Aves

Geobios 49 (2016) 29–36

Available online at

ScienceDirect www.sciencedirect.com

Palaeontology of the upper Miocene vertebrate localities of Nikiti (Chalkidiki Peninsula, Macedonia, Greece)

Aves§ George D. Koufos a,*, Dimitris S. Kostopoulos a, George E. Konidaris a,b a

Aristotle University of Thessaloniki, Department of Geology, Laboratory of Geology and Palaeontology, GR-54124 Thessaloniki, Greece Eberhard Karls University of Tu¨bingen, Palaeoanthropology, Senckenberg Center for Human Evolution and Palaeoenvironment, Ru¨melinstr. 23, 72070 Tu¨bingen, Germany

b

A R T I C L E I N F O

A B S T R A C T

Article history: Received 23 October 2014 Accepted 19 January 2016 Available online 27 January 2016

The upper Miocene avian localities of Greece are rare and the known material is poor. During the last campaigns of excavations in Nikiti 2, a tarsometatarsus of a large struthionid has been unearthed, enriching the poor known Greek sample of late Miocene birds. Morphological and comparative features suggest similarities with the extant and fossil large struthionids referred to Struthio. Several taxa of this genus are known from the upper Miocene but the studied tarsometatarsus is more similar to S. karatheodoris. The latter is known from Samos (Greece, type locality) and Pikermi (Attica, Greece). Albeit their similarity, the lacking of comparable material from the type locality does not allow for a confident determination; it is therefore referred to as Struthio cf. karatheodoris. Based on morphological resemblances, it is quite possible that the north peri-Pontic taxon S. brachydactylus is a junior synonym of S. karatheodoris. Two coexisting Struthio lineages have been recognized in southeastern Europe during the upper Miocene to Pliocene: S. karatheodoris/S. brachydactylus and S. novorossicus. The second lineage is larger and with higher ‘‘degree of didactyly’’ than the former one. The late Miocene southeastern European ostriches were living in a palaeoenvironment, which does not fit with that of the extant African ostriches, apart maybe for S. molybdophanes. ß 2016 Elsevier Masson SAS. All rights reserved.

Keywords: Aves Struthionidae Late Miocene Greece Taxonomy Biogeography Palaeoenvironment

1. Introduction There are a few known Upper Miocene avian localities in Greece, and the available avian remains are very scarce (Fig. 1). Struthionids are seldom found in the Upper Miocene deposits of Greece (Fig. 1), following the general scarcity of avian remains during this time interval. The first struthionid remains originate from the Upper Miocene mammal localities of Samos Island, Greece (Forsyth Major, 1888), where they were described as Struthio karatheodoris Forsyth Major, 1888. Quite later, the presence of struthionids was recognized in Pikermi, Greece, (Bachmayer and Zapfe, 1962) and were described as S. cf. karatheodoris; the presence of the species was recently confirmed (Michailidis et al., 2010, pers. comm. 2014). The discovery of a struthionid in Nikiti 2 (NIK) is the third evidence for the presence of Struthio in Greece and the first in northern Greece, adding new information for this taxon and enriching the list of the bird bearing localities of Greece. NIK is situated in the Chalkidiki Peninsula (northern Greece) and includes a rich mammal fauna, which is studied in the present volume. The fossiliferous site is located in the upper part of the §

Corresponding editor: Gilles Escarguel. * Corresponding author. E-mail address: [email protected] (G.D. Koufos).

http://dx.doi.org/10.1016/j.geobios.2016.01.012 0016-6995/ß 2016 Elsevier Masson SAS. All rights reserved.

Nikiti Fm., consisting mainly of sands, gravels, loose conglomerates and red-brown sandy marls. It is dated to the early Turolian (MN 11) based on its mammal fauna. More information about the stratigraphy, locality and age are given in Koufos (2016) and Koufos et al. (2016).

2. Material and methods The studied specimen is housed in the Laboratory of Geology and Palaeontology, University of Thessaloniki (LGPUT). The measurements on the tarsometatarsus are given in Fig. 2; they were taken using a digital caliper and they are given in mm with an accuracy of 0.1 mm. The comparative material of the modern ostrich is housed in the NMNHS; it has been measured by G.E.K. Abbreviations: Morphology: cmh, crista medialis hypotarsi; cmp, crista medialis plantaris; DAP, antero-posterior diameter; dfi, dorsal fossa infracondylaris; DT, transverse diameter; laf, lateral articular facet of the proximal articular surface; maf, medial articular facet of the proximal articular surface; mdaf, median articular facet of the proximal articular surface; tII, trochlea II; tIII, trochlea III; tIV, trochlea IV. Localities, museums and institutes: LGPUT, Laboratory of Geology and Palaeontology, University of Thessaloniki; MTL, Mytilinii-1, Samos Island, Greece; NIK, Nikiti 2, Macedonia,

G.D. Koufos et al. / Geobios 49 (2016) 29–36

30

Fig. 1. Sketch map with the Neogene avian-bearing localities of Greece and their associated avifauna. 1. Pikermi, PIK (Gaudry, 1862–1867; Michailidis et al., 2010); 2. Chomateres, CHO (Michailidis et al., 2010); 3. Samos, SAM (Forsyth Major, 1888; Martin, 1903; Lydekker, 1891); 4. Kerassia 4, KES (Michailidis et al., 2010); 5. Perivolaki, PER (Boev and Koufos, 2006); 6. Aegina, AGN (Mlı´kovsky´, 1996), 7. Megalo Emvolon, MEV (Boev and Koufos, 2000); 9. Kryopigi, KRY (Zelenkov et al., 2015); 9. Nikiti 2, NIK (this work). Asterisks indicate the localities with struthionids.

