https://orcid.org/0000-0002-1812-4664
Figures
Abstract
The hornless rhinocerotid genus Chilotherium is one of the most common rhinoceroses found in Eurasia during the Late Miocene, even being the most abundant large mammal in some localities. Despite its high frequency, much is still unknown about the anatomy and taxonomy of some representatives of the group, because most studies focus solely on cranial and dental remains. For this reason, we describe the postcranial material of the three well-known chilothere species Chilotherium persiae from the Upper Miocene of Maragheh (Iran), Chilotherium habereri from the Upper Miocene of Kutschwan (China), and Chilotherium schlosseri from the Upper Miocene of Samos (Greece). A comparison to postcranial material of other chilotheres from the literature reveals some characters that can assist the identification of some species, such as the morphology and dimensions of the patella and the metapodials, the morphology of the proximal articulation of tibia, and the arrangement of the articular facets for the calcaneum on the astragalus. The comparison supports previous hypotheses about the relationships of some chilotheres that were only based on cranial features, including the plesiomorphic nature of ‘Chilotherium’ wimani from the Upper Miocene of China and the close relationship of C. persiae and C. habereri.
Citation: Kampouridis P, Svorligkou G, Spassov N, Böhme M (2025) Postcranial anatomy of the Late Miocene Eurasian hornless rhinocerotid Chilotherium. PLoS One 20(12): e0336590. https://doi.org/10.1371/journal.pone.0336590
Editor: Luca Pandolfi, Universita degli Studi della Basilicata, ITALY
Received: July 24, 2025; Accepted: October 28, 2025; Published: December 5, 2025
Copyright: © 2025 Kampouridis et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the manuscript and its Supporting Information files.
Funding: This research received support from the SYNTHESYS+ project http://www.synthesys.info/ which is financed by European Community Research Infrastructure Action under the H2020 Integrating Activities Programme, Project number 823827 (AT-TAF-TA3-9 and AT-TAF-TA4-29 and FR-TAF-TA4-71 granted to PK). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Rhinoceroses are amongst the most iconic large mammals to roam the earth today. Even though only five extant species exist, in the past this group was much more diverse [1–6]. During the Miocene a plethora of rhinoceros species existed and, although the extant representatives are characterised by the presence of horns on their heads, a highly diverse rhinocerotid group were the aceratheriines, also known as the hornless rhinos. One of the most characteristic and species-rich genera of this group was Chilotherium Ringström, 1924 [7] that lived in Asia and Eastern Europe during the Late Miocene [1,8,9]. This genus is characterised by a very wide mandibular symphysis with two large, tusk-like incisors, flat to concave frontal bones, very complex enamel folds in the upper teeth, and the appendicular skeleton is notably shortened and relatively robust, especially the metapodials [7,10].
Ringström (1924) [7] initially attributed over ten species to this newly erected genus, while also erecting several new species, based on material from the Upper Miocene of China. He discussed that the representatives of Chilotherium were extremely similar and that they could only be separated based on the morphology of the upper teeth. However, it has been shown that the morphology of the skull over all is more informative for the taxonomic attribution than the tooth morphology [e.g., 10,11]. Some studies actually suggested that also the postcranial anatomy of some species can be used for their distinction [11,12]. However, only few studies have reported postcranial material for members of this group [e.g., 7,13–15] and even less studies described and analysed this material in detail [11,16]. This creates a large gap in our understanding of the body plan of this rhino group and the potential interspecific variability. However, it was suggested early on that the appendicular skeleton of the genus Chilotherium is very derived [7], which was also confirmed by later studies, suggesting that the limb bones of chilotheres differ from those of other aceratheriines [3,16]. Korotkevitch [11], Krokos [17], and Deng [16] studied large samples of some chilothere species, suggesting that some elements, such as the metapodials, are also important for taxonomic purposes within the group. These findings have not been verified or re-evaluated by any later study, making some assumptions somewhat ambiguous.
The aim of this study is to evaluate the postcranial material of Chilotherium spp. available in different institutions to provide some insight into the taxonomy and interrelationships of this group. For this purpose, we studied collections of three historical fossil localities, Maragheh (Iran), Kutschwan (China), and Samos (Greece), that are known to include Chilotherium in their fauna (Fig 1). The studied specimens were compared with available material in the literature from Europe and Asia (Fig 1), thereby covering almost the complete geographical distribution of this taxon and offering new insight into the anatomy and taxonomy of these animals.
The localities where the herein studied material comes from (Samos in Greece, Maragheh in Iran, and Kutschwan in China) are underlined. The map was generated using Generic Mapping Tools 6 (GMT6) [18].
Abbreviations
Anatomical abbreviations
APD, anteroposterior diameter; APDbec, anteroposterior diameter measured at the level of the posterior astragalar facet; APDcaput, anteroposterior diameter of the humerus head; APDdia, anteroposterior diameter of the diaphysis; APDdist, distal anteroposterior diameter; APDdist art, anteroposterior diameter of the distal articular surface; APDinf, medial anteroposterior diameter; APDolecr, anteroposterior diameter of the olecranon; APDprox, proximal anteroposterior diameter; APDprox art, anteroposterior diameter of the proximal articular surface; APDsommet, anteroposterior diameter of the tuber calcis; DL, distance between the lips of the trochlea; TD, transversal diameter; TDcaput, transversal diameter of the humerus head; TDdia, transversal diameter of the diaphysis; TDdist artic, transversal diameter of the distal articular surface; TDdist max, maximal distal transversal diameter; TD maxi dist, maximal distal transversal diameter; TDmini post, minimal transversall diameter; TDolecranon, transversal diameter of the olecranon; TDprox art, transversal diameter of the proximal articular surface; TDprox max, maximal proximal transversal diameter; TDsommet, transversal diameter of the tuber calcis; TDsust, maximal transversal diameter at the sustentacular tali; H, height; Hant, anterior height; Hface ant, height of anterior facet; H lateral condyle, height of the lateral distal condyle; H medial condyle, height of the medial distal condyle; Hpost, posterior height measured between the tubers calcis and the posterior astragalar facet; Htuberosity, height of the lateral tuberosity; L, length; Lmax, maximal length; L absolute, maximal length; L anatom, length in anatomical position; Lart sup, length of the anterior articular surface; Lart inf, length of the central articular surface; Lcaput, length measured from the humerus head; l, width; lart, width of articular surface; and lart sup, width of anterior articular surface.
Institutional abbreviations
AMNH, American Museum of Natural History, New York (USA); AMPG, Athens Museum of Palaeontology and Geology of the National and Kapodistrian University of Athens (Greece); BSPG, Bayerische Staatssammlung für Paläontologie und Geologie, Munich (Germany); GMM, Geomuseum of the University of Münster (Germany); GPIT, Geologisch-Paläontologisches Institut der Universität Tübingen (Germany); GPIH, Geologisch-Paläontologisches Institut der Universität Hamburg (Germany); HLMD, Hessisches Landesmuseum Darmstadt (Germany); IPUW, Institut für Paläontologie, Universität Wien (Austria); MLU, Martin-Luther Universität Halle-Wittenberg (Germany); MNHN, Muséum national d’Histoire naturelle, Paris (France); NHMW, Naturhistorisches Museum in Wien (Austria); NMB, Naturhistorisches Museum Basel (Switzerland); and SMNS, Staatliches Museum für Naturkunde Stuttgart (Germany); and SNSB, Staatliche Naturwissenschaftliche Sammlungen Bayerns (Germany).
Materials and methods
For the purpose of this study, material of three distinct chilotheres was studied: Chilotherium persiae (Pohlig, 1885) [19] from the Upper Miocene deposits of Maragheh in Iran [13,20–23], Chilotherium habereri (Schlosser, 1903) [24] from the Upper Miocene deposits of Kutschwan in China [25,26], and Chilotherium schlosseri (Weber, 1905) [14] from the Upper Miocene deposits of Samos in Greece [8,27–29] (Fig 1). The material studied herein is housed in different palaeontological collections (GMM, GPIT, IPUW, MNHN, NHMW, and MLU).
Additionally, the studied material is compared to published data from the species ‘Chilotherium’ wimani Ringström, 1924 [7] from the Upper Miocene of China [16], Chilotherium kowalevskii (Pavlow, 1913) [15], Chilotherium sarmaticum Korotkevitch, 1958, both from the Upper Miocene of Ukraine [11,15,17], and Chilotherium orlovi Bayshashov, 1982 [30] from the Upper Miocene of Pavlodar in Kazakhstan [12,30].
Measurements were taken according to Guérin (1980) [31] with some additions. Detailed measurements of all bones from the literature and those measured herein are provided in S1 Tables 1–23 and throughout the main text in the Tables. The anatomical terms used in this study are based on the literature [31–35]. It has to be noted that in the older literature [7,11,12,14,17] no standardised measuring protocol like that of Guérin (1980) [31] has been used. Therefore, it is difficult to make certain and detailed comparisons with some of the species. This is especially true for complex bones like carpals and tarsals, where it is even more difficult to compare measurements if they did not follow the exact same method of measuring.
Taxonomic notes
The taxonomy of chilotheres has been a rather complicated matter for a long time [8,10,17,36–38]. The genus was erected over 100 years ago [7] and over a dozen species have been assigned to it over the years, with many attributions not being valid anymore (see [7,37–39]). Especially the European representatives have been the centre of great confusion, with the erection of overall 10 species, two of which come from Portugal and in fact belong to the elasmotheriine rhinoceros Hispanotherium [37,38]. All other European chilotheres come from Eastern and Southeastern Europe and are validly assigned to the chilotheres. The taxonomy of these other eight species, while representing chilotheres, has been a complicated matter to this day. From the Upper Miocene of Samos alone, four species have been erected, two of which, Chilotherium wegneri (Andree, 1921) [40] and Chilotherium angustifrons (Andree, 1921) [40], have subsequently been synonymised to C. schlosseri [8]. Chilotherium schlosseri and the fourth species, Eochilotherium samium (Weber, 1905), were recently revised and separated on a generic level, with both species belonging to the subtribe Chilotheriina (sensu [10]). The other chilothere species from Greece is Aceratherium kiliasi Geraads and Koufos, 1990 [41], which was erected based on craniomandibular material from the Upper Miocene of Pentalophos in northern Greece. The material from Pentalophos belongs in part to the non-chilothere aceratheriine Acerorhinus neleus Athanassiou et al., 2014 [42]. The holotype and some other specimens, do belong to a chilothere, but represent the species E. samium, making Aceratherium kiliasi a junior synonym of this species [10,42]. The other European chilotheres were all erected based on material from Ukraine. A partial skull from Upper Miocene deposits near Odessa represents the holotype of Teleoceras ponticus Lubicz-Niezabitowski, 1912 [43,44]. Shortly after its’ erection it was suggested to represent a junior synonym of C. schlosseri [36] and is still considered as such [8]. The other two species are Chilotherium kowalevskii (Pavlow, 1913) [15] from the Upper Miocene of Grebeniki and Chilotherium sarmaticum Korotkevitch, 1958 [45] from the Upper Miocene of Berislav (Ukraine). The first of these two has repeatedly been suggested to be a junior synonym of C. schlosseri [11,17,46,47]; however, recent re-evaluations of the Samian chilotheres showed that the two species differ both in their cranial and dental morphology [8,10]. The species C. schlosseri features widely separated parietal crests, with a minimal distance of at least 70 mm (n = 7), whereas the material of C. kowalevskii from Grebeniki exhibits a value range of 40–66 mm (n = 10) for this measurement, thus clearly differing from C. schlosseri. Chilotherium sarmaticum, on the other hand, has received little attention after its initial description and was usually neglected in most comparative studies [8,48]. However, it was shown that C. sarmaticum differs from both C. kowalevskii from Grebeniki and C. schlosseri from Samos [10,11]. Therefore, both C. kowalevskii and C. sarmaticum are herein considered valid species. Regarding the Chilotherium species from China, there are several species that were attributed to the genus by Ringström (1924) [7], including some newly erected species and subspecies. Some of these cannot be considered as valid species, while other do not belong to the genus Chilotherium. Several studies have tried to bring order into this matter [e.g., 37,38]. Heissig (1975) [38] attempted to distinguish different lineages within the genus Chilotherium by proposing subgenera. Only few researchers adopted this scheme [e.g., 48] and it is not used anymore. More recent work on the rhinocerotids from China has shed new light on the issue, separating distinct genera, and for instance erecting new chilothere species, like ‘Chilotherium’ primigenium Deng, 2006 [37] and Chilotherium licenti Sun et al., 2018 [49]. Additionally, it was recently discussed that the two more ‘primitive’ species ‘C.’ primigenium [37] and ‘C.’ wimani [7] may in fact not belong to the genus Chilotherium. It was discussed that they lack the characteristic depression of the frontal region and exhibit some other more plesiomorphic characters that may associate these two species closer with E. samium instead [10]. A detailed re-evaluation of the type material is needed to clarify the generic status of these species. However, for the time being both species will be referred to as ‘C.’ primigenium and ‘C.’ wimani.
Fossil sites
Maragheh.
The locality of Maragheh is situated in northwestern Iran (Fig 1). The Late Miocene fauna is well known since the late 19th century [19,50,51]. Several fossil sites in the area have been excavated over the last century [13,23,50,52,53], recent studies suggested the existence of three distinct fossiliferous horizons and were able to correlate the old fossil sites with these three horizons [21,22]. Material from Maragheh is scattered throughout many collections, such as the MLU, MNHN, and NHMW. The rhino assemblage in Maragheh is quite rich and includes the huge elasmotheriine Iranotherium morgani (Mecquenem, 1908) [50] the large tandem-horned ‘Ceratotherium’ neumayri (Osborn, 1900) [54], and two smaller aceratheriines, Chilotherium persiae and Persiatherium rodleri Pandolfi, 2016 [20]. The last species, was only recently described and is mainly known from its holotype – a subadult skull [20]. Maragheh is the type locality of all four of these rhinocerotid species and the first two are also well known from other localities in Eurasia, showing the importance of the locality for rhinocerotid systematics. Chilotherium persiae is by far the most common representative of the family in the Maragheh fauna and a rich collection of postcranial material is available for this species. Mecquenem (1924) [13] was the first to provide a more detailed description of C. persiae and while shortly describing also some postcranial elements he mainly focussed on the cranial and dental material. The studied material comes from different fossil sites situated in the area north of Maragheh, including ‘Ketschawa’ or Karajabad, Kara Kend, and Kopran, which represent the middle and lower fossiliferous beds of Maragheh [55,56].
As mentioned above, four rhinocerotid species existed in Maragheh and during the course of this project postcranial elements not only of C. persiae but also of the other species were found. The horned rhinoceroses ‘Ceratotherium’ neumayri and Iranotherium morgani are much larger and their postcranial bones are easily distinguishable from Chilotherium. The other hornless species, Persiatherium rodleri, is basically only known from a single skull, apart from it only some isolated dental and fragmentary postcranial elements from the Upper Miocene of Küçükçekmece (Turkey) have been attributed to Persiatherium sp. [46]. However, they are not particularly helpful in distinguishing the species. Based on the phylogenetic analysis of Lu et al. (2023) [57], Persiatherium rodleri is placed in Aceratheriinae, but distinct from Chilotheriina, which is the clade made up by Chilotherium and Shansirhinus in Lu et al. (2023) [57], although not referred to as such by the authors. This further supports the notion that Persiatherium rodleri should not exhibit the shortened limbs that are known in Chilotherium. All herein reported specimens, fit both metrically and morphological the known postcranial elements of Chilotherium.
Kutschwan.
The locality of Kutschwan in China is of Late Miocene age and was discovered by the German geographer and physician Albert Tafel in Shanxi during his trip to China in 1905 [25]. No information about the exact location of the fossil site is known, but the material was collected from horizontal red clay deposits close to the Yellow River (Huang He) in Shanxi (China, Fig 1) [25]. The material was excavated by A. Tafel, comprising of many cranial and postcranial elements. The whole collection was initially prepared at the SMNS (Germany) before being deposited in the collections of the GPIT (Germany) and is still housed there today. Hugo Killgus studied the whole collection for his Ph.D. Dissertation [26] and recognised a quite rich mammalian fauna, based on this limited material, including an ictithere hyaena [58], two rhinocerotids, Chilotherium habereri, Parelasmotherium schansiense Killgus, 1923 [25], a hipparionine horse, the giraffe Schansitherium tafeli Killgus, 1923 [25], one large and up to three small bovids [25]. More recently, Kampouridis et al. (2022) [59] reassessed the taxonomy and phylogeny of Parelasmotherium schansiense by studying its holotype.
The chilothere C. habereri is the most abundant taxon in this collection, including several skulls and a small sample of postcranial material, which cannot be associated with the huge elasmotheriine. The cranial and dental morphology of the species were revised by Ringström (1924) [7] after the original description [24], but not much is known about the postcranial anatomy of the species. The herein reported postcranial elements can only be assigned to the single aceratheriine present in Kutschwan, C. habereri.
Samos.
The island of Samos (Fig 1) has been known to yield Late Miocene vertebrate fossils since the 19th century [60]. Since then, numerous palaeontologists and fossil hunters have travelled to the island to collect fossils [60]. This led to several impressive collections of Samos material in famous natural history museums throughout the world, including among others the AMNH, NHMW, and SMNS. Most recently, the Aristotle University of Thessaloniki, led by Prof. George Koufos, excavated on Samos Island, bringing to light a rich collection of mammalian remains [60,61]. The material was studied in detail, providing crucial information about the (bio-)stratigraphical context and the fauna itself [e.g., 28,29,61,62]. The rhinocerotid material from Samos has been assigned to several different taxa over the years [8,10,27]. Today, it is generally accepted that two large tandem-horned rhinocerotines are present, ‘Ceratotherium’ neumayri and Dihoplus pikermiensis (Toula, 1906) [63], along with two small aceratheriines, Chilotherium schlosseri and Eochilotherium samium [10,14,27,28,64]. The taxonomy of the Samian chilotheres has experienced many difficulties in the past, with the suggested presence of five different species depending on the authors – C. schlosseri, E. samium, C. wegneri, C. angustifrons, and C. kowalevskii [e.g., 8,14,17,27,40,42]. Most recently, Kampouridis et al. (2023) [8,10] suggested that the only valid chilothere species in Samos are C. schlosseri and E. samium, with C. wegneri and C. angustifrons representing junior synonyms of C. schlosseri. The species E. samium is much rarer than C. schlosseri and is currently only known from a single skull (SMF M 3601) with its associated mandible from the locality of Samos.
No postcranial material is assignable to E. samium, whereas C. schlosseri is represented by a large amount of material including, an impressive number of skulls [14,40,65–67] but also postcranial elements, which differ significantly from the much larger ‘Ceratotherium’ neumayri and Dihoplus pikermiensis. This postcranial material is morphologically and metrically much closer to that of typical chilotheres, such as C. kowalevskii, C. anderssoni Ringström, 1924 [7], and C. orlovi, than to the more plesiomorphic ‘C.’ wimani (see also below); therefore, an assignment to E. samium is excluded.
Systematic Palaeontology
Class Mammalia Linnaeus, 1758 [68]
Order Perissodactyla Owen, 1848 [69]
Family Rhinocerotidae Gray, 1821 [70]
Subfamily Aceratheriinae Dollo, 1885 [71] (sensu Lu et al. (2023) [57])
Tribe Aceratheriini Dollo, 1885 [71] (sensu Lu et al. (2023) [57])
Subtribe Chilotheriina Qiu et al., 1987 [72] (sensu Kampouridis et al. (2023) [10])
Included genera
Chilotherium Ringström, 1924 [7], Shansirhinus Kretzoi, 1942 [73], and Eochilotherium Geraads and Spassov, 2009 [48].
