GEOLOGY

The general speleogenesis of Urşilor Cave has been discussed by Vălenaş (1979) and Rusu and Racoviţă (1981) and is modified here based on bone bed genesis and directions of sediment transport in the cave. Dated speleothems from the cave are all of Holocene ages; in some places newer speleothems have covered almost entire older generations of speleothems (Onac et al., 2002) and are not useful for understanding Pleistocene cave development.

The cave lies 100–200 m beneath the surface and is developed in Upper Jurassic (Tithonian) limestone. It consists of 1,500 m of rather large passages developed along two distinct levels, located at (Vălenaş, 1979; Fig. 1). The lower level (= “Scientific Reserve”; Fig. 2A) includes a small stream that has eroded through sediment (fine sand and clay) terraces and hosts most of the cave-bear remains. The upper level (Fig. 2B) is no longer hydrologically active and has most of the older speleothem formations preserved. Although the two present artificial entrances were sealed to preserve the original cave microclimate, the bones in the Scientific Reserve area (Fig. 2A) show recent degradation.

The genesis of the cave, as it is currently understood from the new analysis presented herein of cave sedimentology, bone bed types, and plateau morphology (Figs. 2A–B), includes at least four evolutionary stages. The first three stages are (1) drainage toward the nearby Crăiasa Valley at the level of what is now “Lumânărilor Gallery” (Candlestick Passage); (2) deepening of the underground drainage through the current “Oaselor Gallery” (Bones Passage); (3) a massive collapse at the end of the Emil Racoviţă Passage in the middle of the cave (Fig. 2A), which blocked the normal streamflow and generated a bone bed accumulation in the “Bonebed Chamber.” This chamber received material via inwash from the upper passages (Fig. 2A) as demonstrated by similar gravel types in both levels, which do not appear upstream in the sections of the lower level of the Scientific Research area but several meters higher in the Collapse Chamber. Therefore, similar, gravel-rich sediments are only found in the Collapse Chamber and the Bonebed Chamber; it is unclear where the material originates from exactly, but its origin must to be from a higher cave level. Bones and gravel-rich clays can be followed along a speleothem wall which was washed in small niches reaching from the Collapse Cone Chamber down into the Bonebed Chamber on its western part; the latter chamber seems to be connected with unexplored and mainly refilled cave branches. It seems that the collapse closed the connection to the lower cave area, which was, in the Pleistocene, the original access to the deeper cave hibernation plateaus (Fig. 2A). In this phase of cave development the middle part connecting the middle and lower levels must have become blocked by slumping, which may have happened very quickly as evidenced by an articulated skeleton in the thick, densely packed bone bed of the Bonebed Chamber. Cave-bear remains were washed mainly from this middle to upper cave level into the lower Bonebed Chamber of the cave, where a bone bed up to 1.5 m thick accumulated. The bones did not accumulate from the hibernation plateaus upstream based on sediment types and reconstructed transportation directions. The plateaus were not flooded during the use as hibernation areas, as evidenced by the existence of all the tracks, traces, beds, and articulated skeletons. Cave-bear tracks on the collapse cone sediment surface (cf. Fig. 3.1) indicate that this collapse must have happened during the middle late Pleistocene during occupation by the last cave bear generations to inhabit the cave. This terminal generations no longer used this passage to go deeper into the cave system, but at least they must have used, in the final cave stage, the anterior cave part (= Visitor loop, cf. Fig. 2B). The fourth stage is that the underground stream found an even deeper drainage and now surfaces through Huda de la Chiscau, a small side-valley cave located only a couple of meters above the Crăiasa Valley.

