Introduction

Rock art surveying by a Belgian archaeological mission in March-April 2004 in the el-Hosh area on the west bank of the Nile, about 30km south of Edfu in Upper Egypt, led to the discovery of a hitherto unknown petroglyph locality at the southernmost tip of a Nubian sandstone hill called Abu Tanqura Bahari, about 4km south of the modern village of el-Hosh (Figure 1). This locality (designated ATB11) shows, among other things, several images of bovids executed in a vigorous naturalistic, ‘Franco-Cantabrian, Lascaux-like’ style, which are quite different from the stylised cattle representations in the ‘classical’ Predynastic iconography of the fourth millennium BC. On the basis of patination and weathering, these bovid representations are definitely extremely old. They most probably predate the fish-trap representations and associated scenery previously documented at several locations in the el-Hosh area and AMS ¹⁴C dated to >7000 BP (Huyge et al. 2001; Reference HuygeHuyge 2005). As these el-Hosh bovid images are similar to cattle representations that had been discovered in 1962-1963 by a Canadian archaeological mission (the Canadian Prehistoric Expedition) on the east bank of the Nile, in the Gebel Silsila area, the Belgian mission attempted to retrace the latter images. The attempt was successful and the sites were recovered in October-November 2005 near the modern village of Qurta, along the northern edge of the Kom Ombo Plain, about 40km south of Edfu and 15km north of Kom Ombo (Figure 1). As far as we know, these sites, which are still in pristine condition, have not been visited again by archaeologists since the time of their discovery in 1962-1963.

The Qurta rock art sites

Figure 2.

Scaffolding at Qurta I, locality 1, panel 1 (QI.1.1). Click to enlarge.

Intensive surveying of the Nubian sandstone cliffs immediately east of the village of Qurta in February-March 2007 led to the discovery of three rock art sites, designated Qurta I, II and III (Figure 2). At each of these sites several rock art locations, panels and individual figures were identified. In total there are at least about 160 individual images. The rock art of Qurta consists mainly of naturalistically drawn animal figures. Bovids are largely predominant (at least 111 examples), followed by birds (at least 7 examples), hippopotami (at least 3 examples), gazelle (at least 3 examples), fish (2 examples) and ass (1 example). In addition, there are also (at least) 7 highly stylised representations of human figures (shown with pronounced buttocks, but no other bodily features). All these rock art images are very darkly coloured. They bear a substantially developed patination and/or rock varnish. This, in itself, is already an indication of considerable antiquity. Most of the images also show traces of intensive weathering through aeolian abrasion and/or water run-off. In this respect, the rock art at Qurta is highly homogeneous. In spite of the fact that there are numerous superimpositions of images, it seems to represent one single manufacturing phase.

None of the animals present shows any evidence for domestication. There is little doubt that the bovids represented (Figures 3-6) should be identified as Bos primigenius or aurochs (wild cattle). In general, they seem to be rather short-horned, but there is archaeozoological evidence available (moreover from Late Pleistocene faunal materials collected in the Kom Ombo Plain) that the Egyptian race of Bos primigenius bore relatively smaller horns than the European but was otherwise of about the same size (see Reference ChurcherChurcher 1972).

Figure 3.

Detail of the main rock art panel at Qurta I, locality 1 (QI.1.1), bearing at least 25 bovids.

Figure 4.

A panel with four bovids at Qurta II (QII.4.2). Chalking not done by the Belgian mission.

Figure 5.

A panel with eight bovids at Qurta I (QI.2.1). Chalking not done by the Belgian mission.

Figure 6.

Detail of a bovid at Qurta II (QII.5.1).

In a general way, the rock art of Qurta bears the following main characteristics:

• • as far as the spatial organisation of the art is concerned, there are no evident scenes (compositions displaying a narrative content). Compositions are limited to the juxtaposition of a few images (like two opposed bovids and a bird frieze composed of three drawings at Qurta II). Figures seem rather to be conceived as individual images. In contrast to the rock art of the Predynastic period (fourth millennium BC), there are no imaginary ground lines present. Images can be drawn in all possible directions (and quite often the head is represented upward or downward) (see Figure 5);

• • from a technical point of view, both hammering and incision have been practised to create the images. In a considerable number of cases, both techniques have been combined to create or complete a drawing. Some of the figures are almost executed in bas-relief. A superb example of this is the bovid from Qurta II represented in Figure 6;

• • the dimensions of the drawings are exceptional. Quite often the bovids are larger than 0.80m. The largest example even measures over 1.80m. In this respect the Qurta rock art is quite unlike the rock art of the Predynastic period (fourth millennium BC), in which animal figures are only exceptionally over 0.40-0.50m;

• • often natural features, such as the relief of the rock surface and/or fissures in the surface, have been integrated into images. One typical example of this is a large bovid at Qurta II, where a (only slightly modified) natural vertical crack in the rock surface has been used to render/suggest the back part of the animal (see Figure 7);

• • naturalistic images of animals are combined in this rock art with highly schematised human figures (which are moreover quite closely comparable to stylised human figures as known from the Magdalenian cultural phase of Palaeolithic Europe) (see Figure 8);

• • quite often the drawings are clearly deliberately left incomplete. Elaborately engraved bovids, for instance, lack front legs (see Figure 9) or are otherwise incomplete. In a number of cases animals (both bovids and hippopotami) show numerous scratches over the head and neck, which, evidently, must have some kind of symbolical meaning (see Figure 10).

