1. Introduction
Pliocene deposits are widely exposed in various areas of the Scythian Platform, a unit of the Eastern Carpathian Foreland, located in eastern Romania (Săndulescu Reference Săndulescu1984). These deposits were described as being part of the last sedimentary megacycle of the platform, which accumulated during the Badenian–Pleistocene interval (Ionesi Reference Ionesi1994). The Pliocene fossil assemblages of Mălușteni and Berești (Fig. 1) have been known since the beginning of the 20th century, (Athanasiu Reference Athanasiu1915; Simionescu Reference Simionescu1930, Reference Simionescu1932), remaining to this day the richest continental Pliocene sites of the Dacian Basin (for an updated description of the Plio–Pleistocene units of the basin, see Andreescu et al. Reference Andreescu, Codrea, Lubenescu, Munteanu, Petculescu, Știucă and Terzea2013), and of all Romania, in both large and small mammals.
Figure 1 Geographical position of the fossil sites Berești and Mălușteni in eastern Romania, and the general geological map of the area (after Saulea et al. Reference Saulea, Ghenea and Ghenea1967). Colour legend: (1) Lower Pliocene; (2) Upper Pliocene; (3) Lower Pleistocene; (4) Middle–Upper Pleistocene; and (5) Holocene.
The site of Mălușteni, the more productive of the two, is located along the Româneasca Creek, immediately under the Lacului Hill. According to Ghenea (Reference Ghenea1968), the Pliocene of Mălușteni includes fluvio-lacustrine deposits dominated by sands, where vertebrate fossils accumulated in some channel deposit lenses. Aside from the small mammal taxa (reassessed in this paper), Simionescu (Reference Simionescu1930) describes from Mălușteni fossil vertebrate remains, assigning them to various large mammal groups, such as primates (Macacus florentinus and Dolichopithecus ruscinensis), carnivores (Vulpes donnezani, Lutra rumana, Lynx issiodorensis, Promephitis malustenensis, Mustella aff. martes, Mustella aff. robusta, Mustella aff. putorius, and Phoca sp.), perissodactyls (Tapirus arvernensis, Rhinoceros sp., Hipparion gracile, and Equus cf. robustus), artiodactyls (Sus provincialis, Camelus bessarabensis, Capreolus australis, Cervus cusanus, Cervus buladensis, Cervus cf. ramosus, Cervus cf. arvernensis, and Palaeoryx athanasiui), proboscideans (Mastodon arvernensis and Mastodon borsoni), as well as to reptiles (the chelonians Testudo sp., Clemmys sp., and Trionyx sp.; the snake Coelopeltis cf. laurenti; the lizard Lacerta sp.), and to fishes (Acipenser sp., Esox sp., Lamna sp., Dentex sp., Sargus sp., Chrysophrys sp., and Myliobatis sp.). Except for the chelonians, assigned by Macarovici & Vancea (Reference Macarovici and Vancea1960) to Testudo pregraeca ibera, Testudo grandis, and Clemmys malustensis, and further reassigned by Młynarski (Reference Młynarski1969) to Testudo macarovicii, Geoemyda malustensis, Geoemyda cf. mossoczyi, and Chelydridae gen. et sp. indet., the above-listed vertebrates from Mălușteni were never thoroughly revised, and their occurrence at the site should not be taken for granted until such a revision takes place.
The palaeontological site of Berești is located in the Porcului Ravine, near the railway station. From the lower part of the succession, Macarovici (Reference Macarovici1960) described white sands bearing several bivalve species of the genus Prosodacna (of Dacian, i.e., Early Pliocene, age), but also sands with microconglomerate-like sandstone concretions bearing vertebrate fossils (especially chelonian fragments) from the upper part of the succession. Except for the small mammals (revised in this paper), Simionescu (Reference Simionescu1932) describes large mammals’ remains from Berești, assigning them to primates (Macacus florentinus), carnivores (Machairodus cf. cultridens, Lutra rumana, Vulpes sp., Canis sp., Mustella aff. martes, and Mustella sp.), perissodactyls (Hipparion gracile and Equus sp.), artiodactyls (Sus sp., Capreolus australis, Cervus cf. arvernensis, and Cervus sp.), proboscideans (Mastodon sp.), as well as to reptiles (the chelonians Clemmys sp., Testudo sp., and Trionyx sp.), and to fishes (Scardinius sp., Barbus sp., Carpis sp., and Tinca sp.). As is the case for the material from Mălușteni, the above-listed specimens from Berești were never revised, so their occurrence should be considered only tentative, until a proper revision of the specimens takes place.
The field research of Ghenea (Reference Ghenea1968) represents the last attempt of systematic fossil collection from these sites. Following their original description, the initial fossil material, deposited in the collections of several institutions (see below, section 2. Material and methods) has not been studied intensely, only some small mammal taxa being occasionally reassessed, or just mentioned in general syntheses on Pliocene faunas of Romania (e.g., Radulesco & Samson Reference Radulesco and Samson1967a, Reference Radulesco and Samsonb; Samson & Radulescu Reference Samson and Radulesco1973; Radulescu & Samson, Reference Radulescu and Samson1989a, Reference Radulescu and Samsonb; Rădulescu & Samson Reference Rădulescu and Samson1995; Terzea Reference Terzea, Aguilar, Legendre and Michaux1997, Reference Terzea2005; Radulescu et al. Reference Radulescu, Samson, Petculescu and Stiucă2003, but see below, section 3. Systematic palaeontology, for additional references). Very little novel large vertebrate material has been described since the early days of research in the area – only some new rhinocerotid fossil material was more recently published in several papers (e.g., Codrea Reference Codrea1993; Codrea & Costănel Reference Codrea and Costănel2003).
The faunal lists including small mammals from Berești and Mălușteni were published by Simionescu (Reference Simionescu1930, Reference Simionescu1932), and by other authors, in subsequent syntheses presenting Pliocene continental mammal faunas from Romania (e.g., Terzea Reference Terzea, Aguilar, Legendre and Michaux1997, Reference Terzea2005; Radulescu et al. Reference Radulescu, Samson, Petculescu and Stiucă2003), and they include many taxa that are new to science. Since the two sites have produced the most diverse Pliocene taxonomic assemblages of Romania, they are extremely important in understanding the palaeogeographical and stratigraphical ranges for the small mammal taxa of Central-Eastern Europe, in general, and of Romania, in particular. The aim of this work is to study all available small mammal specimens collected from the above-mentioned sites, now divided into three main collections (see below, section 2. Material and methods), to reassess the taxonomic validity of the taxa described so far from these sites, and to reassess the age of the faunal assemblages by comparing it to updated research on the biostratigraphy of Pliocene small mammal assemblages.
2. Material and methods
The fossils studied here are currently deposited at the Laboratory of Palaeontology, Faculty of Geology and Geophysics, University of Bucharest (Romania) – LPB (FGGUB), at the ‘Emil Racoviță’ Institute of Speleology, Romanian Academy (Bucharest, Romania) – ISER, and at the Department of Geology, ‘Alexandru Ioan Cuza’ University of Iași (Romania) – UAIC.
