Cranial Anatomy of Kryptobaatar Dashzevegi (Mammalia, Multituberculata), and Its Bearing on the Evolution of Mammalian Characters

Home , Chulsanbaatar, Kryptobaatar, Middle clinoid process, Multituberculata, Nemegtbaatar, Premaxilla

CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI (MAMMALIA, MULTITUBERCULATA), AND ITS BEARING ON THE EVOLUTION OF MAMMALIAN CHARACTERS

JOHN R. WIBLE Research Associate, Division of Vertebrate Zoology, American Museum of Natural History; Section of Mammals, Carnegie Museum of Natural History, 5800 Baum Boulevard, Pittsburgh, PA 15206

GUILLERMO W. ROUGIER Research Associate, Division of Paleontology, American Museum of Natural History; Department of Anatomical Sciences and Neurobiology, School of Medicine, University of Louisville, Louisville, KY 40292

BULLETIN OF THE AMERICAN MUSEUM OF NATURAL HISTORY Number 247, 124 pages, 39 ®gures, 3 tables Issued February 28, 2000 Price: $10.90 a copy

Abstract ...... 4 Introduction ...... 5 Materials and Methods ...... 9 Descriptions ...... 16 Premaxilla ...... 16 Nasal ...... 19 Lacrimal ...... 20 Frontal ...... 21 Maxilla ...... 23 Palatine ...... 28 Pterygoid ...... 30 Sphenoid Complex ...... 31 Petrosal ...... 36 Jugal ...... 43 Squamosal ...... 43 Parietal ...... 46 Supraoccipital ...... 46 Exoccipital ...... 47 Basioccipital ...... 48 Endocranium ...... 48 Mandible ...... 59 Vascular Reconstructions ...... 62 Veins ...... 62 Arteries ...... 67 Comparisons ...... 72 Snout and Palate ...... 72 Prenasal Process of Premaxilla ...... 72 Septomaxilla ...... 74 Nasal Overhang and Anterior Nasal Notch ...... 76 Nasal Foramina ...... 76 Infraorbital Foramina ...... 77 Palatine Foramina ...... 77 Facial Process of Lacrimal ...... 78 Orbitotemporal Region ...... 78 Orbital Mosaic ...... 78 Orbital Foramina ...... 80 Postorbital Process ...... 82 Nasoparietal Contact ...... 83 Jugal ...... 83 Basicranium and Lateral Braincase Wall ...... 83 Pterygopalatine Ridges and Troughs ...... 83 Ectopterygoid ...... 84 Alisphenoid ...... 85 Foramina for Branches of the Mandibular Nerve ...... 86 Venous System ...... 87 Arterial System ...... 89 Pterygoid Canal ...... 92 Petrosal ...... 93 Hypoglossal Foramen ...... 96 2 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 3

Endocranium ...... 96 Cavum Epiptericum and Primary Braincase Wall ...... 96 Hypophyseal Fossa, Tuberculum Sellae, and Jugum Sphenoidale ...... 98 Cribriform Plate ...... 99 Conclusions ...... 99 Acknowledgments ...... 104 References ...... 105 Glossary ...... 114 Index of Anatomical Terms ...... 121

We affectionately dedicate this publication to Mike Novacek, for his years of encouragement, support, and friendship. 4 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247

