Extant perissodactyls (horses, rhinos, and tapirs) comprise a small portion of living mammals, but fossil perissodactyls were more diverse and commonly dominated Paleogene faunas. Unfortunately, the taxonomy and distribution of some Chinese Paleogene perissodactyls remain controversial and unclear, hampering the correlation of Asian paleofaunas with paleofaunas from other continents. Here we clarify the temporal and spatial distribution of Paleogene perissodactyl species from the Erlian Basin based on published specimens, archives, and our recent fieldwork. The strata of the Erlian Basin range nearly continuously from the late Paleocene to the early Oligocene, and almost all Eocene Asian Land Mammal Ages (ALMA) are based on corresponding faunas from the Erlian Basin. We revise the most complete section of deposits at Erden Obo (= Urtyn Obo) that range in age from the late Paleocene to the early Oligocene in the Erlian Basin, and correlate it with other type formations/faunas in the basin based mainly on the perissodactyl biostratigraphy and lithostratigraphy. Furthermore, we discuss perissodactyl faunal components and their diversity from the early Eocene to the early Oligocene in the Erlian Basin, as well as the correlation between middle Eocene ALMAs and North American Land Mammal Ages based on perissodactyl fossils. The general decrease in perissodactyl diversity from the middle Eocene to the late Eocene can probably be attributed to a global climatic cooling trend and related environmental changes. The diversity of perissodactyls declined distinctly during the Eocene-Oligocene Transition, when global average temperatures dropped considerably.
INTRODUCTION
The Erlian Basin is located in central Inner Mongolia (Nei Mongol), near the China-Mongolia border (Jiang, 1983; Wang et al., 2012). Deposits of the Erlian Basin range nearly continuously from the late Paleocene to the Oligocene, and produce a large number of fossil mammals and other vertebrates. The lithostratigraphic units in the Erlian Basin include the Houldjin, Arshanto, Irdin Manha, Tukhum, Ulan Shireh, Shara Murun, Ulan Gochu, and Baron Sog formations, named by the Central Asiatic Expedition (CAE), as well as the Nomogen Formation and other lithological units that were proposed by later researchers (Jiang, 1983; Wang et al., 2012). The basin was first investigated by the Russian geologist V.A. Obruchev in 1892, and then during the 1920s, the CAE of the American Museum of Natural History extensively explored the area over five field seasons (fig. 1) (Andrews, 1932; Berkey and Morris, 1927). After the CAE suspended its work after 1930, the Sino-Soviet Paleontological Expedition (SSPE) conducted investigations in 1959–1960 the fossiliferous deposits of the Erlian Basin (Chow and Rozhdestvensky, 1960), followed by the Regional Geological Survey of the Nei Mongol Autonomous Region (1960s–1970s) (Jiang, 1983) and the Institute of Vertebrate Paleontology and Paleoanthropology, in collaboration with the American Museum of Natural History and Carnegie Museum of Natural History, from the 1980s to the present time (Meng et al., 2007; Wang et al., 2010; Jin, 2012; Wang et al., 2012). Derived from that extensive work, the Eocene Asian Land Mammal Ages (ALMA) were established based mainly on the mammalian faunas from the Erlian Basin (Romer, 1966; Tong et al., 1995; Wang et al., 2007; Vandenberghe et al., 2012).
The Eocene mammalian faunas of the Erlian Basin are dominated by diverse perissodactyls (Tong et al., 1995; Wang et al., 2007). Perissodactyls from the Erlian Basin consist of members of Tapiroidea, Rhinocerotoidea, Brontotheriidae, and Chalicotherioidea. Since the first report on perissodactyl paraceratheres from Houldjin (Osborn, 1923a), numerous studies, including many papers and some monographs, have contributed to our knowledge of perissodactyls based on the abundant, relatively complete material from the Erlian Basin. The brontotheres were reviewed by Granger and Gregory (1943) and Mihlbachler (2008) on the basis of numerous, complete skulls in the CAE collection. Wang (1982) described a skeleton of the brontothere Rhinotitan mongoliensis in detail based on material in the SSPE collection from Ula Usu. Radinsky (1965, 1967) studied the tapiroids and hyracodontids from the Paleogene of Asia based principally on the CAE collection. Xu (1966) analyzed many amynodont fossils collected by the SSPE from Inner Mongolia. Qiu and Wang (2007) reviewed the Asian Paraceratheriidae and studied a nearly complete skeleton of Juxia from Ula Usu. However, some perissodactyls from various localities in the Erlian Basin, particularly specimens in the CAE collection, were reported from horizons and formations that had been misinterpreted because of the complexity of correlating terrestrial stratigraphic sections. For instance, the names “Irdin Manha Formation” and “Houldjin Formation” applied by the CAE in the Huheboerhe area (Camp Margetts and neighboring areas) are mostly considered today to be Arshanto and Irdin Manha formations, respectively (Meng et al., 2007). On the other hand, most CAE specimens fortunately retain an association between their field numbers and related field notes, which indicate more specific, accurate layers in the local stratigraphic sections (Meng et al., 2007). Over the past decade, we have recovered many new perissodactyl fossil specimens in the Erlian Basin, and detailed studies of most of them are ongoing. Here, we clarify the distributions of the Paleogene perissodactyls at specific levels from the Erlian Basin based on published specimens, archives, and the preliminary results of our recent fieldwork (appendices 1, 2). In this paper, we focus on biostratigraphic and lithostratigraphic correlations rather than the morphology of perissodactyls from the Erlian Basin. Additionally, a revised correlation of the Eocene ALMAs with their corresponding NALMAs is proposed based mainly on perissodactyl fossils. It also worth noting that a decrease in perissodactyl diversity from the middle Eocene to the early Oligocene likely can be attributed to a cooling trend of global climatic change over that period of time.
FIG. 1.
Paleogene fossil localities in the Erlian Basin of Inner Mongolia, China, and the related routes of Central Asiatic Expeditions during 1920s (modified from Jiang, 1983; Mao and Wang, 2012; Wang et al., 2012). 1, Houldjin; 2, Arshanto; 3, Irdin Manha; 4, Daoteyin Obo (= Overnight Camp, 5 mi east of Camp Margetts); 5, Duheminboerhe (= Camp Margettes); 6, Nuhetingboerhe (= 6 mi west of Camp Margetts); 7, Wulanboerhe; 8, Huheboerhe (= 7 mi west and southwest [235°] of Camp Margetts); 9, Chaganboerhe (= 10 mi southwest of Camp Margetts); 10, Bayan Ulan; 11, Nom Khong (= Holy Mesa); 12, Wulantaolegai (= Viper Camp, 4 mi north of Tukhum Lamasery); 13, Wulanhuxiu (= Chimney Buttes, 8 mi north of Tukhum Lamasery); 14, Erden Obo (= Urtyn Obo); 15, Ganggan Obo (= Ulan Shireh Obo); 16, Heretu (= Spring Camp); 17, Bayan Obo (= Twin Obos); 18, Jhama Obo; 19, Xilin Nor North (= 4 mi north of Baron Sog Lamasery); 20, Ulan Gochu (= 8 mi north of Baron Sog Lamasery); 21, Ula Usu. The black-and-white dashed line represents railway.
MATERIAL AND METHODS
The perissodactyl fossils from the Erlian Basin are housed mainly at the Division of Paleontology, American Museum of Natural History (AMNH), and in the Institute of Vertebrate Paleontology and Paleoanthropology (IVPP). A few CAE collections also are curated in the Field Museum (but sent by AMNH), and some CAE fossils that were sent to the Chinese Geological Survey Museum are missing. The Paleontological Institute of the Russian Academy of Sciences also houses some SSPE specimens from the Erlian Basin.
We checked field numbers of CAE specimens as recorded in Granger's CAE field notes that are stored in the AMNH Library. Those field numbers usually indicate specific and accurate horizons where the specimens were collected, contrasting with the lithological formations that sometimes have been misinterpreted or are otherwise misleading. Although some beds or formations recorded in the field notes were misinterpreted, we still use their original information, denoted with quotation marks. For example, the fossil-bearing layers from Wulanhuxiu originally were interpreted as “Shara Murun beds” in 1925, rather than the Ulan Shireh Formation. In some cases, some specimens' field numbers were not recorded in Granger's field notes, but rather on the labels associated with specimens, and a few specimens probably lack field numbers. On the other hand, the field numbers of a few specimens that were surface collected can be misleading, as exemplified by a single individual of Schizotherium recorded from different horizons with different field numbers (Radinsky, 1964a; Coombs, 1978).
RESULTS
Tapiroidea Gill, 1872
Lophialetidae
Matthew and Granger, 1925c
Minchenoletes Wang et al., 2011
Wang et al. (2011) erected a new genus and species, Minchenoletes erlianensis, from the upper part of the Nomogen Formation at Nuhetingboerhe and Wulanboerhe (fig. 1). The specimens attributed to that taxon include maxillae, mandibles, and isolated teeth, and were unearthed from the NM-3 and 4 horizons of the Bumbanian ALMA.
Schlosseria Matthew and Granger, 1926
Only one species of Schlosseria, S. magister, has been named from the basin (Matthew and Granger, 1926). The holotype (AMNH FM 20241) was unearthed from the Arshanto Forma tion, about 7 mi north of the Telegraph Line (fig. 1). Radinsky (1965) assigned additional dental and postcranial material to Schlosseria cf. S. magister from Huheboerhe area. Those localities include Duheminboerhe, Daoteyin Obo, Huheboerhe, and Chaganboerhe. Our recent work confirms the presence of Schlosseria magister at Huheboerhe (Li and Wang, 2010), which ranges from As-1 to As-6 fossil-bearing horizons and was considered restricted to the Arshanto Formation. However, Radinsky (1965) also considered AMNH FM 81787 and 81788 as Schlosseria cf. S. magister from “?Houldjin gravels” (i.e., Irdin Manha Formation) at Chaganboerhe.
Radinsky (1965) reported a lower jaw with p2-m2 of ?Schlosseria (AMNH FM 26139) from “?Shara Murun beds” at Erden Obo (Urtyn Obo), and the fossil-bearing bed is probably equivalent to the “Basal White.” Li and Wang (2010) described a partial skull of S. magister (IVPP V 16573) from the “Basal White” of Erden Obo. Our recent fieldwork also recovered some Schlosseria specimens from the basal layers at Wulanhuxiu, indicating the presence of Arshantan deposits in the section.
Lophialetes Matthew and Granger, 1925c
Only one species of Lophialetes, L. expeditus, is known from the Erlian Basin (fig. 2A) (Matthew and Granger, 1925c). The holotype (AMNH FM 19163) was found in the Irdin Manha Formation on the Irdin Manha escarpment. Radinsky (1965) also reported the species from both the “Irdin Manha” (i.e., Arshanto) and “Houldjin” (i.e., Irdin Manha) beds in the Huheboerhe area. However, he considered all these specimens as likely deriving from a single horizon. This proposal was supported by Li and Wang (2010) and Wang et al. (2010), who restricted L. expeditus to the IM-1 and IM-2 horizons of the Irdin Manha Formation.
Radinsky (1965) also reported Lophialetes expeditus? from the Ulan Shireh Formation at Wulantaolegai and Wulanhuxiu. These specimens are slightly smaller than Irdin Manha individuals of L. expeditus. He further assigned several larger specimens (AMNH FM 81687, 81690, and 81681) to Lophialetes sp. from the same bed at Wulanhuxiu (Radinsky, 1965). Those specimens are approximately 20% to 25% larger than specimens of L. expeditus? from the same bed, and ~12% larger than Irdin Manha fossils of L. expeditus.
Radinsky (1965) assigned several crushed skulls, jaws, and foot bones (AMNH FM 22091–22095) to Lophialetes expeditus? from the Tukhum Formation, which is overlain by the Shara Murun Formation at Ula Usu. He considered these specimens the same as Ulan Shireh L. expeditus? individuals. A lower jaw with p2-m2 (AMNH FM 26138) from “?Shara Murun beds” at Erden Obo was referred to Lophialetes sp., which corresponds in size to the larger Ulan Shireh Lophialetes fossils (Radinsky, 1965).
Zhongjianoletes Ye, 1983
Only one species of Zhongjianoletes, Z. chowi, known from a left mandible with a partial symphysis (IVPP V 6671), is known from the Ulan Shireh Formation at North Mesa (Ye, 1983). The species is very similar to Lophialetes expeditus, but is larger and lacks the p1 and probably incisors. Ye (1983) further considered that specimens referred to Lophialetes sp. by Radinsky (1965) from the same locality should belong to Z. chowi. Ye (1983) assigned some isolated M3s to Zhongjianoletes sp., which corresponds to the size of Z. chowi from the same horizon.
