using only short series, as they were wanted, and there were more people (proportionally to the low number in a band) who could do any operation. This is a contrast with the few experienced specialised Magdalenian craftsmen, who found it difficult to train apprentices whose awkwardness we can perceive, more especially on peripheric areas.

Utilized flakes and blades

The manufactured tool-ratio presented here above might be criticized. Indeed it includes (for our research on Mesolithic) many retouched flakes, blades and bladelets, often, those were altered very slightly only by that "retouch" - those pieces, which tradition inherited from XIX th century, still in use partly, does not allow to take into consideration, especially when you do the sorting out without glasses or with too poor light. And we do know (Rozoy 1978, p. 28-29, from F. Bordes' experiments and advice) that it is often impossible to distinguish without failing intentional retouch from the result of some work. M. Brézillon (Leroi-Gourhan and Brézillon 1966, p.281-283) showed soundly that "oblique spontaneous little retouch" (which is quite regular) might be caused by oblique transversal actions. The pieces that wear no other alteration on their sides were not tools made on purpose but a posteriori tools. Though they were nothing less than tools and we must take account of them, even (if you are anxious to do so) if we had to make a separate category, but then we should be stopped by a problem with limits, which would be unsolvable. The criterion that we have taken in account is a regular uninterrupted line of "retouches", a short though regular uninterrupted series, knowing that we consider as "retouch" (understand "intentional") any discontinuous (or) irregular removal considered as the result of some work. Those criteria keep (among others) the Brézillon "little spontaneous oblique retouch" among the tools manufactured intentionally but we have found no better criterion. The analyses concerning Le Tillet (Rozoy 2000 c, "Spatial repartition of the common tools", paragraph "using wears") have proved that those criteria have good (relative) efficiency.

Keeping in mind carefully that relative feature we have then to wonder about the comparisons between the various Mesolithic sites and also between those and the Magdalenian sites. We could perceive large differences for the "utilized" pieces between the Le Tillet Mesolithic sites : from 25 to 50% of the manufactured tools (up to 56% of the common kit tools) (Rozoy 2000 c, table 21). That depends on the concerned hunters' technical disposition, even inside one well defined regional group (dialectal tribe, Rozoy 1998 a), consequently with greater reason from tribe to tribe. Besides, the huge ratio of broken blades and bladelets (Rozoy 2000 c, tables 23 and 23 ter, more than 80%) suggests an even more general use. Those variations make it difficult for us to compare with the Magdalenian for which we have only very little knowledge about it. Recognizing as tools all the retouched blades and with greater reason (slightly) retouched flakes is not usually accepted and still less utilized blades (and worst of all, utilized flakes). Nevertheless as soon as 1966 A. Leroi-Gourhan (Leroi-Gourhan and Brézillon 1966, p.304-311) who was far ahead of his scholars, as usually, noticed 62 "cutting blades" in Pincevent Dwelling 1 and described their identification criteria (irregular discontinuous stuff removals). When they are referred to the manufactured 332 tools and to the total amount of actual 394 tools we get a ratio of 8% which corresponds quite well to our lowest notes for Mesolithic ... if we are satisfied with the first sorting out done in poor conditions. But it is far lower than the results of our sorting out with right light and efficient glasses, those are around 25 to 30% at least and sometimes more than 50%. For some time we have found notes and even plans of some retouched flakes in various papers about Magdalenian in the Paris Basin (Olive 1992, p.102, Olive 1997, p. 92, 93), but without any note about the way they were counted, which would have made comparisons easier if the counting had been systematic, no note either on the conditions of the sorting out. F. Audouze (1988, p. 79) published the histograms of the three dimensions of the «neglected blades and the utilized blades in a heap of waste in Verberie". But we could find no number either for this unspecified heap or for the whole site, because most Pr Leroi-Gourhan's scholars did not seem interested in counting. From the graphs it seems that there are nearly 100 neglected blades and 15 or 20 utilized blades, which gives a ratio around 13 to 15 %, i.e. a bit higher than that for Pincevent Dwelling 1 "cutting blades", but lower (half) than those for our Mesolithic sites (far less for Tillet-2-NW which is higher than 50%). Though, M. Jullien (& coll., 1988, p.90) wrote : "The ratios of useful products (manufactured or utilized only) are always lower than 10%". When you take off the manufactured products the ratio of utilized products seems low indeed, only the most obvious cases have been mentioned. Except if a wishful precise inventory happened, now it seems that using chipping-products in that area in the Magdalenian was far less common than in the Mesolithic, especially if we pay attention to what has been shown by breaks on blades and bladelets in the latter. That difference depends certainly on the chipping techniques, the Magadalenians used to do the chipping "from beginning to end without even stirring" as said here above ; the Mesolithics would stop as soon as they had obtained what they wanted and resume later and somewhere else whenever they wanted a new tool.

