restudy of anshunsaurus huangguoshuensis (reptilia ... · the anterior margin of the orbit as in...

Copyright q American Museum of Natural History 2005 ISSN 0003-0082

P U B L I S H E D B Y T H E A M E R I C A N M U S E U M O F N AT U R A L H I S T O RY

CENTRAL PARK WEST AT 79TH STREET, NEW YORK, NY 10024

Number 3488, 34 pp., 13 figures, 2 tables October 27, 2005

Restudy of Anshunsaurus huangguoshuensis(Reptilia: Thalattosauria) from the Middle Triassic

of Guizhou, China

JUN LIU1 AND OLIVIER RIEPPEL2

ABSTRACT

We here describe the complete skeletal anatomy of Anshunsaurus huangguoshuensis thatpreviously remained unknown. The description includes a review of the dorsal side of theskull as well as the first description of the ventral side of the skull and of the postcranialskeleton. A new phylogenetic analysis supports the monophyly of Askeptosauridea and Tha-lattosauridea; Endennasaurus is found to be closely related to Askeptosauridae.

INTRODUCTION

Thalattosaurs are a monophyletic clade ofTriassic marine reptiles with modest special-ization toward an aquatic life (Nicholls,1999). The phylogenetic position of thalat-tosaurs within amniotes is controversial:They have been suggested to be diapsidswith possible affinities to the Lepidosauro-morpha (Romer, 1956; Rieppel, 1998), Ar-chosauromorpha (Evans, 1988), to be Neo-diapsida inc. sed. (Benton, 1985), or else to

1 Division of Paleontology, American Museum of Natural History; Lamont-Doherty Earth Observatory, ColumbiaUniversity, Palisade, New York, NY 10964 ([email protected]).

2 Department of Geology, The Field Museum, 1400 S. Lake Shore Drive, Chicago, IL 60605-2496 ([email protected]).

be the sister taxon of Sauria (Muller, 2004)or of Ichthyopterygia (Muller, 2003).

Thalattosaurs are first known from the Up-per Triassic of California, North America,represented by Thalattosaurus and Nectosau-rus (Merriam, 1904, 1905, 1908). Thalatto-saurus is also known from the Middle Tri-assic of British Columbia. Two additional,but very incompletely known taxa, Agkis-trognathus and Paralonectes, as well as anas yet undetermined form have been reported

2 NO. 3488AMERICAN MUSEUM NOVITATES

from the same deposits in British Columbia(Nicholls and Brinkman, 1993). To the pres-ent day, Askeptosaurus from the Grenzbitu-menzone (Anisian/Ladinian) of Monte SanGiorgio, Switzerland/Italy (Nopsca, 1925;Kuhn-Schnyder, 1952), remains the bestknown thalattosaur (Muller, 2002, in press).In addition, the Grenzbitumenzone of MonteSan Giorgio yielded the two monotypic gen-era Clarazia and Hescheleria (Peyer, 1936a,1936b). Blezingeria from the Ladinian ofGermany may be a thalattosaur (Rieppel,1998), but a definite assignment has to awaitadditional material (Muller, 2002). Enden-nasaurus from the Norian of Italy was iden-tified as a thalattosaur when first described(Renesto, 1992), an assessment that was con-firmed by Muller et al. (2005). Many well-preserved thalattosaur specimens have beendiscovered in Guizhou, China; Anshunsaurusand Xinpusaurus are erected based on thesespecimens (Liu, 1999; Rieppel et al., 2000;Yin et al., 2000; Liu and Rieppel, 2001).Moreover, there are several fragmentary rec-ords of thalattosaurs from the Ladinian ofSpain (Rieppel and Hagdorn, 1998), the Car-nian of the Julian Alps (Dalla Vecchia,1993), the Norian of Austria (Muller, 2002),the Anisian and Carnian of Nevada (Sanderet al., 1994), and the Norian of British Co-lumbia (Storrs, 1991).

The lack of consensus on the phylogeneticrelationships among thalattosaurs is not onlydue to character incongruence, but also a re-sult of fragmentary preservation of fossils orincomplete anatomical knowledge of the well-preserved specimens. For example, only thedorsal view of the skull of Anshunsaurus haspreviously been described (Liu, 1999; Rieppelet al., 2000). Here we describe the completeskeletal anatomy of Anshunsaurus and discussits relationship with other thalattosaurs.

The Triassic marine reptile Anshunsaurushuangguoshuensis was originally describedby Liu (1999) as a sauropterygian on the ba-sis of dorsal view of a skull. The specimenwas recognized as thalattosaurian by Rieppelet al. (2000). Only later were the ventral sideof the skull and the postcranial skeleton ofthe holotype prepared, as well as an addi-tional specimen made available for study.

INSTITUTIONAL ABBREVIATIONS: GMPKU,Geological Museum of Peking University;

Beijing. Gmr, GGSr, collections at the Geo-logical Survey of Guizhou, Guiyang. IVPP:Institute of Vertebrate Paleontology and Pa-leoanthropology, Chinese Academy of Sci-ences; Beijing. PIMUZ, PalaontologischesInstitut und Museum der Universitat Zurich.SPC, Center of Stratigraphy and Paleontol-ogy, Yichang. UCMP, University of Califor-nia Museum of Paleontology, Berkeley.

MATERIAL: IVPP V11835 (holotype), askeleton lacking most of the hind limbs andmost of the caudal vertebrae. The specimenis composed of seven parts, viz., the head(11835-1) (fig. 1), the neck (11835-2) (fig.4), the main body of the trunk and the fore-limbs (11835-3) (fig. 5), the posterior part ofthe trunk (11835-4) (fig. 6), a few caudal ver-tebrae (11835-5) (fig. 7), and the distal partof a fibula (11835-6). The first three blocksare consecutive; blocks 11835-4 and 18835-5 are also consecutive, but some bones aremissing between them. The calcite vein thatcuts across block 11835-4 separates the leftside, which is offset posteriorly relative tothe right side. IVPP V11834 (fig. 3) is anarticulated, perfectly preserved skeleton,with only the last few caudal vertebrae anda few phalanges missing.

SYSTEMATIC PALEONTOLOGY

THALATTOSAURIA MERRIAM, 1904

ASKEPTOSAURIDAE KUHN, 1952

ANSHUNSAURUS LIU, 1999

Anshunsaurus huangguoshuensis Liu, 1999

DIAGNOSIS (revised): An askeptosauridthalattosaur differing from Askeptosaurus bythe maxilla forming part of the anteroventralorbital margin; fusion of the postorbital andpostfrontal; the posterolateral process of thefrontal extending posteriorly far beyond theanterior margin of lower temporal fossa, nar-rowly approaching but not contacting the su-pratemporal; the long and slender ventralprocess of the squamosal extending to thelower margin of the cheek; jugal with anelongate posterior process; lateral exposureof the angular equal to that of the surangular;deltopectoral crest on the humerus devel-oped; fibula expanded.

LOCALITY AND HORIZON: Wayao Member,Falang Formation, Ladinian or Carnian (up-

2005 3LIU AND RIEPPEL: TRIASSIC MARINE REPTILE

per Middle or lower Upper Triassic), Xinpu,Guanling County, Anshun area, GuizhouProvince, southwestern China.

MORPHOLOGICAL DESCRIPTION

The two specimens are similar in size, butthe holotype IVPP V11835 is a little bit lon-ger than IVPP V11834. The holotype ismainly exposed in dorsal view, whereas theskull, pectoral girdle, and pelvic girdle arealso prepared from the ventral side. Speci-men IVPP V11834 is mainly preserved inventral view with most vertebrae in lateralview. All the description of Askeptosaurusrefers to Muller (2002, in press) other thanspecific mention.

SKULL

The description of the skull is almost ex-clusively based on the holotype (IVPPV11835) (fig. 1). The skull is slightly dor-soventrally compressed. The anterior partand two sides of the skull in ventral side arecovered by the closely attached lower jaw.

PREMAXILLA

The paired premaxillae are exposed in dor-sal view only. They are slender and ratherelongated elements, extending over morethan half of the entire skull length as in As-keptosaurus. They compose the main part ofthe rostrum. The lateral margins of the ros-trum gradually assume a straight and parallelcourse before the snout terminates in a blunttip. The premaxillae meet in a straight dor-somedial suture. They form tapering poste-rior processes that contact the anteromedialprocesses of the frontals, thus separating thenasals from each other.

Although the anterior margin of the exter-nal nares is obscured through breakage, it isconceivable that the premaxilla forms the an-terior margin of the external naris as in prim-itive diapsids generally. The left side of theskull shows the premaxillary–maxillary su-ture to extend from the anteroventral (antero-lateral) margin of the external naris in an an-terolateral direction. A few straight, longitu-dinally directed striations are distinct on thedorsal surface of the rostrum.

The left and right premaxillae bear eight

to nine teeth respectively. The height of theteeth ranges from 14 mm to 19 mm (about17 mm); the width at the base ranges from 6mm to 7 mm. The tip of the premaxillae isedentulous. Tooth implantation is subtheco-dont; the teeth are unicuspid and apically re-curved. The enamel surface of each tooth isstriated toward the apical area. The thirdtooth on the right premaxilla is remarkablesmaller than all other teeth; it probably rep-resents a newly erupted replacement tooth. Inspecimen IVPP V111834, the anteriormostpremaxillary tooth is smaller than all the suc-ceeding ones.

MAXILLA

Both maxillae are exposed in dorsal viewonly. They are slender and elongated ele-ments, relatively small in comparison to thepremaxillae. The maxilla extends anteriorlyto a level well in front of the external narisbefore tapering to a point along the lateral(ventral) margin of the rostrum. It formsmost of the ventral (lateral) margin of theexternal naris, and contacts the nasal by wayof a short ascending process directly behindthe external naris. This contact separates theprefrontal from the posteromedial border ofthe external naris. Posteroventrally, the max-illa extends into a process that follows theanteroventral (anterolateral) margin of theprefrontal and meets the suborbital processof the jugal in the anteroventral corner of theorbit. In contrast to Askeptosaurus, the pos-terior end of the maxilla enters the orbitalrim.

The maxillary teeth are incompletely ex-posed in specimen IVPP V11834 (fig. 2).The number of functional teeth is estimatedas seven; the exact number cannot be deter-mined due to the obscure premaxillary–max-illary suture. Tooth implantation again issubthecodont. The shape of the maxillaryteeth resembles those of the premaxillary,and their size is similar except for few mostposterior teeth.

NASAL

The small nasals are exposed in dorsalview, situated close to the midline of theskull. The nasals are slender and elongatedelements, which define the posterior and dor-


Fig. 1. Anshunsaurus huangguoshuensis, head and first three cervical vertebrae in dorsal and ventralview (holotype, specimen no. IVPP V11835-1). See appendix 1 for anatomical abbreviations.


Fig. 1. Continued.

sal (medial) margin of the external nares.Laterally, the nasal meets the ascending pro-cess of the maxillae and the anterior end ofthe prefrontals in a more or less straight, an-terolaterally trending suture. Medially, the

nasal contacts the premaxillary and the fron-tal. Anteriorly, the nasals are separated by thepremaxillae. The pointed posterior tip of thenasal is embraced by the anterolateral andanteromedial processes of the frontal. The


Fig. 2. Anshunsaurus huangguoshuensis,skull and lower jaw in ventral view (specimen no.IVPP V11834).

nasals extend posteriorly to a level behindthe anterior margin of the orbit as in Askep-tosaurus.

