Neanderthals and humans interbred in Europe for 5,400 years

Eagle talons rarely occur at European prehistoric sites and at no Neandertal site have eight been found [18]. Previously, eagle claws have been described as isolated talons at Mousterian sites stretching from about 100,000 kyr at Pech de l'Azé (FR) [6] to older than 44 kyr at Fumane cave (IT) [7]. Other examples of white-tailed eagle bones appear in European Upper Paleolithic and later sites, but they are rare and seldom involve more than one element of the foot. Presence of eight Krapina talons, four showing cut marks, suggests they were disarticulated by cutting the tendons, curated and lost as a unit, probably as a necklace or some other kind of jewelry.

The sandstone rock shelter in NW Croatia was excavated under the direction of Gorjanović-Kramberger between 1899–1905 [19–20]. Today nothing remains of the site except the eroded, sandstone cliff face. But Gorjanović-Kramberger extensively documented the site in his publications and based on the fauna estimated it to be from a warm, interglacial period, now dated to 130,000 years ago or late MIS 5e by ESR and uranium series [21]. According to Miracle [22] this age is concordant with cave bear (Ursus spelaeus) size and morphology. From the earliest to latest deposits the Krapina male cave bears are all small, significantly reduced in size compared to glacial sites in the region with cave bear remains. From this, Miracle suggests the strata at the site were deposited over a warm, short time period, <10,000 years [22]: “The small size of the Krapina male cave bears is best understood as an adaptation to a full interglacial climate … Krapina’s stratigraphy extended neither into later substages of the last interglacial (e.g. MIS 5a-d) nor into the last glacial MIS 3–4” (p. 85).



Only evidence of Neandertals is found at Krapina, confirmed by the appearance of Mousterian tools [23] and human remains [19–20]. Gorjanović-Kramberger and his assistant Osterman collected hundreds of Neandertal bones and teeth, more than 800 stone tools and almost 2800 animal remains [19–20, 24–25]. Everything at Krapina derived from excavations in stratified levels (S1 Fig.) and has been curated in the Croatian Natural History Museum (Zagreb) for more than a century. Animal bones as represented by minimum number of individuals (MNI), consist primarily of large mammals, especially rhinos (21.7%), bears (17.0%), and bison/bos (7.5%), while other species like pigs, deer, and small carnivores make up a lesser extent of the fauna [25]. Beavers (Castor fiber) are the exception, mostly coming from the fluvial base of the shelter and constitute 17.9% of the bones from site [25]. A full listing of mammalian fauna includes 25 taxa and of these 13.6% of show evidence of burning and another 4.1% preserve stone tool cut marks [25]. In addition to the mammals, Gorjanović-Kramberger recorded a turtle (Testudo) humerus, fresh water mussel shells of at least three taxa, several taxa of different land snail species and a few bird remains [19–20, 25]. Except for some of the bird bones, none of the non-mammalian bones or shells shows signs of human manipulation.



The most striking Haliaëtus albicilla claw is Krapina 385.1, a right second talon bearing six cut marks and surface modifications. The three largest cut marks occur across the lateral side of the proximal surface, just distal to the joint with the phalanx. These marks are clearly visible in Lambrecht’s 1915 description (S2 Fig.), but were not recognized by him or Gorjanović-Kramberger even though the latter extensively documented cut marks on human and animal bones [19]. The uppermost cut mark (Fig. 1a) is the shortest (3.2mm) and is a wide, V-shaped cut. The longest cut mark begins at the joint surface where it is 3.7mm long, breaks as it extends across the foramen, then continues 4.1mm onto the projection (Fig. 1b). A third cut mark measures 5.7mm and is angled about 20° to the long cut mark (Fig. 1c). All these cut marks show edge smoothing. On the dorsal surface at the tuberculum extensorium, just perpendicular to the joint is a short (~1.0mm) cut mark. Two slightly curved, parallel cut marks (2.0mm and 1.3mm) appear about 3mm distal from the edge of the articular surface (not shown). Small areas of surface flattening and abrasion appear on the lateral, proximal margin (Fig. 1d) and more polishing occurs around the elevated surface lateral and superior to the vascular foramen.



