Los suoideos de talla pequeña del Mioceno de Europa y Asia - Small suoids from the Miocene of Europe and Asia

The history of study of small suoids from the Miocene of Eurasia is complex for several reasons: scarcity of fossil material, a high degree of dental convergence and parallelism between closely and distantly related lineages, and frequent misattribution of fossils, resulting in the gradual development of a confusing taxonomy. Changes in taxonomy above the genus level, have added to the complexity; European lineages classified in Suidae in 1924 are now arranged into three separate families; Suidae, Palaeochoeridae and Sanitheriidae. Recent studies have considerably clarified the situation, but there remain several problematic issues to resolve, especially among the Palaeochoeridae. The fossil register of some taxa is limited, so it is necessary to put on record newly recognised specimens in order to fill out our knowledge concerning them. This paper includes previously undescribed material of Palaeochoeridae and small Suidae, as well as reinterpretation of some fossils published in “obscure” scientific journals. The latter include some taxa that have priority over more recently proposed names. A systematic revision of these forms is carried out, and the paper ends with a proposal for a revised taxonomy of the Palaeochoeridae, a family that has recently taken on importance in the debate about the origins of Hippopotamidae.

Miocene palaeochoerids of Europe, and the studies of Van der Made (1997a, 1997b, 1998, 2010) on the European Early and Middle Miocene palaeochoerids.There remain several important collections to study, including a fine sample of Propalaeochoerus elaverensis from the type locality Coderet, France.
It is noted that Palaeochoeridae have been linked to the origin of Hippopotamidae (Pickford, 2008) so a more complete understanding of the family is essential, particularly because some researchers consider that hippopotamids emerged from Anthracotheriidae (Boisserie et al., 2005).
Many suoid lineages possess bunodont dentitions, but there are several groups that developed highly lophodont cheek teeth independently from each other, resulting in a plethora of convergences, historically causing confusing between perissodactyls, primates and suoids.In most instances, the systematic and phylogenetic relationships of these lophodont lineages were initially misunderstood.In the case of the suid Listriodon, isolated molars were thought by early workers to belong to tapir-like perissodactyls (Tapirotherium blainvilleanum Lartet (1851), Tapirus pentapotamiae Falconer (1868)) until more complete fossils were found, which showed their suid affinities.Dental and mandibular remains of small palaeochoerids have on occasion been named after their resemblances to lemurs (Sus lemuroides Blainville, 1847, based on a mandibular symphysis of Choeromorus sansaniense, transferred to Pachysimia by Lydekker, 1884) or cercopithecids, as shown by the attribution of their fossils to Colobus grandaevus by Fraas (1870) and Macaca youngi by Gu Yumin (1980).When specimens of Schizochoerus vallesensis Crusafont & Lavocat (1954) were first described, they were interpreted to represent listriodonts (listed as Listriodon n?. gen.by Crusafont & Villalta, 1948), but Golpe-Posse (1972a, 1972b) realised that the species was so divergent from Listriodon that she erected a subfamily Schizochoerini, for it (see discussion in Van der Made, 2010, concerning the validity of this suprageneric name).The genus was transferred to Palaeochoeridae (at the time known as Old World Tayassuidae) by Pickford (1978).When specimens of Namachoerus Pickford (1995) were first described, they were interpreted to represent a small sanitheriid (Xenochoerus moruoroti Wilkinson, 1976) before being transferred to the listriodont genus Lopholistriodon (Pickford & Wilkinson, 1975) itself initially confused with a tapir (Aguirre & Leakey, 1974).When more complete maxillae and mandibles became available for study, they showed that, unlike Lopholistriodon, the species did not share the morphology of the skull with Listriodon; hence the decision to erect a separate genus, Namachoerus, and a separate subfamily Namachoerinae (Pickford, 1995) within the Suidae.When dental remains of Anthraconema dangari Prasad (1970) were first described, they were interpreted to belong to anthracotheres, until Pickford (1978) realised that this species showed strong resemblances to Schizochoerus (now Schizoporcus).Another lineage from the Indian subcontinent was initially thought to represent the palaeochoerid genus Taucanamo, but the material was later put into Schizochoerus (S. gandakasense) by Pickford (1978) because it was more lophodont than the former genus.In the same paper, Pickford (1978) transferred Prasad's species to Schizochoerus as the combination Schizochoerus dangari (Prasad, 1970).Han Defen (1983, 1985) first attributed small lophodont suoid teeth from Lufeng, Yunnan, to Lophochoerus, a small tetraconodont suid, but it was later realised that they represent a lophodont palaeochoerid for which Van der Made & Han (1994) erected the genus Yunnanochoerus.The latter authors did not discuss the status of Prasad's material from India (it is poorly illustrated) but it is now clear that the holotype and additional previously undescribed fossils from the type locality (Hari Talyangar, India) belong to the same genus and species.Thus Yunnanochoerus Van der Made & Han (1994) is a junior synonym of Anthraconema Prasad (1970).However, Anthraconema Prasad   Koenigswald, 1963), so is approximate.In a seminal paper, Van der Made (1996) pointed out that small suids and palaeochoerids from the Miocene of Europe had for many years been confused with each other, partly due to their similar dimensions, but also due to superficial similarities in the occlusal morphology of their cheek teeth.He pointed out the features which distinguished the two groups from each other (notably cheek tooth radicular and crown morphology and some post-cranial characters) and provided a detailed list of localities at which the various species occurred, results which were recently extended by the same author (Van der Made, 2010).Nevertheless, there remain several residual problems concerning the nomenclature of the palaeochoerids in particular, which this contribution attempts to resolve.
The hypsorhizic palaeochoerids Lorancahyus daamsi and Lorancahyus hypsorhizus are extremely rare, each having previously been recorded from a single Spanish locality, but material from a second locality in France (MN 2) is described herein.The fossils indicate a possible descendant relationship from the basal Early Miocene lineage Propalaeochoerus pusillus (MN 1).
The Middle and Late Miocene lophodont palaeochoerid suoids of Europe and Asia have long been in need of taxonomic revision.The first described species from Asia was Microbunodon sminthos Forster-Cooper (1913, 1924) from Bugti, Pakistan, initially interpreted as an anthracothere.It was followed by Pecarichoerus orientalis Colbert (1933) from Chinji, Pakistan, interpreted as a peccary.Then came Anthraconema dangari Prasad (1970) from Hari Talyangar, India, thought by its discoverer to be an anthracothere (Anthraconema Prasad is a homonym of Anthraconema Zur Strassen, 1904).A genus name, Yunnanochoerus Van der Made & Han (1994) is available for this taxon.The fourth species described from Gandakas in the Potwar Plateau, Pakistan, was initially attributed to Taucanamo gandakasense Pickford (1977) and then to Schizochoerus.The fifth described species was "Macaca" youngi Gu Yumin (1980) from Xiaodian, Hubei, China, initially thought to be a monkey, which was followed by the sixth, Yunnanochoerus lufengensis (Han, 1983) from Lufeng, Yunnan, China.More complete material of Yunnanochoerus dangari from the type locality (Hari Talyangar, India) described herein reveals that the Late Miocene Asian species are comfortably accommodated within a single genus, for which the genus name with priority after taking into account the homonymy of Anthraconema Prasad (1970)  Miocene Pecarichoerus orientalis is so similar in morphology and dimensions to Schizoporcus muenzenbergensis (Van der Made, 1998;Orliac et al., 2010b) that I consider the latter to be a synonym of the former.Yunannochoerus differs from Pecarichoerus (including the Asian material of what used to be called Schizochoerus) in possessing molar lophids that are further apart, and it has more enlarged, elongated premolars and a p/4 with one main cusp (Van der Made & Han, 1994).Yunnanochoerus is a lophodont Palaeochoeridae, probably descended from a slightly more bunodont form such as Choeromorus grandaevus, whereas the Schizoporcus lineage is more likely to have been derived from a separate bunodont lineage such as Pecarichoerus primus.
A tooth from Chiang Muan, Thailand previously attributed to Pecarichoerus sminthos is here attributed to the small suid species Lophochoerus nagrii, and a mandible from Przeworno, Poland initially attributed to Taucanamo is identified as the small suid Albanohyus.In order to augment our understanding of European suids, other material from various sites is included and fossil material previously published in "obscure" scientific journals is re-illustrated to make them more accessible to Western scholars.
A revised taxonomy of Neogene Palaeochoeridae ends the paper.

