Figure 6.22 Stratigraphic distribution of Denticulate and Typical Mousterian industries throughout the Mousterian succession at Combe Grenal, according to Bordes 1961b, 1972 etc. The industries from layers 28-31 were originally classified by Bordes (1955a) as representing an 'attenuated Ferrassie' variant, but were later reclassified as 'Typical Mousterian enriched in racloirs.
ticulate and Typical Mousterian assemblages becomes less surprising when the specific typological and technological features which supposedly define these two variants are examined in detail. In terms of the classic Bordes taxonomy, the definition of each variant rests almost exclusively on a single basic typological parameter, namely the overall frequency of racloirs (in relation to the combined frequencies of all other tools) in the tool assemblages as a whole. Thus the Denticulate variant has always been defined primarily by an exceptionally low percentage of racloirs (below 20 percent and usually less than 10 percent) while the Typical Mousterian is defined by a 'moderate' percentage of racloirs ranging between ca 20 and 55 percent (e.g. Bordes 1953a, 1961b, 1963, 1981, 1984). The other features used by Bordes to define these two variants are almost entirely negative - i.e. the absence of features specifically characteristic of the other three better defined variants (i.e. typical hand axes and backed knives, significant frequencies of tools shaped by Ou in a retouch, the scarcity or absence of bifacially retouched racloirs, typical limaces etc). The only possibly more positive feature employed in Bordes' definitions of these two variants is the relatively high frequency of notched and denticulated tools (usually between ca 40 and 60 percent) which is taken as an essential feature of the Denticulate variant (Bordes 1963). Since this is determined mainly by the scarcity in these assemblages of more extensively retouched forms, and is also characteristic of some later [Type B] MTA assemblages, its status as a significant positive feature of the Denticulate assemblages is very tenuous.
The implications of these observations are evident. Defined primarily in negative typological and technological terms the status of the Denticulate and Typical variants as significant or meaningful industrial or taxo-nomic entities must be open to grave doubt. There are indications that Bordes himself was well aware of these problems. In discussing the Denticulate assemblage from the upper levels of the Abri Chadourne, for example, he commented 'il reste possible que le Moustér-ien à denticulés represente un phenomène de convergence, un cul de sac d'où aboutiraient, par dégénérescence, d'autres types d'industries' (Bordes et al 1954: 249). He was even more explicit in discussing the various industries attributed to the Typical Mousterian group. In his 1977 paper he commented 'Typical Mousterian is the most variable, and may represent, in the present state of our knowledge, a kind of "rag bag" in which are put all the assemblages which are neither MTA nor Quina-Ferrassie or Denticulate' (Bordes 1977: 38). On several other occasions he suggested that the industries previously attributed to the Typical grouping certainly required subdivision into two or more separate groups (e.g. Bordes & Sonneville-Bordes 1970: 63, 64, 68; Bordes 1977: 38,1981: 79,1984:151). When viewed in these terms the failure of the Typical and Denticulate industries to occupy a neat, consistent and well defined strati-graphic and chronological position within the total sequence of Mousterian industries in the southwestern French sites is hardly surprising.
A closer examination of the stratigraphie positions of the Denticulate and Typical industries in the Perigord sites, however, reveals a stronger element of chronological patterning than has generally been recognized in the past (Mellars 1969). As several workers have pointed out (e.g. Rolland 1988b; Dibble & Rolland 1992) there are now strong indications that occurrences of tax-onomically Typical Mousterian industries are particularly common during the earliest stages of the last-glacial sequence (i.e. during the various stages of isotope stage 5) and indeed with two or three possible exceptions may well represent the only industrial variants recorded from this part of the last glacial sequence within the Perigord sites - reflected for example in all the well documented successions of Würm I assemblages recorded at
Combe Grenal (Fig. 6.22), Pech de l'Aze sites II and IV, Roc de Marsal, Le Regourdou and La Chaise. The only other well defined occurrences of Typical Mousterian industries date from the middle or later stages of the ensuing Würm II phase and are best represented by the block of assemblages stratified between the uppermost levels of Denticulate Mousterian and the overlying levels of MTA in layers 6-10 at Combe Grenal (Bordes 1972) (Fig. 6.22). As I have discussed elsewhere (1969: 159; 1988), the block of three Typical Mousterian industries stratified wTithin the main sequence of Ferrassie Mousterian levels at Combe Grenal (in layers 28-30) were described in Bordes7 (1955a) initial publication of the site as representing an 'attenuated Ferrassie' variant and differ from the immediately adjacent industries only by a slight reduction in the overall percentage of racloirs in the assemblages. In most respects therefore these industries could be seen as an integral part of the long and continuous succession of Charentian industries on the site.
As regards the Denticulate Mousterian, it now seems clear that these industries are similarly restricted largely to two separate time spans in the Perigord sites. The earlier is represented by single levels of Denticulate Mousterian found stratified within the longer sequences of Typical Mousterian assemblages in the later Würm I levels at
Combe Grenal (layer 38), Pech de l'Aze II (layer 4B), the Roc de Marsal (layers 2-3) and Moulin du Milieu (layer X). The later and better documented horizon is represented by occurrences of characteristically Denticulate assemblages stratified immediately above the rich and typical sequences of Quina-Mousterian industries at Combe Grenal (layers 16-11), Abri Chadourne (layers A and A-B), Hauteroche (layer 3), and (further to the north) Roc-en-Pail (Maine-et-Loire). The single, isolated level of Denticulate Mousterian recorded by Bordes within the uppermost levels of Quina Mousterian at Combe Grenal (in layer 20 - separating the main block of Quina industries in layers 26-21 from the three final Quina levels in layers 19-17) appears to be closely related to this main grouping of post-Quina Denticulate assemblages and stands as the only well documented example of industrial interstra-tification so far recorded in the numerous long and multilayered successions of Mousterian industries in southwestern French sites. In particular it should be recalled that despite the existence of many other long and complex sequences of Charentian industries at several other sites in this region (e.g. Abri Chadourne, Roc de Marsal, Chez-Pourrez, Abri du Chasseur, Abri Caminade, La Ferrassie etc.) no other occurrences of this kind have so far come to light.
Was this article helpful?