Taphonomy and palaeoecology of an assemblage of large mammals: Hyaenid activity in the Lower Pleistocene site at Venta Micena (Orce, Guadix-Baza Basin, Granada, Spain)

TAPHONOMY AND PALAEOECOLOGY OF AN ASSEMBLAGE OF LARGE MAMMALS: HYAENID ACTIVITY IN THE LOWER PLEISTOCENE SITE AT VENTA MICENA (ORCE, GUADIX-BAZA BASIN, GRANADA, SPAIN) ALFONSO ARRIBAS & PAUL PALMQVIST ARRIBAS A. & PALMQVIST P. 1998. Taphonomy and palaeoecology of an assemblage of large mammals: hyaenid activity in tLe Lower Pleistocene site at Venta Micena (Orce, Guadix-Baza, Granada, Spain). [Taphonomie et pal6o6cologie d'un assemblage de grands mammif'eres: activit6 des hy6nes dans un site du P16istoc6ne inf6rieur ~ Venta Micena (Orce, bassin de Guadix-Baza, Grenade, Espagne)]. [Tafonomia y paleoecologfa de una asociaci6n de grandes mamfferos: actividad de hienidos en el Pleistoceno inferior de Venta Micena (Orce, Cuenca de Guadix-Baza, Granada, Espafia]. GEOBIOS, 31, 3, s u p p l 6 m e n t : 3-47. Villeurbanne, le 30.06.1998. Manuscrit d6pos6 le 09.06.1997; accept6 d6finitivement le 28.01.1998. ABSTRACT - We report taphonomic and palaeoecologic data on the rich, diverse and well preserved assemblage of large mammals from lower Pleistocene deposits at Venta Micena (Orce, Granada, south-east Spain). The biostratinomic and diagenetic characteristics of the assemblage are congruous with the sedimentary context deduced from the study of the site, and both confirm that: (i) the assemblage represents an accumulated taphonomic stage, (ii) it was formed by demic, autochtonous palaeobiologic entities, which were preserved and recorded in situ, and (iii) it is the result of biological processes and agents. Interspecific analysis of size/abundance patterns in ungulates shows that the main taphonomic bias affecting the bones was produced by biological destruction before burial, and that the loss of information was ~,~eater for species of smaller body size. Factor correspondence analysis was used to compare the frequencies at which some groups of postcranial elements are represented in several recent and archaeological bone assemblages accumulated by carnivores, rodents and hominids. The results obtained strongly suggest that the bones from Venta Micena were collected mainly by hyaenids, which deposited them near shallow dens excavated around the ponds that surrounded the Pleistocene lake of Orce. An analysis of the abundance of major long bones has shown that differential fragmentation was produced by hyaenas as a function of their structural density and mean marrow content. All these data allow to formulate a descriptive-quantitative model for the characterization of bone assemblages generated from hyaenid activity, in which Venta Micena is an example of bone concentration and modification activities by Pachycrocuta brevirostris. Strong selection of prey by carnivores (which killed preferably juveniles, females and individuals with diminished locomotive capabilities among ungulate prey species of larger body size) is indicated by (i) the abundance of juvenile individuals with deciduous teeth in relation to the average weight estimated for adults in each ungulate species, by (ii) the U-shaped attritional mortality profiles deduced from crown heigth measurements, by (iii) the presence of many metapodials with different osteopathologies, and by (iv) a biased sexual ratio deduced from the metacarpals of large bovids. Comparison between the frequencies in which modern African carnivores kill and scavenge ungulates from different size classes and the abundance of these size categories in the assemblage suggests that the Venta Micena hyaena was a bone-cracking scavenger which fed largely on carcasses of ungulates preyed upon and partially consumed by flesh-eating carnivores such as saber-toothed felids and wild dogs. KEYWORDS: BONE ASSEMBLAGE, TAPHONOMY, PALAEOECOLOGY, MULTIVARIATE ANALYSIS, LOWER PLEISTOCENE, SPAIN. RI~SUMI~ - Ce travail est l'6tude taphonomique et pal6o6cologique d'un assemblage abondant et diversifi6 de grands mammif'eres provenant du gisement de Venta Micena (P16istoc6ne inf6rieur, Orce, Grenade, Sud-Est de l'Espagne). Les caract6ristiques biostratinomiques et diag6n6tiques de l'assemblage s'accordent avec le contexte s6dimentaire d6duit de l'6tude du gisement, les deux approches confirmant (1) que l'assemblage repr6sente un ensemble taphonomique d'accumulation, (2) qu'il fur form6 par des entit6s pal6obiologiques autochtones et d6miques, pr6serv6es et enregistr6es in situ et (3) qu'il est le r6sultat d'agents et de processus bio!ogiques. L'examen intersp6cifique des donn6es, taille/abondance des esp6ces d'ongul6s, montre que les princip~/ux biais taphonomiques sont dus ~ une destruction biologique avant l'ensevelissement et concernent principalement les esp6ces de petite taille. L'analyse factorielle des correspondances est utilis6e pour comparer les fr6quences des diff6rents groupes d'616ments du squelette post-cr~nien rencontr6s dans des ensembles r6cents ou arch6ologiques d'ossements accumul6s par les carnivores, les rongeurs et les hommes. Les r6sultats obtenus sugg6rent que les restes de Venta Micena ont 6t6 r6colt6s principaleme:at par des hy6nes qui les ont d6pos6s autour de l'entr6e de leurs repaires peu profonds, creus6s pr6s

des mares qui bordaient le lac pleistocene d'Orce. L'~tude de l'abondance des principaux os longs conserves dans l'assemblage montre que les hy~nes r~alisaient une fragmentation diff~rentielle des os en fonction de leur densit~ min~rale et de leur contenu moyen en mo~lle osseuse. Avec ces donn~es, nous pouvons presenter un module descriptifquantitatif pour la caract~risation des assemblages osseux r~sultant de l'activit~ des hy~nes dont Venta Micena est un exemple de concentration et de modification r~sultant de l'action de Pachycrocuta brevirostris. L'importante s~lection des proies par des carnivores qui chassaient principalement des individus jeunes, des femelles et des sujets capacit~ locomotrice r~duite dans le cas d'ongul~s de grande taille, est d~montr~e par (1) l'abondance des individus jeunes (presence de la dentition d~ciduale) et la relation avec le poids estim~e pour les adultes de chaque espYce; (2) par la courbe de mortalit~ en U obtenue ~ partir des mesures de hauteur de la couronne dentaire; (3) de la presence de nombreux m~tapodes avec diverses ost~opathologies; (4) de la proportion des sexes d~duite de la morphom~trie des m~tacarpiens des grands bovid~s. Finalement, la comparaison entre la fr~quence observ~e chez les carnivores africains actuels chasseurs et charognards d'ongul~s de diverses classes de taille et l'abondance de ces categories de taille dans l'assemblage fossile sugg~re que la hy~ne de Venta Micena avait un comportement de charognard, s'appropriant les cadavres consommes pour partie par les autres carnivores, tels les f~lid~s ~ dents de sabre et les chiens sauvages. MOTS-CLI~S: ASSEMBLAGE OSSEUX, TAPHONOMIE, PALt~O]~COLOGIE, ANALYSE MULTIVARI]~E, PLt~ISTOCt~NE INFERIEUR, VENTA MICENA. RESUMEN - En este trabajo se efectfia un estudio tafon6mico y paleoecol6gico de la abundante y diversa asociaci6n de grandes mamiferos proveniente del yacimiento de Venta Micena (Pleistoceno inferior; Orce, Granada, sureste de Espafia). Las caracteristicas bioestratin6micas y diagen~ticas de la asociaci6n fSsil son congruentes con el contexto sedimentario deducido del estudio del yacimiento, confirmando ambas que: (i) la asociaci6n presenta un estado tafon6mico de acumulado, (ii) se encuentra formada por entidades paleobiol6gicas d~micas, por entidades conservadas aut6ctonas y por entidades registradas in situ, y (iii) es el resultado de la acci6n de procesos y agentes biol6gicos. E1 an~lisis interespecifico de los patrones de tamafio/abundancia en las especies de ungulados ha puesto de manifiesto que los principales sesgos tafon6micos se produjeron como consecuencia de la destrucci6n por agentes biol6gicos de los restos 6seos en la etapa previa a su enterramiento, afectando la p~rdida de informaciSn m~s a l a s especies con dimensiones corporales reducidas. E1 an~lisis factorial de correspondencias se us6 para comparar las frecuencias en que distintos grupos de elementos esquel~ticos postcraneales se encuentran representados en asociaciones recientes y arqueol6gicas de huesos acumulados por carnivoros, roedores y homlnidos. Los resultados obtenidos sugieren que los restos de Venta Micena fueron recolectados principalmente por las hienas, quienes los depositaron en torno a la entrada de cubiles poco profundos, excavados cerca de las charcas que bordeaban el lago pleistoceno de Orce. Un an~lisis de la abundancia de los principales huesos largos preservados en la asociaci6n ha mostrado que las hienas produjeron una fragmentaci6n diferencial de acuerdo con su densidad mineral y su contenido medio en m~dula 6sea. Todo ello permite presentar un modelo descriptivo-cuantitativo para la caracterizaci6n de asociaciones 6seas generadas por la actividad de los hi~nidos, pudiendo ser Venta Micena el ejemplo de actividad concentradora-modificadora de los representantes de la especie Pachycrocuta brevirostris. Se ha podido deducir una intensa selecci6n de presas por parte de los carnlvoros, los cuales cazaban preferentemente individuos j6venes, hembras y ejemplares con capacidad locomotriz disminuida en el caso de las especies de ungulados con mayor tamafio. Las evidencias que apoyan este modelo son: (i) la abundancia de individuos j6venes, con dentici6n decidua, en relaci6n al peso estimado para los adultos de cada especie; (ii) los perfiles de mortandad gradual con forma de U, inferibles de las medidas de altura de la corona dentaria; (iii) la presencia de abundantes metapodios que muestran diversas osteopatologias; y (iv) la proporci6n de sexos deducible a partir de los metacarpianos de los grandes bSvidos. Finalmente, la comparaci6n entre la frecuencia en que los carnlvoros modernos africanos cazan y carrofiean ungulados de diferentes clases de tamafio y la abundancia de estas categorias en la asociaci6n fSsil sugiere que la hiena de Venta Micena era b~sicamente carrofiera, aprovechando los cad~veres de presas cazadas y consumidas en parte por otros carnlvoros comedores de carney no fracturadores de huesos, tales como los f~lidos con dientes en forma de sable y los perros salvajes. PALABRAS CLAVE: ASOCIACION 0SEA, TAFONOMIA, PALEOECOLOGIA,AN,~LISIS MULTIVARIANTE, PLEISTOCENO INFERIOR, VENTA MICENA.

INTRODUCTION

AND

BACKGROUND

T h e V e n t a M i c e n a site (Orce, G r a n a d a , s o u t h - e a s t Spain) is located in t h e e a s t e r n sector of t h e G u a d i x - B a z a b a s i n (Fig. 1). This b a s i n w a s endorheic u n t i l u p p e r P l e i s t o c e n e times, t h u s facilitat i n g a n e x c e p t i o n a l r e c o r d of t h e P l i o - Q u a t e r n a r y t a p h o c o e n o s e s of l a r g e m a m m a l s , w h i c h w e r e pre-

s e r v e d in s w a m p y a n d l a c u s t r i n e s e d i m e n t s . P r e v i o u s studies i n d i c a t e t h a t t h e V e n t a M i c e n a a s s e m b l a g e is i n c l u d e d in 90-98% p u r e micritic limestone, w h i c h p r e c i p i t a t e d in f r e s h w a t e r p o n d s e m p l a c e d on a caliche paleosol of d i a g e n e t i c origin. This paleosol s u r r o u n d e d a m o r e or less stable a n d shallow lake w i t h s w a m p y m a r g i n a l zones t h a t existed in t h e v a l l e y of Orce d u r i n g t h e

MAIN BETICS INTRAMONTANE segura .........

-

-

~-"

)

- -

%OZll~l)V:

~'~6~adix-8~--~-

- 1 "~ BAZA

.........

::vm, rrA~ac~A'~

ORCE~.jg'~y

f!

u

ff

]_~

U

~\

~ }j_

wEs

~

~

~

~ -~o~

OUADIX~,

N~O@]KNEDEPOSITS

. ~

KXTERNALZONES

II'YFEI~ALZONES

Sierrade Castdl v ~ ~ ~ ~ .~ . E ~ a l ~ t m g r - ~ ~ ~ W

Sierra Nevada

.N

Orce-VentaMic~mSection: fertilestrata: units: legend: ~T~ 35m' ~conglomeratcs

iIIi ~~

!

~

~

B

a

r

r

a

n

c

o

[2_~.~Barranco Ledn ~,

•

.-

~

.

I~BazaFolmation

~

0 10 20km

,----

l%Barranc-odel l)aso

::IV

tes~meNone)

lCortijo D. Alfonso ! /Barranco de los !HI~reen clays ~Conejos H andcarbonates -FuenteNueva-1 0miiI ~redclays

...............

FIGURE i - Geological location of Venta Micena and Orce in the intramountainous basin of Guadix-Baza (province of Granada, Southeast Spain), and stratigraphic sections of the Plio-Pleistocene deposits in the Orce-Fuente Nueva-Venta Micena sector. Localisation g,~ologique de Venta Micena et d'Orce dans le bassin intramontagneux de Guadix-Baza (province de Grenade, Sud-Est de l'Espagne) et succession stratigraphique des ddpSts plio-pldistoc~nes dans le secteur d'Orce-Fuente Nueva-Venta Micena.

Plio-Pleistocene (Vera et al. 1985; Soria et al. 1987)• Biostratigraphic analyses of the faunal assemblage (Martfnez-Navarro 1991, 1992a; Martinez-Navarro & Pahnqvist 1995; Turq et al. 1996; Martinez-Navarro et al. 1997) place this site in the lower Pleistocene, with an estimated age of 1.2 _+0.2 Ma. The 80-120 cm thick Venta Micena stratum is one of the various fossiliferous units with macromamreals in the Plio-Pleistocene sedimentary sequence of Orce, whose surface can be followed along 2.5 km approximately, and stands out topographically in the ravines of the region. All the fertile strata from the Orce-Venta Micena sector are found within a carbonate sequence with mean thickness

of 15 m, horizontal stratification undisturbed by tectonic activity, without breaks or discordances, and with a cartographic extension of approximately 16 kmL This sequence shows the following vertical sedimentary evolution: white clayey marly facies with carbonate concretions of diagenetic origin and freshwater invertebrates (swamp in central facies); clayey muddy facies with mudcracks, abundant coal and bony mammal remains (swamp in marginal facies); white carbonate facies (mudstone) with paleosols, carbonate concretions and abundant skeletal remains of fossil mammals (lake with various periods of partial desiccation, during one of which the Venta Micena assemblage was formed).

The fossils from Venta Micena-are placed on a paleosol which was developed on sediments deposited during a first lacustrine stage, thus evidencing a generalized descent of the water table in the lake, which was characterized by wide emerged zones (mudcracks and rootmarks) with small shallow ponds (< 1 m depth, 2-20 m diameter) (Gibert et al. 1992). The assemblage is embedded in homogeneous and porous micrite sediments, which precipitated during a period of partial expansion of the ponds (restricted swampy biotope of carbonate facies, with plants colonizing the border of the ponds), and it was closed over by a massive precipitation of micrite produced during an immediately subsequent phase of uprising of the w a t e r table in the lake (lacustrine stage two). This welling was rather slow, as suggests the absence of terrigenous, erosive structures, and of any evidence of sediment traction. The micrite which precipitated during both lacustrine stages is very similar, both texturally and structurally. The accumulation of skeletal remains in Venta Micena (Gibert & Caporicci 1989; Gibert et al. 1992; Palmqvist et al. 1992, 1993) was exclusively due to the biotic factors. Geologic processes, such as fluvial transport, can be excluded, since the bones are randomly oriented, and show no traces of abrasion from rolling or similar movements. Furthermore, m a n y bones were for some time exposed to the elements, and a very high percentage of specimens show clear marks of carnivore damage (with almost total dismembering of all elements, and bite m a r k s and breakage in the most vulnerable places). The palaeoecological analyses of this fossil comm u n i t y (Martinez-Navarro 1991, 1992b; Mendoza et al. 1993) have been based both on cenogram methodology (Valverde 1967; Legendre 1986) and on multivariate comparisons with several types of modern m a m m a l i a n communities (Reed 1997). The results obtained suggest that the composition of the palaeocommunity ofVenta Micena, in terms of the relative abundance of groups of species established according to their size and with the type of trophic resources used, was similar to that of present day communities of large mammals inhabiting African savannas with tall grass and spiny trees. The systematic study of macromammals (Mart/nez-Navarro 1991, 1992a; Martinez-Navarro & Palmqvist 1995, 1996; Palmqvist et al. 1996a) reveals three faunal components in the fossil assemblage: 1) Species present in Western Europe during the upper Pliocene (Mammuthus meridio-

nalis, Stephanorhinus etruscus, Homotherium latidens, Lynx aff. issiodorensis, cf. Meles sp. and Ursus etruscus), 2) immigrants from Asia (Praeo-

vibos sp., Bovini cf. Dmanisibos, Soergelia minor, Hemitragus alba, Caprini gen. et sp. indet., Cervidae gen. et sp. indet., Megaloceros [Megaceroides] solilhacus and Canis etruscus), and 3) species from Africa (Megantereon whitei, Equus altidens, Pachycrocuta brevirostris, Canis [Xenocyon] falconeri and Hippopotamus amphibius antiquus). The species of small mammals found at Venta Micena (Agusti et al. 1987) include Desmana sp., Apodemus aff. mystacinus, Eliomys intermedius, Allo-

phaiomys pliocaenicus, Castillomys crusafonti ssp., Hystrix major, Oryctolagus cf. lacosti and Prolagus calpensis. The presence of h u m a n s in the southern Iberian Peninsula during lower Pleistocene times is supported by the finding of a few fossil remains which were tentatively attributed to Homo sp. (a polemic cranial fragment which has generated considerable debate and a h u m e r a l diaphysis from Venta Micena; Gibert et al. 1994; Gibert & Palmqvist 1995; Zihlman & Lowenstein 1996; Palmqvist 1997; a medial phalanx from the karstic site at Cueva Victoria; Palmqvist et al. 1996b), and by undoubted stone artefacts found in situ in the lower Pleistocene site at Fuente Nueva-3a (1.07 Ma), which show a very simple technology not essentially different from t h e Oldowan and Developed Oldowan types found in sub-Saharan Africa (Tixier et al. 1995; Turq et al. 1996; Martinez-Navarro et al. 1997). It is interesting to point out that the h u m a n mandible and lithic industries found in the lower Pleistocene (> 1.6 Ma) deposits from Dmanisi, E a s t Georgia (Gabunia & Vekua 1995) are associated with M. whitei, the African machairodont also present in Venta Micena (Martinez-Navarro & Palmqvist 1995, 1996), what indicates extensive faunal dispersal from Africa to E u r a s i a in t h e lower Pleistocene. The objectives of this study are: (i) to determine whether the faunal assemblage preserved the structure of the original palaeobiocoenoses; (ii) to identify more accurately the biological agents responsible for the modification and accumulation of the assemblage; and (iii) to study the interspecific relationships that existed in the palaeocommunity, such as those produced by the predatory activity of carnivores. THE VENTA MICENA STRATUM

The Venta Micena stratum has a mean thickness of 1 m and presents the following vertical structure from bottom to top (Fig. 2): - A basal level (lacustrine stage 1) whose thickness is between one third and half of the stratum, formed by homogeneus micrite sediments with

FIGURE 2 - Idealized s t r a t i g r a p h i c profile ofVeni:a Micena. Profil stra-

fossil bones

tigraphique schgmatisd de Venta Micena. --F---

7--

.

~,~11~11_1Ldkm~-rl

F.,r.~--Li..~ r',~,,~u

~__

l~-

L.~..,..r

P

I F r[-?"I I I I 3~" "~b..L ="X..T e~L~p LA. I I " ~ 1 A . I I I ~ IAJm !~d..l'-I j l d , ~ J I I I I I I I I I I I I I J I l l ~ 13"JVl I I I I I I~r'-F1 "~ ~ ' ~

0-

+

~ i ~ ~. ~d

0

carbonate nodules edaphic structures hardpan rootmarks

D lacustrine stage 2 C ponds B paleosoil A lacustrine stage 1

geochemical samples:

mudcracks

some carbonate nodules (5-20 cm thick), small m u d banks and microscopic fossil shells of freshwater mollusks (among them, Melanoides tuberculata, an eurythermal species which colonizes a wide range of biotopes according to Anaddn et al. 1987). A 4-15 mm thick calcrete paleosol (hardpan), which forms an irregular surface, subparallel to the bedding plane, and is thicker at the highest topographic heights. -

- A new level of homogeneous and porous micrite sediments (lacustrine stage 2), which continues up to the top of the stratum, showing numerous rootmarks and mudcracks, which are more abundant in its lower part, and a high density of fossil bones of large mammals; this very restricted vertical interval in the distribution of fossils, marks the palaeontological site at Venta Micena. The macroscopic, petrographic, geochemical and mineralogic analyses of the stratigraphic levels have yielded the following results: The basal carbonate deposits are rich in mollusks but sterile in vertebrate :fossils, thus attesting to a first generalized lacustrige stage in the region, in which the micrite was precipitated under a constant, but variable in depth, water sheet (Soria et al. 1987), in a shallow and well oxigenated environment, since it is suggested by the absence of both pyrite and carbonate facies rich in organic matter that the lake was not subject to eutrophic conditions (Wells 1983). The fossils of terrestrial vertebrates preserved a.t Venta Micena are located exclusively above the calcrete level, which is interpreted as being a paleosol, developed on the surface of the micrite sediments previously deposited during the first lacustrine stage. The distribution of lossils follows the pre-existing limnic microtopography. This surface defines a stratigraphic unconformity, which was developed after a period of generalized retraction of the Pleistocene lake (Gibert et al. 1992). The calcrete paleosol thus indicates the emergence of the micrite deposits, which were then exposed to subaerial paedogene-

20 cm

[]

VMSED-1

,~

VMSED-2

Q

VMSED-3

sis in an arid climate with high evaporation rate (Plaziat 1984). The low percentage of iron oxides (0.6% Fe203) suggests that these sediments were hardly affected by the oscillations in the phreatic level. The Pleistocene lake of Orce had r a t h e r limited dimensions both in expanse and depth during the lacustrine stages described above, since the carbonate deposits are restricted in the large lakes to the borders, while in those of minor size they can extend over all the bottom (Kukal 1971), as happened in the studied region (Vera et al. 1985; Soria et al. 1987). The traditional connection between lakes, swamps and calcrete/sebkha (which m a y be masked by subaerial paedogenesis of marshy and lacustrine deposits, in which the evaporation plays an essential role and modifies the original sediments; Plaziat 1984) is transformed in the Orce-Venta Micena sector in the following sequence: swamps (Plio-Pleistocene), lake (lacustrine stage 1, lower Pleistocene), emergence and swampy ponds (paleosol formation; stratigraphic unconformity), lake (lacustrine stage 2) (Arribas 1995). In this case the calcrete type paedogenesis does not hide the sedimentary origin of the studied stratum, but allows the distinction of various sedimentary environments and to identify the inner surfaces which have chronological utility in the region, one of which is the local isochron of Venta Micena (Arribas et al. 1994). GEOCHEMICAL

ANALYSES

AND DIAGENETIC SETTING Geochemical analyses have been performed on bony samples obtained from ten metacarpals of Equus altidens which were u n e a r t h e d in different grids of the Venta Micena quarry. The samples (VME-1 to VME-10) were removed from the cortical bone at half of the diaphysis. Three samples of sediment (Fig. 2), obtained from the bottom of the fertile level of t h e Venta Micena s t r a t u m (VMSED-1, lacustrine stage 2), from the calcareous scab (VMSED-2, emerged environment) and

8 %

from a calcareous nodule (VMSED-3, ponds) were also analysed.