The followed terminologies are: Ziegler (Reference Ziegler1990) and Furió (Reference Furió2007) for erinaceids, Oliver & Peláez-Campomanes (Reference Oliver and Peláez-Campomanes2013) for ‘cricetids’, Cuenca-Bescós (Reference Cuenca Bescós1988) for sciurids, Sarica & Sen (Reference Sarica, Sen, Fortelius, Kappelman, Sen and Bernor2003) for spalacids, Hugueney (Reference Hugueney, Rössner and Heissig1999) for castorids, and López Martínez (Reference López Martínez1984, Reference López Martínez1989), with modifications by Hordijk (Reference Hordijk2010) for lagomorphs. We follow the measuring schemes proposed by Doukas (Reference Doukas1986) for erinaceids, by Oliver & Peláez-Campomanes (Reference Oliver and Peláez-Campomanes2013) for ‘cricetids’, by Cuenca-Bescós (Reference Cuenca Bescós1988) for sciurids, by Sarica & Sen (Reference Sarica, Sen, Fortelius, Kappelman, Sen and Bernor2003) for spalacids, by Rekovets et al. (Reference Rekovets, Stefen and Demeshkant2020) for castorids, and by Hordijk (Reference Hordijk2010) for lagomorphs.
Scanning electron micrographs were taken at the Arheoinvest Research Center (Iași, Romania), using a VEGA II LSH scanning electron microscope (SEM) produced by TESCAN, coupled with a QX2 QUANTAX EDX detector produced by BRUKER/ROENTEC. Larger specimens, that did not fit inside the SEM, were photographed with an Axiocam 105 colour microscopy camera, mounted on a Zeiss Stemi 305 trino stereomicroscope (at LPB[FGGUB] and at UAIC), and, in the sole case of a castorid hemimandible (the largest of all specimens), with a Nikon D7200 DSLR camera (at LPB[FGGUB]).
3. Systematic palaeontology
Phylum CHORDATA Bateson, 1885
Class MAMMALIA Linnaeus, 1758
Order EULIPOTYPHLA Waddell et al. 1999
Family Talpidae Fischer, 1814
Subfamily Desmaninae Thomas, 1912
Desmana Gueldenstaedt, 1977
Desmana verestchagini Topachevski, 1961
Myogale sp. in Simionescu (Reference Simionescu1932)
Desmana verestchagini amutriensis in Radulescu et al. (Reference Radulescu, Samson and Știucă1989)
Desmana amutriensis in Rădulescu et al. (Reference Rădulescu, Samson and Știucă1992)
Occurrence: Early Pliocene (MN14)
Site: Măluşteni
Material: one left p4 (ISER Ml/007a), one left hemimandible with p4–m2 (ISER Ml/007b; Fig. 2A), one right humerus (LPB[FGGUB] 109 – Fig. 2B, C)
Figure 2 Eulipotyphla from Bereşti and Măluşteni. Desmana verestchagini: (A1–A3) left p4–m2 from Măluşteni, in occlusal view; (B, C) left humerus from Măluşteni, in anterior (B), and posterior (C) views. Talpa sp.: (D, E) left humerus from Măluşteni (holotype of Talpa neagui after Radulescu & Samson Reference Radulescu and Samson1989b), in anterior (D), and posterior (E) views; (F, G) complete right humerus from Măluşteni, in anterior (F), and posterior (G) views. Erinaceus sp.: (H) left m1 from Bereşti, in occlusal view.
Measurements: See Table 1
Table 1 Measurements of the small mammals from Bereşti and Măluşteni (in mm).
Description: Material is described in Rădulescu et al. (Reference Rădulescu, Samson and Știucă1992)
Remarks:
These specimens were described by Simionescu (Reference Simionescu1932) as Myogale sp., but they were later used by Radulescu et al. (Reference Radulescu, Samson and Știucă1989) as a basis for the description of a new subspecies, Desmana verestchagini amutriensis. Nevertheless, it was finally described by Rădulescu et al. (Reference Rădulescu, Samson and Știucă1992) as a separate species, Desmana amutrensis (Rzebik-Kowalska Reference Rzebik-Kowalska, van den Hoek Ostende, Doukas and Reumer2005). The genus Desmana, following the amended diagnosis made by Minwer-Barakat et al. (Reference Minwer-Barakat, García-Alix, Martín-Suárez and Freudenthal2020), is characterised by its medium–large size, by the different sizes and morphology of the incisors, as well as by the relative length between the different premolars. Since the incisors are not preserved in the materials studied here, we can only confirm assignment to the genus by using the relative sizes of the p2 and p3 alveoli, with p2 being larger than p3. Moreover, the size of the specimen indicates that it belongs to the medium-sized desmanine group, as already indicated by Rzebik-Kowalska (Reference Rzebik-Kowalska2002).
The holotype of D. amutrensis consists of a single broken M2, in which the parastyle and metastyle are broken (Radulescu et al. Reference Radulescu, Samson and Știucă1989; Rădulescu et al. Reference Rădulescu, Samson and Știucă1992). In spite of the material being incomplete, close morphological similarities to D. verestchagini can be observed, with the latter species being larger. The material found at Măluşteni only consists of lower dentition, the size of which is larger than that reported for the specimens described from Kosyakin quarry (Russia, type locality) and Kardia (Greece; Rümke Reference Rümke1985). However, in the extant species Desmana moschata, such a large variability in size is observed (Minwer-Barakat et al. Reference Minwer-Barakat, García-Alix, Martín-Suárez and Freudenthal2020), that the previously-cited populations of D. verestchagini and those studied here could be included within the same species, each at opposite ends of the size range. Therefore, also taking into account the scarcity of remains, we propose the synonymy of D. amutrensis with D. verestchagini.
A humerus of this species was also found and described by Simionescu (Reference Simionescu1930) as Sorex sp. However, due to its close similarity to the humerus described by Rümke (Reference Rümke1985) as D. verestchagini from Kardia, the humerus described by Simionescu (Reference Simionescu1930) as Sorex sp. is also assigned here to D. verestchagini.
Subfamily Talpinae Fischer von Waldheim, 1817
Genus Talpa Linnaeus, 1758
Talpa sp.
Talpa sp. in Simionescu (Reference Simionescu1930)
Talpa sp. in Simionescu (Reference Simionescu1932)
Talpa minor in Kowalski (1956)
Talpa minor in Sulimski (Reference Sulimski1959)
Talpa minor ssp. (partim) in Rabeder (1972)
Talpa sp. in Samson & Radulesco (Reference Samson and Radulesco1973)
Tapa neagui new species in Radulescu & Samson (Reference Radulescu and Samson1989b)
Occurrence: Early Miocene (MN2)–present
Sites: Bereşti and Măluşteni
Material: six humeri (ISER Br/001 – Fig. 2D, E; Br/002; Ml/018 – Fig. 2F, G; Ml/019; Ml/020; Ml/021)
Description: See Radulescu & Samson (Reference Radulescu and Samson1989b)
Remarks:
Based on the humeri found at Bereşti and Măluşteni, Radulescu & Samson (Reference Radulescu and Samson1989b) erected a new talpid species: Talpa neagui. The new species was defined, and humeri from a few European sites were assigned to it, based only on the relative width of the diaphysis, even in the absence of dentition. As it has no associated dentition, and the relative size of the diaphysis overlaps the ranges seen in other species, such as Talpa minor and Talpa fossilis/Talpa europaea (Rzebik-Kowalska Reference Rzebik-Kowalska2013), we prefer to leave the species T. neagui as nomen dubium, and assign the material listed here to Talpa sp. until additional specimens will be found at Berești and Mălușteni.