Krapina 385.4 is a virtually intact left third talon. On the dorsal edge of the proximal joint surface are two short cut marks, paralleling each other (Fig. 2a). Lateral to these is a broad, V-shaped wedge at the border of the articular surface. The dorsal-most portion shows some polishing and compaction of this surface, which gives it an ‘ivory’ appearance (Fig. 2a-b). The inferior-most portion of this facet shows a spalled-off facet with a hinge fracture, which lacks polish. On the medial side there are two small, abraded areas, which have eroded the elevated morphology (not shown). Neither is as compact nor as shiny as the lateral facet.



Krapina 385.5 (Fig. 3) is a right talon 1, which shows a natural (excavation?) break across the base of the tuberculum flexorium major on the lateral, proximal surface. This is the only talon with large amounts of adhering cave sediment. There may be additional marks under this sediment, but well-defined cut marks occur across the ‘neck’ of the talon. Four cut marks appear on the lateral side, distal to the tuberculum flexorium. These short cut marks, off the articular surface, parallel each other and range in length from 1.3–1.6mm. The proximal is the deepest and widest compared to the others. Compared to other specimens, these cut marks show little edge smoothing.



A right second talon (386.1) shows cut marks along the medial articular margin (Fig. 4). Measuring 2.4mm–3.5mm, the incisions roughly parallel each other, pierce the articular surface, have some preserved cave sediment and show marginal smoothing. Like 385.3 there is a long (3.2mm) burnished facet on the lateral side, ending at the tip of the talon, which is polished (not shown). There is abrasion on the lateral aspect of the blade with a facet covered in its proximal extent by lacquer (not shown). It measures (3.4mm long with a maximum height of 1.4mm). Distal to this is a long narrow facet (3.4mm long with a maximum height of. 6mm), which narrows to the tip.



Krapina 386.18 is an intact third phalanx of digit 3. It was originally assigned to Aquila [30], then to Bubo bubo [31]. Our comparisons with equivalent phalanges of modern Bubo bubo and Haliaëtus albicilla indicate it is a white-tailed eagle. The phalanx is light in color and similar in preservation to the other Krapina eagle talons. At least 21 cut marks appear on the lateral and dorsal surfaces (Fig. 5a-c). The most prominent wraps from the lateral to dorsal face at the proximal end. About 5.0mm long, the incision deeply penetrates the cortical bone and the interior of the mark is stained with cave sediment. It is flanked proximally by another distinct mark 4.0mm long, which curves from the plantar to the dorsal edge. Distal to this cut mark are a series of very shallow, faint incisions on the lateral edge. These range in length from 1.6–2.1mm. Continuing more distal, shorter and deeper cut marks are found proximal to the condyle and another two appear on the lateral wall of the condyle. All measure about 2.5mm. Dorsally, cut marks are only found on the distal region across the bone separating the two dorsal fossae. In this location at least six marks pierce the bone, varying in length, width and depth of penetration into the bone. The longest (2.1mm) and widest (.3mm) is at the medial end and is flanked by a distinctive, more proximal cut mark. In these, there is some smoothing of the cut mark edges. The remaining marks are fainter, but have sharper edges. There are no cut marks on the plantar or medial surfaces (S7 Fig.). This terminal phalanx articulates with the left talon 3 (385.4), so the two bones are likely from the same individual (Fig. 6).

The Krapina human remains found lower in the sequence sometimes have cut marks, but these mainly are sharp-edged and seldom show signs of edge smoothing [19]. The hominid collection is extensively fragmented [24], unlike the eagle talons, which are essentially intact. None of the human remains have densely polished facets or abraded areas like on the talons, but a few the patellae show arthritic, ebrunated surfaces, the result of cartilage breakdown and arthritis. This would not be a likely cause of the polished facets on the talons, but suggests some kind of prolonged contact with another hard surface.

Miracle has extensively documented taphonomic features of the vertebrate fauna and notes that “modification of bone surfaces is extremely rare … [with] only 6.6% of the vertebrate fauna showing evidence of weathering,” p. 236 [25]. Based on his observations and our descriptions of the cut marks and other damages to the talons, there is no evidence that diagenetic factors could have produced the modifications on the eagle talons. We have also surveyed the nonhuman vertebrate faunal remains from levels 8–9 and found no evidence for similar bone modifications, other than the occurrence of cut marks. Review of faunal elements from lower levels at Krapina show no surface modifications like found on the eagle talons. Finally, there are a few other bird remains from Krapina (S1 Table) and, while broken, none show cut marks or other indications of human manipulation.

