Sexual dimorphism in limb bones of Late Pleistocene cave bear. (Austria and ltaly)

Transcription

1 Bollettno della Socetà Paleontologca Italana 37 (1), 199 ISSN Modena, Luglo 199 Sexual dmorphsm n lmb bones of Late Plestocene cave bear ( Ursus spelaeus, Carnvora, Mammala) from three caves n Eastern Alps (Austra and ltaly) Chrstan REISINGER Johann HOHENEGGER lnstrute of Paleontology Unversty ofvenna, Austra KEY WORDS - Cave bear, Ursus spelaeus, Sexual dmorphsm, Herarchcal cluster analyss, Dscrmnant analyss, One-way analyss of varance, Late Plestocene. ABSTRACT - The length and wdth of cave bear (Ursus spelaeus) stylo- and zeugopoda were measured. The matera comes from three Late Plestocene alpne caves: Conturnes Cave n Northern Italy, Gamssulzen Cave n Upper Austra, and Herdengel Cave n Lower Austra. Wth the help of herarchcal cluster and dscrmnant analyss, all of the elements under study may be classfed nto two groups. Sze dffèrences between lmb bones are nterpreted n terms of sexual dmorphsm. Such dmorphsm ofsze may be observed n all recent bear speces. Ths paper argues aganst a "Hgh-Alpne small form" far the remans of the Conturnes Cave as tested by one-way analyses of varance. ZUSAMMENFASSUNG - [Geschlechtsdmorphsmus an Extremtatenknochen jungplestozaner Hohlenbaren (Ursus spelaeus, Carnvora, Mammala) aus dre ostalpnen Hohlen Osterrechs und ltalens] - Verschedene Langen- und Bretenmajìe von Stylo- und Zeugopoden des Hohlenbaren (Ursus spelaeus) wurden vermessen. Das Matera stammt aus dre alpnen Hohlen des jung-plestozdns: der Conturneshohle n Nordtalen, der Gamssulzenhohle n Oberosterrech und der Herdengelhohle n Nederosterrech. Mt Hlfe herarchscher Clusteranalysen und Dzskrzmmanzanalysen konnten alle untersuchten Elemente m zwez Gruppen untertezlt werden. Grof?endfferenzen zwzschen den Extremtatenknochen werden als Sexualdmorphsmus nterpretert. En derartger Groj!enuntersched kann be allen rezenten Barenarten beobachten werden. Deser Artkel wendet sch gegen de n der Lteratur angefohrte "Hochalpne Klenform"for de Conturneshohle. RIASSUNTO - [Dmorfsmo sessuale n ossa artcolar d orso delle caverne ( Ursus spelaeus, Carnvora, Mammala) da tre caverne delle Alp orental (Austra e Itala)) - Sono state msurate la lungezza e largezza d stylo- e zeugopoda d orso delle caverne. Il materale provene da tre grotte alpne del tardo Plestocene: Grotta d Conturnes n Itala settentronale, Grotta d Gamssulzen n Austra settentronale e Grotta d Herdengel n Austra merdonale. Le anals utlzzate <herarchcal cluster> e <dscrmnant analyss>, permettono d classfcare n due ;;rupp tutt gl esemplar esamnat. Le dverse dmenson delle ossa artcolar sono nterpretate n termn d dmorfsmo sessuale. Tale dmorjzsmo nelle dmenson può essere osservato n tutte le spece recent d Orso. Questo artcolo pone n dscussone la "pccola forma Alto alpna" proposta n letteratura per la Grotta d Conturnes. INTRODUCTION In the Ursdae the males are always larger than females (Stubbe, 1993, p. 249). Sex-specfìc sze dfferences can be small (-20%) n some speces, but values may also exceed 50% (Strlng, 1993). Sexual dmorphsm s extreme n the polar bear ( Ursus martmus) and brown bear ( Ursus arctos), confìrmng the correlaton wth absolute body sze (Kurtén, 197, p. 52). In addton to ths sex-determned varaton, bears can also show ntraspecfìc spatal sze varaton. Brown bears n Scandnava are on average much smaller than the eastern Sberan and Alaskan representatves (Kurtén, 1955). Furthermore, bear fossls may demonstrate evolutonary sze ncrease. Analyses of sze dfferences n cave bears are numerous. Studes of cannes have shown evdence of sexual dmorphsm (Bachofen-Echt, 1931; Rode, 1935; Koby, 1949; Kurtén, 1955). In regard to postcranal elements, Kurtén (197, p. 70) generally remarked that "... the sexual dmorphsm n sze s, n fact, seen n all of the body. Almost any skeletal dmenson wll show a bmodal frequency curve of the same type as the cannes (the cheek teeth are an excepton)". Sze dfferences n postcranal elements of cave bears were also studed (e.g., Torres Pérez- Hdalgo, 19; Kunst, 1992; Grandal d'anglade, 1993). Koby (1950) shows the ranges of some long bone elements and compared them wth data dscussed n the lterature. Generally, the average lengths of male cave bear lmb bones are reported to be about 15% longer than female ones. Therefore, the large lmb bones as the best ndcators of body sze are dagnostc far sexual dmorphsm (Resnger, 1995). Statstca! methods offer the possblty to prove dfferences n body sze of foss! anmals and demonstrate the extent of varaton. Unvarate frequency dstrbutons can to a certan degree descrbe sexual dmorphsm as well as the sex rato. However, the smultaneous comparson of a varable set leads to more accurate nterpretatons. Multvarate statstca! methods are used n palaeontology, bology, and

