Palaeodiversity 2: 379–385; Stuttgart, 30.12.2009. 379

Taphonomic reinterpretation of a bone sample of endemic Pleistocene from (Greece): osteoporosis versus regurgitation

ISABELLE ROBERT ATTARD & JELLE W. F. REUMER

Abstract A sample of fossil deer remains (genus Candiacervus) from Mavro Mouri cave, Crete, Greece is studied be- cause a large proportion shows aberrant, seemingly pathological changes to the morphology and the structure of the bones. Here we show that the aberrant appearance of the Mavro Mouri bone sample can be understood through a simple taphonomical explanation: post-mortem damage to the bones inflicted by bearded vultures (Gypaetus bar- batus). Hence they are not osteoporotically altered due to supposed malnutrition, as proposed before. This conclu- sion has a bearing on our understanding of island evolution, as it eliminates a supposed direct relation between persistent malnutrition on the one hand, and island dwarfing on the other. Keywords: Candiacervus, Gypaetus barbatus, taphonomy, malnutrition, osteoporosis, island evolution, , endemism.

Zusammenfassung Es werden fossile Hirsch-Reste (Gattung Candiacervus) aus der Mavro Mouri Höhle, Kreta, Griechenland, untersucht, von denen ein großer Teil ungewöhnliche, scheinbar pathologische Veränderungen ihrer Morphologie und Struktur zeigt. Diese Veränderungen der Knochen können durch einen einfachen taphonomischen Prozess erklärt werden. Es sind durch den Bartgeier (Gypaetus barbatus) verursachte postmortale Beschädigungen. Die Veränderungen sind also keine durch Mangelernährung verursachten Osteoporose-Erscheinungen wie bisher an- genommen wurde. Diese Erklärung beeinflusst unser Verständnis der Evolution von Inselfaunen, da sie eine ver- mutete direkte Beziehung zwischen fortdauernder Mangelernährung und Zwergwuchs ausschließt.

Contents 1. Introduction ...... 379 2. Material and methods ...... 380 3. Description of the material...... 380 4. A taphonomical hypothesis ...... 381 5. Discussion ...... 382 5.1. Taphonomy ...... 382 5.2. Evolutionary theory ...... 384 6. References ...... 384

1. Introduction hopping’ from Asia Minor during phases of low sea-lev- el. Like most islands in the Mediterranean Basin, the is- Mavro Mouri is a karstic locality on the north coast of land of Crete harboured a typical endemic fauna Crete, a few kilometers west of the town of Rethymnon. during several phases of the Pleistocene (REESE 1996 and The site has yielded a large sample of bones of the Late articles and references therein). During Early and Middle Pleistocene endemic fossil deer Candiacervus cretensis Pleistocene times, the island harboured endemic elephants and related taxa of the Cretan cervid species flock (e. g., (Mammuthus spp.), hippo’s (Hippopotamus creutzburgi KUSS 1975; SONDAAR 1977; DE VOS 1979, 1984). A consider- BOEKSCHOTEN & SONDAAR, 1966) and a suite of rat-sized able percentage of these bones show a phenomenon that rodents (Kritimys spp.). During the Late Pleistocene the appears to be persistently misunderstood in the paleonto- fauna consisted of an endemic rodent (Mus minotaurus logical literature, i. e., an aberrant morphology showing, BATE, 1942), a (Crocidura zimmermanni WETTSTEIN, among other, a thin cortex, large openings in the cortex, 1953), a species flock of cervid deer (Candiacervus spp.) and strangely worn articulation surfaces. It has led re- and a lutrine otter (Isolalutra cretensis SYMEONIDIS & searchers to various explanations. The bizarre condition of SONDAAR, 1975). There appear to have been at least two the bones has been explained as being the result of either different immigration waves and subsequent extinctions. malnutrition, of disease (osteoporosis, anaemia), or of the The could have reached Crete by means of ‘island influence of boring insects (or insect larvae). It was 380 PALAEODIVERSITY 2, 2009

