Cranial Anatomy of Citipati osmolskae (Theropoda, Oviraptorosauria ...

Copyright q American Museum of Natural History 2002 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 3364, 24 pp., 13 figures March 26, 2002

Cranial Anatomy of Citipati osmolskae(Theropoda, Oviraptorosauria), and aReinterpretation of the Holotype of

Oviraptor philoceratops

JAMES M. CLARK,1 MARK A. NORELL,2 AND TIMOTHY ROWE3

ABSTRACT

We describe the skull of the holotype of Citipati osmolskae, one of the best preserved oviraptoridskulls known. The skull preserves stapes and epipterygoids, and the mandible preserves a slendercoronoid bone, none of which has been reported before in oviraptorids. The braincase is similarto that of other basal coelurosaurs but possesses extensive recesses presumably occupied by pneu-matic diverticulae; the circumnarial region is highly pneumatized, and a large recess continuesposteriorly from the narial region to invade the frontals and parietals dorsal to the braincase.Circum-otic pneumatic recesses include two dorsal recesses above the otic recess, a posterior recesson the anterior surface of the paroccipital process, and extensive cavities in the basisphenoidbeneath the braincase. The more dorsal of the two dorsal tympanic recesses is very deep, and CTscans suggest that it connected medially across the midline dorsal to the otic region and anteriorlywith the frontoparietal space. The otic recess is unusually shallow. Comparison of the new skullwith the poorly preserved skull of the holotype of Oviraptor philoceratops demonstrates that thebraincase and palate of the latter are similar to those of other oviraptorids. Its rostrum and dentaryare more elongate than in other oviraptorids, however, a more plesiomorphic condition suggestingit may be the most basal oviraptorid. A well-preserved skeleton previously referred to O. philo-ceratops, IGM 100/42, does not belong to this genus or species, and its narial region is verysimilar to that of Citipati osmolskae.

1 Ronald S. Weintraub Associate Professor, Department of Biological Sciences, George Washington University,Washington, DC 20052. Research Associate, Division of Paleontology, American Museum of Natural History. e-mail:[email protected]

2 Chairman, Division of Paleontology, American Museum of Natural History. e-mail: [email protected] J. Nalle Gregory Regents Professor of Geology, and Director, Vertebrate Paleontology Laboratory, Department

of Geological Sciences, The University of Texas at Austin, Austin, TX 78712. e-mail: [email protected]

2 NO. 3364AMERICAN MUSEUM NOVITATES

INTRODUCTION

Among the many specimens collected byAmerican Museum–Mongolian Academy ofSciences expeditions from the Late Creta-ceous deposits of Ukhaa Tolgod, Mongolia,are two new species of oviraptorid theropods,Citipati osmolskae and Khaan mckennaiClark et al., 2001. Citipati is represented byan unusually complete and well-preservedskull, which has been CT scanned to exposethe internal anatomy of the braincase. In ad-dition to providing information on structurespreviously unknown in oviraptorids, this ex-quisite skull provides a basis for reinterpret-ing the poorly preserved holotypic skull ofthe first oviraptorid ever discovered, Ovirap-tor philoceratops (Osborn, 1924). This rein-terpretation of O. philoceratops reveals sev-eral differences between it and other ovirap-torids, suggesting that it may be the mostprimitive member of the group and contra-dicting the previous allocation of other spec-imens to this genus and species.

Oviraptorids were first discovered in 1923at Bayn Dzak (‘‘Flaming Cliffs’’), Mongolia,during the American Museum’s Central Asi-atic expeditions (Andrews, 1932). Oviraptorphiloceratops Osborn, 1924, was initiallycharacterized as an egg eater, due to a mis-identification of the eggs over which the ho-lotypic skeleton was preserved. However, therecent discovery at Ukhaa Tolgod of an em-bryonic oviraptorid in an egg of the sametype as beneath the holotypic skeleton cor-rected this error (Norell et al., 1994, 2001).The interpretation of O. philoceratops as anegg eater was engendered in part by its dis-tinctive skull with a short, edentulous beaksuggesting feeding habits unusual for a non-avian theropod. This bizarre cranial mor-phology has led authors to propose variousdietary preferences (Barsbold, 1986; Smith,1990, 1993). Until recently, oviraptoridswere among the rarest of nonavian dinosaurs,but happily this is no longer true.

Since the discovery of Oviraptor philocer-atops four additional genera and five speciesof oviraptorid have been described (Barsboldet al., 1990; Clark et al., 2001). All are fromthe Late Cretaceous of Mongolia, and thegroup is otherwise reported only from cor-relative deposits in northern China (Dong

and Currie, 1996). Oviraptorid relatives,within the Oviraptorosauria, are known fromthe Early and Late Cretaceous of NorthAmerica and the Late Cretaceous of Uzbek-istan (Sues, 1997; Makovicky and Sues,1998; Currie et al., 1993). Recent phyloge-netic analyses (Sereno, 1999; Holtz, 2001;Norell et al., 2001) of Caudipteryx (Ji et al.,1998; Zhou et al., 2000) suggest that this an-imal is an oviraptorosaur or close relative ofOviraptorosauria. Caudipteryx is notableboth for its early age and for the preservationof fully formed feathers across the body (Jiet al., 1998)

The Djadokhta Formation at Ukhaa Tol-god in south central Mongolia preserves oneof the most abundant, diverse, and well-pre-served theropod faunas yet discovered at anyMesozoic locality (Dashzeveg et al., 1995;Norell et al., 1996; Norell, 1997). Surpris-ingly, among the most common elements ofthis fauna are oviraptorids. Their remains in-clude skeletons of adults, juveniles, and em-bryos, and among the adults are three spec-imens preserved on top of nests of ovirap-torid eggs, indicating brooding behavior (No-rell et al., 1995; Clark et al., 1999). As withother vertebrate remains from the DjadokhtaFormation of Mongolia and China, many ofthe oviraptorid specimens comprise articulat-ed skeletons, often complete or nearly com-plete.

METHODS

The holotype skull of Citipati osmolskae,IGM 100/978, was CT scanned at the Uni-versity of Texas High Resolution X-Ray CTFacility in November 1997. The jaws, hyoid,and stapes had been separated from the cra-nium during initial mechanical preparation,and were not scanned. Interslice spacing was0.5 mm, with a field of reconstruction of 200mm and output levels adjusted to 239 grays.The scanning generated 181 original sagittalslices, which were later digitally resliced into348 coronal and 271 horizontal slices usingNIH Image software.

Contrast in the CT imagery between thevery thin bones and matrix is weak. Never-theless a number of sutures can be traced,and the geometries of the endocranial and

2002 3CLARK ET AL.: CRANIAL ANATOMY OF CITIPATI

several presumably pneumatic cavities in theskull can be traced.

SYSTEMATIC PALEONTOLOGY

THEROPODA MARSH, 1884COELUROSAURIA VON HUENE, 1914

OVIRAPTOROSAURIA BARSBOLD, 1976

COMPOSITION: The Oviraptorosauria as cur-rently constituted includes three taxa: Caen-agnathidae Sternberg, 1940, Microvenatorceler Ostrom, 1970, and Oviraptoridae Bars-bold, 1976 (see Sues, 1997, and Makovickyand Sues, 1998, for discussions of relation-ships among oviraptorosaurs).

GEOLOGIC DISTRIBUTION: The oldest ovi-raptorosaurian is Microvenator celer, fromthe Lower Cretaceous Cloverly Formation ofMontana (Ostrom, 1970; Makovicky andSues, 1998). Caenagnathidae are knownfrom the Upper Cretaceous (Upper Turonian)lower part of the Bissetky Formation of Uz-bekistan (Currie et al., 1993), the Upper Cre-taceous (Campanian) Dinosaur Park Forma-tion and Horseshoe Canyon Formation(Maastrichtian) of Alberta, Canada (Currie etal., 1993; Sues, 1997), and the Late Creta-ceous (Maastrichtian correlative) Hell CreekFormation of South Dakota (Triebold et al.,2000). Oviraptoridae currently are knownfrom reliably identified remains only fromLate Cretaceous formations of Mongolia andChina. Naish et al. (2001) described a pos-sible oviraptorosaur cervical vertebra fromthe Early Cretaceous of the Isle of Wight;however, we find little evidence to refer thisspecimen to this group.

Oviraptoridae Barsbold, 1976

TYPE TAXON: Oviraptor Osborn, 1924.COMPOSITION: Oviraptoridae currently in-

cludes six genera: Oviraptor Osborn, 1924,Ingenia Barsbold, 1981, Conchoraptor Bars-bold, 1986, Nomingia Barsbold et al., 2000,Citipati Clark et al., 2001, and Khaan Clarket al., 2001.

