1 Introduction

The Tamanian faunal complex was firstly established by Gromov (1948) based on the presence of large mammals as one of the biochronological units of the East European Pleistocene (Vislobokova and Titov 2020). After the description of the type site, which is Sinyaya Balka on the Tamanian Peninsula in Russia (Gromov 1948), a number of new mammal localities of this complex have been discovered there and within the north-eastern coastal area of the Sea of Azov in Russia, Azerbaijan and Georgia (Vereshchagin, 1957, 1959; Dubrovo 1963; Alekseeva 1977; Vangengeim et al. 1991; Baigusheva 2000; Tesakov 2004; Dodonov et al. 2006; Tesakov et al. 2007, 2013; Baigusheva and Titov 2008; Bukhsianidze et al. 2014; Vislobokova and Titov 2020).

Based on biostratigraphic and paleomagnetic data, the time range of the Tamanian complex sense strict was estimated as ca. 1.2–0.87 Ma (MIS 36 to the MIS 22–21 boundary, equivalent of Western European Epivillafranchian). It approximately corresponds to the interval between the Cobb Mountain paleomagnetic Subchron and the Brunhes/Matuyama Chron boundary (Markova 2007; Vislobokova and Tesakov 2013; Markova and Vislobokova 2016; Vislobokova and Titov 2020). The key species of the Tamanian complex is Mammuthus (Archidiskodon) meridionalis tamanensis Dubrovo, 1964. Other characteristic herbivores includes Equus aff. suessenbornensis Wüst, 1901, Stephanorhinus etruscus (Falconer 1868), and Elasmotherium caucasicum Borissiak, 1914 (Gromov 1948; Vereshchagin 1957; Dubrovo and Alekseev 1964; Alekseeva 1977; Vislobokova and Titov 2020). Among carnivores, the most important taxa include Canis mosbachensis tamanensis (Vereshchagin 1957), Lycaon lycaonoides (Kretzoi 1938), Lutra simplicidens Thenius, 1965, Homotherium latidens (Owen 1846), Acinonyx pardinensis Croizet and Jobert, 1828, and Pachycrocuta brevirostris Gervais, 1850 (Sotnikova et al. 2002; Titov 2008; Sotnikova and Titov 2008).

Stone artefacts from Sinyaya Balka document one of the oldest Palaeolithic human occurrences in Europe (Shchelinsky et al. 2010). In this context, the study of Tamanian faunas is of particular interest for the reconstruction of the ancient human settlement and environment at that time (Vislobokova and Titov 2020). This task requires a revision of faunal lists from particular sites previously assigned to the Tamanian complex, especially the composition of large mammals, as possible resource and competitor species for the Palaeolithic humans (Vislobokova and Titov 2020).

Although different carnivore species for the Tamanian faunal complex from Ukraine were listed in a number of publications (e.g., Topachevsky 1965; Vereshchagin et al. 1971; Kahlke 1975), they have not been properly revised. In this respect, old collections and newly collected materials require a substantial revision and analysis. The main purpose of this article is the re-study of hyena material from Nogaisk stored in the National Museum of Natural History in Kyiv and Zaporizhzhia Regional Local History Museum. Several skull fragments were found on the coast of the Sea of Azov in 1979. Recently it was possible to combine all remains for our study. The individual from Nogaisk locality provided the first occurrence of P. brevirostris in the modern territory of Ukraine. Additionally, the study provides a new evidence on morphology and evolution of this species within Eurasia.

2 Locality and stratigraphy

A few high coastal sections along the Taganrog Gulf shore expose subaqual sediments such as fluviatile deposits of ancient Don River origin and estuarine deposits of the Sea of Azov. These sediments are overlain by subaerial deposits that primarily belong to the loess-paleosol formation. The underlying subaqual deposits are arranged in three terraces, with the earliest one, the Nogaisk terrace yielding fossil remains of mammals of the Tamanian faunal assemblage. The two others, the Platovo and Voznesenskaya terraces, contain a fauna corresponding to the Middle Pleistocene (Velichko et al. 2010).