2. Material and methods

Material from the Mavro Mouri cave was excavated in the 1960’s/1970’s by several groups, among which a team of paleontologists from Utrecht University under the guid- ance of the late Dr. P. Y. S ONDAAR. Deer material was de- scribed by DE VOS (1979, 1984), while other remains were subsequently published by several researchers (see REESE 1996 for an overview). Even before the formal description of the deer material by DE VOS (1979, 1984) it was apparent that a large proportion of the deer bones from Mavro Fig. 1. First published depiction of the regurgitated bones from Mouri (more specifically from Mavro Mouri IV level C) Mavro Mouri, from SONDAAR 1977. had an aberrant appearance; this was preliminarily de- scribed and illustrated by SONDAAR (1977). Before or since, other authors have written about it (KUSS 1975; BRABER SONDAAR (1977: 694–695) who provided the following de- 1981). The Mavro Mouri Candiacervus material is now in scription: “At one Pleistocene locality on Crete, most of the collection of Naturalis, the National Museum of Natu- the deer bone exhibited osteoporosis. (Osteoporosis is a ral History, Leiden, the Netherlands, and is temporarily rarefying condition of bones which lose much of their being kept in the Natural History Museum in Rotterdam, mineral matter and become fragile and often deformed). the Netherlands. The bones are thin, the nutritive canals are very wide and In order to identify the agent responsible for the bone the articular surfaces are sharp-edged. Evidently there accumulation, we used two criteria: (1) the frequency of was a failure in bone formation while bone loss was pro- skeletal elements, and (2) the presence or absence of marks ceeding normally.” indicating alteration of the bone. The material here pub- Of these different explanations, the suggestion of se- lished was studied by the senior author during a short stay vere malnutrition and subsequent osteoporosis (SONDAAR in Rotterdam. The total amount of material consists of 1977) had led scientists to propose malnutrition as one of 3029 bones and bone fragments. Measurements were the possible reasons for dwarfing in larger island mam- made with manual calipers, photographs were made with mals (e. g., KÖHLER & MOYA-SOLA 2005). It astonished us a Nikon Coolpix 3MPixel camera. that similar malnutrition-related osteological phenomena have not been described from other islands, neither in the Mediterranean Basin nor in other regions. However, this 3. Description of the material should have been expected if the malnutrition caused the dwarfing, because dwarfing is a globally observable char- The number of bones studied is given in Tab. 1, where acteristic of insular larger . A proper under- the different elements are presented in a sequence from standing of the osteological phenomena observed in the the most proximal to the most distal position in the skele- Mavro Mouri sample (and there only and so far not in ton. We based our studies on the quantitative results of other samples from Cretan sites) is thus of importance for BRABER (1981), and in addition we searched through the the theory of island evolution in larger mammals. boxes labeled ‘indeterminate’ in order to retrieve more The original black-and-white line drawings provided material. The bones were then separated in two categories: by SONDAAR (1977, here reproduced as Fig. 1) do not allow with or without indications of alteration. The results are precise identification of the causative agent. New discov- given in Tab. 2, leaving out the cranial and axial material. eries in taphonomy (e. g., ROBERT & VIGNE 2002) allow us The following features are observed on the bones: to re-evaluate the original diagnosis of malnutrition and – thinning of the cortex in the diaphysis osteoporosis, and provide a new explanation for the strange – roughening of the cortical surface, ultimately leading bone deformations. to formation of holes in the bone surface – enlargement of foramina Acknowledgements – deformation of the articular surfaces The comparative study about Mavro Mouri and the bearded – subtle size reduction of the entire skeletal element vulture was supported by the CNRS (APN program) ‘Référen- tiels taphonomiques créés par des agents non humains mod- These features surprisingly conform to the results pro- ernes’. Dr. DAV I D MAYHEW (Rotterdam) and Dr. JOHN DE VOS vided by ANDREWS (1990) for digestion of microfaunal re- (Leiden) critically read the manuscript. mains regurgitated by nocturnal raptors. ATTARD & REUMER, TAPHONOMIC REINTERPRETATION OF PLEISTOCENE DEER BONES 381

