1 | Study Context

1In the course of hominin evolution, the genus Homo acquired increasingly predatory behavior and developed carnivory (e.g., Stanford, Bunn (eds) 2001 ; Hart, Sussman 2011 ; Thompson et al. 2019). As a result, our genus became a full-fledged member of the carnivore guild, in particular as a top predator and meat-eater on the same level as lions, leopards, wolves and hyenas. It then entered into competitive relationships with these species, sharing the same prey spectra, but also seeking out common living areas (e.g., caves) (Brugal, Fosse (eds) 2004a). At this stage, we could also mention the case of bears (Ursus spelaeus, U. arctos), a more omnivorous and vegetarian species, but one with which human groups may have come into competition. It is worth remembering that humans are also omnivorous (Mann 1981; Fischler 1990).

2The relationship between meat-eaters, especially linked to Paleolithic hunter-gatherer groups, is a research topic that has received considerable attention, and the exploitation of non-human predators by these past human groups has been the subject of numerous studies. Such studies concern particularly felids, canids and ursids, as well as smaller carnivores such as foxes, badgers and wolverines (e.g., Tillet, Binford (Eds) 2002; Castel, Madelaine 2003; Chauvière, Castel 2004; Val, Mallye 2011; Fourvel et al. 2013; Mallye et al. 2013; Val et al. 2020; Sanchis, Pascual (eds) 2017; Fosse et al. 2017, 2019, 2024; Romandini et al. 2018; Russo et al. 2023; Lateur et al. 2024). Human exploitation of these carnivores during the Paleolithic had different objectives: recovery of fur, teeth or claws (used as ornaments or retouchers) with utilitarian and/or symbolic scopes. Such practices are known from various cultural contexts, and are associated with several hominin species (i.e., Neanderthals, Anatomically Modern Humans/Homo sapiens).

3At this point, it can be noted that, unlike herbivore species, carnivores have generally been exploited only slightly. One of the reasons is ecological and related to the trophic chain: carnivores, in particular the larger taxa, are potential competitors and conflictual species for humans. Their density on a given landscape is generally lower than that of herbivores. Moreover, they are not the main objective of human groups, which remains essentially of a dietary nature. In fact, the exploitation of large carnivores in the Paleolithic remains relatively unusual, with the clear exception of bears, an omnivorous species that hibernates (Armand 2018; Toniato et al. 2024; and see below). The objectives can be diverse, from an interest for their meat to utilitarian concerns (pelts/furs, teeth) that can touch on symbolic aspects.

5Among these large carnivores, one group remains clearly less represented and less frequent in the fossil record, namely members of the Hyaenidae family, with its most emblematic representative in the European Paleolithic: the cave hyena Crocuta crocuta spelaea.

6Following the discovery, in the faunal assemblage from Coudoulous II, of hyena remain bearing evidence of anthropogenic marks, and after revising partly the faunal collection from Grotte du Renne, Arcy-sur-Cure, we initiated a literature review of published occurrences of cave hyena bones and teeth modified and exploited by humans. We start by a description of the Coudoulous II specimen, with development about the identification and morphology of the marks. We then present all known cases of human-modified hyena remains, including specimens from our second studied site from Arcy-sur-Cure. We provide information on their paleobiological and chronostratigraphic contexts, and we propose a discussion on the objectives of prehistoric people. Finally, we consider the modern use of hyena remains, especially in African societies to feed the discussion about past hyena exploitation.

2 | Coudoulous II (Lot, Tour-de-Faure, France)

2.1 | The site context

7The karstic site of Coudoulous is located on an edge of a plateau overlooking the Lot and Célé rivers. Discovered in 1966, the first fieldwork was led by J. Clottes and E. Bonifay (1978-1982), followed by J. Jaubert, J.-Ph. Brugal (1993-2003) (see Brugal 2006; Brugal, Jaubert 2021). Coudoulous II is a cave with a fossiliferous infilling dated between Marine Isotopic Stages (MIS) 6 to 3 (levels 9 to 3) (Costamagno 1999; Brugal 2000). The middle part of the sequence, levels 6 and 7, yields a diversified temperate fauna with very few lithic artefacts (n=70) and a charcoal horizon of natural origin (Théry-Parisot et al. 2008). These levels are dated to MIS5 a-d (ref. supra). The ungulate assemblage is dominated by abundant Cervids (Cervus elaphus, Capreolus capreolus, Dama sp.), and contains large bovines (Bos/Bison), caprines Capra ibex and suids Sus scrofa as well. The carnivores are few and represented by (in order of abundance): wolf Canis lupus, red fox Vulpes vulpes, cave lynx Lynx spelaea and some bones of badger Meles meles and cave hyena C. spelaea. There is few evidence of hominid actions on this faunal assemblage, mostly of natural origin: only about ten bones, mostly unidentified axial (vertebrae and ribs) elements (probably from herbivores, such as a thoracic vertebra of red deer) show discrete cut-marks (studies in progress). Within this sample, we found a notable and almost surprising exception of a cervical vertebra of a cave-hyena C. spelaea (1997-U19-lev.7a #30 / Excavation JPB). The origin of the fossil assemblage is interpreted as natural, related to a vertical opening (diaclase) functioning as a pit-fall. This site falls into the category of a deposit with low human occupation (or human-frequented deposit) (Brugal, Jaubert 2021).

