Palaeoreas Gaudry, 1861

Geraads, Denis, Spassov, Nikolaï & Kovachev, Dimitar, 2003, Palaeoreas lindermayeri (Wagner, 1848) (Mammalia, Bovidae) from the upper Miocene of Bulgaria, and a revision of the species, Geodiversitas 25 (2), pp. 405-415 : 406-414

Genus Palaeoreas Gaudry, 1861

TYPE SPECIES. — Antilope lindermayeri Wagner, 1848 . DIAGNOSIS. — See Bouvrain 1992: 57.

Palaeoreas lindermayeri ( Wagner, 1848) HOLOTYPE. — Horn-core fragment ( Wagner 1848: pl. 12, fig. 5) (No. 520, Institute of Paleontology, Munich).

TYPE LOCALITY. — Pikermi ( Greece).

DESCRIPTION

Cranial material prevail among the material from Hadjidimovo-1. We have counted 184 fairly complete horn cores, 95 frontlets, 65 skulls and skull fragments with horn cores and 20 neurocranials. This amounts to a total of 369 skull fragments and skulls, possibly representing the largest known sample of a single bovid species from a single locality. The material includes also many mandibles. Limb bones are less common, at least in the Assenovgrad collections, and we have not tried to sort them out. In contrast to the Pikermi material, that of Hadjidimovo-1 is usually not crushed, and the bone surface is not polished as it is so often the case at Pikermi, allowing observation of more details. Palaeoreas lindermayeri has already been described by several authors ( Gaudry 1861; Gentry 1971; Bouvrain 1980, 1992; Geraads & Güleç 1999). Nevertheless, the very rich and well preserved material permits us not only to repeat the main characters of the species, but also to revise and confirm them in a very large sample and to insist upon some other poorly known features.

Skull ( Fig. 1 )

The face is long, with the rear border of M3 at the level of, or slightly behind, the front edge of the orbit. The face is also low, the fronto-nasal profile being deeply concave. In lateral view, the upper nasal line intersects the orbit, as in Pikermi. It is probably in the belief that the Pikermi skulls are deformed that Bouvrain (1992: fig. 7) reconstructed the face higher than it really was, with too straight a facial profile.

The premaxillary-nasal contact is moderately long. The nasals have a slightly convex upper profile, but none is complete anteriorly ( Fig. 1 ). Only one specimen has an ethmoidal fissure, but most specimens do not have any. The anteorbital depression is variable in size and depth, but is always present. The infra-orbital foramen opens above P2 or P2-P3; there are often two or several of them. The anterior border of the choanae is usually at the level of the middle of M3.

The frontal is strongly bent between the horn-cores, with its anterior and posterior parts at an angle of about 90°-100°. The supra-orbital foramina are small, and located in deep and wide depressions. The analysis of male skulls of different individual age leads to the conclusion that the supra-orbital foramina decrease in size with age from subadult to adult individuals and migrate to the upper edge of the frontal pit. This must be born in mind in taxonomic interpretations. The inter-frontal suture is always closed, and often elevated behind the horns. The fronto-parietal suture is also almost always closed. None of the frontals show any evidence of sinuses. There is a large and deep post-cornual fossa, extending onto the post-orbital bar.

The brain-case is of regular width. The temporal lines are weak, but the squamoso-parietal suture is often raised as a well defined ridge.

The occipital is rather low and broad, with both sides facing mostly posteriorly, separated by a broad raised area. The paroccipital processes are curved forwards. The mastoid exposure is of moderate width, its lower part faces laterally, and sometimes produces a small mastoid apophysis. In one case at least, the mastoid is fused with the squamosal in the post-tympanic area. This is of course a component, together with the lack of frontal sinuses, fusion of frontal sutures, and thickening of supra-otic ridges, of a general trend towards strengthening of the brain-case.

The basi-occipital is rather long, and much narrower anteriorly than posteriorly, but the anterior tuberosities are prominent ( Fig. 1C ). All specimens have a central depression, all the way from the foramen magnum to the presphenoid.

The bulla commonly bears a rounded longitudinal ridge, stronger anteriorly along the medial border of the tympano-hyal. The foramen ovale is small and rounded.

The horn-cores diverge by an angle of about 40- 45°. They are usually twisted on their axis, with rarely a tendency towards some slight spiralling. The posterior keel is constant, but variably expressed. There is no anterior keel, but there may be a ridge instead, which usually becomes sharper upwards. The pedicle is always made of compact bone, without sinuses.

Dentition ( Fig. 2 )

All medium-size dentitions from Hadjidimovo must be referred to Palaeoreas , there being no other Bovid of similar size. Their main features agree with those given by Gentry (1971) except that the front and back edges of p3 and p4 are rarely transversal, and that there is usually (but not always) a deep labial groove between protoconid and hypoconid on p4. The metaconid of p4 is more transversal than that of p3, and it may be anteriorly expanded so as to close the middle lingual valley on worn teeth.