PREFRONTAL

The triangular prefrontal, exposed in dor-sal view, forms the anterodorsal (anterome-dial) margin of the orbit. Its anteroventralmargin meets the maxilla, its dorsal contactis with the frontal, and anteromedially itmeets the nasal. The anterior tip of the pre-frontal is rounded and short, whereas its pos-terior process, which extends along the dor-sal margin of the orbit, is more prominentlydeveloped. Although the prefrontal closelyapproaches the postorbitofrontal along thedorsal (medial) margin of the orbit, the twoelements remain separated from one another.

The prefrontals as described here are larg-er than indicated by Rieppel et al. (2000) asthey extend further ventrally along the ante-rior margin of the orbit. The extension of theprefrontal equals that of the prefrontal plusthe lacrimal in Askeptosaurus as describedby Muller (2000, in press). Although the crit-ical area is subject to some breakage, weconclude that a lacrimal is absent in Anshun-saurus.

FRONTAL

The paired slender frontals are exposed indorsal view. Each of the frontals forms ananterolateral and an anteromedial process,both of which taper to a pointed tip; the an-teromedial process of the frontal is longerthan the anterolateral process. The antero-medial processes of the frontals, terminatingslightly in front of the midpoint between theexternal nares and the orbits, together em-brace the pointed posterior tips of the pre-maxillae. In Askeptosaurus (Muller, 2002, inpress), the frontals reach forward to the levelof the external naris, although they do notcontact the external nares. The anterolateralprocesses of the frontals enter between thenasals and the prefrontals, extending forwardnot quite as far as the anterior tips of thenasals.

The paired frontals meet each other in astraight suture and they show a slight lateralexpansion in their middle part between theorbits. The frontal is nearly excluded from


the orbital margin by the prefrontal and post-orbitofrontal. The frontal meets the postor-bitofrontal in a straight suture. Each frontalforms a slender posterolateral process, whichextends backward between the postorbito-frontal and parietal to a level narrowly be-hind the posterior margin of the pineal fo-ramen. The posterolateral process of thefrontal narrowly fails to contact the anteriortip of the supratemopral as in Askeptosaurusand in primitive dispsids generally. In itsmiddle portion, the posterior end of the fron-tal meets the parietal bone in a complex in-terdigitating suture at the level of the post-orbital arch. The suture with parietal remainsunclear on the right side of the skull, how-ever.

POSTORBITOFRONTAL

In contrast to Askeptosaurus, the postor-bital and postfrontal are fused. The postor-bitofrontal forms the posterodorsal (postero-medial) and most of the posterior margin ofthe orbit. Posterodorsally, the postorbitofron-tal draws out into a long and slender process,which extends far back within the skull roof.It meets the mediolateral margin of the pa-rietal between the frontal and supratemporal,and covers the anterior process of the squa-mosal dorsally. Its posterior end overlapswith the anterior end of the supratemporal.The tapering ventral process of the postor-bitofrontal obliquely overlaps the dorsal pro-cess of the jugal within the postorbital arch.

JUGAL

Both jugals are exposed in dorsal, lateral,and ventral views. They are distinctly trira-diate bones as in all other thalattosaurs. Along anterior (suborbital) process forms theventral (lateral) margin of the orbit, meetingthe posterior process of the maxilla. Thepointed dorsal process enters the postorbitalarch, partially defining the anterior margin ofthe lower temporal fenestra. The jugal of An-shunsaurus is peculiar as it draws out pos-teriorly into a long and slender process that,at least to a large degree, closes the lowertemporal fenestra ventrally. Although theposterior processes of both jugals is brokenin the holotype, both symmetrically dis-placed anteromedially, they still seem long

enough to have closely approached the quad-rate in the articulated skull.

PARIETAL

The paired parietals are exposed in dorsalview. The bones meet each other in a straightdorsomedial suture. The parietal contacts theposterior process of the postorbitofrontal be-tween the frontal and the supratemporal. Theupper temporal fossa is completely obliter-ated. The parietals increasingly expand in aposterolateral direction, forming elongate su-pratemporal processes that form the posteriormargin of the deeply excavated occiput.

The parietals enclose a large pineal fora-men that is somewhat displaced anteriorly,located narrowly behind the frontoparietalsuture. The size of pineal foramen differs indifferent specimens of Askeptosaurus (Mull-er, 2002, in press), but even in the largestspecimen the pineal foramen is smaller thanthat in the holotype of Anshunsaurus (Muller,2002).

SUPRATEMPORAL

The supratemporals are broadly exposed atthe posterolateral corner of the skull table indorsal view. They form relatively elongatedslivers of bone. Anteriorly, the supratemporalis reduced to a narrow process, which ex-tends anteriorly between the parietal and thepostorbitofrontal. Ventrally, the supratempor-als contact the squamosal, and possibly alsothe quadrate.

SQUAMOSAL

Both squamosals have been pushed outlaterally during fossilization, and hence ex-pose their lateral side in the dorsal view ofthe specimen. They are hockey-stick shapedand generally slender. The squamosal con-sists of two processes, an anterior one and aventral one. The anterior process follows thelateral margin of the skull table, participatingin the formation of the upper margin of thelower temporal fenestra, underlapping theposterior process of the postorbitofrontal. Ittapers to a pointed tip at a level ahead of theposterior margin of the parietal foramen. Theventral process is much elongated and slight-ly recurved. It must have lined the lateral


edge of the posteriorly concave shaft of thequadrate, extending to the ventral margin ofthe cheek and closely approaching the pos-terior tip of the jugal. The posterior processof the jugal must have closely approachedthe ventral tip of the squamosal in the artic-ulated skull.

QUADRATE

Both quadrates are preserved in articula-tion, but they are partially obscured byneighboring skull elements in both speci-mens due to the dorsoventral compression ofthe skull. The quadrate is a relatively largeand robust element with a distinctly thick-ened shaft but thinning along the margins.The cephalic condyle is located near the ven-trolateral margin of the parietal table, sus-pended in life from the squamosal and su-pratemporal. In dorsal view, the right quad-rate shows an anterior process extendingfrom the medial margin of the cephalic con-dyle. It must represent the dorsal margin ofan extensive anterior quadrate flange, whichis also known in other thalattosaurs. The me-dial lamina is fan shaped and thinner to bor-der. The mandibular condyle is hidden by thelower jaw.

SCLERAL OSSICLES

No scleral ossicle is observed in thesespecimens.

VOMER

Only the posterior part of the vomers isexposed in palatal view. Their preservationsuggests that they are slender and elongateelements. A clear groove separates the pairedvomers from one another (the vomers arefused in Askeptosaurus; Muller, 2002, inpress). The pterygoids enter between the pos-terior parts of the vomers, thus separating thetwo palatines from one another. A concavearea lies behind the vomers, lateral to thepterygoid and anterior to the palatine; thisdepression would have been part of the cho-ana (internal naris). There is no indication ofteeth on the vomers.

PALATINE

The palatines are poorly known; theirshape cannot be determined. The two pala-

tines are separated along the ventral midlineof the skull by the contact of the vomers withthe pterygoids. The way that the palatine en-ters the internal naris remains unclear, asdoes the general shape of the choana. Pos-teriorly, the palatine forms the anterior mar-gin of a large suborbital fenestra, but the ar-rangement of the palatal elements around thelateral and posterior margins of this fenestraremain unclear. There are no palatine teeth.

PTERYGOID

The pterygoids are well preserved and ex-posed in ventral view. They are large ele-ments, meeting anteriorly in front of the nar-row interpterygoid vacuity along the midlineof the skull and becoming progressively nar-rower as they approach their contacts withthe vomers. The length of the plate ramus isnearly twice the length of the quadrate ra-mus. The transverse process of the pterygoidis conspicuous but partially obscured by thelower jaw and the left hyoid; the suture withthe ectoperygoid cannot be observed. Thepterygoid contributes to the medial and pos-terior margins of the suborbital fenestra, butits exact contribution to this opening couldnot be determined. A low but sharp crest pro-jects ventrally from the corner between thetransverse process and the quadrate ramus,extending into the posterolateral direction.This crest extends backward to a level infront of the articulation of the pterygoid withthe basipterygoid process. It must haveserved as a site of origin for parts of the ptery-goideus muscle. Posteriorly, the pterygoidtapers significantly and forms the quadrateramus. The quadrate ramus carries a distinct,ventrally projecting flange. The quadraterami of the pterygoids diverge posteriorly asthey extend backward from the palatobasalarticulation, enclosing an acute angel ofabout 608. The posterior part of the rightquadrate ramus is obscured by overlappingskull elements, but the left one is completeand is applied to the medial side of the quad-rate; the posterior tip of the pterygoid musthave formed a medial contact with the quad-rate just above the mandibular condyle. Thetransverse process and quadrate ramus of thepterygoid flange define the anteromedial and


medial margin of the large subtemporal fe-nestra. There are no pterygoid teeth.

ECTOPTERYGOID

The ectopterygoid is mostly obscured bythe lower jaw.

PARABASISPHENOID

This element is formed by the fusion ofthe dermal parasphenoid to the ventral sur-face of the basisphenoid. It is only exposedin ventral view. Anteriorly, the bone appearsto extend into a long and narrow cultriformprocess that entered the narrow interptery-goid vacuity. As preserved, the cultriformprocess appears to lie close to the left pter-ygoid, with its anteriormost portion obscuredfrom ventral view. The base of the cultriformprocess is slightly expanded. Two prominentdrop shaped ventral expansions lie postero-laterally to the base of the cultriform process,representing the basipterygoid processes. Be-tween them, two slender grooves extend in aposterolateral direction, lined by small me-dial crests. These grooves represent the ven-trally open Vidian or parabasal canals thatcarry the internal carotid artery as well as thepalatine branch of the facial nerve. A fora-men located close to the posteromedial edgeof the basipterygoid processes served as en-try foramen for the cerebral branch of theinternal carotid, as it has also been describedfor Askeptosaurus (Muller, 2002, in press).The ventral surface of the parabasisphenoidshows a shallow triangular depression thatextends to the posterior edge where the bonemeets the basioccipital. Posterolaterally, theparabasisphenoid is drawn out into twoprominent projections. They are broad andwith a relatively sharp distal tip, originallyunderlying the fenestra vestibuli. The suturebetween the parabasisphenoid and the basi-occipital is not very clear in the middle partof the basicranium, but the basisphenoid tu-bercle is not fused with the pronounced ba-sioccipital tubercle.

BASIOCCIPITAL

The basioccipital is again only exposed inventral view. It is a broad element with amore or less trapezoid shape similar to that

seen in Askeptosaurus (Muller, 2002, inpress). Anteriorly, the basioccipital meets theparabasisphenoid in what appears to be anundulating suture. In its anterolateral part,the basioccipital bears prominent knoblikestructures, the basal tubera. They are ventro-laterally directed, their medial surface beingoblique and their lowermost tips formingsemilunar edges. The middle portion of theventral surface of the basioccipital is de-pressed. More posteriorly, the basioccipitalshows a ventral expansion with a slightlyconvex posterior edge. This expansion rep-resents the contribution of the basioccipitalto the formation of the occipital condyle.Laterally, the basioccipital bears a prominentconcave facet between the posterior marginof the basal tubera and the anterolateral edgeof the posteroventral expansion. The facetserves for the contact of the proximal end ofthe opisthotic.