Greece; NHML, Natural History Museum of London; NMNHS, National Museum of Natural History of Sofia, Bulgaria; MGL, Geological Museum of Lausanne; PIK, Pikermi, Attica, Greece. 3. Systematic palaeontology Order Struthioniformes Latham, 1790 Family Struthionidae Vigors, 1825 Genus Struthio Linnaeus, 1758 Type-species: Struthio camelus Linnaeus, 1758 Struthio karatheodoris Forsyth Major, 1888 Origin of the name: The origin of the species name for Samos Struthio is not mentioned by Forsyth Major (1888) but it possibly originates from Alexandros Karatheodoris, who was Samos Overlord at that time. Forsyth Major indicated that the first

person who gave him information about the fossils was the Overlord and probably gave his name to the new Struthio species. Holotype: A complete femur from Samos, Greece, which is referred without description and figures by Forsyth Major (1888); later it was described and figured by Martin (1903: p. 204; textfigs. 30–31). Mlı´kovsky´ (2002) and Boev and Spassov (2009) noted that the location where the holotype is stored is unknown but the keeper of MGL informed us that it is housed in the MGL with catalogue number MGL S 317 (Marchant, pers. com. 2011). Type locality: The holotype is labeled as ‘‘Adriano’’, Samos (Forsyth Major, 1894). The Adrianos ravine is the main fossiliferous area of the Mytilinii Basin in Samos, excavated by many scientists and amateurs. The locality after its rediscovery was named Mytilinii-1 (MTL), including four fossiliferous sites (Kostopoulos et al., 2009); the holotype probably originates from one of them.

G.D. Koufos et al. / Geobios 49 (2016) 29–36

Fig. 2. Struthio camelus, recent, NMNHS-3/2003. Measurements of the left tarsometarsus. M1. Maximal height; M2. Proximal DT; M3. Proximal DAP; M4. DT of the lateral proximal articular facet; M5. DAP idem; M6. DT of the medial proximal articular facet; M7. DAP idem; M8. DT of the median proximal articular facet; M9. DAP idem; M10. DT of the shaft in the middle; M11. DAP idem; M12. DT distal; M13. DT of tIV; M14. DAP idem; M15. DT of tIII; M16. Minimum DAP ˆ . Angle between the longitudinal axis of tIII; M17. DAP of the medial condyle of tIII; a of the bone and that of tIV. Scale bars: 10 cm (a), 5 cm (b, c).

Age: The age of the old Samos collections is a long-time debate among palaeontologists. The recent study of the stratigraphy, as well as the study of the new collection from the Mytilinii Basin suggest that all Adrianos fossil faunas are correlated to the end of middle Turolian, MN 12; the magnetostratigraphic record suggests an estimated age of 7.1 Ma for MTL (Kostopoulos et al., 2003; Koufos et al., 2009). Struthio cf. karatheodoris Fig. 3 Locality: Nikiti 2 (NIK), Chalkidiki Peninsula, Macedonia, Greece. Age: Early Turolian, MN 11, late Miocene. Material: Left tarsometatarsus, NIK-917. Measurements (in mm; measurements are given in Fig. 2): M1 = 485; M2 = 86.3; M3 = 72; M4 = 35; M5 = 41; M6 = 30.6; M7 = 43; M8 = 27.8; M9 = 38.7; M10 = 35.9; M11 = 35.5; M12 = 68.7; M13 = –; M14 = –; M15 = 42; M16 = –; M17 = 46. Description: The studied tarsometatarsus is complete, but its proximal and distal epiphyses are weathered and deformed (Fig. 3). The proximal epiphysis has a triangular aspect with a strong crista medialis hypotarsi (cmh); the crista medialis