GEOLOGIC DISTRIBUTION: Oviraptorids areknown from the Djadokhta, Barun Goyot,and Nemegt formations of Mongolia andChina (Barsbold et al., 1990). The NemegtFormation overlies the Barun Goyot Forma-tion in some areas, but superpositional rela-tions between the Barun Goyot and Dja-

dokhta formations are not known. The ver-tebrate fauna of the Barun Goyot Formationwas considered to indicate a younger agethan the Djadokhta Formation (see Jerzyk-iewicz and Russell, 1991, for a review), butsubsequent discoveries have increased thenumber of taxa shared between the two for-mations (Dashzeveg et al., 1995; Gao andNorell, 2000). The ages of these three for-mations are poorly constrained, but the ver-tebrate faunas suggest that they are within aninterval equivalent to the Campanian toMaastrichtian marine stages (Lillegraven andMcKenna, 1986; Averianov, 1997; Jerzyk-iewicz, 2000).

Citipati osmolskae Clark et al., 2001

HOLOTYPE: IGM 100/978, a nearly com-plete skeleton.

TYPE LOCALITY: Djadokhta Formation atAnkylosaur Flats, Ukhaa Tolgod, Gurvan TesSomon, Omnogov Aimak, Mongolia. Precisecoordinates will be made available to quali-fied researchers on request.

REFERRED SPECIMENS: IGM 100/979, a par-tial skeleton overlying a nest, from Ankylo-saur Flats, Ukhaa Tolgod; IGM 100/971, anembryonic skeleton within an egg.

DIAGNOSIS: Oviraptorid differing fromConchoraptor in having a taller and morehighly pneumatized nasal, from Ingenia inthat metacarpal I is not extremely broad,from Oviraptor philoceratops in having ashorter skull and mandible, from O. mongo-liensis in lacking a convex crest on the fron-tals and parietals, and from Khaan mckennaiin the features listed below. Differs from allother oviraptorids where known in the anter-odorsally sloping occiput and quadrate, theparietal being much longer along the midlinethan the frontal and reaching nearly to thelevel of the anterior end of the orbit, the as-cending process of the jugal being perpen-dicular to the horizontal ramus (rather thanextending posterodorsally), the narial open-ing being more nearly circular and the as-cending process of the premaxilla being ver-tical rather than sloping posterodorsally, andthe cervical vertebrae being elongate (ap-proximately twice as long as they are wide).

REMARKS: This specimen was identifiedpreliminarily as O. philoceratops (e.g., Web-


ster, 1996), but comparison with the holotypeof the latter revealed several significant dif-ferences (see below). A postcranial skeletonthat is preserved on top of a nest, IGM 100/979, is referred to the new species and wasdescribed by Clark et al. (1999). IGM 100/979 is referred to C. osmolskae, pendingpreparation of the postcranial skeleton of theholotype, primarily on the basis of its largesize and its differences from the skeleton ofKhaan mckennai. An embryonic skeleton re-ferred to this species was given a preliminarydescription by Norell et al. (1994) and amore detailed description by Norell et al.(2001). This specimen is referred to Citipation the basis of its premaxilla, which is nearlyvertical rather than sloping posteriorly as inKhaan. An oviraptorid collected from theDjadokhta Formation at Dzamin Khong,IGM 100/42, was previously identified as O.philoceratops (Barsbold, 1981; Barsbold etal., 1990, fig. 10.1A and B), but it is moresimilar to Citipati in the shape of its pre-maxilla and circumnarial region and differsfrom O. philoceratops in the length of themaxilla and dentary (see below). It may rep-resent a second species of Citipati.

DESCRIPTION: The skull of the holotype(figs. 1–10) is complete, including both sta-pes and the paired elements of the hyoid ap-paratus. In lateral view (fig. 2), the skull isalmost rectangular, with a vertical premaxillaanteriorly. The right and left premaxillae arefused and edentulous. Ventrally (fig. 3), thepremaxilla expands transversely to form acurved, U-shaped triturating surface in ven-tral view that supports a series of parallelridges and troughs. The oral margin of thepremaxilla forms a sharp edge that bears aseries of large bony denticles, five on eachside. Posterior to these, the palatal surface ofthe premaxilla supports a pair of robust, lon-gitudinal, parasagittal ridges that extend pos-teriorly along either side of the midline to theposterior edge of the palate. The palatal sur-face broadens toward the back of the mouth,where it bears two additional ridges that lielateral and parallel to the first pair.

The large, elliptical nares are borderedventrally, anteriorly, and anterodorsally bythe premaxilla and posteriorly and postero-dorsally by the nasal. Owing to profoundforeshortening of the face, the naris lies al-

most entirely dorsal to the antorbital fossa,its rear margin lying nearly as far back as theposterior edge of the antorbital fenestra. Theposteriorly fused nasals form a complex ar-ray of pockets along the posterodorsal edgeof the narial opening. The pockets lie dorsalto a horizontal roof to the nares, and another,narrower horizontal roof overlies these pock-ets anterodorsally. A broad elliptical openingin the horizontal narial roof communicatesbetween the nares and the dorsal nasal pock-ets. Dorsal to the naris the nasal forms a ver-tical lamina, and a lip along its dorsal edgewas presumably for the attachment of cir-cumnarial soft tissues.

The large premaxilla forms much of theanterodorsal margin of the antorbital fossa.A fossa on the lateral surface of the ascend-ing ramus anterior to the naris (fig. 4) is sim-ilar to an accessory opening of IGM 100/42described by Barsbold et al. (1990). Posteriorto the naris, the premaxilla extends dorsal tothe maxilla to meet the nasal, separating themaxilla from the narial opening. The antor-bital fossa is triangular with a vertical pos-terior edge and horizontal ventral edge. Thetall, oval antorbital fenestra is separated fromthe large, triangular accessory antorbital fe-nestra by an hourglass-shaped, inset interfe-nestral bar formed by the maxilla. Dorsal tothe accessory fenestra are several smaller fe-nestrae near the dorsal margin of the fossa.The anteroventral margin of the fossa is com-posed of the maxilla, and a small splint ofmaxilla extends posterodorsally above theanterior edge of the fossa. Posteriorly, ventralto the antorbital fenestra, the fossa is bor-dered by the slender anterior process of thejugal.

The lacrimal is transversely broad andforms a concave, posteriorly facing surfacealong the anterior edge of the orbit. It is per-forated by a large lacrimal foramen. The lac-rimal canal passed through this foramen, tra-versing a laterally open trough above the in-ternal antorbital fenestra (Witmer, 1997) be-fore penetrating the side of the face to enterthe nasopharyngeal cavity just behind theposterodorsal extremity of the naris. The an-terior part of the lacrimal dorsal to the an-torbital fossa houses a broad pneumaticpocket opening anteriorly, within which liesseveral recessed pneumatic pockets.


Fig. 1. The holotype skull and mandible of Citipati osmolskae (IGM 100/978) in right lateral viewbefore final preparation. The mandible, scleral ossicles, and hyoid elements were later separated fromthe skull. Abbreviations in appendix 1.

There is no evidence of a prefrontal. Thepaired frontals contact the nasals in a simplesuture (fig. 5). The joint between the frontalsis short but highly complex. Posteriorly, thefused parietals divide the frontals for overone-half of their length. The frontal partici-pates in the supratemporal fenestra, but itlacks the mandibular adductor fossa thatdeeply excavates the dorsal surface of thefrontal in many other nonavian theropods.Laterally, it is overlapped by an anteromedial

process of the postorbital. The frontals con-tact one another along the midline to formthe dorsal surface of the anterior end of thebraincase. The ventral surface of the frontalforming the dorsal edge of the orbit housesseveral small pockets that may have beenpneumatic.

The orbit is bounded ventrally by the jugaland posteriorly by the postorbital, whichmeets the ascending process of the jugal toform a postorbital bar. The infratemporal fe-


Fig. 2. Right lateral view of the holotype cranium of Citipati osmolskae IGM 100/978. Abbreviationsin appendix 1.

nestra is subrectangular, with straight anteriorand ventral margins and rounded dorsal andposterior margins. The anterior process of thequadratojugal forms most of the ventral bor-der, and the ascending process forms most ofthe posterior border of the fenestra. The an-terior process is longer than the ascendingprocess. In addition, the quadratojugal pos-sesses a distinct posterior process, as in dro-maeosaurids. A large quadrate foramen liesbetween the quadrate and quadratojugal justdorsal to the posterior angle of the infratem-poral fenestra (fig. 6). The quadrate is mas-sive and tightly sutured to the squamosal,quadratojugal, pterygoid, and braincase. Thedistal articular surface forms two condyles ofapproximately equal surface and curvaturethat are separated by a longitudinal groove.The pterygoid flange of the quadrate is talland extends posterodorsally to meet the shortdescending process of the squamosal.