Broken blades and bladelets

In Le Tillet, about 80% of the raw or the "utilized" blades and bladelets were broken (calculated from the pieces more than 2cm, now for Tillet-2-NW and SE, Rozoy 2000 c, tables 23 to 23 quater). This means a use of pieces which did not leave use-wears liable to be seen without optical instruments (unfortunately traceology cannot be applied because the coating soil caused too many scratches on the pieces). Taking account of the retouched or utilized blades and bladelets and of the armatures (including those lost outside the camp), more than 90% of the manufactured blades and bladelets were utilized some way or other. And we know from experience and various accounts that it was a general rule, as much in the Middle Mesolithic (Walczak 1997) as in the late and the final Mesolithic (Rozoy 1978, 2000 a). For the Magdalenian the only mention that we have found is also in Leroi-Gourhan and Brézillon (1966, p. 349-350) : 11 out of the 62 utilized "cutting blades" were broken. Consequently a ratio of some 78% of untouched ones opposed to our 80% broken ones. Of course it is impossible to draw any conclusion from that single example, the less as, after all, it was noted quite incidentally about the tool moves  (centrifugal there). Though, B. Schmider (1992, p.92) has given an exhaustive amount for Marsangy but unfortunately it associates blades, bladelets and flakes : out of 20 597 matter-removals, 8 527 are fragments, i.e. 41,4%. That ratio is higher than the one obtained with the utilized blades in Pincevent, but far lower than the one that we obtained, though Marsangy was evidently later than the other three sites, and already tending a little towards mesolithisation. If you consider that breaking makes two pieces instead of one, the ratio of 40% is reduced to 20 out of 80, i.e. 25%, but as for Le Tillet quoted here above, lots of small fragments should be left with the non counted "splinters" and truth might be between the two. We have then here an evocative index, which would be worth a larger study. If that disposition happened to be confirmed for the other Magdalenian sites of the Paris Basin it would mean that the Mesolithics used the chipping products clearly more largely, since they used almost all the blades and bladelets (and an important part of the flakes), which does not seem to be the case in the Magdalenian. The reason is evidently the same as the one said in the previous paragraph.

The ratio tools/chipping

In the Mesolithic the ratio of tools versus the total amount of the found flints is quite steady from one site to another, around 2,5% with light variants : the variants are, on the one hand, in Tillet-6 (1,5%) which is the smallest camp in the site (less than 300 tools), the ratio is then less trustworthy, on the other hand Tillet-8 (4,6%). In the other sites, which were studied by our team, we have always found values between these limits. That ratio depends mainly on the sizes of the meshes of the sieves used on the dig (4 mm for us; in the Sauveterrian with hyper-pygmy armatures we should use 2 mm meshed sieves and the amount of tiny flakes would increase quite amazingly). In the Magadalenian there are huge variations in those ratios : F.Audouze and coll. (1988,p.63) wrote "the tools are less than 5% of the chipping and often 2 to 3%", which corresponds exactly to our results for the Mesolithic, but she specified on the following line : "in Etiolles Dwelling unit W 11 there were 21 tools out of 30 000 flakes (Taborin 1974, p. 17). In P 15 the tool-kits were 0,7% of the chipping and in U 5 they were as much as 2,32%. In Pincevent and Verberie that ratio was generally higher and more than 5%. The following part of this paper confirms with some more examples indicating that the finest blades might have been put apart, saved and taken away outside the dwelling unit or even outside the site itself. This was confirmed by the presence on all the sites of tools and also of raw blades made of foreign flint belonging to the kit they brought with them when arriving. Though, that kit was not generally important enough to change ratios so strongly (and more especially not to reduce them !). It seems that this divergency between the two periods in the ratios of tools was bound to the way of making use of space : the Magdalenians would apply the usual logistical exploitation with sites more or less specialized in such production or the other. So, on any site devoted to the production of flint (at least partly), Etiolles for example, to provide any other camp which might have less, Pincevent for example, it seems quite usual that some units (W 11, P 15) bear almost no tools, whereas others (U 5), where there had been normal life with all the usual activities had a ratio of tools corresponding to that way of life. Then we can understand that Pincevent or Verberie (in the latter they used to go and bring back flint from less far away, from their own area) are higher than 5% of the manufactured tools. But on the one hand we should consider Etiolles as a site partly specialized, not specialized in flint, for quite an important part of the stays they used all the activities. It was just the opposite for Pincevent or Verberie, because if they used imported flint only (from Etiolles or elsewhere) we should have found no, or very few chipping products, and not 95%. Actually specialization there seems to have been exceptional. If we come back to the Mesolithic with its rather steady ratio of tools around 2,5%, it is obvious that the way of exploiting the site was different : it was residential mobility, they used to remove the camp instead of sending a group for flint or anything else. That was the more suitable than the bands (local groups) were smaller, 10 to 15 persons including children, hunting with bows and arrows, instead of some 50 at least for the Magdalenians beating up the game and hunting with spears and spear-throwers. It was easier to remove 15 persons than 50 or 80.

