sign in sign up
<<
Home , Cave bear, Homo ergaster, Ursus deningeri

CAVEBEAR ECOLOGYAND INTERACTIONSWITH

MARYC. STINER, Department of Anthropology,Building 30, Universityof Arizona,Tucson, AZ 85721, USA,email: [email protected]

Abstract:Human ancestors (Homo spp.), ( deningeri, U. spelaeus), andbrown bears (U. arctos) have coexisted in Eurasiafor at least one million , andbear remains and artifacts frequently are found in the same . The prevalenceof cave bearbones in some sites is especiallystriking, as thesebears were exceptionallylarge relative to archaichumans. Do artifact-bearassociations in cave depositsindicate predation on cave bearsby earlyhuman hunters, or do they testify simply to earlyhumans' and cave bears'common interest in naturalshelters, occupied on different schedules?Answering these and other questions aboutthe circumstancesof -cave associationsis made possible in partby expectations developedfrom research on modem bearecology, time-scaledfor paleontologicand archaeologic applications. Here I review availableknowledge on Paleolithichuman-bear relations with a special focus on cave bears(Middle Pleistocene U. deningeri)from YarimburgazCave, .Multiple lines of evidence show thatcave bearand human use of caves were temporallyindependent events; the apparentspatial associations between human artifacts andcave bearbones areexplained principally by slow sedimentationrates relative to the pace of biogenicaccumulation and bears' bed preparationhabits. -linkedbehaviors and population characteristics of cave bears,based on osteometric,isotopic, and age andsex structureanalyses, indicate that they dependedheavily on seasonalfood supplies,which were rich in resistantplant materials and cryptic, gritty foods. Thereis little evidence of direct ecological interactionamong Pleistocene humans and cave bears.

Ursus 11:41-58

Key words: brownbears, cave bears,Mediterranean, mortality, paleodiet, Pleistocene human-bear interactions, sex ratio,Ursus arctos, Ursusdeningeri, Ursus spelaeus

Pleistocene sediments in Eurasiancaves often contain By way of anthropologicalbackground, the chronol- complex records of habitationby predatoryspecies. Of ogy of Paleolithic culturesbegins aroundthe Plio-Pleis- these, early humans ( spp.), (Canis lupus), tocene boundary and lasts until the Holocene. The foxes (Vulpes spp.), spotted hyenas (Crocuta crocuta), Paleolithic traditionallyis divided into 3 major cultural and bears (Ursus spp.) were especially prolific sources phases: the Lower,the Middle, andthe UpperPaleolithic. of bone refuse. Making sense of cave faunas is an inter- The hominid forms associatedwith these culturesvaried esting, but also challenging, enterprise,beginning with considerably,and there is no simple correspondencebe- the questions of how the assemblagesformed, what spe- tween cultural (behavioral) change and hominid mor- cies were active collectors and modifiers of bones, and phological (skeletal) change. Lower Paleolithicartifacts the extent to which sediment chemistryfavored skeletal are attributedto some of the late Australopithecinespe- preservation. These problemsfall in the methodological cies, which were confined to the African continent, as domainof vertebratetaphonomy, an areaof researchcon- as to early variantsof the genus Homo, which by cerned with how bone assemblages become part of the 1.4-1.8 million years ago (MYA) spreadfrom into paleontological and archaeologicalrecords. much of . First appearingsome 250,000 years While clear answers about assemblage formationhis- ago, Middle Paleolithic artifactsare characterizedby in- tory seldom are easy to obtain, thereis a wealth of infor- novations in productiontechniques and arti- mation locked in cave sites about the ecology of early fact forms. They are attributedto archaichumans such humans and their relations with potential competitors, as Neandertalsin general (H. sapiens neanderthalensis) mainly of the order . The story of human- as well as to earliest anatomicallymoder humans (H. bear interactionsis both importantand enigmatic. Con- sapiens sapiens) in the west Asian cave sites of Qafzeh verging patternsof omnivorywould seem reasonenough and Skhul (Bar-Yosef et al. 1986, Valladaset al. 1988, for Pleistocene humans and bears to have stayed out of Bar-Yosef 1989, Vandermeersch1989). The Upper Pa- each other'sway. Yet the remainsof cave bearsor brown leolithic, which began between 42,000 and 35,000 years bears are found in nearlyevery EurasianMiddle and Up- ago (dependingon region), is markedby spectacularra- per Paleolithic site with preserved bone-sometimes diations in materialculture over a relatively short time many bear bones, but more usually a few. In fact, bear span. Artifacts of all 3 periods commonly are found in remains turn up in Paleolithic cave sites more consis- open sites. Cave sites, usually in solutioncavi- tently than the remains of almost any other large carni- ties, were periodically inhabitedduring the Middle and vore, at least prior to 20,000 years ago, when greater Upper Paleolithic culture periods only, and most of the interspecificexclusion among cave residentsis apparent archaeofaunalrecords of these periods come from caves (Gamble 1986, Stiner 1994). for the simple reasonthat cave sedimentsfavor bone pres- ervation. Paleolithic peoples probably were no more 42 Ursus 11:1999 bound to caves than bears, but both were at times at- Middle Pleistocene (U. etruscus) was relatively small- tractedby the prospectof easy shelter. bodied but evolved into largertypes by the Middle Pleis- Here I review availableknowledge on Pleistocene hu- tocene. This presentationconcerns 2 bears in particular, man-bear relations,as gleaned by archaeologicalinves- cave bears (U. deningeri, U. spelaeus, U. rossicus) and tigations of cave sites, with a special focus on cave bears brown bears (U. arctos and its most immediate ances- in southernEurope and western . The rich litera- tors) (Kurten1976, Baryshnikov1998). Once prevalent ture on moder brown and black bears (U. americanus) throughoutEurasia, all cave bearswere extinctby roughly plays a key role in this kind of research. Specifically, 10,000 years ago (Baryshnikov1999), and most popula- wildlife data are used to build testable predictionsabout tions disappearedconsiderably earlier. Brownbears con- how bears may contributeto faunal assemblage forma- tinue to exist and even doubled their geographic range tion. The faunal patternsinside caves that can be ex- duringthe Late Pleistocene by colonizing the Americas, plained by modern bear behavior are played against as did humans not long thereafter. anthropologicalhypotheses about ancienthuman behav- Because recent humans and bears display strong at- ior. A variety of analytic techniques are marshaledfor tractions to meat as well as to energy-richplant foods, this kind of research,including osteometry, skeletal dam- theirevolutionary histories following biogeographiccon- age analysis, taxonomic and body part profiling, and tact must have affected one anotherto some extent. The mortalityand isotope analyses, to obtain rigorous,if in- ecological links between humansand bears may always direct,evidence aboutPleistocene (and human) have been relativelyweak, because both tend to be versa- ecology. For relatedpublications see Stiner(1994, 1998) tile, generalistforagers (sensu Foley 1984). These links and Stiner et al. (1996, 1998), and for information on were perennial, however, due to overlappingneeds for Pleistocene guilds in the MediterraneanBa- foraging territoryand, periodically,for shelter. sin, see Stiner (1990, 1991, 1992, 1993, 1994) and cita- tions therein. PLEISTOCENEHUMAN-BEAR INTERACTIONS:FACT AND FICTION HISTORYOF HOMINID-URSID There is no shortageof bear stories in anthropological COEXISTENCE andpopular literature. Most of them concern"cave men" Between 1.8 and 1.4 million years ago, populationsof (Neandertals)and cave bears. The original site for the a hominid known as Homo ergaster expandedfrom Af- Neandertalcave bear cult is DrachenlochCave in Swit- rica into Eurasia. Hominids' spreadacross southernAsia zerland, made famous by Bachler's radical interpreta- appearsto have been rapid (Klein 1989). Colonization tion of a faunaconsisting almost exclusively of bearbones of continental and the colder regions of north- (Kurten 1976). The sediments of Drachenloch lacked central and eastern Asia by hominids took considerably artifactualmaterial, but allegedly preserveda stone crypt longer, but they certainly reached these regions before with as many as 7 cave bear skulls arrangedneatly in- 500,000 years ago. The first hominids to enter Eurasia side and, nearby,a smaller stone chest packed with bear were with a notable capacity for huntingver- long bones. Bachler's claim was refutedby Koby (1940) tebrateprey, a propertywhich may have been essential and in Kurten's (1976) highly entertainingbook, The to winter survival in northernhabitats (Foley In Press, Cave Bear Story (also Kurten 1958, 1971, 1973). Stiner In Press). Colonization of new continents put The cave bear cult is a engaging story, if only for the hominidsin contactwith diverse environmentsand novel early spiritualismit implies. The idea that archaic hu- floras and faunas. mans manipulatedbear remains in symbolic ways per- Hominids were late arrivalsto the predatoryguilds of sists, at least partlybecause bear bones frequentlyoccur Africa, where they first evolved, and hominid-camivore with stone artifactsin other Paleolithic cave sites. Cut interactionsmust have intensifiedas meat-eatingexposed marksor burningdamage have been found on a few bear hominids to interspecific competition. Hominids were bones in Middle and UpperPaleolithic sites (Barta1989, are rareover- late arrivalsto predatorguilds, yet again, as they invaded Stiner 1994:109-123), althoughexamples known art of Eurasianecosystems characterizedby winters of intense all. Bear images also figure in the earliest but these cases cold, lasting snow cover, and seasonally scarce food. In the Upper Paleolithic in westernEurope, additionto habitualflesh-eaters, the process of coloniza- date to no earlier than 32,000 years ago (Chauvetet al. bearbones and tion broughthominids together with 2 widespreadom- 1996). Many sedimentaryassociations of one of the nivorous genera of the northernhemisphere-pigs (Sus Paleolithic stone artifactsare much older, and spp.) andbears (Ursus spp.). The ancestralbear of Lower- oldest examples of all comes from YarimburgazCave, INVITEDPAPER * HUMAN-BEAR INTERACTIONSIN THEPLEISTOCENE * Stiner 43