SUPRAOCCIPITAL

The supraoccipital is only partially ex-posed in dorsal view. It is seen to projectfrom below the posterior margin of the pa-rietals, carrying a low sagittal crest.

PROOTIC

The anterior of the part right prootic is ex-posed in palatal view, whereas its posteriorpart is coved by the quadrate. It contacts thelateral margin of the parabasisphenoid, po-sitioned between it and the medial margin ofthe quadrate process of the pterygoid.

OPISTHOTIC AND EXOCCIPITAL

The right opisthotic is well exposed inventral and dorsal view. This bone is fusedwith the exoccipital and forms the well-de-veloped paroccipital process. Its medial headarticulated with the concave facet of the ba-sioccipital described above. Its posteriormargin is exposed below the supraoccipitaland the parietal in dorsal view.

STAPES

No stapes can be identified.

LOWER JAW

The lower jaws are mainly exposed in lat-eral view, with only a small portion of the


medial side exposed. The slender mandibularramus has an elongated and straight tooth-bearing portion, whereas the posterior part iscomparatively short and slightly curved.There is no true retroarticular process.

DENTARY

The dentary is the longest element in thelower jaw. It is straight and becomes increas-ingly expanded posteriorly. It covers the an-terior lateral portion of the mandibular ra-mus, but the splenial could also have beenexposed in lateral view. Two dentaries meetanteriorly along approximately one-third oftheir entire length, forming an elongatedmandibular symphysis together with thesplenials. The mandibular symphysis of den-taries tapers to a narrow tip anteriorly. Pos-teriorly, the dentary bifurcates into a shortposterodorsal and an elongated posteroven-tral process, both terminating in a pointed tip.The posterodorsal process overlaps part ofthe lateral surface of the surangular; the re-maining posterior margin of the dentary con-tacts the surangular and the angular. The al-veolar margin of the dentary is covered bythe skull; only four conical teeth are ob-served in the holotype.

SPLENIAL

The splenial is only exposed in ventralview. It is an elongated element located an-teroventrally, but its posterior end reachesthe position of the coronoid process, levelingwith the anterior border of the supratemopralfenestra. Its anteriormost portion contributesto the formation of the symphysis. The splen-ial forms the ventral margin of the lower jawramus in front of the angular and below thedentary.

CORONOID

The coronoid can only be identified on thelateral side of the right mandible of specimenIVPP V11834. Given uncertainties about thedetailed sutural contacts, two alternative in-terpretations obtain. In the first interpretation,the coronoid is an elongated element, meet-ing the dentary anteriorly and the surangularposteriorly. In the second interpretation, it isshorter and in contact with the surangular

only, albeit in a pattern that is comparable toother thalattosaurs. For that reason, we preferthe latter of the two interpretations. The gen-eral shape of the coronoid is that of an elon-gated triangle with a moderately developedcoronoid process. This bone might also beexposed under the orbit in holotype, but thisis difficult to ascertain.

ANGULAR

The angular is a relatively elongated bone,defining the ventral margin of the posteriorpart of the mandibular ramus. Anteriorly, itborders the dorsal edge of the posteroventraldentary process. The straight dorsal marginof the angular meets the surangular along itsentire length. Posteriorly, the angular isslightly curved dorsally and shows a taperingend, which contacts the prearticular on theventral side, level with the anterior border ofthe articular.

SURANGULAR

The surangular covers the posterodorsalpart of the lower jaw ramus. It also is anelongated element. Anteriorly, it meets boththe posterior processes of the dentary; its an-terodorsal margin supported the coronoid.Posteriorly, the surangular closely approach-es the posterior end of the lower jaw, ter-minating in a contact with the articular. Onlya short part of the surangular participates inthe formation of the ventral margin of thelower jaw.

PREARTICULAR

The prearticular would be restricted to themedial side of the posteroventral part of themandibular ramus, but it gained lateral ex-posure due to the deformation of the lowerjaw. Its ventral margin contacts the angularand the surangular; posterodorsally, it con-tacts the articular.

ARTICULAR

The articular expands over the wholewidth of the posterior end of the lower jaw,covering the entire dorsal surface of the low-er jaw between the prearticular and the sur-angular. There are no signs of fusion of thearticular with the prearticular. The foramen


→

Fig. 3. Anshunsaurus huangguoshuensis,specimen no. IVPP V11834.

for the chorda tympani lies close to the su-ture with the surangular.

HYOID

A pair of slender hyobranchial elements,identified as first ceratobrachchials, are pre-served in both specimens and exposed inventral view. The position of ceratobranchi-als is similar in two specimens, that is, to thelateral side of the quadrate. The left hyoidpartially covers the transverse process of thepterygoid, the subtemporal fenestra, and theventral margin of the lower jaw; the right oneis pushed against the paroccipital process andextends backward to the third cervical ver-tebra. The hyoid is constricted in its middlepart, with slightly expanding anterior andposterior ends.

AXIAL SKELETON

Anshunsaurus bears at least 38 presacral,2 sacral, and more than 50 caudal vertebrae(figs. 3–7). The 35th presacral vertebra isrepresented by part of the neural spine onlyin the holotype (fig. 5), and it is separatedfrom the succeeding one by a narrow gap,such that the number of presacral vertebraemight be 39. The cervical vertebrae are ex-posed in both specimens, but the dorsal ver-tebrae are completely concealed by the gas-tral ribs in IVPP V11834, and only a fewcaudal vertebrae are observed in the holo-type.

CERVICAL REGION

Muller (2002, in press) used the morphol-ogy of the ribs to demarcate the cervicalfrom the dorsal region in Askeptosaurus, andthe same method also is used here. The num-ber of the cervical vertebrae is at least 15,because the 15th vertebra still is articulatedwith a double-headed rib in the holotype. Itis difficult to assess whether the 16th verte-bra, located immediately in front of the clav-icle, is the last cervical or first dorsal becauseit is incompletely preserved in the holotype


Fig. 4. Anshunsaurus huangguoshuensis, cervical vertebrae (A) in dorsal view (specimen IVPPV11835-2) (B) in lateral view (specimen IVPP V11834).

and concealed in specimen IVPP V11834.The length of the cervical region is 49 cm inthe holotype. Almost all the cervical verte-brae in the holotype are exposed in dorsalview, but the atlas, axis, and the 3rd and 14thto 16th vertebrae are also prepared in ventralview; in IVPP V11834 the cervical vertebraeare exposed in right lateral view.

The relationship of the various compo-nents of the atlas–axis complex can be ob-served in ventral view in the holotype. Theatlas is ring shaped to match the outline ofthe occipital condyle. A pair of cervical ribs

articulates with the lateral aspect of the atlas.The intercentrum is of roughly rectangularshape in ventral view. The morphology ofthe atlas neural arch remains uncertain. Avertebral element that lies posterior to the in-tercentrum, and appears to project dorsal be-yond the latter, is identified as the first (atlas)centrum. It is a half-saddle-shaped bone withclear margins setting it off from neighboringanterior and posterior elements. This boneseems not to be preserved in Askeptosaurus.Kuhn (1952) noted a possible centrum lyinganteriorly to the atlas neural arches, but it is


Fig. 5. Anshunsaurus huangguoshuensis in dorsal view (specimen IVPP V11835-3). See appendix1 for anatomical abbreviations.


Fig. 6. Anshunsaurus huangguoshuensis in dorsal view (specimen IVPP V11835-4). See appendix1 for anatomical abbreviations.


Fig. 7. Anshunsaurus huangguoshuensis, an-terior caudals (IVPP 11835-5 and part of IVPP V11835-4) in dorsal view.

determined to be a broken part of the brain-case (Muller, 2002). The proatlas, which wasreported in Askeptosaurus (Muller, 2002, inpress), cannot be identified in Anshunsaurus.

The axis is a prominent, elongated ele-ment. The axis centrum is longer than high,and shows a distinctly longitudinal keel ex-tending along the midline of its ventral sur-face. A pair of double-headed ribs articulateslaterally. The neural arch of the axis showsweakly developed pre- and postzygapophy-sis. The neural spine is low and distinctlyelongated; its dorsal margin is straight.

The remaining cervical vertebrae are rath-er small elements carrying a relatively lowneural spine. They slightly increase in size inan anteroposterior direction. The centrum isslightly constricted in its central part andgenerally longer than high. It may show asemilunar depression on its ventrolateral as-pect that faces ventrally. A distinct ventralkeel is present on the third centrum, whichin the last few cervicals is reduced to a faintridge. The remaining lateral part of the cen-trum consists of a concave curvature.

The base of the neural arch of the cervicalvertebrae forms relatively stout pre- andpostzygapophyses. The prezygapophyses aresomewhat more prominently developed, andcover the postzygapophyses in lateral view(fig. 4). They show a broad base, and where-as the prezygapophyses face dorsomedially,the postzygapophyses taper only slightly ina posterior direction and are nearly horizon-tally orientated. The neural spine is of rough-ly rectangular shape. The length and theheight of the neural spine show some in-crease from the 3rd to the 16th vertebra.

Muller (2002, in press) identified pairedrib articulations on the cervical vertebrae ofAskeptosaurus, comprising a diapophysis forthe tuberculum and a parapophysis for thecapitulum. The diapophysis lies directly be-low the neural arch on the dorsalmost part ofthe centrum, more or less in the middle ofthe vertebra, and it consists of an irregularlyrounded or rectangular depression. The par-apophysis is located on the central aspect ofthe centrum and is somewhat more posteri-orly directed than the diapophysis. In An-shunsaurus, only a diapophysis with thesame shape and position as seen in Askep-tosaurus can be identified on the cervicals,but no parapophysis (fig. 4). The capitulumarticulated with the anteroventral aspect ofthe centrum. This position is close to theoriginal position of the intercentrum, whichmay be the reason why Kuhn (1952) sug-gested that the capitulum articulated with acartilaginous intercentrum in Askeptosaurus.

The cervical ribs articulate with vertebraebeginning with the atlas. It remains unknownwhether the ribs articulating with the atlasare dichocephalous or holoceophalous. Theyform a small and slender rod of bone with aslightly expanded anterior head, tapering ina posterior direction and terminating in asharp tip. Dichocephalous cervical ribs arepreserved in articulation with the axis andthe 3rd and 15th cervical in the holotype, andfrom the 3rd to the 10th cervical in V11834.The dorsal tuberculum is broader than theventral capitulum. The cervical ribs becomeincreasingly more elongated posteriorly. Theposterior ends of the third, the fourth, and thefifth ribs do not project beyond the posteriormargin of the respective centrum, whereasmore posterior ribs overlap with the anteriormargin of the succeeding centrum.

DORSAL REGION

In specimen V11834, the vertebrae of thedorsal region are almost completely con-cealed by the ribs and gastralia, such that thefollowing description is mainly based on theholotype. The vertebrae expose their dorsalto lateral side from anterior to posterior. Thenumber of dorsals is minimally 23. No in-tercentra are identified in the dorsal region.