The squamosal is a complex bone thatforms the entire dorsal margin of the infra-temporal fenestra. Medially, it underlies the

parietal, extending nearly to the medial edgeof the supratemporal fossa. Anteriorly, it liesmedial to the postorbital and ventral to itsposterior process. The squamosal curves pos-terolaterally, and its distal end is forkedaround the external auditory meatus anddownturned. A small post-temporal fenestramay be present on the occipital surface be-tween the squamosal and the paroccipitalprocess as seen on the right side, althoughthis may be an artifact. The articulation be-tween the squamosal and the quadrate andquadratojugal is poorly exposed.

The postorbital is a triradiate bone formingmost of the postorbital bar. It overlies thefrontal anteriorly and the squamosal posteri-orly, and it extends lateral to the jugal on thepostorbital bar. The laterosphenoid has ashort contact with the postorbital at the an-terior end of the supratemporal fenestra. Thepostorbital forms the posterior border of afossa dorsal to the orbit extending into thepneumatic recesses of the narial region.

The parietals are fused and form the dorsal


Fig. 3. Palatal view of the holotype craniumof Citipati osmolskae IGM 100/978. Abbrevia-tions in appendix 1.

Fig. 4. Anterior view of the holotype craniumof Citipati osmolskae IGM 100/978. Abbrevia-tions in appendix 1.

surface of nearly half of the skull along themidline (fig. 5). The fused parietals narrowanteriorly except where they expand slightlyat their anterior end. The occipital portion ofthe parietal is very broad and is orientedobliquely at an angle of nearly 458 to faceposterodorsally. The parietal forms little ofthe supratemporal fossa except posteriorly,where a descending flange borders the open-ing of a dorsal extension of the dorsal tym-panic recess. This recess extends anterome-dially parallel to the occipital margin im-mediately posterior to it, and lies almost en-tirely within the parietal. The CT scans ofthe skull suggest that the recess connectswith a large space within the parietal andfrontal dorsal to the brain cavity (figs. 7, 8).

The supraoccipital is poorly preserved, buthas a low vertical midline crest and mayhave been separated from the foramen mag-num by the exoccipitals.

Within the parietal and frontals lies an ex-tensive space dorsal to the endocranial cav-ity, as revealed by CT scans (figs. 7, 8). In-deed, this space is larger in volume than theendocranial cavity itself. It is continuous an-teriorly with the circumnarial pneumatic cav-ities and is largest anteriorly, dorsal to theanterior part of the orbit. There is no appar-ent partitioning of the space, but in the an-terior part of the space radio-opaque areaslaterally indicate what may be internal struts.The space attenuates posteriorly and extendsnearly to the occiput (fig. 8), and it appearsto connect with the dorsal tympanic recessabove the otic regions of both sides.

The jugal is extremely slender. Posteriorly,it terminates about halfway along the infra-temporal fenestra, where it lies lateral to themore robust quadratojugal. It bends anter-oventrally anterior to the orbit and extends


Fig. 5. Dorsal view of the holotype craniumof Citipati osmolskae IGM 100/978. Abbrevia-tions in appendix 1.

Fig. 6. Occipital view of the holotype crani-um of Citipati osmolskae IGM 100/978. Abbre-viations in appendix 1.

anteriorly beneath the anterior end of the an-torbital fenestra. The right element was dam-aged beneath the orbit during the life of theanimal. The postorbital process of the jugalis robust and forms a broad posterior surfaceof the orbit. The quadratojugal forms muchof the bar beneath the infratemporal fenestra,where it is nearly circular in cross section,and extends anteriorly to beneath the post-orbital bar. Dorsally, the quadratojugal ex-tends to meet the squamosal, and it bordersthe quadrate foramen laterally.

The facial portion of the maxilla is verysmall and edentulous but includes a singleosseous denticle in line with the premaxillarydenticles. Posterior to the facial portion, the

maxilla forms a narrow shelf ventral to theantorbital fossa. Below this shelf are a vari-able number of large, anteroposteriorly elon-gate openings in the ventrolateral surface ofthe maxilla, two on the right side and one onthe left. Posteriorly, the maxilla contacts firstthe palatine and then the ectopterygoid alongits ventral edge. Posterolaterally, the maxillais overlain by the jugal.

Within the antorbital fossa, the maxillaforms the medial wall. A slender posterodor-sal process underlies the premaxilla andforms the dorsal edge of the accessory an-torbital fenestra. The nasal overlies the pos-terodorsal part of the maxilla laterally, andthe premaxilla in turn overlies the lacrimallaterally. The interfenestral bar was describedabove.

The palate (fig. 3) is generally similar tothat of an oviraptorid described by Elza-nowski (1999). The lateral portion of the pal-ate is oriented ventrolaterally, and mediallythe palatal portion forms a distinct poster-oventrally directed midline process that sur-rounds the anterior end of the vomer. Themaxilla forms most of the palate anterior tothis, which bears four stout longitudinal ridg-es. A deep groove borders the lateral edge ofthe lateral ridge. The large choanae lie almost


Fig. 7. CT scan transverse sections through the holotype skull of Citipati osmolskae IGM 100/978at three levels showing a large, presumably pneumatized space dorsal to the endocranial cavity and itsconnection to the dorsal tympanic recess. Arrow indicates apparent connection medially from dorsaltympanic recess to dorsal space. Abbreviations in appendix 1.

directly beneath the naris and antorbital fos-sa, such that the nasopharyngeal passagewaywas oriented almost vertically. The choanaeare divided by the vomer posterior to themaxillae.

The palatines are anteroposteriorly shortand highly modified. They meet on the mid-line along the posterior edge of the choanae,where their anterior edge curves dorsally tobecome vertical and transversely oriented.


Fig. 8. CT scan parasagittal section through the right side of the holotype cranium of Citipatiosmolskae IGM 100/978, showing the boundaries of the endocranial cavity and the large, presumablypneumatized space dorsal to it. Abbreviations in appendix 1.

Laterally, the palatine curves dorsally tomeet the maxilla and is separated from theanterior margin of the ectopterygoid by asmall, circular, nearly laterally facing subor-bital fenestra (the ‘‘postpalatine fenestra’’ ofElzanowski, 1999). The ectopterygoid is ver-tical, meeting the maxilla laterally and thepterygoid medially and abutting the palatinethroughout most of its length. The vomersare fused into a single short, solid bone thatdoes not extend posterior beyond its contactwith the pterygoid and palatine at the pos-terior edge of the choana.

The pterygoids have elongate palatal pro-cesses that do not appear to meet along themidline until they contact the vomers. Thepalatal surface of each element has a longi-tudinal concavity, confluent anteriorly withthe choana. The pterygoid flange is greatly

reduced compared to theropods with teeth. Itis anteroposteriorly elongate, and, in lateralview, it has a linear dorsal edge and a ven-trally convex, crescentic ventral edge, the ec-topterygoid forming the anterior half. Theflange does not extend laterally beyond thebody of the pterygoid, indicating that the M.pterygoideus was relatively small. There isno evidence for a subsidiary palatal fenestrabetween the pterygoid and palatine. Posteri-orly, the pterygoids converge towards themidline at the basipterygoid joint, but theydo not come into contact, suggesting that ifa contact was present it was formed by softtissue. The posteromedial edge of the ptery-goid is expanded near the midline, as de-scribed by Elzanowski (1999). The quadrateramus of the pterygoid is anteroposteriorlyshort but very tall, and with the quadrate


forms the lateral surface of a deep pocketlateral to the braincase that narrows anteri-orly. The pterygoid extends posteriorly toend medial to the mandibular articulatingsurface of the quadrate. A small, dorsome-dially opening pocket is formed dorsal to theposterior end at the level of the quadrate fo-ramen. Another pocket on the dorsal surfaceof the pterygoid lateral to the basipterygoidjoint extends anteriorly into the body of thepterygoid.

An isolated oviraptorid quadrate was de-scribed by Maryanska and Osmolska (1997),and the quadrate of Citipati is generally sim-ilar. It is oriented obliquely in lateral view,extending posteroventrally from the squa-mosal. The quadrate has a mediolaterally bi-convex mandibular trochlea. The verticalbody of the quadrate narrows dorsally, butits dorsal articulation with the squamosaland, apparently, the paroccipital process isnot exposed. The pterygoid ramus of thequadrate is very tall, equal in height to abouthalf that of the occipital region.

An epipterygoid is present on both sidesof the skull. It is much broader anteroposte-riorly than it is thick. It has a broad articu-lation ventrally with the pterygoid, where itfaces anteromedially, and it twists laterally asit narrows dorsally. Dorsally, it contacts theventrolateral surface of the co-ossified brain-case where a vertical ridge on the lateros-phenoid converges with a ridge along thecapitate process. On the left side it appearsto be fused to the braincase, but not on theright. The dorsal tip twists dorsolaterally andis expanded in the transverse plane, alignedwith the vertical ridge on the laterosphenoidabove it.