Turkey. Only a few of the remains have cut marks In 1992 I was invited by F Clark Howell (University (<1%) from Paleolithic stone tools. Gnawing damage is of California-Berkeley) and Giiven Arsebtik (Istanbul present on many more of the ungulate bones as well as University) to work on a large, well-preservedfauna that on bear and non-ursid carnivore bones (Fig. 2). Bear they had excavated from the Middle Pleistocene layers remainswere least affected by gnawing ,how- in YarimburgazCave. The site is >250,000 years old ever, and they are the only remainsthat were also gnawed (Blackwell et al. 1990, Farrand1992, Kuhn et al. 1996, significantly by small rodents (Table 3). Stone artifacts Stiner et al. 1996, Santel and von Koenigswald 1998), Table 1. counts taxonomic dating to well before the appearance of anatomically Specimen by general groups represented in the MiddlePleistocene deposits of Yarimburgaz moder humans or Upper Paleolithic cultures anywhere Cave. in the world. The formationhistory is complex, and the Total identified deposits are rich in cave bear remains () skeletal (Table 1) andjust under 1,700 Paleolithic stone artifacts group specimens % (Fig. 1; Kuhn et al. 1996). Also present in low frequen- 151 4 cies are remains of a diverse of Bears 3,920 93 fragmentary array ungu- 3 late and Othercarnivores 109 non-ursidcarnivore (Table 2). Total 4,180 100

~. o'~ 4 .

:t-X

_

o 5 . c im . I cm

r

Fig.1. Examples of Paleolithic stone tools from Yarimburgaz Cave (from Kuhn et al. 1996 with permission of M. Stiner). 44 Ursus 11:1999

Table 2. Macromammal specimen counts by genus or body and the bones of bears, , and non-ursid carni- size category from the Middle Pleistocene of deposits vores are scattered together in the Middle Pleistocene Yarimburgaz Cave. Table The candidates for Total layer (Fig. 3, 4). responsible specimens bone collection and modification in this site are homi- Taxon identified nids, wolves, spotted hyenas, and cave bears. Signs of Cervidae(genus unknown) 36 each candidate are evidenced by the presence of stone Cervus 29 artifacts or skeletal in the Largeungulate 17 representation species profile. Capreolus 13 Dama 13 Equus 9 INTERPLAYBETWEEN RESEARCH ON 6 Bos or Bison 5 MODERNAND PLEISTOCENEBEARS Capra 5 One of the most about the Medium mammal(mostly ungulates) 5 troubling questions Sus 4 Yarimburgaz Cave assemblage is why so many bear bones Mediumungulate 3 occur with the Paleolithic artifacts. One hypothesis pro- Small 2 ungulate that the bear deaths resulted from non-violent Megaloceros 2 poses Hippopotamus 2 causes in the context of hibernation, implying that bears' presence in the cave was not linked in time to human Ursus 3,372 activities there. the bears have been Large mammal(mostly bears) 548 Alternatively, may Panthera 30 hunted by humans, among other predators, implying that Canis or Cuon (?) 25 Vulpes 17 Felis 13 Table 3. Frequencies of gnawing damage from carnivores and Carnivora(genus unknown) 9 rodents on ungulate, bear, and other carnivore remains. Small mammal(mostly small 9 carnivores) Gnawedby (%) Mustela 4 group n Carnivores Rodents Crocuta 2 Ungulates 117 23 6 Total 4,180 Bears 3,129 10 11 Othercarnivoresa 67 18 1 a Consistsprimarily of Panthera,Felis, Canis and Vulpes.

Fig. 2. Gnawed cave bear remains: frontal bone fragment of a young individual punctured by the incisors of a large adult bear (left); phalanges extensively gnawed by small rodents (right), suggesting that these bones lay undisturbed on the cave floor for some time before burial in sediment. Scale in cm. INVITEDPAPER * HUMAN-BEAR INTERACTIONSIN THEPLEISTOCENE * Stiner 45

Fig. 3. Cave bear bone and Paleolithic artifact scatter exposed during the 1988-90 excavations of the Middle Pleistocene layers in Yarimburgaz Cave (Stiner et al. 1996).