A

100

80

\

60

VME-4

l ............

// ~

.......

o vMe-~ I

./ ......

/ .....

- ................

"R/7%,/'/------~

~'-----~

vM.7 I ................

40

. V~E-,, I

.............

~

\

20

Quartz

Dolomite

Calcite

C-F C-A M ineral com position

FI

CI

Hi

B SiO2

~i

TiO 2,..] i

~VME-1 + VME-2 •VME-3 ---VME-4 -x-VME-5/ • -VME-6 ~VME-7 *VME-S OVME-9 ~VME-101

|

MnO. |

I I

I I

MgO.

i

i

i

I I

I ]

I I

I

I

I

The comparison between the geochemical results obtained from the fossil samples and those from the sediment ones (Fig. 4) revealed the following differences and similarities:

CaO I

Na20. " K~,

I I

.

I I

I I

P20~Fe203-

i

I

I

P.C.

OI

10 Total Fe)

20

30 %

I I

I I

I

I

I

I

I

I

I

I

40

50

60

ppm/103

C

5 VM E-1 -{-VME-2 VME-3 VME-4 VME-5 4~ VM E-6 ,~ VME-7 VME-8 O VME-9 VME-10

•J Rb Ba

L _" ~ Sr La Ce Th

-Y

T. Zr Nb Cu Trace elements

Co

Ni

The mineral phases (Fig. 3) which predominated in all the bony samples analysed were fluorineapatite (0-79%) and carbonate-apatite (0-81%); chlorine-apatite was detected in two samples (VME-2,4; 24 and 47%, respectively) and hydroxyapatite in another three (VME-3,6,10; 26, 46 and 13%, respectively). Calcite was found in low percentages (2-16%) in all samples, except VME-10, in which this mineral was very abundant (73%). The prevailing oxides were CaO (46-49%) and P205 (20-33%); SiO~ was found in low proportions (0.5-4.2%), TiO2 was only present in four samples (VME-4,5,6,8) in very low proportions (0.010.05%), I~O was detected in two samples (VME7,8; 0.02 and 0.01%, respectively), Fe~O3 was present in all of them but in minor proportions (<0.4%) and MnO was absent. The predominant trace elements were Sr (2243-4198 ppm) and Ba (121-1348 ppm), followed by V (2-39 ppm), Cu (1319 ppm), Cr (2-12 ppm), Th (1-11 ppm), Ce (0-10 ppm), Co (0-7 ppm), Ni (0-2 ppm) and La (0-2 ppm).

V

Cr

Pb

FIGURE 3 - Percentages of mineral phases (A; C-F: carbonateflurine apatite; C-A: carbonate-apatite; FI: fluorine-apatite; CI: chlorine-apatite; Hi: hydroxyapatite), weight of oxides (B), and a b u n d a n c e (ppm) of trace elements (C) in ten samples (VME1 to VME-10) of equid bones from Venta Micena.

Mineral composition: the three samples of sediment showed great mineral homogeneity (98-99% calcite, 1% quartz), and the sample from the calcareous paleosol also presented dolomite (1%) as a new mineral phase. On the contrary, the mineralogy of the samples obtained from equid fossils was more heterogeneous, since although both the fluorine-apatite and the carbonate-apatite were the prevailing mineral phases, another three phosphatic mineral phases such as carbonate-fluorine-aparite, chlorine-apatite and hydroxyapatite were also present, as well as low proportions of calcite. Abundance of oxides: the percentages in weight for oxides were similar in all the samples analysed from the sediment, showing no significant differences among them. The oxide content of the fossil samples was also homogeneous, with the exception of VME-9 and VME-10, in which P205 constituted about 20%, and their SiO~ (0.5-4.2%) content was lower and more variable than in the sediment samples (8.6-9.1%). Abundance of trace elements: the trace elements which predominated in both the sediment and the Pourcentages des phases mingrales (A; C-F: carbonate-fluorapatite; C-A: carbonate-apatite; Cl: chlorapatite; Hi: hydroxyapatite, poids des oxydes (B) et abondance (ppm) des ~ldments traces pour dix gchantillons (VME-1 & VME-IO) d'os d'~quid~s de Venta Micena.

fossils were Sr and Ba. Both elements showed similar proportions in the sample from the sediment precipitated during the second lacustrine stage, while Sr prevailed over Ba in the samples from the bottom of the fossiliferous level (calcrete paleosol), from the carbonate nodules, and from t h e equid fossils. Ba was detected both in the bones and in the sediment in similar quantities (121-1348 ppm and 79-1328 ppm, respectively), but Sr was less abundant in the sediment (9451325 ppm) than in the fossils (2243-4198 ppm).

% 100 -

-

80

60 •

40

VMSED-1

[]

VMSED-2

[]

VMSED-3 _ _ _

20

Qua.rtz

SiO 2

f

Ti02

Dolomite Mineral composition

_.l<. +VMSED1

Calcite

4.-VMSED2-~

VMSED3

AI203 ~

I

I

I

I

I

MnO-

I I

I I

I I

I I

I I

I

I

I

E

I

f

i

,

MgO CaO -

I

Na2Oqlc"-

I

I

I

I

I

I

205y/

t

I

I

I

I

I

20

30

FeaOs- ~

P.o.

I

0 10 (Total Fe)

~ I 1 I I I I I i I

I

40

50

60

%

ppm

C

1.400

1.000

. . . . . . . . .

800

"-~ VMSED-1

600

-

ee VMSED-3

400 200 0 :,

Rb

Ba

Sr La Ce Th Y Zr Nb Trace elements

Cu Co

~

;it

Ni

V

"~'~

Cr

Pb

Inorganic matter represents approximately 70% of bone and dentine, which is usually composed of hydroxyapatite [Cal0(PO4)G.2OH], and the remaining 30% is organic matter constituted in its majority by collagen, which can also vary in its chemical composition (Francillon-Vieillot et al. 1990). The fossils from Venta Micena showed a mineralogic composition in which the original phase, the hydroxyapatite, had dissapeared in 7 out of 10 samples, and its abundance was very low in the other 3 samples. The diagenetic processes seem to have produced an enrichment in calcite of the bony structure and a chemical modification of the organic apatite into other phases of the apatite group, as a function of the higher or lower substitution of the Ca, P and OH groups by different trace elements or by oxides, which were assimilated from the surrounding sediment. P was in part replaced by Si within the bony structure, and the incorporation of this element could have been produced by chemical substitution during the diagenesis or during the life of the animals, from the silicophytolits of the herbs that they consumed. Si was not found in the bones as a mineral phase (quartz) but in minerals from the apatite group, what seems to indicate that its incorporation into the bone tissue was produced during the life of the animals. In fact, the important amount of SiC2 detected in the micrite sediment (8.6-9.1%), which was generated by chemical precipitation of carbonates, could be due to the concentration in the sediment of the organic silica from the herbs that covered the plains surrounding the paleolake of Orce (Mendoza et al. 1993). The same interpretation is adequate for the enrichment in Sr and Ba, minor elements substituting Ca in the hydroxyapatite, which were more abundant in the fossils than in the sediment. On the other hand, several trace elements such as Rb, La, Y, Zr, Nb and Ni have been detected in the sediment, but they were hardly incorporated in the bones. The X-ray diffraction spectra of the samples analysed and their comparison with those obtained in

FIGURE 4 - Percentages of m i n e r a l phases (A), weight of oxides (B), and abundance (ppm) of trace elements (C) in three samples from the Venta Micena s t r a t u m (VMSED-I: micrite from lacustrine stage 2; VMSED-2: calcrete paleosol; VMSED3: micrite from a calcareous nodule in the ponds).

Pourcentages des phases min~rales (A), poids des oxydes (B) et abondance (ppm) des ~[#rnents traces (C) clans trois gchantillons de la couche de Venta Micena (VMSED-I: micrite de l~tage lacustre 2; VMSED-2: ca[cr~te de pal~osol; VMSED-3: micrite d'un nodule calcaire des mares).

10

VMSED1 VMSED2 VMSED3 I

. . . .

I

CALC

. . . .

I

. . . .

i

. . . .

I

. . . .

]

. . . .

i

. . . .

i

. . . .

~

. . . .

I

. . . .

i

. . . .

i

. . . .

~

. . . .

,

....

' ....

~

. . . .

i

. . . .

j

. . . .

i

. . . .

I

. . . .

I

'

~

i

. . . .

i

. . . .

....

' ....

,

1

ITE .*.

i . . : .

i

. . . .

L.,1

~

. . . .

23. i

. . . .

~

. . . .

I

A

DOLOMITE

0 ....

' ....

10

"

' ....

20

''

'30

40

'

'50

60'

''z°°

'

VME-1 VME-2

I

I

I

CARBONATE-APATITE I

. . . .

i

. . . .

I

. . . .

APATITE 0 ....

' ....

. . . .

~

~

10"''

....

. . . .

0 ....

i

. . . .

.

.

' ....

I

. . . .

r

. . . .

I

. . . .

i

. . . .

i

. . . .

i

. . . .

j

. . . .

~

. . . .

i

. . . .

I

. . . .

i

B

~ J ~ A J ~ L ~ ~ , J ~ , 2 0 " ' " " 3 0

....

"40"'''"50

....

' ....

. . . .

i

. . . .

.

i

.

....

' ....

60"'

'~°"

(DENTINE)

H U EbL A G~O

I

. . . .

~

HUESCAR

E

[

i

~

.

.

. . . .

i

. . . .

i

. . . .

(D . ENTINE) --

.

.

.

.

10"''

i

. . . .

....

i

. . . .

I

. . . .

i

. . . .

.

i

.

....

' ....

i

. . . .

i

. . . .

.

i

.

i

i

. . . .

i

. . . .

.

i

. . . .

I

,,_jh

I

. . . .

i

.

.

.

.

2 0 " ' " " 3 0

.

.

.

.

.

.

.

.

.

40"'''"50

.

.

.

.

.

.

.

.

.

.

.

.

60'''

C

'ze°'~

FIGURE 5 - X-ray diffraction spectra of the samples from Venta Micena. A, comparison of the three samples of sediment with samples of pure calcite and dolomite, what shows the homogeneity in the crystallinity of the spectra from the site, as indicated by their perfect overlap. B, representation of three bone samples (V1V[E-1to VME-3) which show great cristaline homogeneity, compared with the spectra for the two mineral phases of the apatite group with which they show a higher similarity. C, comparison between the bone from Venta Micena (VME-1) and the dentine of equid teeth from Hu~scar; both show analogous composition and similar differences with the spectra of equid dentine from the palaeontological sites at Hu~lago and Cfillar de Baza (which have a new cristaline phase, and are very similar to each other). X-ray diffraction measurements were carried out on powdered bone and tooth samples by means of a Philips diffractemeter monochromatized to Cu-k radiation (X=1540A). Spectre de diffraction X pour des dchantillons de Venta Micena. -4, comparaison de trois dchantillons de sddiment avec des dchantillons de calcite et de dolomite pures montrant l'homogdnditd de cristallinitd du spectre du site comme l'indique leur recouvrement parfait. B, reprdsentation de trois dchantillons d'os (VME-1 h VME-3) montrant la grande homog~nditd cristaIline, comparge avec les spectres de deux phases min#rales du groupe de l'apatite avec lesquels ils montrent la plus haute similarit& C, comparaison d'un os de Venta Micena (VME-1) et la dentine de dents d~quidg de Hudscar; les deux montrent une composition analogue et des diffdrences similaires avec le spectre de dentine d'dquidd des sites palgontologiques de Hudlago et Ccellar de Baza (qui ont une nouvelle phase cristalline et sont tr~s similaires entre eux). Les mesures de diffraction X sont rdalisdes sur de l'os en poudre et des dchantillons de dents ~t l'aide d'un diffractom~tre monochromatique Philips en radiation Cu-k (~= 1540•).

11 samples from other sites in the Baza Formation of the basin (Alberdi & Ruiz Bustos 1989), show interesting results (Fig. 5). The spectrograms of the equid bones from Venta Micena are similar to the spectrum for the sample of tooth dentine obtained from equids of the middle Pleistocene site at Hu~scar-1. Both spectra differ from those obtained for the tooth dentine of equids from two other sites in the basin, the upper Pliocene site at Hu61ago (F,quus stenonis livenzovensis) and the middle Pleistocene site at C611ar de Baza-1, whose samples are very similar. Venta Micena and Hu~scar-1 are v e r y close geographically within t h e c a r b o n a t e Baza Formation (Vera 1970a, 1970b), while Hu61ago and Cfillar de Baza1 are located quite distant (several dozens of km) from the former sites, although both belong to the same basin, and the former is located in the Guadix Formation, which shows clastic characteristics. These comparisons suggest that the mineral composition of the fossilized bony tissue (excluding enamel, which presents both structural and compositional properties clearly different from those of the bone and dentine) does not rely on the taxonomic filiation or on the age of the fossils, b u t on the lithologic characteristics and the hydrogeological evolution of the different formations in which the palaeontological sites are emplaced. The sedimentary infilling affects only those bones which have biostratinomic fractures (spiral and longitudinal fractures, as discussed below). The bones which were preserved complete lack any sediment in their inner surface, even in those areas which are close to large nutrient foramina, w h a t indicates that the bones were buried with the periosteum intact. The sedimentary infilling has the same micrite composition as that of the stratum, and is only present in the medullary cavity of the fractured bones, in form of small mud flows which dissapear towards the interior of the specimens. This type of infilling is not found in the complete bones. All the major long bones which were preserved complete, show one or more diagenetic fractures, orthogonal to their longitudinal axes, which are located either at half of the diaphyses or at their ends. These fractures could have been produced following two distinct processes: during the diagenetic compaction, as a result of the superposition of some elements on others (more than 90% of the bones of the assemblage are in contact with other bones) or due to the decompression produced as a consequence of the erosion of the sediments which covered the stratum. These orthogonal fractures are in all cases delimited by bone (i.e., there is no fracture defining the end of one specimen), what helps to clarify the taphonomic state of the assem-

blage (accumulated, resedimented or reworked), since the finding of bone fragments, isolated and delimited by this type of fractures, would imply taphonomic reelaboration (i.e., burial + mineralization + diagenetic breakage + unearthing by erosion + displacement, dispersal and/or destruction of certain bones + burial of different specimens from the same bone). When these processes take place, the bone assemblage is, at least in part, mixed diagenetically (i.e., reelaborated), which precludes the possibility of m a k i n g detailed palaeoecological analyses, since the bones m a y have different age and may come from different environments. Clearly, this is not the case here.

BIOSTRATINOMY OF THE ASSEMBLAGE: VARIATES, P R O C E S S E S AND AGENTS We have analysed different biostratinomic variates in order to characterize the macrovertebrate assemblage from Venta Micena. Previous approaches to the taphonomic study of this site (Gibert et al. 1992; Palmqvist et al. 1992) include the comparative study of the preservational state of the surface of the bones, and the evidence of the taphonomic bias experienced by the whole assernblage, respectively. In this study we have followed in part the procedure described by Behrensmeyer (1991) for the biostratinomic characterization of vertebrate assemblages. DIVERSITY, SPECIES ABUNDANCE, AGE CLASSES AND SIZE ESTIMATES The Venta Micena collection is composed of 6453 identifiable skeletal remains of 19 macromammalian species (_> 5 kg) and more than 10,000 unidentifiable bone shafts. Table 1 summarizes the raw data on the abundance of species (NISP: number of dental and non dental identifiable specimens; MNI: minimal n u m b e r of individuals; updated from Martlnez-Navarro 1991, 1992a). The palaeontological sample from this site increases each year during systematic excavations. Up to the present moment, the macromammalian assemblage is composed of 19 taxa, which belong to the orders Proboscidea, Artiodactyla, Perissodactyla and Carnivora, represented by the following families: Elephantidae, Bovidae, Cervidae, Hippopotamidae, Equidae, Rhinocerotidae, Canidae, Felidae, Hyaenidae and Ursidae. The presence of two other m a m m a l i a n species of small size, the badger (Meles sp.) and the porcupine (Hystrix major), has been cited in this site, although we have found no fossil remains of the former within the macrovertebrate assemblage. The first species would indicate the presence of mustelids in the palaeocommunity, and the porcupine (which is represented by an isolated maxilla)

12

Species

Mammuthus meridionalis Hippopotamus amphibius antiquus Bovini cf Dmanisibos Soergelia minor Praeovibos sp. Hemitragus alba Caprini gen. et sp. i~det.

Megaloeeros (Megaceroides) solilhacus Cervidae gen. et sp. indet. artiodactyla indet., size 2 artiodactyla indef., size 2-3 artiodactyla indet., size 3

Stephanorhinus etruscus Equus altidens herbivore herbivore herbivore herbivore herbivore

indet., indet., indet., indet., indet.,

size size size size size

2-3 3 3-4 4-5 5-6

Vulpes praeglacialis Canis falconeri Canis etruscus Lynx aff. issiodorensis Megantereon whitei Homotherium latidens Felidae indet., size 2-3

Pachycroeuta brevirostris Ursus etruscus carnivore indet., size 1 carnivore indet., size 2-3 m a m m a l indet., size 1 m a m m a l indet., size 2 mammal indet., size 2-3

NISP (teeth]bones) I 48 (16/32) 58 (19/39) 775 (382/393) 334 (215/129) 6 (3/3) 3O5 (2O9/96) 1(0/1) 962 (557/405) 417 (231]186) 12 (0/12) 9 (O/9) 91 (0/91) 90 (55/35) 2562 (1183/1379) 11 (0/11) 357 (0/357) 20 (0/20) 15 (0/15) 6 (0/6) 24 (19/5) 65 (40/25) 33 (20/13) 6 (2/4) 16 (7•9) 7 (6/1) 24 (0/24) 62 (34/28) 27 (15/12) 2 (O/2) 21 (0/21) 4 (O/4) 11 (0/11) 73 (0/73)

A* (%)

48.0 (0.1)

MNI (inf./ad.)

% infant.

Wmean (range)

Ir (%)

80.0 60.0 59.3 23.1 0.0 14.3 0.0 41.7 15.0

6033.8 (3713.1-9746.9) 3131.4 (2408.7-3978.8) 449.4 (371.4-745.0) 227.6 (127.6-324.0) 316.4 (195.2-388.8) 75.2 (57.2-119.4) 8.0-10.0 383.8 (269.0-567.8) 95.1 (63.5-170.6)

4.26 4.47 15.02 13.45 3.14 11.40

6263.0 (11.3) 7010.9 (12.6)

5 (4/1) 5 (3•2) 27 (16/11) 13 (3/10) 1 (0/1) 14 (2/12) 1(0/1) 36 (15/21) 20 (3/17)

229.7 (0.4) 17 622.1 (31.7)

6 (2/4) 70 (32/38)

33.3 45.7

1521.1 (1074.0-2176.0) 354.0 (243.8-486.7)

5.90 17.34

348O.3 (63) 2748.4 (4.9) 1789.0 (3.2) 501.3 (O.9) 400.8 (0.7) 62.0 (0.1)

1 (0/1) 3 (0/3) 4 (0/4) 1 (0/1) 3 (0/3) 2 (0/2)

0.0 0.0 0.0 0.0 0.0 0.0

3.0-5.0 8.24 29.8 (24.7-34.7) 9.9 (9.3-11.5) 2.83 8.0-10.0 2.13 52.9 (46.1-58.1) 243.9 (183.5-338.0)

1364.6 (2.5) 178.6 (03)

10 (4/6) 3 (1/2)

40.0 333

64.0 (57.8-70.7) 374.9 (309.2-451.6)

90,6 (0.2) 4532.2 (8.2) 3102.2 (5.6) 6015.6 (10.8)

13.99 10.91

7.44

1.89 1.24 2.46 3.04

TABLE 1 - A b u n d a n c e of t h e m a c r o m a m m a l species (_> 5 kg) identified in t h e V e n t a M i c e n a a s s e m b l a g e ( u p d a t e d f r o m M a r t f n e z N a v a r r o 1991, 1992a). N I S P : n u m b e r of identifiable s p e c i m e n s (teeth/bones). A*: a b u n d a n c e corrected for t a p h o n o m i c b i a s e s . MNI: m i n i m u m n u m b e r of i n d i v i d u a l s (juveniles/adults). W: e s t i m a t e d body w e i g h t (in kg) for t h e a d u l t i n d i v i d u a l s ( m e a n a n d r a n g e ) L: i n d e x of r e p r e s e n t a t i o n of s k e l e t a l p a r t s . Abondance des esp~ces de macromammi~res (_>5 kg) identifides dans l'assemblage de

Venta Micena (mise &jour de Martfnez-Navarro 1991, 1992a). NISP: nombre de spgcimens identifiables (dents / os). A *: abondance corrigde des biais taphonomiques. MNI: hombre minimum d'individus (]uv@niles / adultes). W: poids corporel estimd (en kg) pour des individus adultes (moyenne et extension). L: indice de prdsence des parties squelettiques.

stands out since it could be one of the biological collecting agencies of bones.

and three cervids (Criozetoceros ramosus, Eucladoceros cf. senezensis and Cervidae indet.); the