The elements comprising the braincase areco-ossified, although some areas of contactbetween bones are apparent. The basioccip-ital is notable for the posterior extension ofthe occipital condyle well beyond the pos-terior end of the foramen magnum (fig. 5).Indeed, with the floor of the braincase hori-zontal most if not all of the basioccipital liesposterior to the dorsal edge of the foramenmagnum due to the anterodorsal slope of theocciput. The condyle has a nearly flat pos-terior surface with a central depression. Thebasioccipital has a midline depression on itsventral surface. The foramen magnum is

nearly circular except for a ventral midlinedepression (fig. 6). The lateral surface of thebasioccipital is concave, and at least threesmall openings within the concavity presum-ably enter a pneumatic recess in the body ofthe bone. A large opening in this region onthe right side is presumably due to the break-age of thin-walled bone covering a cavity.

A large vacuity is present on the ventralmidline between the basioccipital and basi-sphenoid. The foramen lies mainly within thebasisphenoid, dividing it in two posteriorly.The basisphenoid and basioccipital expand attheir contact, and on the right side they ap-pear to be fused. The ventral surface of thebasisphenoid descends anteroventrally fromthis contact. Basipterygoid processes are ab-sent, although a matrix-filled area betweenthe basisphenoid and pterygoids suggeststhat poorly ossified processes may have beenpresent. The basisphenoid has several pneu-matic openings and cavities on the lateralsurface of the braincase ventral to the trigem-inal foramen. These are poorly exposed, cov-ered laterally by the epipterygoid and be-neath matrix ventrally. Several delicate strutsspan these openings.

A delicate, tall parasphenoid process ex-tends above the length of the interpterygoidvacuity, terminating slightly posterior to thelevel of the vomer. It descends and tapersanteroventrally from its tall posterior base.The lateral surface of the base has a shallowdepression dorsally that opens anterodorsal-ly, presumably the site of origin of an ocularmuscle, perhaps the M. rectus oculi ventralis.

The laterosphenoid portion of the brain-case forms a ventral floor anteriorly, and adiscrete orbitosphenoid ossification is presentanterior to it. The large, horizontal orbito-sphenoids encircle the anterior part of the en-docranial cavity, meeting on the midline dor-sally and ventrally. Posterior to the orbito-sphenoid the laterosphenoid sends a short,ventrally expanding process ventrally alongthe midline. This process forms the anterioredge of a large foramen, presumably for cra-nial nerve (CN) II, that is open ventrally. Aposterior descending process forms the pos-terior edge of this foramen and nearly con-tacts the top of the parasphenoid rostrum.The capitate process of the laterosphenoid islong and slender and extends laterally along


the ventral surface of the frontal and contactsthe postorbital. A small slit separates the pro-cess from the frontal anteriorly. A sharpridge descends posteromedially from the cap-itate process to the epipterygoid contact. Avertical ridge within the supratemporal fossapresumably separated portions of the M. ad-ductor mandibulae. The dorsal contact withthe frontal and parietal is apparent but pos-teriorly the laterosphenoid is fused to theprootic. The posterodorsal contact with theparietal slopes anterodorsally parallel to theocciput. The trigeminal opening (CN V) isrelatively small compared with dromaeosaur-ids (e.g., Velociraptor, IGM 100/976) and issmaller than CN II of this specimen. It isslightly longer than tall and exits ventrolat-erally from the braincase. The lateral surfaceof the laterosphenoid is rugose, presumablyindicating an origin of adductor muscles.

The prootic has a well-developed fossadorsally, homologous to the dorsal tympanicrecess described in dromaeosaurids (Norelland Makovicky, 1997, 1999). A horizontalswelling forms the ventrolateral border of thefossa. Dorsal to this fossa lies a large pneu-matic opening extending medially into thebraincase, as described above. The prooticapparently surrounds this opening dorsally,separating it from the parietal.

The opening in the prootic for CN VII iswell preserved on the right side, but on theleft it is apparently within a crushed area justventral to the horizontal swelling. On theright side a depression with at least two, pre-sumably pneumatic, openings lies dorsal tothe foramen for CN VII, but this depressionis absent on the left side. The otic recess ispoorly preserved on both sides, and the fe-nestrae ovale and rotundum cannot be iden-tified within it. The recess is larger on theright side than the left, and both are smallcompared with other coelurosaurs.

The contacts between the opisthotic andthe prootic and basioccipital appear to befused. A posterior tympanic recess is presentwithin the anterior surface of the paroccipitalprocess on the right side, but this region isbroken on the left. The recess continues me-dially into the braincase from within a shal-low fossa. The fossa penetrates the posteriorsurface of the paroccipital process via a nar-row, mediolaterally elongate opening. In me-

dial view, the jugular foramen and the exitsof two hypoglossal foramina are visible with-in the braincase. The hypoglossal foraminahave a short, simple course ventrolaterallyfrom the foramen magnum, opening into ashallow depression. Two ventrolaterallyopening foramina on the ventral edge of theparoccipital process presumably originatewithin the metotic foramen, and likely con-veyed CN IX and X and the associated vas-culature. The more ventromedial of the twois larger and may be the foramen vagi. Themore lateral foramen is small, similar in sizeto the hypoglossal foramina, and possibly forCN IX. The paroccipital processes are pen-dant and curve medioventrally, attenuatingdistally.

Both stapes are preserved, the left elementincompletely and the right disarticulatedfrom the skull. The stapes is elongate,straight, and slender, with a footplate that isapproximately three times the diameter of theshaft. As implied by the position of the oticnotch in the squamosal, the stapes wouldhave projected posterolaterally from the fe-nestra ovalis at an angle of about 458 andslightly ventrally.

Five scleral ossicles, from a partial scleralring, preserved in the right orbit and belowit were removed during preparation (fig. 1).They are roughly square in shape with irreg-ular margins and a gently concave medialsurface. They are similar to those of otherbasal coelurosaurs such as troodontids anddromaeosaurids.

The edentulous mandible (fig. 9) is dom-inated by a high, arching coronoid eminence.The symphysis is short and broad, with atransversely oriented, upturned anterior edge.It is slightly below the level of the mandib-ular articulation. The dorsal surface of thesymphysis is concave dorsally behind the an-terior edge, and the posterior edge of thesymphysis descends posteroventrally. Thecoronoid eminence rises abruptly from theposterior part of the symphysis and flattensdorsally. The surangular descends graduallyfrom the eminence and extends posteriorly tothe end of the mandible, completely coveringthe lateral surface of the articular. The largemandibular fenestra is divided by an anteriorprocess of the surangular that reaches ap-proximately three-quarters of the distance


Fig. 9. (A) Lateral and (B) medial views of the left hemimandible of the holotype of Citipatiosmolskae IGM 100/978. Abbreviations in appendix 1.

across the fenestra. The lateral surface of thesurangular is gently depressed anterior to themandibular articulation. The adductor fossaon the medial surface of the mandibular ra-mus is extremely large and bordered ven-trally by a slender splenial. The anterior endof the splenial is divided into dorsal and ven-tral rami in the symphysis. It overlaps theprearticular in the middle of the mandibularramus, and the prearticular extends posteri-orly to the end of the mandible, coveringmost of the medial surface of the articular.

The mandibular articulation surface is an-teroposteriorly elongate and unbounded an-

teriorly or posteriorly, suggesting that themandible was capable of anteroposteriormovement (cf. Osmolska, 1976). The surfacebears a longitudinal midline ridge and facesposterodorsally. The retroarticular processdescends posteroventrally from the articularsurface, with which it is coplanar, and itsventral edge descends slightly below the lev-el of the remainder of the mandible. Themandibular articulation is much wider thanthe remainder of the mandibular ramus, andit is bordered ventrally by the prearticularmedially and the surangular laterally.

A very small coronoid element appears to


Fig. 10. Ventral view of the holotype cranium of Citipati osmolskae IGM 100/978 showing a lefthyoid element in natural articulation. Abbreviations in appendix 1.

be present. A splint of bone on the medialsurface of the posterodorsal ramus of thedentary on both sides is in the position of acoronoid bone, although this bone is reportedas absent in other oviraptorids (Barsbold etal., 1990). A bone as small as this could beeasily lost or overlooked in poorly preservedspecimens, suggesting that it may also bepresent in other oviraptorids.

A pair of elements of the hyoid apparatuswere preserved beneath the mandible (fig.10), presumably the ceratohyals (cornuabranchiala I). They are simple structures, es-sentially elongate rods that curve dorsallyposteriorly. The anterior and posterior endsare expanded and slightly compressed me-diolaterally. They were preserved parallelingthe ventral edge of the mandible, possiblytheir natural position.