Simplaletes Qi, 1980
Two species of Simplaletes, S. sujiensis and S. ulanshirehensis, are known from the Irdin Manha and Ulan Shireh, respectively (Qi, 1980). Simplaletes sujiensis was found in the Irdin Manha Formation, and S. ulanshirehensis was discovered in the Ulan Shireh Formation in 1959 by SSPE. Both of those species preserve only lower jaws with relatively heavily worn teeth, and lack the p1. Lucas et al. (1997) considered Simplaletes to be a junior synonym of Schlosseria.
Breviodon Radinsky, 1965
Two species of Breviodon, B. acares and B.? minutus, are known from the Wulanhuxiu and Irdin Manha, respectively (Radinsky, 1965). The holotype of B. acares is a mandible (AMNH FM 26113) characterized by the absence of p1–2 and it is about 30% smaller than Ulan Shireh material of Lophialetes expeditus?. Radinsky (1965) assigned a fragmentary skull (AMNH FM 81751) to Breviodon cf. B. acares from the “Irdin Manhan beds” at Huheboerhe. However, this specimen more likely was recovered from the Arshanto Formation instead of the “Irdin Manha Formation.” A lower jaw with dp4-m2 (AMNH FM 81836) from the same locality may be conspecific with the fragmentary skull (Radinsky, 1965). Reshetov (1975) erected a new genus Parabreviodon based on AMNH FM 81751 as the generic type. A lower jaw (AMNH FM 26115, field no. 645) from Wulanhuxiu and a few isolated teeth (AMNH FM 81839) from the “Irdin Manha beds” (i.e., the Arshanto Formation) at Daoteyin Obo, respectively, were assigned to ?Breviodon based on inadequate material (Radinsky, 1965).
Radinsky (1965) reassigned the holotype of “Lophialetes” minutus (an isolated left upper molar, AMNH FM 20139) and some miscellaneous teeth (AMNH FM 81721A, 81721B, 81722) from the Irdin Manha Formation of the Irdin Manha escarpment to Breviodon? minutus. However, Reshetov (1979) considered Breviodon acares as a junior synonym of B. minutus based on the associated upper and lower cheek teeth. This notion was followed by Qi (1987), who also reported the species from the Arshanto Formation at Huheboerhe and Wulanboerhe.
Rhodopagus Radinsky, 1965
Two species of Rhodopagus, R. pygmaeus and R.? minimus, are known from North Mesa and Ula Usu, respectively (Radinsky, 1965). Rhodopagus pygmaeus is known mainly from the Ulan Shireh Formation, and Radinsky (1965) also assigned a couple of specimens (AMNH FM 81842, 81843) to ?R. pygmaeus from the type Irdin Manha bed. Radinsky (1965) transferred Caenolophus? minimus (AMNH FM 20310) from the Shara Murun Formation at Ula Usu to Rhodopagus? minimus based on its short trigonid and reduced cristid obliqua. Lucas and Schoch (1981) further considered that Rhodopagus pygmaeus is a junior synonym of R. minimus.
The phylogenetic position of Rhodopagus is controversial with its affinities to Lophialetidae, Helaletidae, or even Hyracodontidae (Lucas and Schoch, 1981). Our recent discovery of Rhodopagus from the “Basal White” of Erden Obo will probably clarify this phylogenetic mystery in the future.
Pataecops Radinsky, 1966
Only one species of Pataecops, P. parvus, is known from the upper part of the Nomogen Formation at Nuhetingboerhe and Wulanboerhe (Wang et al., 2011). Wang et al. (2011) pointed out that the holotype of the species from the Kholobolchi Formation of Mongolia is Bumbanian in age, as are specimens from the Erlian Basin. Although Radinsky (1965) assigned Pataecops to Lophialetidae, the phylogenetic position of Pataecops parvus is uncertain. If P. parvus belongs to rhinocerotoids rather than tapiroids (Lucas and Schoch, 1981; Dashzeveg and Hooker, 1997), it would extend the earliest record of rhinocerotoids into the earliest Eocene (Wang et al., 2011).
Deperetellidae Radinsky, 1965
Teleolophus Matthew and Granger, 1925c
Four species of Teleolophus have been reported from the Erlian Basin. Teleolophus primarius and T.? rectus are known from the Arshanto Formation in Huheboerhe area (Qi, 1987). The material of Teleolophus primarius is composed of several fragmentary mandibles, carpals, and phalanges from Wulanboerhe and Huheboerhe. Teleolophus? rectus is known from a left mandible with p4-m1 (IVPP V 5766) from Wulanboerhe (Qi, 1987).
Teleolophus medius, including the holotype, was reported from the Irdin Manha Formation at the Irdin Manha escarpment, and is known mainly from mandibles. Radinsky (1965) also assigned several upper dentitions (e.g., AMNH FM 26286) and a left fragmentary mandible with p3-m1 (AMNH FM 81796) to T. medius? from the Ulan Shireh Formation at Wulanhuxiu and Wulantaolegai. Ulan Shireh fossils of T. medius? are slightly smaller than Irdin Manha species, and have a slightly squarer M3. The material from Ulan Shireh provides a complete record of upper dental morphology of Teleolophus medius?. However, the paucity of lower teeth from Ulan Shireh hinders comparison with Irdin Manha specimens of T. medius, which are preserved primarily as lower dentitions. Recently, a large number of postcrania of Teleolophus medius? were reported from the Ulan Shireh Formation at Wulanhuxiu (Bai et al., 2018a).
Radinsky (1965) also referred several specimens to Teleolophus cf. T. medius from both the Arshanto and Irdin Manha formations in Huheboerhe area. Whether those CAE materials from the Arshanto Formation are conspecific with T. primarius and/or T.? rectus named by Qi (1987) needs further investigation. Radinsky (1965) referred a right mandible with p4-m3 (AMNH FM 81852, field no. 893) to Teleolophus sp. from the “?Houldjin gravels” at Huheboerhe. This specimen is similar to T. medius, but is about 20% larger.
Teleolophus magnum, known from a maxilla and a mandible (AMNH FM 26063), has been reported from the base of the “Middle Red” at Erden Obo (Radinsky, 1965). The morphology of T. magnum is somewhat intermediate between T. medius and Deperetella cristata. However, its stratigraphic horizon is even higher than that bearing D. cristata (Radinsky, 1965). Radinsky (1965) proposed three explanations for this discrepancy, and the possibility that the reported horizon bearing T. mangus is incorrect cannot be excluded.
Deperetella Matthew and Granger, 1925b
Only one species of Deperetella, D. cristata, is known from the Shara Murun Formation at Ula Usu. Deperetella cristata is represented by maxillae, mandibles, and many postcraninal elements (Radinsky, 1965). Radinsky (1965) referred a mandible with dp3-m1 (AMNH FM 26027) to Deperetella cf. D. cristata from “Shara Murun beds” at Twin Obos. This specimen is about 25% larger than Ula Usu fossils of D. cristata. Radinsky (1965) further referred a right mandible with a complete dentition except for p1 (AMNH FM 81807) to Deperetella sp. from the “?Houldjin gravels” at Huheboerhe. This specimen was assigned to Deperetella rather than to Teleolophus mainly based on the presence of complete hypolophids on p2–4.
Helaletidae Osborn, 1892
Heptodon Cope, 1882
Desmatotherium Scott, 1883
Desmatotherium mongoliense was first reported by Osborn (1923b) from the Irdin Manha Formation on the Irdin Manha escarpment. The species has been referred to Helaletes and Irdinolophus (Radinsky, 1965; Dashzeveg and Hooker, 1997), but recent study confirms its affinity with Desmatotherium (Bai et al., 2017). The species has been reported from the Irdin Manha Formation on the Irdin Manha escarpment, Duheminboerhe, Huheboerhe, and Chaganboerhe (Bai et al., 2017).
Paracolodon Matthew and Granger, 1925a
Paracolodon (= Desmatotherium) fissus was first reported by Matthew and Granger (1925c) from the Irdin Manha Formation at Camp Margetts (fig. 2B). The species was assigned originally to Desmatotherium, and then referred to Helaletes or Colodon (Radinsky, 1965; Dashzeveg and Hooker, 1997). Based on a partial skull, Bai et al. (2017) recently assigned the species to Paracolodon, suggesting its ancestral relationship with P. inceptus from the Ergilin of Mongolia. Paracolodon fissus is present in the Irdin Manha Formation at Duheminboerhe and Daoteyin Obo.
Qi (1987) named a new species Helaletes medius based mainly on a fragmentary mandible with p4-m3 (IVPP V 5729) from the Arshanto Formation, Wulanboerhe. Bai et al. (2017) questioned the assignment of this species to Helaletes, and noted some similarities with Schlosseria.
Qi (1987) named a new species Homogalax reliquius based on two m3 specimens (IVPP V 5747, 5748) from Huheboerhe in the Arshanto Formation. The species was assigned initially to Homogalax on the basis of their robust hypoconulids (Qi, 1987). However, the lack of a “metastylid” and the presence of a relatively long trigonid suggest that the species is more similar to Isectolophus than to Homogalax. Furthermore, Lucas et al. (2003) considered Homogalax reliquius to be a junior synonym of Isectolophus latidens.
Rhinocerotoidea Gray, 1825
Hyrachyus Leidy, 1871
Two species of Hyrachyus have been reported from the Erlian Basin. Hyrachyus neimongoliensis (IVPP V 5721) was unearthed in the upper part of the Arshanto Formation at Huhe boerhe (Qi, 1987). Hyrachyus crista is represented by several fragmentary maxillae with P4-M3 (IVPP V 5722), and was discovered in the Arshanto Formation at Bayan Ulan (Qi, 1987).
Qi (1987) assigned a left M3 (IVPP V 5728) to Hyrachyus sp. cf. H. eximius from the Arshanto Formation at Huheboerhe. He also noted some similarities in that species with the M3 of the helaletid Paracolodon (= Colodon) inceptus. Qi (1987) referred other miscellaneous teeth to Hyrachyus sp. from the Arshanto Formation at Daoteyin Obo and Huheboerhe (field no. 77028H2, 77036-2, and 77039). However, whether or not all these specimens should be allocated to Hyrachyus is questionable. Radinsky (1965) assigned a left maxilla with P4-M3 (AMNH FM 81801) to cf. Hyrachyus from the Arshanto Formation at Huheboerhe.
Hyracodontidae Cope, 1879
Triplopus Cope, 1880
Two species of Triplopus have been reported from the Erlian Basin. Triplopus? proficiens was assigned initially to Caenolophus by Matthew and Granger (1925b). The holotype of the species (AMNH FM 20141), a mandible with p1-m3, was found in the Irdin Manha Formation at Irdin Manha. More complete material of T.? proficiens is known from the Ulan Shireh bed in Shara Murun region, and it is smaller with a straight distal border of m3 as compared with the Irdin Manha specimens (Radinsky, 1967). Radinsky (1967) further referred two mandibles (AMNH FM 26674, 26675) and a maxilla (AMNH FM 26673) from the “Irdin Manha beds” at Camp Margetts and Huheboerhe, respectively, to T.? proficiens. If the so-called “Irdin Manha beds” bearing T.? proficiens are indeed the Arshanto Formation, this species extends into the earlier Arshantan.
Triplopus? progressus is known from the Shara Murun Formation at Ula Usu (Radinsky, 1967), and was assigned initially to Caenolophus by Matthew and Granger (1925b). The species is known from a maxilla with M1–3 (AMNH FM 20298), m2–3 (AMNH FM 20309), and a M3 (AMNH FM 81872). Although T.? progressus is smaller than T.? proficiens, it is more advanced than the latter in having a reduced M3 metacone and parastyle (Radinsky, 1967).
Teilhardia Matthew and Granger, 1926
Only one species of Teilhardia, T. pretiosa, is known from the “lower red beds” at the base of the Shara Murun Formation at Ula Usu (Matthew and Granger, 1926). The species is known from a right mandible with p2-m3 (AMNH FM 20299), and the “lower red beds” were later called the Tukhum Formation (Berkey and Morris, 1927). However, whether T. pretiosa was found in the Tukhum Formation is controversial because some typical Sharamurunian taxa, such as Rhodopagus? minimus, Deperetella cristata, Triplopus? progressus, Caenolophus promissus, and Pterodon hyaenoides, were also recorded with the same field number used for T. pretiosa (field no. 191) (Wang et al., 2012). Radinsky (1967) suggested its affinity with amynodontids based on its relatively shorter premolars relative to the molars, proposing that it may be ancestral to Caenolophus promissus from the overlying Shara Murun Formation.
Ulania Qi, 1990a
Qi (1990a) reported Ulania wilsoni from the grayish-green sandstone of the lower part of the “Ulan Gochu Formation” at Erden Obo, and the horizon is probably equivalent to a part of the “Lower White” or the “Middle White” (Qi, 1990b; Wang et al., 2012). The species is known from what is thought to be parts of the same individual IVPP V 8922, including a maxilla, mandibles, carpals, and phalanges (Qi, 1990a). However, Qiu and Wang (2007) doubted the association of the specimens, and referred the maxilla to Caenolophus.