2.2 | The fossil: identification

8In order to characterize the cervical vertebra found in Coudoulous II (level 7a), we conducted morphological and biometrical comparisons. We excluded the first two cervical vertebrae (atlas and axis), and we conducted these comparisons on the five last cervical vertebrae (C3, C4, C5, C6 and C7). We had the opportunity to observe, compare and measure cervical vertebrae of modern spotted hyena C. crocuta, material stored at the Osteological Dpt. of National Museum of Kenya at Nairobi (NMK, May 2024). The fossil specimen from Coudoulous can confidently be attributed to the genus Crocuta. We also based our analysis on the critical study made on hyenas by Ehrenberg (1940); we refer to both his plate (XX and XXI, p. 180-181) and table of measurements (op.cit. p.174) on cervical vertebrae (Halswirbels in German) in the genera Hyaena (brunnea, striata) and Crocuta. We illustrate vertebrae (fig. 1) and present dimension (fig. 2 and tabl. 1).

9The length of the cervical vertebrae decreases steadily from the third to the fifth, while their width increases in the same direction (Barone 1986 p. 161), and the body length of the sixth and the seventh are smaller. Following these criteria, our specimen corresponds to C3, C4 or C5, and most probably belongs to one of the first two. The Coudoulous remains is broken, and the posterior (caudal) disk is missing (unfused), and our specimen is clearly a young adult: however, we estimated the body length at ca.43 mm. The size is similar to the one of modern spotted hyena, especially C3. The Ehrenberg data show the same trend (decrease from C3 to C7), but the overall dimensions are greater. Our comparison allows us to refine the identification of the fossil specimen to a third (C3) or fourth (C4) cervical vertebra.

2.3 | The marks: cut and chop

10The specimen is broken: the body (centrum) is complete with the left part of the vertebra (transversal processes present, but upper one broken), while the entire right part (with articular and transversal processes) is missing. The overall break is relatively oblique with a dry fracture appearance. There is no caudal disk (not fused), while the cranial one is present and shows a line of epiphysis. The individual was a young adult.

11Two unambiguous cut-marks (a & b) are visible on the left cranial transverse process, on the upper surface (opposed to the articular facet) (fig. 3). These marks are short linear incisions [length: a) ca.1.84 and b) ca. 2.90 mm], very fine and parallel, separated by an interval of ca.2.50 mm. Besides, a small tooth score (called α on fig. 3) produced by a carnivore (small-size?) is present on the side opposite to where the cuts are located.

12There is abundant available research, usually based on microscopic approaches, on the identification of striations on bones, in particular those produced by lithic tools (cut- or scrap-marks) (e.g., Binford 1981; Potts, Shipman 1981; Shipman, Rose 1983; Bromage, Boyde 1984; Behrensmeyer et al. 1986; Cook 1986; Oliver 1989; White 1992; Fisher 1995; Fritz 1999; Garralda et al. 2005; to name but a few). The same is true for traces of impact (chop-mark). Several morphological criteria have been highlighted for butchery cut-marks, such as a specific cross-section and other associated traits (e.g. presence of barbs, shoulder-effect, parallel secondary or parasitic striae or microstriations, acute termination). Finally, location and orientation on bones are also important characteristics to consider. Recent works have developed new approaches using experimental and comparative studies. These sometimes involve sophisticated methods including scanning electron or confocal microscopy, profilometer and geometric morphometry, and 3D modelling, assisted by statistics, machine-learning and artificial intelligence (e.g., Domínguez-Rodrigo et al. 2009; Soulier, Costamagno 2017; Palomeque-González et al. 2017; Courtenay et al. 2019; Maté-González et al. 2023; Pobiner et al. 2023. and ref. herein, but see Courtenay et al. 2024).

13The morphology of the incisions on the hyena cervical vertebra from Coudoulous display features that are characteristic of anthropogenic cut-marks (fig. 3), with a relatively V-shaped cross-section, slightly dissymmetrical and deep. Parasite striae, shorter and parallel to the main one, are visible inside and especially outside the main groove, on the cut #a (see profile in fig. 3); their ends are sharp. The cut #b is slightly wider, with one microstriation inside, one steep edge and the other showing traces of material tearing along the ridge (serrated edge); this last modification is often seen when a retouched stone-tool is used (De Araujo Igreja 1999; Val et al. 2017). The cuts #a and b are parallel. The marks observed on our specimen are different from those created during trampling. These are generally more superficial, wide, sometimes numerous and randomly oriented (e.g. Domínguez-Rodrigo et al. 2009). Finally, the left side of the vertebra body shows a zone marked by a straight sub-perpendicular cut accompanied by tearing out of the periosteum. We interpret it as an impact blow, a smashing action made by humans with a stone in order to break the bone open (fig. 3).

14The hyena cervical vertebra from Coudoulous is the closed one to the axis, and the location of these actions (sensu anatomical position) vertebra and marks allow us to identify a dismembering action (blow), as well as a skinning action (cut-marks). An alternative hypothesis to skinning would be that the prehistoric people sought to recover the animal’s head. Such a hypothesis raises the question of the usefulness of this acquisition: to obtain elements such as teeth or soft organs, but also perhaps to demonstrate to their community a kind of control over a predator, without mentioning a possible trophy to display (see Brugal 2016 for this latter interpretation).

