博格達吐魯番獸（二齒獸類）的骨骼學

Osteology of Turfanodon bogdaensis （Dicynodontia）

Abstract Within the dicynodont genus Turfanodon， there are two recognized species， T.bogdaensis and T. jiufengensis. Both species are known by relatively complete cranial materials，but the mandibles and most postcranial bones have been described only for T. jiufengensis. Thispaper reports new dicynodont specimens from Turpan， Xinjiang， referring them to T. bogdaensis.They can clearly be differentiated from T. jiufengensis by the flatter lateral surface of the snoutregion， a prominent swelling on the lateral dentary shelf， and the rounded femoral head. Thediagnosis of Turfanodon is revised. The combination of a flat circumorbital rim， posterior portionof anterior pterygoid rami with converging ventral ridges， and a possible autapomorphy， a deepnotch on scapula forming procoracoid foramen， are confirmed. It also differentiated from alldicynodonts other than Myosaurus， Kembawacela and Lystrosaurus by having accessory ridgeslateral to the median palatal ridge.

Key words Turpan， Xinjiang; Lopingian， Permian; Guodikeng Formation; Dicynodontia，Turfanodon; osteology， postcranial skeleton

Citation Shi Y T， Liu J， 2024. Osteology of Turfanodon bogdaensis （Dicynodontia）. Vertebrata"PalAsiatica， 62（3）： 186–200

1 Introduction

Turfanodon is a large-sized dicynodont of the late Permian Period. It was one of the mostwidely distributed dicynodonts in eastern Pangea and crossed both tropical and temperatezones （Sun， 1973; Li et al.， 2000; Liu， 2021）. The holotype of T. bogdaensis （IVPP V3241）was collected from the Guodikeng Formation of Turpan， Xinjiang， China. This genus wasproposed as a junior synonym of Dicynodon （King， 1988）， but has more recently beenrevalidated （Kammerer et al.， 2011）. Two species， Striodon magnus from Jimusar （Sun， 1978），and D. sunanensis from Gansu （Li et al.， 2000）， have been proposed as junior synonyms of T.bogdaensis （Kammerer et al.， 2011）. Recently， a second species， T. jiufengensis， was namedbased on specimens from the Naobaogou Formation of Nei Mongol， China （Liu， 2021）.Currently， this genus is the most abundant Permian dicynodont in China.

Turfanodon jiufengensis is known from five specimens including crania， mandibles， andrelatively complete postcrania， whereas T. bogdaensis is only known from cranial materials.Here we report a few new specimens of T. bogdaensis， including many postcranial bones，from the Guodikeng Formation of Turpan， Xinjiang， China. They provide more anatomicalinformation on this species.

Institutional abbreviations IGCAGS， Institute of Geology， Chinese Academyof Geological Sciences， Beijing， China; IVPP， Institute of Vertebrate Paleontology andPaleoanthropology， Chinese Academy of Sciences， Beijing， China; NHMUK， The NaturalHistory Museum， London， England， UK; RC， Rubidge Collection， Wellwood， Graaff-Reinet，South Africa.

2 Geological setting， material and methods

The studied specimens were collected in 2012 from the Guodikeng Formation of theSouth Taodonggou locality， Turpan， Xinjiang. A partial archosauriform left hindlimb （IVPPV22764） was excavated with the dicynodonts at the same locality. This specimen was recentlymade the holotype of a new taxon， Vigilosaurus gaochangensis， and its host layer is estimatedto be around 252.2 Ma （Angielczyk et al.， 2022; Chen and Liu， 2022）.

For the dicynodont materials， most bones belong to a nearly complete specimen， IVPPV31402; whereas a smaller humerus， a femur and some vertebrae show the presence of asecond specimen， IVPP V31403.

The nomenclature of the angular follows Olroyd and Sidor （2022）; and someinterpretations of the structure of the postcranial skeleton follow Bishop and Pierce （2023a， b）.

3 Systematic paleontology

Anomodontia Owen， 1860

Dicynodontia Owen， 1859

Dicynodontoidea Olson， 1944

Turfanodon Sun， 1973

Revised diagnosis （expanded from Liu， 2021） Large dicynodontoid with a tall， steeplyslopingsnout， long caniniform process， heavily pitted facial region， flat circumorbital rim，elongated and narrow intertemporal bar， posterior portion of anterior pterygoid rami withconverging ventral ridges. Possible autapomorphy： deep notch on scapula forming procoracoidforamen.