A quite pragmatic way of chipping

Those easy changes of pauses, that cleverness for all to chip without any preparation, that power of adapting oneself with resumption of the chipping by various craftsmen, depend on a plain way of chipping, but which needed to understand the qualities of the material used. Instead of a complex mechanism established empirically and which had to be applied strictly, without derogating for lack of understanding exactly the reactions of the material concerned, they came to a system based on that understanding (Pirnay 1982). In that method (Jérôme Walczak 1997, p.90) "making suitable again convex or angular edges can be learned very quickly : a beginner needs ten days to make Coincy style (personal training) and to work fairly well, without too many accidents, a work-unit on a core. It is so because these features are proper to the physical and mechanical properties of flint, they are not bound to special purposes which would need important pre-determination, as is the case inside a context of highly pre-established production of flakes (Levallois) or blades (“laminaire”). Indeed in those last two contexts it is necessary to guarantee as permanent ones a certain amount of criteria which are liable to obstruct right chipping-proceeding, if all of them were not kept. (...). As for blade-chipping, J. Pélegrin (1986) noted (...) about Roc de Combe Early Perigordian chipping : "The core surface(s) which are or will be concerned by chipping should be regular and more or less convex as much on their axes (hulls) as on their transverse sections (bends). Respecting that restraint is a rule for any "controlled" or systematic chipping process. (That explains) why it is so often necessary to shape a core before proper chipping and to keep its surface suitable when chipping on". Those two remarks can never be applied to Middle Mesolithic chipping. For that one there was only a gradual succession of working units corresponding to the exploitation of favourable physical and mechanical criteria (convexity and angles) and systematically similar to one another. All the working units noticed on the cores (...) are completely independent from one another" (Underlined by J.W.). So, according to Jérôme Walczak, "it is an abstract technical system (...), an efficient technical system, but which has enough with a weak supply of predetermination". That system “can give working units liable to be worked on by different persons (...). In the case of Magdalenian blade-chipping, (...) one and the same person had to manage a whole block to the end of its exploitation. (...) Whenever any beginner or any other craftsman resumed chipping, it was possible to see it at once because the new craftsman had to re-adjust his own method according to the material that he had in his hands. (...) The proceeding was changed all at once and would upset all that had been done before." The new method gave less spectacular products, though easier to be obtained, allowing to use second rate flint. It was associated with the late re-invention and generalization, then, of bow and arrows, which needed microlithic armatures, leading to save utilizable stones, and thus allowing to occupy areas without good flint (Bretagne, Lorraine). It is obvious that the creation of that loose easy method was bound to the development of the mental capacities, as much as the re-invention and generalization of bow and arrows, in consequence of human brain progressing (Rozoy 1995 b, 1997 a, b, c,).