15During our morphometrical study of modern hyenas (NMK, Osteological Section), we noticed a vertebra with butchery marks on the cranial transverse apophysis (fig. 4) made by a metal knife: these marks were exactly at the same location as those observed on the fossil specimen. The other vertebrae showed no marks. We asked the staff at the Osteology laboratory the reason for these traces. They are due to the evisceration and dismemberment of a hyena carcass, whose skeleton is now in the collections. This explanation is similar to the one we are proposing here, according to which percussion marks have been added to continue the action and break the hyena’s neck.

3 | Other documented instances from the Paleolithic record

16The other sites that yielded evidence of hyena exploitation are presented in chronological (from oldest to youngest) and cultural order (tabl. 2).

Sala de los Huesos (Maltravieso Cave, Cáceres, Spain)

17The karstic site of Maltravieso is located in Extremadura, south of Spain. Interpreted as a hyena den, the cave yielded lithic artefacts (n=128) and a faunal assemblage (level 1) composed of (brown?) bear Ursus cf. arctos, cave hyena C. spelaea, Iberian lynx L. pardinus, red fox Vulpes vulpes, horse E. caballus and hydruntine E. hydruntinus, fallow deer Dama sp., aurochs Bos primigenius and wild boar Sus scrofa (Rodríguez Hidalgo 2008, 2010). The infilling is dated between isotopic stages 6/7 to 5, and the hyena specimens approximately dated around 120 ky (speleothems dated by U/Th).

18Anthropogenic action is reflected on a series of remains (n=25) represented by cut-marks and evidence of fracturing by hominids. The hyena is concerned with striae on two remains, namely an ulna and an innominate bone. The latter preserves the acetabulum, as well as part of ischium and ilium. Two incisions are located on the inner side of the ilium: they are interpreted as the result of dismembering action, as are the cut-marks observed on the ulna. Different scenarios are proposed (Rodríguez Hidalgo 2010 p. 76) involving either exokarstic and endokarstic activities (i.e., hyena carcass butchered by human outside the cave, then scavenged and bone transport inside the cave by hyena).

Taubach (SE Germany)

19The open-air travertine site of Taubach in Thuringia, not far from the city of Weimar, is known for its rich and diversified faunal assemblage that comprises more than 20 mammal species (Kahlke 1977; Heinrich 1994), with a dominance (by number of identified specimens or NISP) of rhinoceros Dicerorhinus mercki and brown bear U. arctos, followed by bison B. priscus, red deer C. elaphus, beaver Castor fiber, horse E. caballus, and straight-tusked elephant Paleoloxodon antiquus (Bratlund 1999 tabl. 3). The rhinoceros is especially abundant with almost 70% of juvenile individuals. The brown bear remains are also abundant and, as for the rhinoceros, these show ample evidence for human actions (i.e., cut-marked bones). The site is dated to MIS 5 (‘complexe Eemian’, with travertine dated by U/Th and biochronological data) and yielded a peculiar lithic assemblage characterized by its microlithism (Moncel 2003), named Taubachian. The cave hyena is represented only by one isolated upper tooth and, according to Bratlund (op.cit.), this remains shows cut-marks. Although no further precisions, nor interpretations, are provided about this specimen, Bratlund proposes the following about carnivores: “Since brown bear seem to have been killed by man on a regular basis, it is quite conceivable that other large carnivores would be hunted too. The few available fragments could, however, just as well originate from natural deaths accidentally incorporated into the assemblage”. (op. cit.: p. 93).

Navalmaíllo rock shelter (Pinilla del Valle, Spain)

20This rock shelter is one of the several archaeological sites located at Calvero de la Higuera in Central Spain, 55 km north of the city of Madrid (Baquedano et al. 2012). Navalmaíllo preserves a sequence of karstic and fluvial origin with archeo-paleontological layers F-D; the level F is dated by thermoluminescence (TL) to the end of MIS5a (77.2 ± 6.0 ky; 71.6 ± 5.0 ky) (Moclán et al. 2024). The fossil assemblage is highly fragmented associated with Mousterian lithic artefacts. The ungulates remains are mainly composed of large bovine Bos/Bison, red deer Cervus elaphus, rhinoceros Stephanorhinus hemitoechus, horse Equus caballus and some fallow deer Dama dama, hydruntin Equus hydruntinus, roe deer Capreolus capreolus, chamois Rupicapra pyrenaica (Huguet et al. 2010; Moclán et al. 2021) with few but diverse carnivore species (by order of abundance): cave hyena Crocuta crocuta, red fox Vulpes vulpes, wolf Canis lupus, brown bear Ursus arctos, cave lion Panthera spelaea, lynx Lynx sp. and polecat Mustela putorius. Herbivores represent ca. 94 % of the NISP and carnivore ca. 5 % of the NISP (calculation based on table 1 In: Moclán et al. 2021, which excludes the mesofauna). The site was interpreted as a Neanderthal short-term hunting camp, and human actions on bones are frequent on herbivores. Moreover, one hyena bone (phalanx) displays clear cut-marks (fig. 4b in Moclán et al. 2024), related to fur recovery. These marks are similar to those we observed on the phalanx from Arcy-sur-Cure (see below). Such skinning cut-marks are also present on bovine and equid phalanges, and the use-wear analysis on lithic artefacts confirms their implication in hide removal (Marquez et al. 2016 In: Moclán et al. 2024). Finally, the authors suggest that the use of animal skins (clothes or bed covers) by Neanderthals could be related to environmental stresses induced by cold contexts.