Family Erinaceidae Fischer, 1817
Erinaceus Linnaeus, 1758
Erinaceus sp.
Erinaceus sp. in Simionescu (Reference Simionescu1930)
Occurrence: Late Miocene (MN10)–present
Site: Bereşti
Material: one left m1 (Fig. 2H; ISER, unnumbered specimen, labeled ‘Erinaceus’)
Measurements: See Table 1
Description: The outline of the occlusal surface is sub-rectangular. The V-shaped trigonid is slightly narrower and longer than the talonid. The paraconid is fairly lowered and lies at the end of a long and slightly curved paralophid. The hypolophid and the entocristid have an open V-shape, due to being worn. The anterior cingulum is narrow and connects the labial side of the paraconid to the posterior cingulum. The oblique cristid is straight. The posterior cingulid starts at the entoconid and it is connected to the anterior cingulum.
Remarks:
This molar was cited and preliminarily studied by Simionescu (Reference Simionescu1930). In this article, we describe and illustrate it for the first time. The classification made by the previous author is maintained, as the material is scarce and not very diagnostic. In terms of size, it is grouped with other medium-sized Plio–Pleistocene European species, such as Erinaceus samsonowiczi, Erinaceus ostramosi, Erinaceus olgae, and Erinaceus lechei, as well as with Erinaceus sp. 1 and 2 from Beftia, western Romania (Sulimski Reference Sulimski1959; Rzebik-Kowalska Reference Rzebik-Kowalska2000; Popov Reference Popov2004). Due to the small size of the paraconid, species such as E. lechei and E. samsonowiczi can be discarded (Doukas et al. Reference Doukas, van den Hoek Ostende, Theocharopoulos and Reumer1995; Rzebik-Kowalska Reference Rzebik-Kowalska2000), while E. olgae can also be rejected as it has shorter trigonids (Flynn & Wu Reference Flynn, Wu, Flynn and Wu2017).
Order Rodentia Bowdich, 1821
Family Muridae Illiger, 1811
Subfamily Arvicolinae Gray, 1821
Mimomys sp. or Promimomys sp.
Arvicola pliocenicus in Simionescu (Reference Simionescu1930)
Mimomys moldavicus n. sp., Kormos (Reference Kormos1932)
Promimomys moldavicus in Kretzoi (1955)
Promimomys moldavicus in Samson & Radulesco (Reference Samson and Radulesco1973)
Mimomys silasensis n. sp. in Jánossy (Reference Jánossy1974)
Mimomys gracilis in Sen (1977)
Promimomys moldavicus moldavicus in Gromov (Reference Gromov, Gromov and Poliakov1977)
Mimomys davakosi n. sp. in van de Weerd (Reference Van der Weerd1979)
Promimomys moldavicus in Shushpanov (Reference Shushpanov1985)
Promimomys moldavicus in Aleksandrova (Reference Aleksandrova1986)
Mimomys moldavicus in Radulescu & Samson (Reference Radulescu and Samson1989a)
Occurrence: MN14–early Middle Pleistocene
Sites: Măluşteni
Material: one left hemimandible with m1 + m2 (ISER, specimen not numbered – Fig. 3A–D), one right M2 (ISER, specimen not numbered – Fig. 3E, F)
Figure 3 Rodents (except castorids) from Bereşti and Măluşteni, in occlusal view, except B, D, and F in labial view. Mimomys sp. or Promimomys sp.: (A, B) left m1; (C, D) left m2, (A–D) from the same hemimandible from Măluşteni; (E, F) right M2 from Măluşteni. Allocricetus sp.: (G) right m1–3 from Bereşti. Spermophilus cf. nogaici: (H) right p4 from Măluşteni; (I) left m1/2 from Măluşteni. Pliospalax macoveii: (J) left m1,2 from Măluşteni with advanced wear; (K) left m1–3 from Măluşteni (holotype? of Prospalax rumanus after Simionescu Reference Simionescu1930); (L) right m1–3 from Bereşti; (M) right m1–3 from Măluşteni (holotype? of Prospalax macoveii after Simionescu Reference Simionescu1930).
Measurements: See Table 1
Description: See Radulescu & Samson (Reference Radulescu and Samson1989a).
Remarks:
Simionescu (Reference Simionescu1930) described two arvicolid hemimandibles from Măluşteni, classifying them as Arvicola pliocenicus, later synonymised to Mimomys pliocaenicus. One of them, which is not in the collection under study and is probably lost, was used by Kormos (Reference Kormos1932) to establish the species Mimomys moldavicus Kormos, Reference Kormos1932 (or Promimomys moldavicus according to some authors). While Radulescu & Samson (Reference Radulescu and Samson1989a) revised the material in order to confirm the validity of the species, several attempts have been made to synonymise it to Promimomys cor (Suata-Alpaslan Reference Suata-Alpaslan2015), or to Mimomys davakosi (Fejfar et al. Reference Fejfar, Heinrich and Lindsay1998, Reference Fejfar, Heinrich, Kordos and Maul2011), using as type material a mandible with very worn teeth. In addition, Radulescu & Samson (Reference Radulescu and Samson1996) describe further material assigned to this species in Drănic 0 (M. moldavicus group) and 1 (Mimomys moldavicus/Propliomys), but this material does not fit with the type material of the species in the number of triangles of m1.
The mandible available for this study is characterised by having narrower reentrant angles than the holotype of M. moldavicus described by Kormos (Reference Kormos1932) and by Radulescu & Samson (Reference Radulescu and Samson1989a), as well as presenting an islet on the anteroconid, characters that do not appear in the holotype of M. moldavicus, if the species were to be considered valid. For these reasons, and until they can be compared to each other, we consider that the mandible included in this study could belong to a different species than the former. According to Hordik & de Bruijn (Reference Hordijk and de Bruijn2009), the genera Promimomys and Mimomys are differentiated by ‘A weak or relatively well-developed and persistent LRA4 on m1, a Mimomys-kante on m1 and development of dentine tract ea on m1’. Because of the high degree of digestion by a predator, these characters do not appear in the material under study, making it impossible to assign it to a genus. In any case, more material from this site, other than from senile individuals, will allow further studies aiming to synonymise or better characterise M. moldavicus.
Subfamily Cricetinae Fischer, 1817
Genus Allocricetus Schaub, 1930
Allocricetus sp.