Most elements of the postcranial skeletonof C. osmolskae were described by Clark etal. (1999) for the referred specimen IGM100/979. This specimen preserves only a sin-gle vertebra, however, and lacks the ilia andmost of the scapulae, pubes, and ischia. Itdiffers from Khaan in that the first metacar-pal of the latter is reduced proximally and

the bone does not expand, whereas that ofCitipati is unreduced and expanded.

The postcranial skeleton of the holotype ofC. osmolskae is very well preserved and in-cludes representatives of nearly every ele-ment, but it has not yet been fully prepared.The cervical vertebrae are distinctly moreelongate than those of Khaan and other ovi-raptorids where known. The cervical ribs arealso more elongate, longer than the corre-sponding centrum in Citipati but shorter thanin Khaan.

A REINTERPRETATION OF THE SKULL OF THE

HOLOTYPE OF OVIRAPTOR PHILOCERATOPS

Beginning with its brief description by Os-born (1924), the morphology of the holotypeof Oviraptor philoceratops, AMNH 6517,has been poorly understood. The skull (fig.11) is crushed and abraded, and initially wasincompletely prepared, lying on the originalslab adjacent to the skeleton. It was given itsmost detailed description by Smith (1993),who attempted a reconstruction of the skulland interpreted its functional morphology.Comparison between the new oviraptorid


Fig. 11. Left lateral view of the holotype skull of Oviraptor philoceratops AMNH 6517. Abbrevi-ations in appendix 1.

material described above and the holotype ofO. philoceratops elucidates several featuresthat previously were enigmatic.

The skull of AMNH 6517 has beencrushed mediolaterally, compressing the tem-poral region. Thus, its breadth across the oc-ciput is only about half that of the skull ofthe holotype of Citipati osmolskae, which issimilar in length. The braincase and righttemporal regions have been dislocated as a

unit posteriorly. The broken pieces of the in-dividual bones have been separated from oneanother to varying degrees along a longitu-dinal axis, to a greater degree on the rightside than on the left. This is most evident inthe mandible, especially a large gap withinthe dentary on the right side. The frontals areincomplete dorsally, exposing the roof of thebraincase and the interorbital space. Theright jugal is missing except for its postor-


bital process and its anterior tip, and the leftjugal is missing posterior to the postorbitalbar, as is the left quadratojugal.

Smith (1993) did not have the benefit ofwell-preserved material for comparison, andthe Ukhaa Tolgod material suggests that sev-eral identifications and interpretations of thedistortion of the skull in that paper requiremodification. The dorsal side of the palatalcomplex is not ‘‘rotated as a unit and ex-posed on the left side of the skull’’ (Smith,1993: 367). Instead, the ectopterygoid andthe anterior end of the palatine are in the typ-ical position of oviraptorids in facing later-ally rather than ventrally (Osmolska, 1976;Elzanowski, 1999). Thus the dorsal edge ofthese bones as preserved is comparable to thelateral edge in other theropods, not the me-dial edge as Smith’s interpretation wouldsuggest. Furthermore, a bone identified as theectopterygoid is actually the quadrate (seebelow). The occiput has not been ‘‘rotated. . . so that the dorsal side faces the left sideof the skull’’ (Smith, 1993: 368). The occiputfaces posteriorly in line with the remainderof the skull and is not rotated, although it isslightly compressed mediolaterally.

Smith described the skull as amphikinetic,but it is unclear whether the joints he iden-tified were capable of movement. The ‘‘me-sokinetic hinge between the frontal and pa-rietal bones’’ (Smith, 1993: 369) is not ap-parent, as the contact between frontal and pa-rietal is not preserved and the edges of thesetwo bones are broken. By comparison, thiscontact on the skull of Citipati is oblique oneach side, with the parietal extending far for-ward along the midline, rather than trans-verse as in taxa with a hinge joint, such asmost squamates. Whether or not movementwas possible at the jugal-lacrimal or quad-rate-squamosal contacts is not readily appar-ent from the holotype of O. philoceratops,because the latter contact (preserved only onthe right side) is concealed in matrix and theexoccipital contribution to the joint is notpreserved. Furthermore, the force vectors re-constructed for the mandibular adductormuscles do not take into account the poste-rior displacement of the parietal, and thusshould be more vertically oriented.

The left quadrate was identified as the ec-topterygoid by Smith, from which he de-

duced a significant amount of distortion tothe palate. (This bone is labeled ‘‘ec’’ inSmith’s [1993] figures 2 and 4, but this labelis defined as the ectopterygoid in figure 2and the exoccipital in figure 4 of that paper;in the text [Smith, 1993: 374] it is describedas the ectopterygoid of the right side.) Theunusually round shape of the skull in occip-ital view proposed by Smith (1993: fig. 3B),with a ventrolaterally directed quadrate, isapparently due to this misidentification.However, the intact left quadrate is orientedvertically with a broad dorsal ramus, as inother oviraptorids. Furthermore, most of theventral portions of what is figured as theright squamosal (Smith, 1993: fig. 4H) areinstead parts of the quadrate. On the rightside, the lateral surface of what was identi-fied as the right quadrate is the quadratoju-gal.

The mandibular condyles of the quadrateon the left side are well preserved. The quad-rate is similar to that of IGM 100/978 in hav-ing a well-developed lateral condyle and asmaller medial condyle. The surface gener-ally is oriented transversely and convex ven-trally, but the condyles are separated by agroove running slightly posterolaterally. Asmall process extends laterally from the lat-eral condyle, on which the quadratojugalwould have articulated. Medially, the poste-rior end of the pterygoid is preserved slightlyseparated from the medial edge of the quad-rate condyle.

The body of the pterygoid is longitudinal-ly oriented and convex dorsally. The twobones diverge gradually posteriorly, althoughthe interpterygoid vacuity is not exposed.There is no evidence for an epipterygoid, asin Citipati, although this region has not beenfully prepared. The ventral part of the basi-sphenoid is not preserved, including the re-gion of the basipterygoid joint.

The suborbital process of the jugal is in-complete on both sides, and its contact withthe quadratojugal is not preserved, so thelength of both these bones is indeterminate.The anterior end of the jugal is apparent onboth sides, and does not extend anteriorly be-neath the antorbital fenestra as it does in Ci-tipati, and is thus similar to Khaan. An intactquadrate foramen is preserved on the rightside. It lies mainly in the quadrate, and the


quadratojugal is a simple, vertical strap ofbone. The foramen is moderately large andabout twice as tall as it is wide, similar insize to that of IGM 100/978.

The postorbital is well preserved and sim-ilar to that of other oviraptorids. Anterodor-sally, it curves medially around the anterioredge of the supratemporal fenestra. A portionof the right frontal adheres to the right post-orbital, leading Smith (1993) to mistakenlysuggest that the postorbital extended anteri-orly over the orbit.

Most of the basioccipital is preserved innatural position, and the foramen magnum isintact dorsal to it. The foramen is narrowerthan in other oviraptorids, probably due tomediolateral compression postmortem. Thebones figured and described by Smith (1993)as the exoccipital and supraoccipital corre-spond to this region of the skull if they arerotated 908 out of position. Thus, one of the‘‘knob-like paroccipital processes’’ (Smith,1993: 374) corresponds to the occipital con-dyle. Most of the occiput dorsal to the fora-men magnum is missing, but the position ofthe parietal suggests that this part of the oc-ciput was vertical, as in Khaan, rather thanfacing posterodorsally as in C. osmolskae.

The paroccipital process is well preservedon the right side, but is not preserved on theleft. It extends posterolaterally from the fo-ramen magnum and appears to have been de-flected slightly posteriorly, perhaps due tomediolateral compression. It is slightly pen-dant, but its distal end is missing so the ex-tent of its descent cannot be ascertained.

The shape of the narial opening is difficultto determine, because much of this region isnot preserved. However, as pointed out bySmith (1993), the ventral edge of the narialopening is preserved where it is bordered bythe premaxilla, and it is positioned above theanterior part of the antorbital fossa far fromthe anterior end of the skull. However, it isunclear that the maxilla bordered the narisbecause the posterior part of the nasal, whichseparates the maxilla from the naris in betterpreserved material, is absent. A slender splintof bone described and figured by Smith(1993), but since lost, probably representsthe only part of the nasals preserved, but itdoes not appear to have been very informa-tive. The premaxillae preserve an ascending

process at the anterior end of the skull, butit is unclear how far it rose dorsally. The cir-cumnarial region is quite variable betweenspecies of oviraptorids, so it is not possibleto reconstruct most of this region in the ho-lotype of O. philoceratops. The presence ofa bony ‘‘tooth’’ formed by the maxilla andvomer, as in other oviraptorids, cannot be de-termined as this area has not been exposed.