Diagnosis
Aceratheriine rhinocerotids that feature the following autapomorphic traits: separated parietal crests; upper molars featuring a marked protocone constriction and a moderate to strong antecrochet that tends to project lingually; a mandible that is characterised by a very wide mandibular symphysis with a flat to concave ventral surface; large, sexually dimorphic i2s that exhibit a trend to become more flattened and tusk-like, especially in males, and are separated from each other by a wide diastema and from the p2 by a long diastema with a marked crest. The group also exhibits an upper first deciduous premolar retained into adulthood, whereas the first lower deciduous premolar, when present, is shed and not replaced by a permanent one. Additionally, the group lacks any upper incisor and the upper cheek teeth have generally pronounced secondary enamel folds (i.e., crista, crochet, and antecrochet), and the appendicular skeleton is notably shortened and relatively robust, especially the autopodium (modified after [10]).
Remarks
The clade Chilotheriini was originally established by Qiu et al. (1987) [72] as a tribe, to encompass the genera Acerorhinus and Chilotherium, specifically excluding the genus Aceratherium. However, the relationship between these two genera is not clear, as shown in the contradicting results of some recent phylogenetic analyses [20,57]. Therefore, it has not been possible to prove that these two genera form a monophyletic clade. In fact, Acerorhinus shows many plesiomorphic characters and none of the apomorphic features characterising representatives of Chilotherium, therefore a close relationship cannot be supported. Aceratherium shows characters that are more similar to Chilotherium, like the somewhat shortened limb bones [74], whereas in Acerorhinus zernowi the plesiomorphic state of more elongated limb bone is seen [75]. Another more enigmatic hornless rhinocerotid, Shansirhinus, is much more similar to Chilotherium in many regards, such as the wide mandibular symphysis, the more complicated enamel folds, and the higher tooth crowns [76]. Their probably closer relationship has also been suggested in recent phylogenetic analyses [20,57]. It would therefore be best to remove Acerorhinus from Chilotheriina and to restrict this group to the genera Chilotherium, Shansirhinus, and Eochilotherium, with the latter one having been regarded as belonging to Chilotherium sensu stricto up until recently [8,10,27].
The type genus of Chilotheriini was not designated as such by Qiu et al. (1987) [72]; however, based on ICZN Arts. 29.1 and 64, a suprageneric name must derive from its type genus, which would naturally make Chilotherium the type genus. Following, the recommendation of Kampouridis et al. (2023) [10], we herein use the clade at a subtribe rank (as Chilotheriina) with Chilotherium as the type genus and also including the genera Shansirhinus and Eochilotherium.
Genus Chilotherium Ringström, 1924 [7]
Included species
Chilotherium persiae (Pohlig, 1885) [19], Chilotherium habereri (Schlosser, 1903) [24], Chilotherium schlosseri (Weber, 1905) [14], Chilotherium kowalevskii (Pavlow, 1913) [15], ‘Chilotherium’ wimani Ringström, 1924 [7], Chilotherium sarmaticum Korotkevich, 1958 [45], Chilotherium orlovi Bayshashov, 1982 [30], ‘Chilotherium’ primigenium Deng, 2006 [37], and Chilotherium licenti Sun et al., 2018 [49].
Diagnosis
Aceratheriine rhinocerotids that feature the following autapomorphic characters: flat and wide skull, flattened and depressed frontal region; well-developed postorbital processes; moderately to widely separated parietal crests; highly placed orbits; very wide mandibular symphysis that features a concave ventral side; very large, flattened, tusk-like second lower incisors, with a scalene triangle cross section and upturned, dorsomedially oriented wear facets; reduced premaxillary bones that lack upper incisors; and very strong secondary enamel folds, including a lingually flattened and strongly constricted protocone in the molars. It is also characterised by a relatively short length of the premolars compared with the molars, mainly due to the reduced size of the P2 and p2; and notably shortened metapodials and relative robust appendicular skeleton (modified after [7,10,27,39,48]).
Remarks
Recently, it was shown that ‘C.’ wimani and ‘C.’ primigenium exhibit more plesiomorphic features than the other Chilotherium species [10]. In fact, these two species seem to share some morphological features with Eochilotherium samium, such as the flat to convex frontal region, the thickness of the nasal bones, and the highly elevated nuchal crest It seems most plausible that they do not belong to the genus Chilotherium; however, it is preferred not to include these species into the genus Eochilotherium at the moment. Nonetheless, until the issue about their generic attribution is resolved it is preferable to keep them in their original genus as ‘C.’ wimani and ‘C.’ primigenium [10,39].
Chilotherium persiae (Pohlig, 1885) [19]
Type material
Pohlig (1885) [19] erected the species without fixing a holotype, providing any specimen numbers, photographs, or any information about the collection which housed the material he studied. He mentioned that the material he collected himself in Maragheh (Iran) was sent to Halle (Germany) [51]. This material is now housed in the palaeontological collection of the MLU. Pohlig (1885, 1886) [19,51] specifically mentioned that there were four adult and one juvenile skull that he assigned to this species. He did not specify whether these were collected by himself or if he saw them in some other collection. In the collections of the MLU only one adult and two juvenile skulls of C. persiae were able to be relocated. However, Pohlig (1886) [51] mentioned explicitly the fact that the collection of fossils in the area was continued by Theodor Strauss, as explained also by Rodler (1885) [77], who joined the excavation in Maragheh, collecting for the NHMW. The fact that Pohlig was well aware of these excavations and also specifically mentioned the material housed at the NHMW [51,78], along with the fact that several skulls were deposited in the collection of the NHMW until 1885, make it likely that Pohlig had seen at least some of these skulls and counted them into the, in total, five skulls he mentioned [51]. Until further information becomes available, however, the type material cannot be determined with certainty.
Diagnosis
Medium-sized Chilotherium species with a weakly to moderately depressed frontal region, straight and relatively thick nasal bones, parietal crests that are moderately separated from each other (minimal distance between parietal crests up to 54 mm, n = 11) and a unique combination of dental characters: M3 with a somewhat quadrangular outline; long crochet; small crista sometimes present in the premolars that may close off the medifossette; very strong mesial and distal constriction of the protocone, which is lingually flattened, resulting in a very long antecrochet that may close off the median valley; a strong mesial constriction of the hypocone; crista rarely present in molars but when present may close off the medifossette; and a weak, discontinuous lingual cingulum mainly present at the entrance of the median valley in the premolars.
Referred material
An axis (MNHN.F.MAR3939), five scapulae (MNHN.F.MAR1431, MNHN.F.MAR1433, MNHN.F.MAR2899, MNHN.F.MAR3898, and MNHN.F.MAR3900), 11 humeri (MNHN.F.MAR3901, MNHN.F.MAR3903, MNHN.F.MAR3904, MNHN.F.MAR3905, NHMW-GEO-2020/0014/0115, NHMW-GEO-2020/0014/0122, NHMW-GEO-2020/0014/0123, NHMW-GEO-2020/0014/0124, NHMW-GEO-2020/0014/0125, NHMW-GEO-2020/0014/0148, and NHMW-GEO-2020/0014/0147), 17 radii (MNHN.F.MAR1434, MNHN.F.MAR3906, MNHN.F.MAR3908, MNHN.F.MAR3909, MNHN.F.MAR3911, MNHN.F.MAR3912, MNHN.F.MAR3913, MNHN.F.MAR3963, NHMW-GEO-2020/0014/0107, NHMW-GEO-2020/0014/0108, NHMW-GEO-2020/0014/0109, NHMW-GEO-2020/0014/0110, NHMW-GEO-2020/0014/0111, NHMW-GEO-2020/0014/0126, NHMW-GEO-2020/0014/0127, NHMW-GEO-2020/0014/0128, NHMW-GEO-2020/0014/0149), a scaphoid (MNHN.F.MAR1412), two semilunars (MNHN.F.MAR1401 and MNHN.F.MAR1411), two pyramidals (MNHN.F.MAR1405, MNHN.F.MAR1413), a trapezoid (MNHN.F.MAR1409), six unciforms (MNHN.F.MAR1400, MNHN.F.MAR1403, MNHN.F.MAR1406, MNHN.F.MAR1408, NHMW-GEO-2020/0014/0140, and NHMW-GEO-2020/0014/0155), four second metacarpals (MNHN.F.MAR1375, MNHN.F.MAR1388, NHMW-GEO-2020/0014/0106, and NHMW-GEO-2020/0014/0142), five third metacarpals (MNHN.F.MAR1377, MNHN.F.MAR1379, MNHN.F.MAR1383, MNHN.F.MAR1429, and MLU.GeoS.6242), six fourth metacarpals (MNHN.F.MAR1376, MNHN.F.MAR1387, MNHN.F.MAR1390, MNHN.F.MAR1386, MLU.GeoS.6241, and NHMW-GEO-2020/0014/0143), seven femora (MNHN.F.MAR1416, MNHN.F.MAR1432, MNHN.F.MAR3920, MNHN.F.MAR3921, NHMW-GEO-1911/0005/0273, NHMW-GEO-2020/0014/0114, and NHMW-GEO-2020/0014/0152), 11 patellae (MNHN.F.MAR1463, MNHN.F.MAR3922, MNHN.F.MAR3923, MNHN.F.MAR3924, MNHN.F.MAR3925, MNHN.F.MAR3926, MNHN.F.MAR3927, MNHN.F.MAR3928, MLU.GeoS.6245, NHMW-GEO-2020/0014/0118, and NHMW-GEO-2020/0014/0130), 10 tibiae (MNHN.F.MAR1435, MNHN.F.MAR3931, MNHN.F.MAR3932, MNHN.F.MAR3933, MNHN.F.MAR3990a, MNHN.F.MAR4131, MNHN.F.MAR4132, NHMW-GEO-2020/0014/0112, NHMW-GEO-2020/0014/0113, and NHMW-GEO-2020/0014/0117), 14 astragali (IPUW-MFN21652, MNHN.F.MAR1360, MNHN.F.MAR1366, MNHN.F.MAR1367, MNHN.F.MAR1368, MNHN.F.MAR1369, MNHN.F.MAR1370, MNHN.F.MAR1371, MNHN.F.MAR1372, MNHN.F.MAR1373, MNHN.F.MAR1419, NHMW-GEO-2020/0014/0102, NHMW-GEO-2020/0014/0129, and NHMW-GEO-2020/0014/0137), 19 calcanei (MNHN.F.MAR1421, MNHN.F.MAR1422, MNHN.F.MAR1423, MNHN.F.MAR1424, MNHN.F.MAR1425, MLU.GeoS.6243, MLU.GeoS.6244, NHMW-GEO-2020/0014/0103, NHMW-GEO-2020/0014/0104, NHMW-GEO-2020/0014/0105, NHMW-GEO-2020/0014/0131, NHMW-GEO-2020/0014/0132, NHMW-GEO-2020/0014/0133, NHMW-GEO-2020/0014/0134, NHMW-GEO-2020/0014/0138, NHMW-GEO-2020/0014/0139, NHMW-GEO-2020/0014/0145, NHMW-GEO-2020/0014/0146, and NHMW-GEO-2020/0014/0151,), six naviculars (MNHN.F.MAR1407, MNHN.F.MAR1415, MNHN.F.MAR1427a, NHMW-GEO-2020/0014/0136, NHMW-GEO-2020/0014/0145, and NHMW-GEO-2020/0014/0156), three cuboids (MNHN.F.MAR1414, MNHN.F.MAR1427b, and NHMW-GEO-2020/0014/0145), three ectocuneiforms (MNHN.F.MAR1399, MNHN.F.MAR1427c, and NHMW-GEO-2020/0014/0145), three second metatarsals (MNHN.F.MAR1378, MNHN.F.MAR1381, and MNHN.F.MAR1385), three third metatarsals (MNHN.F.MAR1382, NHMW-GEO-2020/0014/0150, and NHMW-GEO-2020/0014/0154), and four fourth metatarsals (MNHN.F.MAR1452, NHMW-GEO-2020/0014/0121, NHMW-GEO-2020/0014/0144, and NHMW-GEO-2020/0014/0154).
Remarks
Chilotherium persiae, originally erected as Rhinoceros persiae, represents the first described rhinoceros species that was later included in the genus Chilotherium [7,51]. Maragheh, the type locality of the species, has brought to light a very diverse mammalian fauna based on a large amount of fossil bones [21,22,53,79]. Along with many skulls, the species C. persiae is represented also by a large number of postcranial elements housed in different institutions, including the MLU, MNHN, and NHMW. These fossils were excavated at different fossil sites in the area around Maragheh that correspond to different stratigraphical layers [23,e.g., 51,77]. More specifically the studied postcranial material comes from the sites Karajabad, Kara Kend, and Kopran that represent the middle and lower fossiliferous beds in Maragheh [21,55,56].
Chilotherium habereri (Schlosser, 1903) [24]
Lectotype
Schlosser (1903) [24] assigned several teeth to the new species that he erected, Rhinoceros habereri, without assigning a holotype. Therefore, all studied teeth (Schlosser, 1903: plate 5, figs 5–10, 12–21; plate 7, figs 1–3, 6, 8, 10, 11) [24] constitute the syntype of Chilotherium habereri. The associated P3 and P4 (SNSB- BSPG 1900 XII 622), illustrated by Schlosser (1903, plate 5, fig 18) [24], are herein designated as the lectotype of the species under the provisions of ICZN Art. 74. These were also used as the basis for the identification of the species by Ringström (1924) [7], who revised its morphological affinities.
Diagnosis
Medium- to large-sized Chilotherium species with a moderately depressed frontal region, straight nasal bones, parietal bones that are moderately separated and a unique combination of dental characters: long crochet; usually closed off, round medifossette; crista absent in molars; very strong mesial and distal constriction of the protocone, which is lingually flattened, resulting in a long antecrochet that may close off the median valley at a very advanced wear stage; a strong mesial constriction of the hypocone of the molars; medifossette commonly closed in premolars; and a discontinuous lingual cingulum is mainly present in the premolars.
Referred material
A scapula (GPIT/MA/04974), four humeri (GPIT/MA/04831, GPIT/MA/04832, GPIT/MA/04833, and GPIT/MA/04834), three radii (GPIT/MA/04778, GPIT/MA/04786, and GPIT/MA/04791), a scaphoid (GPIT/MA/04784), two second metacarpals (GPIT/MA/04782 and GPIT/MA/04783), two third metacarpals (GPIT/MA/04756 and GPIT/MA/04782), a fourth metacarpal (GPIT/MA/04776) a fifth metacarpal (GPIT/MA/04756), a femur (GPIT/MA/04835), a patella (GPIT/MA/04781), a tibia (GPIT/MA/04836), three astragali (GPIT/MA/04763, GPIT/MA/04779, and GPIT/MA/04792), three calcanei (GPIT/MA/04792, GPIT/MA/04859, and GPIT/MA/04860), a navicular (GPIT/MA/04782), and a third metatarsal (GPIT/MA/04796).
Remarks
Chilotherium habereri is the first chilothere species described from China [24]. It was originally erected as Rhinoceros habereri [24] and later assigned to the genus Chilotherium [7]. However, the type material comprises only isolated teeth and dental characters are evidently not reliable for species identifications in the genus Chilotherium (see for example [47]). Ringström (1924) [7] tried to address this issue by using a skull from a different, but probably closely situated, locality with identical teeth to the type material as a basis to redefine the species. However, the morphological features seen in these teeth that were used to define the species are subject to intraspecific variation within the genus Chilotherium, therefore assigning a skull to the same species based solely on the morphology of the premolars may prove difficult. This complicates the matter and shows that a detailed revision of the type material, along with the material studied by Ringström (1924) [7], is needed. For the purpose of the current study, we tentatively refer the chilothere material from Kutschwan housed in the GPIT to C. habereri, as also proposed in the initial description of the material [25,26] and later supported by Ringström (1924) [7].
Chilotherium schlosseri [Weber, 1905] [14]
Neotype
A well-preserved skull (GPIH 3015) with an associated mandible (GPIH 3015a), designated by Kampouridis et al. (2023) [10].
Diagnosis
A large Chilotherium species characterised by widely separated parietal crests (minimal distance between parietal crests always over 70 mm in adult individuals), notably depressed frontal and nasal bones, nasal bones that bear a central longitudinal groove on the dorsal side, and a unique combination of dental characters: very long crochet; very strong mesial and distal constriction of the protocone, which is lingually flattened, resulting in a very long antecrochet that usually closes off the median valley at an early wear stage in all teeth; a prominent mesial constriction of the hypocone; crista frequently present that closes the medifossette; and a discontinuous lingual cingulum that is occasionally moderately developed in the premolars; often a closed prefossette is present in the P2; in addition to sporadically present enamel plications in the upper teeth; and discontinuous lingual and buccal cingulids in the lower teeth (after [10]).
Referred material
An axis (GMM FO-14), two scapulae (GMM 563 and NHMW-GEO-1911/0005/0245), seven humeri (GMM 495, GMM 496, GMM 599, GMM 601, GMM 602, GMM FO-18, and GMM FO-24), two radii (GMM 561 and GMM 564), two scaphoids (NHMW-GEO-1911/0005/0152 and NHMW-GEO-2009z0089/0001), four femora (GMM FO-10, GMM FO-11, GMM FO-12, and NHMW-GEO-1911/0005/0273), 10 tibiae (AMNH 22818, GMM 497, GMM 594, GMM 595, GMM 596, GMM FO-13, GMM FO-19, GMM FO-27, NHMW-GEO-1911/0005/0497, and NHMW-GEO-1911/0005/0498), six astragali (AMNH Unnumbered, AMPG-SAM516, AMPG-SAM517, AMPG-SAM518, GMM 571, and NHMW-GEO-1911/0005/0424), three calcanei (AMNH-20794, AMNH-95122, and AMNH Unnumbered), a navicular (AMPG-SAM516), a cuboid (AMPG-SAM516), a second metatarsal (GMM 572), four third metatarsals (AMNH-22818, AMPG-SAM520, GMM 572, and NHMW-GEO-1911/0005/0220), and two fourth metatarsals (AMPG-SAM520 and GMM 572)
Description and comparison
Axis.
The studied material includes only one axis belonging to C. persiae from Maragheh (MNHN.F.MAR3939) and one to C. schlosseri from Samos (GMM FO-14) (Fig 2). Both are missing the neural arch and preserve only the centrum of the vertebra. The two bones are very similar to each other, with the axis of C. schlosseri being slightly larger. The dens of the axis is quite narrow and elongated in comparison to the contemporary ‘Ceratotherium’ neumayri, in which the dens is much shorter, wider, and bears a prominent articular facet for the atlas on its ventral side (as seen in GMM FO-21, pers. obs.). In both Chilotherium species, the anterior articular facet of the axis is not as clearly defined, though this might be caused by some damage to the bone. The two vertebrae exhibit some slight differences like the ventral crest is wider and looks more massive in C. schlosseri, whereas in C. persiae it seems narrower and constricted in the middle. Concerning the posterior articular facet for third cervical vertebra, in C. schlosseri it has a high-oval shape with a slight, dorsal indentation, whereas in C. persiae it is more rounded, and the dorsal surface is flat. Additionally, in ventral view there is an indentation on the posterior side in C. persiae (Fig 2C), whereas C. schlosseri lacks such an indentation (Fig 2G). Lastly, the area where the lateral walls of the neural canal of the axis connect to the vertebral body seems to extend more posteriorly in C. schlosseri than in C. persiae. A single axis was described for C. kowalevskii from Grebeniki [15]. Unfortunately, the description and illustration provided for this axis [15] are not sufficient for a comparison to our specimens. Additionally, due to the small sample size it is not possible to assess the intraspecific variability. Nonetheless, it is very likely that at least some of the observed differences are of taxonomic value.