Figure 7.

Large pecked bovid at Qurta II (QII.6.1). The back part of the animal is entirely rendered by a (slightly modified) natural crack.

Figure 8.

Three incised stylised human figures with pronounced buttocks (superimposed by the belly-line of a large bovid) at Qurta II (QII.3.1).

Figure 9.

Elaborately incised bovid lacking front legs at Qurta I (QI.1.1).

Figure 10.

Incised hippopotamus at Qurta II (QII.3.1). Note the scratches over the head and neck.

Age of the Qurta rock art

The Qurta rock art is quite unlike any rock art known elsewhere in Egypt. It is substantially different from the ubiquitous ‘classical’ Predynastic rock art of the fourth millennium BC, known from hundreds of sites throughout the Nile Valley and the adjacent Eastern and Western deserts. The only true parallel thus far known is the rock art previously discovered (in 2004) at Abu Tanqura Bahari (ATB11) at el-Hosh, about 10km more to the north and on the opposite bank of the river.

In 1962-1963, the above-mentioned Canadian Prehistoric Expedition working in the area (the purpose of which was to salvage as much as possible of the prehistoric archaeological remains in the areas around Kom Ombo which were being prepared at that time for cultivation for the Nubians displaced as a result of the construction of the High Dam at Aswan) discovered and excavated several Late Palaeolithic settlements in the vicinity of the rock art sites. The most important of these is GS-III, situated at a distance of only 150 to 200m from the Qurta I rock art site. At this Palaeolithic site fragments of sandstone were found on which linear grooves had been incised; in one case they formed several deep parallel grooves (Smith 1985). This at least proves that the Late Palaeolithic inhabitants of the Kom Ombo Plain practised the technique of incising sandstone.

The GS-III site and similar sites found by the Canadian Prehistoric Expedition and other missions in the Kom Ombo Plain in the early 1960s are currently attributed to the Ballanan-Silsilian culture (see Reference VermeerschVermeersch 1992). Other occurrences of this culture are known from Wadi Halfa in Sudanese Nubia and from the vicinity of Esna (E71-K20) and Nag Hammadi (Arab el-Sahaba). The Ballanan-Silsilian culture is dated to about 16 000 to 15 000 years ago (BP). Climatologically this corresponds to the end of an hyper-arid period, preceding a return of the rains and the ‘Wild Nile’ stage of about 14 000-13 000 BP.

The fauna of these Ballanan-Silsilian and other Late Palaeolithic sites in the Kom Ombo Plain (Reference ChurcherChurcher 1972) suggests a culture of hunters and fishermen with a mixed subsistence economy oriented to both stream and desert for food resources. It is essentially characterised by the following elements: aurochs (Bos primigenius), hartebeest (Alcelaphus buselaphus), some species of gazelle (especially Gazella dorcas), hippopotamus (Hippopotamus amphibius), wading and diving birds (including numerous goose and duck species) and some fish species (especially Clarias or catfish). With the exception of hartebeest, this faunal inventory perfectly matches the animal repertory of the Qurta rock art sites. Both in the Late Palaeolithic faunal assemblages and in the rock art large ‘Ethiopian’ faunal elements, such as elephant, giraffe, and rhinoceros, are conspicuously absent.

Although the Canadian Prehistoric Expedition initially hinted on several occasions at the high antiquity of the rock art at Qurta (see Smith 1964; 1965), it has, in our opinion, ultimately failed to assess the true importance of its finds. In an article in Scientific American, P.E.L. Smith (1976) stated: 'Interesting scenes of wild animals, including cattle and hippopotamus, are engraved on the cliffs near our Gebel Silsila sites, but no one can prove they were the work of a late Palaeolithic group.' And still later, in 1985, he assumed: '… that the Gebel Silsila art […] is of Holocene age like most or all of the art known to date in northern Africa' (Smith 1985). In our opinion, because of the various particularities outlined above, the rock art of Qurta reflects a true Palaeolithic mentality, quite closely comparable to what governs European Palaeolithic art. We accordingly propose an attribution of this Qurta rock art to the Late Pleistocene Ballanan-Silsilian culture or a Late Palaeolithic culture of similar nature and age. In this respect, it can hardly be coincidental that the comparable site of Abu Tanqura Bahari 11 at el-Hosh is also situated at close distance (only at about 500m) from a Late Palaeolithic site that, mainly on the basis of its stratigraphical position immediately below the ‘Wild Nile’ silts, must be of roughly similar age as the Ballanan-Silsilian industry of the Kom Ombo Plain. There remains, in our opinion, therefore little doubt that the rock art of Qurta must be about 15 000 years old. Direct ages for this rock art are not yet available, but analyses are under way to explore its potential for AMS ¹⁴C dating of organics in the varnish rind and/or U-series dating.