Tab. 1. Relative presence of skeletal elements in the Mavro cal changes). This interpretation, viz. the impression that Mouri IV C sample, arranged in an increasingly distal order the observed phenomena are due to disease of some kind, (NISP = number of identified specimens). has haunted literature since. In different publications, the body Total following three explanations were subsequently provided: Skeletal element NISP % region % P o l y c h a e t a . – Boreholes in the bone were suppos- antler 98 5.0 edly made by polychaete worms of the genus Polydora maxillary 42 2.1 cranial 24.2 (Polychaeta, family Spionidae); this must have occurred at mandibula 33 1.7 a time when high sealevel caused the bones in the cave to isolated teeth 306 15.5 be in contact with seawater (KUSS 1975). However, never vertebrae and sacrum 83 4.2 was there an indication found that the sea entered the cave axial 4.2 ribs and sternum – – at any time after the bones were deposited. Therefore this scapula 28 1.4 conclusion seems unrealistic. girdle 1.4 coxal – – Malnutrition and osteoporosis. – A se- humerus 13 0.7 vere osteoporosis was proposed; it was interpreted as a stylo- femur 11 0.6 1.4 result of failure in bone formation due to malnutrition pode patella 2 0.1 (food shortage; SONDAAR 1977). Osteoporosis is “a sys- radius 91 4.6 temic skeletal disorder of diverse ethiology characterised zeugo- ulna 37 1.9 11.4 by decreased bone mass and microarchitectural deteriora- pode tibia and malleolus 97 4.9 tion of bone tissue with a consequent increase in bone fragility and fracture susceptibility” (PRITZKER & KESSLER carpals 96 4.9 1998). There is a decrease in bone build-up, and, more astragalus 52 2.6 basi- 14.7 specifically, a decrease in the amount of trabeculae in the pode calcaneum 76 3.8 spongiosa. The bone surface is not visibly modified. It is tarsals 68 3.4 often caused by a dietary calcium deficiency, and hence meta- metapodia 349 17.7 17.7 the result of malnutrition. Thus, in veterinary medicine, pode one way to handle osteoporosis is firstly to increase the phalange 1 99 5.0 calcium input above the minimal level, secondly to opti- phalange 2 116 5.9 acro- 23.8 mize the calcium absorption, and thirdly to offer the pos- pode phalange 3 256 13.0 sibility for an increased amount of movement to the ani- sesamoids 23 1.2 mals. The result is an increased build-up of the bones, a total of identifiable bones 1976 100.0 retarded resorption of the trabecular structures of the unidentifiable fragments 1053 – bones and a higher resistance against fractures of the total Candiacervus sp. 3029 – bones (PLESKER & ZWERGER 2002). In our samples we did not observe decrease in spongi- ose trabeculae, while, on the contrary, the bone surface is 4. A taphonomical hypothesis strongly modified, which eliminates osteoporosis as the cause of the modifications observed. In addition, it is dif- What are we observing? KUSS (1975) was the first au- ficult to envisage a calcium deficiency in the diet of ani- thor to publish about the Mavro Mouri bones. He wrote in mals living in Crete, which is an island largely consisting a first publication: “Sie erwecken zunächst den Eindruck, of Mesozoic and Cenozoic limestones. als handele es sich hier um krankhafte Veränderungen” (at Malnutrition and insects. – Malnutrition first sight they provide the impression of being pathologi- was supposed to have occurred while the deer were in the

Tab. 2. Relative amount of Candiacervus bones from Mavro Mouri and in the recent reference sample from Corsica showing semi- digestion due to gastric juices. Cranial and mandibular bones are here not included.