A–D, Chilotherium persiae (Pohlig, 1885) [19] (MNHN.F.MAR3939) from Maragheh (Iran) and E–H, Chilotherium schlosseri (Weber, 1905) [14] (GMM FO-14) from Samos (Greece) in anterior (A, E), posterior (B, F), ventral (C, G), and dorsal (D, H) views. Abbreviations: aaf, anterior articular facet; paf, posterior articular facet; and vc, ventral crest. Scale bar equals 5 cm.
Scapula.
The studied material includes five partial scapulae of C. persiae from Maragheh, one of C. habereri from Kutschwan, and two of C. schlosseri from Samos (Fig 3). Unfortunately, none of them are complete, but in all specimens the proximal portion is adequately preserved to compare them, and in some specimens parts of the scapular spine is preserved. The spine is well developed and straight. In the more complete specimens, it is visible that the infraspinous fossa is only slightly larger than the supraspinous fossa. A spinous tuberosity is present, but not very prominent. In all specimens the articular facet for the humerus is oval and anteroposteriorly elongated (as indicated by Guerin (1980) [31]). The articular facet does not seem to differ among the scapulae of C. persiae, C. habereri, C. schlosseri, and C. wimani [16] neither in size, nor in shape. The supraglenoid tubercle and the coracoid process form a continuous tuberosity, the shape of which varies to some extent in the studied specimens. It is more rounded in C. persiae and C. schlosseri (Fig 3C, I) and rather flattened in C. habereri (Fig 3F). Furthermore, is seems to be more protruding in C. schlosseri; however, the small sample size may obscure the variability of these features. The specimens of C. persiae do in fact exhibit some variety in their morphology.
A–C, Chilotherium persiae (Pohlig, 1885) [19] (MNHN.F.MAR3900, right) from Maragheh (Iran), D–F, Chilotherium habereri (Schlosser, 1903) [24] (GPIT/MA/04794, left) from Kutschwan (China), and G–I, Chilotherium schlosseri (Weber, 1905) [14] (GMM 428, right) from Samos (Greece) in anterior (A, D, and G), posterior (B, E, and H), and distal (C, F, and I) views. Abbreviations: daf, distal articular facet; isf, infraspinous fossa; sgl, supraglenoid tubercle; ssf, supraspinous fossa. Scale bar equals 10 cm for A, B, D, E, G, and H and 5 cm for C, F, and I.
Humerus
The studied material includes 12 humeri of C. persiae from Maragheh, four of C. habereri from Kutschwan, and seven of C. schlosseri from Samos (Fig 4, Table 1, S1 Table 1). Most specimens are only partially preserved. Their morphology is rather similar in all available specimens. The diaphysis is relatively straight. In the proximal part, the great and lesser tubercles are of similar size and their processes extend proximally to a similar degree, though the anterior part of the greater tubercle is broken in most specimens. The intermediate tubercle is somewhat less pronounced, without any distinct process. The two intertubercular grooves are smooth and relatively shallow. Anteriorly, the crest of the greater tubercle is prominent and extends distally to the deltoid tuberosity. The deltoid tuberosity is placed at the middle of the diaphysis. Its’ morphology and only be assessed in GMM 601 of C. schlosseri and GPIT/MA/04832 of C. habereri, where it is similarly well developed and posteriorly projecting. In all available specimens, on the posterior side also the tricipital line is well visible, represented by a ridge starting at the middle of the proximal part, approximately at the beginning of the greater tubercle and merged into the deltoid tuberosity, also marking the limit of the muscle attachment area there. In the distal part, the medial epicondyle is relatively weak, especially when compared to the prominent lateral one, with its crest reaching almost the middle of the shaft. The radial fossa is almost triangular and well-developed, placed above the trochlea of the humerus. The medial condyle of the trochlea is larger than the lateral one and they are not separated by a trochlear scar. The olecranon fossa is fairly rounded and deep in all specimens, but it seems to be somewhat narrower transversally in C. schlosseri (Fig 4N) than in C. habereri (Fig 4H). In the latter species, the opening of the olecranon fossa is much wider, and the fossa has overall a transversally slightly elongated oval shape, as seen in the four available specimens.
A–F, Chilotherium persiae (Pohlig, 1885) [19] (MNHN.F.MAR3904, right) from Maragheh (Iran), G–L, Chilotherium habereri (Schlosser, 1903) [24] (GPIT/MA/04831, right) from Kutschwan (China), and M–R, Chilotherium schlosseri (Weber, 1905) [14] (GMM 601, right) from Samos (Greece) in anterior (A, G, and M), posterior (B, H, and N), medial (C, I, and O), lateral (D, J, and P), proximal (E: 3D model, K, and Q), and distal (F, L, and R) views. Abbreviations: dt, deltoid tuberosity; gt, greater trochanter; h, head of the humerus; lc, lateral condyle; le, lateral epicondyle; lt, lesser trochanter; mc, medial condyle; me, medial epidondyle; rf, radial fossa; and of, olecranon fossa. Scale bar equals 20 cm for A–D, G–J, and M–P, and 10 cm for E, F, K, L, Q, and R.
Unfortunately, for C. persiae it was only possible to measure the length from the caput humeri and not the full length of any specimen. For both C. schlosseri and C. habereri two complete humeri were studied respectively, allowing most dimensions to be measured (S1 Table 1).
The measurements provided for a well-preserved humerus of the type species C. anderssoni from its type locality Lok. 30 [7] shows that the humerus of Chilotherium is only little shortened compared to The comparison of the measurements of the herein studied chilotheres (Table 1, S1 Table 1), a well-preserved humerus of the type species C. anderssoni from its type locality Lok. 30 (China) [7], and several other chilothere humeri from the literature [11,12,16,17] to other rhinocerotids, like the aceratheriines Aceratherium incisivum Kaup, 1832 [80] from Höwenegg (Germany) [74] and Acerorhinus zernowi from Tung-gur (China) [75] shows that humerus of Chilotherium is only little shortened compared. Among chilotheres, ‘C.’ wimani from the Linxia Basin [16] and E. samium [14] seem to exhibit the smallest humeri, with the maximal length of the former being 335 mm and the single specimen of the latter being 312 mm. The species with the largest dimensions for the humerus seem to be C. orlovi from Pavlodar, with a length value range between 357 and 378 mm (n = 5) [12]. The dimensions of the humeri of C. schlosseri, C. persiae, C. habereri, C. kowalevskii from Grebeniki [15], and a potential chilothere from Loc. 51 of the Sinap Formation (Turkey) [81], are intermediate, with overlapping ranges.
Radius
The studied material includes 17 radii C. persiae from Maragheh, three of C. habereri from Kutschwan, and two of C. schlosseri from Samos (Fig 5, Table 2, S1 Table 2). Most specimens are not complete and overall, poorly preserved. The proximal articular facet for the humerus is transversally elongated, subtrapezoidal in shape with the medial side being much wider anteroposteriorly than the lateral side. In the middle of the proximal facet, there is an indentation on the anterior side and a protrusion on the posterior side. In anterior view, below the proximal articular facet the insertion for the musculus biceps brachii is represented byan excavation in the middle of the bone, with a rugose surface. This insertion is much deeper in the only specimen of C. habereri where it could be assessed (Fig 5G) than in C. persiae, C. schlosseri (Fig 5A, Q), and C. wimani from the Linxia Basin [16]. However, in C. persiae it seems to cover a wider area. In GMM 564, the only specimen of C. schlosseri in which the proximal part of the radius is preserved, it is not complete, and the morphology of this feature cannot be studied in detail. Medially placed, next to the insertion for the m. biceps brachii, the radial tuberosity is visible. This is also much more prominent in C. habereri (Fig 5G), where it covers almost one third of the bone’s width in anterior view. In C. persiae (Fig 5A), the radial tuberosity is much smaller, and the surface is less rugose. In C. schlosseri, this feature cannot be observed. On the other side of the bone the lateral tuberosity is less prominent but extends towards the posterior side of the bone. On the posterior side, extending distally from the articular facet for the humerus, two articular facets for the ulna are visible. The medial one is a thin, transversally elongated stripe, whereas the lateral one is much larger and has an irregular shape and is larger in C. persiae than in C. habereri (Fig 5B, H). Around these two facets, a wide rugose area covers the whole proximal part of the posterior side of the bone, which represents an attachment area for the interosseous ligaments between the radius and the ulna. Further below a proximal interosseous space of about 2–3 cm is found in all specimens. From there on, the rugose crest for the attachment of the interosseous ligaments, runs down towards the distal part of the bone. A few centimeters above the distal epiphysis of the bone, a distal interosseous space of about 1 cm is placed. Right below this a large rugose area for the attachment of the ulna exists on the lateral side of the bone. At the distal end of this area, a small articular facet for the ulna is found, which seems somewhat larger in C. persiae and C. schlosseri (Fig 5D, T) than in C. habereri (Fig 5O). A discontinuous, rugose, and mediolaterally oriented crest is placed a few centimeters above the distal articulation. Its development varies between the specimens studied, even within the same species. For instance, for C. persiae in MNHN.F.MAR3908 it is rather weak, whereas in MNHN.F.MAR3906 it is much more distinct. In both specimens of C. schlosseri (GMM 561 and GMM 564) it is not very prominent. Concerning the two specimens of C. habereri, the crest is rather weak in GPIT/MA/04791 and almost absent in GPIT/MA/04786. Medially and laterally to this crest, two tuberosities are placed, which also exhibit a significant degree of variability. In C. persiae and C. schlosseri they have the shape of relative small tuberosities (Fig 5B–D, R–T), which may be somewhat proximodistally elongated in the case of MNHN.F.MAR3806 of C. persiae. In both specimens of C. habereri (Fig 5M–O), they actually form strong, proximodistally oriented crests of a few centimeters in length. The distal articular facets have a rather similar arrangement and form among the species. Medially, the articular facet for the scaphoid is large and anteroposteriorly convex. Laterally to the scaphoid facet the articular facet for the semilunar is smaller, subtrapezoidal to rounded in shape, and concave. Next to it the articular facet for the pyramidal is represented by a slightly obliquely placed, small stripe, which is not visible in all specimens. For instance, in MNHN.F.MAR3806 it is very faintly visible, whereas in MNHN.F.MAR3808 there is no trace of it, which may mean that it just is confluent with the facet for the semilunar.
A–F, Chilotherium persiae (Pohlig, 1885) [19] (MNHN.F.MAR3912, right) from Maragheh (Iran), G–P, Chilotherium habereri (Schlosser, 1903) [24] (G–K: GPIT/MA/04878, left and L–P: GPIT/MA/04786, right) from Kutschwan (China), and Q–V, Chilotherium schlosseri (Weber, 1905) [14] (GMM 564, right) from Samos (Greece) in anterior (A, G, L, and Q), posterior (B, H, M, and R), medial (C, I, N, and S), lateral (D, J, O, and T), proximal (E, K, and U), and distal (F, P, and V) views. Abbreviations: ilru, attachment surface for the interosseous ligament between the radius and ulna; lrf, lateral radial facet; lt, lateral tuberosity; mbb, insertion for the musculus biceps brachii; mbb, medial radial facet; rt, radial tuberosity; pyr, articular facet for the pyramidal; rsu; rugose surface for the attachment of the ulna; sc, articular facet for the scaphoid; sem, articular facet for the semilunar; and ul, articular facet for the ulna. Scale bar equals 10 cm for A–D, G–J, L–O, and Q–T, and 5 cm for E, F, K, P, U, and V.
Concerning the dimensions of the radius, C. sarmaticum from Berislav exhibit the lowest values for the length of (247–280 mm) [11,16]. They are followed by those of ‘C.’ wimani from the Linxia Basin and C. kowalevskii from Grebeniki with value ranges for the length of 266–278 mm and 262–286 mm, respectively [11,17]. The radii of C. orlovi from Pavlodar exhibit the largest dimensions, with a value range of 292–320 mm for the length [12]. The only species in which the radii seem to reach or almost reach the lowest values of C. orlovi, are C. schlosseri, C. persiae (Table 2), and a potential chilothere from Loc. 12 of the Sinap Formation [81]. The single radius of C. anderssoni from Lok. 30, seems to show intermediate dimensions, with a length of 280 mm [7]. For C. habereri no complete radius is known, but the measurable dimensions indicate that the radii were quite large, and almost comparable to C. orlovi.
Scaphoid
The studied material includes one scaphoid of C. persiae from Maragheh, one of C. habereri from Kutschwan, and two from C. schlosseri from Samos (Fig 6, S1 Table 4). The specimens are fairly similar, though the Maragheh specimen is partially damaged. In proximal view, the articular facet for the radius is large with a subtrapezoidal outline and is in medial view sigmoidal. A prominent posteriolateral tuberosity is present, the morphology of which somewhat varies, medially to this a long thin stripe represents the proximal articular facet for the semilunar. In distal view, three articular facets for the trapezium, the trapezoid and the magnum are present. The medial facet for the trapezium is the smallest and its shape varies from triangular to semi-oval; this facet seems to be somewhat longer in C. habereri than in the other two species. The central facet for the trapezoid is the largest and has a subtrapezoidal shape with a convex medial border. The lateral articular facet for the magnum is subtriangular and connected to the laterodistal articular facet for the semilunar. The specimens are proximodistally short and lateromedially elongated, similar to ‘C.’ wimani from the Linxia Basin [16], C. sarmaticum from Berislav [11], and C. orlovi from Pavlodar [12]. Although, the measurements provided for the scaphoid of C. orlovi represent the largest values for any chilothere yet [12].
A–D, Chilotherium persiae (Pohlig, 1885) [19] (MNHN.F.MAR1412, left) from Maragheh (Iran), E–H, Chilotherium habereri (Schlosser, 1903) [24] (GPIT/MA/04784, left) from Kutschwan (China), and I–L, Chilotherium schlosseri (Weber, 1905) [14] (NHMW-GEO-2009z0089/0001, left) from Samos (Greece) in anterior (A, E, and I), posterior (B, F, and J), proximal (C, G, and K), and distal (D, H, and L) views. Abbreviations: mag, articular facet for the magnum; pt, posterior tubersotiy; rad, articular facet for the radius; sem, articular facet for the semilunar; td, articular facet for the trapezoid; and tm, articular facet for the trapezium. Scale bar equals 5 cm.
Semilunar
A single partial semilunar (MNHN.F.MAR.1411) of C. persiae from Maragheh has been examined. Only the anterior portion of the bone is preserved. The proximal articular facet for the radius is saddle-shaped, with a posteriorly extending stripe and the anterior border of the facet being convex. In medial view, it bears two suboval articular facets for the scaphoid. In lateral view, two articular facets for the pyramidal are visible, the proximal facet is subtrapezoidal, whereas the distal one is a small longitudinal stripe. In distal view, only two articular facets are visible in the preserved part, the medioproximal one is the one for the scaphoid and the lateral facet articulates to the unciform and the preserves portion is subcircular. The articular facet for the magnum, which would have been placed next to these facets, is not preserved. The measurable dimensions of the Maragheh specimen are almost identical to the measurements provided for the semilunar of ‘C.’ wimani from the Linxia Basin [16]. Interestingly, they are significantly higher than the values provided for the dimensions of the semilunar of C. sarmaticum from Berislav and C. orlovi from Pavlodar [11,12]. It is possible that the measurements of the latter two species do not correspond to the same sections measured in C. persiae and ‘C.’ wimani, which follow those proposed by Guérin (1980) [31].
Pyramidal
Two pyramidals (MNHN.F.MAR.1405 and MNHN.F.MAR.1413) of C. persiae from Maragheh have been examined (Fig 7A–E). They are higher than wide. The proximal facet, for the ulna, is antero-posteriorly concave and transversely slightly convex and indistinguishable from the probably much smaller facet for the radius. In medial view, two articular facets for the semilunar are visible, they are separated by a wide groove and have a semi-oval outline. In lateral view, in the distal half of the bone a well-developed tubercle exists that extends towards the posterior side of the bone. In posterior view, the articular facet for the pisiform has an obtuse subtriangular shape and contacts the articular facet for the ulna. The distal articular facet, for the unciform, is concave and has a suboval outline. The dimensions of the two pyramidals of C. persiae and those of ‘C.’ wimani from the Linxia Basin seem to overlap almost perfectly, despite the very limited sample size [16]. The single measurement for the pyramidal of C. anderssoni from Lok. 30 is also quite close to the height values provided for the other two species [7]. The values for the pyramidal of C. orlovi from Pavlodar are comparably higher [12]. The values provided for C. sarmaticum from Berislav [11] seem somewhat peculiar, as the given transversal diameter value is higher than the height of the bone, which is not the case in the other chilotheres and unusual for rhinoceroses in general. It would be more plausible to assume that distances measured by Korotkevitch (1970) [11], are not the same as those given by Guérin (1980) [31], which was followed in the current study.
A–E, pyramidal (MNHN.F.MAR1413, left), F–J, trapezoid (MNHN.F.MAR1409, right), and K–O, unciform (NHMW-GEO-2020/0014/0140, left) in anterior (A, F, and K), proximal (B, G, and L), distal (C, H, and M), medial (D, I, and N), and lateral (E, J, and O) views. Abbreviations: mag, articular facet for the magnum; McII, articular facet for the McII; McIII, articular facet for the McIII; McIV, articular facet for the McIV; McV, articular facet for the McV; pp, posterior process; pyr, articular facet for the pyramidal; rad, articular facet for the radius; sc, articular facet for the scaphoid; sem, articular facet for the semilunar; tm, articular facet for the trapezium; ul, articular facet for the ulna;and un, articular facet for the unciform. Scale bar equals 5 cm.
Trapezoid
Only one trapezoid (MNHN.F.MAR.1409) of C. persiae from Maragheh was studied (Fig 7F–J). In proximal view, the articular facet for the scaphoid has a subtrapezoidal outline. In medial view, it bears a small, elongated articular stripe for the trapezium. In lateral view, it bears a larger almost rectangular facet for the magnum. In distal view, the articular facet for the McII is oval. Deng (2002) [16] mentioned that the trapezoid of ‘C.’ wimani from the Linxia Basin in China bears two articular facets for the magnum on its lateral side, in contrast to the single articular facet present in the Maragheh specimen. The measurements also show that the Maragheh trapezoid is transversally wider (see S1 Table 7). The dimensions of the trapezoid of ‘C.’ wimani seem to be much closer to those of C. sarmaticum from Berislav [11]. The trapezoid of C. anderssoni from Lok. 30 seems to have dimensions that are more similar to C. persiae [7]. The trapezoid of the C. orlovi from Pavlodar, on the other hand, has generally greater dimensions, with only the transversal width being smaller [12]. It has to be noted that in complex bones like carpals, it is very difficult to compare measured dimensions if they did not follow the exact same method of measuring.