Les Pradelles (Marillac-le-Franc, Charente)

21Les Pradelles (also known as Marillac) is a collapsed karstic gallery with 7 meters of deposits (Maureille et al. 2010) that yielded lithic industry attributed to the Quina Mousterian, and diversified faunal assemblages (Costamagno et al. 2006). The sequence is bracketed between MIS5a to the beginning of MIS3 (dating by TL and U/Th, Frouin et al. 2017). The facies 2a, dated to MIS 4 (72-60 ka BP), provides numerous lithic artefacts and a faunal assemblage largely dominated by reindeer Rangifer tarandus, and few remains of bovines Bos/Bison, horse E. caballus, red deer C. elaphus, marmot Marmota sp., cave hyena C. spelaea, wolf C. lupus and red fox V. vulpes. An incised bone, collected from layer 10 (facies 2a), has been identified as a mesio-distal shaft fragment of a well-preserved left femur that belonged to an adult C. c. spelaea (Rohland et al. 2005; d’Errico et al. 2017 p. 4 fig. 1). The specimen is 53.3 mm length and displays nine regular, parallel, and deep anthropogenic incisions. A convincing interpretation of this series of incisions would denote the earliest evidence for a numerical notation by humans implying a cognitive, or symbolic behavior (and see discussion In: d’Errico et al. 2017; Hussain et al. 2022 p. 13). Moreover, several human bones retrieved from this site have been processed and bear cut-marks as well. Finally, alongside this incised hyena bone, numerous other bones have been modified both by humans (28.4%) and carnivores (3.7 %) (Maureille et al. 2010), suggesting alternating human (Neanderthal) and animal (hyena) occupations.

Les Rochers de Villeneuve (Lussac-les-Châteaux, Vienne)

22It is a relatively small cavity with karstic collapses, and three archaeological layers are recognized (Beauval, Morin 2010; Airvaux et al. 2012). Excavation mainly took place in the upper level J, where it demonstrated a pronounced occupation by hyenas; this layer can be considered as consistent with a hyena den. A limited set of lithic artifacts belong to a final Mousterian (with Levallois and Discoid débitage). The faunal spectrum is diversified with 11 herbivore and nine carnivore species, dominated by horse E. caballus, large bovid (Bison) and cave hyena C. spelaea (Beauval, Morin 2010 fig.7). Furthermore, a gnawed Neanderthal femur is present and dated to 45,200 ± 1,100 years BP (Beauval et al. 2005, 2006).

23The hyena actions on the bone assemblage are strong, with 45.1% of gnawing and/or digested bone, whereas the human modifications are discrete and account for only 1.3 % of the assemblage, mainly as cut-marked and burnt ungulates bones. In addition, one hyena tibia bears a single cut-mark (Beauval, Morin 2010 p. 186); this cut-mark was not described further and no interpretation regarding its purpose has been proposed.

Érd (Transdanubia, Hungaria)

24The Middle Paleolithic open-air site of Érd, discovered in 1961, yielded a rich lithic industry and bone remains from a six levels sequence, from two troughs or vallons. The tool-kit is attributed to a Quina type Mousterian (Charentian) and faunal assemblages are dominated by cave bear U. spelaeus with an average of 86% NISP (see the recent synthesis of Daschek and Mestera 2020). Other main vertebrate species (in decreasing order) are the horse E. caballus, wolly rhinoceros Coelodonta antiquitatis, cave hyena C. spelaea, wolf C. lupus, large bovid Bos/Bison, mammoth Mammuthus primigenius, hydruntine E. hydruntinus, giant deer Megaloceros giganteus and cave lion P. spelaea. Most of the herbivore species have a mixed origin, accumulated by both humans and hyenas, and resulting from predatory/scavenging activities.

25Few hyena bone elements show signs of anthropogenic origin, as reported by Daschek and Mestera from vallon I (2020 plate 4): a radius, one other long bone and one mandible (lingual face) have cut-marks. They are accompanied by other anthropogenic marks on other carnivores, notably on cave bear bones, and this exploitation of carnivores is interpreted as skinning, as well as perhaps consumption.

Bockstein & Hohle Fels (Swabian Jura, Germany)

26The famous Middle and early Upper Paleolithic cave sites from the two valleys (Lone, Ach) in the Swabian Jura have yielded a rich fossil record that documents carnivore exploitation by hominins. Karstic sites have long been used by human and non-human predators as places to live and breed (e.g., Brugal, Fosse 2004b). In a comprehensive article, Kitagawa et al. (2012) present the spectrum of carnivores exploited by prehistoric people, which includes bears, lions, wolves, foxes and hyenas from five German sites. All these large and middle-sized species – especially the cave bear U. spelaeus – are the subject of bone modification and exploitation, as well as representations such as the well-known ivory Lionman (Löwenmensch) from the Aurignacian in Hohlenstein-Stadel (see Conard 2003). However, only two sites preserve evidence for anthropogenic modification on hyenids (Crocuta crocuta spelaea): Bockstein, with one case from a Middle Paleolithic level (MIS 4-3), and Hohle Fels with one remains from the Gravettian level (end MIS 3, 30-27,000 years BP). Faunal assemblages are diversified at both sites: horse E. caballus and reindeer R. tarandus dominate at Bockstein (ca. 66% NISP), and carnivores include cave bears U. spelaeus, hyenas C. spelaea and wolf C. lupus; Hohle Fels has a similar faunal spectrum but, there, cave bears are the most abundant vertebrate species (ca. 63% NISP). At each site, the hyena is represented by single remains, with one tooth used as an item of personal ornament at Hohle Fels, and an unspecified element at Bockstein. The regular and frequent exploitation of Swabian sites by carnivores is remarkable and demonstrates the use of carnivores, and often cave bears, for cultural, symbolic and economic purposes.