Table 4. Pearson correlation matrix of the abundances of tion, and element completeness. Information on the bear, other carnivore, and ungulate remains and lithic artifacts dominantcauses of death and habits can be in- across 17 excavation units in the Middle Pleistocene of feeding layer ferred from the of the Yarimburgaz Cave. mortalitypattern (age structure) Other Lithic fossil population based on eruption, wear, and Bears . t carnivores Ungulates artifacts breakage. Adult sex ratiosand size dimorphismbetween Bears the sexes can be determinedfrom osteometricmeasure- Othercarnivores 0.750 ments. Ungulates 0.722 0.754 Lithic artifacts 0.627 0.625 0.264 Modelsin the Absence of Complete Analogs the spatialproximity of bear bones and stone artifactbe- Cave bears, an extinct subgenus (Spelearctos) of lies a temporaland causal link between them. Ursidae, were versatile enough to inhabit large areas of To addressthese hypotheses, one first needs to know Eurasia during the Middle and Late Pleistocene. Yet (1) whether cave bears were hiberators, and, if so, (2) cave bearshad evolved a specialized dentitionwith large whetherbears would collect bones of prey in dens, and surfacesand powerfulchewing muscles. Their (3) the likelihood that bears would use caves for either multi-cusped,bunodont dentitions emphasized crushing purpose. Constructingtest implications for these and and grinding functions and imply a heavy dietary reli- other propositionsbegins by necessity with information ance on tough, fibrous plant foods (Koby 1940, Kurten on moder bears (Stiner et al. 1996, 1998; Stiner 1998). 1976:11-26, 1998, Stineret al. 1998). Iso- Ursus is of but some of the Baryshnikov deningeri extinct, course, studieshave resultson cave bear behavioraltendencies documented for moder Ursus tope yielded conflicting help diet, however 1994, Hilderbrandet al. 1996, establish what was in the The material (Bocherens possible past. Stineret al. Much of the stems from correlatesof these behaviors can also be modeled over 1998). controversy contradictions that taphonomic and morphologic evi- the long term. Importantaspects of cave bear ecology dence pose to the isotope results (Mattson 1998, Stiner are expressed by the isotopic contents of tooth enamel, et al. 1998). It is likely that cave bear habits varied in patterns of bone damage, skeletal element representa- 46 Ursus 11:1999 response to environmentalcircumstance, but the limits hyaenids deliberatelygather bones at den and rest sites on their abilities to do so remain unknown. Pleistocene (Mech 1970, Kruuk1972, Ewer 1973, Fentressand Ryon cave bears shared with other bears a general tendency 1982) and may amass great quantitiesof prey bones in toward omnivory,but their dental adaptationsreached a some cases (Hill 1980, Binford 1981). Hominids did morphologic extreme and testify to a great dependence essentially the same at residentialsites duringthe Pleis- on seasonal plant foods. This fact, along with the oft- tocene, sometimesin caves. The habitsof bearsare quite notedprevalence of cave bearskeletons in the Pleistocene different, however. Wildlife accounts show that black sedimentsof Eurasiancaves, make a strongcase for cave and brown bears do not normallycarry food of any sort bears as hiberators (Kurten 1958, 1976). to dens, and they consume little food while preparing Although distinct species, modem black and brown their winter beds (McNamee 1984, Rogers 1987:23). bears have much in common biologically, including hi- Bears often mound piles of inedible vegetation in hiber- beration behavior and its nutritionaland reproductive nationdens (McNamee 1984:252-253, Murie 1985:133- contingencies (Garshelis and Pelton 1980; Johnson and 135, Manville 1987, Rogers 1987:20-22, Clevenger Pelton 1980;Nelson et al. 1980;Rogers 1981, 1987;Tassi 1991, Smith et al. 1994), but bones other than those of 1983; Murie 1985; Clevenger et al. 1987; Clevenger unlucky bears generally are not found. 1990, 1991; Hellgren et al. 1990; Clevenger and Purroy The location of a den is secret (Kolenosky and 1991). These basic similaritiespermit a few generaliza- Strathearn1987, Mack 1990, Hayes and Pelton 1994), tions about the relationsbetween bear diet and hiberna- because sleeping bears are vulnerableto attack despite tion behavior, hibernation-related mortality, and the their great size (Kurten 1976, Tietje et al. 1986, Rogers criteria governing den site choice. The metabolic and 1987:53, Ross et al. 1988). Predationon denning bears reproductiveaspects of hibernationin moder bears are by wolves, humans,and other bears is well known, as is thought to be intrinsic (Johnson and Pelton 1980, cannibalismwithin and between bear species (LeCount McNamee 1984:253-257, Watts et al. 1987, Watts and 1987, Mattson et al. 1992). Food debris in and around Jonkel 1988, Hellgrenet al. 1990). We may assume that dens would betray the location of a hibernatingbear to these qualities are unlikely to have been profoundlydif- predatorsand presumablyconflicts with the bear's need ferent in Middle and Late Pleistocene cave bear popula- to remain hidden. The only exception to this generali- tions. The mortality that normally accompanies zation thatI have found is reportedby Rogers (1987:23), hibernationin bears therefore should be comparableas a single case in which a lactatingfemale black bearfound well. a deer carcass near her den in spring and draggedit into the den. It is reasonableto generalize,therefore, that the Do Bears Den in Caves? quantities of non-bear bones that might accumulate in Even if one accepts that cave bears were hibernators, bear dens are nominal, if perceptibleat all, from a pale- we are left to explain their presence in caves. The abili- ontological point of view. ties of modern bears to make artificial shelters are well Yet the bones of a varietyof large mammalsare found known, and excavated dens are the norm in many study in Pleistocene cave sites that also contain the bones of areas(e.g. Rogers 1981, Kolenosky and Strathearn1987, bears. This is true of YarimburgazCave, for example, a Schwartz et al. 1987, Mack 1990, Miller 1990, Hayes fact which is explored in some detail below. Informa- and Pelton 1994, Smith et al. 1994). However, wildlife tion on moder bear behavior nonetheless suggests that studies from Europeand North America show that mod- otherspecies mustbe responsiblefor the presenceof non- ern bears willingly hibernate in natural caves and fis- ursid bones. The presence of scant remains of canids sures where these conveniences are present (Murie and hyaenids in the Yarimburgazspecies profile is quite 1985:133-135, Clevenger and Purroy 1991:113-123, significant in this regard. Hayes and Pelton 1994). The prevalenceof bearremains in Pleistocenecave sedimentsis not necessarilyexplained by heavy use of them over earthendens in the past; lime- BEAR BODY PART PROFILES AND stone caves and fissures are merely better preservation BONE DAMAGEPATTERNS environmentsfor the bones of deceased bears. Bears are likely to contributetheir own remainsto the sediments in hibernationdens, even if new bodies are Do Bears Collect Bones of Prey in Dens? addedonly sporadicallyover many generationsof bears. Carnivore species that rely on natural or excavated In these circumstancesall of a bear's skeletalparts should sheltersseem to divide into 2 behavioralcategories: bone- be presentin an earthenor ,because the bear collectors and non-collectors. Many of the canids and dies in place. The situation is quite different for large INVITEDPAPER * HUMAN-BEARINTERACTIONS INTHE PLEISTOCENE * Stiner 47 mammalbody partsported to the shelterby carnivorous tween the bear remains and other materials in occupants, because these food items are obtained else- YarimburgazCave. The mortality patternof the bears where, are divisible, and are subject to transportcosts. has greaterpotential for distinguishingthe circumstances Body partsnormally are moved to safe havenspiece-meal of bone damage, specifically between huntingand scav- by predatorssuch as wolves and hyenas (Stiner 1991, enging contexts (Stiner 1998). 1994). Although herbivores are the principal prey of large carnivores, inter- and intra-specific aggression Causes of Cave BearMortality nearly guaranteesthat partsof some carnivorecarcasses It is clear from the wildlife literaturethat hibernation will accumulatealong with herbivorebones in the dens dens are places where brown and black bears occasion- and rest sites of hyaenids and canids. If the site is used ally die (Kurten1976; Garshelisand Pelton 1980; Rogers for reproductivepurposes, the skeletons of young canids 1981, 1987). Of course, hibernatingbears may perish or hyaenids may also become part of the paleontologic from violent or nonviolent causes inside dens, just as record,though seldom as complete bodies (Stiner 1994). they do in free-rangingcontexts. Hunting of hibernat- Unless a bear carcass is extensively fed upon, its skel- ing bears is well known among traditionalhuman cul- etal parts should remaininside the cave. However, bear turesof the 19thand 20th centuries(Kurten 1976, Rogers skeletons in den sites may not stay in articulatoryorder, 1981:69, Ross et al. 1988, Binford In Press). Bear hunt- because any bear that later dens in the cave would reno- ing is not a common practice among any of these cul- vate the bedding area (Reynolds et al. 1976, Judd et al. tures and probably could never be so because bear 1986). The scale of this kind of disturbancehinges on populationsare relativelyunproductive (Craighead et al. the frequency with which a shelter is re-used by bears 1976, Glenn et al. 1976, Bunnell and Tait 1981:77). and the rate of sediment accumulation (Kurten 1976, Humans' ability to hunt hibernating bears with tradi- Stiner 1994, Stiner et al. 1996). In the Middle Pleis- tionaltechnology (BinfordIn Press)nonetheless suggests tocene layer of YarimburgazCave, the total bear assem- thathunting must be considereda potentialcause of bear blage computesto a patternof nearlycomplete skeletons, mortalityin dens in the past. but the components of these skeletons are scatteredin a The uniquebiology of moder bears suggests thatrela- nearly random fashion. The remains of ungulates and tively distinctmortality (age structure)patterns will arise non-ursidcarnivores are far less complete and also scat- in hibernationdens from predominantlyviolent versus tered (Stiner et al. 1996). predominantlynonviolent causes. These differences in Gnawing damage on the bear bones, though less fre- mortalitypatterns can be used to test the cave bears' con- quent than on other large mammal remains, suggests nection to prehistorichuman activities in the same shel- additional reasons why bear remains are scattered ter. The net expectation for the cave bear population throughoutthe Middle Pleistocene layer. Denning bears using winter dens should resemble the living age struc- that perish from predationmay be fed upon by their at- ture of the total free-rangingpopulation, because (1) all tackers. But the decaying body of a bear that died from reproducingadult females must hibernate to rear their nonviolentcauses will attractscavengers, often the same cubs, (2) reproductiverates in bearsare low, and (3) other species known to attackbears in other circumstances. It adult bears may also hibernateif the populationis rela- therefore is difficult to distinguish between deaths re- tively vegetarian,as appearsto have been true for cave sulting from attacks on live, dormantindividuals from bears. post-mortemscavenging on the basis of gnawing dam- The living age structureof bears that use hibernation age. Many of the gnawing traces can be attributedto dens each thereforeserves as the model for preda- - or hyena-sized carnivoresand, less often, to larger tion on cave bears by den raiders(Stiner 1998): prime- bears (Fig. 2). aged adult females, old adults, infants, and adolescents The observationsabove raise questions about how the should be affectedrandomly by huntingand thereforebe Yarimburgazcave bears died, but they do not answer representedin the deathassemblage in proportionto their them. Gnawing damage on bear remains-or cut marks living abundances (Fig. 4). Repeated hunting of den- in other cases-is not proof of hunting. Such damage ning bears should be indifferentto the ages of the indi- only indicates consumptionor related use of carcasses. viduals occupying dens for at least 3 reasons. First,there However, bear bones were less frequentlydamaged by is little opportunity for attackers to survey individual carnivoresthan were the remains of ungulates and non- vulnerabilitywhile huntingencrypted prey. Second, adult ursid carnivores,and rodentdamage shows the opposite females normally accompany hibernatingjuveniles in distribution. These observations lend some supportto their first, second, and even third winters of life. Fi- the hypothesis of temporaland causal independencebe- nally, it seems that all members of the denning cluster 48 Ursus 11:1999

40

30-

20

a, )

4/

0 V

I It III IV V VI VII VIII IX

L. I I I J

JUVENILE PRIME ADULT OLD ADULT

enamel worn

dentin exposed

i root pulp cavities e posed

Fig. 4. Age structureof the Yarimburgazcave bear death population(black bars) based on 9 eruption-wearstages for the M2molar (n = 44), compared to an idealized (expected, hatched bars) age structure for a hypothetical living population. are subject to attack (Garshelis 1994), presumablybe- structreal mortalitypatterns with statistical confidence cause young bears are easy victims and their mothersare (Garshelis 1994). likely to become aggressive in tight quarters. Reportsof The alternativeage structuremodel for bear mortality intraspecifickilling are comparativelyfew, but adult fe- in dens is the classic one of attritionaldeath from mal- male bears, often with young cubs, are among the indi- nutrition,disease, and senescence, a patterntypical for viduals that perish during den attacks (on black bears, mammalsin general(sensu Caughley 1977; on bears see Garshelis 1994:3; Rogers 1987). The cumulative mor- reviews by Kurten 1958, Stiner 1998). An attritional tality patternresulting from attackson hibernatingbears deathpattern affects young and old age groupspreferen- should differ from age-dependentmortality in free-rang- tially. Because prime-agedadults are conspicuously un- ing bears because individual characteristicscan be per- der-representedrelative to their living abundance, an ceived in the latter situation and the attackerswill tend attritional mortality pattern displays a bimodal or U- to gravitatetoward young, weak, and old victims. The shaped distributionin histograms. Non-violent mortal- expectationof a relatively nonselective age structurefor ity in moder bears is relatively high toward the end of bears attackedinside dens (though poor in adult males, the hibernationperiod if bears are in poor condition, al- Garshelis 1994, McLellan 1994) is largely theoretical, though starvationis most common in moder popula- however, as too few data are currentlyavailable to con- tions that are not extensively hunted (Kurten 1976; INVITEDPAPER * HUMAN-BEAR INTERACTIONSIN THEPLEISTOCENE * Stiner 49