Differentiated from all dicynodonts other than Myosaurus， Kembawacela andLystrosaurus by having accessory ridges lateral to the median palatal ridge. Differentiatedfrom all dicynodontoids other than Dinanomodon by a contact between the premaxilla andthe frontal. Differentiated from Dinanomodon by having the anterior tip of the snout squared off and nasal bosses present as anteroposteriorly elongated swellings near the posterodorsalmargin of external nares.

Turfanodon bogdaensis Sun， 1973

Holotype IVPP V3241， an incomplete skull.

Referred specimens IGCAGS V296， an incomplete skull （holotype of “Dicynodonsunanensis”）; IVPP V31402， an incomplete skull， mandible， and most of the postcranial bones;IVPP V31403， a humerus， a femur， three dorsal vertebrae and the 3rd–5th sacral vertebrae.

Revised diagnosis A large Permian dicynodontoid with a thick lateral dentary shelf.Differentiated from its congener T. jiufengensis by weakly developed nasal bosses， a mediansnout groove on premaxilla， flat lateral snout surface， a contact between maxilla and nasal（separating lacrimal from septomaxilla）， prefrontal anterior extension not extending anteriorto lacrimal， parietal exposed on midline groove of intertemporal bar， and proximal articularsurface of femur more rounded.

Locality and horizon Turpan， Xinjiang and Sunan， Gansu; Guodikeng Formation andSunan Formation， Lopingian （Changhsingian）， Permian.

Description The skull of IVPP V31402 is nearly complete except for the missingleft postorbital bar， right zygomatic arch， and both caniniform processes （Fig. 1）. The bonesurfaces are poorly preserved， so many sutures and detailed features are hard to discern.

The premaxilla has an incomplete anterior margin （Fig. 1）. The bone surface is stronglyeroded， so neither a groove， as in IVPP V3241 and IGCAGS V296， nor a ridge， as inTurfanodon jiufengensis， can be observed. The premaxilla contacts the frontal posterodorsally，which is a diagnostic character of Turfanodon （Fig. 1B） （Sun， 1973; Liu， 2021）. Thepremaxillary anterior surface is clearly demarcated from the lateral surfaces， and the lateralsurface of the snout region is flat， not a bi-planar surface as in T. jiufengensis. In T. jiufengensis，the maxilla turns more laterally than the premaxillary lateral surface， forming a concave areabelow the external naris （Fig. 2A， B） （Liu， 2021）. The snout is tall and sloping （Fig. 1E， F）.Two faint accessory ridges lie lateral to the median palatal ridge， as T. jiufengensis （Fig. 1C， D）（Liu， 2021）. This character has only been reported as present in Myosaurus， Kembawacela andLystrosaurus previously （Cluver， 1971， 1974; Angielczyk et al.， 2021a）.

The septomaxilla forms a concavity posterior to the naris （Fig. 1E）. Both caniniformprocesses are broken， so the preserved maxillary ventral margin lies above the level of thequadrate （Fig. 1E， F）. The maxillary lateral surface is rather smooth， without the distinctbuttress as in IVPP V3241， similar to IGCAGS V296 （Liu， 2021）.

The nasals are separated from each other by the contact of the premaxilla with the frontal（Fig. 1B）. The paired nasal bosses are weakly developed （Fig. 1E， F）， as in IGCAGS V296，while they are hardly discernable in IVPP V3241. The nasal bosses of all of these specimensare less pronounced than those of T. jiufengensis.

The maxilla， the jugal and the squamosal contribute to the zygomatic arch. Laterally， the jugal is covered by the maxilla anteriorly and the squamosal posteroventrally， and it is exposedas a narrow strip below the orbit. The jugal also extends dorsally to form part of the posterioredge of the postorbital bar （Fig. 1E， F）. Ventrally， the jugal contributes to the subtemporal arch（Fig. 1C， D）. It also forms the dorsolateral margin of the labial fossa （Fig. 1C–F）.

The zygomatic process of the squamosal tapers anteriorly as a narrow wedge， with astraight suture with the maxilla （Fig. 1E， F）， same as in IVPP V3241 （Sun， 1973， Fig. 3） andIGCAGS V296. The same morphology may also be present in T. jiufengensis， although Liu（2021） explained the squamosal inserts in the jugal， and contacting with maxilla only on itsposterior portion.