Materials and methods
Dental nomenclature is based on Pickford (1978Pickford ( , 1988)).Lower molars are abbreviated to m/1, m/2 and m/3 for the first, second and third teeth respectively, upper molars are M1/, M2/ and M3/, likewise for premolars where p/1 or P1/ are used.
In the synonymy lists I have been guided by Matthews (1973).The year of publication to the left of the taxon can be in italic (2010) which implies a mention of the species but without description or illustration, or in roman (2010) which implies that the work contributes to our knowledge of the species.The year can be preceded by an asterisk (*1992) meaning that publication of a name in this work can be regarded as valid in terms of the ICZN, earlier mentions of the name are to be regarded as nomina nuda, or by lower case letter 'v' (vidimus: v1992) meaning that I have seen the specimens mentioned in the paper, and a lower case 'p' (partim: p1870) meaning that the reference applies only in part to the species (on occasion referred to as a 'gryphon' taxon (Van der Made, 2010)).
The geochronology of the species mentioned herein is based predominantly on the European Land Mammal Zonation (Mein, 1990;Rössner & Heissig, 1999) (MN Zones) some of which have been correlated to Asia (India, China).Most authors agree about the position of most of the localities within this zonation, but there are a few localities such as Castelnau d'Arbieu and Sandelzhausen which are positioned in different zones by different authors.For these cases I have taken the estimates published in Rössner & Heissig (1999).
Description: MNHN SG 3603 is a right mandible containing worn p/3-p/4 and m/2, with the alveoli of the canine, p/1-p/2, and m/1 and the mesial part of the alveolus of m/3.There are several unusual features about this specimen.There are short diastemata between the canine, p/1, p/2 and p/3.The p/1 and p/2 are small teeth so that the distance between the rear of the canine alveolus and the front of the p/3 is only 14 mm, which is extremely short in a suoid context.The jaw is also quite slender and gracile, even at the level of the m/2.
There is a prominent mental foramen located about 1/3 of the depth of the jaw beneath the alveolar margin at p/1 level.There are several other nutritive foramina, one behind the mental foramen, and four others lower down the jaw beneath p/3, p/4, m/1 and m/2.In ventral view, the base of the mandible is sharpedged from the genial fossa back to the level of the p/4, after which it is rounded.The depression that occurs above this sharp edge extends from the genial fossa back toward the lingual fossa beneath the molars.The broken section of the mandible at the rear of m/2 shows the presence of a large mandibular canal occupying almost half the height of the mandible.
The mesial and distal roots of m/2 are solidly fused well beneath cervix, although the alveolus of m/1 indicates that its mesial and distal roots were separated from about cervical level.In all three molars the buccal and lingual roots were fused throughout their height, as in Palaeochoeridae in general.
Estudios Geológicos, 67(2), 541-578, julio-diciembre 2011.ISSN: 0367-0449.doi:10.3989/egeol.40634.206 Small suoids from the Miocene of Europe and Asia The canine alveolus is deep, extending as far back as the genial fossa and it is slightly tilted laterally, and this causes a swelling in the lateral profile of the mandible mesial to the p/1.The p/1 was simple rooted, but the p/2 possessed two roots that were fused throughout their height.
The p/3 is worn to the stage where little morphology remains, but it is possible to discern that it has a single main cusp behind which is an expanded distal part showing a low disto-buccal cusplet.The p/4 is deeply worn, but it had a lingual cusp (protoconid) close to the buccal main cusp, as shown by a shallow groove on the lingual side of the crown.There was a prominent disto-buccal cuspid, and on its lingual side a shallow distal shelf.The m/2 is deeply worn to the stage where the crown has almost disappeared, leaving just a ring of enamel surrounding a large dentine lake.The dentine lake shows secondary dentine infilling what used to be pulp cavities, a phenomenon that characterises many palaeochoerids such as Propalaeochoerus, Lorancahyus, and Palaeochoerus, which endure heavy wear on the cheek teeth.
Discussion: The m/2 dimensions in the mandible MNHN SG 3603 are similar to those in a mandible (SMNS 20272) of Propalaeochoerus pusillus from Tomerdingen, Germany, described by Hellmund (1992, pl. 4, fig. 1) but the anterior part of the mandible containing the premolars and canine is considerably shorter than in P. pusillus.The planum alveolare is steeper in the St Gérand-le-Puy specimen than in the Tomerdin-gen one, which indicates a much shorter muzzle in SG 3603 than in the German specimen.Lower molars of Propalaeochoerus pusillus from Tomerdingen, show precocious development of root fusion in the m/3 (Hellmund, 1992, pl.10, fig.3), especially the roots beneath the entoconid, hypoconid and hypoconulid, but there remains a slit between these roots and the one beneath the protoconid -entoconid cusp pair.These specimens suggest a possible ancestor -descendant relationship between Propalaeochoerus pusillus (MN 1) (Hellmund, 1992) and Lorancahyus daamsi (MN 2a) (Pickford & Morales, 1998).Van der Made (2010, p. 114) wrote that the type material of Lorancahyus daamsi may well represent the same species as Palaeochoerus typus.The St Gérand-le-Puy specimen indicates that this possibility can be rejected.The differences in mandibular morphology are too great for them to be classified in a single species, even if the molar morphology is somewhat comparable.
The original description of Lorancahyus daamsi was based on material from a unique site at Moheda in Spain correlated to MN 2a (Pickford & Morales, 1998).The mandible from St Gérand-le-Puy housed at the Muséum National d'Histoire Naturelle is comparable in dimensions and morphology to the holotype from Moheda.The discovery of this species in France enlarges its known distribution and confirms the biochronological correlation between the two sites.Pickford & Morales, 1998 Material: FSL, right mandible fragment containing m/2-m/3, distal left humerus, left talus.