31

plantaris (cmp) is high (its height decreases gradually) and extends across the shaft, covering 1/3 of the bone’s height (Fig. 3(a, h)). The proximal articular surface consists of three articular facets. The medial articular facet (maf) is elliptical, slightly larger than the lateral one and its dorso-ventral axis is slightly oblique to the medio-lateral one of the proximal articular surface (Fig. 3(h)). The lateral articular facet (laf) is also elliptical and concave; its dorso-ventral axis is vertical to the medio-lateral one of the proximal articular surface. It bears a deep hole in the center, which looks like the biting trace of a carnivore; a similar hole is observed in the ventral surface of the proximal epiphysis (Fig. 3(a, h)). Between these two facets, there is a median one (mdaf) which is elevated, slightly convex and rectangular; its plantar half is not preserved (Fig. 3(h)). The proximal epiphysis bears a large and deep dorsal fossa infracondylaris (dfi; Fig. 3(b)). A groove starts from this fossa and extends across the dorsal surface of the shaft, being weaker in the distal part of the shaft. The shaft is elongated and thick; its transverse section is triangular in the 2/3 of its height, being more elliptical at the distal end. The distal epiphysis is strongly compressed dorso-laterally and deformed, while the weathering partly destroyed it; however, it preserves the trochleae of the metapodials. The tII forms a vestigial apophysis in the medial surface of the tIII (Fig. 3(e, g)). The tIII is larger than tIV and consists of two condyles. The tIII of NIK-917 is split longitudinally in two parts by a deep groove giving the impression that the tarsometatarsus bears three trochleae (Fig. 3(a, b, f, g)); this is probably due to the compression and lateral deformation of the bone. The lateral condyle of tIII is larger than the medial one. The tIV is completely separated from the tIII by a deep and wide groove (Fig. 3(a, b, f, g)). It is directed laterally and it is oblique to the long axis of the tarsometatarsus; the angle aˆ is 20o. The angle aˆ corresponds to that between the longitudinal axis of the bone and that of the tIV (Fig. 2). Struthio is didactyl but the degree of didactyly (expressed here by the angle aˆ) varies among species: a small angle aˆ indicates less expressed didactyly than a larger one. The tIV consists of two condyles separated by a shallow groove; the lateral one is the only preserved, and it is large with a large depression in its lateral surface. 4. Discussion The erection of the species S. karatheodoris was originally based on a single femur from ‘‘Adriano’’, Mytilinii Basin, Samos (Forsyth Major, 1888). This author gave limited information about the morphology of the holotype, reporting only its larger dimensions than the modern S. camelus. Few years later, more information was given for the holotype, based on two photographs (Martin, 1903: p. 204; text-figs. 30, 31); this author also described a fragment of pelvis from Samos given to him by Forsyth Major. According to Martin (1903) both specimens belong to S. karatheodoris and their most important differences from the modern S. camelus are:  the neck of the caput femori is stouter in the fossil species;  the fossa intercondylaris in the distal part of the femur is more elongated in S. karatheodoris;  the sacral vertebrae of the fossil Struthio are stouter;  the centra of the two ‘‘true sacral vertebrae’’ are narrower and more rounded in S. karatheodoris. In conclusion, Martin (1903) mentioned that the Samos ostrich has different proportions than the modern one, pointing toward S. karatheodoris as a valid species. Bachmayer and Zapfe (1962) described three distal fragments of tibiotarsus and a basal phalanx of the third digit of Struthio from Pikermi, which are morphologically similar to those of S. camelus but quite larger in size. The authors separated the fossil ostriches in

32

G.D. Koufos et al. / Geobios 49 (2016) 29–36

Fig. 3. Struthio cf. karatheodoris, Nikiti 2 (NIK), Chalkidiki Peninsula, Macedonia, Greece; early Turolian, MN 11. Left tarsometatarsus, NIK-917. a. Ventral view; b. Dorsal view; c. Medial view; d. Lateral view; e. Lateral view of the distal part indicating the vestigial trochlea II; f. Dorsal view of the distal trochleae; g. Ventral view of the distal trochleae; h. Proximal articular surface in ventral view; i. Distal articular surface in ventral view. Scale bars: 10 cm (a-d), 5 cm (e-i).

two groups: a small-sized group, including those close in size to the modern S. camelus, and a large-sized one with larger dimensions than the modern ostrich. The Pikermi material is correlated to the second group; it is referred to S. cf. karatheodoris because there are

not comparable anatomical parts from Samos for a complete comparison, and due to the size and proportion variability of the Pikermi material. Nevertheless, in a recent study based on new, previously unpublished material, S. karatheodoris has been

G.D. Koufos et al. / Geobios 49 (2016) 29–36

confidently recognized in the Pikermi avifauna (Michailidis et al., 2010, pers. com. 2014). 4.1. Comparison with the modern S. camelus The NIK tarsometatarsus is morphologically similar to that of the extant ostrich. Its ‘‘degree of didactyly’’ is similar to that of the modern ostrich as the value of angle aˆ falls within the range of variation for the modern S. camelus (Table 1). The maximal dorso-ventral axis of the proximal medial articular facet (maf) is more oblique in NIK-917 than in the modern ostrich, but this difference could be reinforced due to the deformation of NIK-917. The NIK tarsometatarsus has similar proportions (parallel Log-ratio profiles) but it is slightly longer (M1) and significantly larger than that of the modern ostriches (Fig. 4(a)). The dorso-ventral diameter of the laf (M5), as well as the transverse diameter of the mdaf (M8) and of the shaft (M10) are significantly larger than those of the modern ostriches (Fig. 4(a)). 4.2. Comparison with fossil ostriches Late Miocene ostriches are known from several localities of Eastern Europe and Western Asia, but their material is poor and rare. Their taxonomy is also complicated with several synonymies (Boev and Spassov, 2009). The latter authors recognized four different species of Struthio in the upper Miocene of Eurasia:  S. orlovi Kurocˇkin and Lungu, 1970, which has a size similar to S. brachydactylus, but its morphology is different. It is known from the northern peri-Pontic area and it is the only Vallesian taxon of Eurasia;  S. novorrossicus Alexejew, 1915, which is the largest known late Miocene Struthio, with clear differences from S. karatheodoris (Fig. 4(b)). It is known from the Turolian of Ukraine;  S. brachydactylus Burcˇak-Abramovich, 1949, which is characterized by a small size and low ‘‘degree of didactyly’’. It is known from the northern peri-Pontic area;  S. karatheodoris Forsyth Major, 1888, which is a robust Turolian ostrich known from southeastern Europe and probably Middle East. The Maragheh and Siwaliks Struthio remains possibly belong to this taxon. The Vallesian taxon S. orlovi was erected on some postcranial remains, including a small distal fragment of a tarsometatarsus (no 6-4) from the Moldavian locality of Varnitsa; the specimen preserves only the tIII (Kurocˇkin and Lungu, 1970: figs. 2, 5; Mlı´kovsky´, 2002). S. orlovi is described as less massive than the modern ostrich and S. brachydactylus (Kurocˇkin and Lungu, 1970).

