Info

+

Figure 2.21 Pollen succession recorded through the Combe Grenal sequence. After Bordes et al. 1966.

birch and pine (Bordes et al. 1966), which suggest that climatic conditions never approached the severity of those recorded during the preceding cold phase of isotope stage 6.

The two warm episodes in the sequence (in layers 52-50A and 43-41) reflect a dramatic change from these conditions. In these levels overall arboreal pollen ratios rise to between 60 and 70 percent and now include substantial frequencies of a wide range of warmth-demanding deciduous species such

Layers

Bovids Horse

Layers

Bovids Horse

Weathering (Eemian) Riss 60%

60 80 100% 20 Percentage of identified remains

Figure 2.22 Fluctuating frequencies of the four principal faunal taxa recorded throughout the Combe Grenal sequence, based on counts of numbers of identified specimens (NISP) of the different taxa. The most striking feature is the shift from the heavily red-deer dominated faunas in the earlier, 'Würm V levels of the last-glacial sequence to the mainly reindeer-dominated faunas in the later, 'Würm IF levels. In the Würm I levels, reindeer are represented by only seven identified specimens (from layers 52 and 40) which could well be derived from the underlying 'Kissian levels, in which reindeer remains are abundant. After Bordes & Prat 1965.

60 80 100% 20 Percentage of identified remains

Weathering (Eemian) Riss 60%

Figure 2.22 Fluctuating frequencies of the four principal faunal taxa recorded throughout the Combe Grenal sequence, based on counts of numbers of identified specimens (NISP) of the different taxa. The most striking feature is the shift from the heavily red-deer dominated faunas in the earlier, 'Würm V levels of the last-glacial sequence to the mainly reindeer-dominated faunas in the later, 'Würm IF levels. In the Würm I levels, reindeer are represented by only seven identified specimens (from layers 52 and 40) which could well be derived from the underlying 'Kissian levels, in which reindeer remains are abundant. After Bordes & Prat 1965.

as elm, lime, oak, alder and hazel in addition to the hardier species such as pine and juniper (Fig. 2.21). The same pattern is reflected in the associated sedimentological data which indicate a sharp increase in both temperature and humidity and some degree of soil formation (Laville 1975; Laville et al. 1980). In certain respects, therefore, climatic conditions in these levels approach those of fully interglacial conditions. Paquereau, however, emphasizes that the character of the associated pollen spectra -- still domi nated mainly by pine and with relatively low percentages of oak - is by no means fully 'interglacial' in character and implies that climatic conditions during these intervals were significantly cooler than those experienced either during the preceding Eemian interglacial or at the present day (Bordes et al. 1966: 12, 16). The evidence suggests, in other words, relatively mild interstadial conditions during the formation of these levels.

Despite these climatic fluctuations the composition of the faunal assemblages recor ded throughout these so-called /Würm/1 levels in the Combe Grenal sequence remain remarkably stable throughout the entire sequence of levels 55-38 (Fig. 2.22). In all levels remains of red deer are clearly dominant, accompanied by more sporadic remains of horses, bovids, wild pig and roe deer (Bordes & Prat 1965; Laquay 1981). How one should explain this relative stability of the faunal assemblages in the face of rapidly changing climatic and vegetational conditions raises interesting issues which will be discussed more fully in Chapter 7. In general terms, however, the character of the faunal remains provides further evidence that climatic conditions during the formation of these levels never attained the extremely rigorous, full glacial conditions reflected in both the earlier Rissian levels of the Combe Grenal sequence and in the overlying Würm II levels on the site.

From the general character and strati-graphic position of the various climatic episodes discussed above there is now effectively unanimous agreement as to how these levels should be correlated with the overall climatic sequence for the Upper Pleistocene (Laville et al 1983, 1986; Guadelli & Laville 1990; Mellars 1986a, 1988 etc.). Since these deposits rest directly on top of a presumed last interglacial soil horizon and immediately below a long period of much colder climate (discussed below) they are generally assumed to correlate with the well documented succession of colder and warmer episodes represented by stages 5d-5a of the deep-sea core sequence (Fig. 2.23). Indeed, it could be argued that the patterns suggested by both the sedimentological and pollen evidence for this part of the Combe Grenal sequence coincide almost exactly with what one would predict from the vegetational and climatic sequences recorded at La Grande Pile and Les Echets (Figs 2.6, 2.7). This implies a dating for these levels in the region of ca 115,000-75,000 BP. The only ambiguity lies in the dating of the initial cold episode represented by levels 55-53 of the Combe Grenal sequence. Whilst the obvious correlation of these levels would be with stage 5d of the ocean core sequence (i.e. the first cold phase which effectively initiated the last gla-ciation), it is conceivable that these levels could correspond with the much briefer cold episode which occurred mid-way through the course of the St Germain I interval (i.e. isotope stage 5c) at Grande Pile, Les Echets and elsewhere (i.e. the so-called 'Montaigu' episode) (cf. de Beaulieu & Reille 1984: 125; Reille & de Beaulieu 1990). If this correlation were adopted it would imply that any deposits formed during the very earliest stages of the last glacial period (during stage 5d and the earlier part of stage 5c) had been removed or truncated by erosion in the early stages of the climatic deterioration. But this remaining uncertainty has little effect on the overall correlation of these Würm I levels in the Combe Grenal sequence with the basic pattern of climatic oscillations during isotope stages 5a-5c.