Ardynia Matthew and Granger, 1923b
Two species of Ardynia have been reported from the Erlian Basin: A. praecox and A. kazachstanensis (Radinsky, 1967). Ardynia praecox is known from AMNH FM 26039, which is composed of a skull, a mandible, and a few postcranial elements. The specimens are known from the “Ulan Gochu beds” at Erden Obo, and the field number (field no. 747) indicates their stratigraphic horizon as the “Middle Gray layer” (Bai et al., 2018b). Ardynia kazachstanensis is known from a mandible (AMNH FM 26183) from the “Baron Sog beds” at Nom Khong Obo, and the horizon is equivalent to the “Upper White” (Radinsky, 1967; Wang, 2003).
Proeggysodon Bai and Wang, 2012
Only one species of Proeggysodon, P. qiui, was unearthed from the upper part of the “Middle White” at Erden Obo (Bai and Wang, 2012). The species is known from a mandible with incisors, canine, and cheek teeth (IVPP V 18099), and the species is considered ancestral to the European Oligocene Eggysodon.
Paraceratheriidae Osborn, 1923
Pappaceras Wood, 1963
Three species of Pappaceras, P. confluens, P. minuta, and P. meiomenus, have been reported from the Erlian Basin (Wood, 1963; Lucas et al., 1981; Wang et al., 2016). Although Radinsky (1967) considered Pappaceras a junior synonym of Forstercooperia, the validity of Pappaceras has been supported by other authors (Qiu and Wang, 2007; Wang et al., 2016). The type of P. confluens (AMNH FM 26660) was recovered from the “Upper Gray clays” of the Arshanto Formation at Chaganboerhe. A referred mandible of P. confluens (AMNH FM 26666) was found in the Arshanto Formation at Huheboerhe. A left P2 of P. confluens (AMNH FM 26667) presumably derives from the Arshanto Formation in the Camp Margetts area.
The type of P. minuta (AMNH FM 26672) was found in the “Upper Gray clays” of the Arshanto Formation at Chaganboerhe (Lucas et al., 1981). The species also is known from the Arshanto Formation at Huheboerhe and Daoteyin Obo (Lucas et al., 1981; Wang et al., 2018). A nearly complete mandible of P. minuta (AMNH FM 26056) also was reported from the Shara Murun Formation (?), 6 mi east of Spring Camp at East Mesa.
The type of P. meiomenus is known from a nearly complete skull (IVPP V 20254), which was found in the upper part of the Arshanto Formation at Huheboerhe (fig. 2C) (Wang et al., 2016). Wang et al. (2018) further separated the original specimens of F. minuta into different groups (i.e., P. minuta, P. meiomenus, and P. sp.) mainly based on the presence or absence of p1.
Forstercooperia Wood, 1938
Two species of Forstercooperia, F. totadentata and F. ulanshirehensis, have been reported from the Erlian Basin (Wood, 1938; Wang et al., 2018). Forstercooperia totadentata (AMNH FM 20116) was discovered in the Irdin Manha Formation at Irdin Manha. Recently, Wang et al. (2018) reported a new species F. ulanshirehensis from the Ulan Shireh Formation at Wulanhuxiu and Wulantaolegai, as well as from the Irdin Manha Formation at Irdin Manha. Forstercooperia ulanshirehensis is known from a nearly complete skull, mandibles, and some fragmentary maxillae. Wang et al. (2018) further considered “Forstercooperia” huhebulakensis, erected by Qi (1987), to be a junior synonym of Pappaceras confluens.
Juxia Chow and Chiu, 1964
Two species of Juxia, J. sharamurenensis and J. shoui, have been reported from the Erlian Basin (Chow and Chiu, 1964; Qi and Zhou, 1989; Qiu and Wang, 2007). Juxia sharamurenensis, known from a nearly complete skeleton (IVPP V 2891), was discovered in the Shara Murun Formation at Ula Usu (fig. 2D). Additional specimens housed at the American Museum of Natural History were referred to J. sharamurenensis (Radinsky, 1967), but they have not been studied in detail. Those specimens include: skeletal material (AMNH FM 20286–20288) from the Shara Murun Formation at Ula Usu, a mandible (AMNH FM 26753) from the base of “Lower Gray” at Erden Obo, and a mandible (AMNH FM 26750) from the “Shara Murun Formation” at Twin Obos.
Juxia shoui, known from a fragmentary skull with C-M2 (IVPP V 8757) and a lunar (IVPP V 3268), was recorded from the “Ulan Gochu Formation” at Erden Obo (Qi and Zhou, 1989). The lunar was discovered in the “Lower White” where the skull (V 8757) presumably was also found.
Urtinotherium Chow and Qiu, 1963
Two species of Urtinotherium, U. intermedium and U. parvum, have been reported from the Erlian Basin (Chow and Qiu, 1963; Qiu and Wang, 2007). The type of U. intermedium (IVPP V 2769) was recovered from the “Middle Gray” at Erden Obo. Qiu and Wang (2007) referred a radius (AMNH FM 26062) to U. intermedium from the same horizon and locality. Qiu and Wang (2007) also referred a mandible (AMNH FM 26032) from the “Ulan Gochu Formation” at Jhama Obo to the species, and the specimen was initially assigned to “Indricotherium” parvum by Radinsky (1967). Urtinotherium intermedium is also known from an atlas (AMNH FM 26390) and a right Mc III (AMNH FM 26389) from the “Houldjin Formation” at Huheboerhe and Camp Margetts, respectively. Qiu and Wang (2007) further considered “Dzungariotherium erdenensis” (IVPP V 8803) from the “Ulan Gochu Formation” at Erden Obo (Qi, 1989) to be a synonym of U. intermedium. However, its specific horizon remains unclear.
Urtinotherium parvum is known from a maxilla with P1-M3 (EMM 0146) from the “Lower White” at Erden Obo, and a few postcranial specimens (AMNH FM 26190) from the “Baron Sog beds” at Jhama Obo.
Paraceratherium Forster-Cooper, 1911
Only one species of Paraceratherium, P. grangeri, has been reported from the Erlian Basin (Qiu and Wang, 2007). The following specimens were assigned to the species: AMNH FM 26166 and 26179 from the “Baluchitheres bed” (= “Upper White”) at Erden Obo and AMNH FM 26387 from the “Houdljin gravels” at Daoteyin Obo (Qiu and Wang, 2007).
Aralotherium Borissiak, 1939
Qiu and Wang (2007) assigned some dental and postcranial material (e.g., EMM 0016) to Aralotherium sp. from the type Houdljin Formation near Erlian Hot.
Amynodontidae Scott and Osborn, 1883
Rostriamynodon Wall and Manning, 1986
Only one species of Rostriamynodon, R. grangeri (fig. 2E), is known from the “Irdin Manha Formation,” 2 mi east of Camp Margetts (Wall and Manning, 1986). The species is known from a complete skull and mandible (AMNH FM 107635), and likely was unearthed from the Irdin Manha Formation, rather than the Arshanto Formation (Bai et al., 2017). Rostriamynodon grangeri was considered to be a primitive member of the amynodontids (Wall and Manning, 1986).
Caenolophus Matthew and Granger, 1925b
Two species of Caenolophus, C. promissus and C. obliquus, have been named from the Shara Murun Formation at Ula Usu (Matthew and Granger, 1925b; Radinsky, 1967). Caenolophus promissus is known from a maxilla with P3-M3 (AMNH FM 20297) and a mandible with m1–2 (AMNH FM 20304) (Matthew and Granger, 1925b). Caenolophus obliquus is known from a maxilla with DP3-DP4, P3-M2 (AMNH FM 20296) (Matthew and Granger, 1925b). Caenolophus was originally assigned to the Hyracodontidae (Matthew and Granger, 1925b), but Radinsky (1967) suggested its affinity with the amynodontids.
Sharamynodon Osborn, 1936
Osborn (1936) briefly described a nearly complete skeleton of “Amynodon” monogoliensis (AMNH FM 20278) from the Shara Murun Formation at Ula Usu (fig. 2F). Kretzoi (1942) considered the species as a new genus, Sharamynodon. Xu (1966) reported additional material of the species from the same horizon and locality. Wall (1981) also referred a partial skeleton (AMNH FM 21601) to the species from the Shara Murun Formation (gray beds), 4 mi north of Baron Sog Lamasery.
Sianodon Xu, 1965
Only one species of Sianodon, S. ulausuensis, has been reported from the Shara Murun Formation at Ula Usu (Xu, 1966). The species is known from a nearly complete skull (IVPP V 3215). Sianodon ulausuensis is more primitive than S. bahoensis in having a relatively longer facial part and three upper incisors. Xu (1966) also assigned a maxilla with M2–3 (IVPP V 3221) to Sianodon sp. from the same horizon and locality.
Lushiamynodon Chow and Xu, 1965
Only one species of Lushiamynodon, L. sharamurenensis, has been reported from the Shara Murun Formation at Ula Usu (Xu, 1966). The species is known from an incomplete juvenile cranium with DP1-M1 (IVPP V 2892), a juvenile mandible (IVPP V 2892.1), a posterior part of the cranium (IVPP V 2892.2), and some postcraninal material (Xu, 1966). A mandible with m2–3 (IVPP V 3217) of L. sharamurenensis also has been reported from the “same horizon” at “Ulan Shireh Obo,” which was situated about 30 km northeast of Ula Usu. The so-called “Ulan Shireh Obo” (= Ulan Shireh) is located in the North Mesa, north of Tukhum Sumu, with the deposits of Ulan Shireh Formation roughly equivalent to the Irdin Manha Formation (Chow and Rozhdestvensky, 1960; Wang et al., 2012).
Gigantamynodon Gromova, 1954
Only one species of Gigantamynodon, G. promisus, has been reported from the Shara Murun Formation at Ula Usu (Xu, 1966). The species is known from a mandible with p2-m3 (IVPP V 3218) and many postcranial specimens (IVPP V 3218.1-28). However, Wall (1989) considered Gigantamynodon a nomen dubium (Lucas and Emry, 1996; Lucas et al., 1996).
Cadurcodon Kretzoi, 1942
Two species of Cadurcodon, C. matthewi and C. houldjinensis, have been reported from the Erlian Basin (Wall, 1981; Wang et al., 2009). The type species of Cadurcodon matthewi (AMNH FM 26029) was found in the “Ulan Gochu beds” at Jhama Obo (Wall, 1981). Cadurcodon matthewi is considered to be more primitive than C. ardynensis (Wall, 1981). Cadurcodon houldjinensis, known mainly from isolated teeth (e.g., EMM 0126), was discovered from the Houldjin Formation at Houldjin and near the Erenhot Railway Station (Wang et al., 2009). Wang et al. (2009) considered that three specimens of “Cadurcotherium” sp. (AMNH FM 19183, field no. 36) from the Houldjin Formation, 5 mi south of Iren Dabasu (Matthew and Granger, 1923a) should be assigned to C. houldjinensis (m2) or C. ardynensis (P4 and M1/2).
Xu (1966) referred a maxilla with P2–3 (IVPP V 3222), a right m1 (IVPP V 3222.1), and a canine (IVPP V 3222.2) to Cadurcodon sp. from the Urtyn Obo Formation at Urtyn Obo (= Erden Obo). However, their specific stratigraphic horizons are uncertain.
Amynodontopsis Stock, 1933
Two species of Amynodontopsis, A. parvidens and A. tholos, were named by Wall (1981) from the Erlian Basin. The type of A. parvidens, known from a skull and mandible (AMNH FM 26043), was collected from the top of the “Lower White” at Erden Obo. Wall (1981) recognized two different sizes of the species: the small-size group includes AMNH FM 26045, 26046, and 26050; and the large-size group includes AMNH FM 26041, 26042, 26043, and 26051. He attributed the size differences to sexual dimorphism, because all the specimens were collected from the base of the “Middle Red” at Erden Obo, with the exception of AMNH FM 26046 that was recorded from the “Middle Red” bed and AMNH FM 26043 from top of the “Lower White” at Erden Obo. Wall (1981) also referred AMNH FM 26044 from the base of “Middle Red” at Erden Obo to A. parvidens. The species also is known from the following horizons and localities: AMNH 21599 from the base of upper red bed (Ulan Gochu Formation) 4 mi north of Baron Sog Lamasery; AMNH FM 26053 from the “Ulan Gochu beds” at Nom Khong Obo; AMNH 26178 possibly from the “Baron Sog bed” south of Jhama Obo; and AMNH FM 26038 from the “Ulan Gochu Formation” at Jhama Obo (Wall, 1981).
The holotype of Amynodontopsis tholos, comprising a skull and a lower jaw (AMNH FM 26035), was collected from the top of gray beds (= “Ulan Gochu”) at Ulan Shireh Obo (= Ganggan Obo on East Mesa) (Wall, 1981). The distribution of the species also includes the “Ulan Gochu beds,” south of Jhama Obo (AMNH FM 26031), and the “Lower White or Gray (pink)” at Nom Khong Obo (AMNH FM 26054, 26055, and 26057). Wall (1981) assigned a mandible (AMNH FM 26053, field no. 786) to A. tholos. However, it contradicts the fact that the specimen is also listed as A. parvidens, as discussed above. The specific stratigraphic horizon where AMNH FM 26053 was collected is uncertain, and is probably below the “Upper Red” bed.