If we compare the taxonomic composition of the Venta Micena assemblage with those of another three Plio-Pleistocene lacustrine sites from the Guadix-Baza basin (Alberdi et al. 1989), we see that Venta Micena shows the highest diversity of large mammals, since each of the sites at Hu~lago (upper Pliocene), Hu@scar-1 (early middle Pleistocene) and Cfillar de Baza-1 (middle Pleistocene) only present 10 macromamalian species. The first of these sites lacks carnivores (which constitutes an anomalous peculiarity within the palaeomastologic record from Spain, probably due to the special taphonomic characteristics of this site), has one proboscidean (cf. Mammuthus meridionalis), two perissodactyls (Equus stenonis livenzovensis and Stephanorhinus cf. etruscus), four bovids (Leptobos cf. elatus, Gazella borbonica, Gazellospira torticornis and cf. Hesperidoceras merlae),

assemblage from Hu~scar-1 has four species of carnivores (Canis etruscus, Panthera gombaszoegensis, Homotherium sp. and H y a e n i d a e indet.), one proboscidean (Elephas antiquus), three perissodactyls (Equus stenonis interm, granatensis/ altidens, Equus sussenbornensis and Stephanorhinus etruscus), one cervid (Megaloceros [Megaceroides] cf. solilhacus) and one hippopotamid (Hippopotamus major); finally, the assemblage from Cfillar de Baza-1 has two carnivores (Canis etruscus and Crocuta crocuta), one proboscidean (Mammuthus trogontherii), three perissodactyls (Equus altidens, Equus sussenbornensis and Stephanorhinus etruscus), two bovids (Bison sp. and Capra sp.), one cervid (Dolichodoryceros savihi) and one suid (Sus cf. scrofa). These three sites are also located on sediments which were deposited in s w a m p y and lacustrine environments, under similar conditions to those ofVenta Micena

13 (without evidence of subaerial exposure in the assemblages). The main difference between the assemblage from Venta Micena and those recovered from other sites in the Guadix-Baza basin is the high diversity of carnivore species preserved in the former assemblage, which has all ecological niches of carnivores represented (from opportunistic scaw~ngers to top predators). On the other hand, the taxonomic richness of large mammals in Venta Micena is the same as that recorded in a modern hyaena den (Crocuta crocuta) from the Amboseli National Park (Kenya), which was also developed on a calcrete paleosol (Hill 1981, 1984), in which the skeletal remains from 16 to 18 taxa were represented. Table 1 shows that the herbivore species are those best represented in the Venta Micena assemblage by both NISP and MNI counts, with the following rank abundance: Equus altidens, Megaloceros (Megaceroides) solilhacus, Bovini cf. Dmanisibos, Cervidae gen. et sp. indet., Soergelia minor, Hemi-

tragus alba, Stephanorhinus etruscus, Hippopotamus amphibius antiquus, and Mammuthus meridionalis. The worst represented species are carnivores, in wlhich the most abundant species according to their NISP values are Pachycrocuta brevirostris and Canis (Xenocyon) falconeri (of which the hyaenid stands out if we also consider its high MNI value). The index of skeletal representation (L= [NISP~ 100]/[NEc~MNI~], where NISPi is the minimal number of identifiable specimens recovered from the "i" species, NE~ is the number of osseous elements found in the skeleton of a living individual of this species, and MNI~ is the minimal number of individuals estimated for species "i" in the assemblage) indicates for each species the percentage of skeletal elements which were preserved in relation to the sum of the total number of elements for all individuals identified in the assemblage (i.e., preservational completeness), and thus provides information about the extent of the taphonomic biases which affected the composition of the assemblage. For example, if an assemblage was produced by a catastrophic or mass mortality event, followed by immediate burial, the L index would then take the maximum value (100%) for each species, since all skeletal elements would have been preserved if the diagenetic conditions were not destructive. However, the species in Venta Micena are represented by less than 20% of potentially preservable elements (Fig. 6). The herbivore species may be ordered by decreasing L values as follows (Table 1): Equus altidens, Bovini cf. Dmanisibos, Megaloceros (Megaceroides) solilhacus, Soergelia minor, Hemitragus alba, Cervidae gen. et sp. indet., Stephanorhinus etruscus,

Hippopotamus amphibius antiquus, Mammuthus

meridionalis and Praeovibos sp. The main difference between this ordering and the previous one is that the individuals of Bovini cf. Dmanisibos are better represented by skeletal remains than those of Megaloceros, although the latter species has higher NISP and MNI values, and the same difference is found between Hemitragus and Cervidae indet., respectively. Carnivores are poorly represented in the Venta Micena assemblage, since most of these species have <5% values in the L index, with the exception of the fox (Vulpes praeglacialis) and the large canid (Canis [Xenocyon] falconeri), which are represented mainly by cranial elements. The age estimated for the individuals preserved in the assemblage has been classified within two major groups: immature or juvenile individuals with deciduos teeth, and adults with fully erupted permanent dentition. Inspection of data in Table 1 on the distribution of each species by juvenile and adult individuals reveals that: the species of largest body size (elephant, hippo and buffalo) are mainly represented by juvenile individuals, with the exception of the rhino; the species which are better represented by their NISP and MNI values (horse and large deer) also show high percentages of juvenile individuals (>40%); the carnivore species are represented exclusively by adult individuals, with the exception of the hyaenid and the ursid (40% of the individuals ofPachycrocuta brevirostris are juveniles, which are represented by deciduous teeth not isolated from the maxilla or mandible, which indicates that these cranial eleSpecies Vulpes . Canis etrusous. Canis fafconen. Lynx. Megantereon. Homotherium . Pachycrocuta. Ursus. Cervidae indet, • Megafoceros • Hemitragus . Soergefia . Praeovibos . Bovini indet.. Hippopotamus. Squus . Stephanorhinus. Mammuthus-

0

5 10 15 % preservational completeness

20

FrGURE 6 - Index of representation (Ir: percentage of preserved skeletal elements in relation to the original number of bones estimated from MNI counts) for the species of large mammals identified at Venta Micena. Indice de prdsence (It:

pourcentage d'dldments squeIettiques prdservds en fonction du nombre estimds d'os par ddcomptes MNI) pour les esp~ces de grands rnammifbres identifides & Venta Micena.

14

ments were not produced by tooth replacement, but as a consequence of the death of very young individuals). Body weights (W, in kg) for adults of each species (mean and range) were calculated from several regression equations of weight on craniodental/ postcranial variates in recent species (Janis 1990; Roth 1990; Scott 1990; Van Valkenburg 1990). The mean value estimated for each species and the widest range of weight estimations (i.e., the minim u m and the m a x i m u m of all values obtained with these regressions for craniodental and postcranial remains of a given species) have been included in Table 1. These estimates indicate that the Venta Micena assemblage comprises species weighing between 3-5 kg (fox) and 6000 kg (elephant, approximated range: 3500-10,000 kg), which constitutes a total range of body sizes in the palaeocommunity of more than three orders of magnitude. BONE ORIENTATION, SPATIAL DENSITY AND SKELETAL ARTICULATION The distribution in a rose diagram of the directions of the longitudinal axes of long bones in Venta Micena shows no preferred orientation or alignment (Gibert & Caporicci 1989), which suggests a r a n d o m p a t t e r n in the distribution (Shipman 1981). However, two factors must be borne in mind in the analysis of bone orientation: the presence of a water sheet in the environment in which the assemblage was accumulated, and the fact t h a t the skeletal elements could or could not experience free movements aligning with the direction of the currents (as a function of bone density and size, type of depositional interphase, microtopography, grain size and sediment density; Frison & Todds 1986). In Venta Micena the stratigraphic evidence indicates the absence of channelled currents in the area in which the site was formed, which was the plains that surrounded a lake, with numerous ponds emplaced on a paleosol where the majority of the fossils were concentrated. The marked microtopography, with height differences of up to 40 cm, must have determined the position of the bones on the substrate, as well as their possibilities to be displaced by w a t e r currents (of which there is no objective evidence in the sediment) or by trampling. Therefore, the bones at Venta Micena had fewer opportunities to experience free movements, since their position is determined by the geomorphology of the palaeorelief and the high density of osseous elements preserved in the site, which form an intricated mixture which is very difficult to excavate. The absence of preferred orientations in the bone assemblage is thus determined by the original characteristics of the substrate.

HUMERI •[] GNAWED BONES

....~

¢

FIGURE 7 - Bones outcropping at high d e n s i t y in one grid (3D8-1993) of t h e s u r f a c e e x c a v a t e d in V e n t a Micena.

Affleurement d'os & forte densitd sur un dIdment d u quadrillage (3D8-993) de la surface excavde ~t Venta Micena.

As was explained before, the concentration of bones in this site is very high, and more than 90% of the skeletal elements are in contact with other elements (Fig. 7). The assemblage shows a low degree of horizontal dispersion, with groups of dissarticulated b u t associated e l e m e n t s (i.e., skulls with mandibles, metapodials and phalanges), which represent 80% of all bones, and groups of articulated elements in a lower proportion (approximately 20%). The articulations more frequently preserved are those formed by tibiaet a r s a l - m e t a t a r s al-phalanges, h u m e r u s - r a d i u s / ulna, r a d i u s - c a r p a l - m e t a c a r p a l - p h a l a n g e s and articulated vertebrae. The Venta Micena site has provided a large amount of skeletal elements. The density plot for the distribution of fossils on the excavated surface of the quarry (number of bones and teeth recovered from each m ~) shows that the m e a n density of elements is approximately 60/m ~ (Fig. 8, obtained using the SYSTAT program, version 5.0), although two well defined areas have 80 or even 90 bones/m 2of up to 50 cm of length (i.e., tibiae of Equus and metapodials of Megaloceros). The number of elements per m s is in part a function of their size, since the density of elements reachs higher values in the squares which show a majority of postcranial bones (2-50 cm length), while the spatial density is lower in those squares in which cranial elements (60-80 cm length) predominate, because they have greater surface and volume than the former bones, although the surface of these squares is also almost completely covered by fossils.

15 FIGURE 8 - Density plot for the abundance of skeletal remains (number of bones and teeth per square meter) in the Venta Micena excavation area. Plan de densit#

pour l'abondance des restes squelettiqaes (hombre d'os et de dents par m~tre carrg) sur la surface excav#e de Venta Micena.

~.~

~0

vO

¢-q

O

0

2

REPRESENTATION OF SKELETAL PARTS AND BONE MODIFICATION The palaeontological collection recovered from Venta Micena is composed of 6453 identifiable skeletal elements of large mammals, in which all range sizes of complete elements and bone fragments are represented (i.e., from premolars and third phalanges of Vulpes to complete mandibles of Mammuthus). Fossil remains of micromammals, including teeth and elements from the axial skeleton, are also present in this site, although their study is not included in this work. Descriptive,, taphonomic analysis of the Venta Micena assemblage was based on a well restored sample of 1.339 specimens housed at the Museum of Paleontology of Orce, which represents a random sample of the Venta Micena collection (Palmqvist et al. 1996a). Isolated teeth represent 12.5%, and 1.4% are fragments of deer antlers. In the sample of bone remains (N = 1152; Fig. 9), limb bones dominate (64.7%), followed by vertebrae (15.5%), cranial elements (10.4%), phalanges (6.3%), and ribs (3.1%). The most abundant postcranial bo~Les (Fig. 10) are metapodials (>20%), tibiae and ihumeri (6-8%), calcanei and astragali (3-5%), escapuli, radii, femori, first and second phalanges (2-3%), pelvis fragments and third phalanges (4.3%), and ulnae (<1%). Scapuli are basically represented by proximal fragments; fragments of diaphyses predominate among fossil humeri; the most complete elements are radii; femori are mainly represented by fragments of diaphyses and tibiae by distal epiphyses; the pelvis is only represented by fragments which preserve the ac,etabulum (Fig. 11).

4

6

8

10

12

16m

14

CRANIAL ELEMENT~

VERTEBRATE

RIBS

UMB AND GIRDLE BONES

PHALANGES 0

20

40

60 %

80

100

FIGURE 9 - Representation of different types of bones in the Venta Micena equid and in the assemblage of large mammals (shaded area: abundance of these elements in a horse skeleton), compared with data obtained from six modern dens of Crocuta (Timbavati, South Africa; Brain 1981). Reprdsentation

des diff&ents types d'os de l~quidd de Venta Micena dans l'assemblage de grands mammi[~res (surface hachurge: abondance de ces dl4ments d'un squelette de cheval), comparge avec les donnges obtenues pour six repaires modernes de Crocuta (Timbavati, Afrique du Sud, Brain 1981).

The surfaces of the bones seem to have been exposed to the effects of subaerial weathering for a short time: 89.3% of the skeletal elements show weathering stage 0 (Behrensmeyer 1978), and only 10.7% of the bones (of which two thirds are metapodials) present weathering stage 1, with few, shallow and small split line cracks due to insolation (1-8 in each bone), and without flaking of their outer surface. These results indicate that subaerial weathering was relatively unimportant.

16 FIGURE 10 - Representation of limb and girdle bones in the Venta Micena equid and in the assemblage of large mammals; shaded area: abundance of these bones in a horse skeleton (ESC: escapuli, HUM: humeri, RAD: radii, ULN: ulnae, FEM: femori, TIB: tibiae, AS: astragali, CA: calcanei, MP: metapodials, PH1-PH3: phalanges). Reprdsentation des os de membres et de ceintures de l~quidd de Venta Micena et dans l'assemblage de grands mammif~res; surface haehurde: abondance de ces os dans un squelette de cheval (ESC: scapulaires, HUM: humdrus, RAD: radius, ULN: ulnaires, FEM: fgmurs, TIB: tibias, AS: astragales, CA: calcaneums, ME: mdtapodiaux, PH1-PH3: phalanges.

35 ] I ] 30 25 TOTAL VM

20 ]EQUUS

15

--=- Equus VM

10

ESC

HUM

RAD

ULN

A

I ~ SHAFT DISTAL END PROXIMAL END COMPLETE

0

20

40

60 %

80

100

B HUM [] [] I [] I

ULN FEM TIB 0

20

40

60 %

80

100

FEM

TIB AS CA Postcranialbones

MP

PH1

PH2

PH3

The bones which were preserved complete, lack sedimentary filling, even in those areas of the medullary cavity which are close to nutrient foramina, which indicates that the bones were buried with the periosteum intact. Figure 12 shows the distribution of percentages of bones per weathering stages in Venta Micena and in several control bone assemblages and carcasses of known age since death (Behrensmeyer 1978; Gifford 1977, 1984). The position of Venta Micena in this diagram is close to those of the least w e a t h e r e d assemblages, thus indicating a very short period of subaerial exposure before burial (<1 year). Horse (E. altidens) remains (N = 457) are not well dispersed horizontally: groups of articulated elements represent nearly 20% of this sample, and the remaining 80% of non-articulated bones are found

ESC

RAD

PEL

SHAFT DISTAL END PROXIMALEND COMPLETE

FIGURE 11 - Relative abundance of several limb bone fragments (diaphyses, distal epiphyses and proximal epiphyses) and complete elements (ESC: escapuli, HUM: humeri, RAD: radii, ULN: ulnae, FEM: femori, TIB: tibiae) in the Venta Micena assemblage (A), compared with their frequencies in the sample of equid bones (B). The first graph includes the bones from carnivores and herbivores, while the second one shows only those skeletal elements of one herbivore species; the latter graph differs substantially from the former in the greater abundance of proximal humeri, femoral diaphyses and distal tibiae. Abondance relative de plusieurs fragments d'os de membres (diaphyses, @iphyses distales et @iphyses proximales) et d'ddments complets (ESC: scapulaires; HUM: humdrus, RAD: radius, ULN: ulnaires, FEM: f#murs, TIB: tibias) dans l'assemblage de Venta Micena (A) comparge avec leur fr#quence dans l'gchantillon des os d'dquid~ (B). Le premier graphe comporte les os de carnivores et d'herbivores alors que le second montre seuIement des dldments squelettiques d'une esp~ce d'herbivore; ce graphe diff~re sensiblement du premier par l'abondance plus grande de parties proximales d'humdrus, de diaphyses f#morales et de parties distales de tibias.

17 100% weathering stages 4-5

FIGURE 12 - Frequency distribution of percental~es of bones per weathering stages in Venta Micena (data from Palmqyist et al. 1996a), Olduvai Gorge (data from Potts 1982) and several control bone assemblages a n d carcasses of k n o w n n u m b e r of years since d e a t h (data from B e h r e n m e y e r 1978; Gifford 1977, 1984). Distribution de frdquence des pourcentages d'os clans les couches altdrdes de Venta Micena (donndes d'apr~s Palmqvist et al. 1996), la gorge d'Olduvai (donndes d'apr~s Ports 1982) et plusieurs assemblages d'os et de carcasses de contr6le pour lesquels le nombre d'anndes depuis la mort est connu (donndes d'apr~s Behrenmeyer 1978; Gifford 1977, 1984).

7-15

1O-15 years Q

landscape 5-7

hyaena den <1 year, 1-3 years • FLK 22, N6 • FLK NN2

Z//VVM

1"00% weathering stages 0-1

associated. Biostratinomic fractures are abundant (Figs 13-i[5): only 29.1% (73/251) of major long bones are complete (most of them metapodials), and Type II spiral fractures (Shipman 1981; Lyman 1994) are predominant (100% of fragmented humeri, femuri and radii, 74.4% of tibiae); other types are longitudinal fractures (25.6% of tibiae), undifferentiated fractures (all ribs and vertebrae, with the exception of several vertebrae which only lack their apophyses) and isolated maxilla with both cheek tooth rows (33.3% of cranial elements). On the one hand, the outer surface of the bones is well preserved: no one bone shows signs of abrasion or polish (with the exception of four out of six petrosus bones), and only four elements of the sample (0.9%) are slightly dissolved. On the other hand, gnawing marks are very frequent (Figs 16-18): all cranial fragments, scapuli, humeri, radii, pelvis, femori and tibiae show striations and biting marks produced by carniw)res, the preserved epiphyses have furrows and punctures, and the diaphyses, as well as the skull bones, show scoring and pitting. These marks are also observed in all other taxa identified at Venta Micena. Coprolites are relatively common in Venta Micena, preserved as isolated lobes with diameters between 3 and 6 cm. The results presented up to this moment, allow us to deduce that the very rich and diverse palaeon-

2.5-3years -

100%

4-10years 3-5 years ® •

\

k

weathering stages 2-3

tological site at Venta Micena was formed in the dessicated border of a lake, and that the geological agents m a y be excluded from the bone accumulation process. The geochemical data from the samples obtained in the sediment and in the fossils, as well as the diagenetic information obtained from the analyses of the fossils, are both in keeping with each other, thus corroborating the initial impression that the assemblage represents an accumulated taphonomic stage, without objective evidence of taphonomic reworking. The bone assemblage is composed of demic palaeobiological entities (hyaenids, as it will be shown later) and autochtonous preserved entities (the remaining species), which were recorded in situ (sensu Fern~ndez-L6pez 1991). The results obtained in all biostratinomic analyses indicate clearly that this assemblage shows a differential bias in the representation of certain anatomical parts and bones in relation to others, and that the accumulation of bones was generated by the activity of scavenger carnivores, presumably hyaenids, being buried in a short period of time.

QUANTITATIVE TAPHONOMY The quantitative taphonomic study of the Venta Micena assemblage was based on three different approaches: the analysis of size/abundance pat-

18

2

FIGURE 13 - Sequence of consumption by hyaenids of humeri (1-5) of Equus altidens, showing gnawing of epiphyses and type II spiral fractures. Sdquence de la consommation par des hy~nes d'humgrus (1-5) d'Equus altidens montrant le rongement des dpiphyses et le type H des fractures spirales.

terns in the ungulate species using the model proposed by D a m u t h (1982); the analysis of the abundance of preserved epiphyses and complete major long bones of r u m i n a n t s in the assemblage, and; the s t u d y of bone frequencies in the assemblage in comparison to some recent and archaeological deposits accumulated by carnivores, rodents and hominids. ANALYSIS OF SIZE-ABUNDANCE PATTERNS In a fossil assemblage taphonomically unbiased, the abundance of skeletal remains for a given species will depend exclusively on the number ofindividuals that died while the assemblage was being formed. The relative abundance of each species

(A) is estimated as a function of both its mean population density (D) in the palaeocommunity and its population turnover rate (Tr): A = f(D, T,) (Damuth 1982). Several studies (Damuth 1981, 1987, 1991; Peters 1983; Calder 1982, 1984) have shown t h a t population density (number of individuals per unit area) of a species is inversely and allometrically related to body weight (W) by a slope close to the -0.75 power: D = K1W"°'Ts. With respect to the population turnover rate, related parameters as the birth rate, the duration of postnatal growth or the reciprocal of life expec-

19

cm m

0

3

FIGURE 14 - Sequence of consumption by hyaenids of tibiae (1-6) ofEquus altidens, showing gnawing of epiphyses, spiral and longitudinal fractures. Sdquence de la consommation par des hy~nes de tibias (1-6) d'Equus altidens montrant le rongement des dpiphyses et des fractures spirales et longitudinales.

tancy at birth (Western 1979, 1980; D a m u t h 1982; Peters 1983; Calder 1982, 1984) are allometrically related to species body size, with a slope of around -0.3: Tr = I~W ~°3. As a function of both equations, the original abundance (A) of bones from the different species present in a fossil assemblage is determined by: A = DTr = K1W-°7~K2W-°-3= K3W-1-°5, with a 95% confidence interval for the slope ranging between -0.8 and -1.3 (Damuth 1982).