4. A more dramatic shift in the climatic history of the Combe Grenal sequence is reflected between layers 37 and 35. This episode has been described by Laville et al (1986: 38) as a major threshold ('seuil') in the climatic succession and is marked by the appearance of climatic and vegetational conditions which are far more rigorous than those reflected in any earlier levels in the last-glacial sequence in the site. In the first place, overall arboreal pollen ratios fall to levels substantially and consistently lower than those recorded in any of the earlier Würmian levels in the site and are marked by the total disappearance of all species other than pine - most notably the warmth-loving hazel which had persisted throughout all of the preceding Würm I levels in the sequence (Fig. 2.21) (Bordes et al. 1966). Second, Laville et al.(1986) draw attention to a more general change in the character of the non-arboreal pollen in these levels marked especially by the sharp increase in

Figure 2.23 Proposed correlation between the climatic and vegetational sequence at Combe Grenal and the sequence of oxygen-isotope stages in deep-sea cores (reproduced from Mellars 1986a). The main succession of industries on the site is shoivn on the right.

species such as Ephedra, Galium, Armeria and Poterium which appear to mark the onset not only of more rigorous but much drier and more steppic conditions. Third, this period is marked by a dramatic increase in the frequency of reindeer in the associated faunal assemblages, which rise from effectively zero percent throughout layers 55-37 up to ca 20 percent in layer 35 and then to between 60 and 70 percent in the overlying layers 28-23 (Fig. 2.22) (Bordes & Prat 1965; Guadelli 1987). According to Laville (1975) the associated sedimentological data point equally to the onset of much colder and drier conditions, commencing with layers 36-37 and continuing throughout most of the later levels in the sequence.

There can be little doubt about the correlation of these levels with the climatic sequence in deep-sea cores (Fig. 2.23). All the French workers now agree that the dramatic deterioration in climatic conditions reflected between levels 38 and 35 must coincide with the transition from stage 5 to stage 4 of the isotope record, dated closely in the ocean cores to around 73,000 BP (Laville et al 1983, 1986; Guadelli & Laville 1990; Martinson et al. 1987). Whether this should be regarded as a very abrupt climatic transition is more open to debate. In the vegetational records at La

Grande Pile and Les Echets, for example, the transition from stage 5a to stage 4 is characterized by a sequence of rapid climatic oscillations (the so-called 'Ognon' oscillations) which preceded the emergence of fully 'glacial' conditions at the peak of isotope stage 4 (Figs 2.6, 2.7). A similar pattern may be reflected in the Combe Grenal sequence by the briefly warmer and probably wetter episodes represented in layer 38 and in the clear weathering horizon which separates layers 36 and 35. It could be that the overall climatic sequence at Combe Grenal has been partially truncated over this interval by a phase of erosion between layers 36 and 35. Nevertheless, there can be little doubt as to the general correlation of this climatic transition with the transition from stage 5 to stage 4 in the deep-sea core sequence and there seems no reason to postulate any major break or hiatus at this point in the Combe Grenal sequence.

5. The extremely cold climatic conditions which first appeared in levels 37-35 persisted throughout the great part of the ensuing sequence on the site, i.e. throughout the majority of levels 35-1. In most of these levels overall tree-pollen frequencies remain at extremely low levels (mostly between 5 and 10 percent) and in the majority, pine is the only tree species represented (Fig. 2.21). More significant variations can be detected in some of the non-arboreal species. Thus, relative frequencies of grasses versus composites and heliophiles fluctuate substantially between different levels in the sequence, though perhaps with a general shift from predominantly grass-dominated vegetation in the lower levels (layers 35-26) to mainly composite-dominated spectra in the later levels (layers 25-1 ) (Bordes et al 1966).