Zaisanamynodon Belyaeva, 1971
Lucas et al. (1996) referred some CAE specimens to Z. borisovi from the Erlian Basin. The species is known from the “Middle White” (AMNH FM 26052, 26049) at Erden Obo, the Baron Sog Formation at Baron Sog Mesa (AMNH FM 21602), the “Ulan Gochu beds” at Ulan Shireh Obo (= Ganggan Obo) (AMNH FM 26034), and the “Houldjin gravels” at Camp Margetts. Lucas et al. (1996) further referred a symphysis with i2 and c (AMNH FM 26170) to Z. borisovi from the “Baron Sog Formation” at Erden Obo.
Rhinocerotidae Gray, 1821
Aprotodon Forster-Cooper, 1915
Only one species of Aprotodon, A. lanzhouensis, has been reported from the Houldjin Formation near the Erenhot Railway Station (Wang et al., 2009). The specimens consist of i2 (EMM 0082) and some cheek teeth (e.g., EMM 0079), and are characterized by extremely long and curved lower tusks (Wang et al., 2009).
Symphysorrhachis Beliajeva, 1954
Wang et al. (2009) referred a mandible with m2–3 (EMM 0123) and M1/2 (EMM 0078) to Symphysorrhachis sp. from the Houldjin Formation near the Erenhot Railway Station. Wang et al. (2009) further assigned an M3 (AMNH FM 19184) from Houldjin, originally referred to Caenopus or Praeaceratherium sp. by Matthew and Granger (1923a), to Symphysorrhachis sp.
Brontotherioidea Marsh, 1873
Brontotheriidae Marsh, 1873
Microtitan Granger and Gregory, 1943
Two species of Microtitan, M. mongoliensis and M.? elongatus, have been reported from the Ulan Shireh Formation at North Mesa and the Arshanto Formation at Daoteyin Obo, respectively (Granger and Gregory, 1943; Qi, 1987). Microtitan mongoliensis is known from a mandible with p2-m3 (AMNH FM 22099) and a maxilla with C-M3 (AMNH FM 21611). Microtitan mongoliensis initially was assigned to “Metarhinus?” by Osborn (1925) based a fragmentary mandible with p4-m1 (AMNH FM 20167), lacking diagnostic characters except for its small size. AMNH FM 20167 was discovered in the Irdin Manha Formation on the Irdin Manha escarpment (Osborn, 1925). Later, Granger and Gregory (1943) erected a new genus Microtitan for “Metarhi nus?” mongoliensis and selected AMNH FM 22099 as a neotype. Mihlbachler (2008) further pointed out that AMNH FM 20167 actually was made up of multiple individuals.
Microtitan? elongatus was erected by Qi (1987) based on P3-M3 (IVPP V 5767) from the Arshanto Formation at Daoteyin Obo. However, Mihlbachler (2008) considered the species a nomen dubium. Qi (1987) further referred a M2 (IVPP V 5768) to Microtitan sp. from the Arshanto Formation at Huheboerhe. It is noteworthy that, based on our fieldwork over the past decade, no brontotheres have been found in the Arshanto Formation in the Huheboerhe area. Consequently, the horizon bearing Microtitan? elongatus and Desmatotitan sp. (as discussed below) recorded as the Arshanto Formations is questionable.
Desmatotitan Granger and Gregory, 1943
Only one species of Desmatotitan, D. tukhumensis, has been reported from the Ulan Shireh Formation at Wulantaolegai (Granger and Gregory, 1943). The species is known only from the type AMNH FM 21606, which is composed of a mandible with i1-c and p2-m3. Desmatotitan tukhumensis is probably allied either with Metatelmatherium or Microtitan mongoliensis (Granger and Gregory, 1943; Mihlbachler, 2008).
Qi (1987) referred a single m3 (IVPP V 5769) to Desmatotitan sp. from the upper part of Arshanto Formation at Huheboerhe.
Acrotitan Ye, 1983
Ye (1983) erected a new genus and species Acrotitan ulanshirehensis from the Ulan Shireh Formation at Wulanhuxiu. The species is known from a fragmentary mandible with p3–4 (IVPP V 6686), and is characterized by two pairs of lower incisors (Mihlbachler, 2008).
Epimanteoceras Granger and Gregory, 1943
Only one species of Epimanteoceras, E. formosus, has been reported from the Ulan Shireh Formation at Wulantaolegai (Granger and Gregory, 1943; Mihlbachler, 2008). The type of E. formosus is known from a complete skull with complete dentition (AMNH FM 21613). Granger and Gregory (1943) erected a new genus and species Dolichorhinoides angustidens based on a partial skull with P1-M3 (AMNH FM 21607) from the Ulan Shireh Formation at Wulanhuxiu. However, Mihlbachler (2008) considered the species a junior synonym of E. formosus.
Hyotitan Granger and Gregory, 1943
Only one species of Hyotitan, H. thomsoni, was discovered in the Erlian Basin. The type of H. thomsoni, known from a mandible (AMNH FM 26401), was collected from the “Houldjin gravels” at Camp Margetts. However, Mihlbachler (2008) proposed that H. thomsoni could be a synonym of Qufutitian zhoui.
Metatelmatherium Granger and Gregory, 1938
Two species of Metatelmatherium, M. ultimum and M. parvum, have been reported from the Erlian Basin. Granger and Gregory (1943) erected a new species Metatelmatherium crista tum based on a skull and mandible (AMNH FM 26411) from the “Irdin Manha Formation” at Camp Margetts. Mihlbachler (2008) considered M. cristatum a junior synonym of the Uintan M. ultimum from North America, but Mader (1989, 1998) regarded them as two different species. Furthermore, Bai et al. (2017) pointed out that the horizon bearing AMNH FM 26411 is actually in the Irdin Manha rather than the Arshanto Formation. Metatelmatherium parvum, known from a fragmentary mandible with p3–4 (AMNH FM 20168), was found in the Irdin Manha Formation at Irdin Manha (Granger and Gregory, 1943). However, Mihlbachler (2008) considered the species a nomen dubium because it lacks diagnostic characters.
Protitan Granger and Gregory, 1943
The genus Protitan was erected by Granger and Gregory (1943) and originally included six species: P. grangeri, P. robustus, P. minor, P. bellus, P. obliquidens, and P.? cingulatus. Granger and Gregory (1943) further considered Dolichorhinus olseni and “Manteoceras?” irdinensis named by Osborn (1925) as junior synonyms of P. grangeri. However, Mihlbachler (2008) synonymized P. robustus, P. bellus, and P. obliquidens with P. grangeri, and considered P.? cingulatus a nomen dubium. Thus, according to the revision by Mihlbachler (2008), only two species of Protitan, P. grangeri and P. minor, are valid species.
The type of Protitan grangeri, known from a complete skull and mandible (AMNH FM 20103), is from the Irdin Manha Formation at Irdin Manha (fig. 2G). The species is also known from the following horizons and localities: Irdin Manha Formation at Irdin Manha (AMNH FM 19179, 20104 [holotype of P. “robustus], 20108, 20109 [holotype of “Dolichorhinus olseni”], 20111 [holotype of “Manteoceras? irdinensis”], 20112–20114, 20119, 20120, 20123, 20125 [holotype of P. “obliquidens”], and 20126); “Houldjin gravels” at Camp Margetts (AMNH FM 26421); and “lower red bed” at Spring Camp of East Mesa, Shara Murun region (AMNH FM 26104 [holotype of P. bellus]).
The type of Protitan minor, a skull (AMNH FM 26416), probably is from the top of “Irdin Manha” at 0.5 mi west of Camp Margetts. Mihlbachler (2008) also referred a partial skull (AMNH FM 26417) to P. minor from “?Irdin Manha beds,” 1 mi west of Camp Margetts. However, Bai et al. (2017) have pointed out that the horizon bearing AMNH FM 26416 is indeed within the Irdin Manha Formation.
Four mandibles (AMNH FM 26410, 26415, 26418, and 26408) initially assigned to P. minor or P. grangeri were considered to represent a new species (Camp Margetts “taxa A”) by Mihlbachler (2008). These specimens derive from the “Houldjin gravels” or “Irdin Manha beds” in the Camp Margetts area.
The type of “Protitan?” cingulatus, composed of a mandible with p1-m3 (AMNH FM 26412), was unearthed from the “Houldjin gravels” at Chaganboerhe. Mihlbachler (2008) also referred AMNH FM 26403 from the “Houldjin gravels” at Camp Margetts and AMNH FM 20110 from the Irdin Manha Formation at Irdin Manha to the species.
Qi et al. (1992) referred a single m3 and some postcrania (IVPP V 10104) to Protitan sp. from the so-called “Tukhum Formation” at Erden Obo. However, the horizon is equivalent to the “Low Red” bed.
Gnathotitan Granger and Gregory, 1943
Only one species of Gnathotitan, G. berkeyi, has been reported from the Erlian Basin (Granger and Gregory, 1943; Mihlbachler, 2008). Gnathotitan berkeyi was assigned initially to Telmatherium by Osborn (1925), and later Granger and Gregory (1943) erected the genus Gnathotitan for the species. The species is characterized by its disproportionately large and deep mandible (Mihlbachler, 2008). The type of G. berkeyi is known from a mandible with c-m3 (AMNH FM 20106) from the Irdin Manha Formation at Irdin Manha. Mihlbachler (2008) also referred the following specimens to G. berkeyi from the Irdin Manha Formation at Irdin Manha: AMNH FM 20107, 20115, 20121, and 141231 (formerly part of AMNH FM 20106). As noted by Mihlbachler (2008), two specimens (AMNH FM 20124, 20127) originally referred to G. berkeyi by Osborn (1925) were sent to the Chinese Geological Survey in October 1928 and are unfortunately lost. Both specimens were from the Irdin Manha Formation at Irdin Manha.
Rhinotitan Granger and Gregory, 1943
Two species of Rhinotitan, R. kaiseni and R. andrewsi, have been recognized from the Shara Murun Formation at Ula Usu (Granger and Gregory, 1943; Wang, 1982; Mihlbachler, 2008). Osborn (1923b, 1925) named three new brontothere species from Ula Usu: Protitanotherium mongoliensis, Protitanotherium andrewsi, and Dolichorhinus kaiseni, but Granger and Gregory (1943) later erected a new genus Rhinotitan for these three species. Wang (1982) further considered R. andrewsi as a junior synonym of R. mongoliensis (Takai, 1939). However, Mihlbachler (2008) treated R. mongoliensis as a nomen dubium.
The type of R. kaiseni, known from a skull and mandible (AMNH FM 20252), was unearthed from the Shara Murun Formation at Ula Usu. AMNH FM 20257 and FMNH P 14048 (formerly AMNH FM 20260) were referred to R. kaiseni from the same horizon and locality. The type of R. andrewsi, known from a skull (AMNH FM 20271), is from the Shara Murun Formation at Ula Usu. AMNH FM 20254, 20261, 20263, IVPP V 3254-1, 3254-2, PIN 2198-3, 2198-5, and 7130-3 have been assigned to R. andrewsi from the same horizon and locality as the holotype. Mihlbachler (2008) further referred many specimens to Rhinotitan sp. from the Shara Murun Formation at Ula Usu, 4 mi north of Baron Sog Lamasery.
Pachytitan Granger and Gregory, 1943
Only one species of Pachytitan, P. ajax (AMNH FM 21612), is known from the Shara Murun Formation (gray beds), 4 mi north of Baron Sog Lamasery in the Shara Murun region (Granger and Gregory, 1943; Mihlbachler, 2008). AMNH FM 21612 is a fragmentary skull with left I3, C, and P2-M3. Pachytitan ajax is very close to Rhinotitan, but is more advanced than the latter in having more molarized premolars that resemble those of Embolotherium or Parabrontops from the Erlian Basin (Granger and Gregory, 1943; Mihlbachler, 2008). Cranially, Pachytitan most resemble Rhinotitan andrewsi and Diplacodon elatus (Mihlbachler, 2008).