Cricetulus simionescui sp. nova in Schaub (Reference Schaub1931)
Cricetulus simionescui in Simionescu (Reference Simionescu1930)
Occurrence: Late Miocene (MN9)–Early Pleistocene
Site: Bereşti
Material: Two hemimandibles bearing m1–m3 (ISER NN1; NN2 – Fig. 3G)
Measurements: See Table 1
Description: The m1 has a slightly bifid anteroconid which is connected to the junction of the protoconid–metaconid by a thick, probably double, anterolophulid, resulting in a very worn area of the tooth. The mesolophid is absent and the thick posterolophid is connected to the distal base of the entoconid. The m2 and m3 have the protosinusid as deep as the sinusid; the labial anterolophid is strong, and the lingual one is reduced. The m2 has a short mesolophid, which is more robust than the one found on M3, and is directed forward.
Remarks:
Two mandibles belonging to a medium-sized cricetine, deposited in the ISER collections, were classified by Simionescu (Reference Simionescu1932) as Cricetulus simionescui Schaub, 1932. The identification of cricetines from the Miocene-to-Pleistocene interval is very complicated, as some authors list more than 20 genera, but many of them are morphologically very similar, and, probably, after being revised, many of them will end up being synonymised (McKenna & Bell Reference McKenna and Bell1997; De Bruijn et al. Reference de Bruijn, Doukas, van den Hoek Ostende and Zachariasse2012). Simionescu (Reference Simionescu1990) cites a Cricetus sp. from Măluşteni but it is not known what material was identified as such.
Thanks to their size, and stratigraphical distribution, the hemimandibles available for this study could be not assigned to practically all genera except Allocricetus and Neocricetodon. Sen et al. (Reference Sen, Karadenizli, Antoine and Saraç2019) morphologically differentiate the two genera, the former being characterised by narrow lower molars, single anterolophulids connected to the protoconid, narrow and slightly bifid anteroconid and finally absent or weak mesolophids, but in the case they should possess these structures they should have a greater size of the molar series. On the other hand, according to the same authors, Neocricetodon is characterised by molars bearing broad anteroconids, well-developed mesolophids, and usually double anterolophulids, characters that differ from the material under study, thus making it more similar to Allocricetus. Within Neocricetodon, the following Pliocene species can be excluded: Allocricetus bursae Schaub, 1930, the most common species in the Plio–Pleistocene of Europe, although it has a similar size to the material studied here, can be discarded because of its large m1 anteroconid; A. primitivus Wu & Flynn, Reference Wu, Flynn, Flynn and Wu2017, from China, is also very similar, but has a wider m1; and Allocricetus ehiki Schaub, 1930 can also be excluded because of its large size (Wu & Flynn Reference Wu, Flynn, Flynn and Wu2017; Sen et al. Reference Sen, Karadenizli, Antoine and Saraç2019). Therefore, the species most similar to the material under study is Allocricetus primitivus, of Chinese origin and primitive morphology.
Family Sciuridae Fischer, 1817
Subfamily Xerinae Osborn, 1910
Tribe Marmotini Pocock, 1923
Genus Spermophilus Cuvier, 1825
Spermophilus cf. nogaici (Topachevsky, 1957)
Spermophilus sp. in Simionescu (Reference Simionescu1930)
Spermophilus sp. in Simionescu (Reference Simionescu1932)
Occurrence: Late Pliocene–Middle Pleistocene
Site: Măluşteni
Material: two hemimandibles with a p4 (LPB[FGGUB] 110 – Fig. 3H), one m1/2 (ISER, not registered specimen – Fig. 3I)
Measurements: See Table 1
Description: The p4 is premolariform. The trigonid is moderately labiolingually compressed, which results in the occlusal outline of the tooth crown being subtriangular (in one of the specimens) to subrectangular (in the other). From the main cusps, the anterior two (i.e., the metaconid and the protoconid) are the tallest, and are closely situated; the metaconid is slightly taller than the protoconid. The anterolophid is well developed, but does not close the anterior basin. There is no labial anterolophid and the metalophid is low and well developed. One of the specimens presents a low mesolophid which resembles an elongated cusp and has no contact with the entoconid. The mesoconid is not present, and the entoconid is poorly developed. The entolophid is absent, whereas the hypoconulid is small. Finally, one of the specimens has a straighter posterolophid than the other.
The m1/2 is similar to the p4, but has a more rectangular occlusal outline; the trigonid is shorter, the anterolophid is connected to the metaconid, and the metalophid is less developed.
Remarks:
Two hemimandibles retaining the p4s, and an isolated m1/2, all belonging to a species of ground squirrel, are deposited in the collections of ISER and LPB(FGGUB). One of the p4-bearing hemimandibles was already published by Simionescu (Reference Simionescu1930) as Spermophilus sp., whereas the remaining specimens remained unpublished, until now. Kormos (Reference Kormos1940; as cited by Samson & Radulesco Reference Samson and Radulesco1973) assigns one of the hemimandibles to Citellus primigenius, whereas Samson & Radulesco (Reference Samson and Radulesco1973) reinterpreted that specimen as Citellus sp., and the other p4-bearing hemimandible as Pliosciuropterus sp. The material studied here presents similarities to S. nogaici, species described by Sinitsa & Pogodina (Reference Sinitsa and Pogodina2019) from the uppermost Pliocene–Lower Pleistocene deposits of Ukraine. It is worth mentioning that the basalmost and first known species of the genus in Eurasia, Spermophilus praecox, described from the Late Pliocene–Early Pleistocene interval, differentiates itself by having a larger overall size, a p4 that is more trapezoidal due to its triangular trigonid, and, in general, by lacking the metalophid in the m1/2. It also differs from the other primitive species present in the Late Pliocene, such as Spermophilus polonicus, by having a less triangular p4 (Popova et al. Reference Popova, Lemanik, Ulbricht and Nadachowski2021). Therefore, this would be the first finding of this genus in Europe, outside of Ukraine, making their migration patterns occur earlier than the Late Pliocene, as it was previously established (Sinitsa et al. Reference Sinitsa, Pogodina and Kryuchkova2019).
Family Spalacidae Gray, 1821
Pliospalax Kormos, 1932
Pliospalax macoveii (Simionescu Reference Simionescu1930)
Prospalax rumanus n. f. in Simionescu (Reference Simionescu1930) (in part)
Prospalax Macoveii n. f. in Simionescu (Reference Simionescu1930) (in part)
Spalax (Pliospalax Kormos n. g.) Macoveii in Simionescu (Reference Simionescu1932)
Occurrence: Late Miocene (MN13)–Late Pliocene (MN15)
Sites: Bereşti and Măluşteni
Material: One M1 (LPB[FGGUB] 225-1), one M2 (ISER NN), three hemimandibles m1–m3 (ISER 112.17 – Fig. 3K; ISER, specimen not numbered – Fig. 3L; LPB(FGGUB) 117-3 – Fig. 3M), six hemimandibles m1–m2 (LPB[FGGUB] 117-1; LPB[FGGUB] 117-2; UAIC SM-8-3 – Fig. 3J; UAIC SM-8b; ISER 116-15), three m1 (LPB[FGGUB] 107-11; UAIC SM-8-4; ISER 116-16) and two m2 (UAIC SM-8-1; UAIC SM-8-2; ISER 116-17).