Garshelisand Pelton 1980;Rogers 1981, 1987;McNamee To summarizethus far, informationon species repre- 1984). Bears awaken early if their energy stores are de- sentation in the Yarimburgazfauna and the patternsof pleted and may, as a last resort, make short forays in bone damage,body partrepresentation, and mortalityfor search of food. Females with cubs tend to stay around the cave bears indicate that most or all of the cave bear dens longer, both in dormancyand following first emer- remains are causally unrelated to hominid presence at gence in spring(Miller 1990). Starvingor ill bearssome- the same site. The bearbone accumulationsresulted pri- times collapse and die in the vicinity of winter dens, marily from nonviolent mortalityassociated with hiber- althoughthis is relativelyrare (Garshelis 1994, McLellan nation over many generations of den use. Attacks on 1994). Occasional deaths of this sort nonetheless can living bears may also have occurred,but this appearsto add up to many bodies over millennia. have been uncommon. Canid and hyaenid contact with The main differencebetween the 2 den mortalitymod- bear carcasses certainlyis indicatedby gnawing damage els is the near absence of prime-aged adults in the andscatological bone containingbear cub andhare (Lepus attritional(nonviolent death) model. To test the hypoth- sp.) remains (see Stiner et al. 1996). Direct contact be- esis for the paleontological cave bear population from tween cave bears and humans at this site is unlikely. YarimburgazCave, cheek teeth were age-scored based The faunal and artifactualevidence shows that at least on their eruptionand occlusal wear status (Fig. 4; Stiner 3 distinctbiological agencies contributedto formationof 1998). The Yarimburgazcave bear sample displays an the Middle Pleistocene faunas: hibernatingbears, non- attritional(U-shaped) death pattern,exemplified by re- ursid carnivores such as wolves, and hominids, in de- sults for the lower second molar (Fig. 4). The mortality scendingimportance. The Yarimburgazfaunas represent results for the Yarimburgazbears generally supportthe overlays of many short-termdeposition cycles, most of nonviolent model of bear deaths and refute the violent which were unconnectedin time. The close spatialasso- model. This result, along with the complete absence of ciations of diverse biogenic refuse from bears, wolves, tool marks or burningdamage on cave bear bones, sug- hyenas, and humans are best explained by slow or un- gest that the bear deaths in the cave were entirely inde- even sedimentationinside the cave. Considerablymore pendent of hominid activities there. Because the can be learnedabout the habits of the Yarimburgazcave hypothesis that most of the bear deaths occurred from bears from other aspects of their skeletons. hunting is refuted, the possibility of a consistent tempo- ral or causal link between cave bears and humansat this site is also refuted. Wolves or hyenas may have been CAVEBEAR DIET: EVIDENCE FROM responsible for a minor fraction of the bear deaths, but TOOTHDAMAGE, ADULT SEX RATIO, muchof the gnawingon bearbones by hyenas andwolves AND STABLEISOTOPES appearsto have occurredfrom scavenging. An interesting detail of the mortality patternfor the AdultSex Ratio in Dens and Its Relation Yarimburgazcave bears is the pronouncedpeak in the to Diet late juvenile age stage III (Fig. 4). The deathpeak prob- Hibernationis a bear's for ably representsthe winter when cubs became indepen- strategy raising exception- altricialinfants and food in winter dentof theirmothers. The patternis similarto a condition ally enduring scarcity (Ewer 1973; Garshelis and Pelton noted in some modern bear populations (Glenn et al. 1980; Rogers 1981, 1987; Tassi 1983; McNamee 1984; Murie 1985; 1976, Bunnell and Tait 1981). While the stage III peak et al. 1992). In moder bears, the is a relative measure and can not be tied to an exact age Clevenger importance of hibernation to both sexes increases with the in real years, it suggests relatively long periods of off- on seasonal foods. And the springdependency, such as is seen among moder brown population's dependence of either sex in dens is a func- (as opposed to black) bears. The transitionto indepen- probability dying partly tion of time in these dent life is associatedwith elevated mortality,often from spent places. Mother bears must hibernatefor several months be- predation by dominant adult males. Cave bears were fore cubs are mobile and Pelton In con- much like moder brown and black bears in this regard. (Johnson 1980). trast,adult males and barrenfemales need hibernate The low peak in the youngest age cohort (stage I) may be only as as food is unavailable and Pelton explained partlyby complete ingestion of small cubs by long (Johnson 1980, Schwartz et al. et predatorsor scavengers (Stiner 1998), but the relative Rogers 1987:20-24, 1987, Hellgren al. Miller Van Daele et al. frequenciesof individualsin cohortsII and III can not be 1990:291, 1990, 1990, Smith et al. Weaver and Pelton One explainedaway, because the cubs would have been much 1994, 1994). key to adult sex ratios in dens larger and their bones more robust. appearsto be winter meat from large game (Picton and Knight 1986) availabledur- 50 Ursus 11:1999 ing the months when bears can not find foliage, plant between 30:100 to 100:100 males to females, but thereis mast, tubers, invertebrates,and most small vertebrates. a tendency toward evenness in adult sex ratios overall The degree to which adults other than pregnantfemales (summarizedin Kurten 1976:76-77). The sex ratio of dependon seasonallyavailable foods thereforeaffects the the Yarimburgazcave bearsfrom Turkey therefore is con- adult sex ratio of the hibernatingpopulation. Males of a sistent with the patternfound for the Europeancave bear relatively herbivorouspopulation should have hiberna- samples. tion times approachingthose of pregnantfemales, less if It is interestingthat the sex ratio of adult cave bears reliable food sources can be found in winter. lies at one extreme of the full range of possible patterns The adult sex ratio of a bear populationalso is subject in Ursus. The mild female bias in the Yarimburgazcave to considerableimbalance independent of who hibernates bearsfalls well beyond the rangetypical of moder bears and who does not, often favoring adult females in general and may best resemblethat of highly omnivo- (Craigheadet al. 1974, Rogers 1987). This means that rous black bears. The sex ratio of the cave bears sug- the standardagainst which bear sex ratios in dens are gests 2 important things about them: (1) the living evaluatedmust be temperedby a considerationof the sex populationwas not poor in adultmales in Turkeyor else- structureof the total adult population from which the where;and (2) these animalsdepended very heavily upon death assemblage derives. The adult sex structureof seasonal food resources. The sex ratio specifically con- ancient cave bear populations can not be known abso- tradictsthe possibility of a regular,heavy emphasis on lutely, but it can be inferredin relative terms from indi- large game. Whethercave bears were firmly at the veg- rect evidence. Bear sex ratios in the populationat large etarianend of the dietarycontinuum in every regionwhere tend to be even at birth (Rogers 1987, Kolenosky 1990, their remains are found is unclear, but a dedication to Miller 1994), but sex ratios normally are skewed in fa- highly seasonal food sources is indicated by the adult vor of females in adulthood (for black bears, Rogers sex ratio, dental morphology, and damage to the teeth [1987] reportsM:F 1:2; for brown bears, Smith [1990] (below). reportsM:F 1:3; Clarksonand Liepins [1994] M:F 1:2). Because black bears are highly omnivorous and may ToothDamage and FeedingHabits hibernatefor up to 7 months (Kolenosky and Stratheam The cave bear's dental specializations include cheek 1987, Schwartz et al. 1987, Miller 1990, Smith et al. teeth with many cusps, large occlusal surfaces in rela- 1994)-among the longest hibernationtimes for bears tion to crown height, and a cranialarchitecture mechani- (Miller 1990)-a similar sex ratio in a paleontological cally suited to grinding and crushing (Kurten 1976:18, populationof any bear species would imply a heavy de- Baryshnikov1998). The extent of occlusal wear in cave pendence on plant and invertebratefoods. A greaterdi- bear molars is exceptional relative to that seen in brown etaryemphasis on large (scavenged or hunted) bears: molar crowns were often completely obliterated or Pleistocene habitatscharacterized by mild winters in- in old cave bears across the many regions where they stead should result in substantially greater differences have been studied. Among elderly cave bears from between pregnant female and adult male hibernation Yarimburgaz,several first and second molars were worn times (Kolenosky and Strathearn1987, Miller 1990, Van to stubsand the pulp cavities fully exposed priorto death. Daele et al. 1990, Weaverand Pelton 1994), and, conse- The damage to cave bear cheek teeth corroboratesrou- quently, the overall probabilitiesthat the bones of each tine consumptionof tough plant materialssuch as nuts, sex will become part of fossil assemblages in caves. hard-coatedseeds, tubers and other roots, berries, foli- con- Modem polarbears, which are fully carnivorousand live age, and the like. Mattson(1998) reacheda similar on in very cold regions,represent the oppositeextreme which clusion about cave bear feeding habits based of and proves the rule: adultmales may not hibernateat all and interspecificmorphometric comparisons modern thereforewould seldom if ever die inside ice dens (M/F Pleistocene bear crania (i.e., a heavy emphasis on root 0:100). grubbing). wear and Using an adaptationof Gordonand Morejohn's(1975) Koby (1940, 1953) noted peculiar breakage cave metric technique, the sex ratio of the adult cave bears patterns on the anterior teeth of some European like- from YarimburgazCave is estimated to have been 73 bears. Many of the Yarimburgazcave bear canines males for every 100 females (Stiner et al. 1998). All wise were brokenor chippedduring life andworn through teeth were also juvenile individualswere eliminatedfrom this compari- continueduse. These and other anterior son because the naturallyeven sex ratio at birthin bears seriously damagedby interstitialwear and abrasivefor- into is boundto push the total populationpattern toward even- age. Of the ageable adult canines that had come in some ness, independentof that population's foraging habits. wear by the time of death, 31% had been broken The sex ratio of cave bear samples from Europe varies way. Some modernbrown bear populationsalso display INVITEDPAPER * HUMAN-BEAR INTERACTIONSIN THEPLEISTOCENE * Stiner 51 peculiar wear on and frequent breakage of the canines similar to that of moder herbivoresand carnivoresliv- (A. Clevenger, Banff National Park, Alberta, Canada, ing in regionsdominated by C3vegetation, such as around personal communication, 1995). the MediterraneanSea today. Marineresources were not The frequency of adult canine breakage is very high an importantcomponent of the diets of the Yarimburgaz relative to that observed by Van Valkenburghand Hertel bears, even though YarimburgazCave was situatedin a (1993) for modem and Late Pleistocene carnivores of near-coastalsetting during the Middle Pleistocene. North America, where this phenomenon was related to The carbon isotope results for the Yarimburgazbears bone feeding and feeding competition. The relative in- closely match measurements from some cave bear cidence of canine breakagein the Yarimburgazcave bear samples in western Europe (Bocherens et al. 1994, sample is amplified somewhat by the advanced ages of Hilderbrandet al. 1996). Bocherens et al. (1994) ob- the adults therein,but this bias does not wholly account tained 613Cvalues on cave bear apatite of -14.8?0.7?/ for the high frequencyof damagedcanines. The biome- (n = 15), remarkablyclose to those of the Yarimburgaz chanical argumentof frequentasymmetrical loading on bears at -15.1?0.60?/ (n = 21). Such concordancesug- the canineteeth (VanValkenburgh and Hertel 1993) prob- gests thatthe dependenceof cave bearson terrestrialfood ably also applies to the Eurasiancave bears, but the di- sources was geographically widespread. However, etary causes of the damage probably do not. The Bocherens et al. (1994) and Hilderbrandet al. (1996) unusuallyhigh frequencyof damageto the anteriorteeth disagreedon the extent of herbivoryin cave bears as evi- in cave bears is distinctive (B. Van Valkenburgh,Uni- denced by nitrogenisotope measurements. Our isotope versity of California,Los Angeles, California,USA, per- results cannot speak directly to this issue, as carboniso- sonal communication, 1998) and suggests habitual topic values are not good discriminatorsof trophiclevel contact with and levering of hard matter. Intraspecific and the proteinsources of nitrogenprobably are not pre- competition may have added to the damage wroughtby served in the Yarimburgazspecimens due to their great feeding. The extreme wear and breakage to the front antiquity. Other sources of information on the teeth (canines and incisors) of adult cave bears from Yarimburgazcave bears do speak to this issues: adult YarimburgazCave suggests thatfood frequentlywas ob- sex ratios in dens and dental evidence contradicta high tainedfrom cryptic sourcesrequiring actions such as dig- degreeof carnivoryin the MiddlePleistocene sample from ging and prying (Baryshnikov 1999, Stiner et al. 1998). Turkeyas well as for other fossil cave bear populations. Isotope Signals in Bear Tooth Enamel Quade, Pigati, and Achyuthan of the University of SEXUALSIZE DIMORPHISMIN CAVE Arizona examined the carbonand oxygen isotopic com- BEARS of the enamel of cave and brownbear teeth from position Moder bearsexhibit comparativelyhigh levels of size Cave et al. to address Yarimburgaz (Stiner 1998) ques- dimorphismbetween the sexes. Grossly analogous dif- tions about such as dietarypreferences, between marine ferences in adult statureand weight have been inferred and terrestrial and whetherthe obtained sources, signals for Pleistocene Ursus species of the northernhemisphere from cave and brown bear teeth differ. Tooth enamel based on the dimensionsof canine teeth and weight-bear- was over other skeletal materialsfor preferred this study ing limb bones (e.g. Koby 1949; Kurten 1958, 1976). on that its grounds dense, coarsely crystalline structure The results on the Yarimburgazcave bears (Stiner et al. tends to an better preserve original paleodietary signal 1998) arebased on measurementsof carpalor wristbones than do the less dense mineral structuresof bone and (pisiform and scapholunate; Stiner et al. 1998), using dentin and van der Merwe (Lee-Thorp 1987, Lee-Thorp Josephson et al.'s (1996) method-of-moments (MoM) et al. et al. et al. 1989, Thackeray 1990, Quade 1992). technique. This technique assumes that the total distri- teeth were 21 from cave bears Twenty-three analyzed, bution of a metric traitis composed of 2 underlyingnor- and 2 from brown bears. The carbonanalyses yielded a mal distributions, one for males and one for females. mean 813Cvalue of -15.1?0.7?/0oand a range of -14.1 to Three moments around the mean of the combined-sex -16.4. The results from the brown bears, althoughfrom distributionare used to estimate the means and the com- 2 were fromthose only teeth, essentiallyindistinguishable mon standarddeviation of the 2 underlyingdistributions. for the cave bears. The indicated an oxygen analyses The scale of measurementmost easily controlledin pa- value of in 6180 and a of -4.9 average -6.5+1.00?/ range leontological studies is a simple linear one and, to some to -9.0. The fossil bear teeth from Cave Yarimburgaz extent, areal measurementscomputed therefrom. If the retained an unalteredcarbon isotopic composition very Yarimburgazsize dataare extrapolatedto yet a thirdscale of measurement,analogous to body volume or mass, adult male cave bears would have been roughly twice the lean 52 Ursus 11:1999 21 a