Virtually all of the talons show some evidence of cut mark edge smoothing and surface abrasion at various places on the talon and three have nicks in the medial and/or lateral blade margin. (1) Cut mark smoothing. Compared to other eagle talons documented from Mousterian sites, like Mandrin cave where the cut marks have sharp edges [27], most cut mark edges of the Krapina talons are smoothed (Fig. 7a). Many of the marks are located in positions, which have been reproduced in experimental butchering [27], but subsequent, localized edge alterations appear in the Krapina specimens. Other parts of the same talon show no surface modifications, indicating that the edge wear was focused in its occurrence. (2) Abrasion. All eight talons exhibit shiny surfaces, ranging from small areas of dense, highly polished areas on the proximal ends to longer, shiny areas on the distal talon blades (Fig. 7b). They are always localized, never appear on the plantar aspects of the talons and seem to be the result of repetitive contact with another object. They could not have been produced in the living white-tailed eagle. (3) Nicks. In three of the largest talons (385.1, 385.3, 386.1) small nicks interrupt the sharp edge of the blade’s medial and/or lateral margins. Sometimes these appear on both margins (385.3 and 386.1) or on only the medial margin (385.1) (Fig. 7c). They are located about 10mm distal to the articular margin. We cannot rule out taphonomic causes for these indentations into the sharp medial and lateral talon edges, but they are limited in occurrence and are never found elsewhere on the talons or on any of the smaller talons. In contrast the other five talons clearly show sharp, uninterrupted edges all along the blade margins.

Conclusions

Eagle talons are rare at other Neandertal localities and no sites have yielded eight talons from white-tailed eagles or any other raptor. Since three-four different eagles are represented, they must have been acquired in separate events and were preserved as a unit before they were lost in the sediments. Others have noted [6–7, 27, 41–43] that raptor bones found in late Pleistocene sites signal some kind of symbolic activity. At Krapina, cut marks on the pedal phalanx and talons are not related to feather removal or subsistence, so these must be the result of severing tendons for talon acquisition. Further evidence for combining these in jewelry is edge smoothing of the cut marks, the small polished facets, medial/lateral sheen and nicks on some specimens. All are a likely manifestation of the separating the bones from the foot and the attachment of the talons to a string or sinew. Cut marks on many aspects, but not the plantar surfaces, illustrate the numerous approaches the Neandertals had for severing the bones and mounting them into a piece of jewelry.

As in ethnohistoric-present societies, the Neandertals’ practice of catching eagles very likely involved planning and ceremony [35]. We cannot know the way they were captured, but if collected from carcasses it must have taken keen eyes to locate the dead birds as rare as they were in the prehistoric avifauna. We suspect that the collection of talons from at least three different white-tailed eagles mitigates against recovering carcasses in the field, but more likely represents evidence for live capture. In any case, these talons provide multiple new lines of evidence for Neandertals’ abilities and cultural sophistication. They are the earliest evidence for jewelry in the European fossil record and demonstrate that Neandertals possessed a symbolic culture long before more modern human forms arrived in Europe.

DOI: 10.1371/journal.pone.0119802

The first ‘Neanderthal cave bear bone flute’ from the Middle Palaeolithic was believed to have been discovered in the 1920s from Potočka Zijalka Jama Cave (i.e. Potok Cave), Slovenia [1]. This was a larger cave bear den (cf. [2,3]) and Late Palaeolithic Aurignacian (not Neanderthal) used rock shelter camp site at the entrance (cf. [1]; figure 1). In this cave, cave bear hunts by Cro-Magnon humans seem to be indicated on a cave bear shoulder blade pathology (large diagonal impact hole, partly healed diagonal hole) that seems to have been made by a probable Mladeč projectile bone point [5].


Figure 1.

Other cave bear cub femora with holes (‘bone flutes’) were then reported from the Istállóskö Cave, Hungaria (cf. [6]). This was a smaller cave bear and Ice Age spotted hyena (Crocuta crocuta spelaea) carnivore den which overlaps with another Aurignacian camp site, but again, it has no Middle Palaeolithic Neanderthal occupation signs (cf. [7]). Brodar [8] reported cave bear cub femora and other cave bear bones ‘with holes’ as further proof of the ‘oldest instruments in the world’ from the Mokriška Jama Cave (or Medvedja Jama Cave=Bear Cave), Slovenia. Also, this is a large cave bear den which had again an Aurignacian camp site at the entrance, and again no Neanderthal occupation at all (cf. [9,10]).