Titanodectes Granger and Gregory, 1943
Granger and Gregory (1943) erected a new genus, Titanodectes, comprising T. minor and T. ingens, from the Erlian Basin. However, Mihlbachler (2008) considered both of them syn onymous with Embolotherium grangeri. Considering that the horizons and localities bearing Titanodectes and Embolotherium are different, and that none of the reliable lower dentition of E. grangeri has been reported, we still regard Titanodectes as a valid genus. The type of T. minor, known from an incomplete mandible with front teeth and p2-m1 (AMNH FM 26132), is from the “Shara Murun Formation” at Spring Camp in the East Mesa. Granger and Gregory (1943) further referred AMNH FM 26021 from the “Lower White” of Erden Obo, AMNH FM 26012 from the “Ulan Gochu beds” at Twin Obos, and AMNH FM 21600 from the Shara Murun Formation (gray beds), 4 mi north of Baron Sog Lamasery to the species. However, Mihlbachler (2008) referred AMNH FM 26021 to Parabrontops cf. P. gobiensis. The type of Titanodectes ingens, known from a mandible (AMNH FM 26005), is from the “Ulan Gochu beds” at Jhama Obo. Titanodectes is considered to be in certain respects intermediate between Rhinotitan and Embolotherium (Granger and Gregory, 1943; Mihlbachler, 2008).
Embolotherium Osborn, 1929
Osborn (1929) erected a new genus, Embolotherium, which includes E. andrewsi, E. grangeri, and E. loucksii, from the Shara Murun region and Erden Obo. Embolotherium is characterized by its large, upward extended nasal horns (Osborn, 1929). Granger and Gregory (1943) further named a new species E. ultimum from the Baron Sog Mesa. However, Mihlbachler (2008) considered E. ultimum and E. loucksii as a junior synonyms of E. andrewsi and E. grangeri, respectively. Thus, only two species of Embolotherium are recognized as valid.
The type of Embolotherium grangeri, known from a complete skull (AMNH FM 26002), is from the base of “Middle Red” at Erden Obo (Osborn, 1929). Mihlbachler (2008) referred following specimens to E. andrewsi from the Erlian Basin: AMNH FM 26004 from the base of “Middle Red” layer at Erden Obo; AMNH FM 26040 from the base of “Ulan Gochu beds” at Twin Obos; AMNH FM 21610 (holotype of E. louksii) from the base of Ulan Gochu Formation, 4 mi north of Baron Sog Lamasery; and AMNH FM 26018 from the surface of a plain, 2 mi south of Baron Sog Lamasery.
The type of Embolotherium andrewsi, known from a nearly complete skull (AMNH FM 26001), is from the “Middle White” layer at Erden Obo (fig. 2H) (Osborn, 1929). Mihlbachler (2008) referred the following specimens to E. andrewsi from the Erlian Basin: those largely from the “Middle White” layer at Erden Obo (AMNH FM 26003, 26000 (presented to the Chinese Geological Survey Museum), 26011, 26006–26008, and IVPP V 11959); those from the “Ulan Gochu Formation” at Jhama Obo (AMNH FM 26009 and 26010); those from East Mesa (PIN 2200-1 and 2200-2); and those from the Baron Sog Formation at Baron Sog Mesa (AMNH FM 21604 [holotype of E. ultimum] and AMNH FM 22114). AMNH FM 20352 actually is from the Ardyn Obo Beds at Ardyn Obo (field no. 245), instead of the “Ulan Gochu Formation” at Erden Obo as mentioned by Mihlbachler (2008).
Parabrontops Granger and Gregory, 1943
Only one species of Parabrontops, P. gobiensis, has been reported from the Erlian Basin (Osborn, 1925; Granger and Gregory, 1943). The species P. gobiensis was initially assigned to Brontops by Osborn (1925). The type of the species, known from a distorted skull (AMNH FM 20354), is from the Ardyn Obo beds at Ardyn Obo (field no. 253) (Osborn, 1925; Granger and Gregory, 1943), rather than the Urtyn Obo Formation at Urtyn Obo as mentioned by Mihlbachler (2008). In detail, the locality 253 is situated “near trail, 1 mi northwest of Obo point.” AMNH FM 26020, known from a partial skull, was unequivocally referred to P. gobiensis (Granger and Gregory, 1943; Mihlbachler, 2008) and is from the “Middle White” layer at Erden Obo. Mihlbachler (2008) referred the following specimens to Parabrontops cf. P. gobiensis: AMNH FM 26019 (“Middle White” layer, Erden Obo); AMNH FM 26021 (“Lower Gray” layer, Erden Obo); and AMNH FM 26131 (“Shara Murun beds,” Spring Camp). However, Granger and Gregory (1943) assigned AMNH FM 26021 to Titanodectes minor as discussed above.
Metatitan Granger and Gregory, 1943
Granger and Gregory (1943) named three species within a new genus Metatitan: M. primus, M. progressus, and M. relictus. However, Mihlbachler (2008) erected a new genus, Nasamplus, for M. progressus, and thus only two species are included in Metatitan.
The type of Metatitan primus, known from a partial skull and mandible (AMNH FM 26101), was discovered at Big Red Draw, 1 mi southwest of Chimney Butte Camp at North Mesa. The horizon bearing AMNH FM 26101 is probably equivalent to the Ulan Shireh Formation at Wulanhuxiu. Mihlbachler (2008) further referred a mandible (AMNH FM 26102) to the species from the middle of the upper red beds at the northwest promontory, 1 mi northwest of Chimney Butte Camp.
The type of Metatitan relictus, known from a skull and mandible (AMNH FM 26391), is from the “Houldjin gravels,” 1 mi west of Camp Margetts. Metatitan is considered to be derived from Rhinotitan and to bridge the morphological gap between Embolotherium and most other horned brontotheres with the paired frontonasal horns (Granger and Gregory, 1943). The following specimens from the “Houldjin gravels” in the Camp Margetts area are referred to M. relictus (Granger and Gregory, 1943; Mihlbachler, 2008): AMNH FM 26395–26399; AMNH FM 26402; 26404–26407; 26420; 26427; and 26429. Mihlbachler (2008) further proposed that AMNH FM 26400, originally assigned to Metatitan relictus (Granger and Gregory, 1943), probably represents a new species (Camp Margetts “taxa B”). AMNH FM 26400 is from the “Houldjin gravels,” 1 mi west of Camp Margetts.
Wang et al. (2009) referred a maxilla with P1–3 (IVPP V 15714) and several isolated lower teeth to Metatitan sp. from the Houldjin Formation at Houldjin and near Erenhot Railway Station.
Nasamplus Mihlbachler, 2008
The genus Nasamplus was erected by Mihlbachler (2008) for “Metatitan” progressus (Granger and Gregory, 1943). The type species, known from a fragmentary skull (AMNH FM 26014), is from the “Ulan Gochu beds” at Jhama Obo.
Chalicotherioidea Gill, 1872
Litolophus Radinsky, 1964b
Only one species of Litolophus, L. gobiensis, is known from the basin (Colbert, 1934; Radinsky, 1964b; Bai et al., 2010). Colbert (1934) erected a new species “Grangeria” gobiensis, and later Radinsky (1964b) proposed the new genus Litolophus for the species. The type of L. gobiensis, known from a crushed skull and mandible (AMNH FM 26645), is from the base of the Arshanto Formation at Nuhetingboerhe. Other referred specimens were collected from the same horizon and locality. Recently, many cranial and postcranial specimens have been collected from the “chalicothere pit” (Bai et al., 2011a; Bai et al., 2011b) at the same horizon and locality as CAE Litolophus collections.
Schizotherium Gervais, 1876
Coombs (1978) referred AMNH FM 26061 and AMNH FM 103336 to Schizotherium cf. S. avitum from Erden Obo. The right maxilla with P2-M3 (AMNH FM 26061), in association with left maxilla with P3–4 and M2–3, is composed of two parts: an anterior part with P2–3 recorded from the “Ulan Gochu beds” (probably the “Upper Red”), and a posterior part with P4-M3 found from the “Middle White” (Coombs, 1978). This discrepancy may be the result of drift of weathered-out material down the steep slopes (Radinsky, 1964a; Coombs, 1978). A left mandible with p4-m2 (AMNH FM 103336) was found in the “Baron Sog bed” (= “Upper White”) at Erden Obo (Coombs, 1978). In addition, Colbert (1934) assigned a Mc III (AMNH FM 26188) to Schizotherium sp., and noted that the specimen was presumably from the “Baron Sog beds” at Nom Khong Shireh. Actually, AMNH FM 26188 is recorded as coming from “?Baron Sog beds” at Erden Obo (Wang, 2003).
DISCUSSION
Based on the data about the temporal and spatial distribution of Paleogene perissodactyls from the Erlian Basin (figs. 2–4), we address some long-standing problems about the biostratigraphic correlation of the Paleogene sediments in the Erlian Basin, and discuss perissodactyl diversity through the early Eocene to the early Oligocene in the following section.
“Houldjin Formation” in the Huheboerhe area
Granger and Berkey (1922) named the Houldjin beds on the Houldjin escarpment, which are dominated by yellow, pebbly gravels about 4.6 m (fig. 5A) (Berkey and Morris, 1927). The Houldjin fauna at the type locality includes a variety of perissodactyl taxa such as the brontotheriid Metatitan sp., paraceratheriid Aralotherium sp., amynodontids Cadurcodon ardyensis and C. houldjinensis, rhinocerotids Aprotodon lanzhouensis and Symphysorrhachis? sp. (Matthew and Granger, 1923a; Qiu and Wang, 2007; Wang et al., 2009). The age of the Houldjin fauna remains controversial and varies in age estimates from the late Eocene to the late Oligocene (Qiu and Wang, 2007; Wang et al., 2009). This discrepancy is attributed to the fragmentary material known from the formation and probably a mixture of taxa from different horizons (Wang et al., 2012).
Meng et al. (2007) pointed out that the so-called “Houldjin Formation” and “Irdin Manha Formation” by CAE in the Huheboerhe area are mostly Irdin Manha and Arshanto formations, respectively (Wang et al., 2010). Though Wang et al. (2012) noticed the “yellow gravels” present in Huheboerhe area, it is uncertain whether they belong to the Houldjin Formation. However, some perissodactyls from the “Houldjin gravels” in the Huheboerhe area obviously are more advanced than those from type Irdin Manha Formation, including the deperetellid Deperetella sp., amynodontid Zaisanamynodon borisovi, paraceratheriids Urtinotherium intermedium and Paraceratherium grangeri, and possibly the brontothere Metatitan. It is notable that except for Deperetella sp. and Metatitan, the other species are recorded either by an isolated tooth or postcrania, hampering the comparisons with Houldjin fauna from the type locality. Qiu and Wang (2007, p169) also mentioned a maxilla with P3-M2 of Allacerops turgaica from the “Houldjin gravels” at Daoteyin Obo. These advanced perissodactyls are indicative of a late Eocene Ergilian and/or early Oligocene Hsandagolian age. Whether those taxa recovered from the “yellow gravels” warrant further investigation, strata younger than the Irdin Manha Formation likely are present in the Huheboerhe area (Qiu and Wang, 2007). As noted by Meng et al. (2007: 18), the “Houldjin gravels may be represented only by a thin layer of loose sand and conglomerates forming the Gobi surface” in the Huheboerhe area.
Ulan Shireh Formation at the North Mesa
In 1925, the CAE investigated Wulantaolegai (4 mi north of Tukhum Lamasery) and Wulanhuxiu (8 mi north of Tukhum Lamasery) at North Mesa for the first time (fig. 1), and all collections were recorded from the “Shara Murun beds” (Granger, 1925). Granger (1925: 66) further noted that all fossils were from layers below an “apple red stratum.” During the second investigation of the North Mesa, Granger (1928: 6) divided the sediments at “Chimney Butte” (Wulanhuxiu) into 14 layers from the top to the bottom (fig. 6), with layers 3–5 grouped into an upper “red member” and layers 6–14 grouped into a lower “white member” (fig. 5B). Almost all specimens were collected from the “Buckshot's Quarry” at Chimney Butte, which is equivalent to layer 9 of Granger's (1928) sketch and roughly layer 4 of Li et al. (2016a). The exception is that of the creodont Sarkastodon mongoliensis (AMNH FM 26302, field no. 641) from the relatively lower layer 12 (Granger, 1938). A few specimens, such as AMNH FM 26115 of Breviodon sp. and AMNH FR 6697 of fossil turtle Anosteira mongoliensis (Gilmore, 1931), not otherwise recorded from the “Buckshot's Quarry” in the Chimney Butte section were labeled with the field number 645. In addition, a few specimens of the brontothere Metatitan primus (AMNH FM 26101, 26102) and turtles were collected from Big Red Draw and Northwest Promontory, which are located 1 mi southwest and northwest of the Chimney Butte Camp, respectively. AMNH FM 26102 is from the middle of upper red beds (Granger, 1928). Thus, almost all the previous CAE collections from North Mesa probably were unearthed from the lower “white member” and below the “red member,” with an exception of Metatitan primus (Wang et al., 2012). Similarly, Chow and Rozhdestvensky (1960) noted that most of the fossils from North Mesa (= Ulan Shireh) collected by SSPE were from the lower part of the formation.
