Le Rocher de la Caille: un site magdalénien du plein air au Saut-du-Perron edited by H. DELOGE & L. DELOGE
Société Préhistorique Française Mémoire XXXI. 2003, 252 pages, 96 figs, 29 plates. (€40)

Le site mésolithique des Baraquettes (Velzic, Cantal) et le peuplement de la moyenne mantagne catalienne, des origines à la fin du mésolithique, edited by F. SURMELY
Société Préhistorique Française Mémoire XXXII. 2003, 282 pages, 171 figs, 14 plates. (€35)

2003 saw the publication of two major French sites separated from each other by over 3000 14C years: the Magdalenian site of Le Rocher de la Caille and the Sauveterrian sites of Les Baraquettes (Société Préhistorique Français, Mémoires XXXI and XXXII). The first volume, involving the collaboration of 15 scholars, is a welcome publication of a late Magdalenian site in the Loire. A number of Middle and Upper Palaeolithic sites have been excavated in the Saut-du-Perron, situated on the left bank of the Loire as it enters the Roanne Plain. The Rocher de la Caille was initially investigated from 1970 to 1976, and fully excavated between 1979 and 1982, after which it was flooded due to damming. The site lies between two meanders of the Loire as it passes through steep gorges, and lies in close proximity to the Mousterian site of Champ Grand and Gravettian site of Vigne Brun. At Rocher de la Caille, solifluxion deposited a large amount of limestone cobbles into alluvial sands, and two arcs of these materials were originally interpreted as the remains of Upper Palaeolithic tent structures, although doubt still hangs over these even if it is certain that the site represents a substantial Magdalenian habitation in the style of those further to the north in the Paris Basin. Some limitations of the Rocher de la Caille site are evident. The pollution of the pollen assemblages by Holocene pollen precludes a reliable reconstruction of the palaeoenvironment at the time of habitation. Faunal preservation was very poor, with the total assemblage amounting only to around 50 teeth, all of horse bar one of chamois, and of a number of small, burnt fragments that are unidentifiable to species. The single 14C date for the site (Ly-5645: 12210 ± 480) was therefore measured on carbonised sediment taken from a hearth. This is, at least, in accord with the typological nature of the lithic assemblage, which appears to be transitional between the Middle and Upper Magdalenian. The one date and the typology of the lithic assemblage place the use of the site at the very end of the first half of the Lateglacial interstadial, i.e. the Bölling. After the limitations have been noted, the following 12 chapters present rich evidence of numerous manufacturing, tool-use and artistic activities, and these are brought together in three synthetic chapters by way of conclusion. Overall, despite the limitations of environmental and faunal preservation, the publication is a welcome addition to our understanding of Magdalenian campsites in general.

Raw materials of volcanic origin such as granite and crystalline cobbles were obtained from the immediate vicinity of the site, and no more distant than 10 km. Flint does not outcrop naturally in the gorges of the Loire, although outcrops in large quantities occur to the southeast and especially northwest. Three raw material groups of flint are represented at the site: the first comprises poor-quality but abundant material from a hydrothermal vein within 5km of the site and between 10-40km from the site, on which laminar technology was impossible; secondly, varying Cretaceous flints from the south Paris Basin some 200km to the north which represent 66% of the lithic assemblage; and finally a group of materials at present unidentifiable to source. In all this points to the considerable importance of high-quality allochthonous flints for the laminar technology ubiquitously employed at the site. This attests at the very least to group movement at least 40km downstream and 10km upstream.