Arcy-sur-Cure, Grotte du Renne (Yonne, France)

27The sites of the Cure valley in Burgundy are well known and have been the subject of numerous excavations, notably by A. Leroi-Gourhan in the late 1940s and early 1960s (Leroi-Gourhan 1964; Gaucher 1987; Girard 2018). One of the sites, la Grotte du Renne (the ‘reindeer cave’), in particular level X, contains a Chatelperronian industry associated with a faunal assemblage dominated by the reindeer R. tarandus, followed by the horse E. caballus, woolly mammoth M. primigenius and large bovid Bos/Bison for herbivores, and the cave bear U. spelaeus, cave hyena C. spelaea and wolf C. lupus for carnivores (David, Poulain 1990; David 2004). This cultural layer would be older than 40 ky. Cave hyenas in level Xc represent 3.4% on 1808 total NISP.

28As recalled by David (2004 p. 768) : « …dans le niveau Xc, les traces de découpe sont nombreuses sur les restes d’herbivores, et de plus, près de la moitié des phalanges de hyènes (11 sur 27) portent des traces de silex, marques d’un dépouillement de l’animal (David, Poulain 1990). » She further adds that the spatial distribution of these phalanges inside the space occupied by human would indicate that they were « …incluses dans les peaux posées au sol… ». They are all first phalanges, and the striae are located mainly on the anterior surface; there are numerous and very marked, deep, perpendicular cuts along the shaft (fig. 5). Lastly, two additional hyena specimens have been butchered: the distal end of a left humerus displays seven clear, deep or crossed cut-marks on its medial side, and a proximal end of a radius.

29The anatomical location of the marks, on distal limb bones, constitutes clear evidence for stripping and skinning actions (see experiment of Poplin 1972). The one on the humerus could be related to dismembering or even skinning. Besides, the fact that so many cuts were observed could indicate that the recovery of the pelt took place on an individual that had been dead for some time, and partially desiccated. Such explanation suggests a probable passive acquisition of the animals, unless the butchery action performed on an individual killed by hominins but not immediately skinned. In addition, the root of one hyena canine has been incised and fitted as a pendant (Vanhaeren et al. 2019 fig. 3); such teeth adornments are numerous and made on various mammals at this site.

Les Cottés (St. Pierre-de-Maillé, Vienne)

30The cave site of Les Cottés is located on the south-western margin of the Paris Basin and contains a sequence spanning the Middle to Upper Paleolithic transition (Mousterian Quina to Gravettian) (Soressi et al. 2010; Talamo et al. 2012). The Chatelperronian level (Unit 06) is dated to MIS3 and yielded a faunal assemblage largely dominated by the reindeer R. tarandus, with some remains of large bovid Bos/Bison, ibex Capra ibex, horse E. caballus, wild boar S. scrofa, woolly rhinoceros C. antiquitatis, woolly mammoth M. primigenius, cave hyena C. spelaea and wolf C. lupus, morphologically identified and completed by ZooMS analysis (Welker et al. 2015 tabl. 1). One hind limb element (femur) of hyena bears cut-marks (W. Rendu, pers. comm.), as well as the palmar surface of a hyena phalanx interpreted as skinning (Rendu et al. 2019).

Bacho Kiro (Dryanovo, Province of Gabrovo, Bulgaria)

31The cave site of Bacho Kiro contributes to date first arrival of H. sapiens into Europe to ca. 45.000 years ago, during the Initial Upper Paleolithic (IUP) (Hublin et al. 2020). In layer I, a cave hyena phalanx has cut-marks interpreted as disarticulation by Smith et al. (2021 p. 12, fig. 10). In IUP occupations, the carnivore presence seems to increase with remains showing human modifications, especially on cave bears U. spelaeus that died naturally in the cavity, but also on wolf C. lupus, red fox V. vulpes and one cave hyena C. spelaea. The authors recognize such exploitation to elements used for pendants, as well as for skins and furs.

Pair-non-Pair (Prignac-et-Marcamps, Gironde)

32Discovered in 1881 by F. Daleau, the cave is best known for its decorated walls (engravings). The cave was excavated in the XIX^th century par F. Daleau (during 32 years: 1881-1913) and yielded different Paleolithic techno-complexes: Mousterian, Chatelperronian, Aurignacian and Gravettian (Cheynier 1963 not consulted), and information about the cave is provided in Moisan (1994). The infilling contains almost 30 mammal species and 25 bird species, but the lithic industry is often mixed and lacks stratigraphic data. The entire Daleau collection is stored at the Musée d’Histoire Naturelle de Bordeaux. During the establishment of an exhibition on rare game (see Fosse et al. 2019) at the Musée national de Préhistoire, Les Eyzies (Dordogne), the re-examination of the Daleau collection led to the discovery of one hyena specimen with anthropogenic identification by S. Madelaine.