If the value of the slope t h a t relates A with W in a fossil assemblage is included within this interval (-1.05 _+ 0.25), it could be then concluded t h a t the assemblage h a d not experienced significant taphonomic biases with respect to the quantitative composition of the original palaeobiocoenoses, and t h a t the community structure was preserved during fossilization. If, on the contrary, the value obtained for the slope lies well outside the confidence interval, it should be then deduced t h a t the relative frequencies of the species represented in the bony accumulation do not fit the size/abundance relationship t h a t characterizes recent communities, and t h a t the original composition of the

20

4

cm m

0 3

3

5

FIGURE 15 - Third metacarpal (4) and metatarsals (1-3, 5, 6) ofEquus altidens, complete and fractured by hyaenas, showing orthogonal diagenetic fractures in the diaphyses (1-5), breakage by sediment pressure (4), and longitudinal (5) and spiral (6) fractures made by hyaenid crushing. Troisi~mes mdtacarpien (4) et mgtatarsiens (1-3, 5, 6), complets et fractur4s par des hy~nes, montrant des fractures diagdngtiques orthogonales sur les diaphyses (1-5), la cassure par la pression du sgdiment (4) et des fractures longitudinales (5) et spirales (6) faites par le broyage des hy~nes.

assemblage was then biased by taphonomic processes (or by other factors, such as sampling or curating errors). Body size is one of the factors that seems to have the greatest influence on the fossilization potential of terrestrial vertebrates (Behrensmeyer et al. 1979; Behrensmeyer & Dechant Boaz 1980; Damuth 1982), since the bones of large-bodied species are more resistant to processes of physico-chemical weathering and biological destruction (i.e., exposure to sun radiation, salt precipitation, changes in relative humidity, trampling by ungulates, carnivore gnawing, root growth, etc), due to their smaller relative surface (the ratio of their outer surface to the enclosed volume). To correct for taphonomic loss due

to these processes during the period when the bones were exposed on the surface before their definitive burial, the value d~ = 0.68[Log(Wm)-Log(W~)] should be calculated, where Wl is the body weight estimated for species "i", and Wm is the weight of the largest species in the assemblage. In this way, the amended original abundance (&*) of each species will be estimated (Damuth 1982) as: Log(A~*) = Log(/~)+di. Before the model can be used for the quantitative taphonomic analysis of the Venta Micena palaeocommunity, some considerations are in order. Firstly, the analysis must be restricted to a single trophic level, since the population densities decrea-

21

cm m

0

2

FIGURE 16 - Maxillae of Bovini c£ Dmanisibos (1), Equus altidens (2, 3) and Megaloceros (Megaceroides) solilhacus (4), gnawed by hyaenids, which

show an extreme destruction of the splanchnocranium and neurocranium. Maxillaires de Bovini cf. Dmanisibos (1), Equus altidens (2, 3) et Megaloceros (Megaceroides) solilhacus (4), rongds par des hy~nes montrant une destruction extreme du splanchnocrdme et du neurocr~ne.

se for a given body size as height in the energetic or ecological p y r a m i d increases. P r i m a r y c o n s u m e r species are, as a rule, the m o s t suitable for this analysis, since t h e y are a b u n d a n t in communities, are highly diverse, and differ considerably in size (at least two orders of m a g n i t u d e in size differences are n e c e s s a r y to obtain a statistically reliable fit to the

model). In the Venta Micena a s s e m b l a g e 5558 dental and non-dental r e m a i n s h a v e been identified from 11 species of herbivorous m a c r o m a m m a l s , whose e s t i m a t e d weights r a n g e from a p p r o x i m a t e l y 8 to 6000 kg. Carnivores are less r e l e v a n t for this analysis, as a function of t h e i r scarcity in the original communities and, as a consequence, in the fossil

22

Clll m

0

2

cnl I

0

2

FIGURE 17 - Hemimandibles (1, 2) and pelvis (3, 4) ofEquus altidens, gnawed and fragmented by hyaenids. Hdmimandibules (1, 2) et pelvis (3, 4) d'Equus altidens rongds et fragmentds par des hy~nes.

assemblages (264 remains in Venta Micena, cortesponding to 8 species ranging in body weight from 5 to 375 kg), which increases the randomness of their sampling (Wolf 1975; Palmqvist 1991, 1993; McKinney 1996). Secondly, the total number of skeletal remains (NISP) of each species is the best estimate of its abundance (A) in the assemblage, since

4

the counts of minimal number of individuals (MNI) tend to decrease the values of the common species and to overestimate those of rare species. Thirdly, analysis of only part of the assemblage is still valid, and species about which there is some uncertainty (such as would result from our inability to consistently identify their specimens) may be omitted

23

FIGURE 18 - Tarsal bones (calcanei [1-3] and astragali [4-6]) and third phalanges (7-10) ofEquus altidens. One calcaneum shows marks made by insect larvae (1); two calcanei (2, 3), one astragalus (4) and one phalanx (9) have diagenetic fractures; one calcaneum (3), three astragali (4-6) and two phalanges (9-10) are partially gnawed by hyaenids. Os tarsiens (calcaneums [1-3], astragales [4-6]) et troisi~mes phalanges (7-10) d'Equus altidens. Un calcaneum montre des marques faites par des larves d'insectes (1); deux calcaneums (2, 3), un astragale (4) et une phalange (9) ont des fractures diagdndtiques; un calcandum (3), trois astragales (46) et deux phalanges (9-10) sont partieUement ronggs par des hy~nes.

from the analysis with no loss of accuracy (Damuth 1982). The values of Log(A) and Log(W) for the Venta Micena herbivore species (N=9) are shown in Figure 19A, in which the regression line obtained for both variates using for the adjustment the least squares method is also included. Two species were excluded from the analysis: P r a e o v i b o s sp., a mountain-dwelling species is poorly represented in the assemblage, given that it was d e a r l y allochthonous to the palaeocommunity; Caprini indet, is represented by only one distal epiphysis of a metacarpal. The fit obtained in the adjustment is statistically significant at the 95% confidence level, but the slope is not included within the range of values anticipated by Damuth's model (-1.05 _+0.25):

ce of taphonomic biases (Fig. 19B), the fit obtained is statistically significant at the 99.9% confidence level, and the slope is within the r a n g e of values predicted by the model: Log(A*) = 15.367(_+ 1.626) - 1.278(_+ 0.253) Log(W); r = -0.885, F = 25.422 (p < 0.001) These results confirm t h a t the main loss of information during the taphonomic history of the Venta Micena assemblage was due to destruction of skeletal remains during the period w h e n the bones were exposed on the surface before their definitive burial, and t h a t the effects of this information loss were g r e a t e r in species of smaller body size, thus biasing their q u a n t i t a t i v e representation in the fossil assemblage.

Log(A) = 9.448(_+ 1.626) - 0.598(_+ 0.253) Log(W); r = -0.665, F = 5.564 (p < 0.05).

ABUNDANCE OF E P I P H Y S E S AND COMPLETE MAJOR LONG BONES

In contrast, when we analyse the values for corrected a b u n d a n c e (A*) of the species in the absen-

Two factors seem to have a substantial influence on the frequencies of skeletal parts in an u n t r a n s -

24 F I G U R E 19 - Relationship between bodyweight (W,in kg) and abundance of skeletal remains (A: original abundance = NISP; A*, abundance corrected for taphonomic biases) in the ungulate species (N = 9) of the Venta Micena assemblage (A: regressionline obtained for A vs. W, adjusted by minimum squares analysis; B, regression line for A* vs. W). a, Hemitragus alba, b, Cervidae gen. et sp. indet., e,

Soergelia minor, d, Equus altidens, e, Megaloceros (Megaceroides) solilhacus, f, Bovini cf. Dmanisibos, g, Stephanorhinus etruscus, h, Hippopotamus amphibius antiquus, i, Mammuthus meridionalis. Relations entre le poids corporel ( W e n kg) et l'abondance des restes de squelettes (A: abondance originelle = NISP; A*, abondance corrig#e des biais taphonomiques) pour les esp~ces d'ongulds (N = 9 ) de l'assemblage de Venta Micena (A: droite de rdgression obtenue pour A vs. W, ajustde par l'analyse des carr~s minimaux; B, droite de r@ression pour A*

O

0

'

I

'

I

'

I

'

I

i

• d

"-/cO m 'Oa

OC

h 03

i

4

I

5

I

I

6

I

I

7

Log(Weight)

,

I

8

•J ,

i i

i

5

i

l

6

,

I

7

Log(Weight)

~

I

8

9

VS. W).

ported assemblage (Binford 1978, 1981; Binford & B e r t r a m 1977; Blumenshine & Madrigal 1993; L y m a n 1984, 1994): structural density of bones and within-bone n u t r i e n t utility. Overall meat/ bone utility plus ease of disarticulation are also i m p o r t a n t factors affecting skeletal part representation in accumulations transported by carnivores (see L y m a n 1994, and references therein). Several taphonomie processes, such as carnivore damage (Marean & Spencer 1991), are mediated by the structural density of bones (mechanical and chemical attrition have greater effects on bones with low bulk density), and by the extraction of n u t r i e n t s within bones, particularly marrow. M a n y structural density estimates, obtained using photon absorptiometry, are available in the literature (see L y m a n 1994), but the estimation of within-bone n u t r i e n t s is r a t h e r problematic. Interspecific differences in the distribution of flesh between the hindlimb and forelimb of bovids seem to be related to locomotor type (Blumenschine & Caro 1986). In a similar way, absolute long bone marrow yields differ among species, and among individuals of different size within a species. Most published d a t a (Binford 1978; Blumenshine 1991; Emerson 1990; Jones & Metcalfe 1988) are clearly inadequate for assessing this variability. Only B l u m e n s h i n e and Madrigal (1993) have addressed this issue, using a broad sample of East African ungulates. Figure 20 shows the abundance of long bone epiphyses of r u m i n a n t s in the Venta Micena assemblage and their m e a n bone mineral density (estim a t e d from values obtained using photon absorptiometry in modern bison and deer; L y m a n 1994). The positive relationship observed between both

variates is statistically significant, according with a least squares regression adjustment: Abundance = -68.522(_+23.007) + 270.841 (_+56.458) Density; r = 0.835, F = 23.01 (p < 0. 001). Similarly, Marean & Spencer (1991) found a strong correlation between the frequencies of long bone parts of domestic sheep t h a t survived gnawing by spotted h y a e n a s and their structural density. Figure 21 shows the abundance of major long bone epiphyses of ruminants and their mean fat content (estimated from values for moder bison in Brink 1997). The inverse relationship detected between both variates is also very significant, according with a least squares regression adjustment: Abundance = 70.086(_+7.665) - 0.249(_+0.051) Fat weight; r = -0.839, F = 23.71 (p < 0.001). However, the estructural density of the bone is inversely related with its fat content, which is higher in those epiphyses in which the less denser, cancellous tissue predominate. Figure 22 shows a plot of density and fat content of epiphyses (x and y-axes), where the abundance of long bone epiphyses is represented by contour lines (z-axis) using 20 intervals of increasing value. A multiple regression approach to the estimation of epiphyses as a function of both their density and fat contant has shown a statistically significant relationship: Abundance = 0.706(_+59.438) + 138.176(_+117.883) Density - 0.137(_+0.108) Fat weight; r = 0.862, F = 13.002 (p < 0.002), in which the fat content of the epiphyses is a somewhat better predictor of their abundance, as

25 100

100

d-tib

d-fib

e

©

Abtmdance = -68.522(±23.007) + 270.841(+_56.458)Density r = 0.:835, F = 23.0 l(p = 0.001)

75

.

ga

)- m c

75. •

d-rot

d-hum

.-mr .d-rot ,,d-rot d-mc" d-hum Abundance = 70.086(-+7.665) - 0.249(_+0.051) Fat weight, r =- 0.839, F = 23.71(p = 0.001)

O

©

oFs0.

o 50

.p-rad • p-rad d-rad

==

• d-rad

25"

25,

p -tib. p-fem, 0'|

0.2

q

.

.

"

.

0.3 0.4 0.5 Bone mineral density (g/cm3)

J

0.6

FIGURE 20 - N u m b e r of epiphyses (d: distal, p: proximal) of long bones (hum: h u m e r u s , rad: radius, mc: metacarpal, fern: femur, tib: tibia, rot: metatarsal) of r u m i n a n t species in the Venta Micena assemblage (data from Martinez-Navarro 1991) and m e a n bone mineral density (average of bison and deer; values obtained from L y m a n 1994: Table 7.6). The regression line was adjusted by l e s s t squares analysis. Nombre d'dpiphyses (d: dis-

tales, p: proximales) d'os longs (hum: humerus, rad: radius, me: mgtacarpien, fern: fgmur, tib: tibia, rot: mOtatarsien) d'esp~ces de ruminants de I'assemblage de Venta Micena (donnges d'apr~s Martfnez-Navarro 1991) et moyenne de la densitd minOrale (moyenne de bison et de cerf,; valeurs obtenues d'apr~s Lyman 1994: Tableau 7.6). La droite de r~gression a dt~ ajustde par l'analyse des moindres carr~s.

it m a y be deduced from the values of the standardized coefficients (0.428 and -0.460 for bone density and fat content, respectively). This is confirmed by the values obtained for the first order partial correlation coefficient of the abundance of epiphyses on their estimated estructural density being constant their fat content (r = 0.363), and of the abundance of epiphyses on their fat weight with independence of their density (r = -0.390). The abundance of major long bones of r u m i n a n t s in Venta Micena (complete limb bones and long bone cylinders, lacking one or both epiphyses) and their m e a n wet weights of marrow content (esti-

i

0

• p-ti

\d-b'Qm

p-fern, 100

p-h 200 Fat weight (g)

300

400

FIGURE 21 - N u m b e r of epiphyses (d: distal, p: proximal) of long bones (hum: humerus, rad: radius, me: metacarpal, fern: femur, tib: tibia, mr: metatarsal) of r u m i n a n t species in the Venta Micena assemblage (data from Martlnez-Navarro 1991) and mean bone fat content (estimated from modern bison in Brink 1997: Table 3). The regression line was adjusted by least squares analysis. Reconstruction of Pachycrocuta brevirostris d r a w n by Mauricio Antdn (from T u r n e r & Ant6n 1996). Nombre d'dpi-

physes (d: distales, p: proximales) d'os longs (hum: humdrus, rad: radius, mc: mgtacarpien, fern: fgmur, tib: tibia, rot: mgtatarsien) d'esp~ces de ruminants de l'assemblage de Venta Micena (donndes d'apr~s MartEnez-Navarro 1991) et moyenne du contenu gras des os (estim4e d'apr~s le bison moderne in Brink 1997, tabl. 3). La droite de r4gression a dtg ajustge par l'analyse des moindres carr4s. Reconstitution de Pachycrocuta brevirostris dessin~e par Mauricio Ant6n (d'apr~s Turner & Ant6n 1996).

mated from values for modern wildebeest in Blumenschine & Madrigal 1993) is shown in Figure 23. There is an inverse relationship between the abundance of each long bone and its marrow content, which is linearized by means of logarithmic transformation as: Log(Abundance) = 9.249(_+0.819) - 2.033(_+0.248) Log(Marrow content); r = -0.972, F = 67.29 (p < 0.001). These results indicate the preferential consumption by hyaenas of low-density skeletal parts, dif-

26 FIGURE 22 - Density plot of abundance of long bone epiphyses (d: distal, p: proximal, hum: humerus, rad: radius, mc: metacarpal, fern: femur, tib: tibia, mr: metatarsal) represented by contour lines (zaxis) on mean estructural density (x-axis) and fat content (y-axis). Reconstruction of the face of Pachycrocuta brevirostris drawn by Mauricio Ant6n (from Turner & Ant6n 1996). Graphe de la densitd de l'abondance d'dpiphyses d'os longs (d: distales, p: proximales, hum: humdrus, rad: radius, mc: mdtacarpiens, fern: fdmur, tib: tibia, rot: mgtacarpien) reprdsentde pal" des lignes de contour (axe z) sur lu moyenne de densitd structurale (axe x) et le contenu gras (axe y). Reconstitution de la t~te de Pachycrocuta brevirostris dessinde par Mauricio Ant6n (d'apr~s Turner & Ant6n 1996).

6

t

I

I

I

5

Abundance of major long bone epiphyses:

~20-40

,

2

-1.4

ferential fragmentation leading to less dense bone parts being crushed into unidentifiable fragments in contrast to denser bone parts simply broken into small b u t recognizable pieces during the extraction of marrow (Lyman 1994). The selective transport of certain anatomical parts and the preferential breakage by hyaenas of the richer marrow bones were thus major factors biasing the composition of the Venta Micena assemblage. We can evaluate the extent of this taphonomic bias, which affected the original composition of the Venta Micena assemblage, by comparing the abundance of r u m i n a n t s and equids which may be deduced from the n u m b e r of complete metapodials, the most a b u n d a n t major long bones preserved complete in the fossil assemblage, with the raw abundance of both groups of ungulates, which is easily estimated using MNI counts obtained from dental elements (Table 1). Ruminants are represented in Venta Micena by 87 metapodials, complete or long bone cylinders lacking only one epiphysis (44 metacarpals and 43 metatarsals), and equids by 136 metapodials (60 metacarpals and 76 metatarsals). The ratio of ruminants to equids is thus 0.640 (87/136, approximately 1 ruminant: 1.6 equids). However, ruminants are

-1.0 Log(density)

-0.6

represented in the assemblage by a higher MNI (111) than equids (70), with MNI estimated from teeth counts, and the ratio is therefore 1.586 (111/70, approximately 1 ruminant: 0.6 equids). Both estimations differ widely, but the latter is the most reliable of them, given t h a t we expect no great differences in fossilization potential of teeth from ruminants and equids, while the preservation of complete long bones is highly biased by differential gnawing and fragmentation by h y a e n a s as a function of their structural density, fat and marrow content, as it was demonstrated before (Figs 20-23). Total marrow yields of equid bones are on average five-fold smaller t h a n those from ruminants (Blumenshine & Madrigal 1993): specifically, the mean weight of marrow content from the metapodials of Burchell's zebra is 6.22 g, while this estimate is of 14.23 g for wildebeest metapodials. The Venta Micena equid (E. altidens) was similar in size to modern E. grevy, with an estimated weigth of approximately 350 kg (Palmqvist et al. 1996a). Given the fact that E. burchelli weigths on average 230 kg, we can estimate by simple linear interpolation a mean marrow content of 9.47 g for the metapodials of E. altidens. This means t h a t one

27 FIGURE2 3 - Number of major long bones (complete limb bones and long bone cylinders, lacking one or both epiphyses) of ruminant species in the Venta Micena assemblage (data from Martfnez-Navarro 1991), and mean wet weights (in g) of marrow content (estimated from values for modern wildebeest in Blumenschine & Madrigal 1993: Tabie 2). The cuzare was adjusted by least squares analysis. Reconstruction of Pachycrocuta brevirostris drawn by Mauricio Antdn (from Turner & Antdn 1996). Nombre d'os longs majeurs

Raw abundanceof major long bone epiphyses 5O \ Log(Abundance)= 9.249(+_0.819)- 2.033(+_0.248)Log~larrowcontent)

(os compIets de membres et cylindres d'os longs sans l'une ou les deux @iphyses) d'esp~ces de ruminants de l'assemblage de Venta Micena (donndes d'apr~s Martlnez-Navarro 1991) et moyenne des poids frais (en g) du contenu de moelle (estimations d'apr~s des valeurs ])our le gnou in Blumenschine & Madrigal 1993, tableau 2). La courbe a dtd ajustde par l'analyse des moindres carrds. Reconstitution de Pachycrocuta brevirostris dessinde par Mauricio Antdn (d%er~s Turner & Ant6n 1996).

30

/

I

H

/

r =-0.972, F = 67.29 (p < 0.001)

\ ~metacarpals 4O

20" humerii

10

radii'\N femori

0

.

.

.

.

N~ ~', ..........

etlblae i

0

10

20

30 40 50 60 Mean marrow wet weight (g)

i

70

80

metapodial of the Venta Micena equid had 1.503 times less m a r r o w content t h a n a metapodial of r u m i n a n t (the r u m i n a n t species identified in the assemblage range in body size from 8 kg for Caprini indef, to 450 kg for Bovini cf. D m a n i s i b o s ; the estimation of 14.23 g for the metapodials of m o d e r n wildebeest, a species with a body weigth of 170 kg, is thus appropriate as an average for all r u m i n a n t s preserved in the fossil assemblage).

approximately 1 ruminant: 0.7 equids), which is very close to that calculated from MNI counts (1 ruminant: 0.6 equids), and thus confirms that differential fragmentation by hyaenas was the major factor biasing the representation of r u m i n a n t postcranial bones in Venta Micena.

Using the equation that relates the abundance of complete long bones with their marrow content [log(abundance) = 9.249 - 2.033 log(marrow weigth)] it is possible to predict t h a t a decrease by a factor of 1.503 in the mean marrow content of a given long bone would translate in an increase by a factor of 2.290 of its abundance in the assemblage; in other words, if the m e a n marrow content of the r u m i n a n t metapodials would have been similar to that of the equid ones, the abundance of the former in the assemblage should be of approximately 199 complete elements, instead of 87. This figure gives a ratio of r u m i n a n t s to equids of 1.463 (199/136,

To determine the agents t h a t were mainly responsible for the initial accumulation of bones at Venta Micena, we compared known frequencies of different types of postcranial bones in various fossil and recent assemblages (Table 2). This comparative data base includes bones exposed on the surface or partially buried in Amboseli National Park (Kenya), assemblages found at open feedingplaces of several carnivore species, those from dens and lairs used by leopards, h y a e n a s and porcupines, and also the bones p r e s e n t in both recent h u m a n camps and archaeological assemblages. Bones were clustered in four groups (I: vertebrae;

MULTIVARIATE ANALYSIS OF BONE FREQUENCIES

28 II: ribs; III: limb and girdle bones; IV: phalanges) to facilitate comparisons between data from different bibliographic sources (many of which included complete lists, with the abundance of each skeletal element, but in several references the bones were already clustered in groups similar to those used here). The teeth, cranial elements and hemimandibles were excluded from the analysis, as their frequencies were highly variable (depending, for example, on whether hyaenas could or could not fracture the skulls), and such elements did not show any characteristic pattern that allowed discrimination of assemblages. Figure 24 illustrates the results obtained in a correspondence analysis of the frequencies of the four types of bones (groups I-IV) in these assemblages. This multivariate statistical method was used because it is the most suitable for data on frequencies such as those in contingency tables (Reyment

Assemblages

I

& JSreskog 1993). The two first correspondence axes explain more than 95% of the original variance. The assemblages we analysed show a parabolic or horseshoe distribution in the bivariate scatterplot for correspondence axes I and II (Guttmann effect), which expresses a quadratic relationship between the latent vectors and indicates, among other things, a strong gradient in the frequencies of the different types of bones. The first factor is directly correlated with the abundance of vertebrae and ribs, the least persistant bony elements, with higher factor scores corresponding to assemblages in which these elements are predominant (bones exposed on the surface of Amboseli Park, in open carnivore feeding-places, in the porcupine lair and in the hunter-gatherer camp). This axis is inversely correlated with the frequencies of both phalanges and limb and girdle bones, elements that are more abundant in the remaining assemblages. The

II

III

IV

N

References

35.7% 32.9%

23.7% 15.5%

34.8% 42.1%

5.8% 9.5%

13551 624

Behrensmeyer & Dechant-Boaz 1980 Behrensmeyer & Dechant-Boaz 1980

30.7% 24.2% 23.2% 26.1%

23.3% 19.0% 17.2% 15.6%

35.5% 39.8% 41.7% 40.7%

10,5% 17.0% 17.9% 17.6%

1080 483 2334 804

Richardson Richardson Richardson Richardson

22.0% 25.2% 12.9% 8.3% 12.8% 10.7% 7.0% 3.7% 21.2% 2.6% 35.9%

0.0% 4.9% 10.4% 15.8% 0.9% 1.5% 0.0% 0.7% 5.0% 0.7% 5.9%

473% 40.5% 70.3% 69.4% 83.6% 76.2% 90.8% 93.2% 67.5% 84.5% 49.9%

30.7% 29.4% 6.4% 6.5% 2.7% 11.6% 2.2% 2.4% 6.3% 12.2% 8.3%

114 163 202 552 329 206 357 295 80 2139 726

Brain 1981 Brain 1981 Behrensmeyer & Dechant-Boaz,1980 Bunn 1982 Brain 1981 Skinner et al. 1986 Kerbis-Peterhans & Kolska-Horwitz 1992 Skinner et al. 1980 Scott & Klein 1981 Klein & Cruz-Uribe 1984 Brain 1980

30.1% 9.3% 9.9% 22.6% 7.1% 2.8% 12.7%

19.8% 22.5% 21.6% 10.6% 5.3% 4.1% 10.0%

41.1% 65.4% 57.8% 49.9% 45.8% 44.4% 54.7%

9.0% 2.8% 10.7% 16.9% 41.8% 48.7% 22.6%

601 755 1616 4082 1408 2324 885

Bunn 1982 Brain 1981 Valensi 1991 Brain 1981 Klein & Cruz-Uribe 1984 Klein & Cruz-Uribe 1984 Villa et al. 1986

9.3% 8.8% 42.7% 36.7% 33.6% 23.8%

5.0% 6.4% 8.8% 17.4% 21.4% 6.0%

79.3% 79.5% 47.4% 43.3% 35.7% 58.9%

6.4% 5.3% 1.1% 2.6% 9.3% 11.3%

3208 342 654 1900 571 302

Martinez-Navarro 1992a Bunn 1982 Bunn 1982 Bunn 1982 Bunn 1982 Bunn 1982

Samples from modern ecosystems: 1 2

Bones exposed in Amboseli Park Bones buried 50% in Amboseli Park

Carnivore feeding places: 3 4 5 6

Brown and spotted hyaenas Lions Feral dogs Jackals

1980 1980 1980 1980

Carnivore dens and rodent lairs: 7 8 9 10 11 12 13 14 15 16 17

Leopards Leopards Spotted hyaenas Spotted hyaenas Spotted hyaenas Spotted hyaenas Striped hyaenas Striped hyaenas Indet. hyaena den Upper Pleistocene hyaena den Porcupines

Archaeological and modern human sites: 18 19 20 21 22 23 24

San hunter-gatherer camp Hottentot camp (men and dogs) Middle Pleistocene level of Nazaret Cave Upper Pleistocene cave at Pomongwe Upper Pleistocene cave at Booplaas Upper Pleistocene cave at E1 Juyo Neolithic cave at Fontbr@goua

Pliocene and Lower Pleistocene sites: 25 26 27 28 29 30

Lower Pleistocene Site at Venta Micena FxJj20 level (main+east) at Koobi Fora FxJj50 level at Koobi Fora FLK Zinjanthropus Site level 22 at Oldu FLK North level 6 at Olduvai FKL North North level 2 at Olduvai

TABLE 2 - R e l a t i v e f r e q u e n c i e s of four t y p e s of p o s t c r a n i a l bones (I: v e r t e b r a e ; II: ribs; III: l i m b a n d gi rdl e bones; IV: p h a l a n g e s ) in s e v e r a l fossil a n d r e c e n t a s s e m b l a g e s . N: t o t a l n u m b e r of bones i n each a s s e m b l a g e . Fr@quences relatives de quatre types d'os postcr~t-

niens (I: vertdbr@s; II: cStes; III: os des membres et des ceintures; IV: phalanges) de plusieurs assemblages fossiles et actuels. N: nombre total d'os dans chaque assemblage.