Table 1 Values of the angle aˆ (Fig. 2(a)) in the tarsometatarsus of modern and fossil ostriches. aˆ NIK-917 Struthio camelus NMNHS-1/2001 NMNHS-3/2003 NMNHS-4/2006 NMNHS-5/2006 NMNHS-6/2006 S. cf. karatheodoris, Bulgaria S. novorossicus, Ukraine S. brachydactylus, Ukraine Struthio sp., Ukraine S. coppensi, Namibia

208 188 208 168 158 148 128 288–338 228 258–288 298

33

The single available measurement for the tarsometatarsus of S. orlovi coincides that of S. brachydactylus; it is very close to that of the modern ostriches (Fig. 4(b)), agreeing with Mourer-Chauvire´ et al. (1996a). The studied tarsometatarsus lacks the corresponding measurement, making its comparison with S. orlovi impossible. The species S. novorossicus was originally described on some material found in the locality of Novo-Elisavetovka, Ukraine (Alexejew, 1915; Mlı´kovsky´, 2002). Among the type material, there are two distal fragments of the tarsometatarsus (no 1560 and 1561; Alexejew, 1915: figs. 55, 56). NIK-917 is well separated from S. novorossicus, having smaller size (Fig. 4(b)). Among the late Miocene ostriches, the tarsometatarsus of S. novorossicus is the largest one (Fig. 4(b)). Besides the smaller size of NIK-917, its comparison with the illustrations of Alexejew (1915) indicates that its tIV is relatively weaker and the angle aˆ smaller than that of S. novorossicus (Table 1). The latter difference suggests higher ‘‘degree of didactyly’’ in the Ukrainian fossil ostrich. A similar difference of S. novorossicus is also referred for S. cf. karatheodoris from Kalimantsi, Bulgaria (Boev and Spassov, 2009). The type material of S. brachydactylus has been discovered in Grebeniki, Ukraine; it consists of several remains, including a tarsometatarsus (no 408/359; Burcˇak-Abramovich, 1949: fig. 3; 1953: table X, figs. 1, 2; table II, figs. 1–4; Mlı´kovsky´, 2002). The tarsometatarsus of S. brachydactylus is robust, like the studied one, and relatively short, associated by robust posterior phalanges (Boev and Spassov, 2009). The latter authors also note that it is slightly smaller or similar in size to the modern ostrich; NIK-917 has also a size similar to S. brachydactylus and the modern ostrich (Fig. 4(b)). Comparison of NIK-917 with S. brachydactylus indicates that they have similar morphology; the preserved proximal epiphysis (comparison based on the illustrations of BurcˇakAbramovich, 1953) of S. brachydactylus is also similar to that of NIK-917. The angle aˆ of S. brachydactylus is very close to that of NIK-917 and modern ostriches, indicating a similar ‘‘degree of didactyly’’ (Table 1). Burcˇak-Abramovich (1953) described some remains of an ostrich from Odessa catacombs as Struthio sp. Its tarsometatarsus differs from NIK-917 in having a larger size (Fig. 4(b)), a more robust distal epiphysis and a higher ‘‘degree of didactyly’’; the angle aˆ varies from 258 to 288, being larger than that of NIK-917 (Table 1). A didactyl ratite, named S. coppensi is known from the lower Miocene of Namibia. Among the type material, there is a distal part of tarsometatarsus, no EF 30 94 (Mourer-Chauvire´ et al., 1996a). Its smaller size (Fig. 4(b)) and higher ‘‘degree of didactyly’’ (aˆ = 298; Table 1) clearly distinguish it from NIK-917. As mentioned above, S. karatheodoris is a late Miocene ostrich known from Greece and possibly Bulgaria. The known type material for this taxon (from ‘‘Adriano’’, Samos) lacks remains of tarsometatarsus, preventing any direct comparison with NIK-917. Taking in mind that one of the main characters of S. karatheodoris is its size, larger than the modern ostrich (Forsyth Major, 1888; Martin, 1903; Bachmayer and Zapfe, 1962), the robust NIK-917 might belong to this species (Fig. 4). The distal fragment of a tarsometatarsus (NMNHS-16371) from the Turolian locality of Kalimantsi, Bulgaria, recently described as S. cf. karatheodoris (Boev and Spassov, 2009), is comparable in size to that of NIK-917 and S. brachydactylus (Fig. 4(b)). Its morphology is similar to NIK-917, but with a lesser ‘‘degree of didactyly’’ (aˆ = 128; Table 1); the latter could be due to a small deformation of the bones, especially in NIK917. The most important difference between S. brachydactylus and S. cf. karatheodoris from Bulgaria is its larger minimal DAP of the tIII distal trochlea (meas. M16; figs. 2, 4(b)). NIK-917 lacks this measurement as the bone is badly preserved in that part (Fig. 3). We note, however, that the Bulgarian specimen also seems to be deformed in that part (Boev and Spassov, 2009: fig. 1).