33This element is a lower canine of Crocuta spelaea labelled “PNP-D’.8.8.92 60.1046.127” (fig. 6): it comes from Pair-non-Pair (PNP), level D’ and was found in July 1892 (the other numbers are museum inventory). The level D’ may be attributed to the Aurignacian, and the species associated comprise (in order of importance, tabl. II In: Moisan 1994): the horse E. caballus, a large bovid Bos/Bison, the woolly rhinoceros C. antiquitatis, woolly mammoth M. primigenius, reindeer R. tarandus, cave hyena C. spelaea, cave bear U. spelaeus, deer C. elaphus and Megaceros M. giganteus. The hyena canine displays sign of root piercing consistent with the early steps in the manufacture of a tooth pendant. Both sides of the base of the root show clear evidence of digging, scraping and sawing. However, the work was never completed and the pendant is not functional.

4 | Discussion

4.1 | Spatiotemporal distribution

34Unlike other carnivores, particularly ursids, and to a lesser extent canids and felids, hyenids were rarely exploited in the Paleolithic. Globally speaking, in European Paleolithic sites showing hominid activities on carnivores (314 cases recorded so far), the ursids represent ca.41 %, felids ca.26%, canids ca.27%, hyenids ca.3.2% and mustelids ca.2% (Fosse, Brugal in prep). This contribution stemmed from our observations of cave hyena bones from the sites of Coudoulous II (Lot) and La Grotte du Renne at Arcy-sur-Cure (Yonne) bearing human-made striations. We then extended such cases to other Paleolithic sites from the literature, in order to propose an exhaustive inventory (tabl. 2). Only thirteen sites have yielded evidence of human actions on cave spotted hyenas (fig. 6). Even though some cases are only partially documented (e.g., no information on the modified skeletal elements, no photos, no details about the morphology of the striae), we have chosen to include them in this review. For instance, we have reservations about the cut-marked hyena remains from the sites of Taubach, Rochers de Villeneuve and Bockstein (no photos or descriptive study). However, for Taubach, there are many elements with cuts (34,8% NISP), which belong to all taxa, including herbivores and carnivores (as bear, wolf, cave lion) (Bratlund 1999 tabl. 3).

35With the exception of Taubach, anthropogenic actions on hyenas are mainly documented from cave sites attributed to final Middle Paleolithic (MP) and Early Upper Paleolithic (EUP) industries, and therefore contemporaneous to Neanderthals and early modern humans. These cave sites are mainly found in southern Europe (n=8, mostly France and Spain), with three cases in North (Germany) and two from Central Europe (Hungary, Bulgaria). At this point, it can be reminded that many Paleolithic fossil assemblages are palimpsests or amalgam produced by the alternating activity of humans and carnivores. This means that faunal assemblages may be mixed and include animal remains resulting from human consumption of their prey, as well as from non-human carnivore consumption of herbivores, and even from carnivores dying inside the caves.

36The chronological period of interest goes from MIS 5 to MIS 3. This time period is marked by significant cultural and anthropological transitions, but also by major climatic and biological changes. The number of sites with fossil remains of the cave hyena Crocuta spealea seems to characterize this period, especially between 60 and 40 ky (‘Hyena event’ Discamps 2009, 2011; Hussain et al. 2022 fig. 2). Their frequency seems to decline quickly thereafter, around 40 ka in central Europe and ca.31 ka in north-east and southern Europe (Stuart, Lister 2014; Krajcarz et al. 2023), but with possible refuge areas in northern Europe (Belgium? United Kingdom) until ca. 21 ka (Bamborough 2020). Regardless of the suggested survival time of the cave hyena in Europe, there are no populations known post-28 ka (i.e. important reproductive rate / denning activities) according to both direct radiocarbon dating on bones (and not on coprolites) and from sample counting’s (MNIs).

37Many European cave sites with hyena remains constitute an important record of dens. This is the background to most of our fossil record, sometimes associated with cave bear hibernating sites (Maltravieso, Érd, Pair-non-Pair?, Rochers de Villeneuve, Bockstein, Hohle Fels). Other sites have a natural origin to their bone accumulation processes (Coudoulous II, Taubach?), but most of them are archaeological sites (Les Pradelles, Arcy-sur-Cure/grotte du Renne, Bacho Kiro, Les Cottés) associated with lithic assemblages. Finally, in most of these sites occupied by different carnivores, the human exploitation of bears is often noted, and is generally more pronounced than that of hyenas or other carnivores.

38In fact, regardless of their nature and origin, all the sites considered here have produced lithic artefacts in varying quantities, even in the case of the sites strictly considered as dens. This later case, dens with small lithic series, have been the subject of various interpretations and debates about the possibility of humans using and scavenging animal remains (e.g. Brugal, Jaubert 1991 (2024); Villa, Soressi 2000; Brugal et al. 2006; Discamps et al. 2012; Hussain et al. 2022). Mixed sites used both by non-human predators and by humans are very common in the Middle Paleolithic and early Late Paleolithic periods, and they correspond to combined fossil sets. This raises the question of inter-specific competition for space and food related to the periodicity and frequency of use of the cave by these different predators over time (see Brugal, Fosse 2004b; Enloe 2012; Blasco, Rosell 2009). Such competition can be considered as a human-carnivore conflict, as shown by the presence of human remains in caves used as dens or lairs (e.g., Diez Fernandez-Lomana et al. 2010), without ruling out competition between different species of carnivores visiting or living in the same karstic space. Such a context favors both confrontations and the possibility of scavenging dead animals or recovering skeletal elements.