Figure 2.

All aforementioned femora and other cave bear bones with ‘holes’ (i.e. ribs, humeri and jaws) were compiled and studied by the ethnologist/musician Omerzel-Telep [11], without any natural science, nor palaeozoology background, especially the important ecology of cave bears and their predators/scavengers, non-human top predators of the Ice Age and the wide distribution of cave bear den caves in Europe (cf. [3,12–21]; figures 1 and 2), where always large amounts of damaged and also punctured cave bear bones are present, such as figured with many new examples herein for the northern German Weiße Kuhle Cave and other cave bear dens (figures 3–7).


Figure 3.

Another juvenile bear cub femur with holes from Divje Babe I Cave, Slovenia, a small cave bear den (cf. [25]; figure 5(4)), where also Neanderthal Mousterian layers were believed to be present [26], was declared twice incorrectly as the ‘oldest instrument’, a 43 140 BP old ‘Neanderthal flute’ from layer 8 [26,27] (figure 5(4)). This was already contradictory to the results of the archaeological inventory that is well acceptably declared to be solely of, again, Cro-Magnon human Late Palaeolithic origin, and not of Mousterian (cf. [28]). The Aurignacian lithic material appears also together with cave bear remains [25]. Therefore, there is no evidence for a Neanderthal (Mousterian) context and the cave bear remains, which even occur in several older and younger Late Pleistocene layers (cf. [25]). The only absolute date was made solely on a cave bear bone, the ‘bone flute’, whose age would date into the Neanderthal or ‘cave bear den’ time period. This report of a ‘cave bear femur bone flute’ was not the ‘oldest’, neither historically, nor by stratigraphy. The bone's holes on the dorsal side appear not to line up, whereas on the ventral side another hole was declared as the ‘thumb hole’. The studies even thrilled up to ‘exact musical studies’ [29]. Fink [30,31] declared then to the top of this, without natural scientific studies, that the hole spacing matched a ‘diatonic scale sequence, among the most widespread scales known’—which underlines, also contradictory, that this is not of human origin. Ethnologist/musicians created then a wave of ‘cave bear bone instruments’ based solely on ‘holes in bones’ (compiled in [11]), from all kinds of carnivore punctured cave bear bones, even other than femora. Other authors doubted the ‘flute’ and human origin however (e.g. [32–38]) or were fighting for pro-arguments (e.g. [39,40]). At least, very correctly, the ‘holes’ were mostly discussed to be of ‘carnivore chewing damage’ origin (cf. [32,33,37,41]), whereas X-ray scans did not prove any ‘drill-scratches around the holes’ or any marks of stone tools on the bones, and left again the question of the hole origin open (cf. [42]). The exact carnivore was never estimated, even by newer and fully controversial studies by Turk et al. [24] that lack carnivore ecology knowledge, especially in tooth and jaw function of top predators. Ignoring the top predator bone damage on Ice Age animal bones, again the pseudo-bone flute was not only ‘confirmed’, even more bone flute finds were added by the same Slovenian author (cf. [43]), who misidentified: (a) the site occupation by Neanderthals, as those of Aurignacians [28], (b) the bone, by rotating it upside down (see [44]), the 180° rotation of which is corrected herein (figure 5a), (c) the general bone taphonomy of cave bear bones, and (d) carnivore jaw functions, especially hyenas, correctly presented herein (figure 2).


Figure 4.