The length of the centra is fairly constant


throughout the dorsal region, with just theanteriormost vertebrae being slightly smaller.The length of the centrum ranges from 33mm to 36 mm (with an average of about 35mm). All dorsal vertebrae are larger than anyof the cervicals, mainly as a result of the dor-sal extension of the neural spine. The syna-pophysis is anteroventrally directed, origi-nating in the midventral region of the neuralarch and then extending down to near theanterior margin of the centrum (figs. 5, 6).The attachment site develops into a promi-nent process anteriorly, and posteriorly is re-duced to a small knob. The centrum is gen-erally longer than high, its middle part beingslightly constricted.

The neural arch is relatively long and fullyfused with the centrum. The prezygapophys-es are anteromedially directed. Their dorsaledge is unusually straight, whereas the dis-talmost margin is slightly recurved dorsally.The postzygapophyses are distinctly shorterand not as broad as the prezygapophyses, butthe dorsal recurvation of the distalmost tip isdistinct as well. Below the zygapophyseal ar-ticulation between the succeeding vertebraelie the distinct intervertebral foramens withdistinct triangular outline. The neural spinesare broad and trapeziform in shape, expandedin the dorsal direction. They are more or lessvertically oriented. The last few neural spinesare relatively narrow and higher than the pre-ceding ones.

All the dorsal vertebrae bear ribs, includ-ing the vertebrae immediately preceding thesacrum; there is no true lumbar vertebra. Themorphology of the dorsal ribs differs dis-tinctly from that of the cervical ribs. Theproximal head of the dorsal ribs is holoce-phalous and slightly expanded. With the ex-ception of the posteriormost ribs, all dorsalribs are similar in size (about 16 cm inlength) and notably longer than the cervicalones. All ribs are strongly curved and showa slight distal expansion. The last four dorsalribs decrease abruptly in size, with the lasttwo dorsal ribs being the shortest; they areslender and point almost straight laterally.

A series of gastralia is partially exposedbelow ribs in the holotype, whereas they arewell exposed in IVPP V11834 in ventralview, in the area between the pectoral andpelvic girdle. Gastralia can only be observed

behind the 24th vertebra in the holotype andbehind the posterior tip of the interclaviclein IVPP V11834. They appear to be slenderelements, being slightly expanded mediallyand tapering laterally. Both halves overlapeach other in the ventral midline of the body.More than 35 and 50 pairs of gastralia arepreserved in the holotype and IVPP V11834,respectively, indicating the presence of onepair of gastral ribs per body segment.

SACRAL REGION

Two sacral vertebrae are present in An-shunsaurus (fig. 6). There is no sign of anyfusion between the vertebrae, nor do the sa-cral ribs fuse to the centra or the neural arch-es. The sacral vertebrae are similar in shapeto the preceding dorsals, but the length of thecentrum of the sacral vertebrae is somewhatlonger than that of the dorsal vertebrae, andthe neural spines are reduce in height. Thesacral rib meets the vertebra in the lower halfof the centrum.

The sacral ribs are not fused to the sacralvertebra. They are expanded, especially thedistal portion, and fan shaped, contacting theilium with their distal margin. The left sacralribs measure 57 mm (anterior) and 59 mm(posterior) in length, 32 mm (anterior) and35 mm (posterior) in width of the distal end,and 12 mm (anterior) and 23 mm (posterior)in width of the proximal end.

CAUDAL REGION

Only the anteriormost six caudal vertebraeare preserved in the holotype (fig. 7), where-as 50 are exposed in lateral view in IVPPV11834 (fig. 3). The anterior caudal verte-brae bear ribs, which do not fuse with thevertebra in the anterior six caudals. The ar-ticular facet for the ribs represents a half-rounded lateral depression on the centrum.Where the ribs are fused with the vertebrae,they appear like transverse processes; moreposteriorly, the causal ribs abruptly reduce toa small knob and completely disappear fromthe 12th caudal onward. In more posteriorcaudals, where no ribs attach, the centrashow a vertical depression in the mid region,running from the ventrolateral margin up tothe dorsal edge, in an otherwise smooth lat-eral surface.


The length of all centra is relatively con-stant within the caudal region, around 32 mmin IVPP V11834. However, the height of thecentra gradually decreases in an anteropos-terior direction. In the anterior caudals, thecentrum appears square in lateral view, as itslength approximates its height. More poste-riorly, the centra become relatively slender astheir height decreases.

The neural arches in the anterior caudalsare relatively similar to the condition of thedorsals. They are comparatively high andwith stout pre- and small postzygapophyses.The neural arches of the more posterior cau-dals differ in many respects. The ventral baseis not that high, and almost directly abovethe contact to the centrum the prominent pre-zygapophysis originates. It is distinctly moreelongated than in the anterior portion of thetail and is anterodorsally directed. Converse-ly, the postzygapophysis remains very small,representing only a small, posterolaterallyprojecting triangle.

The neural spines are not as broad as thedorsal neural spines; rather they only reachtwo-thirds of the width of the latter at max-imum. The neural spine becomes increasing-ly shorter in the middle and posterior portionof the tail. Furthermore, they become moreslender and the orientation changes frommore or less dorsal to posterodorsal.

Beginning with the fourth caudal (figs. 3,7), haemal arches are present between ver-tebrae. No haemal arch contacts with thecentra behind of the 13th caudal, and mosthaemal arches are not preserved in IVPPV11834. The haemal arch contacts both cen-tra but only articulates with the posteroven-tral margin of the preceding centrum. Thehaemal arches are posteroventrally directedand V shaped, with the dorsal end being dis-tinctly expanded. The two dorsal knobliketips show small articulation surfaces, where-as the ventral tip of the arch is elongated andrectangular in shape.

At least six pairs of caudal rib are presentin Anshunsaurus. The border between theribs and the vertebrae is clear even if someribs seem to fuse to the vertebrae. The firsttwo pairs of ribs in the proximal caudal re-gion are short and broad, but taper distally,showing a posterior recurvation. The ribs de-crease in size from front to back. In the ho-

lotype, the following ribs are more taperinginto the distal direction, and their straightmargins form a triangular shape in dorsalview. In IVPP V11834, all these caudal ribsshow more or less recurvation and less ta-pering. They are generally slender. All caudalribs point anterolaterally.

PECTORAL GIRDLE

The elements of the pectoral girdle areprepared mainly from the ventral (lateral)side. The shoulder girdle is preserved in situin the holotype but is slightly dislocated inIVPP V11834. It consists of interclavicle,clavicles, coracoids, and scapulae. In the ho-lotype, the interclavicle is almost completeother than the left side of the crossbar; twoincomplete clavicles are exposed in dorsalview; two nearly complete coracoids are pre-served in situ; scapulae are fragmentary (fig.8). In IVPP V11834, all elements of the gir-dle except the left clavicle are completelypreserved and close to their original position,and the left scapula is exposed in dorsome-dial view (fig. 9).

INTERCLAVICLE

The interclavicle is a T-shaped, elongatedelement of flattened appearance. Its crossbaris crescent shaped; the anterior tip is roundedand rostrally directed; the lateral processestaper into the posterolateral direction. Twolateral processes are not symmetric. Theselateral processes are stronger than those inAskeptosaurus. Their anterior margin isslightly depressed for the contact with theback edge of the clavicles. The posterior pro-cess is the main body of the interclavicle; itscaudal elongation is about two times as longas the crossbar. There is no tapering into theposterior direction until the distal tip ends ina sharp angle in the holotype. The distal tipin IVPP V11834 is broken and partly cov-ered by ribs. In contrast, the distal tip appearsto be broad and rounded in Askeptosaurus.

CLAVICLE

The clavicle is a thin, slender, rodlike el-ement, which is significantly recurved intoan anterolateral direction. It can be dividedinto two parts: the anterior part that articu-


Fig. 8. Anshunsaurus huangguoshuensis, pectoral girdle in ventral view (specimen IVPP 11835-3).See appendix 1 for anatomical abbreviations.

lates with the interclavicle and the posteriorpart that does not. The proximal and the dis-tal portions enclose an angle of about 1208,which is more pronounced than that in Xin-pusaurus (Liu, 2001). The anterior part isslightly narrower than the posterior part inventral view, whereas the width of the entireposterior part is nearly the same. This is dif-ferent from the thinning backward in the pos-terior part of this bone in Askeptosaurus. Theanterior part is markedly shorter than theposterior one.

The concave edge of the clavicle, ratherthan the convex edge, articulates with the in-terclavicle (fig. 9). The same should be truefor the articulation with the scapula. In thereconstruction of the pectoral girdle in As-keptosaurus (Muller, 2002: fig 18), the au-thor followed Kuhn-Schnyder (1952, 1960)to position the clavicles curving to the pos-

terior; as a result the convex edge of the clav-icle contacts the interclavicle. If this is theright reconstruction, these two animals arequite different in this structure. However, thesimilar situation as in Anshunsaurus also isobserved in specimen PIMUZ T4842 of As-keptosaurus, so the relationship of the inter-clavicle and the clavicle in Askeptosauruscould be the same as described here in An-shunsaurus. The contact area for the inter-clavicle looks fixed in Anshunsaurus.

SCAPULA

The scapular blade is a broad, low struc-ture, as is the case in Askeptosaurus. Its an-terior margin is narrow and convex. Its dor-sal margin is broad and only slightly convex.The clavicle perhaps contacts only the dorsalmargin of the scapula, not the anterior mar-


Fig. 9. Anshunsaurus huangguoshuensis, pectoral girdle in ventral view (specimen IVPP V11834).See appendix 1 for anatomical abbreviations.

gin as in Askeptosaurus. Its posterior marginis strongly concave. The posteroventral por-tion of the bone is thickened and forms thedorsal part of the glenoid region. The ventralmargin is slightly convex, corresponding tothe morphology of the dorsal edge of the cor-acoid, to which it is sutured.

As mentioned by Liu (2001), there are twodifferent types of scapulae in Thalattosauria:one type that is seen in Thalattosaurus, Nec-tosaurus, and Xinpusaurus and another typecharacteristic of Askeptosaurus, Clarazia,Hescheleria, and Anshunaurus. These twotypes can also be found in the Ichthyoptery-gia (Motani, 1999). The function of theseshape types is still unknown.

CORACOID

The coracoid is a broad and somewhat par-allelogram-shaped bone. The left coracoid isincomplete in the holotype. Its lateral (dor-sal) margin is slightly convex and loosely ar-ticulated with the scapula with a pointed an-terolateral (anterodorsal) corner. The medial(ventral) margin is covered by the interclav-icle, so its exact shape is unclear. The pos-teroventral part extends posteriorly to theglenoid cavity. The coracoid foramen cannotbe observed in the position as in Askepto-saurus; it is perhaps obscured due to crush-ing.

It is difficult to reconstruct the original po-


Fig. 10. Anshunsaurus huangguoshuensis, pelvic girdle in situ in ventral view (specimen IVPP11835-4). See appendix 1 for anatomical abbreviations.

sition of coracoids in IVPP V11834. If theright coracoid is thought to be preserved inarticulation with the right scapula, then theside with a straight margin is ventral andcontacts the interclavicle; nevertheless theshape of the coracoids is quite different fromthat of the holotype. If we accepted thatshape should be at least similar in two spec-imens, then the axial/anteroposterior direc-tion of two coracoids should be reversed.This is less possible than the previous hy-pothesis.