0 20 -

0

? ? E 19- 0 E 0 53* 18- 4- 0 0~~~~* 0 5

17- a 0 e Ca 0~~~~~~~~ m 0 0S S * * 16- S~~~~~~~~~~ 0~~~~~~~~ 15-

14 I I I I I I I i I 22 23 24 25 26 27 28 29 30 31 32 33 34 Length (mm)

Fig. 5. Scatter plot of length (L)and breadth(B) measurements for rightand left mandibularsecond molars (M2)of bears fromthe MiddlePleistocene layer of YarimburgazCave. The isolated point near the graph intercept is a brown bear molar (n = 1) >3 SD smaller in both dimensions than cave bear teeth (n = 64). Maleand female teeth are continuously distributed. body weight of adult females. This level of sexual size (Fig. 5), falls well below the maximum (mean) propor- dimorphismis relativelyextreme among terrestrial mam- tional size difference that can be expected between the mals, but not unusual among large-bodiedpopulations sexes in any terrestrialmammal (roughly males/females of modem brown bears. We infer from this that their = 1.45-1.50 by a linear standard),even in highly size mating system was similar to that of extant Ursus spe- dimorphic lineages such as bears or the great cies. (Pongidae). The other 64 datapoints form a single clus- ter (Fig. 5), representinga continuousdistribution of male and female cave bear tooth measurements. In addition, HOWDID COEXTANT CAVE AND an informal comparison of the fifth metacarpals of a BROWNBEARS DIFFER? mature cave bear and a mature brown bear from Cave dramaticstruc- Osteometrictechniques demonstrate the presence of 2 Yarimburgaz (Fig. 6) exemplifies turaldifferences between the 2 The brown bear MiddlePleistocene bear species in the MiddlePleistocene species. is somewhatshorter than that of the cave deposits of YarimburgazCave-Ursus (Spelearctos) metacarpal only but the difference in is deningeri, a large cave bear, and U. arctos or brown bear, robusticity great. bear-the formerabundant and the latterrare. Acknowl- The limb allometryof cave and brown bears also dif- fered. Cave bears shorterdistal limb edging potentialtime-averaging effects on fossil assem- possess relatively elements and elements than is in blages, it nonetheless seems that cave and brown bears longer upper typical brown bears Mattson The cave coexisted in many areasof Eurasia. Cave bearsmay have (Kurten 1976, 1998). bearbuild is to have over been more prevalentin some regions and periods, possi- interpreted emphasizedstrength or a whereas brown bears are rela- bly at the expense of brown bears, and vice versa. agility quick gate, Tooth Brown bear teeth from YarimburgazCave are easily tively fast runnersover shortdistances. morphol- testifies furtherto differences brownand cave distinguishedfrom cave bear teeth on the basis of size. ogy among least of which were the The only brownbear tooth in the M2sample, for example bears, not extraordinarymilling INVITEDPAPER * HUMAN-BEAR INTERACTIONSIN THE PLEISTOCENE * Stiner 53 tral Asia; their Pleistocene distributionwas considerably greater than was previously supposed (Kurt6n 1976; Baryshnikov 1989, 1996, 1998; Tchernovand Tsoukala 1997; Stiner et al. 1998). Cave bears coexisted with brown bears and early humansthroughout most or all of this range duringthe Middle and Late Pleistocene. The classic cave bearform appearsin the fossil recordroughly 700,000 years ago and persists in many areas until the , 18,000 years ago, and until the end of the Pleistocene (10,000 years ago) in refugia of the Trans-Caucasian highlands (Baryshnikov 1996). Only the brown bear exists in these and other regions today. With its eventual colonization of American eco- systems, the once exclusively Eurasianbrown bear en- joys the widest distributionknown for any bear species in the late historic period (Kurt6n1971). Cave and brown bears have a long history of contact Fig. 6. Gracile and robust bear fifth netacafpals with humans,but evidence for direct interactionis rela- corroboratingthe presence of 2 bearspecies (U. deningeri rare. Of these most of the ex- and U. inthe Cave et al. tively cases, compelling arctos) Yarimburgaz deposits(Stiner direct 1998). amples of interaction are between humans and brown bears. Images of bears are featuredin the earliest capabilitiesevidenced in the latter. The hyper-developed art of Europe, beginning some 32,000 years ago in the dental architectureof the cave bear is arguablyamong Upper Paleolithic. Kurt6n's(1976:91) claim that brown the most telling in its ecology, as it is expressed bears were more often the subject of early art continues by highly conservativeelements in the evolutionof mam- to hold true, even with the recent discovery of the carni- malian skeletons-molar and premolarteeth. vore-richimagery in ChauvetCave (Chauvetet al. 1996). The evolutionaryeffects of limiting similarityare ex- The relativelyfew images of cave bears that exist in Up- pressed in the bone and dental structuresof cave and per Paleolithic cave art-such as in the French sites of brown bears. It is interesting in light of this that we La Colombiere,,and Chauvet-portray cannot distinguish the diets of coextant cave and brown an animal with an enlargedshoulder hump, accentuated bears in the Yarimburgazsample on the basis of carbon stop in the nasal bridge, and relatively fleshy nose and and oxygen isotope analyses. The isotope data do yield lips. Upper Paleolithic art is not without its fanciful el- some insights on the diets of cave bears, but these ap- ements, but some aspects of these paintingsmay be quite proaches may lack much of the resolution required to accurate. differentiatebetween coextant cave and brown bear di- But what about early cases of bear-artifact associa- ets; perhapsmany of the dietarydifferences exist within tions, datingto of UpperPaleolithic cul- the food categories compared by isotope studies (e.g., tures and the earliest preserved art? Taphonomic among C3plants). In additionto anatomicaldifferences, investigations of YarimburgazCave and other sites re- cave bear hibernation sites tend to be concentrated at fute the possibility of causal links between cave bear and relatively high altitudes, thus accounting for their high hominid occupations (Prat 1988, Andrews and Turner density in the moder countriesof Switzerlandand Aus- 1992, Gargett1996). The close proximityof bearbones, tria (Rabederand Nadel In Press). Cave bears were not stone artifacts, ungulate, and non-ursid bones in the confined to highlands-their remainsalso occur in caves Middle Pleistocene layer of YarimburgazCave reflects at sea level in low frequencies (e.g. Stiner 1994)-but contemporaneityonly on a geologic time scale, not at brown bear remains seem to be more common at lower the scale of mammalianlifetimes or annualcycles. Pre- altitudes, as are human archaeological sites dating to served records of these events were superimposed,one before the Upper Paleolithic period. directlyupon another,because the rate of sedimentaccu- mulation was slow relative to the rate of bone and arti- fact accumulation and because of post-depositional DISCUSSION disturbancesby bedding bears and other digging carni- Cave bears inhabiteda vast area stretchingfrom west- vores. Humanoccupations of the early sites were quite ern Europeto the Urals, Siberia, and into west and cen- ephemeral, short visits separated by long intervals of 54 Ursus 11:1999 absence. Underthese circumstances,early humans would areas except certain highland refugia. It is interesting have providedfew if any deterrentsto den-seekingbears. that the earliest grinding and pounding Despite its long history of use by numerouslarge mam- emerges in UpperPaleolithic cultures, Lilliputian imple- mals, it seems thatYarimburgaz Cave was often an empty, ments gearedfirst to preparingmineral but later quiet place. enlargedand adaptedto processing nuts (and eventually The case of Yarimburgazis one of several windows on cereals) in some regions by 13-10,000 years ago. Late cave bearecology, as well as earlyhumans' relations with Upper Paleolithic humans may merely have expanded them. The findings presentedhere confirm a picture of their foraging niche as it was released from competition cave bears as a relatively herbivorousomnivore, highly by large-molaredcave bears, or as favorableplant com- dependent on seasonal food supplies, including plants, munities expanded with climate change. and probably also invertebratesand small vertebrates. The intensity of human occupations in natural shel- Both sexes of cave bears solved the problem of winter ters, which also increasedafter 18,000 years ago (Gamble food scarcityby hibernating.Male andfemale cave bears 1986), may have affected these shelters' attractiveness were as size dimorphic as large-bodiedpopulations of and availabilityto denseeking bears. We do not know, modem brownbears, suggesting similarmating systems, however, if, perhapsowing to their greatbody size, cave specifically the degree of male-male competition for bears depended more on caves as denning sites than mates. Cave bearswere perhapsmost differentfrom other brown bears. We only know that bones preserve better bears in their ability to