The major question hinges on how far one can identify clear traces of the onset of more temperate climatic conditions in the upper part of the Combe Grenal sequence, coincid ing with the shift to the generally milder, 'interstadial' conditions of isotope stage 3. There is apparently pollen evidence for three relatively short-lived episodes of slightly milder climate (represented respectively in levels 22-20, 13-11, and 8-7), in each case reflected by an increase in overall tree-pollen frequencies to around 15-20 percent and by the temporary reappearance of species such as hazel, alder and juniper (Fig. 2.21). The same episodes were also marked by an increase in local humidity reflected in both the sedimentological data and a sharp increase in species such as Cyperaceae and other hygrophiles. Significantly, there is also evidence for a major shift in the composition of the associated faunal assemblages in these upper levels (Fig. 2.22). This is marked by a sharp decrease in the overall frequency of reindeer, which falls from ca 60 percent in levels 19-18 to around 20 percent in levels 15-8, and by a corresponding increase in red deer, horses and bovids (Bordes & Prat 1965; Guadelli 1987). All these shifts can probably be taken to reflect a significant improvement in climatic conditions in the uppermost levels of the Combe Grenal sequence, which almost certainly correlates with the transition from isotope stage 4 to stage 3 (Guadelli & Laville 1990; Mellars 1986a, 1988).

Any attempt at precise dating of the upper levels in the Combe Grenal sequence is difficult from the present evidence. The problems of identifying, separating and correlating the complex succession of separate interstadial episodes which characterizes stage 3 of the ocean core sequence have already been emphasized and any attempt at specific correlation of the events registered at Combe Grenal in these terms might be premature. As discussed below, the evidence from some other sites in southwestern France (especially Le Moustier) suggests that the stratigraphie sequence at Combe Grenal comes to an end well before the end of the overall Mousterian succession in this region, probably by around 50,000 BP (Mellars 1986a, 1988). For more secure evidence of these later stages in the climatic sequence we must rely on findings from other, complementary sites discussed further below.

The question of the absolute chronology of the Combe Grenal sequence has been left to the end of this discussion since this has recently generated a lively controversy. At present there is a series of six thermoluminescence (TL) dates for the sequence produced by the British Museum Laboratory based on samples of burnt flint collected during the excavations by François Bordes in 1958-63 and covering the range from 44,000 to 113,000 BP (Bowman & Sieveking 1983). As several workers have pointed out, if these dates are accepted at face value they not only conflict with the overall chronology and climatic correlations proposed above (i.e. with the correlations now accepted by apparently all French workers) but also imply a pattern of climatic fluctuations within the Combe Grenal sequence which is in obvious conflict with the patterns documented in all other sources of climatic information (i.e. deep-sea cores, ice cores, sea-surface temperatures, long pollen sequences etc.) covering the same time range (Laville et al 1986; Laville 1988; Mellars 1986a,b, 1988). For example, the dates of ca 105,000 and 113,000 BP obtained for layer 60 at the extreme base of the sequence imply that the most severe climatic conditions recorded during the entire Combe Grenal sequence occurred during the period of isotope stage 5d - which emerges in this dating as much colder than isotope stage 4. The dates obtained for layers 49-55 and layer 20 (ca 61,000-68,000 BP and 44,000 BP respectively) similarly reflect the reverse of the climatic patterns expected over this time range, implying exceptionally warm conditions, with full deciduous forest, during the period of isotope stage 4, and very much colder conditions during the interstadial period of isotope stage 3 (Fig. 2.23). As Laville et al (1986: 38-40) have indicated, the existing series of TL dates for the Combe

Grenal sequence is difficult to reconcile either with any reasonable interpretation of the overall climatic and environmental sequence at Combe Grenal itself, or any other climatic sequences documented over this time range.

As Laville and others have pointed out, there are a number of obvious problems with both the provenance and the measurement of the burnt flint samples employed for the TL dating. The six samples employed for the TL measurements were collected during excavations carried out on the site over thirty years ago, were probably exposed to strong sunlight at the time of collection and were stored for nearly 20 years with the archaeological collections at the University of Bordeaux before dating by the laboratory. In addition, it would seem that the levels of background radioactivity in two of the layers involved in the dating were not measured directly (Bowman & Sieveking 1983; Laville et al 1986; Laville 1988). As Aitken and others have emphasized, these are hardly ideal conditions for TL dating of burnt flint samples (Aitken 1985, 1990; Aitken et al 1986). Perhaps more to the point, a series of parallel TL measurements undertaken by the same laboratory on samples from the nearby site of Pech de 1'Aze IV produced results which the laboratory itself described as 'too young to be acceptable' and yielded a date for one of the upper (but not final) Mousterian layers of 19,600±1600 BP - i.e. less than half the known age of the sample (Bowman et al 1982). In the light of these results it would be unwise to place any strong reliance on the existing TL measurements for the Combe Grenal sequence. Given the close correspondence which can be demonstrated between the overall climatic sequence at Combe Grenal and similar patterns in deep-sea cores, ice cores and the long vegetational sequences at La Grande Pile, Les Echets and elsewhere, it seems most reasonable to accept these correlations as providing the best overall chronological framework for the Combe

Grenal sequence, until a new and more tightly controlled programme of absolute dating (preferably employing a combination of several dating techniques) can be applied to the site.

Was this article helpful?

0 0

Post a comment