A new sedimentological overview was made using, in some places, previous excavations together with the natural cave walls. The overview was made to obtain a preliminary understanding of the cave infill and of the collapse feature in the middle section of the cave. Two previous excavations at the beginning of the public part of the Visitor Loop (Fig. 2B) reached depths of only 2–3 meters (cf. Terzea, 1978). The sediments differ in these old excavations, and their micromammal and megafaunal contents suggest late middle Pleistocene (Eemian and Weichselian/Wuermian) (Terzea, 1978). In the Scientific Reserve, in the deeper second half of the cave that is not accessible to visitors, more exposed sections are present (Sections 1–11, Fig. 2A). While the first two sections have similar, fine gravel sediments forming the bonebed in the “Bonebed Chamber” (Fig. 2), the sedimentology changes in Chamber 1 to a more complex section with different sediment layers where the river descends. In Chamber 2 and through to the end of the cave is a third sediment series consisting of repeating similar sequences of red clay, gravel, and brown, laminated clay deposited during flood events, but the sediment thickness decreases toward the end of the cave. The age of these sediments is unclear, but the hibernation plateaus and landslide material on which all cave-bear traces, beds, bones, and articulated skeletons have been found must have already been present during the Weichselian and possibly even in the Saalien. The tracks and bones are all only on the plateau sediments and none in the flood clay deposits, which support the sediment and plateau dating and genesis. Bones therefore were not washed from the plateaus to the “Bonebed Chamber”; rather, they were transported opposite today's stream direction from another higher cave level in the first cave half (Fig. 2). Flooding onto the clay plateaus therefore could not have occurred in the Scientific Reserve during the late Pleistocene: the stream flow must have been more or less stable, no longer flooding the hibernation plateaus, in order for them to provide suitable shelter for cave bears during the winter. The plateaus seem to have been formed in middle or even early Pleistocene times. The cave bears whose remains are found on those plateau sediments have P4 tooth morphologies typical for early to middle Weichselian cave bears, which supports a pre-Weichselian/Wuermian age of the hibernation plateaus.

MATERIAL AND METHODS

Two expeditions into Ursilor Cave in 2007 and 2008 focused first on cave genesis including the sedimentology. Different open sections from the entrance to the end of the cave were measured and documented (Fig. 2). Also, bone taphonomy was used for the cave genesis analyses, including both cave-bear bones and the cave bear ichnology (e.g., footprints, beds). In the first part of the cave, nests are the only ichnological records (Fig. 1B) that were mapped onto using the first published cave map (Jurcsak et al., 1981). The entire cave system was checked for tracks and traces as well as articulated skeletons of cave bears, lions, and so forth (Fig. 2). New maps, with today's preserved cave bear nests (some are destroyed in the Visitor Loop due to the trail works), track areas, and new names of all the chambers and passages, were made for clear, detailed localization of finds (Figs. 2A–B). The main important sites, tracks, and fossils were photographed, and some nests and track areas were mapped by digital photography and computer redrawing. The cave bear bones and tracks and traces were not removed; they remained in place in the cave and so individual remains lacked repositories and formal catalogue numbers. The traces are too fragile to permit traditional replication (casting) techniques. Full surface mappings are planned; testing laser-scan techniques, the optimal method to record data about these tracks, were tested and seem to become with future modifications successful, but mapping of tens of thousands of cave-bear traces would require years of work. Stepping onto the hibernation plateaus in the Scientific Reserve invariably results in the destructions of some traces, which has happened many times over the past 30 years. To document and provide an overview of the undamaged tracks and traces, the best preserved tracks are figured, one of which has already been partly destroyed by more recent caving activities. Most of the tracks presented in this paper are complete, single track imprints from single manus/pes sets all found either in corners of the cave or very close to the cave walls, where such single tracks were able to survive subsequent trampling and overstepping by cave bears or destruction by caving activities during the last 30 years. About 50 single manus and pes imprints have been recorded, as well as many incomplete footprints (which are difficult to count), trampled horizons, and thousands of well-defined claw scratch marks. The cave bear traces are in situ within the federal protected and runned show cave, mainly on the clay flood deposits and hibernation plateaus in the Scientific Reserve area (Fig. 2B) and to a lesser extent with scratch marks on the limestone cave walls.

The first tracks (seven manus and pes imprints) were found in the middle cave area (around the Collapse Cone, Fig. 2A) and are 200 m far of the cave entrance (Figs. 3–4 5).

The surfaces of the hibernation plateau in the Scientific Research area (Fig. 2B) have been heavily trampled, and the upper layer has even become compacted through use by generations of bears (Figs. 5–6). Only in some corners of these trampled surfaces are footprints sufficiently distinct that they can be reasonably well mapped, even though they overlap one another to a considerable degree (Fig. 7). In one area in a small space (Fig. 7), less then 11 footprints can be identified and still more can be present (Fig. 5). It is difficult to decide whether the tracks in this area are manus or pes imprints because overstepping has destroyed the important features by which they can be distinguished.