The fat prefrontal forms the anterodorsal corner of the orbit （Fig. 1）. The bone surfaceis poorly preserved， so its suture with the lacrimal and its anterior extension are hard to trace.Based on bone surface texture， it seems that those two bones extend anteriorly at the samelevel and the lacrimal does not touch the septomaxilla （Fig. 1E， F）.

The frontal contacts the premaxilla anteriorly （Fig. 1B）. The postfrontal is wedge-shaped，as in the holotype （Fig. 1B）. The preparietal is depressed， showing the shape of a ‘three-toe’footprint， with one large median process and two smaller lateral processes （Fig. 1A， B）. Thisshape is similar to that of those relatively small specimens of T. jiufengensis， but differs fromIVPP V3241 and IGCAGS V296 （Liu， 2021）.

The intertemporal bar is formed by the postorbital and the parietal （Fig. 1A， B）.The postorbital orients obliquely， slightly different from the other known specimens of T.bogdaensis. In the holotype， the postorbital is nearly vertical in this region， while in IGCAGSV296， the intertemporal bar has some deformation， but a vertical orientation can be recognized.Turfanodon jiufengensis has a vertically oriented postorbital， as in IVPP V3241. The parietalis bounded by the postorbitals laterally and has a narrow exposure on the dorsal surface. Thereis a shallow midline groove formed by the parietal posteriorly （Fig. 1A， B）， unlike the crestlikeexposure in T. jiufengensis （IVPP V26035） （Fig. 2）. In both IVPP V3241 and IGCAGSV296， the parietal is exposed in the groove formed by the postorbitals （Liu， 2021）. Overall， T.bogdaensis has a wider intertemporal bar than T. jiufengensis.

The palatine extends anteriorly to the level of the posterior margin of the tusks，and it distinctly decreases in width posteriorly （Fig. 1C， D）. The posterior extension ofthe ectopterygoid is shorter than that of the palatine in ventral view （Fig. 1C， D）. Theinterpterygoid vacuity is drop-shaped and expose the parabasisphenoid ventrally. It’s margin iscontributed by the vomer anteriorly and the pterygoid posteriorly.

The anterior pterygoid keel extends for the entire anterior ramus of the pterygoid（Fig. 1C， D）， as in T. jiufengensis （Liu， 2021）. The posterior portion of the keel of the anteriorpterygoid ramus cannot be discerned， but in IVPP V3241， some trace can be observed of thisportion （Fig. 2E）. Only the anterior process of the epipterygoid is preserved on the left side，bordered by the pterygoid （Fig. 1E， F）. The quadratojugal is plate-like. The quadrate has arounded anterior edge， as in most dicynodonts （Fig. 1E， F）. The periotic has a rod-like pilaantotica. The sphenethmoid （“anterior plate”） complex consists of the orbitosphenoid andmesethmoid， but their sutures are unidentifiable （Fig. 1E， F）.

Mandible： The mandible is nearly complete （Fig. 3）， and its shape is nearly identicalto that of Turfanodon jiufengensis. The symphysis is composed of the dentary， the splenialand the angular， and forms an upturned beak （Fig. 3）. The anterior surface of the symphysishas a median ridge and two lateral ridges that follow the lateral edges （Fig. 3A， B）. Thesymphysis and the mandibular rami form a narrow junction anterior to the lateral dentaryshelf. This narrow portion could accommodate the enlarged caniniform processes when thejaw was closed （Fig. 3E）. Dentary tables are present. The posterior dentary sulcus developed as a narrow groove with two sharp ridges bound it （Fig. 3E）. The lateral dentary shelf isdeveloped as a prominent swelling above the anterior margin of the mandibular fenestra，unlike the thin shelf of T. jiufengensis （Fig. 3A， B）. Its morphology is somewhat similar tothat of Kunpania， but Kunpania has more distinct dorsal and ventral surfaces （Angielczyk etal.， 2021b）.

The ventral margin of the reflected lamina is incomplete. A thin ridge runs posterodorsallyabove the dorsal notch （Fig. 3A， B）， without any expansion in its dorsal end， differing fromthe condition in Jimusaria （Shi and Liu， 2023）. The posteroventral fossa is partly preserved onventral margin of the reflected lamina. The angular gap is wide， and forms a distinct angularfossa （Fig. 3A， B）.