In this contribution, not only sole carnivore damage can be demonstrated on all previously published ‘pseudo-bone flutes’, which were already revised in some cases [4,16] (figure 2). Herein, even more of such cave bear bones with holes can be added with focus only on the femora (figures 5–7), from German and Romanian cave bear den sites (therefore not limited to Slovenia at all, as mentioned by Turk et al. [24]; see figures 1, 5–7 and table 1). Their producer, a large carnivore, and the main scavenger/bone destructor of the Ice Age, the Ice Age spotted hyena Crocuta crocuta spelaea, will be discussed as the oval hole producer herein (figure 2), based on the intensive Late Pleistocene central European cave bear and top predator studies in and outside caves of the past years (e.g. [3,12–16–22,51,54,55]). This results in a different viewpoint of modern zooarchaeology, multiple animal/human use of larger cave systems and cave models (figure 2). The Ice Age top predator research in Europe focused these past years on hunting of cave bears in large cave bear dens. There, damage on cave bear bones is now well known and reported in several publications (e.g. [3,4,16,18–21,51,56]; figure 2). All former archaeological, ecological focus cave bear ‘bone flute’ studies forgot all four cave bear predators—steppe lions (Panthera leo spelaea), leopards (Panthera pardus spelaea), Ice Age spotted hyenas (Crocuta crocuta spelaea) and Ice Age wolves (Canis lupus spelaeus)—which are known now to be cave bear killers, and main consumers in mountain regions, where mammoth steppe megafauna were absent [4,18–21]. These predators specialized in consuming mainly (and especially in winter times during cave bear hibernation) cave bears in boreal forest mountain regions, but in different ways and with different impact on the carcasses and bone destruction (cf. figures 2 and 3). However, the main ‘bone destructor’ is known to be the European Ice Age spotted hyena [19] (figure 2), with cave bear bone damage first understood at the overlapping hyena den (cave entrance) and cave bear den of the Perick Caves [50–52], with newer proof at Sophie's Cave [21,22], and Hermann's Cave [16] or Zoolithen Cave [18] and herein best demonstrated and newly added for the Weiße Kuhle Cave (figures 3, 4, 6 and 7).


Figure 5.

All ‘cave bear cub femora bone flute’ sites failed to date into the ‘Neanderthal times’ because all are not of Neanderthal (Middle Palaeolithic) human, but are instead from modern human Aurignacian occupations in ‘archaeological layers’ at entrances of cave bear dens, cf. [1,7,9], or deeper in caves due to cave bear hunt [23]. There, the cave bear layers themselves, which generally span from the MIS3–5d=25,000–113,000 BP, overlap/intercalate with the Cro-Magnon times, mainly Aurignacian, partly Gravettian, cultural layers [5,23,57]. Cave bear bones and archaeological layers are therefore not exactly isochronous in several cases (even mixed due to possibly bioturbation by cave bears building their nests, or burrowing porcupines or digging Ice Age spotted hyenas; cf. [13,20,21]), and ‘cave bear bone flutes’ would have been, if such, from modern human layers, in all cases.


Figure 6.

The pseudo-bone flutes all come from layers of the MIS3–5d (herein added up to MIS 6) and are from smaller early cave bear forms of Ursus spelaeus subsp. (i.e. U. s. eremus, U. s. spelaeus sensu taxonomy of Stiller et al. [58]); interestingly though, alpine Late Pleistocene cave bear forms (U. s. ladinicus) do not show such holes in femora (i.e. indicator of absence of hyenas in alpine regions, and proof of holes made only by hyenas which are found only in middle high elevated mountain regions [19]). As is now well known, Aurignacian humans lived in Europe together with the last and largest cave bear species U. ingressus [16,18,21,23,58,59]. At a few cave sites in Europe, hunting of cave bears with propulsory spears is documented for the Aurignacians–Gravettians [5,22,23,57]. Therefore, the ‘pseudo-bone flutes’ originate from both smaller Ursus spelaeus subsp. (eremus or spelaeus) and the large U. ingressus, and from mountainous regions, where Ice Age spotted hyenas were around all over Europe (cf. map in [19]).



Figure 7.

Drilled holes were produced experimentally for a reconstruction of a ‘cave bear cub bone flute’ (cf. [60]) and are very different also on the hole margins and forms. All herein figured cub femora have, different from drill-holes, distinct characters (figures 5–7): (a) the holes are not fully round, instead oval-shaped, and beside the hole (see also [24]) a breakage-arch indicates an ‘impact’, rather than drilling (cf. also modern hyena impact mark pictures in [61]), (b) the margins are convex in cross-shape, and not steep-straight as with drills, (c) the corners are smooth and do not have drill/cut mark signs, at all, and (d) in most cases (figures 5–7), the antagonistic punctures/tooth marks (lower/upper jaw dentition fit) are present. Such antagonistic tooth marks are found often at different medium-sized hyena prey bones including their own species femora or even Neanderthal femora [19,20], also documented in the modern actualistic spotted hyena bone accumulation record [61–63]. Finally, also X-rays of the ‘bone flute’ hole margins did not verify any ‘drilling’ nor any stone tool work on the bone (cf. [64]).


Figure 8.