Berkey et al. (1929) named the Ulan Shireh Formation for the deposits at North Mesa, which are dominated by red clays, though gray clays, sands, and gravels also are present (fig. 5B). The Ulan Shireh Formation is considered equivalent to the Tukhum Formation (Berkey et al., 1929), and the mammalian fauna from the Ulan Shireh Formation is considered to be homogenous and roughly correlated to the Irdinmanhan (Radinsky, 1964a; Ye, 1983; Wang et al., 2012). However, recent investigation suggests that the upper part of the section probably belongs to the Sharamurunian (Li et al., 2016a), and the basal part may extend down into the Arshantan (Mao and Wang, 2012; Bai et al., 2018a).
FIG. 2.
Cranial and mandibular reconstructions of typical Eocene perissodactyls from the Erlian Basin of Inner Mongolia, China. tapiroid A, Lophialetes expeditus; B, Paracolodon fissus; paraceratheriid C, Pappaceras meiomenus; D, Juxia sharamurenensis; amynodontid E, Rostriamynodon grangeri; F, Sharamynodon mongoliensis; brontotheriid G, Protitan grangeri; H, Embolotherium andrewsi. Scale bar equals 10 cm.
FIG. 3.
Stratigraphic distributions of perissodactyl fossils and taxa in the Nomogen, Arshanto, and Irdin Manha formations on the Irdin Manha escarpment and in the Huheboerhe area. The profile of section and fossil-bearing horizons are modified from Meng et al. (2007) and Wang et al. (2010). The superscripts in front of taxa indicate the localities of holotypes: numbers 1, 2, and 3 refer to Irdin Manha escarpment, Camp Margetts and nearby localities, and Huheboerhe escarpment, respectively. Abbreviations: A., Amynodontidae; Ch., Chalicotherioidea; Depere., Deperetellidae; H., Hyracodontidae; Helalet., Helaletidae; Hy., Hyrachyus; IM, Irdin Manha Formation; and Para., Paraceratheriidae.
FIG. 4.
Stratigraphic distributions of perissodactyl fossils and taxa in the Tukhum, Shara Murun, Ulan Gochu, and Baron Sog formations at Baron Sog Mesa. The profile of the section is based on the sketch at Xilin Nor North (= 4 mi north of Baron Sog Mesa) (Granger, 1925, see also Wang et al., 2012, fig. 2A). The superscripts in front of taxon names indicate the localities of the holotypes: numbers 1, and 2 refer to Ula Usu and 4 mi north of Baron Sog Mesa, respectively. Abbreviations: BS, Baron Sog Formation; Hyra., Hyracodontidae; P., Paraceratheriidae; T, Tukhum Formation; and UG, Ulan Gochu Formation.
Assuming that most previous collections from Wulanhuxiu and Wulantaolegai are from a relatively lower horizon (roughly equivalent to the “white member”), the lower part of the Ulan Shireh Formation can be correlated with the Irdin Manha Formation based on their shared species (fig. 3; appendix 1) including lophialetids Lophialetes expeditus, Zhongjianoletes chowi, and Rhodopagus pygmaeus, the deperetellid Teleolophus medius, the hyracodontid Triplopus proficiens, the paraceratheriid Fostercooperia ulanshirhensis, and the brontotheriid Microtitan mongoliensis. They also share the lophialetid genera Simplaletes and Breviodon. The brontothere Metatitan primus has been reported from the Big Red Draw and North Promontory in North Mesa, and similarly, many specimens Metatitan relictus have been found in the Irdin Manha Formation near Camp Margetts. However, the Irdin Manha Formation also bears the paraceratheriid Forstercooperia totadentata, helaletids, and some brontotheres that have not been reported yet or are different from those of the Ulan Shireh Formation. Teleolophus medius? also has been reported from the middle horizon of upper part of the Ulan Shireh Formation (“red member”) at Wulanhuxiu (Bai et al., 2018a).
Comments on the Shara Murun and Ulan Gochu Formations
The Shara Murun Formation at Baron Sog Mesa was mentioned first by Berkey and Granger (1923) without description. Berkey and Morris (1927) formally named the Shara Murun Formation as consisting of two members: an upper member dominated by white and light-gray sandstone with gravels; and a lower member dominated by sandy clay with varied colors (Berkey and Morris, 1927; Wang et al., 2012) (fig. 4, 7A; see also Wang et al., 2012: fig. 2A). Later, the “Shara Murun Formation” also was applied to some strata at East Mesa, Erden Obo, and Nom Khong Obo (Osborn, 1929; Qi, 1990b). However, these correlations remain controversial (Wang et al., 2012). For the sake of clarity, we listed perissodactyls from the Shara Murun Formation exclusively at Baron Sog Mesa (fig. 4).
Though the mammalian fauna from the Shara Murun Formation is considered to be homogeneous, species may be recovered from different members and represent different evolutionary stages. This hypothesis is supported by the fact that the brontotheres Titanodectes minor and Pachytitan ajax are considered more advanced than Rhinotitan, resembling Embolotherium to some extent (Granger and Gregory, 1943; Mihlbachler, 2008), and their co-occurrence in a single horizon seems improbable. Rhinotitan, Juxia sharamurenensis, and various amynodontids collected by SSPE were found in the lower part of the Shara Murun Formation near the Well of Ula Usu (Chow and Rozhdestvensky, 1960). On the other hand, both Titanodectes minor and Pachytitan ajax were recorded from the “gray beds” of the Shara Murun Formation, 4 mi north of Baron Sog Mesa, and that horizon is more likely to be in the upper, gray and white, member rather than a lower, red and brown, member (Berkey et al., 1929: fig. 4; Wang et al., 2012: fig. 2A). Thus, it is reasonable to infer that the mammalian faunas from the lower and upper parts of the Shara Murun Formation are probably heterogeneous, and they should be treated with caution.
The Ulan Gochu Formation was named by Berkey et al. (1929) as a wedge-shaped mass of red clay along the Baron Sog Mesa (fig. 7B). The Ulan Gochu Formation overlies the Shara Murun Formation, and is overlain by the Baron Sog Formation. The “Ulan Gochu Formation” later was expanded to include a much wider extension of the strata at Erden Obo and Nom Khong Obo (Osborn, 1929), and its boundaries with the underlying “Shara Murun Formation” are uncertain in these areas. The previous mammalian fauna from the “Ulan Gochu Formation” is mainly from East Mesa (Twin Obos and Jhama Obo) and/or Erden Obo (Li and Ting, 1983; Russell and Zhai, 1987; Tsubamoto et al., 2004). For the sake of clarity, we consider the perissodactyls of Ulan Gochu Formation exclusively from Baron Sog Mesa where the formation was named. Only Embolotherium grangeri and Amynodontopsis parvidens have been reported from typical Ulan Gochu Formation outcrops (Osborn, 1929; Granger and Gregory, 1943; Mihlbachler, 2008). Interestingly, Titanodectes minor from the upper part of the Shara Murun Formation even was considered to be a synonym of Embolotherium grangeri (Mihlbachler, 2008), indicating that the perissodactyls from the upper part of the Shara Murun Formation are more similar to those from the overlying Ulan Gochu Formation than to those from the lower part of the Shara Murun Formation.
FIG. 5.
Paleogene outcrops on the Houldjin escarpment and Wulanhuxiu (= Chimney Butte, 8 mi north of the Tukhum Lamasery): A, the yellow and gray, pebbly gravels of the Houldjin Formation on the Houldjin escarpment near the Erenhot Railway Station; B, the outcrops at Wulanhuxiu, showing a steep, upper “red member” and a gradual, lower “white member.”
FIG. 6.
Sketch profile from Wulanhuxiu (= Chimney Butte, 8 mi north of Tukhum Lamasery) at North Mesa (Granger, 1928: 6). The layers 3–5 are grouped into an upper “red member”, and the layers 6–14 are grouped into a lower “white member.”
FIG. 7.
Paleogene outcrops in the Shara Murun region: A, the outcrops at Ula Usu, where the Shara Murun Formation was named, showing an upper member dominated by white sandstones and a lower member dominated by sandy clays with varied colors; B, the outcrops at Xilin Nor North (= 4 mi north of Baron Sog Lamasery), showing the red clays of the Ulan Gochu Formation overlying the grayish white sandstones of the upper part of the Shara Murun Formation.
Correlation of the Erden Obo section
Osborn (1929) first reported the Erden Obo (= Urtyn Obo) section based on Granger's (1928) sketch, and divided the section into eight units from the bottom to the top: “Basal Red,” “Basal White,” “Lower Red,” “Lower White” (19.8 m), “Middle Red” (21.3 m), “Middle White or Gray” (9.1 m), “Upper Red” (29.0 m), and “Upper White” (7.6 m) (figs. 8, 9). Osborn (1929) correlated the “Basal Red” to the “?Arshanto Formation,” the “Basal White” to the “Shara Murun Formation,” the upper part of the “Lower White” to the “Upper Red” collectively to the “Ulan Gochu Formation,” and the “Upper White” to the “Baron Sog Formation.” The “Lower Red” and the lower part of the “Lower White” were considered either the “Shara Murun” and/or “Ulan Gochu” formations. Different correlations of the Erden Obo section in the Erlian Basin have been proposed since Osborn (1929), and were reviewed in detail by Wang et al. (2012). Among them, Jiang (1983) measured and divided the section of Erden Obo into 23 layers, and proposed several new lithological units for the sediments. However, Qi (1990b) applied conventional formational names to the deposits based on the same division of layers of Jiang (1983). Wang et al. (2012) suggested that the “Basal Red” should be Nomogen Formation and the “Upper White” should be the Upper Naogangdai Formation, and the correlation of other layers at Erden Obo remains unclear. Various perissodactyl fossils have been reported from different units at Erden Obo with the exception of the “Basal Red,” and they can contribute to a refinement of the biostratigraphic correlation with other type localities in the Erlian Basin.
The “Basal Red” layer is equivalent to layers 1–4 of Jiang's section, and is dominated by light brownish-red sandy mudstones (Jiang, 1983; Qi, 1990b). The “Basal White” layer is equivalent to layers 5–6 of Jiang's section, and is dominated by grayish white, light-green coarse sandstones and grayish white siltstones and fine sandstones (Jiang, 1983; Qi, 1990b). Recently, on the basis of rodent fossils, Li et al. (2018) correlated upper part of the “Basal Red” to the Nomogen Formation of Bumbanian age, and they further correlated the lower part and middle part of “Basal White” to the upper part of the Arshanto Formation and lower part of the Irdin Manha Formation, respectively. Similarly, Mao and Wang (2012) mentioned the presence of the Arshantan coryphodontid Eudinoceras mongoliensis from the base of “Basal White.” Among the perissodactyls, both Schlosseria magister and Lophialetes expeditus? have been reported from the “Basal White” (fig. 10) (Radinsky, 1965; Li and Wang, 2010), supporting the correlation of the bed with both the Arshanto and Irdin Manha formations as suggested by Li et al. (2017). However, whether the two species are from different horizons is unknown. In recent fieldwork, we collected a partial skull of a relatively primitive deperetellid, which is similar to those from the Arshanto Formation, from the lower part of “Basal White,” in addition to specimens of cranial material of the Irdinmanhan Teleolophus medius from the middle part of “Basal White.”
The “Lower Red” layer is usually considered to be equivalent to layer 7 of Jiang's section, which is dominated by light brick-red sandy clays (Jiang, 1983; Qi, 1990b). However, we consider the “Lower Red” layer also to include the overlying layer 8 of Jiang's section, which is dominated by mudstones with variable colors (see discussion below) (Jiang, 1983; Qi, 1990b). Qi et al. (1992) referred a single m3 and some postcrania to Protitan sp. from so-called “Tukhum Formation” (layer 7 of Jiang's section), which is equivalent to a part of the “Lower Red” bed (Wang et al., 2012) (fig. 10). However, Qi et al. (1992) noted that the m3 also is similar to that of Rhinotitan. Our recent fieldwork uncovered Caenolophus promissus and Teilhardia pretiosa from the “Lower Red” layer. The type specimen of C. promissus is known from the Shara Murun Formation, but that of T. pretiosa is recorded from the underlying Tukhum Formation at Ula Usu (Matthew and Granger, 1925b; 1926) (fig. 4). However, the horizon bearing Teilhardia pretiosa is controversial (Wang et al., 2012). In addition to perissodactyl fossils, the “Lower Red” bed also bears the lagomorph Gobiolagus aliwusuensis (Fostowicz-Frelik et al., 2012), and the rodents Gobiomys exiguus, Gobiomys neimongolensis, and Yuomys magnus (Li, 2017). Gobiolagus and the type specimens of Gobiomys neimongolensis are considered to be Sharamurnian in age (Meng et al., 2005; Fostowicz-Frelik et al., 2012; Li, 2017). As a result, the “Lower Red” bed is most likely correlative to the lower member of the Shara Murun Formation (Qiu and Wang, 2007).