Unciform
Six unciforms from Maragheh can be attributed to C. persiae (Fig 7K–O). In three of them the characteristic posterior process is broken off. In proximomedial view, the saddle-shaped facet for the pyramidal and the trapezoidal-shaped facet for the semilunar meet at an almost right angle. In distal view, the articular facets for the magnum, McIII, and McIV, are represented by a continuing wide stripe, without any well-defined separation. The articular facet for the magnum has posteriorly bent medial edge. The articular facet for the McIII has an irregular outline. The facet for the McIV is subtriangular to subtrapezoidal and forms a small, but notable, hump. In lateral view, the same articular stripe continues and forms the articular facet for the rudimentary McV. At the border between the articular facets for the McIV and McV a very subtle depression is present. In posterior view, a large curved posterior process is visible. The unciforms of C. persiae are much larger than those of ‘C.’ wimani from the Linxia Basin [16], especially the length of the bone in the latter species is significantly smaller. The dimensions reported for the unciform of C. anderssoni from Lok. 30 [7] and C. sarmaticum from Berislav [11] are only slightly smaller than those of C. persiae. The dimensions of the unciforms of C. orlovi from Pavlodar on the other hand are greater than in any other chilothere thus far [12].
Second metacarpal
Four McII of C. persiae from Maragheh and two of C. habereri from Kutschwan were studied (Fig 8, Table 3). Only in a single specimen (NHMW-GEO-2020/0014/0142) the complete length of the bone is preserved and measurable, in the other specimens the distal portion is broken off. In proximal view the articular facet for the trapezoid is suboval to subtrapezoidal, transversally concave and anteroposteriorly convex. Laterally, it shares a wide connection to the obliquely placed longitudinal articular stripe for the magnum. The magnum facet contacts the articular facet(s) of the McIII in both species. In the two specimens of C. persiae where the proximal part is sufficiently preserved, the articular facet for the McIII is a continuous, wide stripe and is in proximal view concave (Fig 8C), whereas in the single specimen of C. habereri we see two small, distinct articular facets for the McIII (Fig 8I). No articular facet for the trapezium is visible on the proximomedial side of the bone. On the anterior side of the McII a prominent tuberosity covers the proximal part of the bone that represents a muscle attachment area for the extension of the metapodials. The tuberosity forms a slight discontinuous groove between the articular facet for the trapezoid and the tuberosity. On the posterior side, the articular facets for the trapezoid and magnum end on a slight protuberance, from which a small but notable crest extends distally to a variable degree. The shaft of the bone is straight, and the distal articular head is somewhat asymmetrical. The median distal keel develops in the posterior half of the trochlea. Deng (2002) [16] described a similar arrangement and overall morphology for the proximal articular facets of the McII mentioning also that the facets for the magnum and the McIII create a marked crest, by meeting at an angle. The length of the sole complete McII from Maragheh (Table 3, S1 Table 9) is slightly lower than the values given for C. kowalevskii from Grebeniki and C. orlovi from Pavlodar but is comparable to the length measured in C. anderssoni from Lok. 30, ‘C.’ wimani from the Linxia Basin, C. sarmaticum from Berislav, and a potential ‘primitive’ Chilotherium from Loc. 12 of the Sinap Formation [7,11,12,16,81]. The maximal transversal diameter of the proximal epiphysis is best comparable to those of ‘C.’ wimani, C. kowalevskii, C. orlovi, and the potential ‘primitive’ Chilotherium from Sinap, while the proximal anteroposterior diameter is closer to C. sarmaticum and C. orlovi.
A–F, Chilotherium persiae (Pohlig, 1885) [19] (NHMW-GEO-2020/0014/0142, left) from Maragheh (Iran), and G–K, Chilotherium habereri (Schlosser, 1903) [24] (GPIT/MA/04783, left) from Kutschwan (China), in anterior (A, G), posterior (B, H), lateral (C, I), medial (D, J), proximal (E, K), and distal (F) views. Abbreviations: mag, articular facet for the magnum, McIII, articular facet for the McIII; and td, articular facet for the trapezoid. Scale bar equals 5 cm.
Third metacarpal
Five McIII of C. persiae from Maragheh and two of C. habereri from Kutschwan were studied (Fig 9, Table 4). In only two of those (MNHN.F.MAR.1379 and MNHN.F.MAR.1383) the complete length of the bone is measurable. The proximal articular facet for the magnum has an almost subtriangular to subtrapezoidal outline, with a sigmoidal anterior border. It is transversally concave and anteroposteriorly convex and meets the small, oval articular facet for the McIII at an almost right angle, on the medial side. Laterally in the anterior part, the facet for the magnum meets the trapezoidal articular facet for the unciform at an almost right angle. This facet, in turn contacts the semi-oval anterior articular facet for the McIV. This facet is separated by a wide and deep groove from the elliptical posterior articular facet for the McIV, which meets the posterior part of the articular facet for the magnum at an acute angle.
A–F, Chilotherium persiae (Pohlig, 1885) [19] (MNHN.F.MAR1379, right) from Maragheh (Iran), and G–P, Chilotherium habereri (Schlosser, 1903) [24] (G–K: GPIT/MA/04796 and L–P: GPIT/MA/04782, both right) from Kutschwan (China) in anterior (A, G, L), posterior (B, H, M), lateral (C, I, N), medial (D, J, O), proximal (E, K), and distal (F, P) views. Abbreviations: mag, articular facet for the magnum; McII, articular facet for the McII; McIV, articular facet for the McIV; and un, articular facet for the unciform. Scale bar equals 5 cm.
All specimens have a rather straight shaft, with a slight concave lateral aspect in anterior view. In the proximal part of the anterior side, just below the proximal articular facet, a strong rugose surface for the insertion of the carpal extensor muscle exists. The distal articular head is fairly symmetrical, with a strongly convex proximal border on the anterior side and a prominent central keel on the posterior sides. The keel reaches approximately the middle of the head. Additionally, above this keel, on the posterior side, a small but prominent rugose crest extends proximally and vanishes before the middle of the bone. The overall morphology of the Maragheh specimens fits the descriptions provided for the McIII of other chilotheres like C. anderssoni from Lok. 30, ‘C.’ wimani from the Linxia Basin, C. sarmaticum from Berislav, C. kowalevskii from Grebeniki, and C. orlovi from Pavlodar [7,11,12,16]. Metrically, the main aspect that could be used to differentiate them seems to be the length of the bones (S1 Table 10). More specifically, the length of the Maragheh specimens (132.9 and 140.7 mm, Table 4) is comparable to the reported values for C. kowalevskii (128.9–142.7 mm, n = 10), C. orlovi (128–153 mm, n = 15) and the potential ‘primitive’ Chilotherium from Loc. 12 of the Sinap Formation (135 mm, n = 1), which are greater than the reported values for ‘C.’ wimani (120–123 mm, n = 2), C. anderssoni (127 mm, n = 1), and C. sarmaticum (113–129 mm, n = 6), having almost no overlap between them [7,11,12,16,81].
APDAPD
Fourth metacarpal.
Six McIV of C. persiae from Maragheh and one of C. habereri from Kutschwan were studied (Fig 10, Table 5), all of which are more or less complete. In proximal view, the shape of the articular facet for the unciform varies from subtriangular to subtrapezoidal. Posteriorly it bends distally, creating additional articulation area for the unciform. In medial view, the two articular facets for the McIII are separated by a groove. The anterior articular facet for the McIII shares a wide contact to the articular facet for the unciform and form a ridge. Its’ shape varies from an elongated stripe to subtriangular. The posterior facet for the McIII either contacts the facet for the unciform only slightly or not at all. Its’ shape varies from almost circular to almost rectangular. On the lateral side, the articular facet for the rudimentary McV is a relatively thin and poorly-defined articular stripe that connects to the posterior extension of the unciform facet. Below the articular facets for the third metacarpal a long, almost triangular rugosity for the attachment of the interosseous ligament is placed. The shaft is slightly laterally curved, and the distal articular head is somewhat asymmetrical.
A–F, Chilotherium persiae (Pohlig, 1885) [19] (MNHN.F.MAR1386, right) from Maragheh (Iran), and G–L, Chilotherium habereri (Schlosser, 1903) [24] (GPIT/MA/04776, right) from Kutschwan (China), in anterior (A, G), posterior (B, H), lateral (C, I), medial (D, J), proximal (E, K), and distal (F, L) views. Abbreviations: McIII, articular facet for the McIII; McV, articular facet for the McV; and un, articular facet for the unciform. Scale bar equals 5 cm.
The McIV of the different chilotheres have a similar morphology overall. Though there are some potential differences found in the literature. For instance, while Korotkevich (1970) [11] and Bayshashov (1993) [12] described the existence of a small articular facet for the McV in C. sarmaticum from Berislav and C. orlovi from Pavlodar respectively, Deng (2002) [16] mentioned that in ‘C.’ wimani from the Linxia Basin there is no articular facet for the McV. However, no lateral view of any McIV is provided to confirm this and it is possible that the McV facet in ‘C.’ wimani remained unnoticed due to its very small size..
The metrical comparison of the McIV (S1 Table 11) indicates that there is a distinction between some species based on the values of the total length of this bone. More specifically, C. persiae (100.6–105.6 mm, n = 5) fits best among the greater values seen in C. kowalevskii (93.2–108.2 mm, n = 10), and is somewhat smaller than C. orlovi (107–115 mm, n = 8) [11,12]. Whereas ‘C.’ wimani (91.5–93 mm, n = 4), C. sarmaticum (90.5–95 mm, n = 6), Chilotherium indet. from Kavakdere in Turkey (89.4–92.5 mm, n = 2) have smaller values [11,16,81]. The single McIVs of C. anderssoni (98 mm) and of the potential ‘primitive’ Chilotherium from Sinap (97 mm) have an intermediate length [7,81]. Additionally, some material from Loc. 26 in the Sinap region referred to as ‘Acerorhinus sp. nov.’, includes a McIV with a length of 95.5 mm [81], which could also be associated with the smaller group, and probably belongs to a chilothere and not to Acerorhinus, as previously suggested [42]. Its’ gracility index (TDdia/L = 30.3%) falls well into the value range of the chilotheres (27.5–34.8%, n = 14) and the overall morphology of this bone also fits the morphology seen in Chilotherium. A McIV (MNHN.F.TRQ329) from Küçükçekmece (Turkey) that was assigned to C. schlosseri is, with a length of 112 mm, longer than most Chilotherium species, being in the range of C. orlovi but very slender (TDprox = 29 mm and TDdia = 25 mm). Its’ gracility index (TDdia/L = 22.3%) is lower than in any available chilothere (27.5–34.8%, n = 14) and even lower than in the non-chilothere Aceratherium incisivum from Höwenegg (24.3–27.4%, n = 4) but falls well into the range of Acerorhinus zernowi from Tung Gur (19.8–23.9%, n = 3), which has a significantly more elongated appendicular skeleton than any chilothere.
Fifth metacarpal
A single McV (GPIT/MA/04756) of C. habereri from Kutschwan was studied (Fig 11). It is small, with a total length of 27.4 mm, and has a cone-like shape with a rounded tip. It has two articular facets: one large and quadrangular one for the unciform and a much smaller striper for the McIV. The two articular facets connect at a slightly acute angle and do not have any well-defined border. No other McV was found in the studied collections, and its morphology remains unknown for C. persiae and C. schlosseri. Ringström (1924) [7] described a simple, rudimentary McV for the type species C. anderssoni from Lok. 30. Deng (2002) [16] mentioned no McV of ‘C.’ wimani from the Linxia Basin and noted that in the studied McIV sample there is no articular facet for the McIV. Unfortunately, no lateral view of any McIV is provided, to confirm the complete lack of this facet. Korotkevich (1970) [11] and Bayshashov (1993) [12] explicitly mentioned the presence of a small articular facet for the McV in the McIV of C. sarmaticum from Berislav and C. orlovi from Pavlodar, respectively. Interestingly, the potential ‘primitive’ Chilotherium from Loc. 12 of the Sinap Formation exhibits a well-developed, and not at all rudimentary, McV [81], comparable to the McV of Aceratherium incisivum from Höwenegg [74]. This is in stark contrast to C. habereri and C. anderssoni, which have a McV that is similarly highly reduced to extant rhinoceroses. Hence, the ‘primitive’ Chilotherium from Loc. 12 of the Sinap Formation is likely to not represent at all a chilothere, but some other aceratheriine. Representatives of Chilotherium overall do not have a functional McV, confirming the notion of Ringström (1924) [7] that Chilotherium has tridactyl limbs.
A–F, McV in lateral (A), posterior (B), proximal (C), medial (D), anterior (E), and distal (F) views. Abbreviations: McV, articular facet for the McV; and un, articular facet for the unciform. Scale bar equals 5 cm.
Femur
Six femora of C. persiae from Maragheh, one of C. habereri from Kutschwan, and two of C. schlosseri from Samos were studied (Fig 12). Of these, only three, one of C. persiae and two of C. schlosseri, were complete enough to measure (almost) all dimensions of the bone. In most specimens the proximal part is either missing or damaged. The femoral head is protruding from the shaft and positioned higher than the greater trochanter, which is restricted to a slightly protruding, prominent tuberosity on the lateral side of the bone. On the posterior side of the femoral head, the fovea capitis is very high and narrow. On the lateral side, the convexity of the greater trochanter extends from the greater trochanter distally, creating the trochanteric fossa that is less prominent in specimen GPIT/MA/04835 of C. habereri (Fig 12G) than in the femora of C. persiae and C. schlosseri (Fig 12B, L). The lesser trochanter is represented by a small crest on the medial side of the bones, below the femoral head. The shaft narrows transversally towards the middle of the bone, with the third trochanter being moderately developed laterally and subtriangular when preserved, but is broken in almost all specimens. Distally, the trochlea for the patella is highly asymmetrical, with the medial lip extending much more proximally than the lateral one, in all specimens. On the anterior side, above the trochlea for the patella a deep fossa is visible (Fig 12A, K). On the posterior side, the two condyles for articulation to the tibia are slightly asymmetrical, with the medial one being wider and more rounded than the lateral one. They are separated by a deep intercondyloid fossa. The femur seems to be morphologically rather conservative within chilotheres, with most species looking almost identical [11,12,15,16]. Metrically, the femora of ‘C.’ wimani from the Linxia Basin [16] exhibits the smallest dimensions, with its length not surpassing 400 mm. On the other hand, the femora of C. orlovi from Pavlodar seem to exhibit the largest dimensions, with a length ranging from 438 mm to 485 mm [12]. The dimensions of C. schlosseri (S1 Table 12) and C. kowalevskii from Grebeniki seem to be somewhat intermediate but closer to C. orlovi, whereas the dimensions of C. sarmaticum from Berislav somewhat closer to those of ‘C.’ wimani [11,17].
A–E, Chilotherium persiae (Pohlig, 1885) [19] (MNHN.F.MAR3921, left) from Maragheh (Iran), F–J, Chilotherium habereri (Schlosser, 1903) [24] (GPIT/MA/04835, left) from Kutschwan (China), and K–P, Chilotherium schlosseri (Weber, 1905) [14] (AMNH-20647, left) from Samos (Greece) in anterior (A, F, and K), posterior (B, G, and L), medial (C, H, and M), lateral (D, and N), proximal (J, and O), and distal (E, and P) views. Abbreviations: fc, fovea capitis, gt, greater trochanter; h, femoral head; mc, medial condyle; lc, lateral condyle; lt, lesser trochanter; pat, articular trochlea for the patella; tf, trochanteric fossa; and tt, third trochanter. Scale bar equals 10 cm.
Patella
In total, 11 patellae of C. persiae from Maraghehand one of C. habereri from Kutschwan were studied (Fig 13, Table 6). The articular surface is asymmetrical, with the medial facet being higher than the lateral one, more so in C. persiae than in C. habereri. In C. persiae, the lateral side is narrow than in C. habereri. The base of the patella is rather high, projecting further proximally in C. persiae than in C. habereri. The apex is pointed in both species, but narrower in C. persiae and more rounded in C. habereri (compare Fig 13A, E). The medial process is rather prominent in both species but seems to be projecting slightly further in C. persiae. Metrically, the two species are very close (Table 6, S1 Table 13). The only other chilothere that has a patella with similar proportions is C. orlovi from Pavlodar [12]. The patella of ‘C.’ wimani from the Linxia Basin differs from the patellae of C. persiae and C. habereri, in being higher proximodistally [16]. More specifically, in ‘C.’ wimani the patella is higher than wide, which is not the case in C. persiae and C. habereri, which coincided with the proportions in C. anderssoni from Lok. 30 and C. sarmaticum from Berislav [7,11]. The measurements for a patella of C. schlosseri from Samos [14] are also close to these species and the proportions are a bit different from the patellae of C. habereri and C. persiae (S1 Table 13). The specimen itself was not figured [14] and was probably destroyed during WWII, along with the rest of the chilothere material from Samos that was housed in Munich [10,27,82]. For the patella of C. kowalevskii from Grebeniki no measurements are provided; based on the illustration provided by Pavlow (1913) [15], it seems to be very similar to that of ‘C.’ wimani and C. sarmaticum [16,45]. However, in C. sarmaticum the apex of the patella is much more rounded than in C. kowalevskii and ‘C.’ wimani. In the latter the tip of the apex is especially acute. The proportions of the patellae of C. anderssoni, ‘C.’ wimani, C. kowalevskii, and C. sarmaticum are closer to those seen in Acerorhinus zernowi from Tung-Gur and Aceratherium incisivum from Höwenegg [74,75]. The morphology of the two illustrated patellae of Aceratherium incisivum differ significantly from each other [74: Abb. 49C, 50]. However, it is mentioned that only one is complete and the outline of this specimen (Fig 13N) is intermediate between the very wide patellae of C. habereri and C. persiae and the much narrower ones of ‘C.’ wimani, C. kowalevskii, and C. sarmaticum (Fig 13I–M). Recently, Mallet and Houssaye (2024) [83] found that the shape of the patella in rhinoceroses and extant perissodactyls in general is strongly linked to their phylogenetic affinities and is less affected by functional constraints like body mass. Chilotheres seem to present an unexpected differentiation in the morphology of their patella and may be an exception to the otherwise rather conservative patella morphology among closely related taxa within Rhinocerotidae and other Perissodactyla.
A–D, Chilotherium persiae (Pohlig, 1885) [19] (MNHN.F.MAR3928, left) from Maragheh (Iran), and E–H, Chilotherium habereri (Schlosser, 1903) [24] (GPIT/MA/04781, left) from Kutschwan (China), in anterior (A, E), posterior (B, F), proximal (C, G), and distal (D, H) views. I–N, outline of the patellae of C. habereri (I, drawn from ), C. persiae (J, drawn from Fig 13A), C. kowalevskii (K, drawn from [15]), C. sarmaticum (L, drawn from [11]), 'C.' wimani (M, drawn from [16]), and Aceratherium incisivum (N, drawn from [74]) in anterior view. Abbreviations: a, apex; b, base; lf, lateral facet; mf, medial facet; and mp, medial process. Scale bar equals 5 cm for A–H. I–N are not in scale.