Species Lorancahyus hypsorhizus
Description: The right mandible of Lorancahyus hypsorhizus (Fig. 3) housed at the University of Lyon was found at St Gérand-le-Puy, France.The jaw contains moderately worn m/2 and m/3.The roots of m/2 and m/3 are solidly fused from side to side and from mesial to distal, the crowns are simplified with the valleys close together, and in the m/3 there is strong backward flare of the hypoconulid and third root.The mesial and distal fovea would have been small and shallow, having been eradicated even in slightly worn teeth.The lingual and buccal notches are narrow but deep.The third root of the m/3 is solidly fused to the first and second ones.Measurements are provided in Table 2.
From the same site there are a distal left humerus and a left talus (Fig. 3C).The humeral articulation is 14.6 mm in mediolateral diameter, and the total breadth of the distal end is 15.4 mm.The capitulum is swollen, almost ball-like and the lateral crest is low but sharp.The supracondylar fossae are deep on both the posterior and anterior aspects with perforation between them.This morphology indicates the probability of extreme flexion and extension movements in the elbow joint of Lorancahyus, while the ball-shaped zona conoidea and the sharpness of the lateral and medial crests indicate a strongly stabilised joint.
The talus has an extremely broad navicular facet (Fig. 3C2), as is usual in palaeochoerids and tayassuids, unlike the narrower facet that occurs in suids, anthracotheres and hippopotamids.In posterior aspect it has the characteristic bent shape that is found in Suiformes.The external length is 23.6 mm, internal length -22.6 mm, proximal breadth -11.2 mm, distal breadth -13.0 mm.