Hair traces were identified during photographic analysis of scratch marks and overlap the photographed scratch marks in various places. They are situated along the small stream, on its slopes, or on the small slippages along its banks. They are also present in a particular corner where cave bears had been digging, with the hair impressions overprinting their scratch marks (Fig. 8). At such digging places, also best documented in the Dancing Cave Bear Chamber (Figs. 9A–B), the bears can be demonstrated here best in their activities. There those were kneeing or those went down onto ther elbows when digging in corners, where was few space.

Ursichnus nov. ichnogen

Ursichnus europaeus nov. ichnosp.

Figs. 3–7

Material: Several single manus and pes imprints, and areas of trampled horizons with numerous footprints and manus-pes sets. The material is preserved in situ in the nonvisitor part of the Urşilor Cave, (Romania) which is under federal protection and a natural public protected monument; therefore, the figured tracks received on place for their identifications and positions in the cave collection numbers.

Stratum typicum: Upper Pleistocene (probably Early to Middle Weichselian/Wuermian). Tracks are printed on brown cave flood clay plateaus on the surfaces ranging from the middle part to the end of the cave in the Scientific Reserve (Figs. 1D, 2A).

Type locality: Scientific Reserve area, Urşilor Cave, west Carpathians, Transylvania, Romania.

Holo- and syntypes: Pes imprint of an adolescent bear (Figs. 3A–B; coll. No. UrsPest-track 1) and manus imprint of a large adult bear (Figs. 4A–B; coll. No. UrsPest-track 2) as best-preserved specimens.

Paratypes: All illustrated manus and pes tracks from Figs. 3–7.

Ethymology: “Ursus,” Greek for “bear,” and “ichnus,” Greek for “trace.” The ichnoscpecies names reflect the presence of such cave-bear tracks across Europe.

Description: Plantigrade manus and pes imprints with manus/pes length ratio of about 1:1.25 (pes maximum length about 25 cm, manus maximum length about 17.5 cm), which differ from each other. The small heel of the front paw is ovoid, while the large heel of the hind foot is triangular. Claw impressions are not always visible. In a normal gait, the manus imprint is in front and quite close to the pes imprint, but the manus prints can be absent if the animal walked bipedally only on its hind feet, which appears to have been quite common in caves such as Urşilor Cave. Both the manus and pes impressions consist of three primary elements: heel pad, digit impressions, and claw marks. However, the expressions of each part vary.

Manus prints: The manus imprints consist of roughly kidney-shaped main pads with five round to oval digit depressions in front of the pad (Figs. 4A–B). The tracks typically have claw marks in front of the digits, although these can be absent if, for example, the prints were left on relatively dry surfaces. The manus impressions are more different, more irregular in softer clays (Figs. 4B–E). In some places, the trackmaker slipped mainly with the fore limb while registering a manus-sliding imprint track (Fig. 7). Because the sizes of the prints depend on the ages of the individual trackmakers, the presence of animals ranging from one year old to mature adults can be inferred from the preserved footprints.

Pes prints: The pes imprint is similar to the manus imprint in the claw and digit impressions but differs in the pad which, in a well-preserved print, can be subdivided into four areas (Figs. 3A–B) within an outline that is more or less triangular. Often only the triangular shaped heel impression can be identified: the heel print of an adult animal can be seen next to the complete footprint holotype (Figs. 3C–D).

Scratch marks: There are several thousand scratch marks typically in sets of four parallel marks (Figs. 7–8), starting in the middle section of the cave. The closest ones to the present day entrance are scratched slightly on the limestone walls above the vertical shaft between the Collapse Cone and the Bonebed Chamber (Fig. 2B), but these are only present in a few places further deep in the cave. Also, on one wall in Chamber 5 (Plateau 11), where a cave bear stood on the hind legs trying to climb up the wall, many large scratch marks were left in the Scientific Reserve Area. Most of the claw imprints and scratches are found in clays on the floors within the Scientific Reserve area of the cave, where the flood-deposit clays formed a perfect, extensive hibernation area. These clays exhibit landslide slippage structures along both banks of the small, active stream, and it is here in particular that cave bears of all ages, ranging from young cubs to old bears, left thousands of scratch marks (Figs. 2 and 8). Some of the parallel scratch sets are only about 5 cm wide, whereas others are about 12 cm wide (Fig. 8), almost certainly made by individual bears of different ages. These marks were made mainly by bears walking down to the stream and then back up to their hibernation places on the clay plateau about 2 m above the level of the stream. Individual scratch marks from large cave bears are each about 4 mm in width, indicating dull claws (Fig. 11). The scratch marks do not receive an ichnogenus or ichnotaxon name because in some cases they represent “footprint preservation types” (“slope track type”), and in some cases they reflect “behavioral traces” (“bed scratching trace type”) if found around the hibernation beds, which let them finally attribute to both defined ichnogenera and ichnospecies.