Tibia
Several tibiae of C. persiae from Maragheh, C. habereri from Kutschwan, and C. schlosseri from Samos have been studied (Fig 14, Table 7). However, only a few are adequately preserved to be able to measure the complete length of the bone and study their morphology in detail. They are rather similar with only a slight variability. The bone itself is relatively short, with the proximal epiphysis being quite wide. The lateral aspect of the tibia, on which the fibula is placed, is concave, whereas the medial side is straighter. In proximal view, the medial articular facet for the femur is larger than the lateral one. Both are concave and bend proximally towards the middle of the bone, creating a prominent ridge. Anterior to the articular facets for the femur a strong tibial tuberosityis placed, which functions as an attachment area for the patellar ligaments and merge distally into the tibial crest. The tibial tuberosity bears a wide and rounded tuberosity groove. The shape of the tibial tuberosity differs between C. persiae and C. schlosseri (compare Fig 13A, E). In proximal view, in C. persiae it is more transversally oriented, whereas in C. schlosseri it is anteroposteriorly oriented, being higher than wide. It is separated from the facet by a shallow groove. The central part of the diaphysis is relatively thin compared to the epiphyses, and relatively straight. On the lateral side of the distal epiphysis the rugose surface for the attachment of the fibula has a triangular outline. The articular surface for the astragalus has a subtrapezoidal outline in distal view, wider anteriorly than posteriorly, with the lateral malleolus of tibia and the posterolateral edge protruding.
A–D, Chilotherium persiae (Pohlig, 1885) [19] (A–B, D: MNHN.F.MAR3931 and C: MNHN.F.MAR3933, both right) from Maragheh (Iran), and E–H, Chilotherium schlosseri (Weber, 1905) [14] (GMM 594, left) from Samos (Greece) in anterior (A, E), posterior (B, F), proximal (C, G), and distal (D, H) views. Abbreviations: a, articular facet for the astragalus; df, distal articular facet for the fibula; lf, lateral articular facet for the femur; mf, medial articular facet for the femur; mm, medial malleolus; tc, tibial crest; tg, tuberosity groove; and tt, tibial tuberosity. Scale bar equals 5 cm.
The dimensions of the tibia are rather similar among chilotheres (Table 7, S1 Table 14), with the main differences being found in the total length and the dimensions of the proximal epiphysis [7,11,12,16]. Chilotherium orlovi from Pavlodar exhibits the greatest values for the length of the tibia, with rather large range of 280–340 mm (n = 11), which overlaps completely with the range for the length of the tibia of C. kowalevskii from Grebeniki (279–322 mm, n = 12). These value ranges, also cover the lengths measured in the tibiae of C. schlosseri from Samos (282.9–317 mm, n = 2), including also the tibia mentioned by Weber (1905) [14] from Samos with a length of 300 or 310 mm. The tibia mentioned by Kiernik (1913) [36] from Odessa has a lower length of 270/275 mm. The skull from the same locality can be assigned to C. schlosseri. The fact that the Odessa tibia is slightly smaller than the value range of C. schlosseri, is probably due to the small sample size (n = 2); especially when compared to the broad value range seen in C. orlovi and C. kowalevskii. The length of C. persiae tibiae from Maragheh is also very close to these values (270.7–302.1 mm, n = 7) and is almost identical to the value range given for C. sarmaticum from Berislav (271–301 mm, n = 11). The tibiae of ‘C.’ wimani from the Linxia Basin seem to be placed at the lowest end of this range or even below it (269–282 mm, n = 7). The only measured tibia of C. anderssoni from Lok. 30 (280 mm) and the only tibia of Chilotherium indet. from Kavakdere on which the length could be measured (276 mm) seem to exhibit a somewhat intermediate length, compared to the other chilotheres.
Astragalus
Concerning the astragalus, 14 specimens of C. persiae from Maragheh, two of C. habereri from Kutschwan, and five of C. schlosseri from Samos were studied (Figs 15–17, Table 8). Most of the specimens are fairly well preserved, allowing detailed descriptions. In anterior view, the articular facet for the tibia is fairly asymmetrical, in all studied specimens. The lateral lip of the trochlea is slightly wider and higher than the medial one. In C. habereri, this asymmetry is much more pronounced than in C. schlosseri, where the lips of the trochlea are much more similar in size (Fig 16A, F, K). In medial view (Fig 16C, H, M), the articular stripe for the medial malleolus of the tibia has a concave distal border and it narrows proximally. In lateral view, the articular stripe for the lateral malleolus of the fibula is very wide in C. persiae and C. schlosseri (Fig 16D, N) and less so in C. habereri (Fig 16I), but in all three species it retains a constant width and establishes a wide connection to the ectal calcaneal facet. The contact between that lateral articular facet for the malleolus of the fibula and the trochlea forms a sharp, almost right angle in C. schlosseri, whereas in C. persiae and C. habereri this angle is more rounded and wider. In distal view, the articular facets for the navicular and the cuboid share a wide contact and form a slight ridge between them (Fig 16E, J, O). The former is very large, subtrapezoidal, anteroposteriorly convex, and transversally concave. The latter is a much thinner stripe, which contacts the distal calcaneal facet. In posterior view, the morphology of the calcaneal facets varies significantly between the different Chilotherium species (Figs 16, 17) and some variability is observed within the same species from a given locality [7,11,12,16]. The sustentacular facet is relatively large and its shape can vary from almost circular to high oval, and in some cases can have an irregular outline. In C. habereri and most studied specimens of C. persiae it remains separated from the articular facet of the navicular, cuboid and the distal calcaneal facet. However, in a few specimens a small contact with these facets can be established. In C. schlosseri, two astragali feature a completely isolated sustentacular facet, in two other specimens this portion cannot be studied, and in the last two astragali (GMM 571 and NHMW-GEO-1911/0005/0424) a unique condition is observed, where the sustentacular facet establishes a wide contact with the ectal calcaneal facet. This kind of connection is not observed in any other chilothere astragalus. The astragalus of ‘C.’ wimani from the Linxia Basin (Fig 17H) variably shows a connection between the sustentacular facet, the cuboid facet and the distal calcaneal facet [16]. In C. sarmaticum from Berislav (Fig 17G) the astragalus exhibits a similar connection between the articular facets, but it is much wider and may also involve the navicular facet; this pattern is observed in 10 of the 12 astragali of C. sarmaticum [11]. Additionally, the sustentacular facet is more rounded in ‘C.’ wimani, whereas in C. sarmaticum it is sub-trapezoidal (Fig 17H, G). The species C. anderssoni, C. kowalevskii, and C. orlovi exhibit the plesiomorphic arrangement of articular facets in rhinoceros’ astragali, in which the sustentacular facet is isolated [7,11,12], similar to C. habereri and most C. persiae astragali and as seen in the astragalus of Aceratherium incisivum from Höwenegg (Fig 17A) and Acerorhinus zernowi from Tung-Gur [75]. Concerning the ectal calcaneal facet, it has a similar subtriangular outline in most species (Fig 17). In lateral view it is only little concave, compared to other rhinocerotids like Aceratherium incisivum from Höwenegg [74], in which it has a more prominently sigmoidal profile in lateral view. In all chilotheres, the ectal facet has a distally projecting tongue, which is usually relatively small and subtriangular to suboval in shape. In C. schlosseri, however, this distal tongue has a much bigger surface than in the other species; in this case, C. schlosseri exhibits a more derived morphology.
A, Chilotherium persiae (Pohlig, 1885) [19] (NHMW-GEO-2020/0014/0145, left) from Maragheh (Iran), and B, Chilotherium schlosseri (Weber, 1905) [14] (AMPG-SAM516, right) from Samos (Greece) in anterior view. Abbreviations: ast, astragalus; calc, calcaneum; cub, cuboid; ect, ectocuneiform; mes, mesocuneiform; and nav, navicular. Scale bars equal 5 cm.
A–E, Chilotherium persiae (Pohlig, 1885) [19] (NHMW-GEO-2020/0014/01437, right) from Maragheh (Iran), F–J, Chilotherium habereri (Schlosser, 1903) [24] (GPIT/MA/04779, right) from Kutschwan (China), and K–O, Chilotherium schlosseri (Weber, 1905) [14] (NHMW-GEO-1911/0005/0424, right) from Samos (Greece) in anterior (A, F, and K), posterior (B, G, and L), medial (C, H, and M), lateral (D, I, and N), and distal (E, J, and O) views. Abbreviations: cub, articular facet for the cuboid; dcf, distal calcaneal facet; dpt; distally projecting tongue; ecf, ectal calcaneal facet; ex, distal expansion of the ectal calcaneal facet; lm, articular facet for thelateral malleolus of the fibula, lt, lateral lip of the trochlea, mm, articular facet for the medial malleolus of the tibia; mt; medial lip of the trochlea; nav, articular facet for the navicular; and scf, sustentacular calcaneal facet; Scale bar equals 5 cm.
A, Aceratherium incisivum Kaup, 1832 [80], B, Chilotherium schlosseri (Weber, 1905) [14], C, Chilotherium persiae (Pohlig, 1885) [19], D, Chilotherium habereri (Schlosser, 1903) [24], E, Chilotherium orlovi Bayshashov, 1982 [30], F, Chilotherium kowalevskii (Pavlow, 1913) [15], G, Chilotherium sarmaticum Korotkevitch, 1958 [45], and H, ‘Chilotherium’ wimani Ringström, 1924 [7]. Abbreviations: afc, articular facet for the cuboid; afn, articular facet for the navicular; dcf, distal calcaneal facet; ecf, ectal calcaneal facet; lt, lateral lip of the trochlea; mt, medial lip of the trochlea; and scf, sustentacular calcaneal facet. Not to scale.
A last difference that can be observed in the astragalus of C. schlosseri in comparison to the other chilotheres is that in anterior view the medial lip is distally connected to the articular facet for the navicular, through a rugose bone growth which is observed in the six specimens where the area is observable (in the seventh this part is covered by sediment) In C. persiae and C. habereri the trochlea is separated by the distal articular facets for the navicular and the cuboid by a prominent groove, as is the case in most rhinoceroses.
Overall, it seems very difficult to differentiate between most of the Chilotherium species, solely based on the astragalus morphology. However, C. schlosseri seems to be the exception, because half of the available specimens exhibit a connection between the sustentacular and ectal calcaneal facets, which is not observed in the other chilotheres. Additionally, the same species also exhibits a larger distally projecting tongue on the ectal calcaneal facet and the medial lip of the trochlea is connected to the navicular facet, which is not the case in the other chilotheres. Thereby, C. schlosseri, could be separated from the other Chilotherium species, based on some features. The variability of these features is, however, not clear and therefore the identification of isolated specimens remains ambiguous. Interestingly, the species ‘C.’ wimani and C. sarmaticum feature a wide connection between the sustentacular and ectal calcaneal facets [11,16]. This is also observed in some specimens of C. persiae, but it is not clear whether this could be used as a diagnostic feature for some species or if it is a variable feature, which could potentially be observed in any given species, given a large enough sample.
Calcaneum
In total, 19 calcanei of C. persiae from Maragheh, three of C. habereri from Kutschwan, and three of C. schlosseri from Samos were studied (Figs 15, 18, Table 9). Most are well-preserved and allow a detailed comparison for this bone. The tuber calcis is well-formed in all species and has a rugose surface. On the posterior side of the bone, the surface from the tuber calcis to the distal articular facet for the cuboid exhibits a very rugose secondary bone growth in most bones. The sustentacular facet for the astragalus is rounded in most specimens, varying in shape from almost circular to subtriangular. It is isolated from the other articular facets in all studied specimen. The ectal facet has a generally subtrapezoidal outline and a sigmoidal profile in lateral view. It shares a wide contact with the semi-oval articular facet for the tibia in all three species. However, the tibia facet seems to be larger in C. persiae than in the other two species. The distal facet for the astragalus is elongated and subelliptical in most specimens. Distally it shares a wide contact with the articular facet for the cuboid, which is large and saddle-shaped.
A–F, Chilotherium persiae (Pohlig, 1885) [19] (NHMW-GEO-2020/0014/0139, right) from Maragheh (Iran), and G–L, Chilotherium schlosseri (Weber, 1905) [14] (AMNH-20794, left) from Samos (Greece) in anterior (A, G), posterior (B, H), lateral (C, I), medial (D, J), proximal (E, K), and distal (F, L) views. Abbreviations: cub, articular facet for the cuboid; df, distal articular facet, ef, ectal articular facet; st, sustentacular tali; stf, sustentacular articular facte; tc, tuber calcis. Scale bars equal 5 cm.
The morphology of the studied specimen does not allow a separation between the three species, due to the extreme similarity of the species and the intraspecific variability observed in the Maragheh sample. For instance, the shape of the tuber calcis may vary, in some specimens of C. persiae it is more posteriorly developed than in others. Based on the measurements found in the literature, it seems that C. orlovi from Pavlodar is larger [12], and more specifically has a greater height, than the other chilotheres. Whereas the calcanei of ‘C.’ wimani from the Linxia Basin [16] and C. sarmaticum from Berislav [11] are clearly smaller. On the other hand, the studied calcanei of C. schlosseri, C. persiae, and C. habereri (Table 9) overlap with the dimensions of C. anderssoni from Lok. 30 and C. kowalevskii from Grebeniki [11] and seem to have an intermediate size (S1 Table 17). Therefore, a specific identification based on the calcaneum is rather difficult.
Navicular
Six naviculars of C. persiae from Maragheh, one of C. habereri from Kutschwan, and one of C. schlosseri from Samos were studied (Fig 15, 19A–B). One navicular of C. persiae (NHMW 2020/0014/0145) and the single specimen of C. schlosseri are found in an articulated tarsus and their morphology cannot be assessed (Fig 15B). In proximal view, the articular facet for the astragalus is subrhomboidal and anteroposteriorly concave. In distal view, the three articular facets for the cuneiforms are separated by weak ridges exist. The medial facet for the entocuneiform is the smallest, obliquely placed, and its shape varies from subcircular to almost triangular. The central facet for the mesocuneiform has a subtrapezoidal outline. The lateral facet for the ectocuneiform is the largest and its shape varies from subtrapezoidal to an upturned “L”. In lateral view, two articular facets for the cuboid exist. The proximal one is a narrow, anteroposteriorly oriented stripe, which is separated from the posterodistal facet by a shallow groove. The latter is obliquely placed and has an almost circular outline. The comparison of the studied material to chilotheres naviculars described in the literature shows that they are very similar. More specifically, Deng (2002) [16] described a very similar morphology for ‘C.’ wimani from the Linxia Basin and even mentioned the fact that it bears two separated articular facets for the cuboid. Comparing the dimensions given in the literature for ‘C.’ wimani [16] and C. sarmaticum from Berislav [11], it seems that C. schlosseri and C. sarmaticum have a similar height, which is lower than in C. persiae and ‘C.’ wimani. However, the overall dimensions of ‘C.’ wimani and C. sarmaticum are smaller [11,16] than in C. schlosseri, C. persiae and C. habereri (S1 Table 18). In the case of C. orlovi from Pavlodar [12] the measurements provided for the anteroposterior diameter of the navicular is greater than the transversal length, which is contrasting all other chilotheres and most rhinoceroses in general, therefore we assume that these values are in fact inversed. This would make the dimensions of C. orlovi fall into the size range of C. schlosseri, C. persiae and C. habereri, which is a much more plausible result.
A–B, navicular (MNHN.F.MAR1427a, right) in proximal and distal views. C–G, cuboid (MNHN.F.MAR1414, left) in proximal, distal, medial, and lateral views. H–I, ectocuneiform (MNHN.F.MAR1424c, left) in proximal and distal views. Abbreviations: ast, articular facet for the astragalus; ect, articular facet for the ectocuneiform; ent, articular facet for the entocuneiform; mes, articular facet for the mesocuneiform; MtIII, articular facet for the MtIII; MtIV, articular facet for the MtIV; nav, articular facet for the navicular; and pp, posterior process. Scale bar equals 5 cm.
Cuboid
Three cuboids of C. persiae from Maragheh and only one of C. schlosseri from Samos were available for study (Fig 19C–G). Of these, only two (MNHN.F.MAR1427b and MNHN.F.MAR1414) of C. persiae were able to be studied in detail, because the third specimen of C. persiae and the only one of C. schlosseri are articulated with other tarsal bones (Fig 15). In proximal view, two articular facets are separated by a slight ridge (Fig 19C). The medial facet for the astragalus has a subtriangular outline. The lateral facet for the calcaneum is subtrapezoidal. Together they form a subtrapezoidal surface with an irregular outline. In medial view, four articular facets are present that are connected in pairs, which are separated by a rugose groove (Fig 19E). In both pairs the proximal facet articulates to the navicular and the distal one to the ectocuneiform. The anterior facet for the navicular is a small triangle that contacts the proximal articular facet for the astragalus. The posterior one for the navicular is almost quadrangular and does not contact the facet for the astragalus. The anterior facet for the ectocuneiform is large, subrectangular, concave, and forms an obtuse angle with the anterior facet for the navicular, while also contacting the distal facet for the MtIV. The posterior facet is much smaller and almost quadrangular, forming an almost right angle with the posterior articular facet for the navicular and being separated from the MtIV facet. In distal view, the facet for the fourth metatarsal is large and almost triangular (Fig 19D). In the posterior part of the bone a large, rugose, and distally bent process is placed. The descriptions and illustrations provided for ‘C.’ wimani from the Linxia Basin [16] and for C. orlovi from Pavlodar [12] are generally similar to the morphology observed in the Maragheh specimens, but many details cannot be compared. In both species the articular facet for the astragalus is separated from the anterior facet for the navicular facet, as in C. persiae. In distal view, the anterior articular facet for the ectocuneiform is visible in C. orlovi [12], but is seems small and not concave, thereby possibly differing from C. persiae. The comparison of the metrical data for cuboids of C. kowalevskii from Grebeniki [17], ‘C.’ wimani [16], C. sarmaticum from Berislav [11], and C. orlovi [12] to the measured specimens of C. schlosseri and C. persiae (S1 Table 19) may offer some insight into the association of these specimen. More specifically, C. persiae has a relatively small cuboid with low length values (49.9 and 51.6 mm, n = 2) similar to C. sarmaticum (50.5 and 54.5 mm, n = 2) and ‘C.’ wimani (53–59 mm, n = 3). Whereas C. schlosseri (62 mm, n = 1) seems to be closer to the larger C. kowalevskii (60.8–66.2 mm, n = 6). The species C. orlovi exhibits a length range that overlaps with the other species (55–61 mm, n = 5). Overall, no clear separation of the species can be observed, and the specific identification would currently be impossible based solely on the morphometry of the cuboid.
Ectocuneiform
Three ectocuneiforms of C. persiae from Maragheh and a single one of C. schlosseri from Samos were able to be studied. Only a single specimen of C. persiae is completely preserved (Fig 19H–I) and the single specimen of C. schlosseri is articulated with the other tarsal bones (Fig 15B). In proximal view, the articular facet for the navicular has an upturned “L” shape and is weakly concave. In medial view, in the proximal part an articular stripe for the mesocuneiform exists, while in the distal part two articular facets for the second metatarsal exist, which are separated by a prominent notch. On the lateral side the ectocuneiform articulates to the cuboid. The morphology of the articular facet(s) for the cuboid cannot be assessed due to the damage in this part of the bone. In distal view, the articular facet for the MtIII is generally triangular and resembles a mirrored, upturned “L”. The studied specimens fit the description for the ectocuneiform of ‘C.’ wimani (therein referred to an entocuneiform) from the Linxia Basin [16]. The ectocuneiform has a rather simple and uniform morphology and does not exhibit any diagnostic features. The metrical comparison of C. kowalevskii from Grebeniki [17], ‘C.’ wimani from the Linxia Basin [16], C. sarmaticum from Berislav [11], and C. orlovi from Pavlodar [12] to the measured specimens of C. schlosseri from Samos and C. persiae from Maragheh (S1 Table 20) shows that the dimensions of the ectocuneiform of most species overlaps even within the very small sample sizes. Therefore, any identification based on this bone seems impossible.