29 FIGURE 24~ - Results obtained in a factor correspondence analysis on the raw i'requencies of different types of postcranial bones in several bony assemblages collected by carnivores, rodents and hominids (data from Table 2). The scatterplot of the factor loading values of the assemblages in the plane defined by the two first factor correspondence axes, which explain more t h a n 95% of the original variance, follows a horse-shoe distribution ( G u t t m a n n effect). The n u m b e r s of the assemblages refer to the n u m b e r s in Table 2. Rdsaltats obtenus par analyse faetorielle des correspondances sur trois frdquences brutes de types diffdrents d'os postcr£tniens dans plusieurs assemblages d'os collectgs par des carnivores, des rongeurs et des hominidds (donndes du Tableau 2). Le tracd dispers~ des valeurs du facteur de charge des assemblages dans le plan ddfini par les premiers axes du facteur de correspondance, Iesquelles expriment plus de 95% de [a variance originelle, suit une distribution en fer & cheval ( effet Guttmann). Les nombres d'assemblages se rapportent aux nombres du Tableau 2.

o0

Second axis of correspondences (32.2% of variance explained) ~i~'.'. /

o

striped hyaena dens

.... ..~.

, / , spotted hyaena dens

i'~ ....... ~

mo d

• 04 o

"..

j

.....

I11(limb and girdle bones)

~) .... ~

....../

porcupine lair

:.• .:

o o

@ I (vertebrae)

•4 / .

archaeological caves

~ o,

\

o

I.,.., -'0""..

""... "'°" 0": \

• • "~"

d ~. o[

11(ribs)

\

carnivore feeding-site s leopard dens

¢ IV (phalanges)

First axis of correspondences (63.1% of variance explained)

0 0

"7 -1.50

hottentot camp

-1.25

-1.00

second factor shows higher projections for bone accumulations in which long bones are well represented, essentially those from spotted and striped hyaena dens, as well as those from the Hottentot camp (in which there are joint effects of Homo and Canis on bones), whereas negative values correspond to bone accumulations showing higher raw frequencies of phalanges (some assemblages accum u l a t e d by the man in caves from upper Pleistocene times, as well as from leopard dens). The distribution of bone assemblages collected by the two main collecting agents (hominids and hyaenids) show quite different behavior: hyaenid accumulations are more homogeneous in composition (Table 2), since they consist mainly of group III skeletal elements (limb and girdle bones, excluding phalanges), whereas other types of bones are scarce in them. The greatest difference among the assemblages which were collected by hyaenas is found between dens of striped hyaenas (which are mainly carrion eaters due to their smaller size and solitary habits) and those of spotted hyaenas (whose larger size and strong social organization make thera efficient hunters). In the former bone assemblages, ribs and vertebrae are almost absent, whereas in the latter ones these elements are somewhat more abundant, thus denoting less full utilization of the carrion. Contrariwise, the composition of bone assemblages due to anthropic action is more heterogeneous, and varies considerably depending on whether they derive from settlements established[ in an open camp or in a cave.

-0.75

-0.50

-0.25

0.00

0.25

0.50

0.75

1.00

Venta Micena lies in the region of the factorial diagram occupied by the bone assemblages originating from spotted h y a e n a dens. With respect to the African Plio-Pleistocene assemblages analysed here, all with purported patterns of anthropic activity on bones (Bunn 1982), the level FxJj20 of Koobi Fora (main+east) is also included within the group of hyaenid-collected assemblages, close in position to Venta Micena, whereas level FxJj50 of Koobi Fora and Olduvai levels F L K 6 and 22 ("Zinjanthropus" site) are placed near the lessbiased assemblages from open habitats. Olduvai level FLK North 2 takes an intermediate position between the two sets. The results of the multivariate comparison suggest that the Venta Micena assemblage originated mainly through accumulation by hyaenas (Pachycrocuta brevirostris) at feeding places nearby their dens, of skeletal remains of the prey hunted by different carnivores. This interpretation is supported by analysis of the distribution of teeth and bones in the Venta Micena excavation quarry (Fig. 8), which shows two areas with greater abundance of skeletal remains, and which probably correspond to the bones dispersed at the entries of shallow dens, whose original structure was not preserved due to the sedimentary characteristics of the stratum and the diagenetic compaction. The recovery of relatively high numbers of deciduous teeth o f P brevirostris reinforces this hypothesis (as other carnivores are represented in the assemblage only by adult individuals) and helps to reject the possibility that bones

30 The ratio of juvenile/adult individuals in a population depends on two factors: the annual birth rate or reproduction rate, and the duration of infancy. The reproduction rate scales approximately to the -0.3 power of adult body w e i g h t of a given species (the slope for births/year ranges from -0.26 [Eisenberg 1981] to -0.33 [Western 1979, 1980]). The duration of infancy is interspecifically related to body mass by a power close to 0.3 (0.29 for the time needed to reach reproductive m a t u r i t y [Calder 1982, 1984] and 0.28 for the rate of incremental growth [Case 1978]). The proportion of juvenile individuals for a given species is thus the product of annual birthrate (Br) and duration of infancy (Di):

were accumulated in open feeding-places located at hunting sites distant from dens, since it can be presumed that infant individuals would not accompany adults on their hunts, but would stay near the dens, as occurs in modern spotted hyaenas (Kruuk 1972). A M O D E L OF P R E Y S E L E C T I O N

Four types of evidence strongly support a selection of prey by carnivores at Venta Micena during lower Pleistocene times (Palmqvist et al. 1996a): the interspecific analysis of the abundance of remains of juvenile individuals among ungulates, in relation to the average weight estimated for adults in each species; the U-shaped attritional mortality profiles deduced from crown heigth measurements in several ungulate species; the presence of many metapodials showing osteopathologies such as arthrosis, and; the sex ratio deduced from the metapodials of buffalos and equids, which is in both cases biased in favour of females. JUVENILE/ADULT

%juvenile individuals = BrDi = K4W-°~KsW°3= K6Wo = constant. This percentage will be t h u s a p p r o x i m a t e l y constant and independent of species body size. However, in the ungulate species from the Venta Micena assemblage we noted a positive relationship between the two variates, which is highly significant statistically (Fig. 25):

RATIOS

MNI counts for each species in the Venta Micena assemblage are shown in Table 1, in which the numbers of young individuals with deciduous teeth (from calfs to subadults or yearlings) and of adults with fully erupted p e r m a n e n t dentition are given. The percentage of juvenile individuals for each ungulate species, in relation to the m e a n weight estimated for adults, is shown in Figure 25. Larger species are represented in the fossil assemblage by a comparatively greater MNI of juveniles t h a n those of smaller species. FIGURE 25 - Least squares regression analysis of the proportion of juvenile individuals on estimated adult body weight (in kg) for ungulate species (N = 9) of the Venta Micena a s s e m b l a g e (data from Table 1): Log(% juveniles) = 1.401 (_+0.540) + 0.346(_+0.084) Log (weight); r = 0.841, F = 16.95 (p < 0.004). a, Hemitragus alba, b, Cervidae gen. et sp. indet., c,

SoergeIia minor, d, Equus altidens, e, Megaloceros (Megaceroides) solilhacus, f, Bovini cf. Dmanisibos, g, Stephanorhinus etruscus, h, Hippopotamus amphibius antiquus, i, M a m m u t h u s meridionalis. Analyse de rdgression des moindres carrds de la proportion des individus juveniles sur le poids corporel estimd (en hg) pour des esp~ces d'ongulds (N = 9) de l'assemblage de Venta Micena (donndes d'apr~s le Tableau 1): Log (% de juveniles) = 1,401 (+0,540) + 0,346 (+ 0;084); Log (poids); r = 0,841, F = 16,95 (p
Log(% juveniles) = 1.401(+0.540) + 0.346(_+0.084) Log(W); r = 0.841, F = 16.95 (p < 0.004). This result suggests strong selection by carnivores, according to the age and size of their ungulate prey, since predation of comparatively larger species such as elephant and hippo is focussed on juvenile and more vulnerable individuals, whereas in smaller species young and adult individuals are captured at similar frequencies.

5 r~

~9 • ~-..I

,f d%e

g,

°c

°i

24

5

6 7 Log(weight)

8

9

31

MORTALITY P R O F I L E S Interspeciiic analysis of juvenile/adult ratios for ungulates in Venta Micena indicates a different age of death depending on the size of the prey, as a consequence of selection by predators, which would increase the proportion of young and more vulnerable individuals hunted of those ungulate prey species of larger body size. However, the juvenile/adult ra~;ios have poor resolution with respect to habitual prey age selection by carnivore species. Given this limitation, mortality profile patterns were deduced for those ungulate species which show greater relative abundance in the assemblage, the horse Equus altidens (MNI = 70, 37.4% of the total MNI of ungulates in the assemblage) and the large deer Megaloceros (Megaceroides) solilhacus (36, 19.3% of the ungulates). Age at death was calculated from dP4 and P~ crown height measurements (after Klein & Cruz-Uribe 1983, 1984). The mortality curve deduced for E. altidens (Fig. 26A) indicates a clear U-shaped age profile, which suggests t h a t death occurred mainly as a result of predation, and that predation was focussed both on very young individuals (most of them showing unworn dP.,) and past prime adults (with medium to heavily worned P4). This pattern of mortality is similar to those observed for modern ungulate prey species h u n t e d in a selective w a y by different carnivores ranging in size from the wild dog (Lycaon pictus) to the lion (Panthera leo) (Fig. 27). The attritional profile of M. solilhacus (Fig. 26B) shows a very pronounced peak corresponding to death in the first 10% of potential lifespan. This difference with the mortality curve deduced for the horse m a y be due to the somewhat greater size estima~;ed for the former species (269-568 kg, in contrast 'with a range of weights for E. altidens of 244-487 kg). The shape of the curve for M. solEhacus is similar to that found for Cape buffalo (Syncerus caffer) from the Middle Stone age layers of I~asies River Mouth Cave 1, South Africa (Klein & Cruz-Uribe 1983), which approximates an idealized attritional profile. OSTEOPATHOLOGIES A s t u d y of the long bones from Venta Micena have revealed m a n y osteopathologies (Fig. 28A,B), the most frequent of which is arthrosis; it was found in the distal epiphysis of a metacarpal of a large deer (M. solilhacus) and in two third metatarsals of equid (E. altidens); both cases were manifested as considerable osseous overgrowths. Another type of abnormality was present in the diaphyses of a goat metacarpal (H. alba) and of an equid metatarsal, both of which present a wrinkled surface with osseous accretions. A third and more

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Percentage of lifespan

FIGURE 26 - M o r t a l i t y profiles of E q u u s altidens a n d Megaloceros (Megaceroides) solilhacus in Venta Mieena, calculated from dP4 and P4 crown height m e a s u r e m e n t s (values for E. altidens kindly provided by V. E i s e n m a n n , unpublished data, N = 81; values for M. solilhacus obtained from Men~ndez 1987, N = 48). The proportion of d e a t h individuals in each 10% of potencial lifespan interval was calculated using the software developed by Klein & Cruz-Uribe (1983). Profils de la mortalitd d~Equus altidens et Megaloceros (Megaceroides) solilhacus & Venta Micena, calculge d'apr~s les mesures de la hauteur de la couronne de dP~ et P~ (valeurs pour E. altidens aimablement fournies p a r V. Eisenmann, donndes ingdites, N = 81; valeurs pour M. solilhacus d'apr~s Men~ndez 1987, N = 48). La proportion d'individus morts pour chaque intervalle de 10% df~ la durge de vie a ~t~ calculge ~t l'aide du logiciel de Klein & Cruz-Uribe (1983).

subtle type of osteopathology was found in the proximal epiphysis of several horse and buffalo (Bovini cf. Dmanisibos) metacarpals and metatarsals, affecting the articular facets of the metapodials, which were absent or showed abnormal growth. Results of a study of part of the Venta Micena collection indicate that these bone abnormalities are rather common: 13 out of 184 metapodials analysed of M. solilhacus, Bovini cf. Dmanisibos and E. altidens show signs of disease. Given that each ungulate has four metapodials, the pooled percentage of crippled animals of all three species would be approximately 28% (31% of equid, 19% of deer and 26% of buffalo). This proportion is higher than w h a t would be expected in wild populations, and thus suggestive that such individuals in less than optimum condition were

32 40%

30%

20%

carpals (Fig. 28C), thus allowing estimation of the ratio of males and females in the fossil assemblage: 26% (5/19) and 74% (14/19), respectively (i.e., approximately 1: 3). These percentages are significantly different, according with a t-test (t = 9.36; p < 0.001), what suggests that predators selected female individuals, given their smaller body size. The metapodials of the horse (E. altidens) do not show a clear bimodal p a t t e r n of size dimorphism as that exhibited by the buffalo, b u t linear discriminant functions developed using osteometric measurements allow to discriminate between the metapodials from males and females of Equus grevyi, the closest living analogue (both in size and shape) to the Venta Micena equid (GuerreroAlba & Palmqvist 1998). When these functions are applied to the complete metapodials of the lossil species, 26% (11/42) of t h e m are assigned to males (9 out of 24 metacarpals and 2 out of 18 metatarsals) and 74% (31/42) to females (15 out of 24 metacarpals and 16 out of 18 metatarsals), which are proportions r e m a r k a b l y similar to those found in the buffalo (i.e., 1 male: 3 females).

10%

0% 40%

30%

20%

10%

DISCUSSION

O%

CalfsYeaflingsTwo Young Primes Past CaBsYearlLngsTwo Young Primes Past years primes primes years primes primes

FIGURE 27 - Percentages of ungulate prey of different age groups (I: calfs, lambs, foals or piglets, less t h a n 12 months old; II: yearlings; III: two years old animals; IV: young prime adults, with full p e r m a n e n t dentition but teeth having only light wear; V: prime adults, showing dentition with medium wear; VI: past prime adults with teeth showing heavy wear) hunted by large carnivores (Panthera leo, N -- 410 kills; Panthera pardus, N = 96; Acinonyxjubatus, N = 33; Lycaon pietus, N = 96) in the Kafue National Park, Zambia (data for 20 ungulate prey species, Mitchell et al. 1965). Pourcentages des ongulds-proies dans des groupes de diffdrents @es (I: veaux, agneaux, poulains et porcelets, dgds de moins de 12 mois; II: animaux d'un an; III: animaux (tg6s de deux ans; I~: jeunes adultes avec dentition permanente mais avec dents ayant seulement une ldg~re usure; V: jeunes adultes avec une denture & usure moyenne; VI: adultes avec denture & usure importante) chassds par des grands carnivores (Panthera leo, N = 410 mises ~t mort; P a n t h e r a pardus, N = 96; Acinonyx jubatus, N = 33; Lycaon pictus, N = 96) dans le parc national de Kafue, Zambie (donndes pour 20 esp~ces d'onguldsproies, Mitchell et al. 1965).

actively selected by the predators, since these pathologies m u s t have limited the ability of the animals to run, thus reducing their chances of escaping from predators. SEX RATIO IN UNGULATES A third line of evidence of prey selection is provided by the analysis of sex ratio for the buffalo (Bovini cf. Dmanisibos), the largest ruminant of the Venta Micena assemblage. This bovid shows m a r k e d sexual dimorphism in the size of meta-

The stratigraphic and mineralogic data presented before allow us to propound the following sedim e n t a r y interpretation of the Venta Micena straturn: generation of a mudstone under a water sheet in a lacustrine environment (marginal facies); partial dessication of the palaeolake and subaerial exposure of the bottom, which preserves the previous microtopography; intense evapotranspiration and formation of an incipient calcrete paleosol (palaeoreiief); development in the lowest topographic heights of ponds formed by micrite mud, which holded a small vegetational cover and experienced succesive stages of dessication (rootmarks and mudcracks restricted to this zone of the stratum); and, finally, a new phase of massive generation of mudstone precipited under water sheet conditions, what sealed the preexisting palaeorelief. The Venta Micena assemblage, in which all size intervals for bones and species of large mammals typical in a Pleistocene palaeocommunity are represented, relies on this calcimorphic paleosol, whose irregular surface defined small trench like ponds and a surrounding area slightly more elevated, composed by a calcrete duricrust which formed the surface of a lacustrine margin. The substrate of these ponds was constituted by calcareous mud, which was dessicated at least three times, and includes the majority of the fossils which were preserved in the assemblage. There is no evidence of transport by w a t e r currents, since the bones show a random p a t t e r n in their orientation, there are no rounded or polished

33 FIGURE 28 -- Osteopathologies in the epiphyses of metapodials (from Palmqvist et al. 1996a). A, arthrosis in a large deer (M. solilhacus) metacarpal (left: pathologic bone, right: healthy bone). B, arthrosis in an equid (E. altidens) metatarsal (left: pathologic bone, right: healthy bone). C, size dimorphism in buffalo (Bovini cf. Dmanisibos) metacarpals (above: male, below: female). OsMopathologies sur des dpiphyses de rndtapodiaux (d'apr~s Palmqvist et al. 1996a). A, arthrose d'un mdtacarpien de grand cerf (M. solilhacus) (~ gauche: os pathologique, & droit!e: os sain). B, arthrose du mdtacarpien d'un gquidd (E. altidens) (& gauche, os pathologique, it droi~e: os sain). C, dimorphisme de taille des mdtacarpiens chez un buffle (Bovini cf. Dmanisibos) (en haul: m~le; en bas: femelle).

edges, a n d d e t r i t i c s a r e n e a r l y a b s e n t f r o m t h e sediment. T h e s u r f a c e of t h e b o n e s s e e m s to h a v e b e e n exposed to t h e effects of i n s o l a t i o n for a v e r y s h o r t t i m e i n t e r v a l , since only 10.7% of t h e p a l a e o n t o l o -

gical s a m p l e s h o w s i n c i p i e n t m a r k s of s u b a e r i a i w e a t h e r i n g , w i t h few s h a l l o w a n d s m a l l split line cracks d u e to insolation, a n d w i t h o u t m o s a i c crack i n g or exfoliation ( w e a t h e r i n g s t a g e 1, B e h r e n s m e y e r 1978), w h a t i m p l i e s t h a t t h e p e r i o d of t i m e