Postcranial skeleton： Vertebrae and ribs At least 26 vertebrae are preserved， althoughmany of them are preserved only as the amphicoelous centra （Fig. 4A–I）. Three articulateddorsal vertebrae and three sacral vertebrae might belong to IVPP V31403， considering theirsmaller size （Fig. 4G， H）. Only those vertebrae that can be referred to specific regions aredescribed here （Fig. 4C–H）.

The axis and the third cervical are articulated （Fig. 4A， B）. The axis features ananteroposteriorly expanded neural spine and the odontoid. The prezygapophysis is broken butthe postzygapophysis extends posteriorly beyond the posterior margin of the centrum. A smallboss is developed near the anterior margin of axis centrum， posterior to the odontoid. King（1981） proposed that is a facet for a cervical rib in “Dicynodon trigonocephalus”， but the rib isnot observed here or in the specimens of T. jiufengensis （Liu， 2021）.

The dorsal vertebrae feature high， nearly vertical neural spines; reduced， more verticallyoriented prezygapophyse; horizontally oriented postzygapophyses and slight overlap of thepostzygapophyses by the prezygapophyses from the lateral side. Their lateral surfaces of thecentra are slightly concave and smooth （Fig. 4C， D）.

Three sacral vertebrae of IVPP V31402 can be identified， likely the first to the third.They are identified by the elongated and vertically oriented pre zygapophyses of the secondone and the third one， the reduced postzygapophyses， the posterodorsally oriented neuralspine， and the fused centra and zygopophyses. The centrum of the third sacral of the seriesis anteroposteriorly elongate and laterally narrow （Fig. 4E）. The third through fifth sacralvertebrae are preserved in V31403 （Fig. 4H）. The fourth sacral vertebra still fused with thethird one， and the base of its neural spine extends more posteriorly than the anterior one，as in T. jiufengensis （Liu， 2021）. The fifth sacral vertebra does not fused with the fourth，and its prezygapophyses are reduced in size compared to those of the more anterior sacralvertebrae.

At least two caudal vertebrae are preserved， consisting of only centra and incompleteneural arches （Fig. 4F）. They lack the transverse processes and rib facets.

Several separate ribs are preserved， most of them are only partially exposed from thesurrounding matrix （Fig. 4J）. One exposed proximal end is holocephalous and grooves arepresent on the shafts of some ribs， indicating those ribs belong to the posterior portion of thedorsal rib series.

Sternum： The incomplete sternum is preserved with the left pectoral girdle， and isexposed ventrally （Fig. 5A）. It is parallelogram-shaped， measuring ～11 cm in anteroposteriorlength. One sternal boss present along the lateral margin （Fig. 5B）.

Pectoral girdle： The preserved pectoral girdles include the scapula， procoracoid， andcoracoid from the left side and the procoracoid and coracoid from the right side. The claviclesare missing. The left pectoral girdle is articulated （Fig. 5A， B）， while the right one is separated（Fig. 5C， D）.

The scapula is distinctly convex laterally， fitting the body contour （Fig. 5A， B）. Its anterior edge projects laterally as a crest， and the origin of the M. triceps scapularis isdeveloped as rugosity. The scapula has a large ventral articular surface for the procoracoid，and a deep notch on the ventral margin. This notch contributes to the border of the coracoidforamen （Fig. 5A， B）. The acromion process is incomplete but well defined （Fig. 5A）.

The procoracoid and coracoid are sutured together as the coracoid plate （Fig. 5B）. Theprocoracoid is a nearly square bone with a notch on the posterodorsal edge （Fig. 5A–C）. Thisnotch forms the procoracoid portion of the coracoid foramen. The procoracoid does notparticipate in the glenoid （Fig. 5B）. The coracoid has a strongly concave medial surface（Fig. 5B， D）. The origin of the triceps coracoideus is thicker than the other portion of thecoracoid edge （Fig. 5D）.

Pelvic girdle： The elements of the pelvic girdle are disarticulated. The ilia and theischia of both sides are preserved， but the right ones are more complete; only the left pubis ispreserved （Fig. 5E–I）.

The dorsal edge of the iliac blade is incomplete （Fig. 5E–G）. The preacetabular processis longer and higher than the postacetabular process. Medially， three facets for the sacral ribscan be observed on the middle （Fig. 5E， F）. According to Liu （2021）， Turfanodon jiufengensispossibly has six sacral ribs. The fossae of T. bogdaensis suggest that at least three ribsare firmly articulated with the ilium and the rest of them are free from it in this species. Aconstricted neck lies ventral to the blade.