Female skulls

Three crushed adult skulls from Hadjidimovo-1 (HD-3454, HD-5134 and HD-unnumbered) are of the same size as the horned skulls described above, but are hornless. Their dentitions are identical, in size and morphology, with those of the horned P. lindermayeri skulls, but these skulls differ from them, besides their hornless condition, by their weaker basicranial angle, weaker anterior tuberosities of basioccipital, and less advanced fusion of the inter-frontal suture.

Female skulls of Palaeoreas were previously ( Gentry 1971; Bouvrain 1980, 1992) thought to bear smaller horns than males, but neither the bivariate plot of basal horn-core diameters ( Fig. 3 ), nor the principal component analysis on the most complete skulls ( Fig. 4 ) show any clear distinction between two groups. Furthermore, these three skulls cannot be referred to any other taxon (the remaining Bovid fauna consists of Boselaphines, Antilopines, and very few cf. Protoryx ). Therefore, it must be concluded that the female of P. lindermayeri was hornless (see part Taphonomy, ecology, ethology and relationships).

DISCUSSION

Variation in Palaeoreas

P. lindermayeri is by far the best known species of the genus. It is mainly known from Pikermi, but also from Samos ( Gentry 1971), Halmyropotamos (Melentis 1967), Vathylakkos 1 and 3, Ditiko 1 and 2, Prohoma ( Bouvrain 1992) and Kemiklitepe ( Geraads & Güleç 1999). It is also present in Kalimantsi. Other reports from Turkey are either based on teeth alone, and are therefore unreliable, or on incorrectly identified horn-cores (such as those from Gülpinar, see Tekkaya 1973). A skull from Titov Veles in Macedonia was referred by Ciric (1957) to this species, but this is also incorrect, because the outline of the horn cores is concave laterally in front view, whereas it is always convex in P. lindermayeri . Therefore, the geographic range of this species is quite restricted: Greece, southern Bulgaria and westernmost Turkey. Thus, interestingly, this species, which is among the dominant ones in the rich Turolian sites of South Bulgaria (see below), has quite a limited range and has not been reported, e.g., from the Northern Pontic region where numerous rich localities are known.

By almost all its dimensions, the P. lindermayeri from Hadjidimovo-1 is larger than that of Pikermi. This is especially true of teeth and horn dimensions, and brain case size ( Table 1; Figs 3-5 ).

However, there does not seem to be any difference in proportions, the Hadjidimovo-1 fossils looking just like larger versions of the Pikermi ones, which are quite similar to the specimens from Kalimantsi. However, a few specimens from the Assenovgrad Museum collection, of which the only measurable one is HD-3155, have a different facies, and are slightly smaller and more similar to those from Pikermi. According to one of us (DK), they most probably come from another Hadjidimovo locality, Hadjidimovo- Tumbichkite, that is from deposits several tens of meters higher than the level of Hadjidimovo-1; they could be contemporaneous with Pikermi. The single specimen from Halmyropotamos cannot be distinguished from the Pikermi sample. A specimen from the latest Miocene of Ditiko ( Bouvrain 1980) differs from the Pikermi sample by its larger size, relatively smaller horns, smaller post-cornual fossa, more waisted basi-occipital, and more inflated bulla. Bouvrain (1980) considered that these differences do not warrant specific distinction. On the contrary, she recognised as P. zouavei Bouvrain, 1980 a skull from the early Turolian of Ravin des Zouaves n° 5 (RZO), which shares with the Ditiko specimen a larger size and relatively smaller horns as compared with the Pikermi material. Its supra-orbital foramina are larger than in P. lindermayeri , but other differences look weak and were perhaps accentuated by deformation. For example the infra-orbital foramina are only very slightly more posterior than at Hadjidimovo-1, the orbit is rather less prominent than in the latter sample, and its relative position to the tooth row is the same. Table 2 lists the most significant differences among of the various Palaeoreas populations; it shows that their classification is not straightforward. The material from Kalimantsi and Kemiklitepe does not differ from the typical Pikermi population. That of Hadjidimovo-1 is characterised by a large body size combined with large horn-cores and small supraorbital pits but we do not think that this warrants specific distinction. The materials from other localities are too poor, at the present time, for their taxonomic positions to be definitely settled, but we believe that the differences between the various samples reflect the intraspecific population variability in time and space, at no more than subspecific level. We think it would be unwise to draw biochronological conclu-

Bovids from the upper Miocene of Bulgaria sions from the inter-populational variations of this species.

The Turkish species Palaeoreas elegans Ozansoy, 1965 and P. brachyceras Ozansoy, 1965 have spiralled (instead of twisted) horn-cores, unfused inter-frontal suture, and supra-orbital foramina of normal shape, and there is no reason to include them in the genus Palaeoreas . P. asiaticus from Garkin ( Köhler 1987) was considered by Gentry & Heizmann (1996: 383) as “doubtfully separate from P. lindermayeri ”, but it is smaller than the latter species, the inter-frontal suture is visible, and the posterior keel is almost straight, instead of spiralled. These are all similarities with “ P. ” elegans , and we think it is closer to this species than to P. lindermayeri , even though the tight spiralling recalls that of the latter species.