2 0 C. REISINGER, ]. HOHENEGGER archaeology n order to ascertan speces and determne sex, such as n recent and foss! prmates (Hohenegger & Zapfe, 1990). Other applcatons nclude the classfcaton of bones n respect to dfferentate between wld versus domestcated anmals, or berween the fore versus back legs of an ndvdua (Benecke, 197). LOCATION AND MATERIAL Ths study s based on measurements of adu!t cave bear stylo- and zeugopods (humerus, radus, ulna, femur, tba, fbula) from the three alpne caves Herdengel (e.g., Rabeder & Mas, 195), Gamssulzen (Khtreber, 1995; Rabeder & Wechenberger, 1995), and Conturnes (Rabeder, 1991). CONTURINES CAVE (CO) Ths cave s stuatec n the Dolomtes, 40 km to the east of Bolzano, ltaly. lt s the hghest cave bear ste n the Alps (2,775 m). Radometrc datng of bones by the uranum-seres method yelded ages from 44,000 to 115,000 years BP (Rabeder, 1995). The dentfcaton labels BC and CU depcted n Text-fg. /1 to /5 mark on the one hand the matera belongng to the Offce for the Preservaton of Hstorcal Monuments n Bolzano, ltaly (BC), and on the other the matera from excavatons by Rabeder (CU). No repons can be found about the stratgraphcal layers. GAMSSULZEN CAVE (GS) Ths cave s locatec 5 km to the west of the vllage Sptal am Pyhrn, Upper Austra (Rabeder, 1995). Radocarbon data ndcate ages from 25,000 to 3,000 years BP (Nagel & Rabeder, 1991; Letner- Wld & Steffan, 1993). The Gamssulzen Cave contans the smallest amount of long bones from the three caves. lts alttude (1,300 m) s ntermedate between the other two caves. No nformaton about stratgraphcal layers are avalable. Unversty ofvenna, Austra. The cave bear remans from the Conturnes Cave are stored at the Offce for the Preservaton of Hstorcal Monuments n Bolzano, ltaly. METHODS The man problem n analysng stylo- and zeugopods of cave bears s the small number, together wth ther fragmentaton. Remans from the Herdengel and Conturnes Caves are relatvely numerous and well rreserved, provdng an adequate bass for bometrca studes. Dfferent lengths and wdths were measured on the stylo- and zeugopods of adult cave bears usng a dal cal per as ndcatec n Dresch (192). The scale s n mllmetres. G. K. Kunst (Unversty of Venna) treated bones from the Gamssulzen and Herdengel Caves, the fìrst author (C.R.) nvestgatec the Conturnes remans. Dfferences n performng measurements were tested and the results correlate. The followng four dstances were measured on the humerus (Text-fg. 1; abbrevatons n parentheses): 'greatest ( GL), 'physologcal length' (PL), 'dsta wdth (DW), and 'wdth of trochlea humer' (WT) The sx dstances on the radus are (Text-fg. 2): 'greatest length '(GL), 'physologcal length' (PL), 'dsta wdth' (DW), 'proxmal wdth' (PW), 'length of caput rad' (LCR), and 'wdth of caput rad' (WCR). Seven dstances were taken on the ulna (Text-fìg. 3): 'greatest length' (GL), 'dsta wdth' (DW), 'proxmal wdth' (PW), 'smallest sagttal daphyss dameter' (SDD), 'smallest transversal daphyss dameter' (TDD), 'dstance between processus anconaeus and tuberculum olecran' (HPAT), and 'dameter of the ncsura trochlears' (DIT) HERDENGEL CAVE (HD) Ths cave s locatec 2 km to the west of Lunz am See, Lower Austra, at 70 m above sea leve! (Rabeder & Mas, 195). Most of the bone remans have been dated by radocarbon and uranum-seres methods, yeldng ages between 3,000 to 7,000 years BP (Letner-Wld & Steffan, 1993; Letner-Wld, Rabeder & Steffan, 1994; Rabeder, 1995). However, two long bones (1 humerus and 1 tba) from the lowest stratgraphcal layers show ages between 12,000 to 13,000 years BP accordng to the uranum-seres method, whch ndcate the Late Rss. The matera from the Herdengel and Gamssulzen Caves s deposted at the lnsttute of Palaeontology,.J c. DW.J C!l Texr-fìg. 1 - Humerus measuremenrs.

3 SEXUAL DIMORPHISM IN LIME BONES OF LATE PLEISTOCENE CAVE BEAR 1 PW WCR slb PW r DW Text-fìg. 2 - Radus measurements. PW DW Text-fìg. 4 - Femur measurements. o.. :e PW Text-fìg. 3 - Ulna measurements. DW Text-fìg. 5 - Tba measurements. Four dstances were taken from the femur (Textfg. 4): 'greatest length' (GL), 'length between trochanter major and epcondylus' (GLT), 'proxmal wdth' (PW), and 'dsta wdth' (DW). The followng dstances characterse the tba (Text-fg. 5): 'greatest length' (GL), 'proxmal wdth' (PW), 'transversa daphyss dameter' (TDD), and 'wdth of cochlea tbae' (WCT). Inference statstca! methods allow to dstngush whether frequency dstrbutons are facrual or accdental (Borrz, 1993; Sokal & Rohlf, 1995). Therefore, the data were treated applyng the follow- ng statstca! methods: 'Frequency dstrbuton dagrams' characterse the sngle varables and nferences were tested usng the 'one-way analyss of varance' n the parametrc case, and the 'Kruskal-Walls test' n the nonparametrc case. Multvarate frequency dstrburons were ganed by 'herarchcal cluster analyss' as the descrptve statstca! method, and nferences were tested by 'dscrmnant analyss'. Snce the chosen classfcaton method (Ward's method on standardzed varables) s based on the 'mnmum varance crteron', ths analyss creates multvarate norma dstrbuted classes, whch can be subsequently tested by parametrc 'dscrmnant analyss' based

