Internet Archaeol. 5. Walker & Gibert

4.5 Cueva Negra Mousterian artifacts

Cueva Negra Mousterian artifacts are numerous (Table 32 - pdf only), despite the intractable nature of raw materials available at the nearby conglomerate outcrop (Figure 9).

fig9
Figure 9: Cueva Negra: distribution of raw materials of artifacts (retouched and unretouched flakes, spalls and fragments

Both despite and because of that, a great deal of knapping took place. Most went little further than producing flakes (Figure 10), or usable fragments for immediate use and disposal, amidst a multitude of waste chips (for both sites, graphs show lithic findings before 1996 and inventories tabulate classifiable finds before 1997).

fig9
Figure 10: Cueva Negra: distribution of raw materials of flakes only

At Cueva Negra, few edges allow more retouch than a single notch or knapping scar before shattering supervenes - even when (not often) chert permits regular primary flakes.

Plate 34: Cueva Negra: scrapers. Scale in cm. (Photo M.J.Walker)

Nevertheless, well-made flat scrapers occur of flint, chert, quartzite, and fine, hard-grained limestone or dolomite (Figures 11, 12: Plate 34).

fig11
Figure 11: Cueva Negra: distribution of tool types

fig12
Figure 12: Cueva Negra: distribution of raw materials used for flatter side scrapers

By contrast, flint or chert characterize most denticulates and end- and keeled scrapers, and sporadic well-made gravers or burins. Several are on good-quality flint and of a size implying that large nodules were sought, and brought perhaps from several kilometres away. Two recurrent centripetal disc cores occur, one of white flint found at the conglomerate outcrop, another of hard fine-grained limestone found near the cave. A scatter of small white flint flakes was excavated in the upper part of unit 2 at the front of the rock-shelter. Perhaps this flint came from the same source as the disc core, though white flint is rare in the conglomerate, so the artifacts likely came from outcrops of white flint 40km away.

Darker flint could, perhaps, come from an outcrop 15km upstream, near a hill ("Carro") whose twin peaks are just visible on clear days from the cave terrace, where they peep just above the southern horizon - giving a bearing from the cave. Perhaps significantly, they are out of sight from rock-shelters across the gorge from Cueva Negra. This might allow an actualistic conjecture that the cold, north-facing cave was chosen partly because of useful views from it. Levallois points seem uncommon here, unlike Sima de las Palomas, though a few triangular flakes occur, perhaps struck from recurrent centripetal disc cores. Unlike Sima de las Palomas, Cueva Negra lacks Mousterian "stubby points" (perhaps thick convergent scrapers). Almost no blades occur struck from prismatic or pyramidal cores, or other signs of upper palaeolithic workmanship. Only a handful of fragments show crazing caused by heat (too small for thermoluminescence dating at Oxford); curiously, burnt bones are extraordinarily rare and ashy soil is nonexistent at Cueva Negra.

Preliminary scrutiny of the entire Cueva Negra collection - including finds from coarse-sieving of disturbed soil thrown up by Civil War and other diggings, which may have skewed the distribution - indicates a preponderance of smallish irregular chert flakes, followed by those of hard, fine-grained limestone or dolomite, showing single-facetted striking platforms far more often than multiple-facetted ones, and usually lacking retouch beyond a sporadical (sometimes notch-like) edge-scar - possibly caused by casual mechanical damage.

This aspect of the collection underpins our conviction that microscopic inspection of use-wear and edge-damage may be more revealing at this early stage of research than attempts to refit stone fragments together - given difficulties for lithic refitting presented by the intractable nature of the raw materials with their tendency to shatter, and the limited and probably incommensurable areas of different units excavated that will continue to thwart spatiotemporal lithic-refitting analysis until further excavation has brought our 25m² to a similar depth.