In addition to the cave bear scratch marks, there are several thousand very small scratches produced by bats, which consist of three to four parallel very thin scratches a few mm apart (Fig. 8). The bat traces often overprint the cave-bear scratches (Fig. 8), indicating that they are younger in age.

Differential diagnosis: The only described fossil bear-like track Platykopus (Miocene) (Sarjeant et al., 2002; Abbassi, 2010) has large plantigrad footprints, but all five digits are close to the manual/pedial pad. In Ursichnus the digital pads are all within distance of the pads, which is not a preservation type feature. Also, those are larger and singular; in Platykopus, the more primitive pedal skeletons have two pad impressions. The manus/pes proportions are in Platykopus nearly 1:1 whereas the length proportion in Ursichnus is (excluding the hallux imprint) about 1:1.25. In addition and absent in Platykopus is the hallux imprint behind the manus tracks, also indicating a more primitive mammalian trackmaker (creodont trackmaker Hirpexipes, after Sarjeant et al., 2002). The pes in Platykopus is antero-posteriorly oval; in Ursichnus, a large triangular. The pads are in the manus and pes in Platykarpus more or less all similar in size; in the Ursichnus footprints the three middle digit pads are larger, the outer ones, smaller. The claw marks in Ursichnus are much larger and wider as in Platykopus (short and pointing similar as in brown bear tracks) and modern brown bear tracks, but are similar to black bear shovel-like claw impact marks. Platykopus tracks are smaller with maximum pes length (13.5 cm) as in Ursilorichnus (max pes length = 25 cm).

Ursalveolous nov. ichnogen

Ursalveolous carpathicus nov. ichnosp.

Figs. 1B, 10A–C

Material: A minimum of 140 bed depressions all over the cave system (Figs. 1–2), of which some are partly destroyed by the modern Visitor Loop trail (Fig. 2A).

Stratum typicum: Upper Pleistocene (probably Early to Middle Weichselian/Wuermian). Depressions are left on brown cave flood clay plateaus on the surfaces but also directly on the coarse gravels ranging from the Entrance (Fig. 1B) to the final end of the cave in the Scientific Reserve in Chamber 8, Plateau 18 (Figs. 1D, 2A–B).

Type locality: Urşilor Cave, west Carpathians, Transylvania, Romania.

Holo- and paratypes: Here those depressions were chosen in which cave bear skeletal material was found in-situ in the beds, or where scratch marks around indicate the cave bears as trackmakers at a minimum of four beds. The track depression in Chamber 2, Plateau 5 (Figs. 10C–D) is the holotype (No. UrsPest-bed 1). The syntypes are the shallow depressions with the adult female skeleton in place (Fig. 2B, No. UrsPest-bed 2), along with the cub skeleton in a medium-deep depression (Figs. 10E–F, No. UrsPest-bed 3), another one without adult bone remains (Figs. 10A–B, No. UrsPest-bed 4), and finally the in-work stage left cave-bear bed with scratch marks and kneeing impressions (Fig. 9A, No. UrsPest-bed 5); the paratypes are all other beds in the cave (e.g., Fig. 1B; e.g., No. UrsPest-bed 6).

Ethymology: “Ursus,” Greek for “bear,” and “alveolus,” Greek for “depression.” The ichnospecies names reflect the presence of such cave bear beds in the Carpathian cave bear den caves.