4 2 C. REISINGER, j. HOHENEGGER Source d.f. S.S. M.S. F-Prob. HUMERUS (jemafe) Betwee11 Grp Statstca! parameters: Wth11 Grp Cave 11 X sd se Tota! CO GS HD RADJUS (male) Tota! Statstca! parameters: Cave 11 X sd se Levene test far homoge11ery of vara11ces: Srarstc d.f. 1 d.f. 2 2-tal Sg Source d.f. S.S. M.S. F-Prob. Betwee11 Grp Wth11 Grp Tota! No11parametrc test far dffere11ces of populatons: Kruskal-Walls Ch-Square d.f. Sg CO GS l HD Tota! Leve11e test far homoge11e ry of vara11ces: Statstc d.f. 1 d.f. 2 2-tal Sgn Source d.f. S.S. M.S. F-Prob. Berwee11 Grp Wth11 Grp Tota! HUMERUS (male) Statstca! parameters: Cave 11 X sd se CO GS HD Tota! Leve11e test far homoge11ery of vara11ces: Statstc d.f. 1 d.f. 2 2-tal Sg11. ULNA (jema!e) Statstca! parameters: Cave 11 X sd se CO GS HD Tota! Leve11e test far homoge11ery of vara11ces: Statstc d.f. 1 d.f. 2 2-tal Sg Source d.f. s.s. M.S. F-Prob. Source d.f. S.S. M.S. F-Prob. Betwee11 Grp Betwee11 Grp Wth11 Grp Wth11 Grp Tota! Tota! RADIUS (jemale) No11parametrc test far dffere11ces of populato11s: Kruskal-Walls Ch-Square d.f. Sg11. Statstca! parameters: Cave 11 X sd se CO ULNA (male) GS HD Statstca! parameters: Tota! Cave 11 X sd se CO Leve11e test far homoge11e ry of vara11ces: GS Statstc d.f. 1 d.f. 2 2-tal Sg11. HD Tota! Tab. 1 - contnua

5 SEXUAL DIMORPHISM IN LIMB BONES OF LATE PLEISTOCENE CAVE BEAR 3 Levene resr for homogenery of varances: So uree d.f. s.s. M.S. F-Prob. Statsrc d.f. 1 d.f. 2 2-tal Sgn. Between Grp Wthn Grp So uree d.f. S.S. M.S. F-Prob. Tota! Between Grp Wrhn Grp TIBIA (female) Tota! Statstca! paramerers: Cave n X sd se Nonparamerrc test for dfferences of populatons: Kruskal-Walls Ch-Square d.f. Sgn. CO GS HD FEMUR (jèmale) Srarsrcal paramerers: Cave n X sd se Tota! Levene resr for homogenery of varances: Srarsrc d.f. 1 d.f. 2 2-tal Sgn CO GS So uree d.f. s.s. M.S. F-Prob. HD Between Grp Tota! Wthn Grp Levene test for homogenery of varances: Statstc d.f. 1 d.f. 2 2-tal Sgn. Tota! TIBIA (male) So uree d.f. S.S. M.S. F-Prob. Statstca! parameters: Between Grp Cave n X sd se Wthn Grp CO Tota! GS 1 HD FEMUR (male) Tota! Statstca! parameters: Levene test for homogenery of varances: Cave n X sd se Statstc d.f. 1 d.f. 2 2-tal Sgn. CO GS HD So uree d.f. S.S. M.S. F-Prob. Tota! Between Grp Wthn Grp Levene test for homogenery of varances: Statstc d.f. 1 d.f. 2 2-tal Sgn. Tota! Tab. 1 - One-way analhss of varance. Descrptve statstcs. Test of homogenery of varance. Non-paramerrc test (Kruskal-Walls) for three cases on y. on canonca! correlaton (e.g., Blackth & Reyment, 197 1; Bortz, 1993; Evertt, 1993; Sokal & Rohlf, 1995; SPSS, 193). Al calculatons and drawngs were done usng the software package SPSS for Wndows (verson.1.3.). UNIVARIATE ANALYSES RESULTS The hstograms of fare lmbs (Text-fg. ) show clear bmodal dstrbutons for al bones n each cave.

6 4 C. REISINGER,]. HOHENEGGER Whle small (ncerpreted as female) bones are pre- However, the hstogram of femora 7) from domnane n Conturnes Cave (and panly n the Herdengel Cave clearly shows a bmoda dstrbu- Gamssulzen Cave as well), larger (nterpreted as male) ton, where the number of male bones s larger. On bones domnate n the Cave. The number the other hand, the Conturnes and Gamssulzen of humer (n = 9) from Gamssu zen s relatvely small, Caves show fewer male femora. In al hstograms but the hstogram fìts very well nco the common from the three caves, however, the comparson of trend vsble n the other caves. In partcular, the modes ndcates clear congruence. Despte the relamodes n al three hstograms Grove the extent of tvely large number of tbae (except from dmorphsm. Besdes the vrtual y dentca! poston Gamssulzen Cave), the frequency dstrbuton of of males the varances wthn sexes are also congruent. ther greatest (Text-fìg. 7) demonstrates a less In genera, sexual dmorphsm of lmb bones s best clear bmodal strbuton than the prevous elendcated by the greatest length of humerus, radus mencs. The Concurnes and Gamssulzen Caves ffjan and ulna. show a clear surplus of female tbae. Complete 1bu- The stuaton s less clear n the hnd lmbs. lae are rarer than any other long bones. Owng to Humerus Var. Wlks' F-rest Sgnf. GS, n=l9 PW CO, n=l GL HPAT PL GL WT DW DW TDD SDD GS, n= DW DIT WT GL PL HD, n=l9 PW HPAT DIT HD, n=24 GL GL PL SDD DW TDD WT DW Radus Var. Wlks' F-rest Sgnf. Femur Var. Wlks' F-test Sgnf. CO, n=21 DW CO, n=23 GL GL PW PL DW PW GLT LCR WCR GS, 11 =7 GLT GL GS, n=13 WCR DW DW PW PL LCR HD, n=32 PW PW GLT GL GL DW HD, 11=24 DW PW LCR Tba Var. Wlks' F-rest Sg11f. GL PL CO, n=32 GL WCR PW WCT TDD Ulna Var. Wlks' F-rest Sg11f. HD, n=35 PW CO, n=30 SDD GL DIT WCT PW TDD TDD DW GL HPAT Tab. 2 - Test for sexual dmorphsm n lmb bone measuremenrs of 3 cave stes usng one-way analyss of varance.