Measurements: See Table 1
Description:
M1: subquadrate occlusal shape, with three labial folds and one lingual reentrant fold. One small labial anteroloph is present, connected to the protocone and the paracone. The mesoloph is connected to the metacone. The posteroloph is long and narrow.
M2: subtriangular occlusal shape, with one labial and one lingual reentrant fold. The anteroloph is connected to the paracone, isolating an islet, which in turn is located in the continuation of the sinus. The mesosinus is L-shaped, and posteriorly directed.
m1: the tooth may be subtriangular in occlusal view (nine out of 11), or it may appear as more rounded (two out of 11); the protosinusid may be curved and directed to the anterior side of the tooth (seven out of 11), it may be straight (two out of 11), or it may appear in the form of an islet (two out of 11). In two specimens, a small labial protosinusid is present. The sinusid may be straight (eight out of 11), it may be curved (two out of 11) and directed to the posterior side, or it may only appear as an islet (one out of 11). An islet may (eight out of 11) or may not (three out of 11) be present on the lingual side of the sinusid.
m2: It is similar to m1, but has a more quadrate outline, and the islet may (four out of 10) or may not (five out of 10) be present.
m3: It is similar to m2, but the tooth may be narrower (one out of three), or it more triangular (two out of three).
Remarks:
Two species of spalacids belonging to the genus Prospalax were described by Simionescu (Reference Simionescu1930) from Mălușteni (Prospalax macoveii and Prospalax rumanus). Subsequently, Kormos (Reference Kormos1932) defined the genus Pliospalax based on this material, separating it from the genus Prospalax, which nowadays is considered an anomalomyid (Nesin & Kovalchuk Reference Nesin and Kovalchuk2021). From that moment on, there has been a large amount of literature doubting whether the two species are different, and many authors synonymise Pr. rumanus with Pl. macoveii (e.g., Hordik & de Bruijn Reference Hordijk and de Bruijn2009). This all came from the fact that the two species are separated by size (Simionescu Reference Simionescu1930), but, after comparing their measurements, they are not sufficiently different to make such separation. On the other hand, some authors, such as Popov (Reference Popov2017), following Topachevskyi (Reference Topachevskyi1969), place this species within the genus Microspalax. Other authors, such as Sulimski (Reference Sulimski1964), synonymise both species with Prospalax priscus, although with some doubts.
Rădulescu & Samson (Reference Rădulescu and Samson1995) describe a new hemimandible belonging to Pl. macoveii, and another one, bearing only two molars, assigned to Microspalax cf. odessanus. However, these specimens are not among the material found by us in the three collections investigated, and, therefore, could not be studied.
One of the major problems of the family Spalacidae is the great variability in their dental morphology, which depends on wearing degree. Although the typical S-form occurs in the extinct genus Pliospalax, as well as in juveniles of the current genus Nannospalax, the two genera can be differentiated by the number of reentrants on the m1's lingual and labial faces, with other characters that can be used for their identification as well (de Bruijn & van der Meulen Reference de Bruijn and van der Meulen1975; Sarica & Sen Reference Sarica, Sen, Fortelius, Kappelman, Sen and Bernor2003; Sen & Sarica Reference Sen and Sarica2011; Popov Reference Popov2017; Erdal et al. Reference Erdal, Erturaç, Dalfes and Sen2018). Despite the large amount of described material, no specimens have been found to have little wear, which makes comparison between specimens, and to specimens belonging to other species difficult. Therefore, taking into account the wide range of different morphologies, and the similar size between Pr. rumanus and Pl. macoveii, we here decided to group both previously-described species into a single one, Pliospalax macoveii.
Family Castoridae Gray, 1821
Genus Trogontherium Newton, 1890
Trogontherium minus Newton, 1890
Steneofiber covurluiensis n. f. in Simionescu (Reference Simionescu1930) (in part)
Castor sp. in Simionescu (Reference Simionescu1930) (in part)
Steneofiber covurluiensis n. f. in Simionescu (Reference Simionescu1932) (in part)
Castor sp. in Simionescu (Reference Simionescu1932) (in part)
Zamolxifiber covurluiensis n. sp. in Radulesco & Samson (Reference Radulesco and Samson1967b) (in part)
Romanocastor filipescui, n. sp. in Radulesco & Samson (Reference Radulesco and Samson1967b) (in part)
Occurrence: Early Pliocene (MN14)–Early Pleistocene
Sites: Bereşti and Măluşteni
Material: five M1/2 (UAIC SM-7-1 – Fig. 4B; UAIC SM-7-3 – Fig. 4D; ISER 27; UAIC SM-7-4 – Fig. 4E; ISER 14), eight M3 (UAIC SM-14-5; ISER 23; ISER 24; ISER 25; ISER 26; ISER 10; ISER 11; ISER 13), five p4 (LPB[FGGUB] 195; UAIC SM14-3; UAIC SM-7-2 – Fig. 4C; ISER 22; ISER 12), one m1/2 (LPB[FGGUB] 104-1), one m3 (UAIC SM-14-4), and one hemimandible m1-2 (UAIC SM-14-1 – Fig. 4A).
Figure 4 Castorids and lagomorphs from Bereşti and Măluşteni in occlusal view, except G, in labial view, mirror-image, to correspond to the orientation in F. Trogontherium minus: (A) left m1,2 from Măluşteni (probably partial holotype of Steneofiber covurluiensis after Simionescu Reference Simionescu1930); (B) right M1,2 from Măluşteni; (C) right p4 from Măluşteni; (D) left M1,2 from Măluşteni; (E) left M1,2 from Măluşteni. Castor fiber: (F, G) left hemimandible bearing p4–m3 from Măluşteni. Trischizolagus dumitrescuae: (H) right p3–4 from Măluşteni; (J) left upper molariform from Măluşteni. Ochotona ursui: (I) right P3 from Măluşteni; (K) left hemimandible bearing p3–m3 from Măluşteni (holotype?); (L) left hemimandible bearing p3–m2 from Măluşteni.
Measurements: See Table 1
Description:
M1/2: subquadrangular tooth, with a narrow posterior side. The metafossette and the parafossette are narrow and long, the mesoflexus and the hypoflexus are open and shorter than the fossettes; and a mesofossette is present in one of the specimens. Two roots are present. One unworn specimen has a wide flexus and an anterior islet.
M3: triangular tooth. The anterior lobe of the tooth may be long (three out of seven), or short (four out of seven). The hypoflexus and the paraflexus may be curved (five out of seven), or straight (two out of seven) and in line. The mesoflexus and the metaflexus may be curved and extend themselves along the tooth (three out of seven), or fossettes may be present (three out of seven). Two small islets may (one out of seven) or may not (five out of seven) be present on the labial side. One of the M3s has divided central lobes. An extra flexus (one out of seven), a fossette (five out of seven), or none of these elements (one out of seven) may exist at the posterior side. Two roots are present.
p4: subrectangular to suboval tooth with narrow and long labial flexids. One of the specimens does not present fossetids, the other two contain the parafossettid and the metafossetid. The paraflexid is the longest flexid, and it is slightly curved. The hypoflexid is short. There are three roots.
m1/2: similar to, but shorter than p4; the paraflexid and the metaflexid are parafossettid and metafossetid.
m3: similar to m1/2, but more quadrangular in occlusal view.