Description: At least 140 cave bear beds in the cave were found mostly in round-oval shape and up to 50 cm deep, several surrounded by scratch marks (Figs. 9–10). The outline of one bed (Figs. 9A–B) appears visible, which would appear to fit an adult hibernating bear lying on its left side. A depression for the front legs can be seen in some beds, including one (Figs. 10E–F) made by a one-year-old cub that did not survive its first hibernation. Unfortunately, the bone material in the trace is highly weathered and poorly preserved, and the skull must have fallen off the slope on which the animal was resting its head. Another cave-bear nest (holotype, Figs. 10C–D) is up to 50 cm deep; in this, male skeletal remains in articulated position with the baculum preserved. This skeleton, however, consists of the distal body part only with the lumbar vertebrae, complete pelvis, and hind limbs; the anterior part is represented by a single cervical vertebra and a humerus. It is unclear what happened to the anterior part of the animal; possibly it was removed by the second bed user, which removed the bones onto the hibernation plateau where many single bones of several specimens are scattered. Lions, which were found with three skeletons around (Diedrich et al., 2009b), have most probably not torn this carcass apart or removed the missing elements, but it cannot be excluded. The present bones are partly pressed into the soft clay, so it seems as if another bear partly cleaned the bed and was finally sleeping on the skeleton part, which he pressed into the ground. Conceivably, the first male cave bear (skeleton) hibernated in a different position than the others, not lying on its side but possibly resting in a squatting position, which would explain the presence of the posterior elements. The scratch marks around several cave-bear beds provide the first reported evidence that cave bears dug out their own beds, including adult and already cubs (large and small scratch marks, Figs. 10A–F). In one case the shape of the bed (Figs. 10A–B) fits the outline of a bear, suggesting that the animal laid on its left side when hibernating. It appears that the bear cleaned out the bones of a less than three-month-old cub from the nest prior to using it, leaving the tibiae of the juvenile in the depression but a humerus outsideit.

DISCUSSION

The large footprints in Urşilor Cave are typical bear tracks and similar in shape to those of modern black and brown bears (cf. Murie et al., 2005; Fig. 11). When compared to Ursus arctos (brown or grizzly bear tracks) and Ursus americanus (black bear tracks), the tracks of adult cave bears (Ursus spelaeus subsp.) are the largest (Fig. 11).

The front paws of U. arctos tracks have small heel pads (Murie et al., 2005), and the toes are closer together and straighter than those of U. americanus or U. spelaeus, in which the toes form a more rounded arc (Fig. 11). An important distinguishing feature between the tracks of U. americanus and U. spelaeus is the difference in the claw mark morphologies, reflecting differences between sharp or dull claw types and different lengths. Brown bear tracks have small claw marks far in front of the toe pads, a result of long, sharp claws (Murie et al., 2005). This is especially true in the manus prints as a result of their long claw-phalanx and resulting long claws. In black bears the claws are much shorter and the claw impressions are therefore much closer to the toe pads (Murie et al., 2005; Fig. 11).

The toe pads of cave bear tracks are generally more oval, not as close together in black bears as they are in cave bears and brown bears. The cave bear claw marks are closer to the toe imprints (Fig. 11), similar to black bear tracks, and the claw phalanx of cave bears is much shorter than in brown bears. Cave bears must therefore have had short but broad claws, more similar to black bears than brown bears. This anatomical detail is important for a palaeoecologically; unlike those of brown bears, the claws of cave bears and black bears were/are not lethal weapons for catching animal prey or fish (e.g., Smith, 1985; Bateman, 1987). The feeding habits, diet, and food-gathering behavior of cave bears must therefore have been more similar to those of black bears than brown bears, with both cave and black bears using their claws more for scratching and digging (Beeman and Pelton, 1976; Smith, 1985). Black bears also leave scratch marks on the walls of their hibernation dens in American caves (Beeman and Pelton, 1976) similar to those of cave bears. Scratch marks from brown bears are narrower and closer together.

The most rare ichnological cave bear records are abundant hair traces in the Urşilor Cave; only at the Slovenian Divje babe I Cave were such hair trace marks reported (Turk, 2007). On the slopes of the Urşilor Cave it must be the stomach fur imprints, when bears climbed up the steep corners touching those with their bellies, whereas such fur imprints seem to have been left during bed excavations.

The few illustrated tracks from other caves in Europe, such as Bijoux Cave (Kurten, 1976), Bruniquel Cave (Rouzaud et al., 1995), and Chauvet Cave (Garcia, 2005), match the variations in Ursichnus europaeus footprints. Many more footprint analyses would be required from European caves to enable the separation of footprints of large brown bears from those of small cave bears, as well as to obtain otherdata.