4.2 | Zooarchaeological considerations

39Evidence of anthropogenic action is always limited and concerns extremely scarce hyenid bones and teeth, often a single specimen per site. These cases raise three questions: i) why have hyenas, unlike others carnivorous, been so rarely exploited and depicted, ii) what is the origin of these acquisitions and iii) what is the purpose of these exploitations?

40It is difficult to answer the first question, and any answer will probably have to remains hypothetical. To do so, we need to look at the modern perception of hyenas by traditional human groups (cf. IV.3), as well as the use of this animal at the utilitarian and symbolic levels, popular beliefs, medical base and superstition (e.g., Moures-Nouri et al. 2023). We can mention here for instance that H. Kruuk (1972, 2002), in his studies on carnivores, mentions a negative perception of African people for hyenids, especially for the spotted hyena, which clearly contrasts with their vision of the lion, for example. This could be the case in the Paleolithic, amplified by the density of their population and their similar behavior (as their occupation in caves). Hyenas were probably a real danger to humans.

41The question of their acquisition by Paleolithic humans seems straightforward, particularly considering the context of the sites (dens, mixed-occupations/hybrid-assemblages). Without totally ruling out the possibility of killing (in a confrontational way), the bone and dental elements used could come from animals that died naturally. It would then be a matter of recovering elements from hyena bodies, which brings us to the third question, about the purpose of such recovery.

42Different elements are concerned: skull (one case), tooth (three cases), axial skeleton (two cases), long bones (six cases) and short distal leg bones (three cases). The teeth clearly have an ornamental purpose, without ruling out a link with possible ritual function (and who was wearing these hyena teeth, considering gender or age…?). In most cases (70 %), the main goal was to recover the skin, as fur was presumably a product desired by hunter-gatherer groups, and marks on the phalanges testify directly to such practices. The presence of cut-marks on long bones such as the femur, humerus, tibia or radius could be also related to skinning processes, but again we cannot rule out the possibility of meat consumption (as suggested from the Maltravieso specimen, Rodríguez Hidalgo 2010), because meat and marrow from carnivores could be edible; especially the bears (and dogs: D’Iatchenko, David 2011). However, this raises questions about the mortality risk due to pathogens (Dutour 2018) and presence of parasites, notably frequent in a scavenging acquisition (Moleòn et al. 2017; and see discussion In: Moclán et al. 2004) or hunting (Ducrocq et al. 2020).

43Finally, it is interesting to note that the cut-marks on hyena phalanges are reminiscent of those on herbivore phalanges, so nicely illustrated by H. Martin (1907-1910) at the Mousterian site of La Quina (ex. p. 226 & Pl. XLIV for skinning). The cuts are numerous and very deep, as the skin adheres strongly on these distal limb parts, especially when the carcass is already dry.

44In Western societies, cultural prohibitions alone, the vision of impurity or cultural taboo keep us from consuming meat of carnivores, although there are many examples of this type of consumption elsewhere in the world (e.g., dogs consumed by the Chinese today; medium-size carnivores: Doughty et al. 2014; polar bears consumed by Inuit: Ducrocq et al. 2020, and other bears: Hallowell 1926; Dyrenkova 1930). Finally, one special case is the ‘notation or cultural exaptation’ visible on a shaft of a hyenid long bone from Les Pradelles (d’Errico et al. 2017). Even though the authors propose a detailed analytical study of these incisions, they do not provide further explanation regarding why these are were made on «… the wrong species and in the wrong place…» (op.cit. p. 7). There might have been no deliberate choice to modify a hyena bone in order to produce a symbolic object and, in our opinion, this may have been more a matter of chance.

4.3 | A short eco-ethnographic approach

45We thought it would be interesting here to propose a short section about eco-ethnology, how hyenas are viewed, and what products s.l. (skin, teeth, bone, organs…) have been used in human societies. Wild carnivores are both attractive and repulsive to human groups, as they engage in potential interspecific competition (as meat eaters) as part of a trophic relationship.

46The family Hyaenidae is represented by four genera in Africa: spotted (Crocuta), striped (Hyaena), brown (Parahyaena), and aardwolf (Proteles) (in Kingdon, Hoffmann 2013). The first one is the most abundant and widespread, as in the Pleistocene record in Europe. It is a top predator in Africa, alongside the lion and the leopard. Spotted hyenas are highly efficient hunters; they live and hunt large ungulates in packs, often in a cooperative way (like wolves); clans can comprise up to 80 members at any given time, and these are dominated by the larger females (Kruuk 1972, 2002; Mills, Hofer 1998). Spotted hyenas are extremely social carnivores; they occupy territories that they defend, and they are notorious man-killers. Sexual dimorphism favors females, which are larger and have a distinctive morphological appearance, with an elongated clitoris similar in shape to the male penis. As a result, it is difficult to differentiate between both sexes. All these eco-behavioral characteristics make hyenas special animals in contrast to other large carnivores, and they certainly must have played a role in the vision and relationship with human groups, even in the Paleolithic period.