The dorsal surface of both frontals is miss-ing along the midline, and the missing por-tion may have been extensive. The fused pa-rietals bear a tall midline crest, and the dor-somedial inclination of the dorsal surface ofeach frontal suggests that the crest may havecontinued anteriorly to above the orbits. Thisinclination is not due to mediolateral com-pression, as the ventral surface of the frontalsis horizontal. Only the posterior part of thefrontals is preserved along the midline, fromthe ventral part of the bone in the anteriorpart of the braincase. These fragments sug-gest that the posterior part of the frontals wasfilled by a large, presumably pneumatic cav-ity as in Citipati.

The crest on the parietal lies only alongthe midline, unlike the domed crest of O.mongoliensis (Barsbold, 1986), which oc-cludes the supratemporal fenestrae. Theheight of the crest may have been accentu-ated by mediolateral crushing ventrally, butdorsally it appears to be solid and bears noindication of crushing. In lateral view, itforms a gentle arc interrupted by a missingportion anterodorsally. The anterior edge ofthe crest is preserved descending to the skullroof, but this area may have underlain thefrontal and any crest that it bore.

The oval antorbital fenestra is unlike thatof other oviraptorids in being elongate lon-gitudinally rather than vertically. A large ac-cessory antorbital fenestra is present in theanterior part of the antorbital fossa (contraSmith, 1993). The entire border is preservedon the left side but is broken on the right. Itis similar in size to that on the right side ofIGM 100/978. A small opening anterior tothe accessory opening on the left side nearthe premaxillary suture is not duplicated onthe right side, but a broader excavation onthe left side is in the same position as a smallfossa on the right side; most likely this areahas been eroded on both sides.


Below the antorbital fossa the suborbitalfenestra is exposed between the vertical ec-topterygoid and palatine. The fenestra islarger than in IGM 100/978 and other ovi-raptorids preserving this region. Much of theposterior part of the left ectopterygoid is vis-ible medial to the mandibular ramus, indi-cating that this bone is similar to that of otheroviraptorids. Thus the pterygoid flangebroadens vertically rather than horizontally.

The lateral surface of the braincase is wellpreserved on the left side but has not beenfully prepared. A shallow fossa is evident onthe prootic dorsal to the fenestra ovale, in aposition corresponding to the dorsal tympan-ic recess. The foramen ovale is exposed an-teriorly on the braincase wall and is longerthan high. The right laterosphenoid is pre-served with the rest of the braincase, whereasthe left adheres to the bottom of the left fron-tal.

The individual bones of the mandible aredifficult to differentiate. The surangular isextensive, but almost certainly does not formthe ventral edge of the posterior part of themandibular ramus (cf. figs. 2 and 3 of Smith,1993) or extend anteriorly beneath the man-dibular fenestra. The articular probably doesnot extend anteriorly beyond the quadrate ar-ticulation on the lateral surface (the bone la-beled by Smith as articular in this region ispart of the surangular). The second mandib-ular foramen, in the posterior part of the ra-mus, may be an artifact, as this feature cor-responds to an area of thin bone in the sur-angular of IGM 100/978. The spine on thesurangular that extends anteriorly into themandibular fenestra is horizontal rather thanextending anteroventrally as in IGM 100/978.

The symphysis is poorly preserved and in-completely prepared. Mediolateral compres-sion has dislocated the symphysial portion onthe left side from the remainder of the den-tary.

Smith (1993: 374) describes ‘‘a sliding ar-ticular surface on the quadrate, at its articu-lation with the articular, permitting an ante-rior/posterior shearing movement of the man-dible.’’ However, in archosaurs, the quadratecondyle is convex and articulates with a con-cave surface on the articular, so a ‘‘sliding’’joint can only be implied by an elongate sur-

face on the articular (as it indeed is in thebetter preserved oviraptorid material; Os-molska, 1976). The articular is preservedwell only on the right side, and it has notbeen prepared sufficiently to discern thelength of the articulation surface, but a dorsalconcavity evident in medial view suggeststhat the surface is not as elongate as in Ci-tipati.

The rostrum and lower jaw of O. philo-ceratops were more elongate than in otheroviraptorids, although the extent of this dif-ference is difficult to determine in its poorstate of preservation. Compared with otheroviraptorids, the premaxilla appears to be in-complete anteriorly, and would have extend-ed one or two centimeters farther than itspreserved length. The region anterior to theantorbital fenestra is longer than in other ovi-raptorids even when the gaps between dis-located pieces of the rostrum are taken intoaccount. This is evident from the greaterlength of the accessory antorbital fenestra incomparison with other oviraptorids (fig. 12).If mandibular length is reconstructed by al-lowing for the large gap between pieces ofthe right dentary (fig. 13), then the portionof the dentary anterior to the coronoid pro-cess is longer than this region in Citipati,which is typical of other oviraptorids. Fur-thermore, the descent of the dorsal edge ofthe dentary anterior to the mandibular fenes-tra is less pronounced than in other ovirap-torids.

DISCUSSION

The skull of C. osmolskae demonstratesthe presence in oviraptorids of epipterygoidand coronoid ossifications and an extensivepneumatic cavity in the dorsal part of thebraincase. An epipterygoid is known to bepresent in several groups of theropods andother fossil archosaurs (Clark et al., 1993)even though it is absent in both groups ofextant archosaurs (crocodylians and birds;De Beer, 1937; Gauthier et al., 1988). How-ever, its uneven distribution among thero-pods suggests that it may ossify only insome, perhaps older, individuals (cf. Madsen,1976). Among tetanuran theropods, thus faran epipterygoid is known only in allosaurids(Madsen, 1976), ornithomimids (Barsbold,


Fig. 12. Rostral regions (not to scale) of (A) the holotype of Citipati osmolskae IGM 100/978, (B)the holotype of ’’Oviraptor’’ mongoliensis (IGM 100/32) (reversed), (C) the holotype of Oviraptorphiloceratops AMNH 6517, and (D) a referred specimen of the caenagnathid Chirostenotes pergracilis(ROM 43250) (from Sues, 1997). The maxilla of O. philoceratops is relatively longer than in theoviraptorids (B and C) but shorter than in the caenagnathid (D), which represents the plesiomorphiccondition for oviraptorosaurians. Arrows indicate premaxilla-maxilla contact anteriorly and palatine-maxilla contact posteriorly. Abbreviations in appendix 1.

1981), tyrannosaurids (Lambe, 1904; Molnar,1973), Dromaeosaurus (Currie, 1995), andtroodontids (Norell et al., in prep.). In othertheropods this bone is generally small, me-diolaterally flattened, and triangular with ahorizontal base. The epipterygoid of C. os-molskae is the largest of any known thero-

pod, and its strongly twisted body and dorsaltip with robust muscle scars are unique.

A coronoid ossification has been consid-ered absent in oviraptorids (e.g., Barsbold etal., 1990), but the presence of a slender cor-onoid in C. osmolskae demonstrates it waspresent, although reduced in size. This sug-


Fig. 13. The holotype mandibles of Citipati osmolskae IGM 100/978 (top) and Oviraptor philocer-atops AMNH 6517 (bottom) in left lateral view (not to scale). The Oviraptor mandible has been digitallycompressed to minimize the separation at two breaks in the dentary. Lines are drawn between topolog-ically comparable points on the two mandibles to illustrate the proportionately greater length of theanterior part of the O. philoceratops mandible.

gests that it may be present in other taxa inwhich it is thought to be absent, such as basalavialans, dromaeosaurids, and troodontids.

The skull of C. osmolskae is notable forthe extensive pneumaticity of the narial re-gion and the dorsal part of the neurocranium.The unusual pneumatic recesses of ovirap-torid skulls have been noted before (e.g.,Witmer, 1997: 46), but this is the first timetheir full internal extent has been revealedwith CT scans. The extremely large spacedorsal to the area occupied by the brain hasnot been reported before, and is reminiscent

of the dorsal pneumatic space in the brain-case of the pterodactyloid pterosaur Anhan-geura (Kellner, 1996), although it is obvi-ously not homologous. The apparent inter-tympanic connection of the dorsal tympanicrecess is another feature hitherto unknown inoviraptorids or in any dinosaurs except neor-nithine birds (Witmer, 1990).

The braincase of C. osmolskae does notdiffer significantly from that of dromaeo-saurids (Currie, 1995; Norell et al., in press),differing mainly in the pneumatic features ofthe dorsal part of the neurocranium, the re-


duced basipterygoid processes, and the au-tapomorphic anterior shift to the occiput anddorsal part of the skull. The dorsal tympanicrecess is deeper than that of velociraptorinedromaeosaurids (Norell et al., submitted) andlies in a slightly more dorsomedial positionin C. osmolskae. The single, large openingbetween the basisphenoid and basioccipitalventrally is comparable to that of some ve-lociraptorines, although in others the openingis divided into two or more openings. Thestruts on the basisphenoid anterior to the tri-geminal opening and medial to the epipter-ygoid are not present in dromaeosaurids, butthey are in the same position as similar strutswithin the ‘‘lateral recess’’ of troodontids andin the otic region of therizinosaurids (Clarket al., 1994).