The Nogaisk locality is located on the northern coast of the Sea of Azov near Primorsk village (Zaporizhzhia Oblast, Ukraine), 2.5–3 km to the east from Obitochna foreland. The association of small and large mammals from this locality is characterised mostly by ancient forms, among which only a few were present in the Middle Pleistocene. As for the species list, a significant bias was established for this locality. Large animals are very rarely found, and their remains are usually fragmentary and in bad condition, whereas small mammal remains are quite abundant and well preserved. Fossils were found in a complex of abandoned channel and lacustrine sediments. Two layers, sometimes described as two sites (Nogaisk 1 and 2) were distinguished, but contained the same fauna and were contemporary (Rekovets and Nadachowski 1996; Kowalski 2001). Deposits consisted mainly of layered, partly ferrugineol clayey sand and gravel with small carbonate concretions and/or sandstone pebbles (Kovalchuk et al. 2017).

The mammal assemblage includes 33 species: Erinaceus sp.; Desmana nogaica Topachevsky and Pashkov, 1990; Spermophilus nogaici (Topachevsky 1957); Trogontherium cuvieri Fischer von Waldheim, 1809; Castor fiber Linnaeus, 1758; Sicista vinogradovi Topachevsky 1965; Sicista sp.; Borsodia newtoni (Forsyth Major, 1902); Clethrionomys kretzoii (Kowalski, 1958); Allocricetus ehiki Schaub, 1930; Cricetus cricetus (Linnaeus 1758); Ellobius palaeotalpinus Schevtschenko, 1965; Lagurodon arankae (Kretzoi 1942); Mimomys rex Kormos, 1934; Mimomys reidi Hinton, 1910; Mimomys savini Hinton, 1910; Plioscirtopoda stepanovi Schevchenko, 1965; Prolagurus pannonicus Kormos 1930; Pygerethmus pumilio (Kerr, 1792); Spalax minor Topachevsky 1965; Lepus sp.; Ochotona sp.; Canis sp.; Mustela stromeri Kormos, 1934; Mustela palerminea (Petenyi 1864); Mustela praenivalis Kormos, 1934; Pachycrocuta brevirostris (Gervais 1850); Mammuthus meridionalis (Nesti 1825); Equus aff. suessenbornensis Wüst, 1901; Elasmotherium caucasicum Borissiak, 1914; Megaloceros sp.; Bison sp.; Pontoceros ambiguus Vereshchagin et al. 1971 (Topachevsky 1957; Garutt 1954; Kretzoi 1965; Topachevsky 1965; Vereshchagin et al. 1971; Kahlke 1975; Garutt and Foronova 1976; Dubrovo and Kapelist 1979; Topachevsky et al. 1987; Garutt et al. 1990; Rekovets 1994; Rekovets and Nadachowski 1996; Kowalski 2001; Velichko et al. 2010; Popescu 2011; Kolfschoten and Markova 2005; Rekovets et al. 2009; Maschenko et al. 2011; Larramendi 2016; Vislobokova and Titov 2020).

Rekovets and Nadachowski (1996) noted that the predominant species from Nogaisk, constituting up to 60% of the micromammal assemblage, were representatives of genera Lagurodon, Allophaiomys and Spermophilus. Other taxa belonging to Ellobius (common), Spalacidae or Cricetidae (rarer) retained ancestral, morphological characters. Allactaginae were diverse, but not very common, also as Spalax and Prolagurus forms (Rekovets and Nadachowski 1996). Lagurid molars with a “praepannonicus” morphology (broadly confluent triangles T4–T5) prevail in earlier faunas. The mean value of the ratio of anteroconid complex length to the total length of the first lower molar from Nogaisk equals 47.7, which suggests an older age than e.g. Zapadnye Kairy (mean value of this index for m1 of Prolagurus is 50.0) (Rekovets 1994; Kolfschoten and Markova 2005). The Nogaisk micromammal assemblage is characterised by the appearance of Prolagurus pannonicus and progressiveness of Allophaiomys pliocaenicus, which is recorded in the change of molar morpho structure pattern. The main indexes of this evolutional level are as follows: anteroconid complex to the total tooth length is 42–44, and the SDQ (Schmelzband-Differenzierung-Quotient, molar enamel thickness) is 90–100. The evolutionary stage of Prolagurus pannonicus resembles forms from the Tamanian faunal assemblage. Primitive morphotypes of Mimomys also indicate the similar age (Velichko et al. 2010). Based on all these data, the fauna from Nogaisk is dated to the late Early Pleistocene, between ca. 1.2–1.1 Ma. It is correlated with an early Tamanian faunal complex (Iosifova and Agadjanian 2010; Krokhmal 2010; Velichko et al. 2010; Kovalchuk 2017).