M. Pickford
Estudios Geológicos, 67(2), 541-578, julio-diciembre 2011.ISSN: 0367-0449.doi:10.3989/egeol.40634.2062010) is not impossible, however, as isolated, worn upper molars from Loranca measure 10 x 10.5 mm and 9.7 x 8.6 mm respectively (Pickford & Morales, in Pickford, 1993) compatible in dimensions with the Cetina M3/.The main difference between the upper molar from Cetina de Aragon and the intermediate upper molars from Loranca is that in the latter specimens the cusps are more isolated from each other and less voluminous, unlike the crowded puffy cusps that occur in the specimen from Cetina (Fig. 4).From this it is concluded that the Cetina tooth does not belong to Lorancahyus but to Propalaeochoerus cf pusillus.
The specimens from Navarette consist of a maxilla with canine in situ and alveoli of P1/ and P2/ separated by a short diastema (7mm).There is a very short gap between the canine and the P1/ alveolus (2 mm).The canine is bucco-lingually compressed, is implanted vertically in the maxilla and has a vertical mesial wear facet.The maxilla is swollen above and behind the canine, forming a canine boss, with a niche mesially  for the lower canine.The lateral surface of the maxilla has a prominent depression above the P2/.The inter-maxillary suture is preserved and measurements of the specimen indicate that the inter-canine palatal breadth was ca 18 mm.There is a prominent palatal groove running from mesial to distal alongside the tooth row, as in Schizoporcus arambourgi (Fig. 11) and the surface of the bone on the lingual side of the groove shows low transverse ridges or swellings which make for an undulating rather than a flat palatal surface.
The canine in the maxilla measures 7.5 x 5.3 mm (md x bl) and has 9.7 mm of exposed crown height.The isolated canine measures 8.8 x 5.6 mm (md x bl) but the total height cannot be ascertained due to damage to the apex.It shows a clear distal crest and in both canines the enamel is thin.
The Navarette suoid specimens are interesting, but without knowledge of the premolars and molars it is not possible to suggest a convincing identification other than that they represent palaeochoerids.
Description: Pecarichoerus was listed in the fauna from the Shanwang Formation (Hsiehchiao (Xiacaowan)) by Li et al., (1984).A lightly worn lower molar from Sihong (Fig. 6b) housed in the IVPP, Beijing is provisionally attributed to this species.The tooth consists of four, tall pointed cusps arranged in two pairs, accompanied by a well formed centrally positioned hypoconulid.The roots beneath each lophid are fused medio-laterally and in lateral view are close together.There is a tiny basal pillar in the median transverse valley between the protoconid and hypoconid.The metaconid shows a weak zygodont cristid.The mesial cingulum is not attached to the anterior accessory cusplet.The precristids of the protoconid and metaconid fuse with the anterior accessory cusplet, enclos-ing a mesial fovea.The postcristids from the protoconid and metaconid meet in the midline of the tooth to close off the rear of the mesial fovea.The sagittal valley between the protoconid and metaconid is deep.The metaconid has a centrally directed cristid that swells basally to form a small cuspid that partly blocks the mesial fovea.The precristid of the hypoconid joins a low median accessory cusplet which partly blocks the median transverse valley.The entoconid is isolated at its apex and, like the hypoconid, shows weak Fürchen.The sagittal valley between the hypoconid and entoconid is deep.The hypoconulid overhangs the cervix distally.The cusps are tall relative to the occlusal dimensions of the tooth (8.4 mm tall at the metaconid which is greater than the breadth of the tooth, 7.1 mm at metaconid).
The upper molar from Xiaodian, China, holotype of Macaca youngi, is a small tooth with moderately worn lingual cusps and unworn buccal ones (Fig. 6a).The mesial cingulum is broad bucco-lingually and narrow mesio-distally.The buccal cusps are isolated from the lingual ones until near their bases.The protocone has a prominent obliquely oriented preprotocrista and the hypocone has a strong obliquely directed precrista that reaches the protocone and an oblique postcrista that extends to the rear of the metacone.The buccal cusps are distinctly smaller than the buccal ones, as in the type species from Chinji.
Discussion: The Xiaodian tooth is not that of a cercopithecid as thought by its describer (Gu Yumin, 1980) but is from a palaeochoerid.As an M1/ it has the right dimensions to belong to Pecarichoerus orientalis of which the M1/ is unknown (the supposed M1/ described by Colbert, 1933, is in fact the M2/).As an M2/ the Xiaodian tooth has comparable dimensions to Choeromorus grandaevus which is similar to Pecarichoerus sminthos (Forster-Cooper, 1913) from Bugti, Pakistan (Pickford, 1987).Given that its meristic position is unknown it is here referred with some doubt to Pecarichoerus sminthos.Previously, Pickford (1987) (Qiu & Qiu, 1995).
Description: GSI K16/341 is an isolated left M3/ from Chinji (Fig. 7A).It is moderately deeply worn, but retains some critical morphology of the rear loph showing that the metacone is isolated from the neighbouring cusps at its apex and most of its height and is reduced in size relative to the hypocone.The hypocone sends a strong postcrista disto-buccally which ends behind the metacone, from which it is separated by a deep cleft.The main part of the hypocone is slightly behind the level of the metacone.The protocone is deeply worn and at its base it is distinctly larger than the paracone and is slightly behind it, giving the tooth a slightly trapezoidal outline.There is a bucco-lingually extensive mesial cingulum.In the base of the buccal notch, there is a low basal pillar, and there is another low cusplet in the notch between the metacone and the distal cingulum, in the disto-buccal corner of the crown.
SMF 6396, a left M3/ from Chinji stored in the Senckenberg Museum, Frankfurt, is in medium wear, and shows a well formed basal cusplet in the lingual end of the median transverse valley (Fig. 7B).The roots are fused for about 5 mm beneath thr cervix as in European specimens of the species., 1998).Not only is the morphology close, but the dimensions are almost identical (Fig. 8) indicating the probability that the Chinji species and that from Münzenberg and Sandelzhausen are the same.The Chinji species shows a small prehypocone crista in the M2/, a structure that is also present in specimens from Sandelzhausen, Ger-many (BSPG 1959 II 8216, skull) (this feature is obliterated in heavily worn specimens).The M3/ in the Sandelzhausen skull is very similar to that of the Chinji holotype (Fig. 9), and the basic structure is close to that of the new specimen described above (GSI K 16/341) (Fig. 10) allowing for the deeper wear in the latter specimen.The two buccal cusps (paracone and metacone) are noticeably smaller than the two lingual cusps (protocone and hypocone) and the mesial and distal cingula are   strongly developed, the distal one curving round towards the rear of the metacone.The protocone and hypocone have robust crista tending mesio-centrally, so that with wear the cusps take on an ovoid contour, pointed obliquely mesially.These two cusps are not opposite the paracone and metacone, but lie slightly distally.The P4/ from Chinji described by Colbert (1933) is worn and damaged buccally, but the preserved parts show the same basic morphology as the material from Sandelzhausen (Van der Made, 2010), especially the lingual valley (or sinus) entering the tooth at a markedly oblique angle pointing towards the rear.This cleft separates the distally positioned protocone from the buccal main cusp and there is a prominent distal cingulum (Fig. 9).This morphology differs radically from that observed in Choeromorus grandaevus (Fig. 24, 28).Colbert (1933Colbert ( , 1935) ) interpreted the worn upper molar from Chinji as an M1/, possibly on account of the much greater degree of wear than in the M3/.However, its dimensions suggest that it is an M2/, and comparison with material from Sandelzhausen, Germany, reveals that the wear gradient in this lineage is indeed steep.For example, the skull BSPG 1959 II 8216 has a similar wear gradient (Van der Made, 2010, fig. 1) as the Chinji holotype of Pecarichoerus orientalis.The similarity between these specimens is striking, even down to details of cusp morphology, cingulum development and wear pattern.There can be little doubt that they are closely related, perhaps even conspecific, the stance adopted here.
Material from Turkey: When Ozansoy (1965) described a mandible (the holotype of Schizochoerus arambourgi) he kept some specimens in open nomenclature -a palate with P3/-M3/ on both sides, a second palatal fragment containing what he considered to be three permanent premolars and a deciduous premolar, an isolated canine and an upper incisor.He attributed the latter specimens to Schizochoerus cf arambourgi.His motive for creating a new species for the Yassiören mandible was related to the appreciable difference in dimensions between it and the holotype mandible of Schizochoerus vallesensis.The teeth in the palate and maxilla, in contrast are more similar in dimensions to the Spanish material, which is admittedly, rather poorly known (a single M1/).
Re-examination of the specimens reveals that the maxilla fragment does indeed carry the three anterior permanent premolars (the P2/ is represented by its roots only), but that in addition, there is an accessory premolar between the P1/ and the canine alveolus (Fig. 11).Ozansoy (1965) argued that the tooth in front of the P1/ was a retained deciduous premolar.However, apart from its smaller dimensions, in its morphology, enamel thickness and root morphology, it closely resembles the permanent tooth behind it, and I take it to represent an accessory permanent tooth.
The presence of accessory premolars is extremely rare among artiodactyls, but they are known to be characteristic of the anthracothere genus Libycosaurus from Northern and Equatorial Africa (Pickford, 1991(Pickford, , 2005(Pickford, , 2006)).Until now their presence in other suiformes has not been documented.Van der Made (2003, Table 13.3)provided measurements of the two anterior teeth in the maxilla, but he thought they were P1/ and P2/ (his measurements labelled P1/ pertain to the P acc / and those labelled P2/ pertain to the P1/).Note also that the measurements of the lower teeth that he labelled Schizochoerus   sinapensis (Van der Made, 2003, Table 13.1) are from the holotype mandible of Schizoporcus arambourgi, which the author considers to be a synonym of Schizoporcus vallesensis (Van der Made, 2010).If the mandible really does represent the same species as S. sinapensis, then the latter name is a junior synonym of S. arambourgi.However, the species name in Van der Made's (2003a) Table 13.1 is likely a lapsusvallesensis was probably meant.Nevertheless, the holotype mandible of Schizoporcus arambourgi is appreciably smaller than specimens of S. vallesensis from Spain, Bulgaria and Moldova (Ozansoy, 1965;Crusafont & Lavocat, 1954;Nikolov & Thenius, 1976;Lungu, 1971) and it is compatible in dimensions with the holotype snout of Schizoporcus sinapensis.In this paper it is considered to be a valid species (see below).
Reassessment of fossils from Turkey: Pickford (1978) described a snout with almost complete dentition from Sinap, Turkey, and attributed it to Schizochoerus cf gandakasensis.It was refigured by Pickford & Ertürk (1979) and was briefly mentioned by Van der Made & Han (1994).Van   Re-analysis of the dimensions of all the fossils currently classified in Schizoporcus reveals that the holotype mandible of S. arambourgi is indeed substantially smaller than material attributed to Schizoporcus vallesensis, not only from the type locality, but also from Bulgaria and Moldavia (Crusafont & Lavocat, 1954;Nikolov & Thenius, 1067;Lungu, 1971).
The dimensions of the upper and lower molars of Schizoporcus show the usual relationships found widely in bunodont suiforms, the lower molars are usually somewhat longer but narrower than their upper counterparts.Plots of the upper and lower molars reveal that the lower teeth of Schizoporcus arambourgi are too small to represent the same species as Schizoporcus vallesensis, but that they are compatible with the upper teeth in the holotype snout of Schizoporcus sinapensis.It is therefore concluded that S. sinapensis (Van der Made, 1997a) is a junior synonym of S. arambourgi (Ozansoy, 1965).Earlier mentions of the species S. arambourgi by Ozansoy (1957aOzansoy ( , 1957b) )