47Hyenas, as some other carnivores, provide direct, material benefits in recent society. We can list their use for fur (clothing, tent…), food, medicine, perfumes, ritual, religious and symbolic aspects, and even tourism in the modern world. Many examples of their use and relationship with human societies are described in Africa (Hailemariam, Mekonen 2021; Kruuk 2002; Malawi: Morris 2000; Namibia: Rust 2015; Algeria: Derouiche et al. 2020; Ourdani Selmoun pers. comm 2024; Nigeria: Chibueze Izah, Seiyaboh 2018), the Near and Middle East (Quarqaz et al. 2004; Handa et al. 2021), and in Asia (Frembergen 1998; Mahawar, Jaroli 2008).

48Through this literature, we can account for the importance and variability of the use (sensu utilitarian) of hyena as a raw material (skin, bones and teeth, organs, blood and urine…), as well as their values in the zootherapeutic and mythological realms. Finally, let us quote Morris (2000 p. 76) «The hyena is viewed by Malawians not only to be associated with witches, but also to be a highly sexual creature », to add a gender dimension. While these ethnozoological examples are diverse in their use of animal products and remains, it is difficult to apply them directly to the Paleolithic period.

Conclusion

49Our review of hyenas (cave hyena C. c. spelaea) bone and dental objects used by Paleolithic human groups in Europe highlights that such practices are limited, with only thirteen sites providing anthropogenically-modified fossil remains. Such objects come from sites considered to be cave hyena dens or cave bear lairs, and sometimes from archaeological sites. They are often bones, and in three cases, teeth were used as pendants. The main objective was certainly to recover the skin, but also to produce adornment, without totally ruling out the consumption of meat, which remains debatable (but the case of bears probably needs to be reinvested, see the instance of Taubach here, and Romandini et al. 2018; Toniato et al. 2024).

50There is a clear discrepancy between human exploitation of cave hyenas and that of other large carnivores, in particular bears, and to a lesser degree felines (especially lions) and wolves, which seems observed from the end of the Middle Paleolithic to the early Upper Paleolithic. The use of smaller carnivores (red fox, badger, etc.) becomes more developed in the Late Upper Paleolithic. It is also interesting to note that there is not a single representation of hyenas in the Upper Paleolithic art (contra Spassov, Stoytechev 2004), unlike other large predators (cave bear, cave lion, or wolf). The status of the hyena in hunter-gatherer societies is therefore clearly different from that of these other predators. In fact, the social organization, behavior and morphology of the spotted hyena are also singular and specific. It can be assumed that this species was considered dangerous and unattractive and carried a negative vision, which could not induce real cognitive or symbolic values.

51The culture-nature relationship however, especially with carnivores, is undoubtedly a more complex phenomenon, and suggests various types of exaptation process. In the contribution of Hussain et al. (2022), the authors analyze the ecocultural relationships between hominins, hyenas and cave bears, using a multispecies/trophic approach. Beyond aspects of antagonism, conflict and competition, they discuss the “…co-adaptive interfaces between hominids and their contemporary hyena neighbors in changing spatiotemporal contexts.” (op. cit. p.10). They define three types of ‘Neanderthal sites’ (op. cit. tabl. 2, p. 12) concerning the human-carnivore relationship, based on the degree of combination and abundance of materials, cultural vs natural (=biological). Incidentally, their third type corresponds to the one we had already highlighted from den contexts (Brugal, Jaubert 1991 (2024)). They mention human’s visits to the dens that they framed within an ‘ecocultural and materiality negotiations’ (op.cit. p. 21) reflecting ideas about sharing the landscape with carnivores, as well as implying a kind of selection of place (i.e. site/cave) by hominids.

52While evidence of a strong relationship between humans and bears in the Paleolithic is frequent, this is not at all the case for the hyena, which represents a species of little « interest » in terms of by-product exploitation, in spite of a strong entanglement with hominins. The large carnivores continued to be used throughout the Upper Paleolithic, while smaller predators (red fox, badger, etc.) were also routinely exploited during that period.

53Finally, the comparison of zooarcheological and ethnographic data is an interesting approach, as it sheds light on human-carnivore interactions and proposes innovative ideas and new avenues of research that were not considered previously (e.g., Plug and Badenhorst 2010: bone use as divining object). In the case of the hyena, in addition to showing the various facets of the use and apprehension of this animal, it remains hypothetical to apply ethnographic information onto the Paleolithic world.

Acknowledgments

54Many thanks to C. Beauval (ArcheoSphere), S. Madelaine (MNP), V. Mistrot (Musée d’Aquitaine, Bordeaux) to provide information about sites and materials, and to T. Guiragossian (TG) for macrophotos of Coudoulous (made with Hirox). The faunal collections from Coudoulous II are stored at UMR 7269 LAMPEA, Aix-en-Provence, those from Arcy-sur-Cure, Grotte du Renne at the Musée National de Préhistoire, Les Eyzies and those from Pair-non-Pair at the Musée d’Histoire Naturelle de Bordeaux. J.Ph.B. wants to thank K. Selmoun-Ourdani (Mouloud Mammeri University, Algeria) and O. Mwebi (NMK, Nairobi) for their help, and exchange about the question of use of hyenas by traditional communities. We warmly thank two anonymous reviewers who greatly help to improve a first draft of our paper, as well to the great editing work by Drs. A. Val & P. Goldberg.

55We wish to pay tribute to Francine David with this contribution, not only for her pioneering zooarcheological work in Arcy-sur-Cure, but beyond (as Pincevent, Mauran), by recognizing the importance of her work that she did not hesitate to share, with her kindness and great humility. Chapeau bas…