3 Material and methods

The identification of P. brevirostris was performed using basic morphometric analysis. Measurements were taken point to point, with an electronic calliper to the nearest 0.1 mm. Each value given here is the mean of three measurements (all in mm). The mean is rounded to one decimal digit. Measuring schemes and morphological terminology were taken and modified from Barycka (2008). Throughout the text, upper teeth are referred to using capital letters (e.g. P4), and lower teeth with lowercase letters (e.g. p4).

The material comprises a few skull fragments of a very large hyena, and all listed remains belong to a single adult individual with slightly worn teeth (Fig. 1). The specimen NMNHU-P 27–1689 is a fragmentary right half of the skull, including a part of the maxilla with I2 and P2-M1, zygomatic arch, palatine and broken basicranium. It is stored in the collection of the Department of Vertebrate Palaeontology (National Museum of Natural History, National Academy of Sciences of Ukraine). The second skull fragment is in the collection of Zaporizhzhia Regional Local History Museum under collection no. P-446/KV-23356. It is the right premaxilla with alveoli of I1 and I3 and including P2 and a part of maxilla with alveoli of P1–P2.

Fig. 1
figure 1

Cranial fragment of Pachycrocuta brevirostris from Nogaisk (Ukraine) NMNHU-P 27–1689 in buccal view (top) and occlusal view (bottom, scale bar 5 cm). Photo by Y. Semenov

Full size image

4 Systematic palaeontology

4.1 Description

The studied skull was previously classified as Hyaena sp. (Topachevsky 1965). The palatine fissure is placed approximately at an angle of 17° to the sagittal line of the skull (Fig. 1). The major palatine foramen is located in the middle part of maxillary palatine process, at the level of the P3 anterior root. The anterior border of the incisura palatine posterior is placed 23 mm behind the line connecting the caudal ends of P4. The length from prosthion to posterior border of hard palate is 173 mm, width at the canines is ca. 80 mm, and ca. 150 mm between the labial surfaces of P4. The zygomatic arch is very robust: the depth of the zygomatic bone anterior to its frontal process is 50 mm, while the maximum depth of the zygomatic process of the temporal bone is 58 mm. The tube of the external auditory meatus is moderately long and pressed closely against the caudal surface of the postglenoid process.

The tooth row is moderately curved, and the incisors are separated from the canine by a relatively long diastema. Shorter diastemas are present between C1 and P1 and P1 and P2, while P2-M1 are tightly appressed. All teeth are large and moderately worn (Fig. 1, Table 1). The I1 alveolus is large and narrow, and strongly flattened laterally. The preserved I2 is a large, strongly laterally flattened tooth, with a strong main cusp with its tip directed distally. Distal to the main cusp, on the distal edge, two smaller, symmetrically positioned cusplets are present. Wear produces an even plane across the apex. Judging from the I3 alveolus, this incisor was canine-like and much larger than the other two incisors. The oval shape of the alveolus shows that C1 was very large, massive and appreciably flattened. The oval P1 alveolus shows a small, reduced tooth, shifted medially in the tooth row. The two-rooted and rectangular P2 is a broad and low crowned tooth, with relatively short and low main cusp (paracone). The mesial and distal margins of the crown are blunt, while the buccal and lingual margins are more or less straight, and slightly curved only in the distal part. Only a delicate concavity of the buccal and lingual sides occurs in the middle part of the crown, in the transition between the mesial and distal portions of the tooth; the transition is slightly marked. The small and low anterior accessory cusplet, the protocone, is shifted moderately mesio-lingually, and weakly associated with the main cusp. It is separated from the paracone by a deep, narrow, V-shaped valley. After the paracone, the posterior accessory cusplet, the hypocone, is placed almost at the midline of the tooth. It is larger than the protocone, but also low and separated from the paracone by a shallow and wide valley. The crocutoid P3 is a massive and moderately high, almost rectangular tooth. It possesses a short and high paracone, which occupies 2/3 the crown length. The mesial, distal and lingual margins of the crown are rounded, while the buccal margin is mostly straight. There are notable concavities in the median part of the crown, in the well-marked transition between the anterior and posterior parts of the tooth. A rudimentary, vestigial anterior cusplet, the protocone, is shifted mesio-lingually. A thin but well-defined ridge runs from its base to the apex of the paracone. The posterior accessory cusplet, the hypocone, is located just behind the main cusp; it is larger than the protocone, and differentiated from the paracone by a deep and V-shaped valley. A crescent-like, shallow, U-shaped valley is present in the disto-lingual part of the crown, between the paracone and the cingulum ridge. A posterior ridge, connecting the hypocone with the apex of the paracone, is sharply pointed and less visible than the anterior ridge. The entire crown is collared by a thick cingulum, which is weakly developed on the buccal side. There is a wide, shallow, U-shaped valley between the cingulum ridge and the paracone on the bucco-mesial and mesial side; it forms a large crescent in occlusal view.