The ischium is a fan-shaped plate， it contributes to the posterior portion of the acetabulum（Fig. 5H）. The outer surface is slightly convex. A notch for the obturator foramen lies anteriorto the neck of the ischium， close to the acetabulum.

The pubis is a relatively small bone （Fig. 5I）. Its body is laterally curved， the posteriorportion has a notch， which forms the anterior margin of the obturator foramen.

Limbs： Three humeri and two different sized femora are preserved together， indicatingthat there are at least two individuals preserved （Fig. 6）. Based on the sizes of the articularfacets of the limb gridles and proximal widths of the limbs， the larger limb bones are referredto the same individual as the limb girdles. The larger humerus is similar in size to a humerus ofT. jiufengensis （IVPP V26035） （Fig. 6E， F）， the skull of which is of comparable size with IVPPV31402 （Liu， 2021） （Fig. 2）. Based on this， the larger limbs and girdles are referred to thesame individual as the cranial material here.

Forelimb： The humeri are well-preserved， including two larger both side humeri （IVPPV31402， Fig. 6A， B， E）， and a smaller complete right humerus （V31403， Fig. 6C）. Dorsally， thehumerus proximal head is rounded， and a rounded， rugose region lies posterior to the condyle，which is the insertion of the M. subscapularis （Fig. 6B）. A bulge developed anterior to thehumerus proximal head， which should be the insertion of the supracoracoideus （supraspinatus）（Ray， 2006; Bishop and Pierce， 2023a）， but it is less developed in Lystrosaurus （Fig. 2C， D）.A thin， curved deltopectoral crest is located anteroventral to the proximal head. This crest isa triangular， thin plate that is convex dorsally， and extends for nearly half the length of thehumerus （Fig. 6A–C）. The insertion of the M. latissimus dorsi （the insertion of the M. teresmajor according to Bishop and Pierce （2023a）） shows a pinna-like shape （Fig. 6A）. Theentepicondylar foramen penetrates the ridge which is elongated by the deltopectoral crest edge（Fig. 6B）. The anteroposteriorly continuous distal articular facet， can be divided， from anteriorto posterior， as the ectepicondyle， the ulnar condyle and the entepicondyle （Fig. 6E）. ContraKammerer （2018）， a “supinator process” （the “brachioradialis process” according to Bishopand Pierce （2023a）） is present in Turfanodon bogdaensis， T. jiufengensis and even Lystrosaurus（Fig. 6A， D）， located proximal to the ectepicondyle， and close to the base of the shaft as in theTriassic stahleckeriids Pentasaurus and Zambiasaurus （Kammerer， 2018）.

The ulna is incomplete， only the proximal portion is preserved. The proximal end isrelatively flat and extends lateromedially （Fig. 6G）. There is a triangular facet for the radiusfacing anteromedially. The proximal end extends laterally as a small olecranon process （Fig. 6G）.The radius is a rod-like bone with a concave proximal end and a convex distal end （Fig. 6H）.

Hindlimb： The preserved two femurs are slightly different in size but nearly identical inshape （Fig. 6I–K）. The femur has a hemispherical head which is continuous with the shaft.Lateral to the femoral head is the trochanter major， which is a flat， broad crest. Distally， it bearsthe insertion of the M. iliofemoralis and the distal end convex anteriorly. The trochanteric crestis about 2/5 length of the femur; this ratio is larger than that of Lystrosaurus but similar to thatof the kannemeyeriiforms Parakannemeyeria and Shanbeikannemeyeria （Sun， 1963）.

The left tibia and fibula are well preserved （Fig. 6L， M）. The tibia is a robust bone with aprominently expanded proximal end and gently expanded distal end. The fibula is a thin， rodlikebone without any crest on the shaft. Medially， the proximal end of the tibia has a notchfitting the lateral condyle of the fibula. The proximal facet for femur of the fibula shows a“quotation mark” shape.

An astragalus （Fig. 6N） and metatarsal （Fig. 6O） are preserved. The astragalus is a nearlyrounded bone， a groove and a concave developed on its surface.