34 elapsed between the death of the individuals and the definitive burial of their skeletal remains was shorter than three years. The remaining 89.3% of the bones have no evidence of subaerial exposure (weathering stage 0, 0-1 year exposed before burial). These data, as well as the absence of sedimentary infilling in the bones which were preserved complete, indicate that the bones were buried with the periosteum intact and with fat content in their microestructure (i.e., they were buried when they were still in fresh condition). THE COLLECTING AGENCY OF BONES The elements from the estilopod, the zeugopod and the metapod predominate among the skeletal regions preserved, in many cases showing the original anatomical connections. The most abundant long bones are metapodials, tibiae, humeri and femori, of which only the metapodials are represented by a high percentage of complete elements. There are strong evidence which indicate that hyaenids were the biological agents responsible of the bone collecting and modifying processes, including the low number of isolated teeth and ribs, the prevalence of limb bones (of which the most abundant is the humerus in the forelimb and the tibia in the hindlimb), the high level of fragmentation of the bones (particularly proximal humeri with spiral fractures and distal tibiae with both spiral and longitudinal fractures), and the high number of gnawing marks made by premolars of large carnivores. All these features are typical of those bone assemblages collected by hyaenids (Skinner et al. 1980; Brain 1981; Horwith & Smith 1988). Additional evidence comes from the presence of many coprolites of large carnivores associated to the bones, and from the fact that Pachycrocuta brevirostris is the only carnivore species which is represented in the Venta Micena assemblage by relatively high numbers of both juvenile and adult individuals. The bony assemblages accumulated by hyaenids are relatively homogeneous in their composition, since they are constituted basically by limb and girdle bones (excluding phalanges), while the remaing types of bones are poorly represented, because they were not collected or due to the selective destruction in the place where the skeletal elements were accumulated. Hyaenids were responsible at Venta Micena of at least four processes: selection and transport/collecting of bones; concentration; selection and alteration by the activity of adult individuals (splintering), and; selection and alteration by juvenile individuals (gnawing). The main difference among bone assemblages collected by modern hyaenids is seen when we compare the dens of striped hyaenas (Hyaena hyaena)

and those of spotted hyaenas (Crocuta crocuta). Striped hyaenas are basically scavenger carnivores due to their small size and solitary habits, which use to transport to their dens parts of the carcasses of ungulates to feed the cubs (Mills 1989); ribs and vertebrae are nearly absent from these assemblages. Spotted hyaenas are efficient hunters, due to their greater size and strong social behaviour, and do not carry regularly carrion to their dens (Ewer 1973; Mills 1989); ribs and vertebrae are slightly more abundant in their assemblages, what may owe to a lower consumption of the bones, or to denote more interest for collecting these anatomical parts. Therefore, the type of preserved assemblage depends on the spatial context where the concentration and modification of the bones took place, and also on the behavior and age class of the bone collecting agency. The former considerations are supported on the differences between the dental pattern shown by both genera of modern hyaenids, because according to Bonifay (1971) the genus Crocuta shows less differenced cheek teeth, while the genus Hyaena presents more evolved dental features, which more closely resemble those of Pachycrocuta. The bone assemblage preserved in Venta Micena would be generated by the selective accumulation of the skeletal remains from carcasses of ungulate prey hunted by different predators in the plains that surrounded the Pleistocene lake of Orce (predators: carnivores sensu lato; bone collecting agency: hyaenids), and by the alteration of the bones, also selective, made by both adult and juvenile hyaenids (bone modificating agency: P. brevirostris, the only eudemic carnivore in the assemblage), as it is usual in modern habitats (Sutcliffe 1970; Kruuk 1972; Brain 1981). The most abundant anatomical parts of the postcranial skeleton of ungulates which were preserved in the bone assemblage are the less nutritious, given their high mineral density and low fat content (for example, metapodials and distal fragments of humerus and tibia), while those bones with low structural density and high fat weight are represented by lower raw frequencies in the Venta Micena assemblage (for example, the distal epiphysis of femur, and the proximal epiphyses of both femur and tibia), because they were preferentially detroyed by hyaenas in order to extract the within bone nutrients. All these data suggest that the activity ofhyaenids took place in resting sites close to their dens. The finding of high numbers of deciduous teeth of P. brevirostris (unworn DP 4 and germs of DP4) supports this hypothesis, and helps to reject the possibility that the skeletal remains were concentrated at hunting sites distant from the dens, since the youngs would not accompany adults on their

35 hunts, but would stay near the dens, as occurs in both modern spotted and striped hyaenas (Kruuk 1972; Mills 1989), in which juvenile individuals live in the dens until the age of 9-12 months (permanent dentition) or 12 weeks (deciduous dentition), respectively. The finding of only germs and unworn deciduous dentition implies that these teeth were not produced by tooth replacement, and it is thus a direct evidence of the death in situ of juveniles, what indicates that P. brevirostris had a similar behaviour to modern Hyaena and not to Crocuta. Such conclusion is also supported by the anatomy of the P4 teeth in Pachycrocuta, which shows a Hyaena like estructural pattern, opposed to that of Crocuta. This anatomical difference contrasts with the results of the bone modifying activity of P. brevirostris, which are analogue to those of fossil Crocuta in Europe. The distribution of teeth and bones in the Venta Micena excavation quarry shows two areas with higher density of skeletal remains, which could correspond to the bones that were dispersed at the entry of the dens; the original estructure of the dens, which were excavated by adult hyaenids in dry micrite mud, was not preserved due to the effects of diagenetic compaction. The model for Venta Micena m a y be analogous to one of the modern dens found in Amboseli National Park (Hill 1981). The results obtained through applying Damuth's model to the interspecific analysis of the size/abundance patterns in ungulate species from the Venta Micena assemblage indicate that the main taphonomic bias was produced by differential collecting and fragmentation by hyaenids of skeletal remains during an interval of time prior to their definitive burial. These biostratinomic agents biased the composition of' the palaeocommunity according to the body sizes of the species it comprised, affecting bones of smaller species more markedly. However, the structure of the original palaeocommunity was partially preserved during fossilization, and it is possible to quantitatively estimate the original representation among the unpreserved fraction. This result is congruent with what is known concerning the origin of the Venta Micena assemblage (Gibert et al. 1992; Palmqvist et al. 1992; Mendoza et al. 1993), since all sedimentologic and taphonomic analyses indicate that bones were accumulated on a paleosol characterized by extensive emerged zones, among which were numerous shallow ponds, places where ungulates came to drink. In this muddy environment skeletal remains of the prey of carnivores were collected by the hyaenas around the entrance of dens dug near the ponds, where they were buried rapidly as can be deduced from the low rate of subaerial weathering.

The most likely hypothesis to explain the accumulation of bones in the assemblage is that in a period of desiccation of the ponds and fall of water level, hyaenas dug small, shallow dens which would then serve almost exclusively to protect their offspring from other carnivores and from the effects of insolation; adult individuals were probably safe from other predators given their large size and social behavior. The dismembered carcasses of animals that were eaten accumulated around the entries of dens. At some time the hyaenas abandoned their dens, probably as a consequence of a slight rise of the water level of the ponds, which caused the dens, dug in dry micritic mud, to be flooded and to collapse. Immediately thereafter the water level rose considerably, precipitating the micritic limestone and sealing the entire bone assemblage. THE EVIDENCES FOR PREY SELECTION The analysis of the juvenile/adult ratios of the ungulate species identified in the assemblage in relationship to the weights estimated for adult individuals suggests different mortality age profiles depending on the size of the prey, as a consequence of selection by predators, which would increase the proportion of young and more vulnerable individuals of the species of larger body size. This interpretation agrees well with available data on prey selection by recent carnivores as a function of size and age of the ungulate prey. The mean percentages of juvenile individuals of each African ungulate species hunted, obtained by averaging the proportions of young killed by the five main predators (wild dog, cheetah, leopard, spotted hyaena and lion), are shown in Figure 29 (Table 3, data compiled from Palmqvist et al. 1996a). Percentages were calculated only for prey species with data available on the predation of juveniles by at least three species of carnivores, except for ungulate species of larger size (>1000 kg) which are only occasionally hunted by lions and spotted hyaenas. These pooled percentages would give the figures expected in an attritional assemblage originating exclusively from the accumulation of bones of hunted animals. A highly significant direct relationship exists between the percentages of young individuals and the body weights of the ungulate species: Log(% juveniles) = 2.966(_+0.201) + 0.203(_+0.036) Log(W); r = 0.805, F = 31.38 (p < 0.0001). The value for the slope in this fit (0.203) is within the 95% confidence interval of the slope that relates the percentages of juvenile individuals and the estimated weights for the ung'alate species ofVenta Micena (Fig. 25; slope = 0.346; standard error of slope = 0.084; confidence limits, with p < 0.05 = 0.156 - 0.536), and reveals that the two

36

parameters do not differ in a statistically significant way. Therefore, the proportions of juvenile individuals representing the ungulate species in the Venta Micena assemblage truly reflect prey selection by the carnivores as a function of body size of the prey hunted. The increase in the value of the juvenile/adult ratios in relation with the weight estimated for the ungulate species allows us to rule out the hypothesis that the fossil assemblage was formed through catastrophic mortality events that would accompany recurrent droughts; in this case the percentage of juvenile individuals of the different species would be approximately constant and size-independent. We can consequently conclude that the immense majority of the skeletal remains accumulated by hyaenas in Venta Micena came

Predators Prey ungulates Thomson's gacelle, 23 kg

from attritional mortality in ungulate populations, caused by selective choice of carnivores. The hypothesis of prey selection in the Venta Micena assemblage is also supported by the finding of various metapodials which show clear indications of osteopathology in the ephiphyseal and diaphyseal regions, which would have affected the locomotive capabilities of the ungulates (Fig. 28A,B). In a study of the condition of prey, assessed from the state of the femoral marrow, Mitchell et al. (1965) found that 27% of ungulates killed by the lions, leopards, cheetahs and wild dogs were in fair or poor condition, as it may be deduced from the low quantities of fat content, while the remaining individuals preyed showed fat contents indicative of good condition. However, given the fact that the marrow test operates only at the lower end of the

Wild dog, 25 kg

Cheetah, 50 kg

(Lycaon pictus)

(Acinonyxjubatus)

Leopard, 50 kg

Spotted hyaena, 55 kg

Lion, 160 kg

(Pantherapardus)

(Crocutacrocuta)

(Pantheraleo )

30%

38%

20%

43%

39%

22%

27%

40%

27%

65%

--

60%

21%

61%

34%

28%

41%

50%

38%

38%

33%

--

33%

33%

(Gazella thomsoni) Springbok, 31 kg

(Antidorcas marsupialis) Reedbuck, 45 kg

(Redunca redunca) Impala, 51 kg

(Aepyceros melampus) Grant's gacelle, 55 kg

(Gazella granti) Puku, 72 kg

(Kobus vardoni) 46%

87%

Warthog, 75 kg

(Phacocherus aethiopicus) Hartebeest, 150 kg

80%

86%

33%

32%

100%

100%

65%

35%

83%

92%

93%

38%

27%

70%

100%

9O%

28%

100%

100%

64%

(Acelaphus buselaphus) Gemsbok, 150 kg

-

-

(Oryx gazella) Wildebeest, 170 kg

(Connochaetes taurinus ) Waterbuck, 210 kg

28%

(Kobus eUipsiprymnus) BurchelI's zebra, 230 kg

42%

37%

6O%

16%

(Equus burchelli) Eland, 560 kg

(Taurotragus ol:yx) Buffalo, 600 kg

23%

100%

(Syncerus caffer) Giraffe, 1,100 kg

87%

71%

9O%

80%

100%

89%

(Giraffa camelopardalis) Black rhino, 1,200 kg

(Diceros bicornis) Hippo, 2,000 kg

(Hippopotamus amphibius) 100%

White rhino, 2,500 kg

(Ceratotherium simum) Elephant, 3,800 kg

100%

100%

(Loxodonta africana) TABLE 3 - M e a n p e r c e n t a g e s of j u v e n i l e individuals of African u n g u l a t e species h u n t e d by the five largest African ~redators (data compiled from P a l m q v i s t et al. 1996a). Moyennes des pourcentages des individus juvdniles d'esp~ces d'ongulds africains chassis par

cinq grands prddateurs africains (donndes compildes d'apr~s Palmqvist et al. 1996a).

37 Log ............................................... (%juveniles)

,

q

.

.

r

.

.

.

.

.

.

.

!

i jk

_ 4

eg ~ CO

a • b 3

]

. . . . . . . . .

i

4

. . . . . . . . .

i

5

. . . . . . . . .

i . . . . . . . . .

6 Log (weight)

i

7

. . . . . . . . . . . . . . . . . .

8

therium iatidens, the medium-sized saber-tooth Megantereon whitei, the large hyaena Pachycrocuta brevirostris and the wild dog Canis (Xenocyon) falconeri. Saber-toothed cats, which belong to the subfamily Machairodontinae, share among others the following craneodental derived characters (see for review and references Biknevicius et al. 1996; Emerson & Radinsky 1980; Marean 1989; Van Valkenburgh & Ruff 1987):

FIGURE 29 - Regression analysis by least squares method of the m e a n percentage of juvenile individuals (% juveniles) of several African u n g u l a t e species (N = 19) h u n t e d by the five large African predators (Lycaon pictus, Acinonyx jubatus, Panthera pardus, Crocuta crocuta and Panthera leo) on m e a n adult body weight (in kg): Log(% juveniles) = 2.966(_+0.201) + 0.203(_+0.036) Log(weight); r = 0.805, F = 31.38 (p < 0.0001). a, Gazella thomsoni, b, Antidorcas marsupialis, c, Redunca redunca, d, Aepyceros melampus, e, Gazella granti, f, Kobus

- Elongate and flattened upper canines of two basic types: Homotherium (tribe Homotherine) had scimitar-shaped canines relatively short and broad, which were serrated bearing coarse crenulations, while Megantereon (tribe Smilodontini) showed dirk-shaped canines, extremely long, narrow, and without serrations.

vardoni, g, Phacochoerus aethiopicus, h, Alcelaphus buselaphus, i, Oryx gazella, j, Connochaetes taurinus, k, Kobus ellipsiprymnus, l, Equus burchelli, m, Taarotragus oryx, n, Syncerus caller, o, Giraffa camelopardalis, p, Diceros bicornis, q, Hippopotamus amphibius, r, Ceratotherium simum, s, Loxodonta africana. D a t a calculated from Table 3. Analyse de rdgression par la mgthode des moindres carr~s du pourcentage moyen des individus juvdniles (% juveniles) de plusieurs esp#ces d'ongulds africains (N = 19 ) chassds par les cinq grands pr~dateurs africains (Lycaon pictus, Acinonyx jubatus, P a n t h e r a pardus, Crocuta crocuta et P a n t h e r a leo) sur la moyenne du poids corporel adulte (en kg): Log (% juveniles) = 2,966 (+_ 0,201) + 0,203 (+_ 0,036) Log (poids); r = 0,805, F = 31,38 (p <0,0001.

Enlarged upper incisors, which are relatively longer, thicker, more pointed and procumbent than in modern felids, and reduced, incisor-shaped lower canines. The incisor row is long and strongly curved, what suggests for saber-tooths a functional emphasis on these teeth for tearing and stripping flesh from carcasses, a task that modern felids perform with the assistance of their stout and conically shaped canines; the extremely large upper canines of the saber-tooths would probably be ineffective at the manipulation of chunks of flesh.

condition scale, the proportion of unhelthy individuals preyed upon by these predators is probably underestimated. The selection of prey in poor condition is higher for those ungulate species of larger body size: 68.4% of buffalo killed by lions, in contrast with only 9.1% of hartebeest, 8.5% of wildebeest, and 23.5% of zebra (Mitchell et al. 1965; Schaller 1972). Similarly, Crisler (1956) found that the caribou the wolf catches is usually the one that slows down: at least 50% of the kills involve crippled or sick individuals, whose incidence is 1.8% or even lower in the caribou herds. THE H U N T E R S AND T H E I R PREY AT VENTA MICENA It is always difficult to determine the role played by difi~rent carnivores in a palaeocommunity. Evidence that supports the model of prey selection proposed here (the results of the interspecific analysis of juvenile/adult ratios in ungulate species, the U-shaped mortality profiles, the presence of m a n y bone diseases, and a biased intersexual ratio of lar~;e bovids and equids) clearly indicates that most of the skeletal remains preserved in the Venta Micena assemblage come from hunted individuals. But, who were the hunters? Four large carnivore species have been preserved in the assemblage: the great saber-tooth Homo-

-

- Upper carnassials (p4) with a reduced or absent protocone (lingual lobe), which is lowered away from the occlusal surface in Megantereon, thus removing it from its role as a h a m m e r for bone crushing (a condition that is only present among extant felids in the hypercarnivorous cheetah), and is lost in Homotherium, in which there is also an anteriorly added accesory cusp. This teeth forms in saber-tooths a long thin blade, which is extremely specialized for slicing flesh, and allowed them to deflesh their prey rapidly. - A lowered glenoid fossa, a reduced height of the coronid process, a laterally shifted angular process, and a shortened zygomatic arch. All these features allow a wider gape than that of modern felids, but suggest that the temporalis muscle was weaker. However, the temporal fossae was shorter and narrower, which indicates that the temporalis was oriented in saber-tooths more vertically and perpendicular to the tooth row than in modern felids. This increased the bite force at the carnassial (M1), which was closer to the mandibular condyle, although it remained significantly lower than in felids. - An enlarged, lowered and ventrally extended mastoid process, which is enormous relative to modern felids, w h a t indicates that the cleido- and sterno-mastoid muscles m u s t have been corres-

38 pondingly large. The occiput is in most sabertooths relatively higher and narrower than in felids, and the temporomandibular joint is located more ventrally. The mastoid process is rotated further below the skull joint so that the leverage of the neck muscle is increased, thus suggesting that a head-depressing motion was involved in the penetration of the canines. The postcranial skeletons of scimitar-toothed and dirk-toothed machairodonts are quite different. Homotherium was a relatively long legged pursuit predator with the size of a modern lion, which had a comparatively large brain with an enlargedment of the optic centre, a condition similar to that of the cheetah (Rawn-Schatzinger 1992). The morphology of Homotherium is unique among extant and past felids, showing relative limb proportions which indicate increased cursoriality and less prey grappling capabilities than other sabertooths. The brachial index (i.e., radius length/ h u m e r u s length) takes values close to or above 100%, w h a t implies that most species of this genus preferred open h a b i t a t s (Lewis 1997). Megantereon was a relatively short limbed ambush hunter, with a comparatively smaller brain, showing olfactory lobes well developed. It had powerfully developed forelimbs, w h a t suggests t h a t a killing bite in the throat m a y have been coupled with the immobilization of the prey by the front limbs. Comparative multivariate analysis of postcranial m e a s u r e m e n t s (Lewis 1997) indicates for Megantereon an overall morphology similar to that of extant jaguars, with tree catching and long distance dragging capabilities; the low value for the brachial index suggests closed habitat preferences. As it will be discussed in detail below, morphofunctional studies currently in progress of African M. whitei from Venta Micena indicate t h a t this predator generated large amounts of carrion, since it would exploit the carcasses of its prey to a smalll degree. All these features indicate; that saber-toothed felids were able to hunt very large prey relative to their own size, and; that they left on the carcasses of the ungulates hunted large amounts of flesh and all within bone nutrients, which were available to be subsequently scavenged by hyaenids and hominids (Marean 1989). Sabertooths became extinct in East Africa by 1.5 Ma ago, what coincides w i t h the emergence of the Acheulean Industrial Complex, but inhabited Eurasia until 0.5 Ma (Turner 1990, 1992). Their persistence m a y then explain the success of both P. brevisrostris and hominids with Oldowan technology in Eurasia, where the Oldowan/Acheulean transition took place much later than in Africa, at approximately 0.5 Ma (i.e., when saber-tooths disappeared in this continent), since the Oldowan

sharp flakes were fully appropiate to scavenge on carcasses partially defleshed by saber-tooths and the cores were used by hominids in breaking bones for their marrow content.

C. falconeri was a hypercarnivorous canid widely distributed during the late Pliocene and early Pleistocene in the Old World (Rook 1994). This species had a large body size, comparable with that of the living northern races of Canis lupus, and was characterized by a relatively short neural cranium and a narrow muzzle. The sagittal crest is strong and the bullae are inflated. Features of the dentition include M2 with a relatively reduced metaconid, and M1 with a stoutly built talonid, consisting of a preeminent hypoconid relative to the entoconid, and a reduced metaconid. The lower premolars usually show accessory cusplets, and the mandibular r a m u s is high and heavy. The upper molars, and specially the upper carnassials, show a marked tendency to ward brachydonty; a wide occlusal basin is present at the base of the metacone and paracone in the M 1, and the M 2 is large. The metacarpal II has a very reduced articular facet for the metacarpal I, w h a t indicates that the latter bone was vestigial if not absent, a condition similar to t h a t ofLycaon pictus, the only extant canid with a tetradactyl forelimb (Rook 1994).

P. brevirostris was an African giant, short-faced hyaena relatively common in lower Pleistocene European assemblages of large m a m m a l s (Howell & Petter 1980). It had a body and skull 10-20% larger than the modern spotted hyaena, Crocuta crocuta, and thus probably had greater ability than the latter species for destroying carcasses and consuming bone (Turner & AntSn 1996). In Europe, P. brevirostris is first recorded in lower Pleistocene deposits at Olivola, and its last appearance is in the early middle Pleistocene site at Sfissenborn. This species differed from other hyaenids in having a relative shortening of the distal limb segments, as reflected in the ratio of radius length to h u m e r u s length (88%; modern hyaenids range between 99% for the brown hyaena, Parahyaena brunnea, and 106% for the striped hyaena, Hyaena hyaena) and in the ratio of femur length to tibia length (74%; values of modern species range from 80% in C. crocuta to 89% in H. hyaena). These differences suggest a less cursorial life style for P. brevirostris, although such shortening could provide greater power and more stability to dismember and carry large pieces of carcasses obtained from aggressive scavenging (Turner & AntSn 1996).

P. brevirostris was replaced in E u r o p e by Pliocrocuta perrieri according with Turner (1992) and Turner & AntSn (1996), but in Spain there is

39 no Pleistocene record of the latter species and P. brevirostris is substituted directly by C. crocuta. However, the finding of P. perrieri in French sites of middle Pleistocene age like Lunel Viel can be also explained by reworking of Pliocene fossils of this species, as occurs in the Spanish karstic site at Cueva Victoria. As discussed before, the extinction in Europe of P. brevirostris seems to have been linked to the decline and subsequent disappearance of machairodonts, particularly M. whitei, what implied the loss of an important source of partly-consumed carcasses, and thus a change in the interactions between flesh-eating and bonecracking species of the carnivore guild (MartinezNavarro & Palmqvist 1996; Turner & Ant6n 1996). A possible way to deduce the respective ecological roles of the fossil carnivores preserved in the Venta Micena palaeocommunity may be to compare their hunting behavior with those of their modern analogs. Figure 30 shows the mean percentages of ungulates of different size classes (<50 kg, 50-150 kg, 150-400 kg, 400-800 kg, and >800 kg) that are killed and scavenged by the five main African predators (lion, spotted hyaena, cheetah, leopard and wild dog) (data compiled in Palmqvist et al. 1996a). As can be appreciated, lions and spotted hyaenas show similar hunting preferences according to prey body size: most kills involve animals that weigh between 50 and 150 kg. However, lions capture more ungulates of 150400 kg, and[ especially of more than 400 kg, while hyaenas prey more often on smaller species (<50 kg). Cheetahs and leopards show similar hunting behaviors, killing mostly ungulates of less than 50 100%i

kg. Prey of wild dogs show two peaks: ungulates of less than 50 kg, mostly Thomson's gazelle, and animals weighing between 150 and 400 kg, like the wildebeest, which they are able to capture thanks to their cooperative hunting techniques. When these graphs are compared with that for the Venta Micena assemblage (Fig. 31), some interesting similarities emerge: the maximum for Venta Micena is in the intermediate size category (150-400 kg), similarly to the prey hunted by the lion and the wild dog, and there are two minima for the highest and lowest weight classes. The abundance in the fossil assemblage of large ungulates (>400 kg) is similar to the proportion of such prey species in lion kills. The relative scarcity of ungulates weighing less than 150 kg in comparison with the abundance of skeletal remains from species of 150-400 kg resembles the proportion of captures by wild dogs. These results suggest that two carnivores -the lion-sized saber-toothed felid H. Iatidens and the canid C. falconeri- played important ecological roles at Venta Micena. However, one question remains: what was the role of the large hyaenid P. brevirostris in this palaeocommunity? Modern spotted hyaenas both kill and scavenge ungulates, depending on factors such as their interaction with lions and other predators (see Kruuk 1972). The remarkably similarity between the mean percentages of ungulates of the five size classes that are scavenged by modern spotted hyaenas (Fig. 30, data from Henschel] & Skinner 1990; Kruuk 1972) and the abundance of these size categories in the fossil assemblage clearly indicates that P. brevirostris fed largely on ungulates preyed upon by other predators. The only differenI

Lion kills

[ Spo~edhyaena kills

75% I 50%L 25% FIGURE 30 - Mean percentages of several ungulate size classes (a: <50 kg, b: 50-150 kg, c: 150-400 kg,

d: 400-800 kg, e: >800 kg) killed

and scavenged by the five main African predators (lion, spotted hyaena, cheetah, leopard and wild dog; data from Palmqvist et al. 1996a). Pourcentage des moyennes des classes de faille de plusieurs ongulds (a: < 50 kg, b: 50-150 kg, c: 150-400 kg, d: 400-800 kg, e: >800 kg) tugs et d#vor#s par les cinq principaux pr,ddateurs africains (lion, hy~ne tachet#e, gugpard, l~opard et chien sauvage; donnges d'apr~s Palmqvist et al. 1996a).