Table 1 Measurements of Pachycrocuta brevirostris from Nogaisk (all in mm)
Full size table

The P4 has a triangular, moderately-developed, relatively narrow and low protocone with centrally positioned tip. Its mesial margin is rounded, and the cusp is situated at an angle of 80° to the rest of the tooth, and is somewhat forwardly sloped. In occlusal view, the tooth is elongated and narrow, with a convex buccal margin and abruptly ending crown narrowing distally. In buccal view, the tooth looks low. The parastyle is large, high and oval. It is well separated from the paracone by a deep and sharply pointed notch, and the constriction between the parastyle and the paracone is noticeable in occlusal view. The anterior wall of the parastyle bears a small but well-defined preparastyle, which forms a vertical ridge on the mesial side, and is connected with an anterior cingulum. The paracone is relatively low and short, while the metastyle is longer than the paracone. Paracone and metastyle form almost continuous cutting surface; there is only a shallow notch between them. The tooth displays a basal triangle where the strong mesialo-lingual keel joins the cingulum. Perpendicularly oriented to the rest teeth, two-rooted M1 is functionally retained and moderately reduced. Its oval trigon is collared with a high and thin cingulum wall, and the trigon base is developed into a broad and shallow surface. The talon is more reduced and narrower, and connected with the P4 metastyle.

4.2 Comparisons

The above described, partially preserved skull belonged to a very large hyena, and its fragmentary condition made taxonomic attribution difficult. However, thanks to mainly complete tooth row and moderately worn teeth, it is possible to provide a clear taxonomic affiliation. From this time interval (late Early Pleistocene) in Eastern Europe, the presence of three large hyenas should be taken acknowledged: Pachycrocuta brevirostris, Pliocrocuta perrieri (Croizet and Jobert 1828), and possibly Crocuta crocuta ssp. Perhaps, a fourth, large hyena species, Chasmaporthetes lunensis Del Campana 1914 was present there. However, C. lunensis vanished much earlier, with its last Eurasian appearance documented from the German site Schernfeld dated to 1.8–1.6 Ma (Qiu 1987; Turner et al. 2008; A.M. pers. obs.). In addition, the teeth of this species are distinctly smaller and narrower than those in the specimen from Nogaisk. As for P. perrieri, the examined individual from Nogaisk also cannot be attributed to this species. The size of Nogaisk hyena exceeds the values known for P. perrieri, where P3 length does not reach 26.0 mm, and P4 length does not exceed 38.0 mm (Schütt 1972; Kurtén 1972; Kurtén and Poulianos 1977, 1981).