Castor Linnaeus, 1758
Castor fiber Linnaeus, 1758
Castor fiber in Simionescu (Reference Simionescu1930)
Castor praefiber in Radulesco & Samson (Reference Radulesco and Samson1967b)
Occurrence: Pliocene–present
Sites: Bereşti and Măluşteni
Material: one M1/2 (LPB[FGGUB] 104-3), one p4 (LPB[FGGUB] 104-2) and one hemimandible bearing the p4–m1 (UAIC SM-6 – Fig. 4F, G).
Measurements: See Table 1
Description:
M1/2: broken tooth, with the connection between the protoloph and the anteroloph at the middle of the tooth. The paraflexus and the hypoflexus are near continuous. The mesoflexus and the metaflexus are connected in a Y-shaped form. There are no roots.
p4: subtrapezoidal tooth with a very long and more irregular flexus than in T. minus. The hypoflexus is short and wide.
m1: similar to p4, but with a shorter and longer hypoflexus.
m2: similar to, but shorter than m1.
m3: similar to m2, but with a less quadrate occlusal shape, and having a more irregular flexus.
Remarks:
Abundant material belonging to beavers has been found at the two sites under study. This material has a long taxonomic history, being first described from Măluşteni (Simionescu Reference Simionescu1930) as Steneofiber covurluiensis Simionescu, Reference Simionescu1930, Castor fiber and Castor sp., and from Bereşti (Simionescu Reference Simionescu1932) as Steneofiber covurluiensis and Castor sp. Subsequently, Radulesco & Samson (Reference Radulesco and Samson1967b) move from Castor sp. to C. praefiber, and define the genus Zamolxifiber, for S. covurluiensis (in part), and Romanocastor filipescui (Simionescu Reference Simionescu1930) from Bereşti as part of S. covurluiensis. Subsequently, this material has been reinterpreted several times, and all reinterpretations agree that Zamolxifiber and Romanocastor should be synonymised to a single, already existing, genus. Thus, Mayhew (Reference Mayhew1978) synonymises the two Romanian species and genera to Trogontherium minutum (von Meyer, 1838), but, eventually, a subgenus was created for the latter species (Euroxenomys), which Hugueney & Duranthon (Reference Hugueney, Duranthon, Peigné and Sen2012) subsequently elevate to genus status. Another less approved interpretation of this material is the one proposed by Korth (Reference Korth2001), who discusses that the morphology of these remains is similar to that of Steneofiber, but, not having studied mandible morphology, he only comments on its synonymy. On the other hand, the species T. minutum has been restricted to the Miocene, while its possible descendant, T. minus, has been restricted to the Pliocene, the taxonomy of these species remaining, however, under discussion (Hugueney et al. Reference Hugueney, Guérin and Poidevin1989), and for this reason the material studied here has been assigned to T. minus. The holotypes of the genera and species Zamolxifiber covurluiensis and Romanocastor filipescui are not found among the material studied. Furthermore, Simionescu (Reference Simionescu1990) mentions C. fiber and C. cf. praefiber Schreuder, 1928 from Măluşteni, but without stating clearly on what material he based these identifications. On the other hand, the species described as Castor praefiber, was synonymised to C. fiber by van der Weerd (Reference Van der Weerd1979).
Lagomorpha Brandt, 1855
Family Leporidae Fischer de Waldheim, 1817
Genus Trischizolagus Radulesco & Samson, Reference Radulesco and Samson1967a
Trischizolagus dumitrescuae Radulesco & Samson, Reference Radulesco and Samson1967a
Lepus valdarnensis in Simionescu (Reference Simionescu1930)
Alilepus sp. in Dice (Reference Dice1931)
Lepus valdarnensis in Simionescu (Reference Simionescu1932)
Alilepus sp. in Kormos (Reference Kormos1934)
Alilepus sp. in Schreuder (Reference Schreuder1937)
Alilepus sp. in Dietrich (Reference Dietrich1942)
Alilepus sp. ( = ? Α. dietrichi) in Fejfar (Reference Fejfar1961)
Alilepus sp. in Samson & Radulescu (1963)
Trischizolagus dumitrescuae n. g. n. sp. in Radulesco & Samson (Reference Radulesco and Samson1967a)
Occurrence: Late Miocene–Early Pliocene of Eurasia
Measurements: See Table 1
Sites: Bereşti and Măluşteni
Material: one hemimaxilla with two molariforms (UAIC NN-24), 19 upper molariforms (LPB[FGGUB] 245-5; LPB[FGGUB] 245-6.1; LPB[FGGUB] 245-6.2; LPB[FGGUB] 245-7.1 – Fig. 4J; LPB[FGGUB] 245-7.2; LPB[FGGUB] 245-8.1; LPB[FGGUB] 245-8.2; LPB[FGGUB] 245.9; LPB[FGGUB] 245-12; LPB[FGGUB] 107-13; LPB[FGGUB] 206; LPB[FGGUB] 211-2; LPB[FGGUB] 211-3; LPB[FGGUB] 211-4; LPB[FGGUB] 211-5; LPB[FGGUB] 211-6; LPB[FGGUB] 101-5; LPB[FGGUB] 101-7; UAIC NN18), four p3 (LPB(FGGUB) 245-1; LPB[FGGUB] 245-2; LPB[FGGUB]; 245-3; LPB[FGGUB] 245-4), one hemimandible bearing the p3–p4 (LPB[FGGUB] 101-8 – Fig. 4H), four hemimandibles bearing the p3–m1 (UAIC SM9b1; UAIC NN03; UAIC NN08; UAIC NN10), two hemimandibles bearing the p3–m2 (UAIC SM11; UAIC NN14), 10 hemimandibles bearing the p4–m2 (LPB[FGGUB] 108-1; LPB[FGGUB] 108-2; LPB[FGGUB] 100-2; LPB[FGGUB] 101-11; UAIC NN01; UAIC NN02; UAIC NN06; UAIC NN07; UAIC NN13; UAIC NN39), 22 hemimandibles with two molariforms (LPB[FGGUB] NN2; LPB[FGGUB] 107-4; LPB[FGGUB] 100-1; LPB[FGGUB] 100-3; LPB[FGGUB] 211-7; LPB[FGGUB] 101-1; LPB[FGGUB] 101-2; LPB[FGGUB] 101-3; LPB[FGGUB] 101-4; LPB[FGGUB] 101-6; LPB[FGGUB] 101-10; LPB[FGGUB] 101-15; LPB[FGGUB] 101-16; UAIC NN04; UAIC NN05; UAIC NN09; UAIC NN11; UAIC NN12; UAIC NN15; UAIC NN17; UAIC NN25; UAIC NN40), 16 lower molariforms (LPB[FGGUB] 245-10; LPB[FGGUB] 245-11; LPB[FGGUB] 100-4; LPB[FGGUB] 211-1; LPB[FGGUB] 211-8; LPB[FGGUB] 211-9; LPB[FGGUB] 101-9; LPB[FGGUB] 101-12; LPB[FGGUB] 101-14; LPB[FGGUB] 101-17; LPB[FGGUB] 101-18; UAIC NN16; UAIC NN19; UAIC NN20; UAIC NN22; UAIC NN23).