Description:
The teeth in the two mandible fragments from Hari Talyangar, India are lightly worn or unworn but have suffered slight damage (missing enamel flakes).Specimen GSI Htu is a juvenile in which the m/3 is erupting; GSI K15/247+249 is adult with lightly worn m/1 and m/2.The m/1 and m/2 have four tall, pointed cusps arranged into two transverse pairs (Fig. 17, 18).The preprotoconid cristid is prominent and curves medially to close off a deep mesial fovea located between the protoconid and metaconid.The premetaconid cristid is weak (in specimen GSI HTu this cuspid is miss-  ing, but it is complete in GSI K 15/247).The posterior cristids linking the front pair of cusps (postprotocristid and postmetacristid) dip cervically and are aligned at the rear of the cusps, closing off the rear of the mesial fovea.The prehypoconid cristid is strongly developed and forms a continuous crest between the hypoconid and the mesial pair of cuspids meeting the distal wall of the mesial fovea in the midline of the tooth.The posthypoconid cristid is oriented transversely and reaches towards the postentocristid but there is a cleft between them.The centrally positioned hypoconulid is well formed but low and it sends a cristid obliquely towards the hypoconid.The lingual and buccal notches are mesio-distally elongated and are deep.The mesial cingulum is weak.
The m/2s are basically enlarged versions of the m/1, the main difference being that the hypoconulid sends a cristid directly mesially, and not as obliquely as in the m/1.The front two lophids of m/3 are similar to the m/2.
Measurements of lower molars are provided in Table 4.
Discussion: Prasad (1970) created the genus and species Anthraconema dangari in the belief that it was an anthracothere.The illustration provided by Prasad (1970, pl. XV, fig. 3) although of poor quality and taken in an oblique orientation, is that of a lophodont suiform with pointed cusp apices.His description of the holotype (GSI 18078) provides a succinct résumé of the main points of the morphology, including the molars with cusps more pointed than bunodont, with well marked ridges connecting the protoconid-metaconid and protoconid-hypoconid, and a minor ridge connecting hypoconid and entoconid, with an anteriorly extending cingulum.Examination of the specimen in the GSI, Calcutta, confirms its identification as a lophodont palaeochoerid and not an anthracothere.The discovery of two additional mandibular specimens in the cabinets of the GSI, described herein, removes any residual doubt that there might have been about the identification of this species (Pickford, 1988).

Description:
The lower left m/1 (PDYV 31) from Yuanmou, Yunnan, China, has four main cusps and a prominent median accessory cusp at the mesial end of the prehypocristid.The mesial cingulum is bucco-lingually broad but mesio-distally narrow.The preprotocristid curves lingually and ends in the midline of the crown.The premetacristid descends towards the mesial cingulum slightly obliquely.The postprotocristid and postmetacristid form the rear wall of the mesial fovea which lies between the protoconid and metaconid.Descending the rear of the protoconid and metaconid are "zygodont" cristids bordered by shallow grooves.This morphology superficially resembles the "M" morphology of the lower molars of tragulids.The buccal and lingual notches are broad and deep, meaning that apices of the mesial pair of cusps are far from those of the distal pair.The median transverse valley is blocked in its centre by the median accessory cusplet, which is strongly joined to the prehypocristid.The latter cristid descends from the apex of the hypoconid towards the centre of the crown.The hypoconid has a small posthypocristid that descends towards the centreline of the tooth where it joins the hypoconulid.The entoconid is detached from the hypoconid at its apex, but it has a buccally directed cristid that links it to the hypoconid, forming the mesial wall of the distal fovea.The distal fovea is small and is located between the hypoconid and entoconid, its rear wall being formed by the hypoconulid, which is low and centrally positioned.
The right m/2 (PDYV 21) is a larger version of the m/1, and although it is basically an upscaled version of this tooth, it shows some differences in morphology from it.The most Estudios Geológicos, 67( 2  notable differences concern the lingual side of the tooth.The lingual surface of the metaconid is adorned with a tall cingular cusplet that tapers apically, almost reaching the apex of the cusp.Behind this structure, there are two lower cingular cusplets at the lingual end of the transverse valley, and there is a slightly taller cingular cusplet at the front of the entoconid. The two right m/3s (PDYV 20 and PDYV 0252) are basically upscaled versions of the m/1 but with the addition of a well developed talonid behind the distal cusp pair.Unlike the m/2 described above, there is no sign of lingual cingular cusplets in either of the m/3s.The mesial fovea is prominent between the protoconid and metaconid, the "zygodont" cristids are present  but almost worn away, their position being shown by the grooves that parallel them.The distal fovea is poorly developed and open to the rear, being more like a groove than a basin.The talonid is comprised of a large hypoconulid about half the height of the crown, accompanied by a prehypoconulid cristid that descends towards a low lingual cusplet or basal pillar.In PDYV 20 the basal pillar is small, but in PDYV 0252 it is enlarged and almost as tall as the hypoconulid.The hypoconulid sends low but sharp cristids buccally and lingually.
Where preserved the roots of these teeth reveal that they are fused bucco-lingually, as is usual in palaeochoerids.In the m/3s the root beneath the hypoconulid slants distinctly towards the rear, but it is strongly fused to the root beneath the hypoconid and entoconid.
Discussion: The four lower molars from Yuanmou, Yunnan, China, are similar in dimensions and morphology to the holotype of Yunnanochoerus gandakasensis (Pickford, 1977).They are similar in morphology to the Lufeng fossils of Yunnanochoerus lufengensis (Han, 1983) but are appreciably larger than them (Fig. 20, Table 4).
Van der Made & Han (1994) separated the European and Asian palaeochoerids into two genera, Schizochoerus for Europe (now Schizoporcus), and Yunnanochoerus for Asia.The main differences between these genera concern the elongation of the premolars (longer in Yunnanochoerus than in Schizoporcus) and the number of cusps in the p/4 (double cusped in Schizoporcus: single cusp in Yunnanochoerus).The separation between the molar lophids appears to be greater in the Asian than in the European forms, but it is clear that the two genera are closely related.The Asian species also appears to have more developed mesial molar cingula than the European species.
There can be little doubt that Van  .Note also the relatively small dimensions of the specimen from Yassiören when compared with those of Schizoporcus vallesensis, which supports its identification as a distinct species, Schizoporcus arambourgi (Ozansoy, 1965).
being congeneric with that from Yunnan (they classified the Potwar holotype as Yunnanochoerus gandakasensis).