In Urşilor Cave, cave bears clearly were using their claws for digging to open or explore corners of the cave, but primarily to excavate their hibernation beds named here Ursalveolus carpathicus. Such cave-bear bed depressions are instead of footprints well known all over Central Europe in different cave bear den caves (e.g., Ehrenberg, 1967; Kurtén, 1976; Chauvet et al., 1996; Rabeder et al., 2000; Garcia, 2005; Fosse et al., 2006; Quilès et al., 2006), best in the French Rouffignac Cave (Plassard, 1999). Alternatively, cave bears were trying to find the way out of the cave by touching the walls, and they certainly could have used markings for orientation purposes within caves as do black bears (Smith, 1985). In Urşilor Cave, however, most of the scratches were left as a result of climbing up and down slopes, possibly while exploring the cave for good places to hibernate, and to approach and leave the stream that runs through the Scientific Reserve cave area. When outside the cave, the bears may have used their claws the way modern black bears do, that is, to scratch the bark of certain trees to get the sap, especially in spring. The shovel-like manus of black bears are also used for excavating roots and to obtain honey from bee hives (Beeman and Pelton, 1976; Smith, 1985).

CONCLUSIONS

Urşilor Cave contains many ichnological marks, including more than 140 beds named here Ursalveolous carpathicus nov. ichnosp., several hundreds of footprints called Ursichnus europaeus nov. ichnosp., and finally thousands of scratch marks belonging to footprint preservation types (slope track type) and to behavioral activities (bed making) and a large number of fur impressions left by cave bears. The partly articulated skeletons and large quantities of bones found in this cave offer a detailed picture of the underground life of Upper Pleistocene cave bears before, during, and after hibernation. The bone accumulations and ichnology of Urşilor Cave have been influenced by environmental changes and the unique cave morphology (Fig. 12). The uppermost and middle elevated cave levels (Level I) were initially filled with pre-Eemian and Eemian sediments as a result of periodic flooding. The presence of large pebbles interspersed with angular and mostly well-preserved bones in the middle level indicates rapid periodic infilling with river terrace gravels mixing with cave-bear bones to create a “Bonebed Layer.” That these bones were already inside the cave is suggested by the fact that they were quite well preserved within and on top of the gravel layer and have clearly only traveled a short distance. At that time, cave bears probably only entered the cave through the main opening and used only the drier conglomerate surfaces. The best hibernation beds are still preserved, but they seem to have used, until the mid-late Pleistocene, the clay plateaus of the lower elevated cave part of the Scientific Reserve area (Fig. 12). The flood-deposit clays of unclear age were used intensively, possibly only in the early to middle Upper Pleistocene, before the middle cave area was blocked by a larger collapsed cone (Fig. 12). An articulated skeleton was embedded in the thick, densely packed bone bed. Cave-bear remains and bones were washed mainly from the middle to upper cave levels (Level I) into another (Level II) in the Bonebed Chamber, where a bone bed up to 1.5 m thick accumulated. Since this bone bed was made up of the same sediments found in the collapsed zone, the bone material is interpreted as having derived from the higher cave level. During this period of slumping and water influence, the distal parts of the cave became disconnected to the upper cave level.

Both hyenas (represented few by bones and prey remains at the entrance area only) and cave bears with their bones and tracks or traces all over the cave system have used the same cave without contact; therefore, cave bear bone damaged by carnivore bites are not seen yet. The small, active stream of the lower level appears to have flowed in the late Pleistocene possibly in the opposite direction of former clay input from the entrance, leaving the anterior part of the cave dry, with drainage to the lowermost cave level of Level III (ponor). During higher water stands, the edges of the flood clay plateau collapsed along the stream margins in small landslides. These areas are covered by tens of thousands of scratch marks, demonstrating that cave bears traveled from the hibernation plateau between 0.5 m and 2 m down to the stream during the late Pleistocene (Weichselian/Wuermian) and back up again. By following the stream, which is indicated by scratch marks and footprints along the stream walls, the cave bears would have reached the exit/entrance, but they possibly also went to the stream during hibernation to drink. Cubs may have gone over the slopes leaving their scratch marks behind, providing evidence that bears were actually born and raised in the cave.

The cave bears also often touched the slopes with their bodies, leaving many fine, parallel hair impressions imprinted in the clay.

Some bears died in the cave hibernation beds, including an adult and a one-year-old cub. Both skeletons were found with “scattered bone beds” on the plateaus in the distal part of the cave (Scientific Reserve), but only the cub was in a sleeping position. This cub, a young male, was found in a slight depression or cave-bear bed surrounded by cub-sized scratch marks, indicating that it was the maker of the bed. A larger cave bear animal lay on the plateau close to the cub but not in any depression, suggesting either that the animal did not die during hibernation or that cave bears may not have always excavated deep depressions to sleep in. A final possibility is its kill by a lion, since a scattered steppe lion P. leo spelaea skeleton was found on the same plateau.