The following table sums up the noticed oppositions between Magdalenian and Mesolithic open-air sites

Comparisons between the structures of Magdalenian and Mesolithic open air sites

Magdalenian camps

Mesolithic camps

Large camps 60 to 200 m2 or more

Small camps 20 to 60 m2

Complex camps with several units (groups of 50-60)

Elementary camps (nuclear families)

Chipping heaps, bare between structures

Homogeneous distribution of chipping and tools

Precise concentrated chipping pauses, sometimes in outside spaces

Chipping on the whole central space

Very painful continuous chipping

Quite free chipping, stopped then resumed

Important central hearth, seldom far from centre, structure origin

Small hearth, far from the centre (leeward)

Tools gathered, often near the hearth

Tools scattered, away from the hearth

Low utilization of chipping

High utilization of chipping

Moving tool ratio (0,1 to >5%)

Steady tool ratio (2,5% with 4mm meshed sieve)

Specialized camps, then logistical exploitation of land

Polyvalent camps, then residential mobility


Audouze (F.) - 1987 - Des modèles et des faits : les modèles de A. Leroi-Gourhan et de L. Binford confrontés aux résultats récents. p. 343-352  in : Bulletin de la Société Préhistorique Française 84, 10-12.

Audouze (F.) et al. - 1988 - Taille du silex et finalité du débitage dans le Magdalénien du Bassin parisien. p. 53-84 in : De la Loire à l'Oder. Les civilisations du Paléolithique final dans le Nord-Ouest européen. Actes du colloque de Liège, décembre 1985, éd. Par M. Otte, B.A.R. 444 (i).

Audouze (F.), Cahen (D.), Keeley (L.) et Schmider (B.) - 1981 - Le site magdalénien du Buisson Campin à Verberie (Oise). Gallia-Préhistoire 24-1, p. 99-143.

Audouze (F.) - 1994 - See Taborin (dir.) 1994.

Bailloud (G.) - 1967 - Informations archéologiques. Circonscription de la Région parisienne. Gallia-Préhistoire X, p.301-316.

Brézillon (M.) - 1971 - Les Tarterets II, site paléolithique de plein air à Corbeil-Essonnes (Essonne). Gallia-Préhistoire XIV-1, p. 3 - 40.

Coudret (P.) et al. - Etiolles. p. 132-146  in : Taborin (Y.), dir., Environnements et habitats magdaléniens dans le centre du Bassin parisien., Documents d'Archéologie Française 43, Paris, Maison des Sciences de l'Homme.

Crombé (Ph.) et Meganck (M.) - 1996 - Results of an auger survey research at the early mesolithic site of Verrebroek "Dok" (East-Flanders, Belgium). Notae Praehistoricae 16-1996,  Louvain-La-Neuve, p. 101-115.

Ducrocq (Th.) et Ketterer (I.) - 1994 - Le gisement mésolithique du "Petit-Marais", La Chaussée-Tirancourt, Somme. Bulletin de la Société Préhistorique Française 92/2,  p. 249-259.

Gaucher (G.) - 1996 - Fouilles de Pincevent II.  Mémoires S.P.F., 30 cm, 235 p.

Hinout (J.) - 1991 - Quelques aspects de la Préhistoire du Tardenois. IV. Le gisement mésolithique de Maurégny-en-Haye (Aisne), lieudit Le Grand Marais. Préhistoire et Protohistoire en Champagne-Ardennes 15, p. 45-59.

Hubert (F.) - 1967 - Un gisement mésolithique à Hergenrath. Archaeologica Belgica 99, 70 p.

Jullien (M.) et al. - 1982 - Organisation de l'espace et fonction des habitats magdaléniens du Bassin parisien, p. 85-123, in : De la Loire à l'Oder. Les civilisations du Paléolithique final dans le Nord-Ouest européen. Actes du colloque de Liège, décembre 1985, éd. Par M. Otte, B.A.R. 444 (i).

Jullien (M.), Karlin (Cl.) et Bodu (P.) - 1987 - Pincevent : où en est le modèle théorique aujourd'hui ? p. 335-342 in : : Bulletin de la Société Préhistorique Française 84, 10-12.

Lausberg-Miny (J.), Lausberg (P.) et Pirnay (L.) - 1982 - Le gisement mésolithique de L'Ourlaine, p. 323-329 in : Le Mésolithique entre Rhin et Meuse, actes du colloque de Luxembourg 1981, A. Gob et F. Spier éd.,  Société Préhistorique luxembourgeoise.