Only a hypothetical presence of C. crocuta in Nogaisk could be proposed, even if this species appeared in European faunas much later, ca. 0.8 Ma in Trinchera Dolina (Madurell-Malapeira et al. 2014). P. brevirostris is regarded as the largest hyena ever existed, approaching the size of a female lion, Panthera leo (Linnaeus 1758), with an average weight of ca. 110–120 kg, and the maximal weight up to 180 kg (Turner and Antón 1996; Turner 2001; Palmqvist et al. 2011). This is twice the weight of the extant spotted hyena, and much larger than the Middle and Late Pleistocene Crocuta crocuta spelaea Goldfuss, 1823. Large teeth were often regarded in the past as a diagnostic feature of the large body size. However, contrary to the canines or the first lower molar (m1), the upper carnassials (P4) are not useful for determining the sexual dimorphism or body size. Their measurements are not correlated with those of the body. As remarked by Lundholm (1952): “…the animal may have been comparatively small, as compared with other members of the same population, however small individuals have had relatively bigger teeth than those in larger specimens”. Other possibly important factors that may affect the size of the teeth are climatic and geographical factors, sexual dimorphism and changes over time, which further complicates the interpretation of such data. Contrary to Ursidae, Felidae, and Mustelidae, and partially similarly to Canidae, Hyaenidae are virtually monomorphic in linear measurements, and sexual dimorphism can be cautiously rejected. Comparison of large subspecies of the cave hyena Crocuta crocuta praespelaea with P. brevirostris reveals that the latter possesses in average a larger and narrower P4 with proportionally short metastyle. Beside the size, the present teeth differ also in various diagnostic characters such as a relatively low P3 crown (more brachydont) with stronger hypocone and the development of a triangular surface between the paracone and the strong cingulum ridge on the medio-lingual side. The main cusp is proportionally shorter and lower. The P4 of the Nogaisk hyena differs from P4 of C. crocuta in the presence of well-developed preparastyle, larger parastyle, which is better separated from the paracone, broader blade and proportionally shorter metastyle. Less reduced M1, with a size (13.9 mm) exceeding for C. crocuta, in which this tooth is rarely larger than 8 mm. The C1 alveolus is still informative even if the tooth is missing. Its length (25.4 mm) and width (16.1) exceed the diameters obtained for the very large Middle Pleistocene cave hyenas from Mosbach 2 or Petralona cave (Schütt 1971; Kurtén 1972; Kurtén and Poulianos 1977, 1981). For these reasons, we attribute the analysed specimen from Nogaisk to P. brevirostris.

5 Discussion

Pachycrocuta brevirostris is first recorded in South Turkwel (northern Kenya, Africa) ca. 3.6–3.5 Ma (Werdelin 1999; Werdelin and Lewis 2000, 2005, 2008), Laetoli ca. 3.5 Ma (Leakey and Hay 1979; Arribas and Palmqvist 1999), West Turkana sites LO5 and LO10 (3.5–3 Ma), possibly Koobi Fora, Tulku Bor Mb (3.4–2.6 Ma). This species was also recorded in Makapansgat (ca. 3.1–3.0 Ma; Howell and petter 1979, 1980; Randall 1981; Arribas and Palmqvist 1998, 1999). The earliest Asian record from Nihowan is slightly younger and dated to ca. 2.0 Ma (Teilhard de Chardin and Piveteau 1930).

Their ancestors likely evolved from the late Miocene Hyaenictitherium (Howell and Petter 1980). P. brevirostris coexisted with Crocuta in Africa during more than 1 million years, and disappeared in East Africa in the Late Pliocene but dispersed and was still present in Southern Africa (Toerien 1952; Randall 1981; Werdelin 1999), where it was recorded among others in Sterkfontein 4–5 (2.6–2.4 Ma; Howell and Petter 1980; Turner 1987; Turner and Antón 1996), Bolt’s Farm (2.5–2.3 Ma; Mutter et al. 2001). P. brevirostris is still present in southern Africa during the middle and late Early Pleistocene, and their remains were found in Kromdraai A (1.6–1.5 Ma; Ewer 1954; Howell and Petter 1980; Turner 1987) and Swartkrans 3 (1.5–1.4 Ma; Mutter et al. 2001). The latest African record of the species is known from the Gladysvale cave (1.1–1.0 Ma; Mutter et al. 2001).