7 ;"Ì &, ((Q C\ Humerus Radus Ulna a.:; "'...,., o a.:; (1) ::; ;:;- ::; C1Q 5- es.e...,., o O:rtures ca.e (N=2'2) '.J l J O:rtures ca.e (N=31) 200 ZIO O:rtures ca.e (N = 33) 12 3Xl '.J t-... b a è:;; (1) ' c:r c:r o :: C?arnstlZ1 ca.e (N=9) Garsst..ZEll ca.e (f\1=19) 12 CHrssUzE11 ca.e (N=21) t-... b;; b:l a Vj 2 &;: t;j '.J ) J (N=a:ì) 200 ZlO (l) (N=44) 3Xl '.J 400 ca.e (N=25) 12 t:.1 V, Cl Q Q '.J ) J :,_._,,_,..,_,....,...._,...-L-..--L ZIO Xl m '.J o Vl

8 X &, C(Q '-.J Femur Tba Fbula o G'\ {1Q..., o O.Q ";:; Qrtl.lres ca.e (N=25) Qrtl.lres ca.e (N=3) Qrtl.lres ca.e (N=22) ;:;-- ::> {1Q 3- o.e o..., ::r 5 O- ' 0-0'" o :: '.l CffssuZ1 ca.e (N=) J CffssUZE!'l ca.e (N=) 220 ZJO J CffssUZ1 ca.e (N=9) n <.?: ':-.. gs : J ZJO J 200 'ZTO Gì ::t (N=39) :._,..., '.l (N=4) J ca.e (N=17) ctjb n n 220 ZJO J

9 SEXUAL DIMORPHISM IN LIME BONES OF LATE PLEISTOCENE C4VE BEAR 7 ther rod-shaped desgn, fbulae tend to fragment more easly than other lmb bones. Nevertheless, there are more female bones n Conturnes and Gamssulzen, whle male fbulae are more numerous n Herdengel. These sex ratos correspond very well to those of the tbae because fbula length drectly correlates wth tba n the same ndvdua!. MULTIVARIATE ANALYSES Al lmb bones were examned usng herarchcal cluster analyss, except of the fbula. The dendrograms (Text-fg. / 1 to /5) of every ste show a clear separaton of the fve lmb bones nto two classes, the one s nterpretable as female, the other as male ndvduals. One-way analyses of varance confrm the sgnfcant dfferences between sexes far all lmb bones n each cave (Table 2). Accordng to the dendrograms of humer (Textfg. /1), samples from the three caves dfferentate very clearly nto two groups correspondng to sex. Even the small sample from Gamssulzen Cave (n = ) fallows ths trend. Dendrograms of rad from Conturnes, Gamssulzen and Herdengel Cave also allow a clear separaton nto rwo groups. Ths confrms the concluson from unvarate comparson, that rad are sutable to dstngush sexual dmorphsm. The dstncton of sexes usng the ulnae s smlar to that of the rad, as the two bones are functonally and postonally connected. Furrhermore, ulnae and rad are more common n all stes than other lmb bones. The femora and tbae dendrograms also allow a clear dstncton nto sexes. Note that only one male tba from Gamssulzen Cave can be separated (Text-fg. /5) usng cluster analyses. The sex rato obtaned by cluster analyss also corresponds to the results of unvarate analyss (Text-fgs and 7). Cave bears from Conturnes and Gamssulzen Caves clearly show more female lmb bones. In Herdengel Cave, male bones are predomnant or the sex rato s relatvely balanced (radus). Dscrmnant analyses based on the groups obtaned by cluster analyss strengthen the clear dfferentaton nto sexes as obtaned by unvarate analyses. Ths s well represented by frequency hstograms of dscrmnant scores whch are smlar to the unvarate frequency dstrbutons (Text-fg. 911 to 9/3). Addtonally, the canonca! dscrmnant functons far each cave and skeleta! element are hghly sgnfcant (p < 0.05). Even the small sample sze of humer at Gamssulzen (e.g., only 5 female and 3 male) or male rad from the same cave (number of 4) stll reveals the man trend of a clear classfcaton nto 2 groups (Text-fg. 9/1). The number of male ulnae from Gamssulzen (= 3), the male femora from Conturnes (= 3) and the female and male femora from Gamssulzen Cave ( = 3 and 4, respectvely) s also relatvely small (Text-fg. 9/2), but dscrmnant analyses can be calcu!ated on the bass of at least three ndvduals per class. Therefare, ths analyss was appled for tbae only at Conturnes and Herdengel Caves (Text-fg. 9/3), because only one case was observed n the male subset of Gamssulzen Cave sample. Dfferences berween sexes are very small n the case of ulnae from Conturnes and tbae from Herdengel Cave, but ther dstncton nto rwo groups s nevertheless clear. Dscrmnant analyses also allows to test the results of cluster analyss by postonng each bone to defned classes accordng to probabltes. Ths type of classfcaton resulted n a rate of 0% correctly desgnated skeletal elements, thus msclassfcaton rate s 0% far every long bone at each cave. Addtonally, the canonca! dscrmnant functon far each cave and skeleta! element show no sgnfcance (p = 0.05). SEX RATIO Sex ratos of the fve lmb bone elements are calculated for each cave by relatng the counted numbers berween the rwo classes as obtaned by cluster analyss. Dfferences n sex ratos berween the Herdengel and Conturnes Cave are obvous. Whle n the hgh alpne cave the females domnate, whch s confrmed by the range of female/male-ratos between 2.0 and.7, the males domnate n the deepest cave (female/male-rato from 0.17 to 0.4) except the radus, where both sexes are equally represented. The sex rato for the Gamsulzen Cave bears s smlar to that of the Conurnes Cave wth a domnance of small-szed bones, despte the low content of cave bear remans. Sex ratos of humerus data: Cave CO GS HD n 1 24 Sex ratos of radus data: Cave CO GS HD n Sex ratos of ulna data: Cave CO GS HD sex rato (female/male) 13: 5 or : 3 or 1.7 : 1 or 0.33 sex rato (female/male) 1 : 5 or : 4 or : 11 or 1.1 sex rato (female/ male) 22 : or : 3 or 5.33 : 13 or 0.4