PELVIC GIRDLE

The pelvis consists of three elements, ili-um, ischium, and pubis. It is completely pre-served in articulation in the holotype. In theholotype (figs. 6, 10), only the ilia can beseen in dorsal view; the ischiums and pubesare covered by the vertebrae and ribs dorsal-ly, but they are completely exposed ventrally.A vein of calcite splits the sacral rib and alsogoes across the right pubis and right ischium.The bones are displaced along this vein. Thepelvic girdle is exposed in ventral view in


IVPP V11834. The following description ismainly based on the holotype. No thyroid fe-nestra presents in Anshunsaurus, it shows aweak notch in the ventral margin of thepubo-ischiadic plate where the two bonesmeet. This is similar to Hescheleria (Rieppel,1987), corroborating the conclusion that thethyroid fenestra is absent in Askeptosaurus(Muller, 2002).

ILIUM

The ilium is similar to that of Askeptosau-rus in shape and mainly exposed in medialview (fig. 6). It forms the dorsal part of theacetabulum and contacts the sacral ribs at itsmedial side. The ilium does not fuse with theribs. The dorsal part of the ilium consists ofan elongated process that is directed stronglyposterodorsally. Posterodorsally, the processbecomes increasingly flattened but slightlynarrowing into the distal direction, with asharp dorsal and ventral edge. The dorsaledge of the ilium is rugose, which may haveserved for muscle attachment, that is, iliofe-moralis and iliofibularis muscles. There is nodevelopment of a preacetabular process. Theposterodorsal wing of the ilium spans abouttwo vertebral centra. The basal part of theilium, sutured to the ischium and pubis andcontributing to the formation of the acetab-ulum, is distinctly expanded and presents aconvex ventral margin.

PUBIS

The pubis is the most expanded elementof the pelvic girdle. It is a broad plate ofbone with an enlarged ventral portion. Theposterodorsal part of the pubis contributes tothe formation of the acetabulum. The small,oval obturator foramen is positioned antero-ventrally to the contact with the ilium. Theventral margin of the bone is convex, and thedorsal margin is nearly triangular. Its poste-rior margin, shorter than the anterior one,meets the ilium along its complete extension.The anterior margin of the pubis is stronglyconcave, whereas the outline of the posterioredge is only slightly concave. The right pubisseems to be wider than the left one in ventralview; this is caused by the offset along thecalcite vein.

ISCHIUM

The ischium is a more or less semicircularplate of bone. Its narrow anterodorsal portionforms the posteroventral portion of the ace-tabulum. Its anterodorsal margin is slightlyconcave, the posterodorsal margin is stronglyconcave, and the ventral margin is convex.The outline of the anterior and posterior mar-gins is not so obvious, but it seems that theanterior one is slightly convex, correspond-ing to the posterior margin of the pubis,whereas the posterior margin is concave asin IVPP V11834. The ischium is thickenedon the anterodorsal area; the bone becomesincreasingly flattened into the ventral direc-tion.

LIMBS

Anshunsaurus bears well-developed fore-and hindlimbs, although both are relativelyshort in relation to the length of the trunk.The forelimb is shorter than the hindlimb. Incontrast to the terrestrial reptiles generally,several details of the extremities such asmuscle scars, processes and epiphyses arenot present or at least very difficult to detect.

FORELIMB

Preservation allows both the flexor (dor-sal) and the extensor (ventral) side of theforelimb to be described. The forelimb isnearly complete in IVPP V11834 (fig. 9),whereas most of the hand is lost in the ho-lotype (fig. 5).

HUMERUS

The humerus (fig. 11) is the most robustand longest element of the forelimb skeleton.There is a notable torsion in the bone so thatboth heads are twisted to one another. Theproximal and distal articular surfaces arestrongly rounded. The proximal head pinchesto a narrow shaft and the shaft rapidly ex-pands to the distal side, forming a broad,convex end. A prominent crest, the deltopec-toral crest, originates near the head andmainly lies in the middle of the ventral sur-face. The crest is most distinct on its centralpart, fading away in the distal and proximaldirection, but the crest is more slender prox-imally than distally. There is no significant


Fig. 11. Anshunsaurus huangguoshuensis,right humerus of holotype in (A) ventral and (B)dorsal view.

development of a supinator process, ectepi-condyle, or entepicondyle, nor is there anytrace of a respective foramen. Radial and ul-nar condyles are not well developed either.

RADIUS

The radius is a relatively stout bone. It isstouter and broader than the ulna. The radiusis characterized by a slight proximal expan-sion and marked distal expansion. The prox-imal margin is almost straight whereas thedistal one is convex. The medial margin ismore concave than the lateral one.

ULNA

The ulna is slightly shorter than the radiusand lacks an ossified olecranon. The widestpart of the distal portion is nearly the sameas that of the proximal one. The proximalhead varies in shape in the two specimens:The margin is round and slightly convex inthe holotype but is nearly straight or evenslightly concave in IVPP V11834. A differ-ent degree of ossification may be the cause.The distal articular surface shows a convexmargin. The ulna is significantly constrictedin the shaft area, and shows concave lateraland medial margins. The ulna encloses a pro-nounced interosseal space with the radius.

MANUS

Seven rounded ossified elements are pres-ent in the left carpus of the holotype whereasonly four and half elements are preserved inthe right carpus; six and five, respectively,are present in IVPP V11834. Two proximalelements are present, ulnare and interme-dium. A small element, the centrale, is situ-ated between the intermedium and the distalcarpals 2 and 3 in the holotype. Four distalcarpals, distal carpals 1 to 4, which are ad-jacent to the metacarpals, form a distallyarched carpus. The intermedium is the largestelement of the carpus, and of rectangularshape with rounded edges. The ulnare, con-tacting the intermedium and distal carpal 4,is generally smaller than distal carpal 4, butit is slightly larger than the latter in the righthand of IVPP V11834. The centrale is ap-proximately half as large as the intermediumin Askeptosaurus, but it is much smaller thanthe intermedium and is the smallest bone inthe carpus here. Distal carpal 4 is the largestof all distal carpals. Distal carpal 3 is similarto the centrale in size in the holotype, but itis notably bigger than distal carpals 1 and 2in IVPP V11834. These kinds of variationalso are observed in Askeptosaurus.

Five metacarpals are shown in Anshunsau-rus as is general in tetrapods. Metacarpal 1is the shortest bone, being prominently wid-ened both on its two ends and in the shaftarea. The proximal articular surface is themost expanded part. The remaining metacar-pals are relatively slender. All are nearlyequal in size; only metacarpal 2 is slightlyshorter.

The phalanges are not so well preserved,but a least formula of 2-3-4-4-4 can becounted. Compared to the general formula of2-3-4-5-3, it is notable that the fifth digit pos-sesses one additional element. The phalangesare generally smaller than the metacarpals.The proximal phalanges of the first and thefifth digits are slightly longer than that of thesecond and the third, and the first two areslender whereas the last two are stout. Thestate of the phalanges in the fourth is unclear.The phalanges tend to become somewhatsmaller distally. The ungual phalanges aretriangular in shape with a sharp distal tip.


HINDLIMB

The hindlimb is completely preserved inIVPP V11834 (fig. 12), whereas only partsof the femur and the fibula are present in theholotype (figs. 6, 10). The left hindlimb ofIVPP V11834 is exposed in dorsal (anterior)view, whereas the right limb is observed inmedial view.

FEMUR

The incomplete femora in the holotype aremainly observed in ventral view whereastheir dorsal sides are partly covered by theilium. The right femur in IVPP V11834 isbroken and exposed in medial view; the leftfemur is mainly exposed in dorsal view withthe head partially covered by the pelvic gir-dle.

The femur is an elongated bone of cylin-drical shape that is longer than the humerus.The femoral heads are not so well preserved,but it is clear that they are rounded, weaklyconvex, and at least about the same width asthe distal end. A well-developed internal tro-chanter is shifted somewhat posteriorly (lat-erally) on the ventral aspect of the bone. Thiscrest can also be observed in IVPP V11834,developed as a modest expansion at the me-dial edge of the proximal articular surface.This suggests that the two heads were af-fected by a slight torsion. There is little ev-idence of a fourth trochanter, and the adduc-tor crest is absent. The intertrochanteric fossais reduced. The anterior edge of the femur issmoothly concave; the posterior edge isstraight to form a pronounced angulation.The femoral distal end shows two facets: amedial major one, which is slightly concave,for receiving the convex head of the tibia anda lateral small one, which is angled with themajor one, for articulating to the fibula.

TIBIA

The tibia is a cylindrical bone like the fe-mur. The proximal end, clearly expanding, iswider than the little-expanded distal end. Theproximal end is slightly convex, but the dis-tal end is flat. The shaft does not significantlyconstrict, and shows slight curvatures both inmediolateral and dorsoventral views.

FIBULA

The fibula is roughly as long as the tibia,but shows a distinctly different shape. Theproximal head shows torsion related to otherparts, and is only slightly broader than theshaft. The bone becomes more flattened to-ward the distal end, and bears a greatly ex-panded distal head. The latter is fan shapedand nearly three times as broad as the shaft,with the medial edge being slightly sharperthan the external one. The margin of the dis-tal articular surface is convex.

PES

The tarsus presents six bones, two proxi-mal and four distal elements. The proximalbones are interpreted as astragalus and cal-caneum; the distal elements represent distaltarsals 1–4. The astragalus is the most prom-inent element, being approximately the sizeof the calcaneum plus the fourth distal tarsal.It is broad and roughly kidney shaped, itsproximolateral facet articulating with the fib-ula, and the proximomedial edge contactingthe tibia. Laterally, the astragalus meets thecalcaneum; distally, it articulates with allfour tarsals with a convex margin. The cal-caneum is a rounded bone, with the proximalside articulating with the fibula, the lateralside with the astragalus, and the distal sidewith the fourth distal tarsal. A notch for theperforating artery cannot be observed. Thedistal tarsals vary in size. Distal tarsals 1–3are more or less equally large, although thefirst distal tarsal is sometimes slightly moreexpanded. Distal tarsal 4 is the largest ele-ment. The distal tarsals all contact theirneighboring elements and meet their respec-tive metatarsal, although the fourth also per-haps articulate with metatarsal 5.

The five metatarsals are longer and rela-tively more slender than the complementaryelements in the manus. The first metatarsalis shorter and stouter than the remaining el-ements and shows a distinct proximal expan-sion. The fifth metatarsal is only slightly lon-ger than the first one, expanded at least at itsproximal end. It is partially concealed by thehaemal arch. Metatarsals 2–4 are similar insize and shape. They are slender, elongatedelements with slightly expanded heads; the


Fig. 12. Anshunsaurus huangguoshuensis, hindlimb of specimen IVPP 11834, left side in dorsalview, right side in medial view. See appendix 1 for anatomical abbreviations.


proximal heads are slightly wider than thedistal ones.

The phalangeal formula of the pes is 2-3-4-5-4, as is the general pattern of reptiles.The proximal phalange of the first digit is thelongest, the remaining proximal phalangesare shorter and stouter, and that of the fifthdigit is the shortest. All phalanges decreasein size toward the distal direction. The un-gual phalanges are of the same shape as inthe manus.