The Early Pleistocene records of this hyena from Central and Eastern Asia are common: (1) Pinjor Formation (2.5–2.4 Ma; Arribas and Palmqvist 1999); (2) Mohui (2.5–2.4 Ma; Louys 2014); (3) Djetis (2.0–1.9 Ma; Dubois 1908; Schütt 1972; Geraads 1979); (4) Tologoi (2.0–1.8 Ma; Turner and Antón 1996); (5) Volga river (2.0–1.5 Ma; Turner and Antón 1996); (6) Kopaly (1.6–1.5 Ma; Kojamkulova et al. 1987; Sotnikova et al. 1997; Vislobokova 2005); (7) Liucheng (1.2–1.0 Ma; Louys 2014); (8) Zasukhino (1.1–1.0 Ma; Sotnikova 1989; Turner and Antón 1996); (9) Gongwangling (1.1–1.0 Ma; Louys 2014); (10) Lakhuti (1.1–1.0 Ma; Sotnikova and Vislobokova 1990; Turner and Antón 1996); (11) Nalaikcha (1.1–1.0 Ma; Sotnikova 1989; Turner and Antón 1996). P. brevirostris has become quite rare in Asia already in the early Middle Pleistocene and present only in the south-eastern and eastern parts of the continent, being recorded for this period from Bukuran, Ngebung level B and Kedung Brubus, all dated on 0.9–0.8 Ma (Bouteaux et al. 2007; Bouteaux and Moigne 2010; Louys 2014). The species survived during the late Middle Pleistocene, with the latest occurrence at Zhoukoudian 1, dated on 0.5–0.4 Ma (Turner and Antón 1996; Boaz et al. 2000, 2004; Dennell et al. 2008).

During the Early Pleistocene, the species reached Europe (most probably from Asia) (Werdelin 1999; Palmqvist et al. 2011). In a short time, P. brevirostris replaced two other large hyaenid species, the large bone-crushing Pliocrocuta perrieri and the gracile pack hunter Chasmaporthetes lunensis (Turner and Antón 1996; Turner et al. 2008). It was present in Europe between 2.0 and 0.7 Ma, and finally disappeared during the early Middle Pleistocene (Fig. 2; Table 2; Turner and Antón 1996; Arribas and Palmqvist 1999; Turner et al. 2008; Palmqvist et al. 2011).

Fig. 2
figure 2

Distribution of Plio-Pleistocene sites with P. brevirostris in Europe and adjoining areas. For locality numbers – see Table 2

Full size image
Table 2 Occurrence of P. brevirostris in the Plio-Pleistocene of Europe and adjoining areas. Locality numbers (Map) correspond to those in Fig. 2
Full size table

The extinction of P. brevirostris in Europe was probably linked to the decline and subsequent extinction of saber-tooth cats, particularly Megantereon whitei and Homotherium crenatidens (Martínez-Navarro and Palmqvist 1996; Palmqvist et al. 1996a, b; Palmqvist and Arribas 2001a, b; Palmqvist et al. 2011). Disappearance of these cats implied the loss of an important source of partly consumed carcasses and thus a change in the interactions between flesh-eating and bone-crushing species of the carnivore guild.

P. brevirostris was a member of the large carnivore paleoguild comprising also Lycaon falconeri/lycaonoides, Canis etruscus/mosbachensis, Ursus etruscus/arctos, H. crenatidens/latidens, M. whitei, Panthera toscana/gombaszoegensis, Acinonyx pardinensis, Puma pardoides, and Lynx issiodorensis/pardinus. This carnivore assemblage existed in Eurasia for more than one million years (between 2.0 and 0.9–0.8 Ma), and its paleohierarchy was more or less stable. Thanks to its large size and social lifestyle, P. brevirostris was one of the dominant species (Turner and Antón 1996; Turner 2001; Palmqvist et al. 2011).