According to Granger's (1928) sketch, the “Lower Gray” is composed mainly of “white sands” with a “conglomerate layer” in the upper part and a “concretionary layer” at the top (figs. 8, 9A) (Wang et al., 2012: fig. 2B). The base of the “Lower White” first was considered to be equivalent to layer 10 of Jiang's section and correlated to the upper part of the Shara Murun Formation (Wang, 2001). Wang (2007a) later considered the “Lower White” to be equivalent to layers 8–9 and 10–13 of Jiang's section, and grouped the “Lower White” into the “Ulan Gochu Formation,” spanning from the “Lower White” to “Upper Red” (Qiu and Wang, 2007; Wang et al., 2012). Although Wang (2007a) noted that layers 8–9 of Jiang's section are same as layers 10–13 in lithology (grayish white sandstone), the layer 9 is mainly dominated by the grayish green mudstones with two lens of feldspathic quartz coarse sandstones (fig. 9A) (Jiang, 1983). Based on our recent fieldwork, we regard the “Lower White” as likely equivalent to layers 9–10 of Jiang's section (Jiang, 1983; Qi, 1990b).
FIG. 9.
Paleogene outcrops at Erden Obo: A, the outcrops of the “Lower Red,” “Lower White,” and “Middle Red.” The lower-right part of the photo is mainly the “Lower White,” which overlies the “Lower Red” and is overlain by the “Middle Red” as shown by the hill, far left; B, the outcrops of the “Middle Red,” “Middle White,” “Upper Red,” and “Upper White.” The lower-right part of the photo is mainly the basal part of the “Upper Red,” which overlies the “Middle White” and is overlain by the “Upper White (or yellow).” The top of the “Upper White” forms the Gobi surface.
FIG. 10.
Stratigraphic distributions of perissodactyl fossils and taxa from Erden Obo. The profile of section is based on the sketch by Granger (1928) and modified from Li (2017). The asterisks in front of taxa indicate where the holotype occurred in the section. Abbreviations: Amyno., Amynodontidae; BR, Basal Red; BW, Basal White; Chali., Chalicotherioidea; LR, Lower Red; MW, Middle White; and UW, Upper White.
The perissodactyl fossils from the “Lower White” include five genera and six species (fig. 10). Juxia sharamurenensis (AMNH FM 26753, field no. 734) is recorded from the base of the “Lower White” (Radinsky, 1967), and Amynodontopsis parvidens is known from the top of “Lower White,” and by abundant material from the base of overlying “Middle Red” bed. Urtinotherium parvum is known from the “Lower White” with IVPP field number 199102, which is the same horizon as IVPP field number 199104 bearing abundant small mammals according to B.Y. Wang's field notes. The horizon bearing Ulania wilsoni was reported as the base of “Ulan Gochu Formation” with a grayish green sandstone (Qi, 1990a), most likely is the “Lower White” (Qi, 1990b). The horizon producing the type of Juxia shoui was reported as the “Ulan Gochu Formation” (Qi and Zhou, 1989). However, Qiu and Wang (2007) suggested that layer may be the “Lower White” based on a lunar found in that layer, which presumably belong to J. shoui. Titanodectes minor (AMNH FM 26061, field no. 769) also was reported from the base of the “Lower White” and probably unearthed from the base according to Osborn's sketch (Osborn, 1929: fig. 2). Thus, the “Lower White” has produced Amynodontopsis parvidens, Urtinotherium parvum, Juxia shoui, J. sharamurenensis?, Ulania, and Titanodectes minor. It is necessary to note that whether the mandible with p2, p4-m3 (AMNH FM 26753) should be assigned to Juxia sharamurenensis remains controversial, because J. shoui, which is preserved only as a partial skull with a similar size, is also present in the same layer (Qi and Zhou, 1989). Considering that Urtinotherium parvum and Juxia shoui are more derived than J. sharamurenensis from the lower member of the Shara Murun Formation, and Titanodectes minor also is known from the upper member of the Shara Murun Formation, we believe that the “Lower White” can be correlated with the upper member of the Shara Murun, as suggested by Wang (2001).
In addition to perissodactyl fossils, various small mammalian fossils have been reported from the “Lower White,” including the rodents Gobiomys exiguus, G. asiaticus (Wang, 2001), Pappocricetodon antiquus, P. sp., and P.? sp. (Wang, 2007b), the lagomorph Desmatolagus vetustus (Wang, 2007a), the primate Eosimias sp. (Wang, 2008b), the insectivores Anatolechinos neimongolensis and Erinaceidae gen. et sp. indet., and an indeterminate chiropteran (Microchiroptera) (Wang, 2008a). All the small mammalian fossils were collected from a single pocket (IVPP Loc. 1991004), dominated by grayish-white sandstone with conglomerates. The horizon is considered to be equivalent to layer 10 of Jiang's section (Jiang, 1983; Wang, 2001), belonging to the “Lower White.”
The “Middle Red” layer has been considered equivalent to layers 14–15 of Jiang's (1983) section (Qiu and Wang, 2007; Wang et al., 2012). However, the “Middle Red” layer is more likely equivalent to layers 11–15 of Jiang's section, which is dominated by light brick-red silty mudstones and deep brownish red mudstones with grayish-white fine sandstone and siltstone at the base (fig. 9A) (Jiang, 1983; Qi, 1990b). The perissodactyl fossils from the “Middle Red” are relatively rare, and all were discovered from the base of the unit. These taxa include Teleolophus magnum, Amynodontopsis parvidens, and Embolotherium grangeri (Osborn, 1929; Radinsky, 1965; Wall, 1981; Mihlbachler, 2008) (fig. 10). The latter two species also are known from the base of the Ulan Gochu Formation at Baron Sog Mesa, and considering Embolotherium loucksii as a junior synonym of E. grangeri (Mihlbachler, 2008) (fig. 4). Thus, the “Middle Red” is likely correlative to the Ulan Gochu Formation at Baron Sog Mesa (Qi, 1990b; Qiu and Wang, 2007). The rodents Gobiomys exiguus and G. neimongolensis have been reported from “Middle Red” (Li, 2017). However, only the cylindrodont Ardynomys olsoni has been reported from the Ulan Gochu Formation, 4 mi north of Baron Sog Mesa (Wang and Meng, 2009). It is noteworthy that the perissodactyls from the “Lower White” are more similar to those from the overlying “Middle Red” in sharing Amynodontopsis parvidens and similar brontothere taxa than to those from underlying “Lower Red.”
The “Middle White” layer is equivalent to layers 16–17 of Jiang's section (Qiu and Wang, 2007; Wang et al., 2012), which is dominated by grayish white sandstones with gravels (fig. 9B) (Jiang, 1983; Qi, 1990b). The “Middle White” has a relatively diverse set of perissodactyl fossils (fig. 10), including Ardynia praecox (Radinsky, 1967; Bai et al., 2018b), Proeggysodon qiui (Bai and Wang, 2012), Urtinotherium intermedium (Chow and Qiu, 1963; Qiu and Wang, 2007), Zaisanamynodon borisovi (Lucas et al., 1996), Embolotherium andrewsi, and Parabrontops gobiensis (Osborn, 1925; Osborn, 1929; Granger and Gregory, 1943; Mihlbachler, 2008). Zaisanamynodon borisovi and E. andrewsi (= E. ultimum) also are known from the Barson Sog Formation at Baron Sog Mesa (Granger and Gregory, 1943; Lucas et al., 1996; Mihlbachler, 2008) (fig. 4). In addition, the artiodactyl Lophiomeryx anagare was recorded from both the “Middle White” bed (AMNH FM 26259, field no. 752) and the Baron Sog Formation at Baron Sog Mesa (AMNH FM 22113, field no. 593) (Wang, 2003). Thus, the “Middle White” can be correlated with the Baron Sog Formation, partly in agreement with Qi (1990b)'s proposal. Recently, Li (2017) reported the presence of the rodent Gobiomys neimogolensis in the “Middle White.”
The “Upper Red” layer is equivalent to layers 18–20 of Jiang's (1983) section, which is dominated by variegated mudstones, brick-red sandy clays with calcareous nodules, and deep brownish red, bright red mudstones (fig. 9B) (Jiang, 1983; Qi, 1990b). Jiang (1983) named the Lower Naogangdai Formation for both the “Middle White” and “Upper Red” layers. Only the chalicothere Schizotherium cf. S. avitum is known from “Upper Red” (Coombs, 1978) (fig. 10). In contrast, rodent fossils are relatively much more diverse, including cricetids (Eucricetodon wangae, E. sp., and Pappocricetodon siziwangqiensis), dipodids (Heosminthus nomogenesis, Sinosminthus sp., Allosminthus cf. A. majusculus, Allosminthus ernos, and A. cf. A. diconjugatus), and ctenodactyloids (Gobiomys exiguus, and G. neimongolensis) mainly from the lower and middle parts of the “Upper Red” beds (Li et al., 2016a, 2016b, 2016c; 2017). Li et al. (2016b, 2016c) suggested that the “Upper Red” is the late Eocene and correlative with the Ergilian.
The “Upper White” layer is equivalent to layers 21–23 of Jiang's section, which is dominated by grayish white, yellow coarse sandstones, liver-colored mudstone, and loose yellow sandy gravels (fig. 9B) (Jiang, 1983; Qi, 1990b). Jiang (1983) named the Upper Naogangdai Formation for the “Upper White” layer. Considering that the “Upper White” bed at Erden Obo can be correlated to that at Nom Khong (Wang, 2003), the perissodactyl fossils from the “Upper White” of both localities include Ardynia kazachstanensis, Paraceratherium grangeri, Zaisanamynodon borisovi, and Schizotherium cf. S. avitum (Coombs, 1978; Lucas et al., 1996; Wang, 2003) (fig. 10). In addition, tsaganomyid rodents Tsaganomys altaicus and Cyclomylus intermedius have been reported from the bed (Wang, 2003). The “Upper White” is considered to be late early Oligocene in age and different from the “Baron Sog Formation” (Wang, 2003), which is correlative to the underlying “Middle White.” The late early Oligocene age for the “Upper White” is further supported by the occurrence of Ardynia kazachstanensis, the type specimen of which was unearthed from the late early Oligocene deposits of Chelkar Teniz, Kazakhstan (Russell and Zhai, 1987; Qiu et al., 2004; Bai et al., 2018b). Although Zaisanamynodon borisovi is indicative of the late Eocene Ergilian, material of the species from the “Upper White” is represented only by a symphysis with i2 and canine (AMNH FM 26170) (Lucas et al., 1996), which is too incomplete to allow a confident taxonomic assignment.
Diversity of Perissodactyls during Paleogene ALMAs
In the Gashantan, only some fragmentary teeth have been assigned to Perissodactyla without further determination (Meng et al., 1998). However, Meng et al. (1998) noted their similarities with the brontothere Lambdotherium. However, Tong et al. (2004) considered these materials to be similar to the condylarth Olbitherium.
In the early Eocene Bumbanian (Bowen et al., 2005; Ting et al., 2011), the lophialetid Minchenoletes erlianensis and presumably the hyracodontid Pataecops parvus are known (Wang et al., 2011) (fig. 11). However, Pataecops initially was included in the Lophialetidae (Radinsky, 1965).
Perissodactyls from the middle Eocene Arshantan are dominated by members of Tapiroidea (figs. 3, 11), including lophialetids (Schlosseria magister and Breviodon cf. B. acares), deperetellids (Teleolophus primaries and T.? rectus), and the helaletid Heptodon minimus. The basal rhinocerotoid Hyrachyus and paraceratheriid Pappaceras also appeared in the Arshantan, but chalicotheres are represented only by Litolophus gobiensis from the base of the Arshanto Formation (Colbert, 1934; Bai et al., 2010; 2011a). Brontotheres are scarce with only some fragmentary material of Microtitan and Desmatotitan known (Qi, 1987). However, there is question as to whether those brontotheres were recovered in the Arshanto Formation. Based on our recent fieldwork, some new taxa of ceratomorphs have been found in the Arshanto Formation, indicating that the current diversity of perissodactyls in the Arshantan is probably underestimated. The Arshantan is usually correlated to the Bridgerian NALMA (Tong et al., 1995; Wang et al., 2007; Sun et al., 2009). Furthermore, the fauna from the upper part of Arshanto Formation exhibits some transitional features. Schlosseria magister and the coryphodontid Eudinoceras represent handover taxa from earlier strata, but the appearances of Pappaceras and Simplicimys bellus are indicative of a close relationship with that of the overlying Irdin Manha Formation (Li and Wang, 2010; Mao and Wang, 2012; Li and Meng, 2015; Wang et al., 2016).