A few artifacts show definite signs of bipolar knapping - perhaps unsurprising given the small size of many pebbles from the conglomerate. There are also rejuvenation flakes that imply reduction of Levallois blanks, though no blank has been found. A start has been made of measuring spalls, chips and fragments, with a view to statistical exploration of hypotheses about core-reduction sequences. Although there are flakes obviously made by rejuvenation of blanks, the tendency of local chert to shatter impedes their reconstruction: the nearby conglomerate contains many subangular nodules, often only 50-100mm across, in which chert and calcite are interleaved (in 'twin'-like fashion, so to speak) brought about by metamorphism induced by sedimentary pressure. Whilst this macroscopic interleaving is not true mineralogical twinning at a microscopical level of petrographic inspection, it nonetheless highlights the incomplete nature of the metamorphous process, as perhaps do also the incompleteness of cortical development over the chert nodules and their blocky, subangular shape which contrasts with the many well-rounded cobbles of fine, hard-grained, metamorphosed Jurassic limestone and dolomite in the conglomerate outcrop.

Another aspect of the predominance of unretouched artifacts, and of the spatial and volumetric differences of our excavations of units 2 and 3, is the inappropriateness, at the present stage of research, of attempting to draw classical 'Bordes'-type contrasts by using cumulative percentage frequency graphs. Apart from very sound scientific reasons for rejecting this approach as little more than scientistic humbug anyway (Kerrich and Clarke 1967), it is hard to see what advantages might possibly ensue when it is well-nigh impossible to identify more than a handful of François Bordes' 60-odd clear-cut types (as revised in the very useful guide by Debénath and Dibble 1994) and where these amount to only a trivial numerical proportion of our collection.

Particular classificatory problems of denticulates and end- and keeled scrapers are presented by the nature of the raw materials and vagaries of its fracturing. It can be hard to tell apart notched flakes, with a single large scar, from possible denticulates where a small notch-like flake-scar is accompanied by another doubtful one. Another difficulty concerns pieces that run the gamut from side scrapers with diminutive steep retouch to others with invasive retouch or semi-invasive steep retouch, from end-scrapers on flakes to thumbnail scrapers, and from keeled scrapers to what look for all the world like small pyramidal blanks for microlithic bladelets (though absence of microburins assuredly rules out leptolithic knapping). Clear-cut end-scrapers on flake-blades and flakes, and also thumbnail scrapers, occur both here and at Sima de las Palomas. However, it is particularly hard at Cueva Negra to distinguish trimming undertaken in order to rejuvenate either nodules or flake-blanks, from retouch of either in order to produce clearly identifiable tool types from them: this is because of both petrological diversity and irregular fracturing. Thus, we find flake scars on keeled fragments that cannot always be separated unequivocally into scrapers, blanks, or chunky rejuvenation débitage. Nor can sharp distinctions always be drawn between side scrapers and steep scrapers. Another problem is posed by a possibility that some artifacts were themselves objects of further reduction, for which we have evidence in the form of patinated artifacts with patina on some flake scars, that were subsequently retouched without patina having formed on these later retouch scars.

For reasons mentioned above, prudence limits our comparisons and contrasts between at most half-a-dozen unambiguous categories: flattish side scrapers, denticulates (including notched pieces), carinated (keeled) pieces and very steep or end-scrapers, struck flakes without retouch, hammer-stones, and other cores (nodules, cobbles). Their rank order in unit 2 resembles that in unit 3, though the few hammer-stones in unit 3 don't show up in the graph whereas its denticulates seem to outnumber those in unit 2. We have, nevertheless, excavated pebbles and cobbles >50mm across in unit 3, as well as their reduction products. In order to reach unit 3, it is hard to envisage such large cobbles as having fallen down through retraction fissures in unit 2, which later sealed up somehow, even if smaller stones could have done so. Although only 6m² of unit 3 have been excavated, and only 1m² of unit 4, the presence of cobbles implies human activity, albeit intermittently, over a long time. Some cobbles with pitted surfaces were hammer-stones; unpitted cobbles are fractured, suggesting they were raw materials for knapping. Visibility of raw materials not far away, not to mention availability of water and game or carrion nearby, doubtless favoured repeated use of the rock-shelter.