Description:
The lower molar from Chiang Muan, Thailand, CMu 426, is a left m/1 lacking its roots.The cusp tips of the mesial cusp pair are lightly worn, exposing small dentine lakes.There is a bucco-lingually broad mesial cingulum which is narrow mesio-distally.The preprotocristid is well developed and descends mesio-centrally and links to the premetacristid forming a loop of enamel that slightly overhangs the mesial cingulum.The median transverse valley is moderately broad and is obstructed in its centre by the median accessory cusplet.At its buccal end, the median transverse valley has low basal pillar.The entoconid is isolated from the hypoconid except at its base.The hypoconulid is small and joined to the hypoconid by the posthypocristid (Fig. 21).The base of the tooth is deep, which distinguishes it from molars of lophodont palaeochoerids.The slight flattening of the lingual cusps agrees with Pilgrim's (1926) description of the molars of this genus.
Because the upper dentition of Lophochoerus nagrii is extremely poorly known, it is appropriate to provide measurements and illustrations of a maxilla from Nagri, Pakistan, collected by G.H.R. Von Koenigswald in 1954.The specimen (Fig. 21B, Table 5) is housed in the Senckenberg Museum, Frankfurt.
Discussion: In a previous paper on the biochronology of Chiang Muan this tooth was identified as a lower molar of Pecarichoerus sminthos (Pickford et al., 2004).However, the median transverse valley is considerably narrower than is usual in lophodont palaeochoerids, the cristids are not as well developed, the Fürchen are open and shallow, and the base of the tooth is deep, all of which are features that indicate that the specimen represents a suid rather than a lophodont palaeochoerid (Fig. 21).Comparisons with specimens from India (Pilgrim, 1926;Pickford, 1988) reveal that it is a left m/1 of Lophochoerus nagrii, a tiny tetraconodont suid.
Von Koenigswald (1963) described briefly and illustrated two upper molars that he obtained from Chinese Drug Stores in Hong Kong, designating the upper first or second molar (Von Koenigswald, 1963, fig. 12) as the holotype of Lophochoerus pilgrimi (the tooth is likely an M2/) and a first upper molar as Lophochoerus cf pilgrimi (Von Koenigswald, 1963, fig. 13).The specimens have the right dimensions and morphology to belong to Lophochoerus nagrii (Table 5).A third lower molar described in the same paper (Von Koenigswald, 1963, fig. 10) as the holotype of Sus officinalis is likely the same species.Its breadth is similar to that in a specimen of Lophochoerus nagrii from the Siwaliks (Table 5) housed in the Yale Peabody Museum (Pickford, 1988) but the Drug Store specimen is somewhat longer (20.5 mm versus 16.7 mm).These records extend the range of the genus to unknown localities in southern China where much of the fossil material in Chinese Drug Stores was obtained (Fig. 1).Pickford (1988) estimated the age range of Lophochoerus nagrii to be 12-9 Ma.The species is exceedingly rare, so this age range is provisional, but it agrees with the age estimate of the Chiang Muan, Thailand, deposits made on the basis of the proboscideans and other mammals (Pickford et al., 2004).The Thai record extends the geographic range of the species 2,700 km to the southeast of its previously known distribution in the Indian Subcontinent (Fig. 1).
Description: The m/3 in the Przeworno mandible is in medium wear and measures 11.8 x 7.0 mm.The buccal and lingual roots are not fused to each other, the lophs are closer together than is the case in Taucanamo, and the precristid of the hypoconid is weaker than in palaeochoerids.The buccal and lingual notches are not deep and broad, but are about half the height of the crown and narrow.
The three lower third molars from La Grive-St-Alban are unworn, and none of them preserves the roots (Fig. 23).The lophs are close together, the buccal and lingual notches are narrow, and shallow, the bases of the teeth being relatively tall.
Estudios Geológicos, 67( 2   Discussion: The mandible fragment from Przeworno housed in the Stuttgart Museum was previously attributed to Taucanamo but it belongs to the small suid Albanohyus and not to a palaeochoerid.This is revealed not only by the morphology of the molar and its roots, but also by the high position of the mandibular condyle (Fig. 22).Previous authors (Kubiak 1981(Kubiak , 1982) ) attributed the small Przeworno suoids to Taucanamo, but Chen (1984) referred them to Barberahyus Golpe-Posse (1977) at the time thought to be a tayassuid, but now known to be a suid.Barberahyus is a junior synonym of Albanohyus Ginsburg (1974) according to Van der Made (1996).The latter author (Van der Made, 1996, 2010) considered that Przeworno has yielded two small suoids, the suid Albanohyus and the palaeochoerid Taucanamo grandaevum, the latter attributed in this paper to Choeromorus grandaevus.The same species occurs at Atzelsdorf, Austria (Daxner-Höck & Bernor, 2009).

Systematics
The following systematic arrangement and synonymy is proposed for Middle and Late Miocene Palaeochoeridae.