34

G.D. Koufos et al. / Geobios 49 (2016) 29–36

Fig. 4. Simpson diagram of Log-ratios based on the dimensions of the NIK tarsometatarsus with modern ostriches (a) and fossil Struthio from various localities (b). Reference: S. camelus, NMNHS-1/2001. Data source: S. camelus: orig. meas. taken by G.E.K.; S. coppensi: Mourer-Chauvire´ et al. (1996a); other taxa: Boev and Spassov (2009).

Overall, the size of NIK-917, larger than the modern ostrich; its resemblance to S. cf. karatheodoris from Bulgaria; its differences from all other fossil taxa; and its geographic position indicate that it may be assigned to S. karatheodoris. On the other hand, the absence of tarsometatarsus from the type locality – or femur from NIK –, preventing any direct comparison with the type material, makes the specific determination delicate. Therefore, waiting for additional material it is currently safer to refer the NIK ostrich as S. cf. karatheodoris.

5. Biogeographic and palaeoecological remarks Although widely distributed in the Old World, fossil ostriches, especially during the Neogene period, are poorly represented in the fossil record (i.e., very low number of specimens), making their current taxonomy doubtful and any attempt for tracing their origin, historical dispersal patterns, and relations between African and Eurasian taxa quite difficult. Given the oldest known occurrence of a definite ostrich, S. coppensi from the lower

G.D. Koufos et al. / Geobios 49 (2016) 29–36

Miocene of Namibia (Mourer-Chauvire´ et al., 1996a), an African origin is suggested by most palaeo-ornithologists (MourerChauvire´ et al., 1996b). Earliest Eurasian evidence of Struthio sp. comes from the early middle Miocene site of C¸andır in Turkey (Sauer, 1979), implying possibly that ostriches invaded Eurasia at the same time (end Burdigalian; Karpatian) and under the same circumstances as hominoid primates (e.g., Begun et al., 2003). However, whether this first ostrich invasion gave rise to all, or only some of the later Eurasian Struthio remains unknown, as the next occurrence datum of the genus comes from the Moldavian locality of Varnitsa, of roughly Vallesian age (Sen, 1997; Vangengeim et al., 2006; Vangengeim and Tesakov, 2008), delineating a > 6-ma gap in the fossil record. Mourer-Chauvire´ et al. (1996b) proposed a two-stage colonization scheme for Eurasia: a first, middle Miocene one evidenced by an ‘‘aepyornithoid’’ eggshell pattern (found up to the Pliocene), and a second, upper Miocene phase directly linked to Struthio and related to a ‘‘struthious’’ eggshell pattern. This scenario fits well with the ‘‘aepyornithoid’’ type of eggshells found at C¸andır (Sauer, 1979). During the upper Miocene to Pliocene, ostriches greatly expanded in Eurasia from southeastern Europe to China (Mlı´kovsky´, 2002; Boev and Spassov, 2009: fig. 3). Some ‘‘aepyornithoid’’-type eggshells have been found in the lower Pliocene of the Teruel Basin, in Spain (Mein and Dauphin, 1995), indicating that Struthionidae have reached western Europe at the beginning of Pliocene, at least. Albeit there is no taxonomic consensus on late Miocene species (e.g., Mlı´kovsky´, 2002; Boev and Spassov, 2009), mostly due to the absence of directly comparable anatomical parts, the southeastern European fossil record provides sufficient evidence for the co-occurrence of two lineages, at least. Hence, S. brachydactylus coexists at the Ukrainian locality Grebeniki (MN11) with the larger species Palaeostruthio sternatus Burcˇak-Abramovich, 1939, that, according to Boev and Spassov (2009), would represent S. novorossicus, originally known from the slightly younger (MN12) site of Novo-Elisavetovka, again in Ukraine. The latter lineage retains during the late Miocene a much higher ‘‘degree of didactyly’’ (aˆ = 28–338) than the former (aˆ = 228), which already is closer to S. camelus (aˆ = 14–208; Table 1). The older and much more primitive, but again poorly known S. orlovi from Varnitsa, Moldova, may stand as a possible ancestor of both lineages, though Kurocˇkin and Lungu (1970) stressed out important morphological differences with S. brachydactylus that rather preclude direct relationships. Additionally, Sauer (1979) recognized in the earlier C¸andır species close morphometrical resemblance with S. brachydactylus, which would suggest that this lineage may have a different phylogeographic origin. Standing in contrast with previous works, we recommend synonymy between S. brachydactylus and S. karatheodoris, the latter name having date priority. Among criteria commonly used in the distinction of these two species are size difference (Sauer, 1979) and geographic provenance (Mlı´kovsky´, 2002). Though there are no comparable anatomical elements among the Samos (Martin, 1903) and Grebeniki (Burcˇak-Abramovich, 1953) samples, the study of the Nikiti 2 tarsometatarsus shows great morphological and dimensional similarities with S. brachydactylus. The distal tibiotarsus and the pedal phalanx 1/III from Pikermi (Bachmayer and Zapfe, 1962), as well as the distal tarsometatarsus from Kalimanci (Boev and Spassov, 2009) also resemble S. brachydactylus. On the other hand, the alleged size difference between these two species (Sauer, 1979; Boev and Spassov, 2009) actually does not exceed those seen among S. camelus individuals of different sex and/or ontogenetic age (data in Sauer, 1979). Deeming et al. (1996: table 1) also provide