MEASUREMENTS AND RATIOS

The size and proportions were analyzed inAskeptosaurus to document individual vari-ation (Muller, 2002). Because only two spec-imens are available for this study, this aim isnot suitable here. Here we present the mea-surements and ratios for further study.

The skull length was measured from theanterior tip of the snout to the level of theposterior tip of the supratemporal, whereasthe mandible length represents the distancefrom the symphysis to the caudal tip of thearticular. The standard length is the length offour articulated vertebrae in the posterior partof the trunk. Following Rieppel (1989), thefour posteriormost dorsals should be used forstandard length. However, as in the speci-mens studied by Muller (2002), the four pos-teriormost dorsals are not so well preserved,so the measurement was based on well-pre-served posterior dorsal vertebrae. As thelengths of the dorsals are subequal, thiswould not be a problem. The glenoid–ace-tabulum length represents the distance fromthe posteromedial corner of the proximal hu-merus head to the posteromedial corner of theproximal femur head. If possible, measure-ments were taken for both sides of the re-spective animal. A complete list of measure-ments and ratios is presented in appendix 2.

PHYLOGENETIC POSITION OFANSHUNSAURUS

Anshunsaurus is clearly distinct from thethalattosaur genus, Xinpusaurus, which oc-curs in the same locality and horizon. Theyare easily differentiated by the number ofcervical vertebrae, the direction of the clav-icle, and the shape of scapula; even the long

snout of Xinpusaurus bamaolinensis tapersto a pointed tip.

In all thalattosaurs, Anshunsaurus is sim-ilar to Endennasaurus and Askeptosaurus inthe presence of a long-necked body and along-snouted skull. Anshunsaurus especiallylooks like Askeptosaurus, although they aredifferentiated by some detailed characters. Itwas interpreted as the sister taxon of Askep-tosaurus based on 16 skull characters byRieppel et al. (2000). This hypothesis of amonophyletic long-necked Askeptosauridaeis not supported by some later analyses (Liuand Rieppel, 2001; Jiang et al., 2004), al-though it remains the result of Muller (2005).In Muller’s cladogram, Endennasaurus clus-ters with Anshunsaurus and Askeptosaurus.

To evaluate the phylogenetic relationshipswithin thalattosaurs, and to test the mono-phyly of Askeptosauridae in particular, a newanalysis was conducted, including all validthalattosaur genera and species (Agkistro-gnathus campebelli, Anshunsaurus huangguo-shuensis, Askeptosaurus italicus, Claraziaschinzi, Endennasaurus acutirostris, Hes-cheleria ruebeli, Nectosaurus halius, Para-lonectes merriami, Thalattosaurus alexan-drae, Thalattosaurus borealis, Xinpusaurusbamaolinensis, Xinpusaurus suni). Araeos-celidia and Younginiformes were selected asoutgroups in the phylogenetic analysis ofMuller (2005), but they are specified as Pe-trolacosaurus kansensis and Youngina ca-pensis in this study; Prolacerta broomi isalso added as an outgroup. Thus compositetaxa are avoided in this phylogenetic analy-sis, following the suggestion by others (Bi-ninda-Emonds et al., 1998; Norell et al.,2001). In sum, 12 ingroups and 3 outgroupsare applied in this analysis.

The following characters were selected forthis test; most of them are adapted from pre-vious analyses. The characters are coded onthe basis of personal observation and litera-ture sources (tables 1, 2).

CHARACTER LIST

The abbreviations that appear withinsquare brackets in the list below refer to thepublished reference and the number of timesin which the character or a similar characterhas been applied in numerical phylogenetic


TABLE 1Taxon-Character Data Matrix

Character number: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Petrolacosaurus kansensisYoungina capensisProlacerta broomiAgkistrognathus campebelliAnshunsaurus huangguoshuensis

000?2

000?2

000?0

000?0

000?1

000?1

000?1

000?0

000?1

000?0

000?0

000?0

000?0

010?1

011?0

Askeptosaurus italicusClarazia schinziEndennasaurus acutirostrisHescheleria ruebeliNectosaurus halius

21211

21111

01022

00?0?

10100

1?00?

01?11

00?00

10???

01???

0????

0?001

01???

201?0

001?1

Paralonectes merriamiThalattosaurus alexandraeThalattosaurus borealisXinpusaurus bamaolinensisXinpusaurus suni

12221

11111

21101

?1011

?0000

?0?10

???11

00011

?0?00

??211

?1?11

?1?11

?1?00

?0?00

?1?11

studies. Reference abbreviations are: N 5Nicholls 1999; R 5 Rieppel et al. 2000; L5 Liu and Rieppel 2001; J 5Jiang et al.2004; M 5 Muller et al., 2005.

SKULL

1. Rostrum absent (0); rostrum present butpreorbital region of skull not distinctly lon-ger than postorbital region of skull (1) (dis-tance from tip of snout to anterior margin oforbit shorter than distance from anterior mar-gin of orbit to posterior tip of supratempor-al); rostrum present but preorbital region ofskull distinctly longer than postorbital regionof skull (2) (distance from tip of snout toanterior margin of orbit longer than distancefrom anterior margin of orbit to posterior tipof supratemporal). [R1, L1, J1, M1]

2. Rostrum absent (0); rostrum tapering topointed tip, that is, with convergent lateralmargins in front of external nares (1); ros-trum tapering to blunt tip, that is, with par-allel lateral margins in front of external nares(2). [R1, L1, J1, M1]

3. Tip of snout (rostrum) straight (0), tipof snout (rostrum) slightly deflected ventrally(1), or strongly deflected ventrally (2). [N19,R2, L2, M2, J2] Definition revised.

According to Rieppel et al. (MS), a verti-cally descending premaxillary rostrum is aunique morphology shared by Hescheleriaand Nectosaurus, perhaps also Paralonectes;so they are coded as ‘‘2’’.

4. Anterior narial margin formed by pre-maxilla (0) or largely or exclusively by thenasal (1). [J19]

A possible septomaxilla was suggested be-fore the external naris in Endennasaurus(Muller et al., 2005), but this part is broken,so coded as ‘‘?’’ here. The nasal projects an-teriorly of the external naris in new specimenof Xinpusaurus suni (Rieppel and Liu, MS),and it looks like this also is the case inV11860, so it is coded as ‘‘1’’.

5. Premaxilla has a distinct posteroventralprocess extending far beyond the anteriornarial margin in ventral side (0) or no sucha structure (1) [J21]. Character definition andpolarity has been changed.

Coding Clarazia as (0) following Muller(2002).

6. Maxilla is longer (0) or shorter (1) thanpart of premaxilla forming upper jaw max-illa.

7. Maxilla does not (0) or does (1) formpart of the outer orbital rim.

8. Anterior part of alveolar margin of max-illa straight (0) or distinctly curved upward(1). [L18, J17]

9. Nasals do not (0) or do (1) extend back-ward to level behind anterior margin of orbit.[R6, L6, J5, M11]

In the skull reconstruction of Youngina(fig. 6 in Gow 1975), it seems that the nasalsextend backward to a level behind the ante-


TABLE 1(Extended )

16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32

000?1

000?0

000?1

001?0

00010

0?1?0

000?0

000?0

00000

00000

000?1

000?1

000?0

000?0

000?1

000?0

000?0

01111

01???

0????

01???

0101?

10??1

00100

01?11

01???

01?10

101?0

1010?

00011

0101?

11110

11010

01012

?1?11

?1?0?

?1?00

????1

11?11

?1?11

01100

11?00

11?00

11111

00000

???00

?1?11

?1??1

?0?00

?0?11

?2?22

rior margin of the orbit, but this is not truebased on original figures.

10. Anterolateral processes of frontal re-main broadly separated from external naris(0) or anterolateral processes of frontal close-ly approach (1) or even enter the posteriormargin of external naris (2). [R7, L7, M12,J6]

It is clear that only the anteromedial pro-cess nearly reaches the external naris in As-keptosaurus, not the anterolateral process.This character was miscoded for derivedstate by Muller et al. (2005).

11. Nasal in contact with prefrontal (0),nasal separated from prefrontal (1). [N14,R8, L8, J7, M13]

The lacrimal is viewed as absent in Askep-tosaurus, so prefrontal in Askeptosaurus isthe same as that in Anshunsaurus, so thischaracter is coded as ‘‘0’’.

12. Prefrontal without (0) or with (1)shieldlike extension along the anterodorsalorbital margin. [J20]

13. Frontal separate from supratemporal(0), in contact with supratemporal (1). [R11,L11, J10, M15] Character definition hadbeen revised.

For the holotype of Xinpusaurus bamao-linensis (SPCV30015), it is clear showingthat frontal separate from supratemporal inthe description and figure 1a (Cheng, 2003),but the reverse state is present in the recon-struction. This part is poorly preserved in

‘‘X. kohi’’ (GMPKU2000/005). So the cod-ing for both species of Xinpusaurus is ‘‘0’’.

14. Posterolateral processes of frontal donot extend beyond anterior margin of lowertemporal fossa (0), extend far beyond ante-rior margin of lower temporal fossa (1), areabsent (2). [R10, L10, J9, M14]

The posterolateral processes are moderate-ly developed in Petrolacosaurus and Prola-certa, well developed in Youngina. So thepolarity is changed.

15. Frontoparietal suture interdigitating,oriented transversely for most of its part (0)or deeply embayed in the shape of a broadV, the apex pointing forward (1). [L17, J16,M16]

16. Postfrontal and postorbital separate (0)or fused (1). [N16, R12, L12, J11, M17]

17. Posterior end of squamosal with (0) orwithout ventral process (1). [R14, L14, J14,M19]

18. Quadrate without (0) or with (1) dis-tinct medial lamina. [M20]

19. Pineal foramen large and located infront of midpoint of parietal skull table (0)or small and located at center or somewhatbehind parietal skull table (1). [R16, L16,J15, M22] The polarity is changed.

LOWER JAW

20. Dorsal margin of dentary symphysisstraight (0), recurved (1). [N22]


TABLE 2Specimens and Literature Consulted for Construction of Phylogenetic Data Matrix in Table 1

Taxon Information sources

Petrolacosaurus kansensisYoungina capensisProlacerta broomiAgkistrognathus campebelliAnshunsaurus huangguoshuensis

(Reisz, 1981)(Gow, 1975)(Gow, 1975; Modesto and Sues, 2004)(Nicholls and Brinkman, 1993)IVPP V11834, IVPP V11835

Askeptosaurus italicusClarazia schinziEndennasaurus acutirostrisHescheleria ruebeliNectosaurus halius

(Kuhn-Schnyder, 1952, 1960, 1971; Muller, 2002; Nopsca, 1925)PIMUZ A/III 211 (Muller, 2002; Peyer, 1936a; Rieppel, 1987)(Muller et al., 2005)PIMUZ T2469 (Peyer, 1936b; Rieppel, 1987; Rieppel et al., MS)UCMP 9124 (Nicholls, 1999; Rieppel et al., MS)

Paralonectes merriamiThalattosaurus alexandraeThalattosaurus borealisXinpusaurus bamaolinensisXinpusaurus suni

(Nicholls and Brinkman, 1993)UCMP 9085 (Nicholls, 1999)(Nicholls, 1999)SPCV 30015, GMPKU 2000/005 (Cheng, 2003; Jiang et al., 2004)IVPP V11860, IVPP V14372, Gmr 010, GGSr001

21. Surangular and angular equally wellexposed on the lateral side of lower jaw (0)or surangular exposure much more extensive(1). [J24] Coding for Askeptosaurus as ‘‘1’’.