0%

lO0%~Wild kills' dog 75°/°

50%

L

1

25% 0%

<50 kg 150-400 kg >800 kg <50 kg 150-400 kg >800kg <50 kg 150-400 kg >800 kg 50-150 kg 400-800 kg 50-150 kg 400-800kg 50-150 kg 400-800 kg

40 I

I

100%f VentaMicenaassemblage 75%f 50%

1

25% 0%

<50kg 150-400kg >800kg 50-150kg 400-800kg

FIGURE 31 - Mean percentages of several ungulate size classes (a: <50 kg, b: 50-150 kg, c: 150-400 kg, d: 400-800 kg, e: >800 kg) in Venta Micena. Pourcentage des moyennes des classes de taille de plusieurs ongulds (a: < 50 kg, b: 50-150 kg, c: 150-400 kg, v~ 400-800 kg, e > 800 kg) & Venta Micena.

ce between both graphs is in the abundance of small ungulates (<50 kg), which are practically absent from Venta Micena. This may be explained if we consider that these carcasses were almost entirely consumed on the spot where the large extinct hyaenas found them, rather than being carried back to their dens. A similar selective behavior has been reported among modern spotted hyaenas (Brain 1981; Kruuk 1972). If hyaenas failed to transport the carcasses of individuals of small species to their dens, this preservational bias could then explain the scarcity of ungulate prey of very small size, which are represented in the fossil assemblage by only one distal epiphysis of a metacarpal of Caprini gen. et sp. indet. (a species weighing between 8 and 10 kg). Such selective behavior in the transport to dens of carcasses by hyaenas could then lead to poorer preservation of the young of small species in the assemblage than the juveniles of large species, thus affecting the interpretation of differential size effects on hunting success proposed here. However, this bias is only important for species of less than 50 kg, since those ungulate species weighing 50-150 kg (Hemitragus alba and Cervidae gen. et sp. indet.) are well represented in the Venta Micena collection (MNI = 14 and 20, respectively), in a proportion similar to that found among ungulates scavenged by spotted hyaenas (Fig. 30). Consequently, the increasing value of the juvenile/adult ratios as a function of body size for ungulate species weighing more than 50 kg (Fig. 25) is not an artefact of size effects on differential bone transport by hyaenas.

The results obtained clearly indicate that P. brevirostris was a bone-cracking scavenger in the Venta Micena palaeocommunity. It fed mainly on carcasses of animals killed and partially consumed by other flesh-eating carnivores, in contrast with the behaviour of modern spotted hyaenas, which are both hunters and scavengers (Kruuk 1972). The strong representation of juveniles among ungulate species identified at Venta Micena, 40.4% (80/198) of all individuals, suggests that the bone assemblage was an accumulation made by primary predators (Brain 1981; Shipman 1981; Vrba 1980). However, in primary assemblages most individuals tend to fall into a relatively restricted body size, according with the predator's preferences, and the carnivore/ungulate ratio is usually high, while at Venta Micena the body weight distribution of ungulate species is very wide (approximately 8-6000 kg: i.e., nearly three orders of magnitude) and the value of the carnivore/ungulate ratio is rather low, only 14.1% (28/198), what indicates that the bone assemblage was predominantly a non-primary, scavenged one. These data support the previous conclusion that the Venta Micena hyaena was specialized in scavenging ungulate prey hunted selectively by other carnivores. With respect to very large ungulate prey species identified at Venta Micena (>1000 kg), H. latidens was probably the only such carnivore species capable of hunting them, since only modern lions and spotted hyaenas can hunt juveniles or critically ill individuals of such large species. So far as other ungulates are concerned, one of the most suitable candidates for predation seems to be the canid C. falconeri, as suggested by the high frequency of osteopathologies. These canids must have pursued their prey over long distances, like modern African wild dogs, thus leading to intense selection of individuals unable to withstand prolonged running. This hypothesis is supported by the finding of numerous juveniles of ungulate species of intermediate size (46% horse, 42% large deer, and 60% buffalo), since lions and spotted hyaenas do not select young individuals among modern ungulate species of comparable sizes (Table 3) as strongly as they are represented in the Venta Micena assemblage. The high proportion of juveniles of larger species (elephant and hippo) in the assemblage indicates that the hunting behaviour of the saber-toothed felid H. latidens was similar to that of the modern African lion. The proboscidean M. meridionalis is represented in the fossil assemblage by five individuals, four of which are juveniles. The worn molars of the remaining individual suggest that it was old and probably died of starvation, being subsequently scavenged by hyaenas. Similarly, a

41 study by Rwan-Schatzinger (1992) of the Friesenhahn Cave assemblage, which was accumulated by the great scimitar cat Homotherium serum, has revealed a very high selection (almost 100%) of juveniles among the two proboscidean species identified in this site (Mammuthus americanus and Mammuthus cf. columbi). In contrast, other ungulate species of smaller body size such as Mylohyus nasutus, Odocoileus virginianus and Bison sp. are represented in this assemblage by much lower percentages of young individuals (40%, 33% and 50%, respectively). As noted above, lions are only able to hunt very young elephants, because the large size of adults makes them virtually invulnerable to predation. The high selection of juveniles of large ungulate prey species in Venta Micena suggests then that the hunting behavior of large machairodonts would be similar to that of recent large felids, which first subdue prey with their claws while biting the neck, as opposed to specialization in the capture of large proboscideans by stabbing them with their elongated canines and waiting for the prey to bleed to death, as suggested in some classic studies. Otherwise, the proportion of adult individuals of large ungulate species in the assemblage would have been much higher. With regard to the possible ecological role of M. whitei in this paleocommunity, we can offer some morphofunctional considerations which have been published elsewhere (Martine~-Navarro & Palmqvist 1995, 1996; Palmqvist et al. 1996a). The results of a comparative study of the skeletal remains of M. whitei suggest that the dimensions of this machairodont differ markedly depending on whether they are estimated from the teeth or from the postcranial skeleton. When body size is calculated with minimum squares regression analysis (Van Valkenburg 1990) of lower carnassial tooth (M~) length on body weight in modern species of felids (i.e., the procedure followed in this work; see Table 1), the value obtained is of only 55 kg. This suggests that this African species was leopardsized (Mart:inez-Navarro & Palmqvist 1995). On the other hand, surface area of the diaphyseal cross section of the humerus in this species is approximately half of that in Homotherium, thus suggesting a weight of at least 100 kg for M. whitei. However, the width of the distal epiphysis of the humerus, which articulates with the radius, is greater in M. whitei than in a leopard or even in the male lion, which would suggest that the animal was somewhat larger than this latter species (i.e., around 200 kg). Obviously, these three independent estimates differ widely, although the most reliable of them is probably that obtained from the diaphysis of the humerus, given that the section of this long bone bears the weight of the forepart of

the body. If we consider this estimate correct, we are then dealing with a predator of about 100 kg, whose muscular strength used in immobilizing prey (estimated from the width of the distal epiphysis of the humerus) while it used its elongated canines to kill was four-fold greater than would be suggested by its food requirements or the speed at which it could eat (deduced from its markedly reduced carnassials). We therefore have a hypercarnivorous felid which would presumably generate large amounts of carrion, since it would exploit the carcasses of its prey to a small degree, thus leaVing enough meat and all within bone nutrients for the large hyaena P. brevirostris and for the hominids. In the light of this likely situation, the recent discoveries of evidence on human presence both at Southern Spain and Georgia associated with African M. whitei (Martinez-Navarro & Palmqvist 1995, 1996) are not surprising, since this latter species would have made the first dispersal of hominids to Eurasia in the lower Pleistocene possible, due to the greater scavenging opportunities it provided. Figure 25 shows the direct relationship between the abundance of juvenile individuals of those ungulate species preserved in the Venta Micena assemblage and the weight estimated for the adults; however, there are two distinct trends in this graph: the proportion of young increases gradually from 14.3% in H. alba to 59.3% in Bovini cf. Dmanisibos, decreases to 33.3% in Stephanorhinus etruscus, and increases again from the latter species to 80.0% in Mammuthus meridionalis. According with the above discussion, the first group of species (< 1000 kg) were presumably hunted by M. whitei and C. falconeri, while those of the second one (> 1000 kg) were probably predated by the large scimitar cat H. latidens. Figure 32 shows the results of separate regression analyses in both groups of species. The statistical fit obtained for the first set of ungulates (r = 0.967, p = 0.002) is better than that obtained for all species in the former analysis (r = 0.841, p = 0.004), which tends to strengthen the assumption of differences in the optimum prey-size for these carnivores. CONCLUSIONS

The macrovertebrate assemblage of Venta Micena presents an accumulated taphonomic stage and is constituted by demic and autochtonous palaeobiological entities recorded in situ (sensu Fern~ndez-LSpez 1991). This bone assemblage shows the highest diversity of large mammals recovered from any Plio-Pleistocene lacustrine site in the Guadix-Baza basin. The density of bones is very high, their orientation shows a random pattern, and more than 90% of the skeletal

42

Log(%juveniles) = -0.777(±0.546)+ 0.770(_+0.101)Log(weight of adults) r = 0.967, F = 58.271 (p = 0.002)

oB~,~ni,'~en, et sp. indet.

Megantereonw h z ~ e i ~ f e g ~ o ~ solilhacus

Stephanorhinus~ et~/sscus i

©

He~s~i~aidae,

gen. et sp. indet.

c-,/

Canis~Xenocyon) ~ ~ falconer1

N~~ Log(%juveniles) = -1.136(_+0.873) + 0.639(_+0. lOS)Log(weight of adults) r = 0.986, F = 34.726 (p = 0.107) I

2,

5

I

6 7 Log(weight of adults)

9

FIGURE 32 - Results of separate regression analyses of the abundance of juvenile individuals on m e a n adult weight for two groups of u n g u l a t e species from the Venta Micena assemblage, the first of which (< 1000 kg) were presumably h u n t e d by Megantereon whitei and Canis CXenocyon)falconeri, and the second one (> 1000 kg) by Hornotheriumlatidens. Reconstruction of Megantereon from A n t 6 n (1996); reconstruction of Homotherium from M a r t i n (1989). Rdsultatsd'analysesde rdgressionsdpardesde l'abondance

des individusjuvdniles sur le poids moyen adulte pour deux groupes d'esp~cesd'onguI~sde l'assernblagede VentaMicena, le premier d'entre eux (< 1 000 kg) dtaitprobablementchassdpar Megatereon whitei et Canis (Xenocyon) falconeri et le second (> 1 000 kg) par H o m o t h e r i u m latidens. Reconstitutionde Megantereon d'apr~sAntdn (1996);reconstitutiond ' H o m o t h e r i u m d'apr~sMartin (1989). elements are in contact with others. The surface of the bones was exposed to the effects of subaerial weathering for a very short time i n m o s t cases (0-1 years). The bones are not dispersed horizontally, since 20% of them are articulated and the remaining 80% are found associated. Biostratinomic fractures, gnawing marks and coprolites are very abundant. Differential fragmentation of major long bones by hyaenas is suggested by the close relationship between the abundance of epiphyses and complete elements, and their structural density and marrow content, respectively. Quantitative analysis of size/abundance patterns of ungulate species at Venta Micena indicates t h a t the p r e s e r v a t i o n a l bias produced during the taphonomic history of this assemblage affected predominantly species of smaller body

size. Once the original abundance of each species is estimated, the size/abundance relationship shows a good fit to the predictions of Damuth's model for a fossil assemblage formed by attritiohal mortality. The preservational characteristics of the bones from this fossil assemblage and the results of a multivariate comparison of the relative abundance ofpostcranial bones in assemblages from hyaena, leopard and porcupine lairs, as well as carnivore open feeding grounds and bone accumulations made by the man, suggest t h a t the Venta Micena assemblage was formed by the accumulation of skeletal remains near the entries of shallow dens dug by h y a e n a s near the ponds that surrounded the Orce lake. The use of descriptive and quantitative taphonomic analyses for the study of Venta Micena allows to propound a general model for the characterization of fossil

43 a s s e m b l a g e s g e n e r a t e d b y t h e activity of h y a e nids, for w h i c h this p a l a e o n t o l o g i c a l site is probably t h e b e s t e x a m p l e of t h e collecting-modifying b e h a v i o u r of Pachycrocuta brevirostris. - I n t e r s p e c i f i c a n a l y s i s of t h e p r o p o r t i o n o f j u v e n i le i n d i v i d u a l s a m o n g u n g u l a t e s in r e l a t i o n to t h e b o d y w e i g h t s e s t i m a t e d for a d u l t s indicates a s t r o n g selection of p r e y b y carnivores. This selection is c o r r o b o r a t e d b y t h e f i n d i n g of m a n y bones w i t h different o s t e o p a t h o l o g i e s s u c h as arthrosis, w h i c h l i m i t e d t h e locomotive capabilities of t h e a n i m a l s a n d t h e r e f o r e t h e i r ability to escape predation. T h e sex r a t i o of large bovids a n d equids, w h i c h is h i g h l y b i a s e d in f a v o u r of females, also p o i n t s to this h y p o t h e s i s of p r e y selection, as f e m a l e s are m o r e v u l n e r a b l e to p r e d a t i o n given t h e i r s m a l l e r b o d y size. T h e s e r e s u l t s t h u s fully confirm t h a t t h e V e n t a M i c e n a a s s e m b l a g e was f o r m e d b y a t t r i t i o n a l m o r t a l i t y p r o d u c e d by carnivores on t h e u n g u l a t e p o p u l a t i o n s , a n d allow us to rule out; t h a t t h e a s s e m b l a g e o r i g i n a t e d from c a t a s t r o p h i c m o r t a l i t y events. - C o m p a r i s o n of t h e r e l a t i v e f r e q u e n c i e s at w h i c h u n g u l a t e s of different sizes are killed a n d scavenged b y t h e m a i n A f r i c a n p r e d a t o r s , a n d t h e prop o r t i o n s in w h i c h different sized u n g u l a t e s are f o u n d in t h e V e n t a M i c e n a a s s e m b l a g e , s u g g e s t s t h a t h y a e n a s t h e r e fed l a r g e l y on c a r c a s s e s of anim a l s h u n t e d b y t h e l a r g e m a c h a i r o d o n t H. latidens a n d b y t h e h y p e r c a r n i v o r o u s canid C. falconeri. M o r p h o f u n c t i o n a l a n a l y s i s of the m e d i u m sized s a b e r - t o o t h M. whitei shows t h a t it m a y h a v e also p l a y e d a s i g n i f i c a n t ecological role in this p a l a e o c o m m u n i t y as a source of u n g u l a t e carcasses for b o t h h y a e n a s a n d hominids. Acknowledgments - This research was supported by a general grant from the "L.S.B. Leakey Foundation" and by a Spanish governmental major research grant, Projects PB94-1222-C02-02 and PB97-1082, from the "Direcci6n General de Investigaci6n Cientffica y T~cnica". Professor Marfa Teresa Alberdi allowed us to make the geochemical analyses in the laboratory of "Museo Nacional de Ciencias Naturales" of Madrid. We gratefully acknowledge to Professor Jordi Martinell (University of Barcelona) his encouragement to write this article for Geobios, to Mr. Mauricio AntSn (Museo Nacional de Ciencias Naturales, Madrid) his permission to use his drawings of Pachycrocuta and Megantereon, i~o Dr. Bienvenido Martlnez-Navarro (Museo de Paleontologfa de Orce, Granada) and to Dr. Alain Turq (Mus~e National de Pr~histoire, Les Eyzies de Tayac) their helpful comments. And, last but not least, we want to express our acknowledgment to Professor Claude Gu~rin (Universit~ de Lyon) and Professor Richard A. Reyment (University of Uppsala) their constructive review of the original manuscript. This is contribution number 10 of the "Orce Research Project". REFERENCES

AGUSTi J., AREIOLS. & MARTiN-SUAREZE. 1 9 8 7 - R o e d o r e s y l a g o m o r i b s ( M a m m a l i a ) del P l e i s t o c e n e i n f e r i o r d e

Venta Micena (depresidn de Guadix-Baza, Granada).

Paleontologia i Evolucio, Mem. esp., 1: 95-107. ALBERDI M.T. & Ruiz-BusTos A. 1989 - Taxonomfa y Biostratigraf/a de Equidae (Mammalia, Perissodactyla) en la Cuenca de Guadix-Baza (Granada). Trabajos Nedgeno-Cuaternario, 11: 239-271. ALBERDI M.T., ALCALAL., AZANZAB., CERDENOE., MAZO A.V., MORALESJ. & SESE C. 1989 - Consideraciones bioestratigrfificas sobre la fauna de Vertebrados fSsiles de la cuenca de Guadix-Baza (Granada, Espafia). Trabajos Nedgeno-Cuaternario, 11: 347-355. ANAI)0N P., JULIA R., DE DECKKER P., Rosso J.-C. & SOULm-MARsCHEI. 1987 - Contribucidn a la Paleolimnologfa del Pleistocene inferior de la cuenca de Baza (Sector Orce-Venta Micena). Paleontologia i Evolucio, Mere. esp., 1: 35-72. ANTONM. 1996 - Reconstruccidn de carnivores fdsiles. In GARCiA-PEREAR., BAQUEROR.A., FEt~N~NDEZ-SALvADOR R. & GISBERTJ. (eds), Carnivores: Evolucidn, Ecologia y Conservacidn. Consejo Superior de Investigaciones Cientificas, Madrid: 137-152. ARRmASA. 1995 - Consideraciones paleogeogrfificas preliminares de los dep6sitos plio-pleistocenos deI sector Orce-Venta Micena (Granada, Espafia), en relaci6n con los yacimientos paleontolSgicos. In GIBERT J. (coord.), Comunicaciones del Congreso Internacional de Paleontolog[a Hurnana de Orce. Ayuntamiento de Orce, Granada: 40-41. ARRIBASA., GIBERTJ., MARTiNEz-NAVARROB., GIBERTL. & ALBADALEJOS. 1994 - Geologia y paleontolog~a del sector Orce-Venta Micena (Cuenca de Guadix-Baza, Granada): estado actual de los conocimientos. In FERNANDEz-L(~PEZS. (coord.), Comunicaciones X Jornadas de Paleontologia. Universidad Complutense, Madrid: 23-25. BEHRENSMEYER A.K. 1978 - Taphonomic and ecologic information from bone weathering. Paleobiology, 4: 150-162. BEI-mENSMEYERA.K. 1991 - Terrestial Vertebrate Accumulations. In ALLISONP.A. & BRm~ D.E.G. (eds),

Taphonorny: Releasing the Data Locked in the Fossil Record, Plenum Press, New York: 291-327. BEHRENSMEYERA.K., WESTERN D. & DECHANTBOAZ D.E. 1979 - New perspectives in vertebrate paleoecology from a recent bone assemblage. PaIeobiology, 5: 12-21. BEHI~ENSMEYERA.K. & DECHANTBOAZ D.E. 1980 - The recent bones of Amboseli Park, Kenya, in relation to East African Paleoecology. In. BEHRENSMEYERA.K. & HmL A.P. (eds), Fossils in the Making. The University of Chicago Press, Chicago: 72-93. BIKNEVICIUSA.R., VANVALKENBURGHB. & WALKERJ. 1996 - Incisor size and shape: implications for feeding behaviors in saber-toothed "cats". Journal of Vertebrate Paleontology, 16: 510-521. BINFORD L.R. 1978 - N u n a m i u t Ethnoarchaeology. Academic Press, New York. BINFORD L.R. 1981 - Bones: Ancient Men and Modern Myths. Academic Press, New York. BINFORDL.R. & BERTRAMJ. 1977 - Bone frequencies and attritional processes. In BINFORDL. R. (ed.), For Theory Building in Archaeology. Academic Press, New York: 77-156.