Description: See Radulesco & Samson (Reference Radulesco and Samson1967a).
Family Ochotonidae Thomas, 1897
Genus Ochotona Link, 1795
Ochotona ursui Simionescu, Reference Simionescu1930
Ochotona ursui n. f. in Simionescu (Reference Simionescu1930)
Ochotona ursui in Simionescu (Reference Simionescu1932)
Occurrence: Only Bereşti and Măluşteni, after Sen (Reference Sen2020)
Sites: Bereşti and Măluşteni
Material: One P3 (LPB[FGGUB] 245-13 – Fig. 4I), three hemimandibles bearing the p3–m3 (UAIC SM9b-2 – Fig. 4L; UAIC SM10-1; UAIC SM10-2 Fig. 4K), one hemimandible bearing the p3–m2 (LPB[FGGUB] 242-4), one hemimandible bearing the p4–m3 (LPB[FGGUB] 107-14), 15 hemimandibles bearing the p4–m2 (LPB[FGGUB] 107-1; LPB[FGGUB] 107-3; LPB[FGGUB] 106; LPB[FGGUB] 211-10; UAIC NN27; UAIC NN29; UAIC NN31; UAIC NN34; UAIC NN35; ISER 107-1; ISER 107-2; ISER 107-4; ISER 107-5; ISER 107-6; ISER 107-7), 14 hemimandibles with two molariforms (LPB[FGGUB]107-2; LPB[FGGUB] 107-5; LPB[FGGUB] 107-9; LPB[FGGUB] 211-11; LPB[FGGUB] 211-13; UAIC NN21; UAIC NN26; UAIC NN28; UAIC NN30; UAIC NN32; UAIC NN33; UAIC NN37; UAIC NN38; ISER 107-3), and 13 lower molariforms (LPB[FGGUB] 107-6; LPB[FGGUB] 107-7; LPB[FGGUB] 107-8; LPB[FGGUB] 107-10; LPB[FGGUB] 107-12; LPB[FGGUB] 242-1; LPB[FGGUB] 242-2; LPB[FGGUB] 242-3; LPB[FGGUB] 211-12; LPB[FGGUB] 211-14; LPB[FGGUB] 211-15; LPB[FGGUB] 101-13; UAIC NN36).
Measurements: See Table 1
Description:
P3: molariform tooth, having a trapezoidal shape. Its mesial hyperloph is more or less convergent to the distal hyperloph, and the precone ends in front of a small mesoflexus. The mesial hyperloph is straight. The paraflexus is poorly developed and it has a distinctive J-shape. A secondary inverted J-shaped enamel wall appears at the lingual side of the paraflexus and extends to the posterior side. This secondary paraflexus has well-developed enamel on the lingual side. A very shallow mesoflexus is present, marking the distinctive well-developed lagicone and a broken postlobus. A shallow and wide hypoflexus is present. Crown cement is present only at the secondary paraflexus. The lagicone is large and oval, and lacks the lagiloph and the centrocone. It is oriented in a lingual direction. The postlobus is damaged.
p3: the anteroconid is diamond-shaped and well-developed. The paraflexid and the protoflexid may be well-developed and connected to each-other (one out of three), or may appear in a narrow anterolophid (two out of three), which in turn separates the antero-, the proto-, and the metaconid from each-other. The anteroconid is larger than the metaconid. The anteroflexid is shallow. The metaconid and the protoconid are small and oval in shape. The hypoconid is more developed than the entoconid. The mesoflexid is very shallow, well-developed in one specimen, and the hypoflexid is long and narrow. The isthmid is wide, and serves as connection for the protoconid and the metaconid with the entoconid–hypoconid lobe.
Lower molariforms: the trigonid and the talonid are of the same width (trigonid width = taloned width), but the trigonid is longer. The trigonid appears to be massive when compared to the talonid, as well as prominent, and with a rounded labial side. The connection between the trigonid and the talonid is narrow. The talonid labial reentrants at the m1,2 is more developed than the ones at p4. The enamel is more developed in the posterior side of the trigonid and talonid.
m3: this tooth is formed by only one lobe; it may be similar to the trigonid of the m2, but smaller (two out of six), or it may be rounded (four out of six).
Remarks:
Two species of lagomorphs are present in the available material, and can be clearly separated by their size, among other characteristics. Much of the material found at the studied sites, described or listed in previous papers, is lost. The two present species were already differentiated by Simionescu (Reference Simionescu1930) for Măluşteni, although they were described as Lepus valdarnensis (now T. dumitrescuae), and O. ursui. Both species were also identified at the locality of Bereşti (Simionescu Reference Simionescu1932). Kormos (Reference Kormos1934), describes a p3 as Alilepus sp. Later, based on the material of L. valdarnensis from both localities, Radulesco & Samson (Reference Radulesco and Samson1967a) described the new genus and species, T. dumitrescuae, and synonymised the former p3 of Alilepus sp. to this new species. This species was amended by Averianov (Reference Averianov1995). Subsequently, Rădulescu & Samson (Reference Rădulescu and Samson1995) based on the larger size of the specimen, identified another species at Bereşti, Pliolagomys cf. gigas, using a mandible that had been previously classified as O. ursui. Unfortunately, this mandible has not been found among the material studied. Although it is not known from which materials the descriptions have been made, Samson & Radulesco (Reference Samson and Radulesco1973) also cite Proochotona gigas ( = Pl. gigas) for both sites, whereas Simionescu (Reference Simionescu1990) cites five species from Măluşteni: T. dumitrescuae, O. ursui, Lepus cf. timidus, Lagomys corsicanus, and Lagomys rumanus.
While T. dumitrescuae is a well-characterised species, and is found in more than one site (e.g., Averianov Reference Averianov1995), O. ursui presents the opposite case, and it does not have a good photographic record (Sen Reference Sen2020). Pidoplichko (Reference Pidoplichko1956; according to Samson & Radulesco Reference Samson and Radulesco1973), suggests that synonymy exists between O. ursui and Ochotona antiqua. Subsequently, Čermák & Rekovets (Reference Čermák and Rekovets2010) give three different possibilities: firstly, that O. ursui and O. antiqua are two different species; secondly that O. ursui is a subadult of Pliolagomys (these authors compared three mandibles without p3, but with their alveolus, that had been deposited in the Natural History Museum of Basel); or, lastly, that O. antiqua and O. ursui are the same species. The lectotype is not in the collections being studied, but instead they contain three p3s. From the three options listed above, the hypothesis of them belonging to a representative of the genus Pliolagomys, due to having a paraflexid, can be discarded, as, in the latter genus, the paraflexid tends towards the anterior part (Čermák Reference Čermák and Nowakowski2010) – this is not the case in the three studied specimens. One of the characters that can be used to discriminate O. antiqua from O. ursui is the size of the p3, as a similar morphology can be observed between both species; the remains under study have a larger size than in O. antiqua. There are specimens that fall within the size range of O. antiqua, and others are found to be larger; the P3R (p3 ratio = p3 length × 100/p3 width) is somewhat larger in O. ursui (104 in this species, and 96 in O. antiqua), although based on a poor sample (it may be altered by its low number of specimens). While O. ursui material has the MR (mandible ratio – buccal height of mandible at m2 × 100/alveolar length of p3–m3) between 83.29 and 89.97, the O. antiqua from the Novopetrovka site (Ukraine; Čermák & Rekovets Reference Čermák and Rekovets2010) has an MR of 66. Therefore, although both species are very similar in dental morphology, the mandibles of O. ursui are much more robust, confirming the first hypothesis, which supports that they are different species. Nonetheless, additional material from this species is needed for a clear conclusion to be drawn.