TEETH

22. Premaxilla dentition is present (0) orabsent (1) (state 1 includes also pseudodontprojections). [N29, M3]

23. Diastema between premaxillary andmaxillary teeth absent (0), present (1). [N28,R3, L3, J3, M4]

24. Anteriormost dentary teeth upright (0),procumbent as their implantation curvesaround anterior end of dentary (1). [R4, L4,M5]

25. Posterior teeth on dentary conical andpointed (0), bulbous and blunt (1). [part ofR5, L5, J4, M6]

The character definition is revised, so thecoding has changed for Xinpusaurus (1).

26. Palatal dentition (on vomer and pter-ygoid) present (0) or absent (1). [N25, 26, 27M7, 8, 9]

POSTCRANIAL

27. Number of cervical vertebrae is lessthan 10 (0) or higher than 10 (1). [N35, M24]

28. Neural spines of posterior cervicalsand dorsals are relatively low (0) or at leasttwo times higher than broad (1). [M28, J27]

29. Proximal caudal neural spines are rel-

atively low (0) or distinctly elongated and atleast three times higher than broad (1).[M29]

Because the postcranium of Xinpusaurus‘‘kohi’’ appears to be a composite and it isnot known in the holotype of X. bamaoli-nensis, coding is ‘‘?’’ in this species.

30. Scapula is slender and elongate (0) orbroad and rounded (1).

31. Deltopectoral crest is well developed(0) or reduced (1). [N38, M26]

32. Radius is slender (0), slightly expand-ed (1), or strongly expanded and roughly kid-ney shaped (2). [J28]

33. Femur proximal and distal ends aboutequal in width (0) or distal end markedlywidened (1). [J29]

34. Fibula distal end narrow, that is, widthless than half of length (0), expanded (1).

Some characters included in former stud-ies are deleted from this analysis becausethey are uninformative for the analysis of in-group relationships. These characters in-clude:

1. Nasals meet each other medially (0) orare separated due to the posterior extent ofthe premaxilla (1). [M10]

2. Anteromedial processes of frontals thatenter in between nasal and premaxilla areshorter (0) or longer (1) than anterolateralprocesses of frontal. [R9, L9, J8]

3. The quadratojugal is present (0) or ab-sent (1). [M21]


Fig. 13. Phylogenetic relationships of Anshunsaurus within Thalattosauria based on the analysis of34 characters (see table 1 for data), strict consensus of 15 shortest trees, tree length 5 62, CI 5 0.65,RI 5 0.77.

4. Lacrimal present (0) or absent (1).[N12, J18]

5. Upper temporal fenestra present andlarge (0), reduced and slitlike (1), or absent(2). [N1, R13, L13, J12, M18]

All multistate characters were treated un-ordered and unweighted. Implementingbranch-and-bound search option of thePAUP* 4.0b10 (Swofford, 2001), the 15most parsimonious trees resulted, with alength of 62 (consistency index 0.65, reten-tion index 0.77). In the strict consensus tree(fig. 13), the monophyly of Askeptosauridae(see Rieppel et al., 2000) and Thalattosau-roidea (Nopsca, 1928; see also Rieppel 1987)are recovered, Endennasaurus is the sistertaxon of Askeptosauridae, Thalattosaurus

and Xinpusaurus are monophyletic genera,Paralonectes and Nectosaurus from NorthAmerica are closely related to Clarazia andHescheleria from Monte San Giorgio.

Askeptosaurus and Anshunsaurus are un-equivocally diagnosed by the rostrum taper-ing to a blunt tip (Char. 2[2], CI 5 1), themaxilla is shorter than the part of the pre-maxilla forming the upper jaw maxilla (Char.6[1], CI 5 0.50), and the frontoparietal su-ture interdigitating, oriented transversely formost of its part (Char. 15[0], CI 5 0.33).Based on DELTRAN character optimization,the shared characters also include the nasalextends backward to a level behind the an-terior margin of the orbit (Char. 9[1], CI 51).


The family Askeptosauridae (Kuhn-Schnyder, 1952) included only one genus andspecies for a long time. Anshunsaurus wasincluded as the second member of this group(Rieppel et al., 2000), although the mono-phyly of this family is in doubt. Endenna-saurus was proposed as being closely relatedto thalattosaurs, especially Askeptosaurus, bythe nominator (Renesto, 1992). This opinionwas accepted by Nicholls (1999), but not ac-cepted and excluded in our previous analysis(Rieppel et al., 2000). Subsequent researchby Muller et al. (2005) supported its positionas a thalattosaur. In this study, it also turnsout to be closely related to Askeptosaurusand Anshunsaurus. So it is reasonable to in-clude Endennasaurus in Askeptosauroidea(Kuhn-Schnyder, 1971). The unequivocalsynapomorphies include: premaxilla has nodistinct posteroventral process extending be-yond the anterior narial margin in ventralside (Char. 5[1], CI 5 1); posterolateral pro-cesses of frontal extending far beyond ante-rior margin of lower temporal fossa (Char.14[1], CI 5 0.67); palatal dentition (on vo-mer and pterygoid) absent (Char. 26[1], CI5 1); the number of cervical vertebrae morethan 10 (Char. 27[1], CI 5 1); and the scap-ula broad and rounded (Char. 30[1], CI 50.5).

The monophyly of Thalattosauroidea issupported by all analyses including this one(Nicholls, 1999; Rieppel et al., 2000; Liu andRieppel, 2001; Muller, 2002; Jiang et al.,2004). An unequivocally shared derivedcharacter of this group is dorsal margin ofdentary symphysis recurved (Char. 20[1], CI5 1). Based on ACCTRAN character opti-mization, the shared characters also includetip of snout (rostrum) slightly deflected ven-trally (Char. 3[1], CI 5 0.5), anterolateralprocesses of frontal closely approach exter-nal naris (Char. 10[1], CI 5 1), nasal sepa-rated from prefrontal (Char. 11[1], CI 5 1),prefrontal with shieldlike extension along theanterodorsal orbital margin (Char. 12[1], CI5 0.5), pineal foramen small and located atcenter or somewhat behind parietal skull ta-ble (Char. 19[1], CI 5 0.5), neural spines ofposterior cervicals and dorsals at least twotimes higher than broad (Char. 28[1], CI 50.5), proximal caudal neural spines distinctlyelongated and at least three times higher than

broad (Char. 29[1], CI 5 1), the radiusstrongly expanded and roughly kidneyshaped (Char. 32[2], CI 5 1), the femur distalend markedly widened (Char. 33[1], CI 5 1),the slender fibula (Char. 34[1], CI 5 0.5).

The intrarelationships of Thalattoauroideaare different from all previous results, al-though the exact relationships of the fourgenera are still unclear. Although Claraziidae(Peyer, 1936b) may be monophyletic, Thal-attosauridae (Merriam 1904) is definitely not.A sister relationship of Xinpusaurus withNectosaurus is not supported by this study.

ACKNOWLEDGMENTS

We are grateful to Li Jinling and Li Chunfor generous access to the thalattosaur ma-terial kept at the Institute of Vertebrate Pa-leontology and Paleoanthropology, ChineseAcademy of Sciences. Ding Jinzhao andZhang Hong prepared the specimens; ZhangJie and Zhang Fucheng helped to photograph.Donald Brinkman and Johannes Muller kind-ly reviewed an earlier draft of this paper. Thiswork is partly supported by National ScienceFoundation of China (No. 40302007).

REFERENCES

Benton, M.J. 1985. Classification and phylogenyof the diapsid reptiles. Zoological Journal of theLinnean Society 84: 97–164.

Bininda-Emonds, O.R.P., H.N. Bryant, and A.P.Russell. 1998. Supraspecific taxa as terminalsin cladistic analysis: implicit assumptions ofmonophyly and a comparison of methods. Bi-ological Journal of the Linnean Society 64:101–133.

Cheng, L. 2003. A new species of Triassic Thal-attosauria from Guanling, Guizhou. GeologicalBulletin of China 22: 274–277.

Dalla Vecchia, F.M. 1993. Reptile remains fromthe Middle-Upper Triassic of the Carnic andJulian Alps (Friuli-Venezia Giulia, northeasternItaly). Gortania-Atti del Museo Friulano di Sto-ria Naturale 15: 49–66.

Evans, S.E. 1988. The early history and relation-ships of the Diapsida. In M.J. Benton (editor),The phylogeny and classification of the tetra-pods: 221–260. Systematics Association Spe-cial Publication. Oxford: Clarendon Press.

Gow, C.E. 1975. The morphology and relation-ships of Youngina capensis Broom and Prola-certa broomi Parrington. Palaeontologia Afri-cana 18: 89–131.


Jiang, D.-Y., M.W. Maisch, Y.-L. Sun, A.T. Matz-ke, and W.-C. Hao. 2004. A new species ofXinpusaurus (Thalattosauria) from the upperTriassic of China. Journal of Vertebrate Pale-ontology 24: 80–88.

Kuhn-Schnyder, E. 1952. Die Triasfauna der Tes-siner Kalkalpen. XVII. Askeptosaurus italicusNopcsa. Schweizerische Palaeontologische Ab-handlungen 69: 1–52.

Kuhn-Schnyder, E. 1960. Uber einen Schultergur-tel von Askeptosaurus italicus Nopsca aus deranisischen Stufe der Trias des Monte San Gior-gio (Kt. Tessin, Schweiz). Ecologae GeologicaeHelvetiae 53: 805–810.

Kuhn-Schnyder, E. 1971. Uber einen Schadel vonAskeptosaurus italicus Nopcsa aus der Mittler-en Trias des Monte San Giorgio (Kt. Tessin,Schweiz). Abhandlungen des Hessischen Lan-desamtes fur Bodenforschung 60: 89–98.

Liu, J. 1999. New discovery of sauropterygianfrom Triassic of Guizhou, China. Chinese Sci-ence Bulletin 44: 1312–1315.

Liu, J. 2001. The postcranial skeleton of Xinpu-saurus. In T. Deng and Y. Wang (editors), Pro-ceedings of the Eighth Annual Meeting of theChinese Society of Vertebrate Paleontology: 1–8. Beijing: Ocean Press.

Liu, J., and O. Rieppel. 2001. The second thalat-tosaur from the Triassic of Guizhou, China.Vertebrate PalAsiatica 39: 77–87.

Merriam, J.C. 1904. A new marine reptile fromthe Triassic of California. University of Cali-fornia Publications in Geological Sciences 3:419–421.

Merriam, J.C. 1905. The Thalattosauria: a groupof marine reptiles from the Triassic of Califor-nia. California Academy of Sciences Memoirs5: 1–38.

Merriam, J.C. 1908. Notes on the osteology of thethalattosaurian genus Nectosaurus. Universityof California Publications in Geological Sci-ences 5: 217–223.