% affected in skeletal category n total n semi-digested % semi-digested reference sample girdle bones 28 5 17.9 5.8 long bones incl. ribs and sternum 588 251 42.7 22.5 compact bones (from Tab. 3) 316 152 48.1 53.6 phalanges 471 366 77.7 77.5 total of identifiable bones 1403 774 55.2 52.5 unidentifiable fragments 1053 259 24.6 38.0 382 PALAEODIVERSITY 2, 2009 juvenile stage, leading to thinner bones during life (but: tabolism. Occasionally it happens that some swallowed see above), in combination with post-mortem attack of the bones escape complete digestion and are regurgitated in- bone by boring insects, e. g. Lepidoptera (BRABER 1981). side the nest. Regurgitation year after year creates an ac- The different sizes of the holes in the bone were attributed cumulation of semi-digested bones within the nesting to different species of insects. BRABER based her conclu- caves. Bearded vultures still breed in Crete, Corsica and sion on a single observation of insect larvae emerging the Pyrenees and have recently been re-introduced in the from a gnu’s horn (KRUUK 1975). However, firstly this is a Alps. Fossil Gypaetus barbatus material was found in different substance (keratin). Secondly, if insect larvae Crete in three cave-sites (WEESIE 1988): Liko cave, Gerani would be implicated, the exit openings would either all be IV and Mavro Mouri. of similar size and shape, or – if different insect species H y p o t h e s i s . – Based on (1) the resemblance of are involved – the sizes and shapes of the exit holes would our material with known cases of small mammals being group in two, three or even four categories. Such is not the regurgitated by predators, (2) the presence in Crete of no case in the Mavro Mouri material. We consider the insect other mammalian predators or avian raptors than the option to be an unlikely explanation. bearded vulture capable of devouring (partial) ungulate These possibilities eliminated, we still have to find the carcasses, and (3) the presence of Gypaetus barbatus ma- responsible agent. As the phenomena correspond to diges- terial in several caves, including Mavro Mouri itself, we ANDREWS tion traces as described for microfauna by (1990), hypothesize that the Mavro Mouri aberrant deer bones are we propose an agent that digested the deer bones to in fact regurgitated bones deposited by bearded vultures. a certain degree. We therefore have to look among the faunal elements known from Pleistocene Crete, and/or to find similarly altered (recent) bones elsewhere with known 5. Discussion responsible agent. P r e d a t o r . – Looking for predators capable of in- 5.1. Taphonomy gesting larger mammals (i. e., small deer-sized) the choice is among mammalian and avian candidates (reptilian can- The contents of Corsican bearded vulture nest sites didates such as crocodiles, varanid monitors or large py- were used (by the first author) to create a taphonomic thons were not present in the context of a Pleistocene model with the aim of recognising bone accumulations in Mediterranean island). The only mammalian predator on paleontological and archaeological sites (ROBERT & VIGNE Pleistocene Crete was an otter (Isolalutra cretensis), quite 2002). The cervid bones from Mavro Mouri were com- incapable of devouring deer. Avian raptors of appropriate pared to this reference sample and the results are com- size were present, however. Here we exclude nocturnal pleted with Pyrenean material. avian raptors that concentrate on mice, and small Of the total amount of Cretan Candiacervus bones re- as being unlikely candidates. Of other (diurnal) rap- covered (Number of Identified Specimens = 1403), 55.2 % tors, WEESIE (1988) mentioned Aquila chrysaetos simurgh showed semi-digestion marks, such as the mentioned en- WEESIE, 1988, Haliaetus albicilla (LINN., 1758), Gyps larged foramina and very thin cortical surfaces (Tab. 2). melitensis (LYDEKKER, 1890), Gyps fulvus (HABLIZL, 1783), This proportion is similar to that in the reference material Aegypius monachus (LINN., 1766), and Gypaetus barbatus from modern nests (52.5 %). The most severely attacked (LINN., 1758). skeletal elements are phalanges, carpals and tarsals, and Aquila chrysaetos simurgh (Falconidae) is the endem- metapodials, i. e., the distal parts of the limbs (see also ic, somewhat larger form of the golden eagle. The recent Tab. 3), while proximal parts are considerably less vulner- golden eagle is capable of handling rabbit-size prey (e. g., able to (semi)digestion; this is probably because they less rabbit, hare, juvenile fox, marten, grouse), and a slightly easily detach from the body (Tab. 2). larger animal might have handled slightly larger prey, per- There is quite a difference in degree of damage to the haps juvenile dwarf deer, but certainly nothing more volu- various parts of the skeleton. Sometimes compact bones minous. The same applies to Haliaetus albicilla. are only etched on the surface. Mostly the effect is more Gyps melitensis, Gyps fulvus and Aegypius monachus severe, e. g., the calcaneum is often perforated (Fig. 2-1a, are scavengers, preferentially eating soft bodyparts. On b); the astragalus shows articulation surfaces with sharp the other hand, Gypaetus barbatus is a known scavenger ridges that could never have functioned as proper articula- of medium sized mammal skeletons. The bearded vulture tion surfaces in a living animal. This is an additional proof is a very large cave nesting . Its diet consists of bones against a possible pathological origin of our phenomena. removed from carcasses of ungulates (TERRASSE 2001). Phalanges are among the most heavily attacked bones: Some of these bones measure up to 25 cm long. The highly 77 % both in the Mavro Mouri sample and in the recent acidic (pH = 1) gastric juices in its stomach extract organic material show digestion marks (Tab. 2; Fig. 2-2b). The substances from the bones, necessary for the bird’s me- only significant difference between the Mavro Mouri ATTARD & REUMER, TAPHONOMIC REINTERPRETATION OF PLEISTOCENE DEER BONES 383

Tab. 3. Relative amount of compact Candiacervus bones from Mavro Mouri showing semi-digestion.

type of compact bone n total n semi-digested % semi-digested scaphoid 21 10 47.6 other carpals 75 14 18.6 patella 2 1 50.0 astragalus 52 46 88.5 calcaneum 76 59 77.6 navico-cuboid 67 15 22.4 sesamoids 23 7 30.4 total 316 152 48.1