Phylogenetic considerations
Some authors (Pickford, 1988;Van der Made, 1997a) have considered that lophodont palaeochoerids evolved from bunodont ancestors.The transition from bunodont to lophodont morphology has caused debate in the literature, with some forms originally classified in Taucanamo by various authors (see synonymy lists) being transferred to Schizochoerus by Van der Made (1997a) (now Schizoporcus) and Pecarichoerus.The relationships between the end products of the lophodont trends (Schizoporcus, Yunnanochoerus) initially proved to be difficult to disentangle because of the high degree of convergence and the relatively poor fossil record.Historically, this has meant that many of the lophodont palaeochoerid fossils were originally thought to represent suids related to Listriodon or even to be anthracotheres, perissodactyls or monkeys.Once better material was found, it became clear that the lophodonty in palaeochoerids was derived independently from that in lophodont suids such as Listriodontinae and Namachoerinae (Pickford, 1995).
Van der Made (2010, fig.32) recognised a deep split between the Schizoporcus and Yunnanochoerus lineages, dating back to the middle of MN 5 (ca 15 Ma), and he recognised Pecarichoerus orientalis as an offshoot of the Schizoporcus lineage.Among the important features differentiating these lineages are the molar and premolar proportions (Van der Made, 1997aMade, , 2010;;Van der Made & Han, 1994).Another is the morphology of the P4/ which in the Choeromorus group has an anteriorly positioned protocone in line with the paracone, with the lingual sinus oriented obliquely forwards, whereas in the Pecarichoerus group the protocone is posteriorly positioned and the lingual sinus is oriented obliquely posteriorly (Fig. 9) (Colbert, 1933;Van der Made, 2010).Orliac et al., (2010b) reached much the same conclusion, highlighting the "sharp oblique ridge running between the anterior and posterior pairs of cusps of the upper molars" that occurs in Pecarichoerus ori-entalis and "Taucanamo" primum from the Early Miocene, but not in species of Taucanamo (i.e.Choeromorus sansaniensis and Choeromorus grandaevus).As these authors pointed out, this ridge is comprised of the premetacristule and the centroconule.

Conclusions
My assessment of the relationships and taxonomy of Miocene Palaeochoeridae taking into account the information yielded by the newly available fossils from the Late Miocene of India and China, is that Van der Made's (2010) scheme has much to recommend it, but it is evident that some changes of nomenclature are needed.Early members of the schizoporcine group are here attributed to Pecarichoerus rather than to Schizoporcus.If these two should eventually prove to be congeneric when their fossil record improves, then Pecarichoerus would prevail as the valid name.I also include Pecarichoerus primus in the genus on account of its shared derived molar morphology with other species of the genus (Orliac et al., 2010b), different from the morphology found in species of Choeromorus.Schizoporcus sinapensis is here considered to be a junior synonym of Schizoporcus arambourgi.
The systematic arrangement adopted in this paper is shown in Table 7, subject to change when poorly understood taxa become better known.Subfamily and tribe names are omitted on account of uncertainty about the utility of some recently proposed subfamily and tribe names ( Van der Made, 1997a, 2010) and of doubts about their contents.Van der Made (2010) subdivided the Palaeochoeridae into three subfamilies (Unknown, Taucanaminae Van der Made, 1997a, andPalaeochoerinae Matthew, 1924) the two named subfamilies being further subdivided into two tribes each (Schizoporcini Van der Made, 2010, and Taucanamini Van der Made, 1997a, for the former and Doliochoerini Simpson, 1945, andPalaeochoerini Matthew, 1924, for the latter).There is a slight possibility, evoked by Van der Made (1997a) and explained in detail by Orliac et al., (2010) that Taucanamo and Pecarichoerus could be synonyms, in which case Pecarichoerus would be the valid name.
This uncertainty impinges on the nomenclature of the Sansan palaeochoerids, Van der Made (1997a) having proposed that, in the interests of stability, Taucanamo Simpson, 1945, should be retained (instead of Choeromorus Gervais, 1850).Given the possible synonymy between Taucanamo and Pecarichoerus (Orliac et al., 2010) then the most stable solution is to recognise Choeromorus Gervais 1850, as the valid genus name for the Sansan palaeochoerids.
Choeromorus Gervais (1850) (type species Choeromorus mamillatus Gervais, 1850) has priority over Taucanamo Simpson (1945).Simpson (1945) recognised the validity of Choeromorus, but erroneously classed it in the Cebochoeridae, and erected the genus Taucanamo to replace Choerotherium (type species Choerotherium sansaniense Lartet, 1851) in ignorance of the fact that Choerotherium sansaniense was a junior synonym of Choeromorus mamillatus, thereby creating a homonym.Van der Made (1997a) (1933) was erected before Taucanamo Simpson (1945) then, if he is right, it ought to be called Pecarichoerus.He also made a plea, in the interests of stability, to suppress Choeromorus Gervais (1850) in preference to Taucanamo Simpson (1945) but given the possibility that Taucanamo and Pecarichoerus may be synonyms (as recognised by Van der Made (1997a) and as discussed by Orliac et al., 2010b) then it is preferable, in the interests of stability, to recognise that the most stable nomenclature flows from using Gervais' (1850) name for these suoids (Choeromorus: type species Choeromorus mamillatus, type locality, Sansan, France).Among the Suoidea, analyses that are based exclusively or mainly on dental comparisons tend to produce weak scenarios of relationship.Liu (2003) considered the genus Schizochoerus to be closely related to Taucanamo and Yunnanochoerus a finding with which I agree, but she grouped the ensemble into an enlarged family Suidae, a suggestion not retained by Orliac et al., (2010a) who concluded that Schizochoerus is part of the Suidae sensu stricto (including some cranial characters) while the other two genera are not (Palaeochoeridae), in contrast to the position demonstrated by Pickford (1988) and supported by Van der Made (1997a).The new material of Yunnanochoerus dangari from Hari Talyangar and Yunnanochoerus gandakasensis from Yuanmou, China, adds support to the views of the latter two authors, as it is clear that, even though fragmentary, the remains belong in the same subfamily as Schizoporcus and Pecarichoerus.In addition, the mandibles from Nsebar, Bulgaria (Nikolov & Thenius, 1967) and Kalfa, Moldova (Lungu, 1971(Lungu, , 1974) ) show marked differences from Suidae, and, apart from size, exhibit closer similarities to Palaeochoeridae such as Choeromorus mamillatus Gervais (1850) from Sansan.
The species Schizoporcus arambourgi (Ozansoy, 1965) is considered to be the same as Schizoporcus sinapensis (Van der Made, 1997a).The holotypes of both species are from the Sinap area, Turkey, which evidently contains two species of Schizoporcus, a small one, S. arambourgi, in the Middle Miocene beds, and a large one, S. vallesensis, in the Late Miocene strata.Some individuals of the younger of the two species, S. vallesensis, possess five upper premolars, an unusual trait among artiodactyls, but known to occur in the anthracothere genus Libycosaurus (Pickford, 1991).
The relatively common European species previously classified in Taucanamo muenzenbergensis (or Schizoporcus muenzenbergensis) (Van der Made, 1998) is transferred to the genus Pecarichoerus, and because the dimensions and morphology of its teeth are similar to those of the type species from Chinji, it is considered to be synonymous with Pecarichoerus orientalis Colbert (1933) at the species level as well.
The hitherto poorly known taxon Yunnanochoerus dangari (Prasad, 1970) from India is now better represented by two mandible fragments from the type locality, Hari Talyangar.The species Yunnanochoerus lufengensis (Han, 1983) is smaller than Y. dangari and Y. gandakasensis, and is therefore valid.
New fossils attributed to Lorancahyus daamsi and Lorancahyus hypsorhizus from St Gérand-le-Puy, France, extend the geographic ranges of these species and suggest a possible ancestor in Propalaeochoerus pusillus.The humerus and talus of L. hypsorhizus reveal its appurtenance to Palaeochoeridae, and indicate that its post-cranial articulations were stabilised in the parasagittal plane.The two species of Lorancahyus are possibly ancestral to a tubulident palaeochoerid from Corcoles, Spain (Alferez et al., 1988).
The origin of the family Hippopotamidae is still the subject of debate (Pickford, 2008).The earliest described members of the family occur in the Middle Miocene of Kenya (Pickford, 2007) represented by Palaeopotamus ternani, which gave rise to Kenyapotamus Pickford, 1983.All the Neogene Eurasian palaeochoerids are too specialised to represent hippopotamid ancestors although they retains several characters such as extremely short neurocranium compared to the splanchnocranium that are also found in hippos..This indicates that the origin of hippos needs to be sought among the Oligocene palaeochoerids; lineages such as Propalaeochoerus and its relatives holding some potential in this respect.
maxilla was provided by Gertrud Rössner, München.I would like to thank Nina Jablonski for translating the article by Gu Yumin and for images of the specimen.Thanks to Christine Argot, MNHN, for access to fossils from St Gérand-le-Puy and to Jorge Morales, Museo Nacional de Ciencias Naturales, Madrid and to Marguerite Huguenay and Abel Prieur, Faculty of Science Museum, University of Lyon for access to fossils from Loranca and Coderet respectively.This study was carried out with support from the CNRS UMR 7207, the Département Histoire de la Terre, Muséum National d'Histoire Naturelle and the Collège de France, 541-578, julio-diciembre 2011.ISSN: 0367-0449.doi:10.3989/egeol.40634.206