extractnutrients from tough, re- in caves than in the open. The necessary links between sistant, and often gritty plant tissues. In so doing, old caves, cave men, and cave bears are largely of a geologi- adults frequentlywore their massive molars to pegs, a cal natureand say little or nothing about the possibility conditionnot unknownin brown bears, but certainlynot of habitationand den sites elsewhere on landscapes. as typical nor as widespread among populations. Al- More significant than the possibility of interspecific though more work is needed on the subject, cave bears competition between early humans and caves bears is appearto have been terrestriallyoriented, even in terri- the peculiarnature of cave bear skeletal structure. Cave tories adjacent to productive seas (Hilderbrandet al. bearswere a specialized,highly derivedtype of bear,aris- 1996). ing from a larger lineage which otherwise emphasized As for cave bears' ecological relations with early hu- generalizedcharacters. The fact that cave bears special- mans, we have had to deal first with how archaeological ized underpredominantly glacial conditions(most of the and paleontologicalrecords form and the possibility that Pleistocene) itself may provide a more compelling ex- separatebiogenic events were recordedsurfacially as one. planation for their at the beginning of the Knowledge of how moder bears behave and die makes modern interglacial(Holocene). The dentition of cave Paleolithic artifact-bearassociations difficult to under- bears is exceptional for a bear. Specialists are particu- stand if taken at face value. Only a cross-disciplinary larly vulnerableto extinction when the conditionsof life perspective, informed by wildlife data on modem ani- become unstable, and those following the Last Glacial mal behaviorand by the natureof sedimentaryprocesses, Maximum, a rapid process of global warming, were allows us to explain this phenomenon. The possibility among the most trying for large Pleistocene mammals. of a cave bear cult among the Neandertalshas been falsi- Whatevereffect these factorshad on cave bears,they did bears fied, and in its place is a more interestingstory of coex- not affect brown bears in the same way. Brown istence between omnivorous lineages whose resource have retained many generalist features and habits, and weath- interestsoccasionally overlapped. this flexibilityno doubtaided their great success in As omnivores,humans' and bears' options for switch- ering the Holocene transition. on modern are im- ing among alternativefood sources have always been Wildlife data, such as that bears, the considerable. While it is possible that humans eventu- portantreferents for zooarchaeologicalresearch on human as well as on the ally drove cave bears to extinction, this is not at all cer- evolutionaryhistory of beings tain. It is not clear,for example,that cave bearscompeted species with whom they may have interacted. Archae- of the directly with humans for food, even if they may have ologists do not have the luxury directly observing our a series competed with humansfor territory. Humanpopulation phenomenathat drive discipline. Through from densities were very low duringmost of the Middle Pale- of inferentialsteps, we extractinformation coarse, The olithic and earlier (Stiner et al. 1999). Humandensities imperfect behavioral and morphological records. on clear increasedduring the UpperPaleolithic period, especially efficacy with which we do so depends expecta- de- following the Last Glacial Maximum (18,000-20,000 tions about the relationsamong key variables,many but which also find years ago), when cave bearswere disappearingfrom most veloped in the contemporaryworld, INVITEDPAPER * HUMAN-BEARINTERACTIONS INTHE PLEISTOCENE * Stiner 55 expression in fossil animal assemblages. The present physiology, and ecology of fossil mammalsas inferredfrom frequentlyinforms research on past. Insightsabout mod- stable carbon and nitrogen isotope biochemistry: em conservationissues may also be gained from the long implications for Pleistocene bears. Palaeogeography, and 107:213-225. view of humans' relations with coextant species. Palaeoclimatology, Palaeoecology BUNNELL,F.L., ANDD.E.N. TAIT.1981. Populationdynamics of bears-implications. Pages 75-98 in C.W. Fowler and T.D. Smith,editors. Dynamics of largemammal populations, John ACKNOWLEDGMENTS Wiley and Sons, New York,New York,USA. I am grateful to M. Pelton (University of Tennessee) CAUGHLEY,G. 1977. Analysis of vertebratepopulations. John for inviting me to participate in the 1998 IBA meetings Wiley and Sons, London, England. and to A. Clevenger (Banff National Park, Wildlife Sec- CHAUVET,J.-M., E.B. DESCHAMPS,AND C. HILLAIRE.1996. Dawn tion) for several years of inspiring correspondence on of art:the ChauvetCave. HarryN. Abrams,New York,New USA. the subject of bears. I also thank T. Fuller and an anony- York, CLARKSON,P.L., AND I.S. LIEPINS. 1994. bear mous reviewer for many constructive comments on an Grizzly population estimate and characteristicsin the Anderson and Horton earlier draft of this paper and F.C. Howell, G. Arsebiik, Rivers area,Northwest Territories. International Conference K. Juell, S. Josephson, J. Quade, J. and H. Pigati, on Bear Research and Management9(1):213-221. for their contributions to the Achyuthan many CLEVENGER,A. 1990. Biologia del Oso Pardo(Ursus arctos) en Yarimburgaz cave bear study. la Cordillera Cantabrica. Memoria de Tesis Doctoral, Universidadde Leon, Leon, Espafia. (In Spanish.) . 1991. The phantombear of the SpanishSierras. Wildlife LITERATURECITED Conservation94:34-45. ANDREWS, P., AND A. TURNER. 1992. Life and death of the , AND F.J. PURROY.1991. Ecologia del Oso Pardo en Westburybears. Annalidi Zoologica Fennici 28:139-149. Espafia. Monographias del Museo Nacional de Ciencias BARTA,J. 1989. Hunting of brown bears in the Mesolithic: Naturales,Madrid, . (In Spanish.) evidence from the Medvedia Cave near Ruzin in Slovakia. , ,_ , AND . 1992. Food habits of brown bears Pages 456-460 in C. Bonsall, editor. The Mesolithic in (Ursus arctos) in the CantabrianMoutains, Spain. Journal Europe. John Donald, Edinburgh,. of Mammalogy73:415-421. BAR-YOSEF,0. 1989. Geochronologyof the LevantineMiddle , ___ , ANDM. SAENZDE BURUAGA. 1987. Status of the Palaeolithic.Pages 589-610 in P. Mellarsand C.B. Stringer, brown bear in the Cantabrian Mountains, Spain. editors. The human revolution. Princeton University, InternationalConference on BearResearch and Management Princeton,New Jersey, USA. 7:1-8. , B. VANDERMEERSCH,B. ARENSBURG, P. GOLDBERG, H. CRAIGHEAD,J.J., F.C. CRAIGHEAD,AND J. SUMNER. 1976. LAVILLE,L. MEIGNEN,Y. RAK, E. TCHERNOV,AND A.-M. TLLIER. Reproductive cycles and rates in the grizzly bear, Ursus 1986. New data on the origin of moder man in the Levant. arctos horribilis, of the Yellowstone ecosystem. CurrentAnthropology 27:63-64. InternationalConference on BearResearch and Management BARYSHNIKOV, G. 1989. Les mammiferes du Paleolithique 3:337-356. Inferieurdu Caucase. L'Anthropologie(Paris) 93(4):813- , J.R. VARNEY, AND F.C. CRAIGHEAD.1974. A population 830. (In French.) analysis of the Yellowstone grizzly bears. Bulletin of the . 1996. Distribution of cave bears in the Pleistocene of MontanaForest and Conservation Experiment Station 40:1- Asia. Journal of VertebratePaleontology 15(supplement 20. 3):17A. EWER,R.F. 1973. The carnivores.Weidenfeld and Nicholson, . 1998. Cave bears from the Paleolithic of the Greater London, England. .In J.J.Saunders, B.W. Styles, andG. Baryshnikov, FARRAND,W.R. 1992. Geoarchaeologyof YarimburgazCave, editors. Quaternary Paleozoology in the Northern Turkey. Pages 19-36 in J. Bermudez, J. Arsuaga, and E. Hemisphere.Scientific papersof the Illinois State Museum, Carbonell, editors. Evoluci6n humana en Europa y los Springfield, Illinois, USA. yacimientos de la Sierrade Atapuerca,Actas, Vol. 1. Junta BINFORD,L.R. 1981. Bones: ancient men and moder myths. de Castilla y Le6n, Leon, Spain. (In Spanish.) Academic Press, New York,New York, USA. FENTRESS, J.C., AND J. RYON. 1982. A long-term study of _ . In Press. Linking ethnographic information on man- distributedpup feeding in captive wolves. Pages 238-260 bear interactions. In T. Tillet and L.R. Binford, editors. in F.H. Harringtonand P.C. Paquet, editors. Wolves of the L'Homme et L'Ours (Man and Bear). GIRPPA Colloque world:Perspectives of behavior,ecology, and conservation. International,Auberives-en-Royans, France. Noyes Publications,Park Ridge, New Jersey, USA. BLACKWELL,B., H.P. SCHWARCZ,N. PORAT,EC. HOWELL,AND F. FOLEY,R. 1984. Puttingpeople into perspective:an introduction ARSEBUK.1990. Electron spin resonance (ESR) dating of to communityevolution and ecology. Pages 1-24 in R. Foley, Ursus teeth from YarimburgazCave, Turkey.Abstracts of editor.Hominid evolution and community ecology. Academic the Geological Society of America 22(3201):A120-121. Press, London, England. BOCHERENS, H., M. FIZET, AND A. MARIOTTI. 1994. Diet, . In Press. The evolutionaryconsequences of increased 56 Ursus 11:1999