A specimen previously identified as O.philoceratops (e.g., Barsbold et al., 1990),IGM 100/42, is more similar to Citipati os-molskae than to the former in the shape ofits narial region and in the presence of anaccessory opening on the lateral surface ofthe ascending process of the premaxilla. Fur-thermore, its maxilla and dentary are as shortas those of other oviraptorids except O. phil-oceratops. We therefore consider it to belongto Citipati rather than Oviraptor, although itis unclear whether it represents a species dif-ferent from C. osmolskae. Because it is muchbetter preserved than the holotype of O. phil-oceratops it has often been relied upon foranatomical details of the species, so cautionshould be used in referring to previous char-acterizations of this species.

A second species of Oviraptor has beendescribed, O. mongoliensis Barsbold, 1986,but there is little evidence to place it in thisgenus. It shares with O. philoceratops a pa-rietal crest (albeit much larger), but differsfrom it in having a maxilla and dentary asshort as those of all oviraptorids except O.philoceratops. The name ’’Rinchenia’’ wasapplied to this species by Barsbold (1997),but this name has not been formalized.

Relationships among oviraptorids arepoorly understood (see Barsbold et al., 1990,2000), but our study of the holotypic skullO. philoceratops suggests that this speciesoccupies a primitive position in the family.The relatively elongate maxilla and dentaryof this specimen (figs. 12, 13) contrast with

those of all other oviraptorids, and are moresimilar to the condition in oviraptorid out-groups such as the oviraptorosaurs Chiros-tenotes (Sues, 1997) and Microvenator (Ma-kovicky and Sues, 1998) (i.e., a dentary ten-tatively identified for the latter). A primitiveposition for O. philoceratops does not nec-essarily contradict the hypothesis, based onfeatures of the ilium, that Nomingia gobien-sis may be the most primitive oviraptorid(Barsbold et al., 2000), because the skull ofNomingia and the pelvis of O. philoceratopsare not yet known.

ACKNOWLEDGMENTS

We thank the Mongolian Academy of Sci-ences, especially D. Baatar, A. Boldsukh, andR. Barsbold; the leader of the Mongolianside of the expedition, D. Dashzeveg; andour many friends with whom we haveworked in Mongolia over the years. In NewYork, we thank the members of our fieldcrews over the years and Dr. Richard Shep-ard. Amy Davidson prepared the specimen,and Michael Ellison illustrated this paper ex-cept for fig. 7, which was prepared by F.Ippolito. CT scans were done by R. Ketchamand C. Denison, and the image was pro-cessed by J. Maisano. Richard, Lynette, andByron Jaffe are thanked for their generoussupport through the years. This work is sup-ported by NSF DEB 9407999, NSF DEB9300700, NSF IIS 9874781, the NationalGeographic Society, the Epply Foundation,IREX, the Philip McKenna Foundation, theFrick Laboratory Endowment, and the De-partment of Vertebrate Paleontology at theAmerican Museum of Natural History.

REFERENCES

Andrews, R. C. 1932. The new conquest of Cen-tral Asiatic. A narrative of the Central AsiaticExpeditions in Mongolia and China 1921–1930. New York: American Museum of NaturalHistory, 678 pp.

Averianov, A. O. 1997. New Late Cretaceousmammals of southern Kazakhstan. Acta Pa-laeontologia Polonica 42(2): 243–256.

Barsbold, R. 1976. On a new Late Cretaceousfamily of small theropods (Oviraptoridae fam.n.) of Mongolia. Doklady Akademia NaukSSSR 226: 685–688. [in Russian]

Barsbold, R. 1981. Toothless dinosaurs of Mon-


golia. Joint Soviet–Mongolian PaleontologicalExpedition Transactions 15: 28–39. [in Rus-sian]

Barsbold, R. 1986. Raubdinosaurier Oviraptoren.In E.I. Vorobyeva (editor), HerpetologischeUntersuchungen in der Mongolischen Volksre-publik: 210–223. Akademia Nauk SSSR Insti-tut Evolyucionnoy Morfologii i Ekologii Zhi-votnikhim. Moskva: A.M. Severtsova. [in Rus-sian, German summary]

Barsbold, R. 1997. Oviraptorosauria. In P.J. Currieand K. Padian (editors), Encyclopedia of di-nosaurs: 505–509. New York: Academic Press.

Barsbold, R., T. Maryanska, and H. Osmolska.1990. Oviraptorosauria. In D.B. Weishampel, P.Dodson, and H. Osmolska (editors), The Di-nosauria: 249–258. Berkeley: University ofCalifornia Press.

Barsbold, R., H. Osmolska, M. Watabe, P. J. Cur-rie, and K. Tsogtbaatar. 2000. A new ovirap-torosaur (Dinosauria, Theropoda) from Mon-golia: the first dinosaur with a pygostyle. ActaPalaeontologica Polonica 45: 97–106.

Clark, J. M., M. A. Norell, and R. Barsbold. 2001.Two new oviraptorids (Theropoda: Ovirapto-rosauria) Late Cretaceous Djadoktha Forma-tion, Ukhaa Tolgod, Mongolia. Journal of Ver-tebrate Paleontology 21: 209–213.

Clark, J. M., M. Norell, and L. Chiappe. 1999. Anoviraptorid skeleton from the Late Cretaceousof Ukhaa Tolgod, Mongolia, preserved in anavian-like brooding position over an ovirapto-rid nest. American Museum Novitates 3265: 1–36.

Clark, J. M., A. Perle, and M. Norell. 1994. Theskull of Erlicosaurus andrewsi, a Late Creta-ceous ‘‘segnosaur’’ (Theropoda: Therizinosaur-idae) from Mongolia. American Museum Nov-itates 3113: 1–39.

Clark, J. M., J. Welman, J. A. Gauthier, and J. M.Parrish. 1993. The laterosphenoid bone of earlyarchosauriforms. Journal of Vertebrate Paleon-tology 13: 48–57.

Currie, P. J. 1995. New information on the anat-omy and relationships of Dromaeosaurus al-bertensis (Dinosauria: Theropoda). Journal ofVertebrate Paleontology 15: 576–591.

Currie, P. J., S. J. Godfrey, and L. Nessov. 1993.New caenagnathid (Dinosauria: Theropoda)specimens from the Upper Cretaceous of NorthAmerica. Canadian Journal of Earth Sciences30(10–11): 2255–2272.

Currie, P. J., and D. Russell. 1988. Osteology andrelationships of Chirostenotes pergracilis(Saurischia, Theropoda) from the Judith RiverFormation of Alberta, Canada. Canadian Jour-nal of Earth Sciences 25: 972–986.

Dashzeveg, D., M. J. Novacek, M. A. Norell, J.

M. Clark, L. M. Chiappe, A. R. Davidson, M.C. McKenna, L. Dingus, C. C. Swisher III, andA. Perle. 1995. Unusual preservation in a newvertebrate assemblage from the Late Cretaceousof Mongolia. Nature 374: 446–449.

De Beer, G. R. 1937. The development of the ver-tebrate skull. Oxford: Clarendon Press, 554 pp.

Dong, Z.-M., and P. J. Currie. 1996. On the dis-covery of an oviraptorid skeleton on a nest ofeggs at Bayan Mandahu, Inner Mongolia, Peo-ple’s Republic of China. Canadian Journal ofEarth Sciences 33(4): 631–636.

Elzanowski, A. 1999. A comparison of the jawskeleton in theropods and birds, with a descrip-tion of the palate in the Oviraptoridae. Smith-sonian Contributions to Paleobiology 89: 311–323.

Gao, K., and M. A. Norell. 2000. Taxonomiccomposition and systematics of Late Creta-ceous lizard assemblages from Ukhaa Tolgodand adjacent localities, Mongolian Gobi Desert.Bulletin of the American Museum of NaturalHistory 249: 1–118.

Gauthier, J., A. G. Kluge, and T. Rowe. 1988.Amniote phylogeny and the importance of fos-sils. Cladistics 4: 105–209.

Holtz, T. J. 2001. Arctometatarsalia revisited: theproblem of homoplasy in reconstructing thero-pod phylogeny. In J. Gauthier and L.F. Gall (ed-itors), New perspectives on the origin and earlyevolution of birds: proceedings of the interna-tional symposium in honor of John H. Ostrom:99–121. New Haven: Peabody Museum of Nat-ural History, Yale University.