10 C. REISINGER, J HOHENEGGER C A S E cu 13-2 BC 172 BC 327 BC 331 BC 335 BC 329 se 314 BC 32 BC 31 BC 131 BC 33 BC 2 BC 170 BC 315 BC 10 cu BC 330 BC 337 Rescaled Dstance Cluster Combne C A S E GS 34/4 GS 129/20 GS 129/5 GS 99/1 GS 39/1 GS 129/4 GS 2/147 GS 723/1 Rescaled Dstance Cluster Combne E IGS (female), (male), n=3 C A S E HD /3 HD 5/3 HD 171/2 HD 34/3 HD 54/5 HD 52/5 HD 339/9 HD 339/7 HD 451/1 HD 0/2 HD 53/ HD 170/1 HD 335/1 HD 407/1 HD 54/1 HD 132/2 HD 20 HD 339/ HD 2/1 HD 1/1 HD 341/1 HD 337/4 HD 7/1 HD 40/1 Rescaled Dstance Cluster Combne Texr-fìg. /1 - Herarchcal cluster analyss of humerus. 4 measurernent varables (GL, PL, WT, DW). CO Gamssulzen Cave, HD = Herdengel Cave. Conrurnes Cave, GS

11 SEXUAL D!MORPHISM IN LIME BONES OF LATE PLEISTOCENE C4VE BEAR 9 C A S E BC 123 BC 290 BC 12 7 BC 92 BC 7 BC 155 BC 7 BC BC 125 BC 71 CtJ 1-31 CtJ 13-5 BC BC 25 BC 22 BC 99 BC 7 BC 17 BC 134 BC 70 BC Rescaled Dstance Cluster Combne C A S E GS 1 1 /124 GS 129/25 GS 2/12 GS 50/1 GS 11/125 GS 2/12 GS 45/ GS 2/2 GS 2/130 GS 2/15 GS 2/2 GS 2/145 GS 31/1 Rescaled Dstance Cluster Combne C A S E HD 45/1 HD 4/1 HD 343/1 HD 15/1 HD 91I1 HD 1/1 HD 53/14 HD 51/1 HD 57/1 HD 340/3 HD 441/1 HD 71/2 HD 312/1 HD /2 HD 305/3 HD 70/1 HD 52/1 HD 457/1 HD 52/2 HD 33/1 HD 35/1 HD 403/1 HD 337/1 HD 92I 1 Rescaled Dstance Cluster Combne Texr-fìg. /2 - Herarchcal clusrer analyss of radus. measuremenr varables (GL, PL, PW, DW, LCR, WCR). CO = Conrurnes Cave, GS = Gamssulzen Cave, HD = Herdengel Cave.

12 1 C. REISINGER, j HOHENEGGER C A S E BC 299 BC 551 BC 14 BC 17 BC 721 BC 124 BC 723 BC 54 BC 720 cu x5 BC 51 BC 174 cu BC 559 BC 50 cu 1-32 cu BC 573 BC 59 BC 59 cu 13-4 BC 51 BC 57 BC 2 BC 592 BC 135 BC 19 BC 553 BC 130 BC 301 Rescaled Dstance Cluster Combne C A S E GS 133/44 GS 11/127 GS /3 GS 30/ GS 11/12 GS 45/9 GS 55/21 GS 2/25 GS 2/17 GS /35 GS 2/117 GS 2/1 GS 41/3 GS 52 /1 GS 5/5 GS 129/2 GS 2 I 24 GS 45/ GS 513/2 Rescaled Dstance Cluster Combne C A S E HD 94/3 HD 144/2 HD 1/2 HD 77 /1 HD 52/3 HD 419/1 HD 55/2 HD 123/1 HD 54/1 HD 337/ HD 55/1 HD 1/5 HD 407/3 HD 9/12 HD 4/1 HD 35/4 HD 4/1 HD 10/1 HD 142/2 Rescaled Dstance Cluster Combne Texr-fìg. /3 - Herarchcal cluster analyss of ulna. 7 measurement varables (GL, PW, DW, SDD, TDD, HPAT, DIT). CO= Conmrnes Cave, GS = Gamssulzen Cave, HD = Herdengel Cave.