4 Discussion

IVPP V31402 can be referred to Turfanodon based on the combination of a narrowintertemporal bar， large temporal fenestra， premaxilla contacting the frontal， and nasalsseparated by the premaxilla and the frontal. A ‘three-toe’-shaped， depressed preparietalwas previously proposed as an autapomorphy of T. jiufengensis （Liu， 2021）. Although thepreparietal of V31402 is ‘three-toe’-shape rather than oval， it can be referred to T. bogdaensisbased on the presence of slightly raised nasal bosses， prefrontal anterior extension notextending anterior to lacrimal， and maxillary contact with nasal， indicating that preparietalshape does not allow confident distinction between the two species of Turfanodon.

We found more differences between T. bogdaensis and T. jiufengensis， including thatthe lateral surface of the snout region is nearly flat rather than concave （Fig. 2） and thelateral dentary shelf forms a prominent swelling rather than a thin sheet in T. bogdaensis. T.bogdaensis （IVPP V31402） has a more rounded femoral head， while T. jiufengensis （V26035）has a less distinct femoral head， although the latter is larger （and thus one would presume itwould be better ossified） than V31402.

Within T. bogdaensis， different specimens exhibit different degrees of parietal exposure.IVPP V3241 and IGCAGS V296 have deep grooves overlapped by the parietals， but inIVPP V31402 the groove is relatively shallow， and is limited on the posterior portion of theintertemporal bar （Fig. 1A）. This difference could be due to the ontogenetic variation. Thenarrowing of the intertemporal bar and the parietal exposure is a common ontogenetic variationamong the dicynodonts （Kammerer et al.， 2015）.

This study presents the osteology of the mandible and postcranial skeleton of Turfanodonbogdaensis， including characters that were previously unknown for this genus. The postcranialbones are generally similar to those of other Dicynodon-grade dicynodontoids except for theshape of the ventral margin of the scapula： its basal portion is deeply notched by the largecoracoid foramen and it appears there is a distinct anteroventral projection below the acromionprocess （Fig. 5A， B）. A scapular contribution to the coracoid foramen is also present in， e.g.，Diictodon， Dicynodontoides， Lystrosaurus， and Ischigualastia， but the scapulae in thosetaxa do not have such a deep notch as in IVPP V31402 （Young， 1935; Cox， 1959; Cox andParrington， 1965; Ray and Chinsamy， 2003; Angielczyk， 2007）. This suggests that the deeplynotched scapula is autapomorphic for the genus Turfanodon. The large coracoid foramenmight house a robust subclavian artery passing through （Romer， 1956）. The insertion of theM. supracoracoideus is strongly developed on the humerus （Fig. 2C）， indicating that thismuscle was well-developed in Turfanodon. Both the more abundant blood supply and strongM. supracoracoideus suggest well-developed anterodorsal movement of the forelimb （Ray andChinsamy， 2003; Ray， 2006; Bishop and Pierce， 2023a）.

Variation in the preparietal： Referral of IVPP V31402 to Turfanodon bogdaensis indicatesthat the “three-toe” shaped preparietal is not exclusive to T. jiufengensis， and the preparietalshape shows substantial variation in T. bogdaensis specimens. Because the oval-shapedpreparietal appears in relatively larger skulls， this variation may be related ontogeny. In T.jiufengensis， the preparietal shape is unknown in the large skull， but it shows some variationin the small-sized skulls （Liu， 2021）. Based on the condition in T. bogdaensis， we nowhypothesize that T. jiufengensis also had an oval-shaped preparietal in the adult.

Two other dicynodont specimens show the similar preparietal shape to that of smallTurfanodon： NHMUK R4955 and RC 64. NHMUK R4955 is the holotype of Palemydopsplatysoma and has a trident-shaped preparietal with anterior and lateral processes ofcomparable length， plus a unique posterior elongation of the pineal foramen （Broom， 1921;Boos et al.， 2016）. RC 64 is the holotype of Dicynodon duvenhagei， in which the preparietalhas three similar processes， but the suture is more tortuous than in Turfanodon （Broom， 1947）.The validity of both species is questionable. NHMUK R4955 was considered a possible specimen of Pristerodon mackayi， but later regarded as a nomen dubium due to its incompletepreparation by Keyser （1993）， although King （1988） considered it valid but of uncertainrelationships. RC 64 was referred to Kingoria by Cluver and Hotton （1981）， and subsequentlyreferred to the species Dicynodontoides recurvidens （Angielczyk et al.， 2009; Kammerer etal.， 2011）. A preparietal with three processes is unknown in the other published specimens ofDicynodontoides， and is also rare within Bidentalia.