Fig. 2. Regurgitated goat (Capra hircus) and chamois (Rupicapra rupicapra) bones from bearded vulture nests in the French Pyre- nees compared to Pleistocene Cretan deer (Candiacervus cretensis) bones from Mavro Mouri. – 1. Calcaneum. 1a. Capra hircus; 1b. Candiacervus cretensis. 2. Phalange I. 2a. Capra hircus; 2b. Candiacervus cretensis. 3. Metatarsus. 3a. Rupicapra rupicapra; 3b. Candiacervus cretensis. 384 PALAEODIVERSITY 2, 2009 sample and the recent bone sample is in the percentage of digested deer bones (hyena scats?)”; thus interpreting metapodia preserved: 17.7 % in Mavro Mouri versus 5.1 % them as of mammalian coprolithic origin. in the recent material. Bearded vultures are the only spe- cies known capable of swallowing and regurgitating entire metapodials without breaking them in the process. Given 5.2. Evolutionary theory the high percentage of attacked metapodia in Mavro Mou- ri, metapodia were apparently indispensible in the bird’s SONDAAR (1977) provided the first illustration of an diet. But, apart from this difference, damage locations and inflicted bone from Mavro Mouri in his seminal paper on damage intensity are identical in our fossil material when mammalian island evolution. He then used the – in his compared to modern ungulate bones regurgitated by the view – diseased osteoporotic bones as evidence for mam- bearded vulture. malian island evolution, i. e., dwarfism in larger mam- We wonder why scavenging by bearded vultures has so mals, when he wrote: “There is fossil evidence indicating far escaped from researchers’ notice when evaluating the mass starvation and malnutrition which could point to Mavro Mouri bones, especially given the fact that the bird overpopulation through lack of predators and the conse- still lives on Crete nowadays. WEESIE (1988), unfortunate- quent overgrazing and destruction of otherwise suitable ly, did not analyse the full depth of the consequences of the biotopes. After this the population density declined drasti- composition of his avian faunules. He supposed that dwarf cally which might have been an important factor promot- elephants were transported through the air by the large ing the rapid evolution of island mammals.” Thus, his line eagles of which remains were found: the very young ele- of reasoning was as follows: large herbivorous mammals phants and hippopotamuses were envisaged as eagles’ arrive on an island (through a sweepstake route), they pro- prey. For this, he used the mythical Roc as an example: in liferate in the absence of suitable predators, then they a Persian legend a large bird of prey carried three live overpopulate the island, and suffer from a lack of food and (baby dwarf?) elephants in its claws at the same time. Re- malnutrition. This malnutrition shows its effects in the cent Aquila chrysaetos preys preferably on animals of bones, viz. the ‘osteoporotic’ bones from Mavro Mouri. around rabbit-size (marmot, hare, neonate chamois, alpine This stage is followed by a population collapse, the popu- grouse, etc., see DUQUET 1995), so, in Pleistocene Cretan lation then goes through renewed genetic bottleneck(s), reality, the eagles would have nourished themselves pref- favouring animals that require less food, i. e. herbivores erably with birds (prey like rails, ducks, crows, etc.), and with smaller body mass. Dwarfed mammals are the final probably the available mice and shrews instead. WEESIE result. (1988: 63) hypothesized that the large Aquila chrysaetos The peculiar aspect of the deer bones from Mavro simurgh took mainly endemic deer as its prey and that the Mouri is, however, not due to any disease, and neither to extinction of the deer led to the extinction of the large ea- malnutrition leading to severe osteoporosis. All the deer gle. present in this cave assemblage died naturally and part of In his discussion WEESIE appears to have missed the the material was carried inside by the vultures. As malnu- dietary habits of bearded vultures: “However, vultures trition is not the reason of the phenomena, it cannot be usually feed on carrion and they never carry away entire used as an explanation for endemic morphological devel- prey” (WEESIE 1988: 61). Apparently, the feeding habits of opments such as dwarfism, or for final extinction of the vultures (Gyps, Aegypius) that preferably feed on soft car- endemic deer in the island of Crete. We consider island rion are mistakenly also applied to the bearded vulture dwarfism to have adaptive reasons that may certainly be that habitually does transport entire prey carcasses (TER- related to the limited size of the islands, to the absence of RASSE 2001). terrestrial predators and/or to the potentially limited It thus appears that the seemingly pathological changes amount of available nutrition, but it is not triggered by of the Cretan bones have been caused by the bearded vul- malnutrition. ture Gypaetus barbatus. It is therefore of post-mortem origin, and a case of taphonomy, not of pathology. Analys- ing the quantity of Candiacervus bones, we can conclude 6. References that successive pairs of bearded vulture probably nested in ANDREWS, P. (1990): Owls, Caves and Fossils. 231 pp.; London the Mavro Mouri cave during a few centuries, and in this (British Museum (Natural History)). period produced the remarkable faunal accumulation that BRABER, F. (1981): Afwijkingen bij hertebotten uit het Pleisto- has caused so much misinterpretation. ceen van Kreta: osteoporotisch of postmortaal ontstaan? 30 Bones that were digested by bearded vulture’s gastric pp.; Unpublished MSc. dissertation, Utrecht University, Utrecht. juices may have escaped attention more often. As an ex- DE VOS, J. (1979): The endemic Pleistocene deer of Crete. – Pro- ample, VALENTE (2004: 618) depicts digested deer bones ceedings Koninklijke Nederlandse Akademie van Weten- (phalanges, carpalia) with the caption “Carnivore damage: schappen, B 82 (1): 59–90. ATTARD & REUMER, TAPHONOMIC REINTERPRETATION OF PLEISTOCENE DEER BONES 385