35

data indicating that size variability in tibiotarsal and tarsometatarsal length of adult S. camelus may reach 30%. Differences in the angle aˆ (Fig. 2(a)) between the Kalimantci tarsometatarsal and S. brachydactylus or NIK-917 are also within the range seen in S. camelus. As Grebeniki and Nikiti 2 are older than Samos, Kalimanci, Pikermi and Maragheh, we may assume a late Miocene shift toward more cursorial adaptations within this (single) taxon that would allow for more slender and longer phalanges (and distal leg in general; compare femur proportions and size in Burcˇak-Abramovich and Vekua, 1990) through time, in analogy with several ungulate mammals of the same time interval. Modern ostriches are rather typical elements of arid habitats such as the coastal areas of South Africa, desert grasslands, semiarid savannah or even true deserts. Open plains of low altitude with short-grass and semi-arid regions and deserts with annual grasses are preferred contra woodlands and forested areas, as well as regions with grass exceeding 1 m (Deeming, 1999; Verlinden and Masogo, 1997; Cooper et al., 2009). The palaeoenvironment of the late Miocene ostriches from southeastern Europe is hard to be traced based on available evidence but certainly does not fit with that of the living species. The C¸andır vertebrate fauna, showing the first known Eurasian occurrence of Struthio, indicates a mosaic environment of open woodlands close to the lake margin (Geraads et al., 2003). A forested character has been suggested for Varnitsa and contemporaneous fossil sites from north peri-Pontic region where the next occurrence of the genus is recorded (Vangengeim and Tesakov, 2008). During the Turolian, S. karatheodoris/S. brachydactylus shows a southern biogeographic expansion occupying territories, such as Southern Balkans-Asia Minor and Iran. Ostrich-bearing Greek and Bulgarian vertebrates faunas at that time have been proved to represent a less forested and more arid or seasonal condition, generally corresponding to marching open bushlands (Bonis et al., 1992; Koufos, 2006; Kostopoulos, 2009; Merceron et al., 2006, 2007; Boev and Spassov, 2009). All these data favor a different palaeoecological profile for the Miocene southeastern European ostriches when compared to the living ones and may suggest that grazing habit of living Struthio could be a new achievement gained much more recently (e.g., after ca. 4 Ma when the main split within the living species occured; Freitag and Robinson, 1993). The phylogenetically distinct Somali ostrich (S. molybdophanes Reichenow, 1883, regarded as a subspecies by some authors) is, however, a bushland browser (Freitag and Robinson, 1993) and may serve as a better ecological analogue for the Neogene fossil taxa. The difference in taxonomic diversity between late Miocene north peri-Pontic and living taxa is also worth noting and possibly related to the previous discussion. Thus, during approximately the same timespan (i.e., roughly 4 Ma, the Somali taxon being excluded), living Struthio shows great stability in morphological structure and ecological preferences, whereas three distinct fossil lineages at least appear in a region representing less than half of the smallest distribution area of an extant race (e.g., S. c. massaicus Neumann, 1888, the Masai ostrich). This would be indicative of a high ecological pressure due to the fractioned (i.e., mosaic) upper Miocene palaeoenvironment of the north peri-Pontic region, that may allowed one of the lineages (S. karatheodoris/S. brachydactylus) to adopt a more ‘‘open’’ ecological profile and occupy southern territories while the other (S. novorossicus) adapted to the local conditions. Within this context, it would be useful to explore and test the morphometrical fit of the first lineage to the Pliocene Struthio sp. from C¸alta (Janoo and Sen, 1998) and the later Indian S. asiaticus, and those of the second lineage to the Dmanisi taxon (Burcˇak-Abramovich and Vekua, 1990) and S. pannonicus Kretzoi, 1953, but such comparisons clearly fall beyond the scope of the present study.

36

G.D. Koufos et al. / Geobios 49 (2016) 29–36

6. Conclusions The late Miocene ostrich remains from Greece are known by limited bone fragments from the localities of ‘‘Adriano’’ from Samos and Pikermi, both samples attributed to S. karatheodoris. The morphological and metrical features of the studied tarsometatarsus from Nikiti 2 indicate a species slightly larger than modern ostrich and similar to S. cf. karatheodoris from Bulgaria but also to S. brachydactylus from Grebeniki. The absence of tarsometatarsus from Samos, the type locality of S. karatheodoris, allows us referring the Nikiti 2 species to as S. cf. karatheodoris; we point out, however, that the north peri-Pontic S. brachydactylus would most likely represent a junior synonym of the Greek and Bulgarian species S. karatheodoris. We suggest that the ecological profile and palaeoenvironment of late Miocene southeastern European ostriches do not fit with that of the living African species, except perhaps that of the Somali ostrich. Based on available evidences, we also assume that S. karatheodoris/S. brachydactylus represents a distinct lineage that possibly originated during middle Miocene times in Anatolia; during the upper Miocene, it gradually adapted to more open conditions, expanding from the North peri-Pontic region to the south. Acknowledgements Thanks to Dr. Z. Boev for giving us access to the modern material of ostriches housed in the Natural History Museum of Sofia. Many thanks to C. Mourer-Chauvire´ and two anonymous reviewers for their constructive comments, which greatly improved this article.