Acknowledgements We thank LI Lu， LIU Yu-Feng， LI Xing-Wen， and XU Xu for thefieldwork， GAO Wei for the photograph， XU Yong for optimizing drawings. We also thankChristian Kammerer and Kenneth Angielczyk， their comments and edit improved this article.This work is supported by Strategic Priority Research Program of Chinese Academy ofSciences （XDB26000000） and The Innovation Training Programs for Undergraduates， ChineseAcademy of Sciences.

摘要：吐魯番獸（Turfanodon）是一類生活在二疊紀晚期的二齒獸類，曾廣泛分布在新疆至華北一帶。目前該屬下建立有兩個種——博格達種（T. bogdaensis）與九峰種（T. jiufengensis）。兩個種都具有相對完整的頭骨材料，但只有九峰種報道了下頜及頭后骨骼材料。一批產(chǎn)自新疆吐魯番地區(qū)，可以歸入到博格達吐魯番獸的新材料，具有一系列可以明顯區(qū)別于九峰種的特征，包括頭骨的鼻吻部側(cè)面平坦，下頜齒骨側(cè)架成腫突狀，以及股骨頭為近半球狀。吐魯番獸屬的鑒別特征也有更新，包括眼框周緣平坦，兩條翼骨前支上的嵴向后匯聚，以及一個可能的自有衍征：肩胛骨上參與塑造烏喙孔的凹缺很大。吐魯番獸腭部的中嵴兩側(cè)發(fā)育有副嵴，這一特征區(qū)別于除肌龍獸（Myosaurus）、掘鐵獸（Kembawacela）和水龍獸（Lystrosaurus）外的其他二齒獸類。

關(guān)鍵詞：新疆吐魯番盆地，二疊紀樂平世，鍋底坑組，二齒獸類，吐魯番獸，骨骼學，頭后骨骼

中圖法分類號：Q915.864 文獻標識碼：A 文章編號：2096–9889（2024）03–0186–15

References

Angielczyk K D， 2007. New specimens of the Tanzanian dicynodont “Cryptocynodon” parringtoni von Huene， 1942（Therapsida， Anomodontia）， with an expanded analysis of Permian dicynodont phylogeny. J Vert Paleont， 27（1）：116–131

Angielczyk K D， Sidor C A， Nesbitt S J et al.， 2009. Taxonomic revision and new observations on the postcranial skeleton，biogeography， and biostratigraphy of the dicynodont genus Dicynodontoides， the senior subjective synonym of"Kingoria （Therapsida， Anomodontia）. J Vert Paleont， 29（4）： 1174–1187

Angielczyk K D， Benoit J， Rubidge B S et al.， 2021a. A new tusked cistecephalid dicynodont （Therapsida， Anomodontia）from the upper Permian upper Madumabisa Mudstone Formation， Luangwa Basin， Zambia. Pap Palaeontol， 7（1）：405–446

Angielczyk K D， Liu J， Yang W， 2021b. A redescription of Kunpania scopulusa， a bidentalian dicynodont （Therapsida，Anomodontia） from the ？Guadalupian of northwestern China. J Vert Paleont， 41（1）： e1922428

Angielczyk K D， Liu J， Sidor C A et al.， 2022. The stratigraphic and geographic occurrences of Permo-Triassic tetrapods in"the Bogda Mountains， NW China—implications of a new cyclostratigraphic framework and Bayesian age model. J"Afr Earth Sci， 195： 104655

Bishop P J， Pierce S E， 2023a. The fossil record of appendicular muscle evolution in Synapsida on the line to mammals： Part"I–Forelimb. Anat Rec， 1–62

Bishop P J， Pierce S E， 2023b. The fossil record of appendicular muscle evolution in Synapsida on the line to mammals： Part"II–Hindlimb. Anat Rec， 1–71

Boos A D， Kammerer C F， Schultz C L et al.， 2016. A new dicynodont （Therapsida： Anomodontia） from the Permian of"southern Brazil and its implications for Bidentalian origins. PLoS One， 11（5）： e0155000

Broom R， 1921. On some new genera and species of anomodont reptiles from the Karroo Beds of South Africa. Proc Zool"Soc Lond， 1921： 647–674

Broom R， 1947. XXI.—A contribution to our knowledge of the vertebrates of the Karroo Beds of South Africa. Trans R Soc"Edinb， 61（2）： 577–629

Chen J， Liu J， 2022. A late Permian archosauriform from Xinjiang shows evidence of parasagittal posture. Sci Nat，110（1）： 1