DE VOS, J. (1984): The endemic Pleistocene deer of Crete. – al research: Diseases: 415–460; San Diego (Academic Verhandelingen der Koninklijke Nederlandse Akademie Press). van Wetenschappen, afd. Natuurkunde, Eerste reeks, 31: REESE, D. S. (ed.) (1996): Pleistocene and Holocene Fauna of 1–100. Crete and its First Settlers. – Monographs in World Archae- DUQUET, M. (1995): Inventaire de la faune de France. Vertébrés ology, 28: 1–422. et principaux Invertébrés. 416 pp. ; Paris (Muséum national ROBERT, I. & VIGNE, J.-D. (2002): The Bearded vulture (Gypaetus d’histoire naturelle, Nathan). barbatus) as an accumulator of archaeological bones. Late KÖHLER, M. & MOYA-SOLA, S. (2005): Reduction of brain and glacial assemblages and present-day reference data in Cor- sense organs in the fossil insular bovid Myotragus. – Brain, sica (Western Mediterranean). – Journal of Archaeological Behavior and Evolution, 63: 125–140. Science, 29 (7): 763–777. KRUUK, H. (1975): Hyaena. 80 pp.; Oxford (Oxford University SONDAAR, P. Y. (1977): Insularity and its effects on mammal evo- Press). lution. – In: HECHT, M. K., GOODY, P. C. & HECHT, B. M. KUSS, S. E. (1975): Die pleistozänen Hirsche der ostmediterranen (eds.): Major patterns in vertebrate evolution: 671–707; New Inseln Kreta, Kasos, Karpathos und Rhodos (Griechenland). York (Plenum Publishing Corporation). – Berichte der Naturforschenden Gesellschaft zu Freiburg i. TERRASSE, J.-F. (2001): Le gypaète barbu. 208 pp. ; Lausanne, Br., 65: 25–79. Paris (Delachaux et Niestlé). PLESKER, R. & ZWERGER, C. (2002): Rickets in Primates. – Pri- VALENTE, M. J. (2004): Humans and Carnivores in the Early Up- mate Report 62, January 2002: 69–78. per Paleolithic in Portugal: Data from Pego do Diabo Cave. PRITZKER, K. P. H. & KESSLER, M. J. (1998): Diseases of the – Revue de Paléobiologie, 23 (2): 611–626. Musculoskeletal System. – In: BENNETT, B. T., ABEE, C. R. WEESIE, P. D. (1988): The Quaternary Avifauna of Crete, Greece. & HENRICKSON, R. (eds.): Nonhuman primates in biomedic- – Palaeovertebrata, 18 (1): 1–94.

Adresses of the authors: ISABELLE ROBERT ATTARD, Centre Guillaume le Conquérant, 13 bis rue de Nesmond, 14402 Bayeux, France E-mail: [email protected] Corresponding author: JELLE W. F. R EUMER, Faculty of Geosciences, Utrecht University, P. O. Box 80021, 3508 TA Utrecht, the Netherlands and Natural His- tory Museum, Rotterdam, the Netherlands E-mail: [email protected]

Manuscript received: 13.12.2008, accepted: 6.3.2009.