ABSTRACT

The cranial anatomy of the Mongolian Late Cretaceous multituberculate Kryptobaatar dashzevegi is described based on exquisitely preserved specimens collected from Ukhaa Tolgod and Tugrugeen Shireh in the Gobi Desert by joint expeditions of the American Museum of Natural History and the Mongolian Academy of Sciences. Most sutural relationships are preserved, enabling a bone-by-bone description of the skull and lower jaws exclusive of the nasal fossa and paranasal sinuses. A reconstruction of the principal components of the cranial nervous, arterial, and venous systems is facilitated by specimens with exposed endocranial surfaces. Comparisons with previously described multituberculates, other Mesozoic mammaliaforms, and extant mammals allow an assessment of major topics in the evolutionary morphology of the multituberculate and mammaliaform skull, as well as identi®cation of Kryptobaatar as an appropriate model regarding most aspects of the multituberculate skull for future phylogenetic studies. Elements previously unknown or poorly known in multituberculates are described. Included are a complete jugal on the internal surface of the zygoma; the orbital mosaic and foramina, including the optic foramen, the metoptic foramen, the transverse canal, and the foramen for the pituito-orbital vein; and the endocranium with an extensively ossi®ed primary braincase wall formed by the pilae metoptica and antotica. The latter pila is very robust compared with its ossi®ed remnants in non-mammalian cynodonts and monotremes, suggesting that it is a derived multituberculate condition. The co-occurrence of the pilae metoptica and antotica in multituberculates is thus far unique among mammaliaforms, but agrees with the morphology expected to be primitive for Mammalia. This in turn implies an independent loss of an ossi®ed pila metoptica in monotremes, marsupials, and Vincelestes or the loss of the pila metoptica in the ancestry of multituberculates and therians combined with the independent reacquisition of a neomorphic pila metoptica in multituberculates and eutherians. The absence of several elements from the multituberculate skull, controversial in nature, is con®rmed, including the prenasal process of the premaxilla, the septomaxilla, the ectopterygoid, and the orbital process of the palatine. Also con®rmed is the presence of several controversial elements in the multituberculate skull, including an alisphenoid with a reduced contribution to the braincase and an anterior lamina expanded dorsal to the alisphenoid. Competing anatomical hypotheses for several elements are addressed, including the function of the lateral pterygopalatine trough as muscle attachment and not for the auditory tube, the homology of the postorbital process on the parietal with that on the frontal, the identity of foramina in the anterior lamina as for mandibular nerve branches and not for the mandibular and maxillary nerves, and the function of the jugular fossa as primarily having housed a diverticulum of the cavum tympani and not large cranial nerve ganglia. The cranial arterial system in Kryptobaatar generally resembled that restored for other multituberculates and for other mammaliaforms, in particular the prototribosphenidan Vincelestes. Both Kryptobaatar and Vincelestes had a transpromontorial internal carotid artery, a stapedial artery that ran through a bicrurate stapes, ramus inferior, ramus superior, and an arteria diploeÈtica magna. The cranial venous system in Kryp- tobaatar resembled that described for other Mongolian Late Cretaceous multituberculates and for monotremes with the major exits of the dural sinuses having been the prootic canal and the foramen magnum. A revised diagnosis of Kryptobaatar distinguishes it from other djadochtatherians (the grouping that includes 10 of the 11 genera of Mongolian Late Cretaceous multituberculates) by a pterygoid canal either con¯uent with or barely separated from the carotid canal and a separate hypoglossal foramen.

RESUMEN

Se describe la anatomõÂa craneana del multituberculado Kryptobaatar dashzevegi, del Cre- taÂcico TardõÂo de Mongolia. Los especõÂmenes, muy bien conservados, han sido colectados en las localidades Ukhaa Tolgod y Tugrugeen Shireh, en el desierto del Gobi por expediciones conjuntas del American Museum of Natural History y la Mongolian Academy of Sciences. La mayor parte de las suturas de Kryptobaatar estaÂn conservadas permitiendo una descripcioÂn hueso por hueso del craÂneo y mandõÂbulas con excepcioÂn, de la fosa nasal y los senos para- 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 5

nasales. La reconstruccioÂn de los principales nervios craneanos, sistema arterial y venoso que se presenta es facilitada por especõÂmenes presentando la super®cie endocranial expuesta. La determinacioÂn de los toÂpicos maÂs importantes en la evolucioÂn morfoloÂgica del craÂneo de los multituberculados y otros mamaliafomes es posible a traveÂs de la comparacioÂn de Kryptobaatar con multituberculados descritos previamente, otros mamaliaformes y mamõÂferos vivientes. Estas comparaciones permiten identi®car a Kryptobaatar como un modelo apropia- do, en la mayorõÂa de los caracteres craneanos, para representar a los multituberculados en anaÂlisis ®logeneÂticos. Elementos previamente desconocidos, o mal representados en multituberculados, son descritos. Ellos incluyen: el jugal completo en la cara interna del arco cigo- maÂtico, los elementos constituyentes de la orbita y sus foramenes, incluyendo el foramen oÂptico, metoÂptico, canal transverso y vena orbitopituitaria. AdemaÂs de estos caracteres, se describe en detalle el endocranium. Este posee una pared primaria extensamente osi®cada, formada por las pilas metoÂptica y antoÂtica. La uÂltima de ellas es muy robusta si se la compara con los restos osi®cados de esta estructura en cinodontes no mamalianos y monotremas. La ocurrencia simultanea de un pila metoÂptica y antoÂtica en multituberculados es hasta ahora una caracterõÂstica uÂnica de multituberculados, pero estaÂ de acuerdo con la morfologõÂa que se pre- sume primitiva para Mammalia. Esto a su vez implica una peÂrdida independiente de una pila metoÂptica osi®cada en monotremas, marsupiales y Vincelestes,olapeÂrdida de la pila metoÂptica en el ancestro comuÂn de multituberculados y terios, combinado con la readquisicioÂn de una pila metoÂtica neomoÂr®ca en euterios. Varios elementos del craÂneo de los multituberculados cuya presencia ha sido discutida son aquõÂ reconocidos como ausentes. Estos rasgos incluyen la ausencia de un proceso prenasal del premaxilar, septomaxila, ectopterigoides y la laÂmina orbital del palatino. Hemos con®rmado tambieÂn la presencia de varios elementos o rasgos controvertidos. Entre ellos se incluyen un aliesfenoides reducido y la presencia de una extensa laÂmina anterior expandida dorsalmente al aliesfenoides. HipoÂtesis alternativas para diversos caracteres son discutidas, incluyendo la funcioÂn de los surcos pterigopalatinos, aquõÂ interpretados para insercioÂn muscular y no para el tubo auditivo; la homologõÂa del proceso postorbitario en el parietal de los multitubercualdos de Mongolia con aquel del frontal de otros multitubercualdos; la identidad de los foraÂmenes en la laÂmina anterior, interpretadas como salidas de las ramas mandibulares del trigeÂmino y no para las ramas mandibulares y maxilares de este nervio. AdemaÂs, la gran fosa jugular de los multituberculados habrõÂa albergado un divertõÂculo de la cavidad auditiva y no grandes ganglios ex- tracraneanos. El patroÂn general de las arterias craneans de Kryptobaatar es semejante a aquel restaurado en otros multituberculados y algunos mamaliaformes, en particular el prototribos- feÂnido Vincelestes. Tanto Kryptobaatar como Vincelestes tienen una caroÂtida interna de curso transpromontorial, una arteria estapedial que atraviesa un estribo con dos cruras; las ramas superior e inferior del sistema estapedial, y la arteria diploetica magna estan presentes. El sistema venoso del craÂneo de Kryptobaatar, es similar a aquel descrito en otros multituberculados del CretaÂcico TardõÂo de Mongolia y en monotremas, en los cuales los senos durales son drenados a traveÂs del canal prooÂtico y el foramen magnum. Se presenta aquõÂ una nueva diagnosis de Kryptobaatar, agregando los siguientes caracteres diagnoÂsticos: canal pterigoideo con¯uente o estrechamente asociado al canal carotideo y un foramen hipogloso separado del foramen jugular. Estos caracteres lo distingen de otros Dja- dochtatherians, el grupo que incluye con Kryptobaatar diez de los once generos de multituberculados del CretaÂcico TardõÂo de Mongolia.