Second metatarsal
Three MtII of C. persiae from Maragheh and one of C. schlosseri from Samos were studied (Fig 20, Table 10). Most of them are more or less complete and allow a detailed description and metrical comparison. In proximal view, the articular facet for the mesocuneiform has a semioval outline and laterally contacts the articular facet(s) for the ectocuneiform, forming a slightly obtuse angle. Anterior to the mesocuneiform facet a rugose protrusion is present in the proximal part of the shaft, which extends towards the medial side and continuesdistally. There it connects to a slight tuberosity that is placed in the proximal part of the shaft. In lateral view, the MtII bears two separated articular facets in the only specimen of C. schlosseri (GMM 572) and two specimens of C. persiae (MNHN.F.MAR1378 and MNHN.F.MAR1381). In the third specimen of C. persiae (MNHN.F.MAR1385), a single continuous facet for the ectocuneiform exists. The facets for the ectocuneiform are connected to two separate, small articular facets for the MtIII. In C. schlosseri, the articular facets for the MtIII seem to be somewhat larger compared to those of C. persiae. In both species, the shaft has a subtriangular cross-section in the proximal part and is more oval and transversally elongated in the distal part; however, in C. persiae the cross-section is somewhat more rounded than in C. schlosseri. In medial view, on the proximal half of the shaft of C. persiae a well-developed, bulbous rugosity for the attachment of the interosseous ligament is placed (Fig 20A); in C. schlosseri, this rugosity covers a similar surface but is much more weakly developed (Fig 20G). On the anterior side of the bone, above the distal trochlea two bilateral protuberances are placed for the attachment of the ligaments of the fetlock joint. These are much more pronounced anteriorly in C. schlosseri than in C. persiae. The distal articular head is relatively rounded, asymmetrical with a convex proximal border on the anterior side. The sagittal keel of the articular head is rather weak and located mainly in the posterior part of the bone. On the posterior side, above the articular head two bilateral rugose protrusions are located.
A–F, Chilotherium persiae (Pohlig, 1885) [19] (MNHN.F.MAR1385, left) from Maragheh (Iran), and G–L, Chilotherium schlosseri (Weber, 1905) [14] (GMM 572, right) from Samos (Greece) in anterior (A, G), posterior (B, H), medial (C, I), lateral (D, J), proximal (E, K), and distal (F, L) views. Abbreviations: ect, articular facet for the ectocuneiform; mes, articular facet for the mesocuneiform; MtIII, articular facet for the MtIII Scale bars equal 5 cm.
The morphological differences observed in the MTII of the two species are rather slight. Metrically, however, there seems to be a distinction between larger and smaller species (Table 10, S1 Table 21). This is evident in species, where a larger sample size has been reported, like C. kowalevskii from Grebeniki, C. sarmaticum from Berislav, and C. orlovi from Pavlodar [11,12,17]. All three of these species, exhibit similar values for the transversal and the anteroposterior diameter of the diaphysis (TDdia and APDdia) of the MtII, but differ in some other dimensions of the bone and most prominently in the length. More specifically, C. kowalevskii and C. orlovi exhibit a higher value range for the length of the MtII (94.6–114.2 mm, n = 13 and 98–106 mm, n = 7 respectively) than C. sarmaticum (86.2–93 mm n = 6). Based on these well-sampled species, it is possible to associate C. schlosseri (102–102.4 mm, n = 2), C. persiae (93.4–97.3 mm, n = 3), C. anderssoni (97–99, n = 3), and a potential primitive Chilotherium (100.7 mm) with the larger species C. kowalevskii and C. orlovi [7,14,81]. Additionally, C. schlosseri from Odessa (95–110 mm) also falls into this size range. On the contrary, ‘C.’ wimani (89–94 mm, n = 3) and Chilotherium indet. from Kavakdere (85.2 mm) may be closer to the smaller C. sarmaticum [16,81]. It is rather surprising, that the MtII of the potential primitive Chilotherium from Loc. 72 of the Sinap Formation [81] has a length that is close to the length of species that are considered derived, such as C. schlosseri, C. anderssoni, and C. orlovi (S1 Table 21).
Third metatarsal
Three MtIII of C. persiae from Maragheh and four specimens of C. schlosseri from Samos were studied (Fig 21, Table 11). Most of them are more or less complete and allow a detailed description and metrical comparison. Three specimens of C. schlosseri and one of the MtIII of C. persiae are completely preserved, except for some slight damage. The other specimens are lacking the distal half. In proximal view, the articular facet for the ectocuneiform covers the complete proximal surface of the bone and is almost triangular. The anterior border of the articular facet for the ectocuneiform is almost straight in C. schlosseri and has a very slight indentation in MNHN.F.MAR1382 of C. persiae. In medial view, two distinct facets for the MtII are visible in all specimens of both species; both facets are small and have a semi-oval outline. On the lateral side, three out of the four specimens of C. schlosseri feature an articular facet for the cuboid. In the fourth specimen (AMNH-22818) the relevant portion is somewhat damaged, and the potential presence of a cuboid facet is not clear. In the three specimens of C. schlosseri where it can be observed, the cuboid facet is small and trapezoidal, placed between the articular facet for the ectocuneiform and the anterior one for the MtIV (Fig 21G, J). In contrast to that, all three studied C. persiae specimens lack a cuboid facet (Fig 21B, D). In lateral view, two articular facets for the MtIV are placed below the ectocuneiform facet. These two facets are separated by a deep groove, which is much wider in MNHN.F.MAR1382 of C. persiae. In anterior view, the shaft widens distally towards the bilateral protuberances for the attachment of the ligaments of the fetlock joint. In posterior view, the sagittal keel of the articular head for the proximal phalanx is well developed. It placed in the posterior part of the articular head and in distal view, reaches the level of its medial rim in distal view.
A–F, Chilotherium persiae (Pohlig, 1885) [19] (MNHN.F.MAR1385, right) from Maragheh (Iran), and G–L, Chilotherium schlosseri (Weber, 1905) [14] (GMM 572, right) from Samos (Greece) in anterior (A, G), posterior (B, H), medial (C, I), lateral (D, J), proximal (E, K), and distal (F, L) views. Abbreviations: cub, articular facet for the cuboid; ect, articular facet for the ectocuneiform; MtII, articular facet for the MtII; and MtIV, articular facet for the MtIV. Scale bars equal 5 cm.
The presence of a cuboid facet in MtIII may present a diagnostic feature separating C. schlosseri from at least some of the other species like C. persiae (Fig 21D, J). The type species C. anderssoni from Lok. 30 also lacks the cuboid facet [7]. Deng (2002) [16] mentioned that in both C. anderssoni and ‘C.’ wimani from the Linxia Basin (China) the MtIII does not bear an articular facet for the cuboid. Similarly, no cuboid facet has been described for C. orlovi from Pavlodar and in the illustrated MtIII, no such facet seems to be present between the articular facet for the ectocuneiform and that for the MtIV [12]. A MtIII from the Upper Miocene locality Küçükçekmece was assigned to C. schlosseri but was described as not having a cuboid facet. It is likely that this specimen, as well as the other Chilotherium specimens from Küçükçekmece actually belong to a different species and not C. schlosseri, which bears a cuboid facet. Unfortunately, for most chilotheres this condition has not been described or illustrated.
Metrically, however, there seems to be a slight distinction between larger and smaller species. More specifically, in C. kowalevskii from Grebeniki, C. sarmaticum from Berislav, and C. orlovi from Pavlodar the length of the MtIII seems to reach up to 120 mm, in the largest specimens [11,12,17]. The measured specimens of C. schlosseri (111.2–120 mm, n = 6) and C. persiae (110 mm) seem to have similar values to these species, with C. schlosseri having a mean value of about 116 mm, which is close to, but even higher than, the mean value of 113 mm seen in C. kowalevskii [11,17]. Respectively, C. anderssoni from China has a similar range for the length of the MtIII (110–118 mm, n = 3) as in these larger species [7]. The other Chinese species, ‘C.’ wimani, on the other hand exhibits a rather low value range for the length of the MtIII (100–109 mm, n = 3), not surpassing 110 mm. Although the sample size is rather small for most species – in many instances three specimen or less – it seems that ‘C.’ wimani differs somewhat from the other species, in having smaller dimensions. Similarly, the measured MtIII of a potential primitive Chilotherium from Loc. 72 of the Sinap Formation (106 mm) and of Chilotherium indet. from Kavakdere (101.5 mm) are below the value ranges of C. kowalevskii and C. orlovi or barely reaching the lowest values. It seems that the material from Turkey [81] has more similar values to ‘C.’ wimani and are also within the value range of C. sarmaticum.
Fourth metatarsal
Four MtIV of C. persiae from Maragheh and two specimens of C. schlosseri from Samos have been studied (Fig 22, Table 12). Of these, only one specimen of each species (NHMW-GEO-2020/0014/0144 and GMM 572, respectively) is adequately preserved. The others are missing their distal parts in most cases. In proximal view, the articular facet for the cuboid is subtrapezoidal in GMM 572 of C. schlosseri and almost kidney shaped in NHMW-GEO-2020/0014/0144 of C. persiae. Below the articular facet for the cuboid, on the anterior side a small tubercle is present medially and a large one laterally, that covers the complete lateral side of the bone and weakens towards the posterior side. In medial view, the two articular facets that articulate to the MtIII vary in morphology between the two species. In C. schlosseri the anterior one is more elongated and thinner, while the posterior one is larger and rounded. In C. persiae the posterior one is similar to C. schlosseri, while the anterior one differs in being semi-oval and not elongated. Additionally, in C. persiae on the medial side, between the cuboid facet and the anterior MtIII facet, a small rectangular facet for the ectocuneiform is placed (Fig 22A, C, E). This differs from the morphology observed in C. schlosseri, where this facet is not found (Fig 22I). However, in both species the posterior MtIII facet seems to curve proximally, creating a very minute oblique surface, which could represent a ectocuneiform facet (Fig 22B, C, E, H, I, K), but cannot be confirmed, since there is no articulated set of tarsals and metatarsals. However, in both species the proximal portion of the posterior articular facet for the MtIII curves up, possibly creating a small articular surface for the ectocuneiform. Below the articular facets for the MtIII a prominent rugosity for the attachment of the interosseous ligament extends slightly beyond the middle of the shaft. This rugosity is more pronounced in C. persiae than in C. schlosseri (Fig 22A–C, G–I). The shaft of the bone is more or less straight. On the articular head for the proximal phalanx, there is only a very weak sagittal keel on the posterior side. In anterior view, the articular head is slightly asymmetrical.
A–F, Chilotherium persiae (Pohlig, 1885) [19] (NHMW-GEO-2020/0014/0144, right) from Maragheh (Iran), and G–L, Chilotherium schlosseri (Weber, 1905) [14] (GMM 572, right) from Samos (Greece) in anterior (A, G), posterior (B, H), medial (C, I), lateral (D, J), proximal (E, K), and distal (F, L) views. Abbreviations: cub, articular facet for the cuboid; ect, articular facet for the ectocuneiform; and MtIII, articular facet for the MtIII. Scale bars equal 5 cm.
The morphology of the articular facets for the tarsals and the MtIII may provide some insight into the separation of the species, as C. persiae and C. schlosseri exhibit significant differences. Most importantly, the existence of an ectocuneiform facet in the anterior part of the MtIV in C. persiae (Fig 22A, C, E) or the lack of it in C. schlosseri (Fig 22G, I, K) may present a diagnostic feature. Unfortunately, the presence of this facet cannot be assessed in the other chilotheres, as the descriptions and illustrations in the literature are not sufficient. However, this articulation is present in the non-chilothere aceratheriine Aceratherium incisivum from Höwenegg [74: Abb. 63]. No illustrations of the medial side of the MtIV is provided for C. kowalevskii from Grebeniki, but in anterior view it looks like such a articular facet for the ectocuneiform might be present [15,17]. The metrical comparison on the other hand seems to be more comparable (S1 Table 23). The species C. kowalevskii from Grebeniki and C. orlovi from Pavlodar have larger dimensions (length: 91.8–100.5 mm, n = 9 and 90–105 mm, n = 4, respectively) than C. sarmaticum from Berislav (length: 83.5–87 mm, n = 4), which do not overlap [11,12,17]. The length of the MtIV of C. schlosseri (91.7–96 mm, n = 4) and C. persiae (94.3 mm) fit much better the value ranges of the larger C. kowalevskii and C. orlovi. Whereas the dimensions of ‘C.’ wimani from the Linxia Basin (83–90 mm, n = 4), the potential primitive Chilotherium from the Sinap Formation (106 mm), and Chilotherium indet. from Kavakdere (83.4 mm), fit better those of the smaller C. sarmaticum [16]. Therefore, despite the small sample size in most species, a separation into two size classes can be observed, similar to observation by other authors [84].
Discussion
Detailed comparisons of the cranial and dental material of different chilothere taxa [e.g., 8,10,49] suggest that their separation is a rather delicate topic. Nonetheless, based on certain features in the skull most species can be distinguished. For instance, it was recently shown that Eochilotherium samium differs significantly from most members of the genus Chilotherium and should be placed in its own genus within the Chilotheriina [10]. It was further established that two other species, ‘C.’ wimani and ‘C.’ primigenium, deviate from the typical Chilotherium morphology as exemplified by the type species C. anderssoni and probably also represent distinct genera [10,39]. Nonetheless, the tooth morphology of these species is remarkably consistent with Chilotherium and differs only in few features.
Within Chilotherium sensu stricto, the differentiation of species is less clear, as the skulls are very similar. The species C. schlosseri proves to be among the most recognisable species of the genus based on features such as the very prominent dorsally depressed skull and the very widely separated parietal crests (>70 mm, n = 12 [39]). Even the position of the nasal bones is affected by the frontal depression, and they are placed comparatively lower than in the other species, with the nasal bones being placed below the level of the dorsal border of the orbit in lateral view, in contrast to the other chilotheres [10,39]. However, there are characters that are able to also separate some of the other species. For instance, in the species C. schlosseri and C. anderssoni the skull is flattened, and the dorsal surface has a straight profile [10]. Whereas in species like C. persiae, C. kowalevskii and C. sarmaticum the skull is less flattened with the nuchal region being raised. Another morphological feature that could be used to separate some species is found in the frontal depression. More specifically, in some species like C. persiae and C. kowalevskii a faint longitudinal ridge is placed along the middle of the skull and is visible within the frontal depression [39]. This feature is not present in C. schlosseri and C. anderssoni [39]. Based on these features also a skull from the Upper Miocene locality of Reghiu (Romania), is morphologically closer to C. kowalevskii than to C. schlosseri. The two species were considered synonyms in the most recent re-evaluation of this skull, which led to the identification of the specimen as C. schlosseri, despite its closer affinities to C. kowalevskii [47]. This skull includes both a somewhat raised nuchal crest and a longitudinal ridge in the middle of the frontal depression like C. kowalevskii, also the parietal crests are 64.2 mm apart, thereby closer together than in C. schlosseri (>70 mm, n = 12), falling into the range of C. kowalevskii (40.1–66, n = 10) [10,17,39,47].
Unfortunately, in most cases within chilotheres the tooth morphology cannot solve the species determination. The lower teeth have an especially conservative morphology and are characterised by a notable uniformity within the group [8]. Despite that uniformity, the lower teeth can be used for the separation of chilotheres from other rhinoceroses, because in chilotheres the premolars are relatively shortened, resulting in a lower ratio of premolar to molar length [e.g., 42]. The morphology of the upper teeth is much more characteristic and can be used to separate chilotheres from all other rhinoceroses, but also includes some diagnostic features for the identification of some species within chilotheres. For instance, in the species E. samium, ‘C.’ wimani and ‘C.’ primigenium the premolars have either weakly constricted or unconstricted protocones. In the case of the P2, both the protocone and the hypocone are unconstricted and the median valley remains open even at a very late wear stage in these three species [10]. Also, the upper tooth morphology of Shansirhinus deviates from other chilotheres, exhibiting a prominent paracone fold and often a large number of enamel plications [73,76]. The sporadic occurrence of enamel plications is also observed in other chilotheres like C. schlosseri, where they are rather frequent, but not as abundant as in Shansirhinus [10].
Therefore, the taxonomy of the chilotheres has historically been very complicated and inconsistent but has started to be illuminated. In the present work we compare the postcranial material of chilothere species across Europe and Asia, in order to further elucidate the taxonomy of this group. Despite the relative abundance of cranial material from members of this group, the postcranial material is much more limited and rarely described in detail. This is probably due to a sampling bias towards skulls, which was especially common in the past. For instance, the Chilotherium material from the Upper Miocene of Kutschwan includes 11 cranial elements, but only 23 postcranial ones. Similarly, in the case of C. schlosseri, there are eight complete or almost complete skulls in different collections [39], but only very few collections house postcranial material of the species, namely the AMNH, AMPG, GMM, and NHMW.
Nonetheless, when Ringström (1924) [7] coined the genus Chilotherium, he distinctly mentioned the rather peculiar anatomy of the appendicular skeleton that characterises this rhinocerotid group. He specifically noted that the extremities are strongly shortened, both the manus and pes are tridactyl with short metapodials, and that the abaxial metapodials are obliquely posteriorly oriented in comparison to the central one [7]. However, both in the work of Ringström (1924) [7] and the rest of the literature that was used in this context included rather short descriptions and comparisons of the postcranial elements of Chilotherium and their usefulness for taxonomic purposes was rather limited, as mainly the cranial and dental material was used to identify the species [e.g., 7,14,15]. However, the herein conducted comparison of the postcranial material of chilotheres confirms that there are several distinct features that characterise this genus, as already proposed by Ringström (1924) [7]. Most of these features are related to the shortening of the limbs. For instance, among the carpal bones the scaphoid is proximodistally notably shortened when compared to non-chilothere rhinocerotids for instance (Fig 23), with the only exception being teleoceratines like Brachypotherium brachypus. More specifically, most chilothere scaphoids have a height to length ratio between 50% and 69%, whereas in other Miocene rhinoceroses like Dihoplus pikermiensis and Acerorhinus zernowi this ratio can be higher, usually being above 70% and reaching up to 85% (Fig 23A). It has to be noted that this ratio is highly affected by the strongly proximodistally developed posterior tuberosity. In direct comparison it is clearly visible that the scaphoid of chilotheres is much shorter than that of extant horned rhinoceroses for instance (Fig 23B–E). More specifically, the anterior height is much lower than the posterior height in Chilotherium. This is a phylogenetically informative character also used in current morphological character matrices for rhinocerotids [34]. If only the anterior height was measured and used for the calculation, the ratio would be much lower. For C. habereri specifically instead of 66% it would be 53%, whereas in rhinocerotines the difference between the anterior and the posterior height in much less pronounced (Fig 23D–E). In most cases in the literature, only the maximal height is provided, without stating where it was measured, which complicates direct comparisons (Fig 23). In Aceratherium incisivum from Höwenegg (Germany) this ratio (70–73.4%, n = 4) is close to that of ‘C.’ wimani, which shows the least degree of shortening of the limbs among any chilothere (Fig 23). In the European teleoceratine Brachypotherium brachypus the ratio (58.7–60.5%, n = 4, based on [85] and own data) is close to that of the chilotheres. This is also expressed in its morphology, which is much closer to that of C. schlosseri (compare Fig 23B, C). Though, the posterior tuberosity is not as strongly developed in Brachypotherium brachypus from Steinheim (Fig 23C). In the American short-limbed hornless rhinocerotid Teleoceras, the scaphoid seems to be similarly compressed (50–55%, n = 3) as in Chilotherium orlovi [86], which exhibits the lowest value of all chilotheres (Fig 23).