13 SEXUAL DIMORPHISM IN LIM B BONES OF LATE PLEISTOCENE CAVE BEAR 111 Rescal ed D stance Cl uster Combne C A S E BC 139 BC 141 cu 53-3 BC 302 BC 3 7 BC 391 BC 23 BC 445 BC 33 BC 295 BC 35 BC 3 BC 34 BC 3 BC 132 BC 372 BC 392 BC 39 BC 379 BC 3 1 BC 140 BC 3 04 BC 27 Rescaled Dstance Cluster Combne C A S E GS 129/ GS 129/9 GS 70/1 (ma l e), GS 2/11 IGS GS 34/5 GS 133/42 GS (female), GS 525/ 1 Rescal ed Dstance Cluster Combne C A S E HO 0/1 HD 150/1 HD 331/5 HD 339/ HD 312/ 4 HD 54 /7 HD 2/1 HD 12/1 HD 144/3 HD 75/3 HD 119/1 HD 74/4 HD 54/ HD 13/4 HD 0/1 HD 339/ HD 0/2 HD 402/1 HD / HD 315/3 HD 513/1 HD 5/4 HD 3/1 HD 430/1 HD 53/1 0 HD 99/2 HD 142/4 HD 1 / 1 HD 9/20 HD 54/ HD 75/1 HD 525/1 Texr-fìg. /4 - Herarchcal cluster analyss of femur. 4 measuremenr varables (GL, PW, GLT, DW). CO Gamss ulzen Cave, HD Herdengel Cave. Conrurnes Cave, GS

14 C A S E BC 450 cu x3 BC 15 BC 455 BC 451 BC 73 BC 737 BC 49 BC 41 BC 45 BC 472 BC 154 BC 452 BC 470 BC 473 BC 297 BC 303 BC 29x BC 43 BC 454 BC 457 BC 734 BC 51 BC 147 BC 42 BC 45 BC 300 BC 733 BC 13 BC 293 BC 12 BC 40 Rescaled Dstance Cluster Combne C A S E GS 2/19 t GS 133/43 GS 129/23 GS 129/90 GS 51/5 GS 703/ GS 129/21 GS 525/2 I'',...,.,.. (male), Rescaled Dstance Cluster Combne C A S E HD 70/3 HD 9/22 HD 90/2 HD 50/1 HD 9/23 HD 121/3 HD 321/5 HD 7/2 HD /5 HD 430/2 HD 4/2 HD 5/5 HD /1 HD 70/4 HD 337/3 HD 0 I 3 HD 4 /13 HD 2 I 4 HD 47/4 HD 74/5 HD 4/2 HD 09/2 HD 341/23 HD 1/2 HD 09/1 HD 4 /14 HD 401/1 HD 57/1 HD /4 HD 35/7 HD 34 O/ 2 HD 340/1 HD 72/1 HD 140/1 HD 11/3

15 SEXUAL DIMORPHISM IN LIMB BONES OF LATE PLEISTOCENE CAVE BEAR 113 Sex ratos of femur data: Cave n sex rato (female/male) CO 23 20: 3 or.7 GS 7 3: 4 or 0.75 HD 32 9: 23 or 0.39 Sex ratos of tba data: Cave n sex rato (female/male) CO 32 27: 5 or 5.40 GS 7: 1 or 7.00 HD 35 4: 31 or 0.17 On the bass of the three nvestgated caves from dfferent alttudes, a shfr n the sex ratos correlated to alttude can be detected. Whle larger szed males are numerous at low alttudes, females seem to become more abundant n hgher regons. If lmb bones are not dfferentated accordng to sex, then the summarsed sze dstrbutons may show a sgnfìcant decrease n populaton means correlated to alttude, whch led to the assumpton of a small-szed 'hghalpne form' (Rabeder, 1991). The exstence of such a small szed local populaton must be proven by sgnfìcant dfferences wthn both or even one of the sexes consderng the three caves. Parametrc and nonparametrc analyses of varance were performed for each sex from the three stes usng the greatest length of examned lmb bones (Table 1). Testng homogenety of varances as an assumpton for parametrc one-analyss of varances showed homogenety, but demonstrated sgnfìcant dfferences for only three bones: the female humerus, the female ulna, and the male ulna. The nonparametrc Kruskal-Walls test was addtonally performed for these elements. Results of both types of analyses demonstrate the statstcally sgnfìcant homogenety of al lmb bones for each sex. Thus, there are no sgnfìcant dfferences n al fìve skeletal elements between the three caves for each sex. In the followng, the assumpton of a small-szed form of cave bear from Conturnes Cave must be rejected. DISCUSSI ON Sexual dmorphsm can be expressed as bmodal frequency dstrbutons n long bones of cave bears. Other explanatons of bmodalty are possble, such as effected by large- and small-szed sympatrc speces occupyng dfferent ecologca! nches. Ths nterpretaton, however, would lead to the assumpton that there was no sexual dmorphsm n ether speces present, a weak argument consderng the stuaton n modem bear speces (Kurtén, 197, p. 5; Scrlng 1993). Tumer (194) has <leale wth che extent of sex dmorphsm n recent and fossl large felds by examnng che teeth and remans of the postcranal skeletons of modem and Plestocene lons. He argues that che nterpretaton of bmodal sze dstrbutons as two dstnct speces nstead of two sexes s msleadng. Tumer concludes that "... moreover, attempts usng che fossl record as a gude to tempo and mode n evoluton rsk consderable error f taxa have been proposed wchout regard to sexual varacon" (Tumer, 194, E ). Dfferences n body sze of cave bear sexes calls for a bologca! explanaton. The larger body sze of males may best be accounted for terrtoral fìghts durng matng perods, as has been observed n modem bears. In contrast, dmorphsm may perhaps also be explaned by females havng reached fertlty and stopped growng earler chan ther male counterparts, also a well-known fact n modem bears (Strlng, 1993). Other hypothess such as food supply shortages argues aganst a sze-ncreasng trend. However, such developments can also lead to the stablsaton of sexual dmorphsm. Comparson of dstrbuton parameters for al lmb bone measuremenrs between the stes of dfferent alttude resulted n constancy for each sex, thus no dfferences n ether males or females correlated to alttude could be detected. The problem wth sex rato of cave bears s chat the matera ofren represents a small number of che exstng remans. Addtonally, che sze of caves can change rapdly durng some,000 years. The nvestgaced skeletal elements are echer wthout any stracgraphcal nformaton (Conrurnes and Gamssulzen Caves) or show only a small pan of the cave bear contents of the stes (Herdengel Cave), thus an nrerpretaton becomes dffìculc. Nevertheless, che total number of female elements n the Conturnes and Gamssulzen Caves exceeds the male ones. By contrast, che rato of lmb bones n che Herdengel Cave s moscly well-balanced, whle male elements are otherwse predomnant. Therefore, the obvous shfr to smaller szed specmen lvng n hgh alttudes s only caused by che change of sex ratos n favour of che smaller szed females. An explanaton of female domnance n ecologcally extreme regons may be found n hdng cubs aganst males. For example, Kurtèn (197, p. 77) suggests that female cave bears wch cubs show preference for small caves as sleepng places, because they are easer to defend aganst male ntruders. Texr-fìg. /5 - Herarchcal cluster analyss of tba. 4 varables measurement (G L, PW, TDD, WCT). CO = Conrurnes Cave, GS = Gamssulzen Cave, HD = Herdengel Cave.