The lithic assemblage includes 1479 formal tools among some 18,400 pieces, most on the good quality allochthonous flint. It is typically Magdalenian, the formal tools dominated by armatures and gravers. The abundance of blade and bladelet cores, in addition to numerous manufacturing waste, attest to considerable amounts of manufacture on-site. This presumable relates to a necessarily economical use of imported material in the context of local rocks that were unsuitable to laminar technology. Most lithic technology at the site involved the use of blades or blade fragments as supports; flakes are very rare. The identification of characteristic signs of pressure flaking on the cores was facilitated by experimental knapping by Jacques Pelegrin, and the complex laminar chaîne opératoire is amply described and illustrated. The overall impression is of considerable technological homogeneity overwhelmingly making use of pressure flaking. The tool count – around 8% of the total amount of flint, is high, and retouched blades and bladelets dominate, although there is considerable typological diversity that presumably attests to the numerous tasks performed at this habitation site. Backed bladelets are abundant (101 complete; 427 fragmentary), and take lunate, straight and obliquely-truncated forms varying from 1.3 to 3.4cm in length. Some of the obliquely-truncated forms resemble northern European Creswell/Cheddar/Tjonger points. The 229 recovered burins are highly variable in form, and form the next important category of tools after the armatures. Of these, burins dièdres dominate (37%), followed by burins sur troncature (22%) and burins sur cassure (18%). The recovery of a number of number of chutes de burin attest to their production and use on-site. The dominance of backed bladelets over burins is exceptional for the later Magdalenian of the wider area, although in other respects the assemblage is broadly comparable to contemporary (i.e. Bölling) and slightly younger (Dryas II) Magdalenian sites of the Paris Basin such as Marsangy (Schmider 1987; 1992). The broken and fragmentary nature of a number of pieces esquilles attests to their prolonged use in, one assumes, heavy-duty tasks. 159 perçoirs, microperçoirs and becs attest to precision boring/scoring, possibly of hard animal tissue, and it would seem that this occasionally involved the use of ochre given its recovery on a number of these pieces. Presumably the 80 recovered lames appointees largely functioned as general knives. Only 48 grattoirs and 45 pièces tronquées were recovered.

Given a degree of downslope movement of material that is evident, reconstructing spatial patterning of lithics is not without problems. Overall, the lithics were recovered from 112m2 of deposits and a degree of patterning is observable. The greater concentration, i.e. the primary activity area, comprises some 30m2 within this, in which tools on bladelets are particularly abundant, and appears to represent, in the main, activity centred around a large hearth. Other clustering by tool type is apparent, for example the burins to the centre and northeast of the principle activity area, where they were used and discarded following their manufacture nearer to the hearth. Large blocks of the soliflucted material also formed foci for activity. Scattered across the occupation area were some 30 cobbles of volcanic rock which served as hard hammers for knapping, pounders and anvils for bone smashing, mallets for bone and antler working and probably as pestles for grinding vegetal matter. Twenty-nine fragments bowls carved largely from steatite were recovered across much of the primary activity area. Two (possibly three) of these served as lamps and were intimately associated with hearths. At least sixteen appear to have served as containers, of which eleven bore engraved decoration. Four of these containers bore traces of a black residue in their concave interiors, which spectrometric analysis revealed to be fatty acids probably constituting mineral colourants. Of the decorated containers, it is the slightly raised rims on which the decoration is found, either in the form of square bosses, or, most commonly, incised lines either radiating outwards or criss-crossing. Given the recovery on the site of a fragment of fossil ammonite, the source of which appears to have been at least 70km to the northeast, it is tempting to interpret the decoration of the containers as deliberate replications of the ammonite, perhaps with some ritual meaning given their appearance at other Magdalenian sites, although this of course remains speculative. Traces of shaping by scraping and polishing are present on all pieces, and experimental replication of these revealed that most were abandoned during the course of manufacture as opposed to having been used and abandoned.

Forty fragments of engraved schist plaquettes were recovered, again mainly in the primary activity area and particularly at its southern periphery. Most of these take the form of very small fragments bearing only partial traces of the images they originally bore, which were depicted with fine and shallow engraved lines. Most depictions are figurative, with horse perhaps not surprisingly dominating (6 depictions). These certainly fall into the cannon of Middle/Upper Magdalenian engraved plaquettes such as those recovered from Enléne, La Marche, Lortet, Gönnersdorf and elsewhere. As with others these appear to have been executed with rapid, confident strokes suggesting that the production of such engravings was perhaps frequently undertaken. Other decorative or symbolic activity is indicated by the staining of sediments by red ochre, which occurs in no particular cluster although appears to be restricted to areas away from the hearths.

Overall, the distribution of burnt fragments of bone, the hearths and the lamps suggest that activity was concentrated largely around the three hearths about four metres apart, and the refitting of some material between these may suggest that they were contemporary. One of these (from which the 14C date came) was surrounded by large limestone blocks that appear to have functioned to delineate and shield the hearth. Overall, the Rocher de la Caille report contributes to our understanding of the organisation of open-air Magdalenian campsites, and in particular to decorative, artistic and symbolic activity.