Leroi-Gourhan (A.) et Brézillon (M.) - 1966 - L'habitation magdalénienne n° 1 de Pincevent près Montereau (S. et M.). Gallia-Préhistoire  IX, 2, p. 263-385.

Leroi-Gourhan (A.) et Brézillon (M.) - 1972 - Fouilles de Pincevent. Essai d'analyse ethnographique d'un habitat magdalénien, la section 36.  VII° suppl. à Gallia-Préhistoire, C.N.R.S., Paris.

Newell (R.R.) - 1978 - Mesolithic dwelling structures : Fact and fantasy. 2° International Symposium "Mesolithikum in Europa", Potsdam 1978, p. 235-284.

Olive (M.) - 1992 - En marge des unités d'habitation d'Etiolles : les foyers d'activité satellites. Gallia-Préhistoire 34,  p. 85-140.

Olive (M.) - 1997 - Foyer domestique ou foyer annexe. Les modes d'occupation de l'espace des Magdaléniens d'Etiolles. Gallia-Préhistoire 39, p. 85-107.

Olive (M.) - see Taborin, Olive et Pigeot 1979, Coudret in Taborin 1994.

Parent (R.) - 1967 - Le gisement tardenoisien de l'Allée Tortue à Fère-en-Tardenois (Aisne). : Bulletin de la Société Préhistorique Française. p. 187-207.

Parent (R.) - 1973 - Fouille d'un atelier tardenoisien à la Sablonnière de Coincy (Aisne). : Bulletin de la Société Préhistorique Française p. 337-351.

Pigeot (N.) - 1987 - Eléments d'un modèle d'habitation magdalénienne (Etiolles). : Bulletin de la Société Préhistorique Française 84 10-12, p.358-363.

Pigeot (N.) - 1979 - 1994 -  SeeTaborin, Olive et Pigeot 1979, Coudret  in Taborin 1994.

Pirnay (L.) - 1982 - Méthodes de taille utilisées à l'Ourlaine. p. 73-75 in : Le Mésolithique entre Rhin et Meuse, actes du colloque de Luxembourg 1981, A. Gob et F. Spier éd.,  Société Préhistorique luxembourgeoise.

Pirnay (L.) et Straet (H.C.) - 1978 - Le site mésolithique des Mazures (commune de Pepinster, province de Liège, Belgique). B.S.P.F  75/10, p. 327-340.

Rozoy (Dr J.-G.) - 1978 a - Les derniers chasseurs. L'Epipaléolithique en France et en Belgique. Essai de synthèse.  Charleville, chez l'auteur, 3 vol., 1500 p.

Rozoy (Dr J.-G.) - 1982 - Les éclats retouchés des Mazures. p. 77-88 in : Le Mésolithique entre Rhin et Meuse, actes du colloque de Luxembourg 1981, A. Gob et F. Spier éd.,  Société Préhistorique luxembourgeoise.

Rozoy (Dr J.-G.) - 1988 - Le Magdalénien supérieur de Roc-La-Tour I. Helinium XXVIII, p. 157-170.

Rozoy (Dr J.-G.) et Slachmuylder (J.-L.) - 1990 - L'Allée Tortue à Fère-en-Tardenois (Aisne, France), site éponyme du Tardenoisien récent. En hommage à la mémoire de René Parent. Contributions to the Mesolithic in Europe (The Mesolithic in Europe IV),  p. 423-434. Leuven, P. Vermeersch éd., 30 cm, 645 p.

Rozoy (Dr J.-G.) - 1990 - La Roche-à-Fépin et la limite entre l'Ardennien et le Tardenoisien, p. 413-422 in  : Contributions to the Mesolithic in Europe. Papers presented at the fourth international  symposium. Leuven 1990.  Leuven, Univ., 1990, P. Vermeersch et Ph. Van Peer, ed. 30 cm, 474 p.

Rozoy (Dr J.-G.) - 1992 - Le propulseur et l'arc chez les chasseurs préhistoriques. Techniques et démographies comparées. Paléo 4,  (Les Eyzies), p. 175 - 193.