16 114 C. REISINGER, J HOHENEGGER 5 4 g3 2 al" 1 L: Humerus Calturnes Cave (n=1) O.O Dscrmnant scores : 2 al" 1 L: Radus Calturnes Cave (n=21) O.O Dscrmnant scores al" L: Humerus Garssulzen Cave (n=) O.O Dscrmnant scores Radus Garssulzen Cave (n=13) 5 4 >. 3 2 al" 1 L: O.O Dscrmnant scores Humerus Herclerga Cave (n=24) [ 4 u O.O Dscrmnant scores Radus Herdengel Cave (n=24) 5 4 al" 1 L: O.O Dscrmnant scores Text-fìg Frequency hstograms of dscrmnant scores from humerus and radus. Hollow bars represent female elements, shadowed bars male elements. The number n parentheses (besde the subtde) represents sample sze. A last ssue of ths nvestgaton s the stablty of lmb-bone dmensons over a relatvely long perod of tme for both sexes. Such stablty s ndcated by frequency dstrbutons of the nvestgated bones, constant over a duraton of about 90,000 years (from 25,000 to 115,000 years BP). Only 2 bones from the Herdengel Cave show older ages (from 12,000 to 13,000 years B.P), but both fìt well mo the dmenson range of the younger bones.

17 SEXUAL DIMORPHISM IN LIMB BONES OF LATE PLEISTOCENE CAVE BEAR 115 Ulna Croturnes Cav' (n=30) Dscrmnant scores Femur Croturnes Cav' (n=23) : u Dscrmnant scores Ulna Garssulzen Cav' (n=19) Dscrmnant scores Garssulzen Cav' ar U: Femur Dsamnant scores Ulna Herdengel Cav' (n=19) ar 1 U: O.O Femur Herdengel Cav' (n=32) O.O Texr-fìg. 9/2 - Frequency hsrograms of dscrmnanr scores from ulna and femur. ACKNOWLEDGEMENTS We are grareful to K. É. Thomanek for edtng rhs paper. G. K. Kunst was so knd to make avalable hs data from rhe Herdengel and Gamssulzen Caves, and C.R. s grateful to G. Rabeder for gvng rhe opporrunry to work wrh cave bear mareral from rhe Conmrnes Cave. Except for K. É. Thomanek, borh researchers menroned above are affìlated wrh rhe Instmte of Palaeontology of the Unversry ofvenna. Fnally, our thanks go to the Offce for the Preservaton of Hstorcal Monuments n Bozen (Bolzano), South Tyrol (Italy), whch enabled the work wth an extensve array of Conturnes matera at the Venna lnstmte. That matera was examned and measured n the same manner as the bear bones from the other caves, a fortunate crcumstance allowng for drect comparsons between the dfferent stes.