INTRODUCTION Multituberculates are an extinct group of a determination accepted by some authors early mammals with a temporal range that (e.g., Godefroit, 1997). However, one of us probably spans from the Middle Jurassic (G.W.R.) has studied the specimen and be- (Freeman, 1976, 1979; McKenna and Bell, lieves it to be a theroteinid (see Sigogneau- 1997) to the Late Eocene (Kristhalka et al., Russell et al., 1986). Multituberculates also 1982; Prothero and Swisher, 1992). A pur- have a very wide geographic distribution ported multituberculate has been described and, with recent discoveries in Morocco (Si- from the Late Triassic by Hahn et al. (1987), gogneau-Russell, 1991), are known from all 6 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247 major land masses with the exception of tion, and the probably equivalent Barun Go- South America, Australia, and Antarctica. If yot Formation (Novacek et al., 1996): Bugin- gondwanatheres from the Late Cretaceous of baatar Kielan-Jaworowska and Sochava, Argentina (Bonaparte, 1986; Krause et al., 1969; Bulganbaatar Kielan-Jaworowska, 1989; Kielan-Jaworowska and Bonaparte, 1974; Catopsbaatar Kielan-Jaworowska, 1996), Madagascar (Krause and Grine, 1996; 1994; Chulsanbaatar Kielan-Jaworowska, Krause et al., 1997), and India (Krause et al., 1974; Djadochtatherium Simpson, 1925; 1997) are multituberculates as considered by Kamptobaatar Kielan-Jaworowska, 1970; the above authors, but not in the most recent Kryptobaatar Kielan-Jaworowska, 1970; Ne- account by Pascual et al. (1999), then the megtbaatar Kielan-Jaworowska, 1974; Nes- geographic distribution of Multituberculata is sovbaatar Kielan-Jaworowska and Hurum, even broader. Currently, more than 70 genera 1997; Sloanbaatar Kielan-Jaworowska, of multituberculates are recognized (McKen- 1971; and Tombaatar Rougier et al., 1997a. na and Bell, 1997), but the overwhelming Regarding the state of preservation, 6 of the majority of these are known only from iso- 11 genera are described from well-preserved, lated teeth and fragmentary jaws (Simmons, nearly complete skulls and lower jaws (Kie- 1993). lan-Jaworowska, 1970, 1971, 1974; Kielan- Among Mesozoic multituberculates, the Jaworowska et al., 1986; Kielan-Jaworowska most notable exceptions to the usually in- and Hurum, 1997), and 5 of these are also complete morphological record for the group represented by some postcranial elements are taxa collected from the continental Late (Kielan-Jaworowska, 1969; Kielan-Jawo- Cretaceous formations of Mongolia (Kielan- rowska and Gambaryan, 1994). The most Jaworowska, 1970, 1971, 1974; Kermack completely described taxa from the stand- and Kielan-Jaworowska, 1971; Kielan-Ja- point of anatomy are Nemegtbaatar gobien- worowska and Gambaryan, 1994; Dashzeveg sis and Chulsanbaatar vulgaris. They are et al., 1995; Kielan-Jaworowska and Hurum, known from most aspects of cranial anatomy 1997; Kielan-Jaworowska et al., 2000). To (Kielan-Jaworowska, 1974), including the date, 13 monospeci®c genera have been brain endocast (Kielan-Jaworowska, 1983, named from these formations, the majority of 1986), the basicranium and the cranial vas- them by Zo®a Kielan-Jaworowska, the leader culature (Kielan-Jaworowska et al., 1984, of the Polish±Mongolian Paleontological Ex- 1986; Hurum et al., 1995, 1996), the brain- peditions of the late 1960s and early 1970s. case (Hurum, 1998a), the inner ear (Hurum, Two of the 13 genera, Gobibaatar and Tu- 1998b), the nasal fossa and paranasal sinuses grigbaatar, subsequently have been regarded (Hurum, 1994), and the masticatory appara- as junior synonyms of a third, Kryptobaatar tus (Gambaryan and Kielan-Jaworowska, (see below). Another taxon, Buginbaatar 1995); in addition, the postcranial skeleton of (Kielan-Jaworowska and Sochava, 1969), both is almost fully known (Kielan-Jawo- has been said to be of Late Cretaceous or rowska and Gambaryan, 1994). early Paleocene age (Tro®mov, 1975; Clem- Although Nemegtbaatar and Chulsanbaa- ens and Kielan-Jaworowska, 1979). Howev- tar are currently the most fully described er, Martinson (1982), followed by Jerzykie- Mongolian Late Cretaceous multitubercu- wicz and Russell (1991), considered the lo- lates, Kryptobaatar dashzevegi, the subject cality where Buginbaatar was collected to be of this report, is the taxon represented by the of Late Cretaceous age, representing ``Ne- most specimens. It is the most abundant megt times.'' The discovery of dinosaur egg- mammal both in the traditional localities of shells at this locality by recent Joint Pale- the Djadokhta Formation (Kielan-Jaworow- ontological Expeditions from the Mongolian ska, 1974) and in Ukhaa Tolgod (®g. 1), and, Academy of Sciences and the American Mu- in fact, is represented by more skulls than seum of Natural History supports a Late Cre- any other Mesozoic mammal. Ukhaa Tolgod taceous age for Buginbaatar. Therefore, we is the locality showing the highest concen- currently recognize 11 genera of Mongolian tration of mammalian skulls and skeletons Late Cretaceous multituberculates from the from any Mesozoic site; it was discovered in Nemegt Formation, the Djadokhta Forma- July 1993 by the Joint Mongolian±American 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 7