It is no wonder then that further incorrectness about cave bear bone taphonomy at Divje Babe Cave 1 was published (cf. [73]), because all ‘fragmented’ bones were simply declared as due to ‘sediment pressure’. Indeed, some are naturally weather-cracked. A studied ulna of a cave bear at the site is one of the best examples of bone crushing by hyena premolar teeth. Even the puncture marks in the upper shaft area are visible, demonstrating the scavenging/bone cracking activities also in the Divje Babe Cave 1, similar to that found in German caves (cf. [4,16,18,23]; cf. figures 5–7). His final arguments that ‘hyenas are absent’ at this site (cf. also carnivore fauna in [25]) are none, because as ignored in intensive cave bear den cave site taphonomy studies of Europe, the models of presence and absence of any large predator are well known [4,13,18,19,23]. Hyenas and other carnivores are rarely found at the ‘scavenging sites’, including caves and cave bear dens, because they are only found there when they occupied the cave entrances as (a) cub raising, (b) communal or (c) prey depot dens (cf. definitions and discussions in [4,14,18–20,20,21,54,74].

Conclusion

The ‘cave bear cub femora with holes’ are, in all cases, neither instruments nor human made at all. All cave bear pseudo-bone flutes are not dated to Neanderthal Middle Palaeolithic Mousterian layers, but instead, if possible to date, to Late Palaeolithic, Aurignacian/Gravettian layers. There, where they are dated absolutely (Divje Babe Cave 1) are without archaeological context at all, and simply of cave bear den use during the MIS 3–5d. At these times, different cave bear subspecies Ursus spelaeus subsp. eremus (smallest cave bear) and spelaeus (i.e. Neanderthal times) and U. ingressus (largest cave bear, i.e. Aurignacian/Gravettian times) used caves all over Europe for cub raising and hibernation. All the large carnivore punctured cave bear cub femora (and other punctured bones) appear always in small to large cave bear den cave/cave entrance contexts. This sometimes overlaps with hyena dens and human camp sites at cave entrances only, where cave bear den, carnivore den and human remains are even mixed up (partly separated in layers), all over Europe due to competition for and seasonal use of cave entrances/rock shelters. The cave bear bones with round–oval, larger puncture marks can be well attributed solely to the main cave bear scavenger of Europe—the Ice Age spotted hyena Crocuta crocuta spelaea. This main Late Pleistocene bone destructor in Europe is known recently with more than 150 den sites (95% are cave sites) all over Europe. At cave bear dens hyenas left, by periodic scavenging, up to 20% of damaged bones, whereas also lions (cave bear killers), leopards and wolves played a larger role in the cave bear hunting/scavenging, even deep in caves. Those indeed also left, in some cases, round–oval, larger punctures in cave bear bones, but with their canines only in soft spongiosa (pelvis, vertebrae), and never in any bone shaft compacta. Neither carnivores nor cave bears (herbivorous) used their canine teeth to crush longbones, or any other bones. Therefore, all other top predators—except hyenas—can be excluded, at least for the round–oval punctures in cave bear longbone shafts. Only hyenas have developed a carcass destruction and butchery strategy, also for cave bears. This strategy is demonstrated, herein in detail, on cave bear femora destruction (especially material from Weiße Kuhle Cave, Germany), which is presented in three stages and for different aged individuals—cubs (less than 1 year), subadults (1–2 years) and adults. Cub bones are ‘soft’ and thin-walled in the bone shaft compacta, which increases in subadults to adults. This explains why puncture marks are found only in cub (less 1 year) femora, and partly in subadults, whereas they are absent completely in adults, because hyenas cracked those bones into pieces with the premolar triangle teeth (i.e. bone crushing teeth) for access to the bone marrow and easier swallowing of those pieces for the bone collagen use. Hyenas left, therefore, ‘pseudo-bone flutes’ during the Late Middle to Late Pleistocene all over Europe in cave bear dens, and on different cave bear species/subspecies. This is known due to lack of breakage on most of the cave bear cub femora, which generally show additional diagonal zigzag margins (from chewing joints by scissor teeth of hyenas) or have triangular or smaller scratch tooth marks. This even allows reconstruction, in some cases in detail, the tooth mark of the upper and lower jaw teeth of hyenas—the last tooth mark of the premolar bone crushing triangle of the powerful jaws of the last hyenas of Europe. Finally, some flakes and refitted cub femora, both with tooth mark holes, prove the bone cracking activities at cave sites.

DOI: 10.1098/rsos.140022 Published 1 April 2015