There are many examples of Palaeolithic portable engravings that have been discovered, long after their excavation, among the collections stored in museums. For example, a remarkable pair of bear figures was spotted in the mid-1980s on a rib fragment housed with the bone industry from the Magdalenian cave of Isturitz in the western Pyrenees; the rib came from a level excavated by the St Périers in 1931 (Esparza & Mujika 2003). It is far rarer, however, for a new engraving to be found among faunal material curated within a palaeontological collection. We report here the discovery by one of us (LMK) of a horse engraving in the collection of the Palaeontology Department of the Natural History Museum (NHM), London, some 140 years after the excavation and acquisition of the specimen. The new engraving was found among the horse remains from the Late Magdalenian site of Roc du Courbet, Bruniquel, France.
The Roc du Courbet is one of a series of Upper Palaeolithic rockshelters near the village of Bruniquel, in France's Tarn region. Located within the face of a massif or cliff (now known as Courbet), it was first explored by Vicomte de Lastic Saint-Jal in 1863, and then by de Lastic, M. des Serres and R. Owen in 1864 (Owen1869a). It is believed that the majority of the remains recovered were derived from de Lastic's black layer or 'couche noire', which is thought to date to Magdalenian V or VI; however, the provenance of remains within this layer is not clear (Sieveking 1987). Furthermore, Owen (1869a, 1869b) notes that both human and animal remains were found within not only a black layer ('limon noir'), but also a red layer ('limon rouge') and a breccia deposit. Owen took such an interest in the cave itself, as well as the faunal and artefactual remains which had been recovered from it, that he purchased de Lastic's first collection for the British Museum in 1864 (Cook & Welté 1995).
The faunal collection is now housed at the NHM, while most of the artefacts are housed at the British Museum (Franks House). According to Owen (1869b) the horse remains acquired by him in 1864 consisted of at least 30 individuals from the limon noir, limon rouge and breccia. LMK has recently examined all available horse specimens (teeth, cranials and postcranials) from Owen's 'de Lastic collection' at the NHM. Many retain the cut marks associated with butchery, especially on the teeth and maxillary/mandibular elements. It is also evident that the remains have been treated with an unknown preservative.
Owen (1869b) designated the Courbet horses taxonomically as Equus spelaeus, determining by detailed comparison that they were caballine horses rather than zebras or asses. He considered, based on the size of third molars, that two types or varieties were present. Nowadays the remains are referred to E. ferus and biometric analysis is consistent with the presence of a single species (Kaagan 2000).
Radiocarbon dating of artefactual remains in the British Museum collection was undertaken at the Gif laboratory—this yielded two dates for the 'limon rouge' and one for the 'limon noir', all of which are equivalent at 1s and lie within the range 13 380–13 490 BP (uncalibrated median values), within the early part of the Late Glacial (Cattelain 2005). Radiocarbon dating of purified collagen from a lower P3 of horse (NHM Palaeontology Department no. 39325) from the faunal collection yielded a date of 13 230±90 BP (OxA-6667), consistent with the dates obtained previously.
Many art objects (engravings or drawings) have been found at Courbet, utilising bones of deer, birds and other animals. These include a bird bone and deer rib, both with engraved reindeer, figured by Owen (1869a), and a bone fragment bearing three horse heads on one side and two on the other (Sieveking 1987: pl. 68; here Figure 1).
Owen (1869b) had taken a particular interest in the horse remains from Courbet, and expended considerable effort on cataloguing and interpreting the collection. However, the fresh study of this material by LMK (Kaagan 2000) led to the discovery of a previously unnoticed engraving on a horse's right fourth metatarsal (no. 38475). This unusual choice of material is the delicate 'splint bone' that in life lies lateral to the cannon bone (third metatarsal) of the hind leg; the bone is complete and is 15.9cm long.
The engraving (Figures 2–4) depicts a horse's head in left profile located on the bone's dorsal surface at the proximal end. The image was partly obscured by the specimen identification sticker, confirming that it had not been noticed by previous workers. It is a good example of a 'naturalistic' depiction, in typical Magdalenian style, of a Late Pleistocene wild horse. Engraved lines above the ear suggest that these animals had a forelock. This is a common trait in domestic horses, but many wild Przewalski's horses also grow a forelock before their mid-summer moult, especially in old age or due to lack of fitness (Mohr 1971). Compared to other Palaeolithic depictions of horses, the muzzle in the Courbet engraving is unusually square in shape; other horse-heads from the site seem to share this trait, albeit to a lesser degree.
The horse is the most important animal in Ice Age iconography, and past studies expended a great deal of effort in the attempt to establish which 'races' were depicted—some scholars even claiming up to 37 varieties (see Bahn & Vertut 1997: 141). However, all such exercises were eventually abandoned, being based primarily on the shape and size of the figures, their manes, and the colour and pattern of their coats, all of which could have been subject to artistic whim. All that can be stated with confidence is that most depictions of Late Pleistocene horses strongly indicate that the wild (Przewalski's) horse is their closest modern analogue (Mohr 1971), although even the latter has been under human selection for 'primitive' appearance (M. Bower, pers. comm. 2010). Genetic data have indicated that Przewalski's horse is closely related to, but was not the direct ancestor of, modern breeds; they derived from a common ancestor which is no longer living (Lau et al. 2009).
We are grateful to Marion Duffin for drawing Figure 4, Kevin Webb for photography (Figures 2 & 3), and Jill Cook for discussion.
*Author for correspondence