Eight scholars contribute towards the publication of the Mesolithic sites at Les Baraquettes in the mountainous Cantal region of the Auverne, some 120km southwest of Saut-du-Perron. A number of small Mesolithic sites are known in the region, and several rockshelters are known from Les Baraquettes itself, of which four were excavated in the 1990s. The apparent density of occupation following deglaciation is perhaps not surprising given the ecotonal richness that the Cantal uplands offers in the form of abundant water, variety of altitudes, rich soils, varied microclimates and a high faunal taxonomic diversity which are outlined in the first section of the volume. The second section describes 11 cave and rockshelter sites that, along with the Baraquettes sites themselves allow a good understanding of early and middle Mesolithic activity in the region. While short, these synopses of sites are an invaluable summary of their main characteristics, which span the early (i.e. Sauveterrien) and later (Tardenoisian) Mesolithic. The four Baraquettes shelters are situated on the right bank of the Jordanne river at 780 metres above sea level and some 80m above the river itself, and probably sample archaeological deposits that are effectively continuous along the rockshelter line. Organic preservation was poor, a general phenomenon in montane areas. At the Baraquettes 1 shelter, fauna was preserved. At least nine species were recognised, dominated by chamois (MNI=7), which, on the basis of dental evidence were hunted in spring. Most of the lithic raw materials derived from within 5km, but a few pieces originated over 200km distant. Bladelet cores dominate, and are generally small and worked down. The technology followed a typically Middle Sauveterrien chaîne opératoire geared to the production of rectilinear bladelets on which Sauveterre points, backed points, scalene and isosceles triangles and other forms were produced. A number of points were fractured in a manner suggestive of impact damage and attesting their use as armatures. The Baraquettes 2 shelter is smaller but essentially similar, i.e. with Sauveterre points and a 14C date of 8250 ± 50 (Beta-122219). Baraquettes 3 is so small it evades classification beyond the Middle Mesolithic although it is not inconsistent with the material from the other three shelters. The largest assemblage, both numerically and spatially, is the Baraquettes 4 shelter, with over 3200 lithic pieces clearly attributable to the Middle Sauveterrian. 14C dates average around 8800 BP, although the range of these suggests at least two main periods of occupation. As with the shelters 2 and 3, fauna is not preserved apart from small fragments of wild boar teeth. Wood charcoal suggests the presence of mixed oak forests, and carbonised hazelnut and acorn shells are abundant. This attests to the importance of gathered resources, as are hinted at by two fragments of leguminous grains and a macrofossil of Prunus. The rich lithic industry is, once again, clearly Middle Sauveterrian in character, dominated by triangular armatures. Refitting of bipolar blade and bladelet cores was possible, attesting the integrity of the assemblages.

The Baraquettes sites form an important reference for the understanding of the Mesolithic of the Massif Central, which was hitherto poorly understood only through small occupation sites that presumable functioned as simple halts. The lithic analyses allow the reconstruction of an evolution from the Azilian through early Mesolithic to Sauveterrian, geared around the production of narrow blades and increasingly towards triangular armatures, with the Baraquettes material demonstrating high levels of production standardisation presumably relating to some extent to arrow aerodynamics. Economically, the sites demonstrate the importance of gathered resources within the ecotones including mixed oak forest, and hunting of wild boar (and probably aurochs and deer too from the shelter 1 fauna) with arrows generally tipped with scalene triangles. Rather than supporting traditional models of the montane environment supporting the specialised hunting of chamois and ibex, the data indicate the importance off ecotonal areas often rich in primary and secondary biomass to Mesolithic economy, in which flint was readily accessible.

P. B. Pettitt
University of Sheffield

Schmider, B. 1987. Le Magdalénien dans le centre du Bassin Parisien: les gisements, l’industrie lithique. In Rigaud, J.-Ph. (ed.) Le Magdalénien en Europe. Liège: ERAUL 38, 219-35.
Schmider, B. 1992. Marsangy: Un Campement des derniers Chasseurs Magdaléniens, sur les Bords de l’Yonne. Liège: ERAUL 55.

Review Submitted: March 2004

The views expressed in this review are not necessarily those of the Society or the Reviews Editor.

Home Page The Prehistoric Society Home Page