Fig. 4 .
Fig. 4.-A) MNCN 35306, rootless left M3/ from Cetina de Aragon, Spain, attributed to Propalaeochoerus cf pusillus.The cusps are more crowded and puffier than is the case in upper molars of Lorancahyus hypsorhizus (B, C) but the tooth is compatible in dimensions with this taxon (scale : 10 mm).

Fig. 8 .
Fig. 8.-Bivariate plots of upper second and third molars of Palaeochoeridae.Note that the fossils from Chinji, Pakistan (Pecarichoerus orientalis) plot close to those from Sandelzhausen, Leoben and Neudorf (grey ovals) previously classified as Schizoporcusmuenzenbergensis, but here considered to belong to the same species as that from Chinji.The Xiaodian tooth is here plotted as an M2/, but it could be an M1/ of Pecarichoerus orientalis instead.

Fig. 16 .
Fig. 16.-Relationship between upper and lower molars of species of Schizoporcus (V -Schizoporcus vallesensis, A -Schizoporcus arambourgi).The holotype mandible of S. arambourgi is compatible in size with the holotype snout of S. sinapensis, indicating that the two species are synonyms (? signifies that the m/3 of S. arambourgi is unknown, but that it is likely to have approximately the dimensions indicated on the basis of the size of the m/2).
Fig. 20.-Length-breadth plots of the lower molars of lophodont Palaeochoeridae.The Pasalar specimens have until recently been attributed to Choeromorus inonuensis (ex Taucanamo inonuensis) but the cusps are rather lophodont, suggesting a possible transfer to Pecarichoerus, close to Pecarichoerus orientalis.Molars of Yunnanochoerus tend to be narrower than those of Schizoporcus.(N.B.the breadth measurement of Prasad's specimen from Hari Talyangar is probably too narrow -arrow in figure).Note also the relatively small dimensions of the specimen from Yassiören when compared with those of Schizoporcus vallesensis, which supports its identification as a distinct species, Schizoporcus arambourgi(Ozansoy, 1965).

Table 1 .-Distribution of lophodont Palaeochoeridae in Eurasia arranged by country (Vos et al., 2007, recorded Taucanamo? / Yunnanochoerus at Kastelios Hill 2 (MN 9-10), Crete. The material basis for this listing seems not to have been published)
Locality(1970) is pre-occupied by Anthraconema Zur Strassen (1904) a free living nematode, itself a synonym of Siphonolaimus De Man (1893).Thus Yunnanochoerus is the valid name for this genus of palaeochoerid.

in mm) of the teeth of Lorancahyus hypsorhizus (FSL and LO) and Lorancahyus daamsi (MOH and SG) (+ = measurements are provided for the alveolus of m/1, the crown would be somewhat larger than these measurements)
considered that the fossils from Bugti, Chinji and Hubei could belong to a single species, but with the better samples now available, especially those from Europe (Van der Made, 2010), it is clearer that the bulk of palaeochoerid fossils from Bugti, Chinji, Sihong and Xiaodian belong to a species smaller than Pecarichoerus orientalis, although the Estudios Geológicos, 67(2), 541-578, julio-diciembre 2011.ISSN: 0367-0449.doi:10.3989/egeol.40634.206

in mm) of the upper molars of Eurasian Palaeochoeridae
Colbert (1933)but it are slightly narrower.What is intriguing is that the morphology is close to that of Schizoporcus muenzenbergensis (Van der Made

in mm) of lower molars of Yunnanochoerus dangari Prasad (1970) from Hari Talyangar, Himachal Pradesh, India, other material of the genus and of Pecarichoerus, Choeromorus and Schizoporcus from Asia and Europe (e = estimated measurement)
Table 4 (continuation).-Measurements(