Taphonomy, ecology, ethology and relationships The large number of frontlets and horn-cores in the taphocoenosis provide the opportunity to calculate the minimum number of individuals at about 260. They were concentrated in isolated rock lenses with an area of tens of square meters and a maximum thickness of less than 1 m. The very high male/female ratio can be explained by

Geraads D., Spassov N. & Kovachev D.

differential preservation, the male skulls with their very thick braincases being far more robust than female ones. The same bias presumably acted against juvenile specimens. Similarly biased sex-ratios are found, for the same reason, in Tragoportax , both in Pikermi and Hadjidimovo-1. However, it is unlikely that taphonomic bias was the sole factor for this concentration of the remains of hundreds of male individuals, and it is likely that in the thanatocoenosis there were mainly herds of males.

Palaeoreas lindermayeri is the dominant species of Hadjidimovo-1 among 29 taxa ( Spassov 2002). Such antelopes of medium to large size living in relatively large herds are usually mixed feeders or grazers (groups 3-5 of Gagnon & Chew 2000) living in open landscapes. According to Solounias et al. (1999) this species was mostly a browser. However, some of its morphological features fit better those of a grazer. Its teeth, although not hypsodont by modern standards, are clearly not brachyodont. Furthermore, although no skull is perfectly preserved in this area, some of the best specimens show that the premaxillae are wide, and that they are inclined downwards, with their anterior tip below the occlusal plane. These two features denote grass-feeding, according to Spencer (1995). It is clear, in any case, that Palaeoreas did not inhabit a forest-like, tree-dense woodland.

The taphonomic data as well as the frontal morphology and sexual dimorphism in horns could also suggest a social structure close to that of some Reduncini ( Kobus kob ) including territorially dominant males as well as both male and female herds. Territoriality in Palaeoreas is shown by the presence of an ante-orbital depression, hence of a pre-orbital gland. The horns probably played an important role in display and intraspecific tournaments. The development of such a socialization usually takes place in Bovids inhabiting spaces open enough, living in herds and having a high degree of visual communication. It is found neither in the low-social inhabitants of dense woods, nor in the super-social Bovids of

Bovids from the upper Miocene of Bulgaria the fully open spaces (see Geist 1974; Janis 1986). The best Recent analogue is perhaps Antilope cervicapra , a species where the horns are highly evolved and specialized, with an important social function ( Geist 1966). They usually gather in herds of 20-30 to several hundreds, living in open plains covered with scrub; they penetrate open forests, but avoid dense forests ( Prater 1971). This agrees with the environment that we have suggested above for Palaeoreas ( Spassov 2002) .

The lack of sinuses in the frontal bone, pedicle, and horn-core, as well as the lack of horns in females, provide a few more clues about social structure and “agonistic” behaviour. These morphological features are strongly correlated in modern Bovids. Species with hornless females usually lack any frontal sinus (the main exception being Aepyceros ), and in the species lacking frontal sinus the horns are small or absent in females (the main exceptions being Tragelaphus euryceros and, less clearly, Taurotragus ). Therefore, absence of horns in female Palaeoreas is not unexpected. According to Roberts (1996), it denotes a low level of competition among females.

The thickness of horn-cores and overall robustness of the rear part of the cranium suggest intensive use of horns in intra-specific fighting. This is in contrast with modern Tragelaphines, which share with Palaeoreas the fused frontal suture and spiralled horns, but which use their large horns primarily for display. Another behaviour that can be definitely ruled out is violent frontal clashing (“Rammkampf”) such as performed by some Caprines and Bovines, because it is always associated with extensive frontal and cornual sinuses. Lack of sinuses and torsion of horn-cores in P. lindermayeri are again similarities with Antilope , although the overall shape of the skull is more derived than in this genus. Overall, the cranio-cornual morphology of Palaeoreas has no modern equivalent, and it is hard to draw definitive conclusions about its fighting behaviour. We could suggest at least that its horns had an active display role and also were used in tournament fights of frontal pressure type, similar to Antilope ( Geist 1966) .

This has certainly no bearing on the systematic position of Palaeoreas , whose skull is quite distinct from that of Antilope . Reduncines could be a modern equivalent for the social structure, but their forwardly directed horn tips and related very prominent anterior tuberosities of basioccipital, typical for the tribe, are missing here. It is more reasonable not to refer Palaeoreas to any modern tribe.

Figures

FIG. 1. — Palaeoreas lindermayeri (Wagner, 1848), Hadjidimovo-1; A, skull HD- 5112, oblique antero-lateral view; B, same specimen, front view; C, skull HD-5120, oblique latero-ventral view of the cranial base. Scale bar: A, B, 10 cm; C, 5 cm.

FIG. 2. — Palaeoreas lindermayeri (Wagner, 1848) from Hadjidimovo-1; A, lower tooth row (HD-2443); B, upper tooth row (HD-3664). Scale bar: 5 cm.

FIG. 3. — Plot of antero-posterior vs. transverse diameters of Palaeoreas lindermayeri (Wagner, 1848) horn-cores. Mean measure- ments for each site are underlined. They include also those specimens where only one measurement can be taken.

FIG. 4. — Plot of upper premolar vs. upper molar length of Palaeoreas lindermayeri (Wagner, 1848). Means and symbols as for Fig. 3.