A, univariate diagram of the height (H) to maximal length (L) ratio of the scaphoid of chilotheres compared to other rhinocerotids. B, left scaphoid of Chilotherium schlosseri (Weber, 1905) [14] (NHMW-GEO-2009z0089/0001) from the Upper Miocene of Samos (Greece); C, right scaphoid of Brachypotherium brachypus (GPIT/MA/03454, mirrored) from the Middle Miocene of Steinheim in Germany; D, left scaphoid of extant Dicerorhinus sumatrensis (NHMW-1500); and E, left scaphoid of extant Ceratotherium simum (NHMW-3086) in anterior view. Scale bar equals 5 cm. Measurements can be found in S1 Tables 4.
The most characteristic portion of the shortened limbs of chilotheres are the metapodials [7]. These are rather stout, especially when compared to other rhinocerotids (Fig 24) like the contemporaneous tandem-horned rhinocerotines Dihoplus pikermiensis and ‘Ceratotherium’ neumayri or the Middle Miocene aceratheriine Acerorhinus zernowi from Tung Gur (China) [75]. Aceratherium incisivum from Höwenegg [74] seems to be somewhat closer to Chilotherium than to Acerorhinus zernowi, especially concerning the dimensions of the MtIII (Fig 24). In the MtIII diagram specifically, it is visible that there is a broad range in the values of ‘C.’ wimani, which is however lower than that of C. kowalevskii for instance (Fig 24). The Middle Miocene European teleoceratine Brachypotherium exhibits similar ratios to the chilotheres (McIII: 39–46.1%, n=9, and MtIII: 38–47%, n=12 [85]), which overlap to an important degree with the ratios of C. kowalevskii (Fig 24). The American teleoceratine Teleoceras also exhibits similarly short or even shorter metapodials than Chilotherium [87], as indicated by the single data points for each species (Fig 24). A McIV (MNHN.F.TRQ329) from Küçükçekmece (Turkey) that was attributed to C. schlosseri [46], exhibits proportions that do not fit any chilothere. This McIV from Küçükçekmece is more slender (TDdia/L=22.3%) than the McIV of the chilothere compared herein (27.5–34.8%, n=14). Instead, its’ proportions fit much better those of Acerorhinus zernowi from Tung Gur (19.8–23.9%, n = 3) [75]. In the same locality, the presence of another hornless rhinocerotid, Persiatherium sp., has been reported, which is not a chilothere and most likely had a more elongated appendicular skeleton. Therefore, this McIV (MNHN.F.TRQ329) could possibly belong to Persiatherium instead of Chilotherium. The MtIII from the same locality (MNHN.F.TRQ335) that was also assigned to C. schlosseri was described as not bearing a cuboid facet. However, in all C. schlosseri specimens from Samos that preserve the relevant portion a cuboid facet is indeed present. Therefore, it is most plausible that the MtIII MNHN.F.TRQ335, does not belong to this species. On the other hand a McIV from Loc. 26 in the Sinap region referred to as ‘Acerorhinus sp. nov.’, has an index of 30.3% that falls well into the value range of the chilotheres (27.5–34.8%, n = 14) and the overall morphology of this bone also matches the morphology seen in Chilotherium. Therefore, it is most likely that this bone actually does belong to Chilotherium or another chilothere and not Acerorhinus.
Univariate diagrams of the proximal transversal diameter (DΤprox max) to length (L) ratio of the McIII and MtIII of chilotheres compared to other rhinocerotids. Measurements can be found in S1 Tables 10 and 22.
The trend of shortening the limbs is also demonstrated in other bones of the appendicular skeleton. For instance, the humeri and femora in chilotheres are also relatively stouter than in other aceratheriines, like Aceratherium incisivum and Acerorhinus zernowi (Fig 25). Within rhinocerotids only teleoceratines have reached a similar degree of limb shortening, as exemplified by Teleoceras major and Teleoceras fossiger in Fig 25. Although, in the Middle Miocene European teleoceratine Brachypotherium brachypus the femur is not quite as shortened (15.8%, n = 1 [85]), still being comparable to the lowest values of Chilotherium (Fig 25).
Univariate diagrams of the transversal diameter of the diaphysis (TDdia) to length (L) ratio of the humerus and femur of chilotheres compared to other aceratheriines. Measurements can be found in S1 Tables 1 and 12.
Within chilotheres there seems to be a gradual increase in size and further shortening of the limbs. More specifically, E. samium and ‘C.’ wimani, which were considered as more “primitive” chilotheres, are somewhat smaller based on the dental and cranial material [e.g., 10]. This is also reflected in the appendicular skeleton, as confirmed by the comparison of the published postcranial data of ‘C.’ wimani from the Linxia Basin (China) to that of other chilotheres (Figs 23–25, S1 Table). The only other well-sampled chilothere that seems to be comparable in size is the chronologically oldest European representative, C. sarmaticum from the Vallesian of Berislav (Ukraine) [11,45]. Whereas the younger, middle to late Turolian chilotheres C. schlosseri from Samos (Greece) and C. orlovi from Pavlodar (Kazakhstan) seem to be the largest chilotheres [84], while at the same time C. orlovi also seems to be among the species with the most pronounced shortening of appendicular skeleton. The only chilothere of younger geological age is Shansirhinus, which is, however, exclusively known from craniodental material. For the longest time, only teeth were known from this species [24,73,88] and only recently complete skulls were described from Upper Miocene and Lower Pliocene deposits from China [76,89,90], making it the youngest reported representative of the chilotheres until now and the only chilothere surviving into the Pliocene. Nonetheless, no postcranial material has been described for any Shansirhinus species to be used in our current comparisons.
The possible attribution of the aceratheriine material (a partial forelimb) from Loc. 12 of the Upper Miocene Sinap Formation in Turkey [81: fig. 12.15] to the genus Chilotherium is herein rejected, due to the existence of a well-developed, functional, and not at all reduced McV. It is known that in Chilotherium the McV is represented by a small and sesamoid-like rudimentary bone similar to what is observed in extant rhinoceroses and also seen in the herein studied C. habereri material from Kutschwan. That is why Chilotherium has been widely considered as tridactyl [3,7,16]. The condition of the Sinap specimen is more similar to Aceratherium incisivum from Höwenegg, which also includes a well-developed McV [74]. The shape and size of the McV from Sinap and those from Höwenegg are extremely similar. Also, the other associated metacarpals from Loc. 12 of Sinap, are overall similar in proportions and form to the metacarpals of Aceratherium incisivum from Höwenegg. The gracility index of these bones (McII: 28.5%, McIII: 27.9%, and McIV: 25.8%) is very close to the values in Aceratherium incisivum from Höwenegg (McII: 27.2–27.6%, McIII: 29.2–29.6%, and McIV: 24.3–27.4%, n = 4). Nonetheless, the Sinap material, with a length of 122 mm in the McIII from Loc. 12 [81], is smaller than Aceratherium incisivum from Höwenegg (140–144 mm, n = 4) [74]. It is more likely that the associated forelimb from Loc. 12 represents a non-chilothere aceratheriine, possibly closely related to Aceratherium incisivum; however, solely based on such material it is currently impossible to attempt any taxonomic attribution.
Bringing together the information from the literature based on the craniodental material that can be used for the identification of the chilotheres species and combining it with the results of the current study of the postcranial material refines their differentiations. The current study showed that one of the most promising elements for taxonomic distinction may be the astragalus. More specifically, the arrangement of the articular facet for the calcaneum seem to differ in some species. In C. anderssoni, C. habereri, C. persiae, C. kowalevskii, and C. orlovi all articular facets are well separated from each other [7,11,12], similar to Aceratherium incisivum from Höwenegg (Fig 17). The species ‘C.’ wimani and C. sarmaticum, exhibit a connection between the sustentacular facet, the cuboid facet and the distal calcaneal facet (Fig 17) [11,16]. In C. schlosseri, the condition of these articular facets varies intraspecifically, with two astragali (GMM 571 and NHMW-GEO-1911/0005/0424) exhibiting a wide connection between the sustentacular facet and the ectal calcaneal facet, which is not observed in any other chilothere species (Fig 17B). The astragalus morphology of C. schlosseri further differs from the other chilotheres by having a connection between the medial lip for the articulation of the tibia and the articular facet for the navicular (Fig 16K). This character is observed in all available astragali of C. schlosseri; in other chilotheres there is a well-developed groove separating the medial lip and the navicular facet. Thereby it can be assumed that these features are apomorphic for C. schlosseri and can be added to the diagnostic characters for the species.
The studied axes of C. persiae and C. schlosseri also differ slightly from each other (Fig 2). The ventral crest of the axis is wider and looks more massive in C. schlosseri, whereas in C. persiae it seems compressed in its’ middle. The posterior articular facet for third cervical vertebra has a high-oval shape with a slight, dorsal indentation in C. schlosseri, whereas in C. persiae it is more rounded, and the dorsal surface is flat. Additionally, in ventral view there is an indentation on the posterior side in C. persiae, whereas C. schlosseri does not feature such an indentation. Lastly, the area where the lateral walls of the neural canal of the axis connect to the vertebral body seems to extend more posteriorly in C. schlosseri than in C. persiae. Due to the small number of specimens, it is not possible to assess the intraspecific variability of these characters or whether they are also present in other species. In the scapula, the shape of the medial tuberosity, is more rounded in C. persiae and C. schlosseri and rather flattened in C. habereri but seems to be more protruding in C. schlosseri (Fig 3). In the humerus, the dimensions of the species seem to be most informative, since ‘C.’ wimani and E. samium have the smallest humeri [14,16], C. orlovi has the largest humeri [12] and most other species, C. persiae, C. habereri, C. schlosseri, and C. kowalevskii, have humeri with intermediate dimensions. Additionally, C. schlosseri exhibits the most proximodistally compressed humerus among the compared species, being comparable to Teleoceras (Fig 25).
The patella also could be helpful for the taxonomic separation of chilotheres (Fig 13). The current comparison showed that there are two different morphotypes in the patella of chilotheres. The probably plesiomorphic ‘C.’ wimani has a more proximodistally elongated patella than C. habereri and C. persiae. In the latter two species, the patella is much wider. In the case of C. persiae, the wider form of the patella corresponds to the mediolaterally oriented tuberosity in the proximal epiphysis of the tibia. If this is true, it would mean that C. schlosseri, that has a more anteroposteriorly oriented tuberosity in the tibia, would have a patella that is more similar to that of ‘C.’ wimani. Besides the significance for taxonomy and phylogeny, patellar morphology is known to be associated with ecological adaptations [91]. In primates for instance, it has been shown that a proximodistally elongated patella is associated with running, climbing, and leaping behaviour, whereas the wider patella found in great apes probably reflects a more mobile knee joint [e.g., 92,93]. Recently, it was suggested that in extant Perissodactyla patellar morphology is rather conservative, being correlated to their phylogenetic affinities [83]. Chilotheres seem to be an exception to this rule, and the patellar morphology may well be informative also for their ecology, with species like ‘C.’ wimani that have a more proximodistally elongated patella being potentially somewhat more cursorial. However, additional material and analyses are needed to confirm such correlations.
Another postcranial element that seems to offer some taxonomic insight is the tibia. The dimensions of the bone already seem to help in the identification of the taxon, as some species like ‘C.’ wimani are fairly small. More interestingly, the morphology of the lateral tuberosity of the proximal articulation, for the attachment of the patellar ligaments, is differently shaped in C. schlosseri and C. persiae (Fig 14C, G). In proximal view, in the former one it is anteroposteriorly oriented, being higher than wide, and in the latter one it is more mediolaterally oriented being wider than high. Unfortunately, the proximal view of the tibia is not provided for any other chilothere, therefore it is not possible to assess the state of this character in any other species.
Lastly, metapodial bones can be useful for the separation of some species. Their dimensions overall, and most prominently the length, are diagnostic. Korotkevitch (1970) [11], showed that C. sarmaticum from Berislav has shorter metapodials than C. kowalevskii from Grebeniki, based on the comparison of his own measurements to those provided by Krokos (1917) [11,17]. However, this study did not take into account data available from other localities such as Samos [14], Odessa [36], and the Red Clay localities from China [7]. Additionally, since then, there have been some new studies that have described postcranial material of chilotheres [12,16,46,81], which need to be considered. The combined comparison of the data from the literature and the newly studied material of C. persiae from Maragheh, C. habereri from Kutschwan, and C. schlosseri from Samos does not directly confirm the assumption of Korotkevitch (1970) [11], but shows that the chilothere species exhibit varying degrees of shortening of the limbs. Instead of C. sarmaticum having significantly shorter metapodials, it seems that the degree of shortening is similar, possibly even lower than that seen in C. kowalevskii (Fig 24). However, it seems that there is a general separation of chilotheres into two size classes. The large size group is represented by C. kowalevskii, C. schlosseri, C. anderssoni, C. persiae, and C. orlovi, whereas C. sarmaticum, ‘C.’ wimani, Chilotherium indet. from Kavakdere, and possibly E. samium have smaller dimensions. A similar observation was also noted by Bayshashov et al. (2024) [84], who mentioned three size groups, with C. sarmaticum representing the small one, C. schlosseri (considered as the senior synonym of C. kowalevskii by these authors) and C. anderssoni representing the intermediate one, and only C. orlovi representing a distinctly larger size group. While C. orlovi seems to be the largest chilothere known to date, in our current study we divide the chilotheres in only two size groups, with C. orlovi being part of the large size group, for the purpose of this comparison. Additionally, there seems to be a temporal trend towards further shortening of the limb and becoming larger through time within the chilotheres. This is exemplified by the fact that the stratigraphically oldest chilotheres, like C. sarmaticum and ‘C.’ wimani from the Vallesian, are in fact the smallest and have relatively longer limbs compared to younger chilotheres like C. schlosseri and C. orlovi from the middle to late Turolian.
Overall, these new findings that are based on the postcranium confirm previous assumptions about chilotheres being characterised by shortened limbs compared to other rhinocerotids [e.g., 3,7,16]. The only comparable group that independently acquired a similar bodyplan is the Teleoceratini [e.g., 86,87]. When comparing chilotheres with other Eurasian rhinoceroses that lived during the Late Miocene, like ‘Ceratotherium’ neumayri, Dihoplus pikermiensis, and Aceratherium incisivum [27,74,94–96], it becomes clear that chilotheres differ significantly from the other members of the Rhinocerotidae and that despite often living in the same environment they were adapted to a different lifestyle. The results of our current study seem to also generally agree with what has been inferred from skull comparisons, concerning the relationship between different chilotheres. More specifically, it was already suggested that ‘C.’ wimani represents a more plesiomorphic chilothere, based on the very little separated parietal crests, the convex frontal region, the concave dorsal profile of the skull, the highly raised caudal portion of the skull, and the relatively simple tooth morphology [9,10,37]. This idea is corroborated by the comparatively small postcranial elements that are not quite as shortened as in other chilotheres (Fig 24). On the other hand, C. schlosseri was suggested to be a highly derived chilothere based on the strongly depressed frontal and nasal region of the skull, the generally flat and wide skull shape, and the complicated secondary enamel folds in the upper dentition [8,10]. This is further supported by the relatively large size and short appendicular elements, along with the derived morphology of the astragalus that differs from all other chilotheres.
Unfortunately, for many non-Chilotherium chilotheres, such as E. samium, Shansirhinus spp., and ‘C.’ primigenium, no postcranial material is available [10,14,37,76,89,90]. This makes it impossible to assess how widespread the shortening of the limbs was within the chilotheres. Nonetheless, the fact that ‘C.’ wimani – a species that is generally considered to be a “primitive” chilothere – exhibits shortened limbs in comparison to non-chilothere rhinocerotids, suggests that it is a plesiomorphic character within the chilothere group. It was recently suggested that ‘C.’ wimani should actually represent a different genus within the chilotheres, because it lacks one of the key autapomorphies of the genus, which is the depression in the frontal region of the skull, similar to E. samium and ‘C.’ primigenium, and unlike C. anderssoni and C. schlosseri [10]. The results of our study also support the notion that ‘C.’ wimani differs from the typical species of Chilotherium. In the case where ‘C.’ wimani is removed from the genus Chilotherium sensu stricto, the shortened limbs may actually prove to be a shared feature within chilotheres in general.
Conclusion
The postcranial material of the hornless rhinocerotid group Chilotheriina has received little attention throughout their research history. Most studies have focussed on the skull and more specifically on the dental morphology. Herein, we describe the postcranium of three chilothere species: Chilotherium persiae from the Upper Miocene of Maragheh (Iran), Chilotherium habereri from the Upper Miocene of Kutschwan (China), and Chilotherium schlosseri from the Upper Miocene of Samos Island (Greece). We compared the newly described material with respective material of other chilotheres that is available in the literature and found some characters that seem to be taxonomically informative, including features in the scaphoid, tibia, patella, astragalus, and metatarsals. Additionally, we pointed out features that vary between the species, but due to the limited material their significance can presently not be evaluated. The new findings confirm that the shortening of the limbs is a key feature shared within chilotheres and support assumptions about the relationships between some species that were based only on cranial material until now.
Supporting information
S1 Table. Measurements of the studied material and species compared from the literature.
https://doi.org/10.1371/journal.pone.0336590.s001
(XLSX)
Acknowledgments
We would like to thank J. Meng (AMNH), S. Roussiakis, G. Theodorou (AMPG), S. Trümper (GMM), U. Kotthoff (GPIH), I. Werneburg (GPIT), R. Rozzi, M. Albrecht (MLU), C. Argot (MNHN), U. Göhlich, F. Zachos, A. Bibl (NHMW), L. Costeur, F. Dammeyer (NMB), D. Nagel, M. Maslo (IPUW), R. Brocke, L. Kraus (SMF), and G. Rössner (SNSB-BSPG) for allowing us to study material under their care. We thank A. Athanassiou (Ephorate of Palaeoanthropology-Speleology at the Hellenic Ministry for Culture, Greece) and B. U. Bayshashov (Institut of Zoology in Akademgorodok, Kazakhstan) for providing information and literature about chilotheres. We kindly thank N. Kargopoulos (University of Zaragoza) for providing measurements and photos. We would like to thank J. Tissier and an anonymous reviewer for their useful comments that significantly improved our work and the Editor L. Pandolfi for the thorough and reliable handling of the manuscript.
References
- 1.
Heissig K. Family Rhinocerotidae. The Miocene land mammals. Munich: Pfeil. 1999. p. 175–88.
- 2.
Prothero DR. The evolution of North American rhinoceroses. Cambridge, UK: Cambridge University Press. 2005.
- 3.
Heissig K. The Rhinocerotidae. The evolution of perissodactyls. New York, NY: Oxford University Press. 1989. p. 399–417.
- 4.
Prothero DR, Guerin C, Manning E. The history of Rhinocerotoidea. The evolution of Perissodactyls. New York, NY: Oxford University Press. 1989. p. 321–40.