The correlation of the Ulan Shireh Formation is still debatable, although we propose correlating it with the Irdin Manhan Formation. The perissodactyls from the Irdinmanhan are restricted to those from type Irdin Manha escarpment and Huheboerhe area. In the Irdinmanhan, perissodactyls reached their highest diversity and were dominated by brontotheres (figs. 3, 11), including six genera and seven species: Microtitan mongoliensis, Hyotitan thomsoni, Metatelmatherium ultimum, M. parvum, Protitan grangeri, P. minor, P.? cingulatus, Gnathotitan berkeyi, and Metatitan relictus. It is interesting to note that a large number of specimens of Protitan grangeri are known from Irdin Manha escarpment, and the contemporaneous Metatitan relictus is common at Duheminboerhe. Tapiroids from the Irdinmanhan become more diverse than those from the Arshantan, and they include seven genera and species: lophialetids (Lophialetes expeditus, Simplaletes sujiensis, Breviodon minutus, and ?Rhodopagus pygmaeus), the deperetellid Teleolophus medius, and helaletids (Desmatotherium mongoliense and Paracolodon fissus). Paraceratheriids are represented by Forstercooperia totadentata and F. ulanshirhensis. The hyracodontid Triplopus and amynodontid Rostriamynodon emerged during the Irdinmanhan. Bai et al. (2017) correlated the Irdinmanhan roughly with Ui1 and Ui2 subages of the Uintan NALMA based solely on perissodactyl fossils. Similarly, based on the small mam mals from the lower fossil-bearing horizon of the Hetaoyuan Formation in the Liguanqiang Basin, Tong (1997) suggested that Irdinmanhan can be correlated to the early Uintan NALMA.
FIG. 11.
The diversity of perissodactyl fossils in the Erlian Basin from the early Eocene to the early Oligocene with superimposed reconstructed temperature curves, and revised correlation between the Eocene Asian Land Mammal Ages (ALMA) and North American Land Mammal Ages (NALMA). The left column showing magnetic Polarity Chrons and NALMAs modified from Vandenberghe et al. (2012). The average global temperature curve was derived from stable oxygen isotopes and is modified from Zachos et al. (2001). The gray, blue, green, and orange areas represent Tapiroidea, Rhinocerotoidea, Brontotheriidae, and Chalicotherioidea, respectively. Abbreviations: Ck., Clarkforkian; Du., Duchesnean; EECO, Early Eocene Climatic Optimum; Ga., Gashatan; Hs., Hsandagolian; Or., Orellan; Pa., Paleocene; PETM, Paleocene-Eocene Thermal Maximum; Shara., Sharamurunian; Ta., Tabenbulakian; Th., Thanetian; Ul., Ulangochuian; and Wt., Whitneyan.
The Sharamurunian and Ulangochuan are complex issues to address. As discussed above, the perissodactyls from the upper member of the Shara Murun Formation and “Lower White” of the Erden Obo are more similar to those from the overlying Ulan Gochu Formation and “Middle Red,” respectively, than to those from their underlying strata. However, previously the Sharamurunian probably was based largely on the entire fauna from both the lower and upper members of the Shara Murun Formation, although most fossils likely were found in the lower member (Russell and Zhai, 1987). On the other hand, previously the Ulangochuan mainly was based on the fauna from East Mesa and Erden Obo rather than those from the type Baron Sog Mesa (Russell and Zhai, 1987), and was a mixture of faunas from various ages and horizons. The following discussion on the Sharamurunian and Ulangochuan is based on the perissodactyls from Baron Sog Mesa and Erden Obo. However, the material from Twin Obos, Jhama Obo and nearby localities, and Nom Khong Obo were excluded, pending further investigation and correlation.
Perissodactyls from the lower member of the original Shara Murun Formation and “Lower Red” of the Erden Obo (both may represent restricted Sharamurunian) are dominated by amynodontids (figs. 4, 10, 11), comprising five genera and six species: Sharamynodon mongoliensis, Sianodon ulausuensis, Lushiamynodon sharamurenensis, Gigantamynodon promisus, Caenolophus promissus, and C. obliquus (Matthew and Granger, 1925b; Osborn, 1936; Xu, 1966). However, Wall (1989) suggested that Lushiamynodon and some species of Sianodon should be placed in Sharamynodon, and Gigantamynodon likely is a nomen dubium. Tapiroidea is restricted to two species: the lophialetid Rhodopagus? minimus and the deperetellid Deperetella cristata (Radinsky, 1965). The hyracodontid Triplopus? progressus and the paraceratheriid Juxia sharamurenensis also are known (Radinsky, 1967; Qiu and Wang, 2007). The brontotheriidae is represented by Rhinotitan kaiseni and R. andrewsi (Granger and Gregory, 1943; Mihlbachler, 2008). The Sharamurunian is usually correlated to the late Uintan and Duchesnean NALMAs (Wang et al., 2007; Vandenberghe et al., 2012). However, Wang (1982) correlated the Sharamurunian with the late Uintan NALMA (Ui3) based on the similar evolutionary stage between Rhinotitan andrewsi and the late Uintan (Ui3) Protitanotherium (Robinson et al., 2004; Mihlbachler, 2008; Gunnell et al., 2009).
Regarding the perissodactyls from the upper member of the original Shara Murun Formation, overlying Ulan Gochu Formation, and the “Lower White” and “Middle Red” of the Erden Obo (all may represent restricted Ulangochuan), the diversity of tapiroids radically decreased over time with only the deperetellid Teleolophus magnum remaining. The hyracodontid Ulania wilsoni and abundant amynodontid Amynodontopsis parvidens were present (figs. 4, 10, 11). Paraceratheres are represented by the more advanced Juxia shoui and Urtinotherium parvum rather than the underlying Juxia sharamurenensis (Qiu and Wang, 2007). Brontotheres are slightly more diverse than those from the restricted Sharamurunan, including Pachytitan ajax, Titanodectes minor, and Embolotherium grangeri (Granger and Gregory, 1943; Mihlbachler, 2008). Among perissodactyls from the restricted Ulangochuan, only Amynodontopsis is distributed in North America as the genotype A. bodei (Stock, 1933; Wall, 1989). Ulangochuan is usually correlated to the early Chadronian NALMA (Wang et al., 2007; Vandenberghe et al., 2012). However, considering that Amynodontopsis is known only from the Duchesnean NALMA (Lucas, 1992; Robinson et al., 2004), we tentatively correlate the restricted Ulangochuan with the whole Duchesnean NALMA.
The fauna from the Baron Sog Formation and the equivalent the “Middle White” of Erden Obo contain roughly equally diverse perissodactyl faunas as those from the Ulangochuan (figs. 4, 10, 11), including brontotheriids (Embolotherium andrewsi and Parabrontops gobiensis), abundant specimens of the amynodontid Zaisanamynodon borisovi, hyracodontids (Ardynia praecox and Proeggysodon qiui), and the paraceratheriid Urtinotherium intermedium. Since P. gobiensis, Z. borisovi, and A. praecox are indicative of late Eocene Ergilian (Dashzeveg, 1991; Dashzeveg, 1993), the Baron Sog Formation can be correlated to the Ergilian, which in turn can be correlated to the Chadronian NALMA.
Both brontotheres and amynodontids disappeared in the Oligocene Hsandagolian, as represented by the fauna from the “Upper White” at Erden Obo and its equivalent bed at Nom Khong Obo (Wang, 2003) (figs. 10, 11). The perissodactyl diversity decreases considerably, and consists of the hyracodontid Ardynia kazachstanensis, the paraceratheriid Paraceratherium grangeri, some rhinocerotids, and the chalicothere Schizotherium (fig. 11). The age of the “Upper White” at Erden Obo is considered late early Oligocene, as discussed above.
On the basis of the distribution of the general numbers of perissodactyls from the Erlian Basin from the early Eocene through the early Oligocene, we inferred that the changes in perissodactyl diversity that occurred over that time were in relation to coincedent paleoclimatic change (fig. 11). The biodiversity greatly increased during the Arshantan as compared to the early Bumbanian, which is probably attributable to the Early Eocene Climatic Optimum (EECO) (Zachos et al., 2001, 2008; Vandenberghe et al., 2012). Although global temperatures decreased after EECO, perissodactyls reached their greatest diversity in the Irdinmanhan. Alternatively, the underestimated diversity of perissodactyls in the Arshantan may account for its relatively lower diversity than in the Irdinmanhan. The gradual decrease in perissodactyl diversity from the middle Eocene Irdinmanhan to the late Eocene Ergilian likely can be attributed to the steady cooling trend of global temperature as indicated by a 7° C drop in deep-sea temperatures during the middle and late Eocene (Zachos et al., 2001; Prothero, 2006). In the early Oligocene Hsandogolian, the diversity of perissodactyls distinctly declined, and may be the result of the dramatic 5°–6° C drop in deep-sea temperatures across the Eocene-Oligocene boundary (Prothero, 2006). After the Eocene-Oligocene transition, the perissodactyl-dominant fauna in a warm, humid Eocene “greenhouse” was replaced by a rodent/lagomorph-dominant fauna in a dry, cold Oligocene “icehouse” (Meng and McKenna, 1998; Sun et al., 2014).
CONCLUSIONS
We clarify and discuss the temporal and spatial distributions of the dominant perissodactyl fossils from the Erlian Basin at the species level from the early Eocene to the early Oligocene, spanning about 26 million years. Over the past decade, our investigations in the Erlian Basin have resolved some long-standing problems, particularly in the Huheboerhe area, and have produced many new mammalian fossils, including perissodactyls from various localities and stratigraphic horizons. On the basis of this present paper, future study of these new materials may improve or alter our understanding of the evolutionary history of perissodactyls. Furthermore, although we propose a correlation of the fossiliferous sedimentary deposits at Erden Obo with other formations/faunas from type localities, evidence from our new discoveries of perissodactyls and other mammalian fossils may support or augment the present argument. Most importantly, the Eocene Asian Land Mammal Ages, which are based primarily on the faunas from the Erlian Basin, are in need of a more finely calibrated definition, subdivision, and intercontinental correlation that will be accomplished in future studies of the entire mammalian fauna and magnetostratigraphy.
ACKNOWLEDGMENTS
We appreciate the discussion with Zhan-Xiang Qiu, Ban-Yue Wang, Chuan-Kui Li, and Zhao-Qun Zhang (all IVPP). We thank Xun Jin, Wei Zhou, Shi-Jie Li, Qi Li, Yong-Xing Wang, Yong-Fu Wang, Yan Li, Wei Chen, Qiang Cao, Qiang Li (all IVPP), K.C. Beard (University of Kansas), and D.L. Gebo (Northern Illinois University) for assistance in the field; Thomas Stidham (IVPP) for assistance with the English text; Yu Chen (IVPP) for the drawing of figure 2; and S. Bell, T. Baione, M. Reitmeyer, A. Springer, and A. Gishlick (all AMNH) for providing literature access and assistance at the AMNH. We are grateful to M. Mihlbachler (New York Institute of Technology) and an anonymous reviewer for helpful comments, and editor R. Voss and M. Knight (both AMNH) for instructive editoral comments. Funding was provided by grants from the Strategic Priority Research Program of Chinese Academy of Sciences (grant no. XDB26000000), the National Natural Science Foundation of China (41672014, 41572021), Youth Innovation Promotion Association CAS, State Key Laboratory of Palaeobiology and Stratigraphy (Nanjing Institute of Geology and Palaeontology, CAS) (No. 163103), China Scholarship Council, and Frick Fund from the Division of Paleontology, American Museum of Natural History.
REFERENCES
Appendices
APPENDIX 1
List of Perissodactyl Fossils from the Ulan Shireh Formation at North Mesa
Asterisks indicate the holotype discovered from the North Mesa.
Perissodactyla
Tapiroidea
Lophialetidae
*Zhongjianoletes chowi Ye, 1983
*Simplates ulanshirehensis Qi, 1980
*Breviodon acares Radinsky, 1965
*Rhodopagus pygmaeus Radinsky, 1965
Lophialetes expedius? Matthew and Granger, 1925c
Lophialetes sp.
Deperetellidae
Teleolophus medius? Matthew and Granger, 1925c
Rhinocerotoidea
Hyracodontidae
Triplopus proficiens (Matthew and Granger, 1925c)
Paraceratheriidae
*Forstercooperia ulanshirehensis Wang, Bai, Meng, and Wang, 2018
Brontotheriidae
*Microtitan mongoliensis (Osborn, 1925)
*Desmatotitan tukhumensis Granger and Gregory, 1943
*Acrotitan ulanshirhensis Ye, 1983
*Epimanteoceras formosus Granger and Gregory, 1943
*Metatitan primus Granger and Gregory, 1943
APPENDIX 2
Field Records of Paleogene Perissodactyl Fossils by the CAE, SSPE, and IVPP from the Erlian Basin of Inner Mongolia, China
The field numbers of CAE expeditions in 1922, 1923, 1925, 1928, and 1930 include 32–107, 130–404, 507–616, 631–812, and 813–937, respectively. Abbreviations: AM, AMNH FM; BSL, Baron Sog Lamasery; EMM, Erenhot Dinosaur Museum, mammal fossils; SS, Sino-Soviet Paleontological Expedition specimens; TLC, Telegraph Line Camp; TKL, Tukhum Lamasery; and V, IVPP V.
Continued
Continued
Continued
Continued
Continued
Continued
Continued
Continued
Continued
Continued
Continued
Continued
Continued