4. General discussion and conclusions
The site of Măluşteni is classically interpreted as being geologically younger than Bereşti due to the presence of the arvicolid Mimomys moldavicus, and Bereşti is interpreted to be older due to the presence of the cricetine Cricetulus simionescui (Samson & Radulesco Reference Samson and Radulesco1973). According to Terzea (Reference Terzea, Aguilar, Legendre and Michaux1997), and to Radulescu et al. (Reference Radulescu, Samson, Petculescu and Stiucă2003), although both sites have similar faunas, thanks to the presence of the above-mentioned arvicoline, Măluşteni could be dated as MN15a (final part of the Ruscinian), while Bereşti, taking into account the absence of the arvicoline and the presence of the cricetine, could be dated as MN14b (middle part of the Ruscinian). However, other authors, such as Simionescu (Reference Simionescu1990), group the faunas from both sites as late Early Pliocene, without specifying which of the two sites is older.
In our new review of the material (Table 2), both deposits can be grouped at a very similar age, within the Early Pliocene. This is because the cricetine and arvicoline genera present at both sites have a very broad biostratigraphical range, spanning the entire Pliocene. Thus, for Măluşteni, the presence of Desmana verestchagini indicates that this deposit accumulated during MN14, meaning that it belongs to the Early Pliocene, leaving Spermophilus cf. nogaici as the only taxon that does not fit within this age, as it belongs to the Late Pliocene (MN16). Nevertheless, more material would allow us to confirm or infirm the presence of this species, suggesting that the age of the deposits would be in the proximity of the MN14/15 transition. On the other hand, the site of Bereşti yielded specimens assigned to Trogontherium minus and Pliospalax macoveii, which allows us to place this site across MN14–15, a stratigraphical range that includes the entire Early Pliocene (Fig. 5).
Figure 5 Stratigraphical range of the species studied here, and the traditional and new hypothesis of the stratigraphical position of the sites Bereşti and Măluşteni. Neogene Mammal (MN) zones follow Mein (Reference Mein and Senes1975).
Table 2 Faunal list published by Simionescu (Reference Simionescu1930, Reference Simionescu1932), Kormos (Reference Kormos1940), Samson & Radulesco (Reference Samson and Radulesco1973), Radulescu & Samson (Reference Radulescu and Samson1989a, Reference Radulescu and Samsonb), Rădulescu et al. (Reference Rădulescu, Samson and Știucă1992) and Rădulescu & Samson (Reference Rădulescu and Samson1995) compared to the one published in this paper for small mammals; the presence of a ‘?’ indicates that the material previously assigned to that species was not found in the studied collections; the presence of a ‘–’ indicates the absence of this taxon in the site; when two cells in horizontal are joined together, it means the presence of that taxon in both sites; and when two or more upright cells are joined together, it signifies the synonymy of those taxa to a single taxon in the new taxonomic assignment.
1 Material lost after Samson & Radulesco (Reference Samson and Radulesco1973).
2 Cited in Simionescu (Reference Simionescu1990).
3 Pliolagomys cf. gigas after Rădulescu & Samson (Reference Rădulescu and Samson1995), but only present in Bereşti.
In other Romanian and Ukrainian sites of similar age, the presence of sites belonging to MN15 is abundant (Ciuperceni 1 and 2, Drănic 1, 2, and 3, Lupoaia 8, and Odessa Catacombs, among others; Vislobokova et al. Reference Vislobokova, Erbajeva and Sotnikova1993; Rădulescu et al. Reference Rădulescu, Samson, Sen, Stiucă, Horoi, Aguilar, Legendre and Michaux1997; Terzea Reference Terzea, Aguilar, Legendre and Michaux1997; Rădulescu & Samson Reference Rădulescu and Samson2001), the sites belonging to MN14 are very scarce (only Drănic 0, MN14-15; Rădulescu & Samson Reference Rădulescu and Samson2001). In general, the fauna studied here is characterised by a great diversity and/or abundance of talpids, lagomorphs, beavers, and spalacids, with scarce cricetines, arvicolids, sciurids, and other insectivores. Something similar occurs in Ciuperceni 2 (Terzea Reference Terzea, Aguilar, Legendre and Michaux1997), while in other sites such as Drănic 0, 1, 2, and 3, there is a great diversity of soricids and arvicolids, while the murines are present, although lagomorphs and beavers are scarcer (Rădulescu et al. Reference Rădulescu, Samson, Sen, Stiucă, Horoi, Aguilar, Legendre and Michaux1997). In the other hand, the Odessa Catacombs site would have a combination of both types of sites, there neither soricids nor talpids are present, while there is a great diversity of lagomorphs, beavers, cricetines, and murines, but there are only a few species of arvicolids and spalacids (Vislobokova et al. Reference Vislobokova, Erbajeva and Sotnikova1993). This difference may be due to three reasons: for palaeoecological reasons – different types of the environments in which the sediments were deposited; for biostratigraphical reasons – all the sites are more modern than those studied here; or for methodological reasons – many of the missing species are small in size, and there may have been a bias in collecting the material. For this reason, the study of Bereşti and Măluşteni is very important, as it is a unique example for understanding the Early Pliocene small mammal faunal assemblages in the northern Black Sea region.
5. Acknowledgements
The authors thank Ana Drob (Arheoinvest, Iași) and Sergiu Loghin (‘Alexandru Ioan Cuza’ University of Iași) for the help provided during specimen scanning electron microscope and trinocular photographing. We also thank Sevket Sen and an anonymous reviewer for their invaluable comments that helped improve the quality of the original manuscript.
6. Financial support
This work was supported by two grants of the Romanian Ministry of Research and Innovation, CNCS – UEFISCDI, project numbers PN-III-P1-1.1-TE-2021-0664 (Ș.V., B.G.R., B.S.H.) and PN-III-P4-ID-PCE-2020-2282 (Ș.V.), both grants within PNCDI III. V.D.C. thanks the Stimulus of Scientific Employment, Individual Support – 2021 Call grant by the Fundação para a Ciência e a Tecnologia (Portugal, CEECIND/03080/2021) and GeoBioTec. A.P. acknowledges funding by the Romanian Research Authority (UEFISCDI) through grants PCCF 16/2016 (DARKFOOD), PCE 2282/2020 (ECHOES), and EEA Grant 126/2018 (KARSTHIVES).
7. Competing interests
None.