References Alexejew, A.K., 1915. [Animaux fossiles du village Novo-Elisavetovka]. Zapiski Novorosiyskogo institute, Tekhnik publishing house, Odessa 9, 273–453 (in Russian). Bachmayer, F., Zapfe, H., 1962. Reste von Struthio aus Pikermi. Proceedings of the Academy of Athens 37, 247–253. Begun, R.D., Gu¨lec¸, E., Geraads, D., 2003. Dispersal patterns of Eurasian hominoids: implications from Turkey. Deinsea 10, 23–49. Boev, Z., Koufos, G.D., 2000. Presence of Pavo bravardi (Gervais, 1849) (Aves, Phasianidae) in the Ruscinian locality of Megalo Emvolon, Macedonia, Greece. Geologica Balcanica 30 (1–2), 69–74. Boev, Z.N., Koufos, G.D., 2006. Aves. The Late Miocene Vertebrate locality of Perivolaki, Thessaly, Greece. In: Koufos, G.D. (Ed.), Palaeontographica, Abt. A 276, 11–22. Boev, Z., Spassov, N., 2009. First record of ostriches (Aves, Struthioniformes, Struthionidae) from the Late Miocene of Bulgaria with taxonomic and zoogeographic discussion. Geodiversitas 31 (3), 493–507. Bonis, L.de, Bouvrain, G., Geraads, D., Koufos, G.D., 1992. Diversity and palaeoecology of Greek Late Miocene mammalian faunas. Palaeogeography, Palaeoclimatology, Palaeoecology 91, 99–121. Burcˇak-Abramovich, N.I., 1949. [Fossil ostrich (Struthio brachydactylus sp. nov.) from the Maeothian deposits of the Grebenniki village, South Ukraine] Dokadi Akademii nauk SSSR 67 (1), 141–144 (in Russian). Burcˇak-Abramovich, N.I., 1953. [Fossil ostriches of Caucasus and Southern Ukraine]. Trudy Estestvenoritoricheskogo muzeya imeni G. Zardabi 7, 1–206 (in Russian). Burcˇak-Abramovich, N.I., Vekua, A.K., 1990. The fossil ostrich Struthio dmanisensis sp. n. from the Lower Pleistocene of eastern Georgia. Acta Zoologica Cracoviensia 33, 121–132. Cooper, R.G., Mahrose, K.M.A., Horban´czuk, J.O., Villegas-Vizcaı´no, R., Kennou Sebei, S., Mohammed, A.E., 2009. The wild ostrich (Struthio camelus): a review. Tropical Animal Health Production 41, 1669–1678. Deeming, D.C. (Ed.), 1999. Introduction. The Ostrich: Biology, Production and Health. Oxon, Wallingford. Deeming, D.C., Sibly, R.M., Magole, I.L., 1996. Estimation of the weight and body condition of ostriches (Struthio camelus) from body measurements. Veterinary Record 139, 210–213. Forsyth Major, C.J., 1888. Sur un gisement d’ossements fossiles dans l’ile de Samos, contemporains de l’aˆge de Pikermi. Comptes rendus de l’Acade´mie des sciences Paris 107, 1178–1181.