Cluver M A， 1971. The cranial morphology of the dicynodont genus Lystrosaurus. Ann S Afr Mus， 56： 155–274

Cluver M A， 1974. The cranial morphology of the Lower Triassic dicynodont Myosaurus gracilis. Ann S Afr Mus， 66：35–54

Cluver M A， Hotton N， 1981. The genera Dicynodon and Diictodon and their bearing on the classification of the"Dicynodontia （Reptilia， Therapsida）. Ann S Afr Mus， 83： 99–146

Cox C B， 1959. On the anatomy of a new dicynodont genus with evidence of the position of the tympanum. Proc Zool Soc Lond， 132（3）： 321–367

Cox C B， Parrington F R， 1965. New Triassic dicynodonts from South America， their origins and relationships. Philos Trans"R Soc Lond B， 248： 457–514

Kammerer C F， 2018. The first skeletal evidence of a dicynodont from the lower Elliot Formation of South Africa. Palaeontol Afr， 52： 102–128

Kammerer C F， Angielczyk K D， Fr?bisch J， 2011. A comprehensive taxonomic revision of Dicynodon （Therapsida，Anomodontia） and its implications for dicynodont phylogeny， biogeography， and biostratigraphy. J Vert Paleont，31（sup1）： 1–158

Kammerer C F， Angielczyk K D， Fr?bisch J， 2015. Redescription of the geikiid Pelanomodon （Therapsida， Dicynodontia），with a reconsideration of ‘Propelanomodon’. J Vert Paleont， 36 （1）： e1030408

Keyser A W， 1993. A Re-evaluation of the Smaller Endothiodontidae. Pretoria： Government Printer. 1–53

King G M， 1981. The functional anatomy of a Permian dicynodont. Philos Trans R Soc Lond–Biol Sci， 291： 244–322

King G M， 1988. Anomodontia. Stuttgart： Gustav Fischer Verlag. 1–174

Li P X， Cheng Z W， Li J L， 2000. A new species of Dicynodon from Upper Permian of Sunan， Gansu， with remarks on"related strata. Vert PalAsiat， 38（2）： 147–157

Liu J， 2021. The tetrapod fauna of the upper Permian Naobaogou Formation of China： 6. Turfanodon jiufengensis sp. nov.（Dicynodontia）. PeerJ， 9： e10854

Olroyd S L， Sidor C A， 2022. Nomenclature， comparative anatomy， and evolution of the reflected lamina of the angular in"non-mammalian synapsids. J Vert Paleont， 42（1）： e2101923

Olson E C， 1944. Origin of mammals based upon cranial morphology of the therapsid suborders. Spec Pap， Geol Soc Am，55： 1–136

Owen R， 1859. On some reptilian remains from South Africa. Edinb N Philos J， 10： 289–291

Owen R， 1860. On the orders of fossil and recent Reptilia， and their distribution in time. Rep Br Assoc Adv Sci， 29： 153–166

Ray S， 2006. Functional and evolutionary aspects of the postcranial anatomy of dicynodonts （Synapsida， Therapsida）.Palaeontology， 49（6）： 1263–1286

Ray S， Chinsamy A， 2003. Functional aspects of the postcranial anatomy of the Permian dicynodont Diictodon and their"ecological implications. Palaeontology， 46（1）： 151–183

Romer A S， 1956. Osteology of the Reptiles. Chicago： University of Chicago Press. 1–772

Shi Y T， Liu J， 2023. The tetrapod fauna of the upper Permian Naobaogou Formation of China： 10. Jimusaria monanensis"sp. nov. （Dicynodontia） shows a unique epipterygoid. PeerJ， 11： e15783

Sun A L， 1963. The Chinese kannemeyerids. Palaeontol Sin， New Ser C， 17： 1–109

Sun A L， 1973. Permo-Triassic dicynodonts from Turfan， Sinkiang. Mem Inst Vert Paleont Paleoanthrop， Acad Sin， 10：53–68

Sun A L， 1978. Two new genera of Dicynodontidae. Mem Inst Vert Paleont Paleoanthrop， Acad Sin， 13： 19–25

Young C C， 1935. On two skeletons of Dicynodontia from Sinkiang. Bull Geol Surv China， 14（4）： 483–518

中國科學院戰(zhàn)略性先導科技專項（B類） （編號：XDB26000000）和中國科學院大學生創(chuàng)新實踐訓練計劃資助。