Huene, F. von. 1914. Das naturliche System derSaurischia. Centralblatt fur Mineralogie Geo-logie und Palaontologie (B)1914: 154–158.

Jerzykiewicz, T. 2000. Lithostratigraphy and sed-imentary settings of the Cretaceous dinosaurbeds of Mongolia. In M.J. Benton et al. (edi-tors). The age of dinosaurs in Russia and Mon-golia: 279–296. New York: Cambridge Univer-sity Press.

Jerzykiewicz, T., and D. A. Russell. 1991. LateMesozoic stratigraphy and vertebrates of theGobi Basin. Cretaceous Research 12: 345–377.

Ji, Q., P. J. Currie, M. A. Norell, and S.-A. Ji.1998. Two feathered theropods from the UpperJurassic/Lower Cretaceous strata of northeast-ern China. Nature 393: 753–761.

Kellner, A. W. A. 1996. Description of the brain-case of two Early Cretaceous pterosaurs (Pter-odactyloidea) from Brazil. American MuseumNovitates 3175: 1–34.

Lambe, L. M. 1904. On Dryptosaurus incrassatus(Cope) from the Edmonton Series of the North-west Territory. Contributions to Canadian Pa-leontology 3: 1–27.


Lillegraven, J. A., and M. C. McKenna. 1986.Fossil mammals from the ‘‘Mesaverde’’ For-mation (Late Cretaceous, Judithian) of the Big-horn and Wind River basins, Wyoming, withdefinitions of Late Cretaceous North Americanland-mammal ‘‘ages’’. American MuseumNovitates 2840: 1–68.

Madsen, J. 1976. Allosaurus fragilis: a revised os-teology. Bulletin of the Utah Department ofNatural Resources 109: 1–163.

Makovicky, P., and H.-D. Sues. 1998. Anatomyand phylogenetic relationships of the theropoddinosaur Microvenator celer from the LowerCretaceous of Montana. American MuseumNovitates 3240: 1–27.

Marsh, O. C. 1884. Principal characters of Amer-ican Jurassic dinosaurs. Pt. VIII. The orderTheropoda. American Journal of Science (ser.3) 27: 329–340.

Maryanska, T., and H. Osmolska. 1997. The quad-rate of oviraptorid dinosaurs. Acta Palaeonto-logia Polonica 42: 377–387.

Molnar, R. 1973. The cranial morphology and me-chanics of Tyrannosaurus rex (Reptilia: Saur-ischia). Ph.D. diss., University of California atLos Angeles.

Naish, D., S. Hutt, and D. M. Martill. 2001. Saur-ischian dinosaurs. 2: theropods. In D.M. Martilland D. Naish (editors), Dinosaurs of the Isle ofWight: 242–309. London: The PalaeontologicalAssociation.

Norell, M. A. 1997. Ukhaa Tolgod. In P.J. Currieand K. Padian (editors), Encyclopedia of di-nosaurs: 769–770. New York: Academic Press.

Norell, M. A., J. M. Clark, and L. M. Chiappe.1996. Djadokhta series theropods: a summaryreview. Dinofest International Symposium Pro-gram and Abstracts. Phoenix: ASU Press.

Norell, M. A., J. M. Clark, and L. M. Chiappe.2001. An embryo of an oviraptorid (Dinosau-ria: Theropoda) from the Late Cretaceous ofUkhaa Tolgod, Mongolia. American MuseumNovitates 3315: 1–17.

Norell, M. A., J. M. Clark, L. M. Chiappe, andD. Dashzeveg. 1995. A nesting dinosaur. Na-ture 378: 774–776.

Norell, M. A, J. M. Clark, D. Dashzeveg, R. Bars-bold, L. M. Chiappe, A. R. Davidson, M. C.McKenna, and M. J. Novacek. 1994. A thero-pod dinosaur embryo, and the affinities of theFlaming Cliffs dinosaur eggs. Science 266:779–782.

Norell, M. A., J. M. Clark, and P. J. Makovicky.2001. Relationships among Maniraptora: prob-lems and prospects. In J. Gauthier and L.F. Gall(editors), New perspectives on the origin andearly evolution of birds: proceedings of the in-ternational symposium in honor of John H. Os-

trom: 49–67. New Haven: Peabody Museum ofNatural History, Yale University.

Norell, M. A., J. M. Clark, P. J. Makovicky, R.Barsbold, and T. Rowe. In prep. A revision ofSaurornithoides.

Norell, M. A., and P. Makovicky. 1997. Importantfeatures of the dromaeosaur skeleton: informa-tion from a new specimen. American MuseumNovitates 3215: 1–28.

Norell, M. A., and P. Makovicky. 1999. Importantfeatures of the dromaeosaur skeleton II: infor-mation from newly collected specimens of Ve-lociraptor mongoliensis. American MuseumNovitates 3282: 1–45.

Norell, M. A., P. Makovicky, and J. Clark. Inpress. The braincase ofVelociraptor. GravesMuseum of Archeology and Natural HistoryPublications in Paleontology.

Osborn, H. F. 1924. Three new Theropoda, Pro-toceratops zone, central Mongolia. AmericanMuseum Novitates 144: 1–12.

Osmolska, H. 1976. New light on skull anatomyand systematic position of Oviraptor. Nature262: 683–684.

Ostrom, J. H. 1970. Stratigraphy and paleontologyof the Cloverly Formation (Lower Cretaceous)of the Bighorn Basin area, Wyoming and Mon-tana. Peabody Museum of Natural History YaleUniversity Bulletin 35: 1–234.

Sereno, P. 1999. The evolution of dinosaurs. Sci-ence 1999: 2137–2147.

Smith, D. K. 1990. Osteology of Oviraptor philo-ceratops, a possible herbivorous theropod fromthe Upper Cretaceous of Mongolia. Journal ofVertebrate Paleontology 3(supplement): 42A.

Smith, D. K. 1993. The type specimen of Ovirap-tor philoceratops, a theropod dinosaur from theUpper Cretaceous of Mongolia. Neues Jahr-buch fur Geologie und Palaontologie Abhan-dlungen 186(3): 365–388.

Sternberg, R. M. 1940. A toothless bird from theCretaceous of Alberta. Journal of Paleontology14: 81–85.

Sues, H.-D. 1997. On Chirostenotes, a Late Cre-taceous oviraptorosaur (Dinosauria: Theropo-da) from western North America. Journal ofVertebrate Paleontology 17: 698–716.

Triebold, M., F. Nuss, and C. Nuss. 2000. Initialreport of a new North American oviraptor. TheFlorida Symposium on Dinosaur Bird Evolu-tion: abstracts. Graves Museum of Archaeologyand Natural History Publications in Paleontol-ogy 2: 25.

Webster, D. 1996. Dinosaurs of the Gobi. NationalGeographic 190(1): 70–89.

Witmer, L. M. 1990. The craniofacial air sac sys-tem of Mesozoic birds (Aves). Zoological Jour-nal of the Linnean Society 100: 327–378.


Witmer, L. M. 1997. The evolution of the antor-bital cavity in archosaurs: a study in soft-tissuereconstruction in the fossil record with an anal-ysis of the function of pneumaticity. Journal ofVertebrate Paleontology Memoir 3: 1–73.

Zhou, Z.-H., X. L. Wang, F. C. Zhang, and X. Xu.2000. Important features of Caudipteryx—evi-dence from two nearly complete new speci-mens. Vertebrata Palasiatica 2000 10: 241–254.

APPENDIX 1

ANATOMICAL ABBREVIATIONS USED IN FIGURES

aaof accessory antorbital fenestraacc accessory openingan angularaof antorbital fenestraar articularbs basisphenoidch choanacor coronoidd dentarydtr dorsal tympanic recessec ectopterygoidemf external mandibular fenestraendo endocranial cavityeo exoccipitalep epipterygoidf frontalfm foramen magnumfo fenestra ovalefoo foramen ovalehy hyoid elementj jugall lacrimal

lat laterosphenoidm maxillan nasalnar narial openingo orbitoc occipital condyleor orbitosphenoidp parietalpa palatinepar parasphenoidpath pathologypm premaxillapnc pneumatic cavitypo postorbitalpre prearticularpt pterygoidq quadrateqf quadrate foramenqj quadratojugalrar right articularrd right dentaryrl right lacrimalrlat right laterosphenoidrpop right paroccipital processrpre right prearticularrpt right pterygoidrq right quadratersp right splenialsa surangularso supraoccipitalsp splenialsq squamosalst stapest bony ‘‘tooth’’ formed by maxilla

and vomerv vomerII, V, VII, XII openings for cranial nerves II, V,

VII, and XII

Cranial Anatomy of Citipati osmolskae (Theropoda, Oviraptorosauria ...

Documents

Transcript of Cranial Anatomy of Citipati osmolskae (Theropoda, Oviraptorosauria ...