Rozoy (Dr J.-G.) - 1995 -Evolution récente du cerveau humain. p. 1007-1042 in  : "Nature et Culture", colloque international de Liège, déc. 1993,  Liège, Université, ERAUL 68.

Rozoy (C. et Dr J.-G.) - 1996 -Fouilles sur sable au Tillet. Notae Praehistoricae 16,  p. 123 - 144  (Louvain-la-Neuve, Univ.)

ROZOY (Dr J.-G.) - 1997 a - La fin et les moyens. Quelques mécanismes, causes et significations des changements dans les industries des chasseurs préhistoriques. : Bulletin de la Société Préhistorique Française 94/4,  p. 483-502.

ROZOY (Dr J.-G.) - 1997 b - Nature et origine des variations régionales des industries mésolithiques,  p. 99-107 in  : Tardenoisien et Ardennien,  Bulletin n° 4  du Centre Ardennais de Recherche Archéologique C.A.R.A., Charleville-Mézières, chez l'auteur, 30 cm, 110 p.

Rozoy (Dr J.-G.) - 1997 c - Les capacités mentales des artistes paléolithiques de l'Ardenne. L'Anthropologie,  p. 83 - 113.

Rozoy (Dr J.-G.) - 1998 a - Stratégies de chasse et territoires tribaux au Mésolithique. : Bulletin de la Société Préhistorique Française p. 525-536.

Rozoy (C. et Dr J.-G.) - 1998 b - Le Tardenoisien moyen de Tigny-Les Marnières, analyse topographique. Revue Archéologique de Picardie,  p. 3-29.

Rozoy (C. et Dr J.-G.) - 2000 a - Les camps mésolithiques de l'Allée Tortue à Fère-en-Tardenois. Bulletin de la Société Préhistorique Française 97, p. 5-56

Rozoy (C. et Dr J.-G.) - 2000 b - Conditions d'étude des sites préhistoriques. Le cas des sites mésolithiques de plein-air. Soumis à la Revue archéologique de Picardie.

Rozoy (C. et Dr J.-G.) - 2000 c - Les camps mésolithiques du Tillet. Soumis au Bulletin de la Société Préhistorique Française (« Travaux »).

Rozoy (C. et Dr J.-G.) – 2000 d - L'Ardennien moyen de la Roche-à-Fépin, analyse topographique. Soumis à la Revue Archéologique de Picardie

Rozoy (C. et Dr J.-G.) - 2001 - Le Magdalénien de Roc-la-Tour I, analyse topographique. En préparation.

Schmider (B.) - 1992 - Marsangy. Un campement des derniers chasseurs magdaléniens sur les bords de l'Yonne. Liège, ERAUL 55, 30 cm, 275 p.

Schmider (B.) - 1994 - Marsangy, p. 147-153 in  Taborin 1994.

Taborin (Y.) - 1974 - Note préliminaire sur le site paléolithique d'Etiolles. Cahiers du Centre de recherches préhistoriques de l'Université de Paris-I, 3, p. 5-22.

Taborin (Y.) - 1987 - Une certaine lecture des sols d'habitation, p. 353-357 in : : Bulletin de la Société Préhistorique Française 84, 10-12.

Taborin (Y.), dir. - 1994 - Environnements et habitats magdaléniens dans le centre du Bassin parisien. D.A.F. 43, Paris, Maison des Sciences de l'Homme, 30  cm, 189 p.

Taborin (Y.), Olive (M.) et Pigeot (N.) - 1979 - Les habitats paléolithiques des bords de Seine : Etiolles (Essonne, France). p. 773-782 in : La fin des temps glaciaires en Europe. Chronostratigraphie et écologie des cultures du Paléolithique final. Colloque de Talence 1977, C.N.R.S.

Walczak (J.) - 1997 - Les industries de silex des sites du Mésolithique moyen de Tigny-Les Marnières, commune de Parcy-et-Tigny (Aisne) et de la Roche-à-Fépin , commune de Haybes (Ardennes). Approche comparative, p. 3-98 in  : Tardenoisien et Ardennien,  Bulletin n° 4  du Centre Ardennais de Recherche Archéologique C.A.R.A., Charleville-Mézières, 30 cm, 110 p.

Résumé des abréviations utilisées dans les articles : consulter la liste.

© Jean-Georges Rozoy - Tous droits réservés 2016