carivory in hominids. In C. Stanford and H.T. Bunn, .1953. Modificationsque les ours des cavernes ont fait editors. Early human diet and the role of meat. Oxford subir a leur habitat. Premier Congr6s International de University, Oxford, England. Sp6leologie, Paris,France, Tome IV(4): 15-27. (In French.) GAMBLE,C. 1986. The Palaeolithic settlement of Europe. KOLENOSKY,G.B. 1990. Reproductivebiology of black bears in CambridgeUniversity, Cambridge, England. east-central Ontario. International Conference on Bear GARGETT,R.H. 1996. Cave bears and modem human origins: Research and Management8:385-392. The spatial taphonomy of Pod Hradem Cave, Czech , AND S.M. STRATHEARN.1987. Winter denning of black Republic.University Press of America,Lanham, Maryland, bears in east-centralOntario. International Conference on USA. Bear Research and Management7:305-316. GARSHELIS,D.L. 1994. Density-dependentpopulation regulation KRUUK,H. 1972. The spotted hyaena. University of Chicago, of black bears. Pages 3-14 in M.K. Taylor,editor. Density- Chicago, Illinois, USA. dependentpopulation regulation in black, brown, and polar KUHN,S.L., G. ARSEBUK, AND F.C. HOWELL.1996. A Middle bears. International Conference on Bear Research and Pleistocene lithic assemblage from YarimburgazCave, ManagementMonograph Series No. 3. Turkey.Paleorient 22(1):31-49. , ANDM.R. PELTON.1980. Activity of black bears in the KURTtN,B. 1958. Life and deathof the Pleistocene cave bear:a Great Smokey Mountains National Park. Journal of study in paleoecology. Acta Zoologica Fennica 95:4-59. Mammalogy61:8-19. 1971. The age of mammals.Columbia University, New GLENN,L.P., J.W. LENTFER,J.B. FARO,AND L.H. MILLER.1976. York,New York,USA. Reproductivebiology of female brownbears (Ursus arctos), ___ . 1973. Transberingian relationships of Ursus arctos McNeil River, Alaska. InternationalConference on Bear Linne (brown and grizzly bears). Commentationes Research and Management3:381-390. biologicae, Societas ScientiarumFennica, Helsinki, Finland. GORDON,K.R., AND G.V. MOREJOHN.1975. Sexing black bear . 1976. The cave bear story. ColumbiaUniversity, New skulls using lower canine and lower molar measurements. York,New York,USA. Journalof Wildlife Management39:40-44. LECOUNT,A.L. 1987. Causes of black bear cub mortality. HAYES,S.G., ANDM.R. PELTON.1994. Habitatcharacteristics of InternationalConference on BearResearch and Management female black bear dens in northwestern Arkansas. 7:75-82. InternationalConference on BearResearch and Management LEE-THORP,J.A., ANDN.J. VANDER MERWE. 1987. Carbon isotope 9(1):411-418. analysis of fossil bone apatite. South African Journal of HELLGREN,E.C., M.R. VAUGHAN,R.L. KIRKPATRICK,AND P.R. Science 83:712-715. SCANLON.1990. Serial changes in metabolic correlatesof ___ , ,__ ANDC.K. BRAIN. 1989. Isotopic evidence for hibernationin female black bears. Journalof Mammalogy dietarydifferences between two extinctbaboon species from 71:291-300. .Journal of 18:183-190. HILDERBRAND,G.V., S.D. FARLEY, C.T. ROBBINS,T.A. HANLEY, MACK, J.A. 1990. Black bear dens in the Beartooth Face, K. TITUS,AND C. SERVHEEN.1996. Use of stable isotopes to southcentralMontana. InternationalConference on Bear determinediets of living andextinct bears. Canadian Journal Research and Management8:273-277. black bears in of Zoology 74:2080-2088. MANVILLE,A.M. 1987. Den selection and use by HILL,A.P. 1980. Hyaena provisioning of juvenile offspring at Michigan's northern Lower Peninsula. International the den. Mammalia44:594-595. Conferenceon Bear Researchand Management7:317-322. of extantand JOHNSON, K.G., AND M.R. PELTON. 1980. Environmental MArrSON,D.J. 1998. Diet andmorphology recently on Bear relationships and the denning period of black bears in extinct northernbears. InternationalConference Tennessee. Journalof Mammalogy61:653-660. Research and Management10:479-476. 1992. Cannibalism JOSEPHSON,S.C., K.E. JUELL,AND A.R. ROGERS.1996. Estimating ___, R.R. KNIGHT,AND B.M. BLANCHARD. bears in the by method-of-moments. American and predation on black bears by grizzly Journalof Journalof Physical Anthropology100:191-206. Yellowstoneecosystem, 1975-1990. Mammalogy JUDD,S.L., R.R. KNIGHT,AND B.M. BLANCHARD.1986. Denning 73:422-425. 1994. grizzly bears in the Yellowstone National Park area. MCLELLAN,B. Density-dependentpopulation regulation InternationalConference on BearResearch and Management of brown bears. Pages 15-24 in M. K. Taylor, editor. 6:111-117. Density-dependentpopulation regulation in black, brown, on Bear Research KLEIN,R.G. 1989. The human career: human biological and and polarbears. International Conference Series No. 3. cultural origins. University of Chicago, Chicago, Illinois, and ManagementMonograph USA. MCNAMEE,T. 1984. The grizzly bear. Alfred A. Knopf, New KOBY,F.E. 1940. Les usures seniles des canines d'Ursus York,New York, USA. L.D. 1970. The wolf: The and behavior of an spelaeus et al prehistoire. Verhandlungen der MECH, ecology NaturforschendenGesellschaft in Basel, ,LI:76- endangered species. Natural History Press, Garden City, 95. New Jersey,USA. of brown bears in . 1949. Le dimorphisme sexuel des canines d'Ursus MILLER, S.D. 1990. Denning ecology with a black arctos et d'Ursus spelaeus. Revue suisse de Zoologie. southcentralAlaska andcomparisons sympatric on Bear Research 56:675-687. (In French.) bearpopulation. International Conference INVITEDPAPER * HUMAN-BEARINTERACTIONS INTHE PLEISTOCENE * Stiner 57