44 BLUMENSHINE R.J. 1991 - H o m i n i d carnivory, foraging s t r a t e g i e s , a n d t h e socio-economic function of e a r l y archaeological sites. Philosophical Transactions of the Royal Society (London), B334: 211-221. BLUMENSHINE R.J. & CAROT.M. 1986 - U n i t flesh weights of some E a s t African bovids. African Journal of Ecology, 24: 273-286. BLUMENSHINE R.J. & MADRmALT.C. 1993 - Variability in long bone m a r r o w yields of E a s t African u n g u l a t e s a n d i t s zooarchaeological implications. Journal of Archaeological Science, 20: 555-587. BONIFAY M.F. 1971 - C a r n i v o r e s q u a t e r n a i r e s du SudE s t de l a France. M6moires M u s 6 u m d'Histoire Naturelle, N.S. serie C, 21: 43-274. BRAIN C.K. 1980 - Some criteria for the recognition of bone-collecting agencies in African caves. In BEHRENSMEYER A.K. & HILL A.P. (eds), Fossils in the Making. The U n i v e r s i t y of Chicago Press, Chicago: 108-130. BRAIN C.K. 1981 - The hunters or the hunted:An introduction to African cave taphonomy, 365 p., The U n i v e r s i t y of Chicago Press, Chicago. BRINK J.W. 1997 - F a t content in leg bones of Bison bison, a n d a p p l i c a t i o n s to archaeology. Journal of Archaeological Science, 24: 259-274. BUNN H. T. 1982 - Meat-eating and human evolution:

Studies on the diet and subsistence patterns of PlioPleistocene hominids in East Africa. Ph.D. Thesis (Anthropology), 514 p., University of California (unpublished). CALDER W.A. 1982 - The pace of growth: a n allometric a p p r o a c h to c o m p a r a t i v e e m b r y o n i c a n d poste m b r y o n i c growth. Journal of Zoology, 198: 215-225. CALDER W.A. 1984 - Size, Funtion, and life History. H a r v a r d U n i v e r s i t y Press, Cambridge. CASE T.J. 1978. On t h e evolution a n d a d a p t i v e significance of p o s t - n a t a l g r o w t h r a t e s in t h e t e r r e s t r i a l v e r t e b r a t e s . Quarterly Review of Biology, 53: 243282. CRISLER L. 1956 - O b s e r v a t i o n s of wolves h u n t i n g caribou. Journal of Mammalogy, 37: 337-346. DAMUTH J. 1981 - P o p u l a t i o n d e n s i t y a n d body size in m a m m a l s . Nature, 290: 699-700. DAMUTH J. 1982 - A n a l y s i s of t h e p r e s e r v a t i o n of comm u n i t y s t r u c t u r e in a s s e m b l a g e s of fossil m a m m a l s . Paleobiology, 8: 434-446. DAMUTH J. 1987 - Interspecific allometry of population density in m a m m a l s and other animals: the independence of body m a s s and p o p u l a t i o n energy-use. Biological Journal of the Linnean Society, 31: 193-246. DAMUTH J. 1991. Of size a n d abundance. Nature, 351: 268-269. EISENBERG J.F. 1981 - The mammalian radiationa U n i v e r s i t y of Chicago Press, Chicago. EMERSON A.M. 1990 - Archaeological implications of

variability in the economic anatomy of Bison bison. Ph.D. d i s s e r t a t i o n . W a s h i n g t o n S t a t e University, A n n Arbor, U n i v e r s i t y Microfilms. EMERSON S.B. & RAD]NSKYL. 1980 - F u n c t i o n a l a n a l y s i s of s a b e r t o o t h c r a n i a l morphology. Paleobiology, 6: 295-312. EWER R.F. 1973 - The Carnivores. 494 p., Cornell Univ e r s i t y Press, Ithaca.

FERNANDEz-LOPEZ S. 1991 - Taphonomic concepts for a t h e o r e t i c a l biochronology. Revista EspaFtola de Paleontologfa, 6: 37-49. FRANCILLON-VIEILLOTH., DE DUFFRI~NILV., CASTANETJ., GI~RANDUJ., MEUNIER F.J., SIRE J.Y., ZYLBERBERGL. & DE RICQULI~SA. 1990 - M i c r o s t r u c t u r e a n d m i n e r a l i zation of v e r t e b r a t e s k e l e t a l tissues. In CARTER J.G. (ed.), Skeletal biomineralization: patterns, processes and evolutionary trends. Van N o s t r a n d Reinhold, New York: 471-530. FRISON G.C. & TODDS L.C. 1986 - The Colby Mammoth

Site: taphonomy and archaeology of a Clovis kill in northern Wyoming. The U n i v e r s i t y of N e w Mexico Press, Albuquerque. GABUNIA L. & VEKUAA. 1995 - A Plio-Pleistocene hominid from Dmanisi, E a s t Georgia, C a u c a s u s . Nature, 373: 509-512. GIBERT J. • CAPORICCIR. 1989 - Tafonomia y paleoecologia del y a c i m i e n t o de V e n t a Micena. In GIBERT J., CAMPILLO D. & GARCIA-OLIVARESE. (eds), Los restos humanos de Orce y Cueva Victoria. Diputaci5 de Barcelona, Spain: 241-268. GIBERT J., CAPORICCIR., MARTINEz-NAVARROB. & ARRIBAS A. 1992 - Estudio tafon6mico de los met~podos de caballo de V e n t a Micena. In GIBERT J. (coord.),

Proyecto Orce-Cueva Victoria (1988-1992): Presencia humana en el Pleistoceno inferior de Granada y Murcia. A y u n t a m i e n t o de Orce, G r a n a d a , Spain: 115-132. GIBERT J., S~NCHEZ F., MALGOSAA. & MARTINEz-NAVARRO B. 1994 - D6couvertes des r e s t e s h u m a i n s d a n s les g i s e m e n t s d'Orce ( G r a n a d a , E s p a g n e ) . N o v e a u x r e s t e s h u m a i n s & Orce. Comptes Rendus de l'Acaddmie des Sciences de Paris, 319: 963-968. GIBERT Z. & PALMQVISTP. 1995 - F r a c t a l a n a l y s i s of t h e Orce skull sutures. Journal of Human Evolution, 28: 561-575. GIFFORD D.P. 1977 - Observations on modern human

settlements as an aid to archaeological interpretation. Ph.D. d i s s e r t a t i o n , U n i v e r s i t y of C a l i f o r n i a a t Berkeley, A n n A r b o r (unpublished). GIFFORD D.P. 1984 - Taphonomic specimens, Lake Turkana. National Geographic Research Reports, 17: 419428. GUERREROS. & PALMQVISTP. 1998 - Estudio morfom6trico del caballo de V e n t a M i c e n a (Orce, G r a n a d a ) y sn comparaci6n con los 6quidos a c t u a l e s y del PlioPleistoceno de E u r o p a y Africa. Paleontologia y EvoluciS, 30 (in press). HENSCHEL J.R. & SKINNERJ.D. 1990 - The diet of s p o t t e d h y a e n a s Crocuta crocuta in K r u g e r N a t i o n a l Park. African Journal of Ecology, 28: 69-82. HILL A.P. 1981 - A modern h y a e n a den in Amboseli National Park, Kenya. Proceedings Pan-African Congress on Prehistory and Quaternary Studies: 137-138. HILL A.P. 1984 - Modification b y m o d e r n s p o t t e d h y a e n a s (Crocuta crocuta). Abstracts from the First

International Conference on Bone Modification. N e v a d a , USA. HORWlTZ L.K. & SMITH P. 1988 - T h e effects of t h e s t r i p e d h y a e n a on h u m a n r e m a i n s . Journal of Archaeological Science, 15, 11 p.

45 HOWELLF.C. • PETTER G. 1980 - The Pachycrocuta and Hyaena lineages (Plio-Pleistocene and extant species of the Hyaenidae). Their relationships with Miocene ictitheres: Palhyaena and Hyaenictitherium. Geobios, 13: 579-623. JANIS C.M. 1990 - Correlation of cranial and dental variables with body size in ungulates and macropodoids. In DAMUTHJ. & MAcFADDENB.J. (eds), Body

Size in Mammalian Paleobiology: Estimation and Biological Implications. Cambridge University Press, Cambridge: 255-299. JONES K.T. t~ METCALFED. 1988 - Bar bones archaeology: bone marrow indices and efficiency. Journal of Archaeological Science, 8: 315-335. KERBIS-PETERHANSJ.C. & KOLSKAHORWITZL. 1992 - A bone assemblage from a striped hyaena (Hyaena hyaena) den in Negev Desert, Israel. Israel Journal of Zoology, 37: 225-245. KLETNR.G. & CRuz-URIBEK. 1983 - The computation of ungulate age (mortality) profiles from dental crown heights. Paleobiology, 9: 70-78. KLEIN R.G. & CRuz-URIBE K. 1984 - The Analysis of Animal Bones from Archaeological Sites, 266 p., The University of Chicago Press, Chicago. KRUIJKH. 1972 - The spotted hyena. A study of predation and social behavior, 335 p., The University of Chicago Press, Chicago. KUKAL Z. 1971 - Geology of Recent Sediments, 490 p., Academic Press, NewYork. LEGENDRE S. 1986 - Analysis of mammalian communities from the late Eocene and Oligocene of Southern France. Paleovertebrata, 16: 191-212. LEWIS M.E. 1997 - Carnivoran paleoguilds of Africa: implications for hominid food procurement strategies. Journal of Human Evolution, 32: 257-288. LYMANR.L. ].984 - Bone density and differential survivorship of fossil classes. Journal of Antropological Archaeology, 3: 259-299. LYMAN R.L. 1994 - Vertebrate Taphonomy, 524 p., Cambridge University Press, Cambridge. MAREANC.W. 1989 - Sabertooth cats and their relevance for early hominid diet and evolution. Journal of Human Euolution, 18: 559-582. MAREANC.W. & SPENCERL.M. 1991 - Impact of carnivore ravaging on zooarchaeological measures of element abundance. American Antiquity, 56: 645-658. MARTIN L.D. 1989 - Fossil history of the terrestrial Carnivora. In GITTLEMAN J.L. (ed.), Carnivore Behavior, Ecology, and Evolution. Chapman and Hall, London: 536-568. MARTfNEZ-NAYARROB. 1991 - Revisi6n sistem~itica y estudio cuantii;ativo de la fauna de macromamiferos del yacimientc de Venta Micena (Orce, Granada). Ph.D. Thesis. 264 p., Universitat Authnoma de Barcelona, Spain (unpublished). MARTINEZ-NAVARROB. 1992a - Revisi6n sistem~itica y estudio cuantitativo de la fauna de macromamfferos del yacimiento de Venta Micena (Orce, Granada). In GIBERT J. (coord.), Proyecto Orce-Cueva Victoria

MART~NEZ-NAvARROB. 1992b - Estudio cuantitativo y consideraciones paleoecolhgicas de la comunidad de mamfferos del yacimiento de Venta Micena (Orce, Granada). In GIBERTJ. (coord.), Proyecto Orce-Cueva

Victoria (1988-1992): Presencia humana en el Pleistoceno inferior de Granada y Murcia. Ayuntamiento de Orce, Granada, Spain: 155-187. MARTINEZ-NAVARROB. & PALMQVISTP. 1995 - Presence of the African machairodont Megantereon whitei (Mammalia, Carnivora, Felidae) in the Lower Pleistocene site of Venta Micena (Orce, Granada, Spain), with considerations about the origin, dispersal and evolution of the genus. Journal of Archaeological Science, 22: 569-582. MARTiNEz-NAVARRO

B. &

PALMQVIST P. 1996 - Presence of the

African saber-toothed felid Megantereon whitei (BRooM 1937) (Mammalia, Carnivora, Machairodontidae) in Apollonia-1 (Mygdonia Basin, Macedonia, Greece). Journal of Archaeological Science, 23: 869-872. MARTiNEZ-NAVARROB., TURQA., AGUST~J. & OMS O. 1997 - Fuente Nueva-3 (Orce, Granada, Spain) and the first human occupation of Europe. Journal of Human Evolution, 33: 611-620. McKINNEYM.L. 1996 - The biology of fossil abundance. Revista Espahola de Paleontologia, 11: 125-133. MENDOZA M., PALMQVIST P., GUERRERO S., MARTiNEZNAVARROB., PI~REZ-CLAROSJ.A., GIBERTJ. & EISENMANN V. 1993 - Consideraciones tafonhmicas, paleoecolhgicas y paleoetolhgicas sobre la comunidad de macromamiferos de Venta Micena (Orce, Granada). In GONZALEz-DoNosoJ.M. (ed.), Comunicaciones de las IX Jornadas de PaleontologEa. M~laga, Spain: 73-78. MENI~NDEZE. 1987 - C~rvidos del yacimientos del Pleistoceno inferior de Venta Micena-2, Orce (Granada, Espafia). Paleontologia i Evoluci6, mere. spec. I: 129181. MILLS M.G.L. 1989 - The comparative behavioral ecology of hyenas: The importance of diet and food dispersion. In GITTLEMANJ.L. (ed.), Carnivore Behavior, Ecology and Evolution. Chapman and Hall Ltd., London. MITCHELL B.L., SHENTON J.B. & UYs J.C.M. 1965 Predation on large mammals in the Kafue National Park, Zambia. ZoologicaAfricana, 1: 297-318. PALMQVIST P. 1991 - Differences in the fossilization potential of bivalve and gastropod species related to their life sites and trophic resources. Lethaia, 24: 287-288. PALMQVISTP. 1993 - Trophic levels and the observational completeness of the fossil record. Revista Espa~ola de Paleontologfa, 8: 33-36. PALMQVISTP. 1997 - A critical re-evaluation of the evidence for the presence of hominids in lower Pleistocene times at Venta Micena, Southern Spain. Journal of Human Evolution, 30: 83-89. PALMQVIST P., GIBERT J. & MARTINEz-NAVARRO B. 1992 Implicaciones tafonhmicas y paleoeeoldgicas de la relacidn tamafio-abundancia en la paleocomunidad

(1988-1992): Presencia humana en el Pleistoceno inferior de Granada y Murcia. Ayuntamiento de

de Venta Micena. In GIBERTJ. (coord.), Proyecto OrceCueva Victoria (1988-1992): Presencia humana en el Pleistoceno inferior de Granada y Murcia. Ayunta-

Orce, Granada, Spain: 21-85.

miento de Orce, Granada, Spain: 155-187.

46 PALMQVIST P., GIBERT J. & MARTiNEZ-NAvARROB. 1993 P r e c i s i o n e s s o b r e l a b i o e s t r a t i n o m l a de V e n t a Micena. Revista Espaftola de Paleontologia, 8: 206. PALMQVISTP., MARTINEz-NAVARROB. & ARRIBASA. 1996a P r e y selection by t e r r e s t r i a l carnivores in a lower Pleistocene paleocommunity. Paleobiology, 22: 514534. PALMQVISTP., Pl~REZ CLAROSJ.A., GIBERT J. & SANTAMARiA J.L. 1996b - C o m p a r a t i v e m o r p h o m e t r i c s t u d y of a h u m a n p h a l a n x from t h e Lower Pleistocene site a t C u e v a Victoria (Murcia, Spain), by m e a n s of F o u r i e r analysis, s h a p e coordinates of l a n d m a r k s , principal a n d r e l a t i v e w a r p s . Journal of Archaeological Science, 23: 95-107. PLAZIATJ.-C. 1984 - Le biotope palustre: m6thode d'identification pal6o6cologique. Geobios, M6m. sp. 8: 313-320. PETERS R.H. 1983 - The ecological implications of body size, 329 p., C a m b r i d g e U n i v e r s i t y Press, Cambridge. POTTS R. 1982 - Lower Pleistocene site f o r m a t i o n a n d h o m i n i d activities a t Olduvai Gorge, Tanzania. Ph. D. d i s s e r t a t i o n , H a r v a r d University, A n n Arbor: U n i v e r s i t y Microfilms. RAWN-SCHATZINGERV. 1992 - The s c i m i t a r cat Homotherium serum COPE: osteology, functional morphology, a n d p r e d a t o r y behavior. Illinois State Museum Reports of Investigations, 47, 80 p. REED K. E. 1997 - E a r l y h o m i n i d evolution a n d ecological c h a n g e t h r o u g h t h e African Plio-Pleistocene. Journal of Human Evolution, 32: 289-322. REYMENT R.A. & JORESKOG K.G. 1993 - Applied Factor Analysis in the Natural Sciences, 371 p., C a m b r i d g e U n i v e r s i t y Press, Cambridge. RICHARDSONP.R.K. 1980 - C a r n i v o r e d a m a g e to antelope bones a n d its archaeological implications. Palaeontologia Africana, 23: 109-125. ROOK L. 1994 - The Plio-Pleistocene Old World Canis (Xenocyon) ex. gr. falconeri. Bollettino della Societd Paleontologica Italiana, 33:71-82. ROTH V.L. 1990 - I n s u l a r dwarf elephants: a case study in body m a s s estimation and ecological inference. In DAMUTH J. & MACFADDEN B.J. (eds), Body Size in

Mammalian Paleobiology: Estimation and Biological Implications. Cambridge University Press, Cambridge: 151-179. SCHALLERG.B. 1972 - The Serengeti Lion. A study of predator-prey relations, 480 p. The U n i v e r s i t y of Chicago Press, Chicago. SCOTT K.M. 1990 - Postcranial dimensions of ungulates as predictors of body mass. In DAMUTH J. & MAC FADDEN B.J. (eds), Body Size in Mammalian Paleobiology: Estimation and Biological Implications. Cambridge U n i v e r s i t y Press, Cambridge: 301-335. SCOTT L. & KLEIN R.G. 1981 - A h y e n a - a c c u m u l a t e d bone a s s e m b l a g e f r o m L a t e Holocene d e p o s i t s a t Deelplan, O r a n g e F r e e State. Annals of the South African Museum, 86: 217-227. SHIPMANP. 1981 - Life history of a fossil. An Introduction to Taphonomy and Paleoecology. H a r v a r d U n i v e r s i t y Press, C a m b r i d g e , M a s s a c h u s s e t t s . SKINNER J.D., DAVIS S. & ILANI G. 1980 - Bone collecting b y s t r i p e d h y a e n a s , Hyaena hyaena, in I s r a e l . Palaeontologia Africana, 23: 99-104.

SKINNER J.D., HENSCHEL J.R. & VAN JAARSVELDA.S. 1986 - Bone-collecting h a b i t s of s p o t t e d h y a e n a s Crocuta crocuta in t h e K r u g e r N a t i o n a l Park. South African Journal of Zoology, 21: 303-308. SORIA F.J., LOPEZ-GARRIDO A.C. • VERA J.A. 1987 An~lisis estratigr~fico y sedimentol6gico de los dep6sitos n e 6 g e n o - c u a t e r n a r i o s en el sector de Orce (depresi6n Guadix-Baza). Paleontologia i Evolucio, Mem. esp., 1: 11-34. SUTCLIFFE A.J. 1970 - S p o t t e d - h y e n a : crusher, gnawer, digester a n d collector of bones. Nature, 227: 11101113. TIXIER J., ROE D., TURQ A., GIBERT J., MARTINEz-NAVARRO

B., ARRIEAS A., GIBERT L., MAILLO L. 8~ IGLESIAS A. 1995 - Pr4sence d'industries lithiques dans le Pl4istoc~ne inf~rieur de la r~gion d'Orce (Granada, Espagne): E t a t de la Question. Comptes Rendus de l'Acaddmie des Sciences de Paris, 321, 2a: 71-78. TURNER A. 1990 - The evolution of t h e guild of l a r g e r terr e s t r i a l carnivores in t h e Plio-Pleistocene of Africa. Geobios, 23: 349-368. TURNER A. 1992 - Large carnivores a n d e a r l i e s t Europ e a n hominids: c h a n g i n g d e t e r m i n a n t s of resource a v a i l a b i l i t y d u r i n g t h e lower a n d m i d d l e Pleistocene. Journal of Human Evolution, 22: 109-126. TURNER A. & ANTON M. 1996 - The g i a n t h y a e n a , Pachycrocuta brevirostris ( M a m m a l i a , C a r n i v o r a , H y a e n i dae). Geobios, 29: 455-468.

TURQ A., MART~NEz-NAvARRO

B., PALMQVIST P., ARRIBAS A.,

AGUSTIA. & RODRIGUEZ-VIDALJ. 1996 - Le Plio-Pl~istoc~ne de la r~gion d'Orce, province de G r e n a d e , E s p a gne: Bilan et perspectives de recherches. Paleo: 161204. VALENSI P. 1991 - l~tude des s t r i e s de b o u c h e r i e s u r les ossements de cerf 41aphe des n i v e a u x s u p 4 r i e u r s de la grotte du L a z a r e t (Nice, A l p e s - M a r i t i m e s ) . L'Anthropologie, 95: 797-830. VALVERDE J.A. 1967 - Estructura de una comunidad mediterrdnea de vertebrados terrestres, 218 p., Ediciones del C.S.I.C., M a d r i d . VAN VALKENBURGB.V. 1 9 9 0 - S k e l e t a l a n d d e n t a l predictors of body mass in carnivores. Pp. In DAMUTH J. & MAcFADDEN B.J. (eds), Body Size in Mammalian

Paleobiology: Estimation and Biological Implications. Cambridge U n i v e r s i t y Press, Cambridge: 181-205. VAN VALKEN~URGHV. & RUfF C.B. 1987 - C a n i n e tooth s t r e n g t h a n d killing b e h a v i o u r in l a r g e carnivores. Journal of Zoology, 212: 379-397. VERA J.A. 1970a - E s t u d i o estratigr~fico de l a Depresi6n de Guadix-Baza. Boletin Geoldgico y Minero, 81: 429462. VERA J.A. 1970b - Facies del Plioceno de l a Depresi6n de Guadix-Baza. Cuadernos de Geologia de la Universidad de Granada, 1: 23-25. VERA J.A., FERNANDEZ J., LOPEz-GARRIDO C.A. & RODRiGUEZ-FERNANDEZJ. 1985 - Geologla y e s t r a t i g r a f i a de los m a t e r i a l e s plioceno-pleistocenos del sector de Orce-Venta Micena (Prov. G r a n a d a ) . Paleontologia i Evoluci5, 18: 3-11. VILLA P., BOUVILLE C., COURTIN J., HELMER D., SHIPMANP., BELLOUMINI G. & BRANGA M. 1 9 8 6 - C a n n i b a l i s m in the Neolithic. Science, 233: 431-437.

47 VRBA E.S. 1!980 - The significance of bovid r e m a i n s as i n d i c a t o r s os e n v i r o n m e n t a n d p r e d a t i o n p a t t e r n s . I n BEHRENSMEYERA.K. & HILL A.P. (eds), Fossils in the making. Vertebrate taphonomy and paleoecology. The U n i v e r s i t y of Chicago Press, Chicago: 247-271. WELLS N.A. 1983 - C a r b o n a t e deposition, physical liranology a n d e n v i r o n m e n t a l l y controlled c h e r t formation in Paleocene-Eocene l a k e Flagstalf, C e n t r a l UTAH. Sedimentary Geology, 35: 263-296. WESTERN D. 1979 - Size, life h i s t o r y and ecology in m a m mals. African Journal of Ecology, 17: 185-204. WESTERN D. 1980 - L i n k i n g t h e ecology of p a s t a n d pres e n t m a m m a l communities. In BEHRENSMEYERA.K. & HILL A.P. (eds), Fossils in the Making. The U n i v e r s i t y of Chicago Press, Chicago: 41-54. WOLF R.G. 11975 - S a m p l i n g a n d s a m p l e size in ecologi-

cal a n a l y s i s of fossil m a m m a l s . Paleobiology, 1: 195204. ZIHLMAN

A.L.

&

LOWENSTEIN

J.M.

1996

- A

Spanish

Olduvai? Current Anthropology, 37: 695-697. A. ARRIBAS

Museo Geominero Instituto TecnolNgico Geominero de Espafia (ITGE) R~os Rosas, 23 E-28003 Madrid P. PALMQVIST

Departamento de Geologla y Ecologia (Area de Paleontologla) Facultad de Ciencias, Universidad de Mfilaga E:29071 Mfilaga E-maih [email protected].