Nevertheless, the evolutionary history of this species was ended during the latest Early Pleistocene, where P. brevirostris was replaced by C. crocuta. The earliest European records of this species are known from Spanish sites at the Gran Dolina of Atapuerca (García 2003). In general opinion, Trinchera Dolina 4 (TD4) is dated to the Early-Middle Pleistocene boundary (ca. 0.78 Ma). However, new ESR dates suggest that the age of this horizon is closer to 0.9 Ma, whereas Trinchera Dolina 4 (TD4) is close to the Jaramillo subchron (Moreno-García 2011; Madurell-Malapeira et al. 2014). This new chronological background suggests co-existence in time of P. brevirostris and C. crocuta in the Iberian Peninsula during the latest Early Pleistocene (Madurell-Malapeira et al. 2014). This should not be surprising, since the co-existence of these two hyaenids is well documented for more than one million years in East Africa (Werdelin and Lewis 2005). Although it should be noted that the African C. crocuta that co-existed with P. brevirostris was a much smaller and gracile species than the European Middle Pleistocene form (Madurell-Malapeira et al. 2014).

It is interesting that P. brevirostris disappeared more or less simultaneously in most of Europe, and survived only in Western and Central Europe (Turner and Antón 1996; Turner 2001; Palmqvist et al. 2011). But even there sites younger than 0.7 Ma are unknown. The only exception is the Hungarian site Vertesszöllös 2, but the presence of this species is based on questionable remains (Jánossy 1990). Also, some admixture of older layers also cannot be excluded (Burdukiewicz pers. comm.). Numerous species, like L. lycaonoides or P. gombaszoegensis, which co-occurred with P. brevirostris, survived in Central and Eastern Europe longer, even until the late Middle Pleistocene (MIS 11–10) (Jánossy 1990; Baryshnikov 2011; Marciszak 2014). The exact answer for this difference is unknown and needs further research.

P. brevirostris disappeared relatively fast likely due to the competitive pressure from C. crocuta, but the exact reconstruction of this scenario is still unresolved. It is known that, despite its larger body size, P. brevirostris has proportionally smaller teeth than C. crocuta, which was not only better adapted for bone crushing (proportionally wider and larger P3 and p3), but also for slicing the meat. Upper carnassials (P4) of P. brevirostris had a proportionally shorter metastyle (cutting blade) and narrower crown, while its lower carnassials (m1) were smaller (compared to body size) with shorter major cusps (Turner and Antón 1996; Turner 2001; Barycka 2008). Both species are characterised by other adaptations to bone crushing like zigzag-like enamel structure (Stefen and Resenberger 2002, Barycka 2008). In aggressive encounters one to one, C. crocuta had likely no chance against P. brevirostris, but C. crocuta lived and hunted in larger clans than P. brevirostris (Turner and Antón 1996; Turner 2001; Barycka 2008; Palmqvist et al. 2011). These two reasons, better adaptation for crushing bones and meat slicing and more numerous clans, were probably the main factors responsible for the success of C. crocuta. However, other factors might also have played some role in this process. As was pointed by Vinuesa et al. (2016), the ancestral/plesiomorphic condition in bone-crushing hyenas were limited social behavior with putative main scavenger behavior (i.e. similar to extant brown hyenas). The extreme social behavior displayed by the extant spotted hyenas probably was related with unknown selective pressures occurred during the Middle Pleistocene in East Africa or idiosyncratic biological characteristics of this species. Taking this into account, the scenario discussed above showed not to be so clear and still needs further verification (Vinuesa et al. 2016).

The first Ukrainian record described here, dated to 1.3–1.2 Ma, fits well into this scenario. Beside the single record from Ubeidyia, dated to 1.5–1.4 Ma (Martínez-Navarro et al. 2009), there are no other known Eastern European or Western Asian records of C. crocuta. In this context, the co-occurrence P. brevirostris with C. crocuta seems unlikely in this area, but needs further investigation. Specimens of P. brevirostris from various earlier sites located in neighbouring areas to the so called “Khapry fauna” of Nogaisk are morphologically similar to the Asian forms. This might indicate that P. brevirostris may have appeared in Eastern Europe earlier than in western regions (Sotnikova et al. 2002). In general, the Asian influence on the fauna of the Azov region was reasonably strong. Apart from carnivores, which appeared here earlier than in Western Europe by the invasion from the east, the Khaprovian faunal assemblage consists of typical Asian elements (Sotnikova et al. 2002).