18 lba Cmturnes Cave (n=32) :)' 4 j Dscrmnant scores lba Herdengel Cave (n=35) Dscrmnant scores Text-fìg. 9/3 - Frequency hstograms of dscrmnanr scores from tba. REFERENCES BACHOFEN-ECHT, A., 1931, Beobachtungen.ber de Entwcklung und Abn.tzung der Eckzahne be Ursus spelaeus. In Abel, O. & Kyrle, G., De Drachenhhle be Mxntz: Spelaolog. Monogr., 7: , Wen. BENECKE, N., 197, Zur Bedeutung und Anwendung mathematsch-statstscher Verfahren n der Archaozoologe: Ethnogr. Archaol. Z., 2: , Berln. BLACKJTH, R. E. & REYMENT, R. A., 197 1, Multvarate Morphometrcs: 412 pp., Academc Press, London/New York. BORTZ, ]., 1993, Statstk f.r Sozalwssenschaftler: 4.Aufl., 753 pp., Sprnger Verlag, Berln. DRIESCH, A. VON DEN, 192, Das Vermessen von Terknochen aus vor- und frlhgeschchtlchen Sedlungen: Insttut f.r Palaoanatome, Domestkatonsforschung und Geschchte der Termedzn der Unv. M.nchen, 2nd. ed., 144 pp., M.nchen. EvERITT, B. S., 1993, Cluster analyss. 3rd edton: 170 pp., Hodder & Stoughton, London. GRANDAL D'ANGLADE, A., 1993, El oso de las cavernas en Galca: e! yacmento de Cova Eròs: O Castro, 25 pp., La Corufa. HOHENEGGER, J. & ZAPFE, H., 1990, Cranometre nvestgatons on Mesopthecus n. comparson wth two recent colobnes: Betr. Palaont. Osterr., 1: , Wen. KOBY, F.-E., 1949, Le dmorphsme sexuel des cannes d' Ursus arctoset d' Ursus spelaeus. Rev. Susse Zoo!., ': 75-7, Geneve. -, 1950, Les dmensons mnma et maxma des os longs d' Ursus spelaeus-. Eclog. Geo!. Helvet., 43: 27, Base!. KuNST, G. K., 1992, Hoch- und spatglazale GrC?fÌsaugerreste aus dem Nxloch be Losensten-Ternberg (O.O.). In Nagel, D. & Rabeder, G., (eds.), Das Nxloch be Losensten- Ternberg: Mtt. Komm. Quartarforsch., : 3-127, Wen. KOHTREIBER, T., 1995, Ablauf und Methodk der palaontologschen und archaologschen Grabungsarbeten n der Gamssulzenhhle, Totes Gebrge. n Rabeder, G., (ed.), De Gamssulzenhhle m Toten Gebrge: Mrt. Komm. Quartarforsch. Osterr. Akad. Wss., 9: 1-133, Wen. KURTÉN, B., 1955, Sex Dmorphsm and Sze Trends n the Cave Bear, Ursus spelaeus Rosenm.ller & Henroth: Acta Zoo!. Fennca, 90: 1-4, Helsnk. -, 197, Plestocene Bears ofnorth Amerca. 2. Genus Arctodus, short-faced bears: Acta Zoo!. Fennca, 117: 1-0, Helsnk. -, 197, The cave bear story: 13 pp., Columba Unv. Press, New York. LETTNER-WILD, E., RABEDER, G. & STEFFAN, I., 1994, Determnaton of the evolutonary mode of Austran alpne cave bears by uranum seres datng: Hstorcal Bology, 7: & STEFFAN, I., 1993, Uranum-seres datng of fossls bones from alpne caves: Archaeometry, 35: NAGEL, D. & RABEDER, G., 1991, Exkursonen m Plozan und Plestozan Osterrechs: Osterr. Palaont. Ges., 44 pp., Wen. RABEDER, G., 1991, De Hhlenbaren der Conrurnes-Hohle. Enrdeckung und Erforschung ener Dolomten-Hohle n 200 m Hohe: 124 pp., Athesa, Bozen. 1995, Evolutonsnveau und Chronologe der Hohlenbaren aus der Gamssulzen-Hohle m Toren Gebrge (Obersterrech). In Rabeder, G. (ed.), De Gamssulzenhohle m Toten Gebrge: Mtt. Komm. Quartarforsch. Osterr. Akad. Wss., 9: 1-133, Wen & MAIS, K., 195, Erste Grabungsergebnsse aus der Herdengelhohle be Lunz am See (Nedersterrech): De Hohle, 3 (2): 35-41, Wen. & WETCHENBERGER, J., 1995, De Gamssulzenhohle be Sptal am Pyhrn m Toten Gebrge. Lage, Morphologe und Forschungsgeschchte. In Rabeder, G. (ed.), De m Toten Gebrge: Mtt. Komm. Quartarforsch. Osterr. Akad. Wss., 9: 1-133, Wen. REISINGER, C., 1995, Untersuchungen am Stylo- und Zeugopodum vom Hohlenbaren ( Ursus spelaeus, Carnvora, Mammala): 93 S., unpubl. Dplomarbet, Unv. Wen. RODE, K., 1935, Untersuchungen.ber das Gebfì der Baren: Monogr. Geol. Palaont., 7 (2): 1-12, Lepzg. SoKAL, R.R. & ROHL F, F.J., 1995, Bometry: 3rd. ed., 7 pp., WH.Freeman, San Francsco. SPSS, X, 193, User's gude: McGraw Hll, New York. STIRLING, I., 1993, The Lvng Bears. In Strlng, I. (ed.), Bears: 240 pp., Harper Collns, London. STUBBE, M., 1993, Famle Ursdae GRAY, 125. In Stubbe, M. & Krapp, F. (eds.), Handbuch der Saugetere Europas. Raubsauger (Tel 1): 52 pp., Wesbaden. TORRES PÉREZ-HIDALGO, T. ]., 19, Osos (Mammala, Carnvora, Ursdae) del Plestoceno de la Pennsula Ibérca: Bol. Geo!. y Mn., 99 (1-): 1-31, Madrd. T URNER, A., 194, Dental sex dmorphsm n European lons (Panthera leo L.) of the Upper Plestocene: palaeoecologcal and palaeoethologcal mplcatons: Ann. Zoo!. Fennc, 21: 1-, Helsnk. (manuscrpt receved August 1, 1997 accepted Apr 29, 199) Chrstan RETSTNGER Johann HOHENEGGER Instture of Paleontology Unversty ofvenna Althanstrasse 14, A-90 Venna, Austra