Modesto, S.P., and H.-D. Sues. 2004. The skull ofthe Early Triassic archosauromorph reptile Pro-lacerta broomi and its phylogenetic signifi-cance. Zoological Journal of Linnean Society140: 335–351.

Motani, R. 1999. Phylogeny of the Ichthyoptery-gia. Journal of Vertebrate Paleontology 19:473–496.

Muller, J. 2002. A revision of Askeptosaurus ita-licus and other thalattosaurs from the EuropeanTriassic, the interrelationships of thalattosaurs,and their phylogenetic position within diapsidreptiles (Amniota, Eureptilia). Ph.D. thesis, Jo-hannes Gutenberg-Universitat, Mainz.

Muller, J. 2003. Early loss and multiple return of

the lower temporal arcade in diapsid reptiles.Naturwissenschaften 90: 473–476.

Muller, J. 2004. The relationships among diapsidreptiles and the influence of taxon selection. InG. Arratia, R. Cloutier, and V.H. Wilson (edi-tors), Recent advances in the origin and earlyradiation of vertebrates: 379–408. Munich: Dr.Friedrich Pfeil.

Muller, J. in press. The anatomy of Askeptosaurusitalicus from the Middle Triassic of Monte SanGiorgio, and the interrelationships of thalatto-saurs (Reptilia, Diapsida). Canadian Journal ofEarth Sciences.

Muller, J., S. Renesto, and S.E. Evans. 2005. Themarine diapsid reptile Endennasaurus (Reptilia:Thalattosauriformes) from the Late Triassic ofItaly. Palaeontology 48: 1–16.

Nicholls, E.L. 1999. A reexamination of Thalat-tosaurus and Nectosaurus and the relationshipsof the Thalattosauria (Reptilia: Diapsida).PaleoBios 19: 1–29.

Nicholls, E.L., and D.B. Brinkman. 1993. Newthalattosaurs (Reptilia: Diapsida) from the Tri-assic sulphur formation of Wapiti Lake, BritishColumbia. Journal of Paleontology 67: 263–278.

Nopsca, F.V. 1925. Askeptosaurus, ein neues Rep-til aus der Trias von Besano. Cbl. Min. Geol.Pal. B: 265–267.

Nopsca, F.V. 1928. The genera of reptiles. Palaeo-biologica 1: 163–188.

Norell, M.A., J.M. Clark, and P.J. Makovicky.2001. Phylogenetic relationships among coelu-rosaurian theropods. In J. Gauthier and L.F.Gall (editors), New perspectives on the originand early evolution of birds: proceedings of theInternational Symposium in honor of John H.Ostrom: 49–67. New Haven, CT: Peabody Mu-seum of Natural History, Yale University.

Peyer, B. 1936a. Die Triasfauna der Tessiner Kal-kalpen. X. Clarazia schinzi nov. gen. nov. sp.Abhandlungen der schweizerischen palaonto-logischen Gesellschaft 57: 1–61.

Peyer, B. 1936b. Die Triasfauna der Tessiner Kal-kalpen. XI. Hescheleria rubeli nov. gen. nov.sp. Abhandlungen der Schweizerischen Palaon-tologischen Gesellschaft 58: 1–48.

Reisz, R.R. 1981. A diapsid reptile from the Penn-sylvanian of Kansas. University of Kansas Spe-cial Publication of the Museum of Natural His-tory 7: 1–74.

Renesto, S. 1992. The anatomy and relationshipsof Endennasaurus acutirostris (Reptilia, Neo-diapsida), from the Norian (Late Triassic ofLombardy). Rivista Italiana di Paleontologia eStratigrafia 97: 409–430.

Rieppel, O. 1987. Clarazia and Hescheleria, a re-investigation of two problematical reptiles from


the Middle Triassic of Monte San Giorgio,Switzerland. Palaeontographica A 195: 101–129.

Rieppel, O. 1989. A new pachypleurosaur (Rep-tilia: Sauropterygia) from the Middle Triassicof Monte San Giorgio, Switzerland. Philosoph-ical Transactions of the Royal Society of Lon-don Series B 323: 1–73.

Rieppel, O. 1998. The systematic status of Han-osaurus hupehensis (Reptilia, Sauropterygia)from the Triassic of China. Journal of Verte-brate Paleontology 18: 545–557.

Rieppel, O., and H. Hagdorn. 1998. Fossil reptilesfrom the Spanish Muschelkalk (Montral andAlcover, Province Tarragona). Historical Biol-ogy 13: 77–97.

Rieppel, O., and J. Liu. Submitted. On Xinpusau-rus (Reptilia: Thalattosauria). Journal of Ver-tebrate Paleontology.

Rieppel, O., J. Liu, and H. Bucher. 2000. The firstrecord of a thalattosaur reptile from the LateTriassic of southern China (Guizhou Province,

PR China). Journal of Vertebrate Paleontology20: 507–514.

Rieppel, O., J. Muller, and J. Liu. Submitted. Ros-tral structure in Thalattosauria (Reptilia, Diap-sida). Canadian Journal of Earth Sciences.

Romer, A.S. 1956. Osteology of the reptiles. Chi-cago: University of Chicago Press,772 pp.

Sander, P.M., O. Rieppel, and H. Bucher. 1994.New marine vertebrate fauna from the MiddleTriassic of Nevada. Journal of Paleontology 68:676–680.

Storrs, G.W. 1991. Note on a second occurrenceof Thalottosaur Remains (Reptilia, Neodiapsi-da) in British-Columbia. Canadian Journal ofEarth Sciences 28: 2065–2068.

Swofford, D.L. 2001. PAUP*. Phylogenetic Anal-ysis Using Parsimony (* and Other Methods).version 4.0b10. Sunderland, MA: Sinauer As-sociates.

Yin, G., X. Zhou, Z. Cao, Y. Yu, and Y. Luo.2000. A prelimary study on the early late Tri-assic marine reptiles from Guanling, Guizhou,China. Geology Geochemistry 28: 1–23.

APPENDIX 1

ANATOMICAL ABBREVIATIONS

A angularArt articularas astragalusBo basioccipitalca calcaneumCl clavicleCo coracoidcr caudal ribcul pr cultriform processD dentarydc distal carpaldt distal tarsalF frontalFe femurFi fibulaHu humerusHy hyoidI iliumIcl interclaviclein intermediumIs ischiumJ jugalM maxilla

N nasalobt. fo obturator foramenOp opisthoticP parietalPl palatinePm premaxillaPof postorbitofrontalPra prearticularPrf prefrontalPro prooticPs-Bs parabasisphenoidPt pterygoidPu pubisQ quadrateR radiusSa surangularSc scapulaSo supraoccipitalSp splenialSq squamosalsr sacral ribSt supratemporalTi tibiaU ulnaul ulnareV vomer


APPENDIX 2

Anshunsaurus huangguoshuensis Measure-ments (L 5 left; R 5 right)

Skull and lower jaw of holotype, IVPP V11835(in mm)

Distance from tip of snout:

to anterior margin of orbit LR

236.0232.5

to anterior margin of upper temporalfossa

LR

309.5309.0

to posterior margin of parietal skulltable

345.0

to posterior margin of supraoccipital 355.5

to posterior tip of supratemporal LR

394.0397.0

to the posterior margin of basi-occipital

361

Width of skull across posterior tips ofsupratemporal

96.5

Longitudinal diameter of orbit LR

63.362.8

Longitudinal diameter of upper tempo-ral fossa

L 69.0

Distance from posterior margin of ex-ternal naris to anterior margin of orbit

LR

32.531.5

Distance from tip of dentary to posteriormargin of articular

LR

401400

Cervical region (in mm)IVPP V11835 Atlas Axis 3 10

Length of centrumHeight of vertebraLength of ribLength of neural spine

28

.322424

.32

29242620 20

IVPP V11835 14 15Sacral

1Sacral

2

Length of centrumHeight of vertebraLength of ribLength of neural spine

4059

40

27

34

57

34

59

IVPP V11834 Axis 3 4 5 6 7 8

Length of centrumLength of rib

21 2618

26 2720

2822

29 30

IVPP V11834 9 10 11 12 13 14 15

Length of centrumLength of rib

2726

2828

31 31 30 32

APPENDIX 2

(Continued )

Left sacral ribs

Firstrib

Secondrib

Length 57 mm 59 mmWidth of distal endWidth of proximal end

32 mm12 mm

35 mm23 mm

Limb girdle

V11835 V11834

Length of interclavicleWidth of interclavicleLength of clavicle

22 cm

.17 cm

21 cm12 cm17 cm

Length of coracoid (axial)Width (or height) of coracoidLength of iliumWidth of basal ilium

119 mm.73 mm122 mm

55 mmLength of L pubis (axial)Height of L pubisLength of ischiumHeight of ischium

72 mm82 mm73 mm77 mm

Stylo- and zuegopodium (in mm)

IVPP V11835

Humerus

L R

Radius

L R

LengthWidth of proximal endMinimal width

1524635

1524435

913622

913623

Width of distal end 73 72 45 44

IVPP V11835

Ulna

L R Femur Fibula

LengthWidth of proximal endMinimal width

893720

883522

.15654

Width of distal end 37 58

IVPP V11834

Humerus

L R

Radius

L R

LengthWidth of proximal endMinimal widthWidth of distal end

141

3067

134413066

8530?2037

85302137.5

IVPP V11834

Ulna

L RFemur

LTibia

LFibula

L


823418

79361832

160?35?2853

100371824

101201751


APPENDIX 2

(Continued )

Measurements taken on specimenIVPP V11834

Length of lower jaw 37.5 cmLength of cervical vertebrae 44 cmLength of caudal vertebrae .170 cmPresacral length (including skull) 175 cmLength from glenoid cavity to acetabulum 73 cm

Metacarpal

I

L R

II

L R

III

L R


40291318

39271219

45

1015

44161117

4915

915

47171016

Metacarpal

IV

L R

V

L R


47111013

47191116

1847151116

Leftmetatarsal I II III IV V


45251219

5321?1114

5519?1117

54181115

42

15

APPENDIX 2

(Continued )

Some ratios in two specimens of Anshunsaurus

MeasurementsIVPP

V11835IVPP

V11834

Head length to glenoid-acetabulum length 0.52

Head length to humeruslength 2.60 2.70 2.84

Humerus length to glenoid-acetabulum length 0.19 0.18

Humerus length to standardlength 0.98

Humerus length to femurlength 0.88

Humerus length to length ofmetacarpal 3 2.88 2.85

Humerus length to radiuslength 1.67 1.67 1.66 1.58

Humerus length to distalwidth 2.08 2.11 2.10 2.03

Radius length to standardlength 0.59

Humerus minimal width tohumerus distal width 0.23 0.23 0.21 0.22

Femur length to glenoid-acetabulum length 0.22

Femur length to fibulalength 1.58

Complete lists of all issues of the Novitates and the Bulletin are available at World Wide Website http://library.amnh.org/pubs. Inquire about ordering printed copies via e-mail [email protected] or via standard mail from: American Museum of Natural History, Library—Scientific Publications, Central Park West at 79th St., New York, NY 10024. TEL: (212) 769-5545. FAX: (212) 769-5009.

a This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper).

restudy of anshunsaurus huangguoshuensis (reptilia ... · the anterior margin of the orbit as in...

Documents