and Management8:279-287. 18:455-482. . 1994. Black bear reproductionand cub survivorshipin . 1992. Overlappingspecies "choice" by Italian Upper south-central Alaska. International Conference on Bear Pleistocene predators.Current Anthropology 33:433-451. Research and Management9(1):263-273. _ . 1993. Modem human origins-faunal perspectives. MURIE,A. 1985. The grizzlies of Mount McKinley. University Annual Review of Anthropology22:55-82. of Washington,Seattle, Washington,USA. . 1994. Honoramong thieves: A zooarchaeologicalstudy NELSON, R.A., G.E. FOLK, E.W. PFEIFFER,J.J. CRAIGHEAD,C.J. of Neandertalecology. PrincetonUniversity, Princeton, New JONKEL,AND D.L. WELLIK.1980. Behavior, biochemistry, and Jersey,USA. hibernationin black, grizzly, and polar bears. International . 1998. Mortalityanalysis of Pleistocene bears and its Conferenceon Bear Researchand Management4:284-290. paleoanthropologicalrelevance. Journal of HumanEvolution PICTON,H.D., ANDR.R. KNIGHT.1986. Using climate data to 34:303-326. predict grizzly bear litter size. InternationalConference on . In Press. Camivory, coevolution, and the geographic Bear Research and Management6:41-44. radiation of the genus Homo. Journal of Archaeological PRAT,F. 1988. Les Ursides de la grotteVaufrey. Pages 291-318 Research. in J.-P. Rigaud, editor. La Grotte Vaufrey: _ , H. ACHYUTHAN,G. ARSEBUK,F.C. HOWELL,S. JOSEPHSON, paleoenvironnement, chronologie, activites humaines. K. JUELL,J. PIGATI,AND J. QUADE. 1998. Reconstructing cave Memoires de la Societe PrehistoriqueFrancaise, France, bearpaleoecology from skeletons:a cross-disciplinarystudy Tome XIX, C.N.R.S. (In French.) of MiddlePleistocene bears from Yarimburgaz Cave, Turkey. QUADE, J., T.E. CERLING, J.C. BARRY, M.E. MORGAN, D.R. Paleobiology 24(1):74-98. PILBEAM, A.R. CHIVAS, J.A. LEE-THORP, AND N.J. VAN DER , G. ARSEBUK,AND F.C. HOWELL.1996. Cave bears and MERWE.1992. A 16Marecord of paleodietusing carbonand Paleolithicartifacts in YarimburgazCave, Turkey:dissecting oxygen isotopes in fossil teeth from Pakistan. Chemical a palimpsest. Geoarchaeology 11(4):279-327. Geology (Isotope Geoscience Section) 94:183-192. , N.D. MUNRO, T.A. SUROVELL, E. TCHERNOV, AND O. RABEDER,G., AND D. NAGEL.In Press. Ursid chronology of BARYOSEF.1999. Paleolithic population growth pulses Austriancave faunas. In T. Tillet and L.R. Binford,editors. evidenced by small game exploitation. Science 283:190- L'Homme et l'ours (Man and Bear), GIRPPA Colloque 194. International,Auberivesen-Royans, France. TASSI,F. 1983. L'orso bruno Marsicano.Panda 12 (Estratto), REYNOLDS,H.V., J.A. CURATOLO,AND R. QUIMBY. 1976. Denning Parco Nazionale d'Abruzzo, Italia. (In Italian.) ecology of grizzly bearsin northeasternAlaska. International TAYLOR,M.K., EDITOR.1994. Density-dependent population Conferenceon Bear Researchand Management3:403-409. regulation in black, brown, and polar bears. International ROGERS,L.L. 1981. A bear in its lair. Natural History 90:64- Conferenceon Bear Researchand ManagementMonograph 70. Series No. 3. 1987. Effects of food supply and kinship on social TCHERNOV,E., ANDE. TSOUKALA.1997. Middle Pleistocene (early behavior,movements, and populationgrowth of black bears Toringian) carnivore remains from northern Israel. in northeasternMinnesota. Wildlife Monographs97. QuaternaryResearch 48:122-136. Ross, P.I., G.E. HORNBECK,AND B.L. HOREJSI.1988. Late denning THACKERAY,J.F., N.J. VANDER MERWE, J.A. LEE-THORP,A. SILLEN, black bears killed by grizzly bear. Journalof Mammalogy J.L. LANHAM, R. SMITH, A. KEYSER, AND P.M.S. MONTEIRO. 69:818-820. 1990. Changes in carbon isotope ratios in the late SANTEL,W., AND W. VONKOENIGSWALD. 1998. Preliminary report recordedin therapsidtooth apatite. Nature 347:751-753. on the middle Pleistocene small mammal fauna from TIETJE,W.D., B.O. PELCHAT,AND R. RUFF.1986. Cannibalismof Yarimburgaz Cave in Turkish Thrace. Eiszeitalter u. denned black bears. Journalof Mammalogy 67:762-766. Genegwart48:162-169. VALLADAS, H., J.L. REYSS, J.L. HORON, G. VALLADAS, O. BAR- SCHWARTZ,C.C., S.D. MILLER, AND A.W. FRANZMANN. 1987. YOSEF, AND B. VANDERMEERSCH.1988. Thermoluminescence Denning ecology of threeblack bear populationsin Alaska. dating of "Proto-Cro-Magnon"remains from InternationalConference on BearResearch and Management Israel and the origin of moder man. Nature 331:614-616. 7:281-291. VAN DAELE, L.J., V.G. BARNES, JR., AND R.B. SMITH. 1990. SMITH,B.L. 1990. Sex weighted point system regulatesgrizzly Denning characteristicsof brown bears on Kodiak Island, bear harvest. InternationalConference on Bear Research Alaska. InternationalConference on Bear Research and and Management8:375-383. Management8:257-267. SMITH, M.E., J.L. HECHTEL,AND E.H. FOLLMANN. 1994. Black VANDERMEERSCH,B. 1989. The evolution of modem humans: bear denning ecology in interior Alaska. International recent evidence from southwest Asia. Pages 155-164 in P. Conference on Bear Research and Management9(1):513- Mellars and C.B. Stringer,editors. The human revolution: 522. Behavioural and biological perspectives on the origins of STINER, M.C. 1990. The use of mortality patterns in modern humans, Volume 1. Edinburgh University, archaeological studies of hominid predatory adaptations. Edinburgh,Scotland. Journalof AnthropologicalArchaeology 9:305-351. VAN VALKENBURGH,B., AND F. HERTEL.1993. Tough times at La 1991. Food procurementand transportby human and Brea: tooth breakage in large carnivores of the Late non-human predators.Journal of Archaeological Science Pleistocene. Science 261:456-459. 58 Ursus 11:1999

WATTS,P.D., ANDC. JONKEL.1988. Energetic cost of winter conditions.Physiological Zoology 60:687-691. dormancyin grizzly bear.Journal of Wildlife Management WEAVER, K.M., AND M.R. PELTON. 1994. Denning ecology of 52:654-656. black bears in the Tensas River Basin of Louisiana. , N.A. 0RITSLAND,AND R.J. HURST. 1987. Standard InternationalConference on BearResearch and Management metabolic rate of polar bears under simulated denning 9(1):427-433.