Forsyth Major, C.J., 1894. Le gisement ossife`re de Mytilinii et catalogue d’ossements fossiles recueillis a` Mytilini, ˆıle de Samos, et de´pose´s au Colle`ge Galliard a` Lausanne. Georges Bridel & Cie, Lausanne. Freitag, S., Robinson, T.J., 1993. Phylogeographic patterns in mitochondrial DNA of the ostrich (Struthio camelus). Auk 10, 614–622. Gaudry, A., 1862–1867. Animaux fossiles et ge´ologie de l’Attique. Savy, Paris. Geraads, D., Begun, D., Gu¨lec¸, E., 2003. The middle Miocene hominoid site of C¸andır, Turkey: general palaeoecological conclusions from the mammalian fauna. Geology and Vertebrate Palaeontology of the Middle Miocene hominoid locality C¸andır (Central Anatolia, Turkey). In: Gu¨lec¸, E., Begun, D., Geraads, D. (Eds.), Courier Forschung-Institut Senckenberg 240, 241–250. Janoo, A., Sen, S., 1998. Pliocene vertebrate locality of C¸alta, Ankara, Turkey. 2. Aves: Struthionidae. Geodiversitas 20 (3), 339–351. Kostopoulos, D.S., 2009. The Pikermi Event: temporal and spatial resolution of the Turolian large mammal fauna in SE Europe. Palaeogeography, Palaeoclimatology, Palaeoecology 274, 82–95. Kostopoulos, D.S., Sen, S., Koufos, G.D., 2003. Magnetostratigraphy and revised chronology of the Late Miocene mammal localities of Samos, Greece. International Journal of Earth Sciences 92, 779–794. Kostopoulos, D.S., Koufos, G.D., Sylvestrou, I.A., Syrides, G.E., Tsompachidou, E., 2009. Lithostratigraphy and fossiliferous sites. The Late Miocene Mammal Faunas of the Mytilinii Basin, Samos Island, Greece: New Collection. In: Koufos, G.D., Nagel, D. (Eds.), Beitra¨ge zur Pala¨ontologie 31, 13–26. Koufos, G.D., 2006. Palaeoecology and chronology of the Vallesian (Late Miocene) in the Eastern Mediterranean region. Palaeogeography, Palaeoclimatology, Palaeoecology 234, 127–145. Koufos, G.D., 2016. History, stratigraphy and fossiliferous sites. Palaeontology of the upper Miocene vertebrate localities of Nikiti (Chalkidiki Peninsula, Macedonia, Greece). In: Koufos, G.D., Nagel, D. (Eds.), Geobios 49 (1-2) , http://dx.doi.org/ 10.1016/j.geobios.2016.01.007 (this volume). Koufos, G.D., Kostopoulos, D.S., Vlachou, T.D., 2009. Chronology. Late Miocene Mammal Faunas of the Mytilinii Basin, Samos Island, Greece: New Collection. In: Koufos, G.D., Kostopoulos, D.S. (Eds.), Beitra¨ge zur Pala¨ontologie 31, 397– 408. Koufos, G.D., Kostopoulos, D.S., Vlachou, T.D., Konidaris, G.E., 2016. Synthesis. Palaeontology of the upper Miocene vertebrate localities of Nikiti (Chalkidiki Peninsula, Macedonia, Greece). In: Koufos, G.D., Kostopoulos, D.S. (Eds.), Geobios 49 (1-2) (this volume). Kurocˇkin, E., Lungu, A., 1970. [A new ostrich from the Middle Sarmatian of Moldavia]. Paleontologicheskiy zhurnal 1, 118–123 (in Russian). Lydekker, R., 1891. Catalogue of the fossil birds in the British Museum (Natural History). British Museum (Natural History), London368p. Martin, R., 1903. Note on some remains of Struthio karatheodoris Major on the Island of Samos. Proceedings of the Zoological Society of London 1903 (1), 203–210. Mein, P., Dauphin, Y., 1995. Des coquilles d’œufs de type Aepyornis dans le Bassin de Teruel (Plioce`ne basal, Espagne). Neues Jahrbuch fur Geologie und Pala¨ontologie, Monatschefte 1995, 182–191. Merceron, G., Blondel, C., Viriot, L., Koufos, G.D., Bonis, L.de, 2007. Dental microwear analysis of bovids from the Vallesian (Late Miocene) of Axios valley in Greece: reconstruction of the habitat of Ouranopithecus macedoniensis (Primates, Hominoidea). Geodiversitas 29 (3), 421–433. Merceron, G., Zazzo, A., Spassov, N., Geraads, D., Kovachev, D., 2006. Bovid paleoecology and paleoenvironments from the Late Miocene of Bulgaria: evidence from dental microwear and stable isotopes. Palaeogeography, Palaeoclimatology, Palaeoecology 241, 637–654. Michailidis, D., Roussiakis, S., Theodorou, G., 2010. Palaeoavian remains from the Late Miocene localities of Pikermi, Chomateri and Kerassia; palaeoecological implications. Geologica Balcanica 39 (1–2), 250–251. Mlı´kovsky´, J., 1996. Tertiary birds of Greece. Acta Universitatis Carolinae Geologica 39, 649–653. Mlı´kovsky´, J., 2002. Cenozoic Birds of the World. Part 1: Europe. Ninox Press, Praha. Mourer-Chauvire´, C., Senut, B., Pickford, M., Mein, P., 1996a. Le plus ancien repre´sentant du genre Struthio (Aves, Struthionidae), Struthio coppensi n. sp., du Mioce`ne infe´rieur de Namibie. Comptes rendus de l’Acade´mie des sciences, Paris 322, II a, 325-332. Mourer-Chauvire´, C., Senut, B., Pickford, M., Mein, P., Dauphin, Y., 1996b. Ostrich legs, eggs and phylogenies. South African Journal of Sciences 92, 492–495. Sauer, F.G.E., 1979. A Miocene ostrich from Anatolia. Ibis 121, 494–501. Sen, S., 1997. Magnetostratigraphic calibration of the European Neogene Mammal Chronology. Palaeogeogaphy, Palaeoclimatology, Palaeoecology 133, 181–204. Vangengeim, E.A., Tesakov, A.S., 2008. Late Sarmatian mammal localities of the eastern Paratethys: stratigraphic position, magnetochronology, and correlation with the European Continental Scale. Stratigrafiya Geologicheskaya Korrelyatsiya 16, 95–107. Vangengeim, E.A., Lungu, A.N., Tesakov, A.S., 2006. Age of the Vallesian Lower Boundary (Continental Miocene of Europe). Stratigrafiya Geologicheskaya Korrelyatsiya 14, 81–93. Verlinden, A., Masogo, R., 1997. Satellite remote sensing of habitat suitability for ungulates and ostrich in the Kalahari of Botswana. Journal of Arid Environments 35 (3), 563–574. Zelenkov, N., Boev, Z., Lazarides, G., 2015. A large ergilornithine (Aves, Gruiformes) from the Late Miocene of the Balkan Peninsula. Palaeontologische Zeitschrift, http://dx.doi.org/10.1007/s12542-015-0279-z (in press).