- 5. Antoine P-O, Downing KF, Crochet J-Y, Duranthon F, Flynn LJ, Marivaux L, et al. A revision of Aceratherium blanfordi Lydekker, 1884 (Mammalia: Rhinocerotidae) from the Early Miocene of Pakistan: postcranials as a key. Zoological Journal of the Linnean Society. 2010;160(1):139–94.
- 6.
Antoine P-O, Becker D, Pandolfi L, Geraads D. Evolution and Fossil Record of Old World Rhinocerotidae. Fascinating Life Sciences. Springer Nature Switzerland. 2025. p. 31–48. https://doi.org/10.1007/978-3-031-67169-2_2
- 7. Ringström T. Nashörner der Hipparion-Fauna Nord-Chinas. Palaeontologia Sinica. 1924;1:1–156.
- 8. Kampouridis P, Svorligkou G, Kargopoulos N, Augustin FJ. Reassessment of ‘Chilotherium wegneri’ (Mammalia, Rhinocerotidae) from the late Miocene of Samos (Greece) and the European record of Chilotherium. Historical Biology. 2021;34(3):412–20.
- 9. Deng T. New material of Chilotherium wimani (Perissodactyla, Rhinocerotidae) from the Late Miocene of Fugu, Shaanxi. Journal of Vertebrate Paleontology. 2001;39:129–38.
- 10. Kampouridis P, Svorligkou G, Kargopoulos N, Spassov N, Böhme M. Revision of the Late Miocene hornless rhinocerotids from Samos Island (Greece) with the designation of neotypes and implications for the European chilotheres. Journal of Vertebrate Paleontology. 2023;43(1).
- 11.
Korotkevich OL. The mammals of the Berislav late Sarmatian Hipparion-fauna. 5th ed. Naukova Dumka. 1970.
- 12.
Bayshashov BU. Neogene rhinoceroses of Kazakhstan. Almaty: Academy of Sciences of Kazakhstan. 1993.
- 13. Mecquenem de R. Contribution à l’étude des fossiles de Maragha. Annales de Paléontologie. 1924;13:135–60, 1–36.
- 14. Weber M. Über tertiäre rhinocerotiden von der Insel Samos. Bulletin de la Société Imperériale des Naturalistes de Moscou. 1905;18:344–63.
- 15. Pavlow M. Mammifères tertiaires de la Nouvelle Russie, 1. Partie: Artiodactyla, Perissodactyla (Aceratherium kowalevskii n.s.). Nouveaux Mémoires de la Société Impériale des Naturalistes de Moscou. 1913;17:1–68.
- 16.
Deng T. Limb bones of Chilotherium wimani (Perissodactyla, Rhinocerotidae) from the Late Miocene of the Linxia Basin in Gansu, China. 2002;40:305–16.
- 17. Krokos WI. Aceratherium schlosseri Web. du village de Grebeniki du gouvernement de Kherson. Memoires of the Agricultural Society of Southern Russia. 1917;82:1–96.
- 18. Wessel P, Luis JF, Uieda L, Scharroo R, Wobbe F, Smith WHF, et al. The Generic Mapping Tools Version 6. Geochem Geophys Geosyst. 2019;20(11):5556–64.
- 19. Pohlig H. Ueber eine Hipparionen-Fauna von Maragha in Nordpersien, über fossile Elephantenreste Kaukasiens und Persiens und über die Resultate einer Monographie der fossilen Elephanten Deutschlands und Italiens. Zeitschrift der Deutschen Geologischen Gesellschaft. 1885;37:1022–7.
- 20. Pandolfi L. Persiatherium rodleri, gen. et sp. nov. (Mammalia, Rhinocerotidae) from the upper Miocene of Maragheh (northwestern iran). Journal of Vertebrate Paleontology. 2015;36(1):e1040118.
- 21.
Ataabadi MM, Bernor RL, Kostopoulos DS, Wolf D, Orak Z, Zare G, et al. Recent Advances in Paleobiological Research of the Late Miocene Maragheh Fauna, Northwest Iran. Fossil Mammals of Asia. Columbia University Press. 2013. p. 546–65. https://doi.org/10.7312/columbia/9780231150125.003.0025
- 22. Ataabadi MM, Kaakinen A, Kunimatsu Y, Nakaya H, Orak Z, Paknia M, et al. The late Miocene hominoid-bearing site in the Maragheh Formation, Northwest Iran. Palaeobio Palaeoenv. 2016;96(3):349–71.
- 23. Mecquenem R de. Contribution à l’étude du gisement des vertébrés de Maragha et de ses environs. Annales d’histoire naturelle. 1908;1:27–98.
- 24. Schlosser M. Die fossilen Säugethiere Chinas nebst einer Odontographie der recenten Antilopen. Abhandlungen der Königlichen Bayerischen Akademie der Wissenschaften. 1903;22:1–221.
- 25. Killgus H. Unterpliozäne Säuger aus China. PalZ. 1923;5:251–57.
- 26.
Killgus H. Die unterpliocaenen chinesischen säugetierreste der tafelschen sammlung zu tübingen. Eberhard-Karls University of Tübingen. 1922.
- 27.
Giaourtsakis IX. The Fossil Record of Rhinocerotids (Mammalia: Perissodactyla: Rhinocerotidae) in Greece. Fossil Vertebrates of Greece Vol. 2. Springer International Publishing. 2021. p. 409–500. https://doi.org/10.1007/978-3-030-68442-6_14
- 28. Koufos GD, Kostopoulos DS, Vlachou TD, Konidaris GE. A synopsis of the late Miocene Mammal Fauna of Samos Island, Aegean Sea, Greece. Geobios. 2011;44(2–3):237–51.
- 29. Kostopoulos DS, Sen S, Koufos GD. Magnetostratigraphy and revised chronology of the late Miocene mammal localities of Samos, Greece. International Journal of Earth Sciences. 2003;92(5):779–94.
- 30.
Bayshashov BU. A new rhinoceros species of the genus Chilotherium from Pavlodar. Mesozoic and Cenozoic vertebrate fauna and flora of North-Eastern and southern Kazakhstan. Almaty: Academy of Sciences of Kazakhstan. 1982. p. 72–83.
- 31.
Guérin C. Les rhinocéros (Mammalia, Perissodactyla) du Miocène terminal au Pleistocène supérieur en Europe occidentale: comparaison avec les espèces actuelles. Documents des Laboratoires de Géologie de la Faculté des Sciences de Lyon. 1980. 1–1185.
- 32.
Barone R. Anatomie comparée des mammifères domestiques. Paris: Vigot Frères. 2010.
- 33.
Heissig K. Les Rhinocerotidae (Perissodactyla) de Sansan. Mammifères de Sansan. Paris: Muséum national d’Histoire naturelle. 2012. p. 317–485.
- 34.
Antoine PO. Phylogénie et évolution des Elasmotheriina (Mammalia, Rhinocerotidae). 2002.
- 35. Mallet C, Cornette R, Billet G, Houssaye A. Interspecific variation in the limb long bones among modern rhinoceroses-extent and drivers. PeerJ. 2019;7:e7647. pmid:31579585
- 36. Kiernik E. Über einen Aceratheriumschädel aus der Umgebung von Odessa. Bulletin international de l’Académie des sciences de Cracovie. 1913;1913:808–64.
- 37. Deng T. A primitive species of Chilotherium (Perissodactyla, Rhinocerotidae) from the Late Miocene of Linxia Basin (Gansu, China). Cainozoic Research. 2006;5:93–102.
- 38. Heissig K. Rhinocerotidae aus dem Jungtertiär Anatoliens. Geologisches Jahrbuch (B). 1975;15:145–51.
- 39. Svorligkou G, Kampouridis P, Kargopoulos N, Pandolfi L. Craniodental anatomy of the hornless rhinocerotid Chilotherium schlosseri (Mammalia, Perissodactyla) from the Late Miocene of Samos Island, Greece. J Mammal Evol. 2025;32(4).
- 40. Andree J. Rhinocerotiden aus dem Unterpliocän von Samos. Paläontologische Zeitschrift. 1921;3:189–212.
- 41. Geraads D, Koufos G. Upper Miocene Rhinocerotidae (Mammalia) from Pentalophos-1, Macedonia, Greece. Palaeontographica Abteilung A. 1990;210:151–68.
- 42. Athanassiou A, Roussiakis SJ, Giaourtsakis IX, Theodorou GE, Iliopoulos G. A new hornless rhinoceros of the genus Acerorhinus (Perissodactyla: Rhinocerotidae) from the Upper Miocene of Kerassiá (Euboea, Greece), with a revision of related forms. pala. 2014;303(1–3):23–59.
- 43.
Lubicz-Niezabitowski E. Teleoceras ponticus n. sp. Vorläufige Notiz. Nowy Targ. 1912. 2.
- 44. Lubicz-Niezabitowski E. Über das Schädelfragment eines Rhinocerotiden (Teleoceras ponticus Niez.) von Odessa. Bulletin International de l’Académie des Sciences de Cracovie, Series B. 1913;:223–35.
- 45. Korotkevich OL. A new Chilotherium species from the Sarmatian deposits of the Ukraine. Dopovidi Akad Nauk Ukrainskoi RSR. 1958;12:1372–6.
- 46. Antoine P-O, Sen S. Rhinocerotidae and Chalicotheriidae (Perissodactyla, Tapiromorpha). Geodiversitas. 2016;38(2):245–59.
- 47. Țibuleac P, Tissier J, Petculescu A, Becker D. Chilotherium schlosseri (Weber, 1905) (Rhinocerotidae, Mammalia) from the late Miocene of the foreland of the Eastern Carpathians in Romania. Comptes Rendus Palevol. 2023;(36).
- 48. Geraads D, Spassov N. Rhinocerotidae (Mammalia) from the Late Miocene of Bulgaria. pala. 2009;287(4–6):99–122.
- 49. Sun DH, Li Y, Deng T. A new species of Chilotherium (Perissodactyla, Rhinocerotidae) from the Late Miocene of Qingyang, Gansu, China. Vertebrata PalAsiatica. 2018;56:216–28.
- 50. Mecquenem de R. Le lac d’Ourmiah. geo. 1908;17(92):128–44.
- 51. Pohlig H. On the Pliocene of Maragha, Persia, and its resemblance to that of Pikermi in Greece. Quarterly Journal of the Geological Society of London. 1886;42:177–9.
- 52. Bernor RL. Mammalian biostratigraphy, geochronology, and zoogeographic relationships of the Late Miocene Maragheh fauna, Iran. Journal of Vertebrate Paleontology. 1986;6(1):76–95.
- 53. Campbell BG, Amini MH, Bernor RL, Dickinson W, Drake R, Morris R, et al. Maragheh: a classical late Miocene vertebrate locality in norhtwestern Iran. Nature. 1980;287(5785):837–41.
- 54. Osborn HF. Phylogeny of the rhinoceroses of Europe. Bulletin of the American Museum of Natural History. 1900;12:229–67.
- 55. Kampouridis P, Mirzaie Ataabadi M, Hartung J, Augustin FJ. The easternmost occurrence of the Late Miocene schizotheriine chalicothere Ancylotherium pentelicum at the classical locality of Maragheh (Iran). J Mammal Evol. 2024;31(3).
- 56. Salminen J, Paknia M, Kaakinen A, Ataabadi MM, Zare G, Orak Z, et al. Preliminary magnetostratigraphic results from the late Miocene Maragheh Formation, NW Iran. Palaeobio Palaeoenv. 2016;96(3):433–43.
- 57. Lu X-K, Deng T, Pandolfi L. Reconstructing the phylogeny of the hornless rhinoceros Aceratheriinae. Front Ecol Evol. 2023;11.
- 58. Kargopoulos N, Kampouridis P, Hartung J, Böhme M. Hyaenid remains from the Late Miocene of Kutschwan (Shanxi Province, China). PalZ. 2023;97(3):653–66.
- 59. Kampouridis P, Hartung J, Ferreira GS, Böhme M. Reappraisal of the late Miocene elasmotheriine Parelasmotherium schansiense from Kutschwan (Shanxi Province, China) and its phylogenetic relationships. Journal of Vertebrate Paleontology. 2021;41(6).
- 60. Koufos GD. The late Miocene mammal faunas of the Mytilinii basin, Samos Island, Greece: new collection: 1. history of the samos fossil mammals. Beiträge zur Paläontologie. 2009;31:1–12.
- 61. Kostopoulos DS, Koufos GD, Sylvestrou IA, Syrides GE, Tsombachidou E. The late Miocene mammal faunas of the Mytilinii basin, Samos Island, Greece: New collection: 2. Lithostratigraphy and fossiliferous sites. Beiträge zur Paläontologie. 2009;31:13–26.
- 62. Koufos GD, Kostopoulos DS, Vlachou TD. The late Miocene mammal faunas of the Mytilinii basin, Samos Island, Greece: New collection 16. Beiträge Paläontol. 2009;31:397–408.
- 63. Toula F. Das gebiss und reste der nsenbeine von Rhinoceros (Ceratorhinus Osborn) hundsheimensis. Abhandlungen der k.k Geologischen Reichsanstalt. 1906;20:1–38.
- 64. Weber M. Ueber tertiäre Rhinocerotiden von der Insel Samos. Bulletin de la Société Impériale des Naturalistes de Moscou. 1904;17:477–501.
- 65. Drevermann F. Aus der Zeit des dreizehigen Pferdes. Natur und Museum. 1930;60:2–13.
- 66. Leonardi P. Resti fossili inediti di rhinoceronti conservati nelle collezioni dell’instituto geologico dell’università di Padova. Memorie dell’Instituto Geologico. 1948;15:2–30.
- 67. Gürich G. Fossile Säugetierreste aus Samos. Verhandlungen des Naturwissenschaftlichen Vereins in Hamburg. 1911;3:79.
- 68.
Linnaeus C. Systema naturae. Stockholm: Laurentii Salvii. 1758.
- 69. Owen . Description of Teeth and portions of Jaws of two extinct Anthracotheroid Quadrupeds (Hyopotamus vectianus and Hyopbovinus) discovered by the Marchioness of Hastings in the Eocene Deposits on the N. W. coast of the Isle of Wight: with an attempt to develope Cuvier’s idea of the Classification of Pachyderms by the Number of their Toes. QJGS. 1848;4(1–2):103–41.
- 70. Gray JE. On the natural arrangemenet of vertebrose animals. London Medical Repository. 1821;15:297–310.
- 71. Dollo L. Rhinocéros vivants et fossiles. Revue de Questions Scientifiques. 1885;17:293–9.
- 72. Qiu ZX, Xie JY, Yan DF. A new chilothere skull from Hezheng, Gansu, China, with special reference to the Chinese “Diceratherium”. Scientia Sinica B. 1987;:545–52.
- 73. Kretzoi M. Bemerkungen zum System der Nachmiozänen Nashorn-Gattungen. Földtani Közlöny. 1942;72:4–12.
- 74. Hünermann KA. Die nashornreste (Aceratherium incisivum) aus dem jungtertiär vom höwenegg im hegau. Andrias. 1989;6:5–116.
- 75. Cerdeno E. Rhinocerotidae from the Middle Miocene of the Tung-gur Formation, Inner Mongolia (China). American Museum Novitates. 1996;44.
- 76. Deng T. New cranial material of Shansirhinus (Rhinocerotidae, Perissodactyla) from the Lower Pliocene of the Linxia Basin in Gansu, China. Geobios. 2005;38(3):301–13.
- 77. Rodler A. Verhandlungen der kaiserlich-königlichen geologischen Reichsanstalt. 1885;14:333–7.
- 78. Gaudry A. V.—Palæontology in Germany and Austria. Geol Mag. 1885;2(12):556–9.
- 79. Bernor RL, Semprebon GM, Damuth J. Maragheh Ungulate Mesowear: Interpreting Paleodiet and Paleoecology from a Diverse Fauna with Restricted Sample Sizes. Annales Zoologici Fennici. 2014;51(1–2):201–8.
- 80. Kaup J. Über Rhinoceros incisivus Cuv., und eine neue Art, Rhinoceros schleiermacheri. Isis von Oken. 1832;:898–904.
- 81.
Fortelius M, Heissig K, Saraç G, Sen S. Rhinocerotidae (Perissodactyla). In: Fortelius M, Kappelman J, Sen S, Bernor RL, editors. Geology and paleontology of the Miocene Sinap Formation, Turkey. New York, NY: Columbia University Press. 2003. p. 282–307.
- 82.
Nothdurft W, Smith JB. The lost dinosaurs of Egypt. New York: Random House. 2002.
- 83. Mallet C, Houssaye A. Deciphering the influence of evolutionary legacy and functional constraints on the patella: an example in modern rhinoceroses amongst perissodactyls. PeerJ. 2024;12:e18067. pmid:39469593
- 84. Baishashov BU, Nigmatova SA, Seidali AL, Madiyarova IT, G. Lucas S. Paleogeographical reconstruction of the late neogene in the north-eastern part of Central Kazakhstan (The «Goose Passage» Site). EJSU. 2024;146(5):32–44.
- 85. Cerdeño E. Étude sur Diaceratherium aurelianense et Brachypotherium brachypus (Rhinocerotidae, Mammalia) du Miocène moyen de France. Bulletin du Muséum national d’histoire naturelle, Paris. 1993;15:25–77.
- 86. Prothero DR, Manning EM. Miocene rhinoceroses from the Texas Gulf Coastal Plain. J Paleontol. 1987;61(2):388–423.
- 87. Short RA, Wallace SC, Emmert LG. A new species of Teleoceras (Mammalia, Rhinocerotidae) from the late Hemphillian of Tennessee. FLMNH. 2019;56(5):183–260.
- 88. Ringström T. Über quartäre und jungtertiäre rhinocerotiden aus China und der Mongolei. Palaeontologia Sinica. 1927;4:1–21.
- 89. Lu X, Chen S, He W. New skulls of Shansirhinus ringstroemi from the upper Neogene of the Linxia Basin, and their paleoenvironmental context. Quaternary Research. 2015;35:539–49.
- 90. Deng T, Lu X, Sun D, Li S. Rhinocerotoid fossils of the Linxia Basin in northwestern China as late Cenozoic biostratigraphic markers. Palaeogeography, Palaeoclimatology, Palaeoecology. 2023;614:111427.
- 91. Samuels ME, Regnault S, Hutchinson JR. Evolution of the patellar sesamoid bone in mammals. PeerJ. 2017;5:e3103. pmid:28344905
- 92. Pina M, Almécija S, Alba DM, O’Neill MC, Moyà-Solà S. The Middle Miocene ape Pierolapithecus catalaunicus exhibits extant great ape-like morphometric affinities on its patella: inferences on knee function and evolution. PLoS One. 2014;9(3):e91944. pmid:24637777
- 93. Ward CV, Ruff CB, Walker A, Teaford MF, Rose MD, Nengo IO. Functional morphology of Proconsul patellas from Rusinga Island, Kenya, with implications for other Miocene-Pliocene catarrhines. Journal of Human Evolution. 1995;29(1):1–19.
- 94. Antoine PO, Saraç G. Rhinocerotidae (Mammalia, Perissodactyla) from the late Miocene of Akkașdaği, Turkey. Geodiversitas. 2005;27:601–32.
- 95.
Svorligkou G. Study and taxonomical identification of carpal and metacarpal bones of rhinocerotids Dihoplus pikermiensis and “Diceros” neumayri (Mammalia, Perissodactyla) from the Turolian of Pikermi. National and Kapodistrian University of Athens. 2018.
- 96. Heissig K. The fossil rhinoceroses of Rudabánya. Palaeontographia Italica. 2004;90:217–58.