Museum Expeditions (Dashzeveg et al., neous with Bayn Dzak. These authors noted 1995). The Djadokhta Formation is thought many resemblances between Tugrigbaatar to be of early Campanian age (Jerzykiewicz and Kryptobaatar, and subsequently the for- et al., 1993; Dashzeveg et al., 1995; Rougier mer was declared a junior synonym of the et al., 1997a; Averianov, 1997). The age of latter (Rougier et al., 1997a; Kielan-Jawo- Ukhaa Tolgod relative to Djadokhta and Ba- rowska and Hurum, 1997). In 1980, Kielan- run Goyot is under study (Dingus et al., in Jaworowska recognized Gobibaatar parvus prep.), but the faunal contents of these units (Kielan-Jaworowska, 1970), originally de- are more uniform than previously thought scribed as the only ptilodontoid from the (Novacek et al., 1996). Late Cretaceous of Mongolia, to be a junior Kryptobaatar dashzevegi was named by synonym of K. dashzevegi. Further remarks Kielan-Jaworowska (1970) based on an in- on the masticatory apparatus of K. dashze- complete skull (missing the braincase) and vegi have been published by Gambaryan and lower jaws recovered in the Djadokhta For- Kielan-Jaworowska (1995), and other incom- mation, Bayn Dzak, Shabarakh Usu (the plete skulls from Bayn Dzak have been ®g- ``Flaming Cliffs''; ®g. 1). It was allocated to ured by Kielan-Jaworowska and Gambaryan the suborder Taeniolabidoidea, family Eucos- (1994) and Kielan-Jaworowska and Hurum modontidae, as were most of the other mul- (1997). A revised diagnosis of K. dashzevegi tituberculates from the Djadokhta and Barun has been offered by Kielan-Jaworowska and Goyot Formations (Clemens and Kielan-Ja- Hurum (1997). worowska, 1979). Recently, in fact, Rougier Regarding the postcranium, a partial skel- et al. (1997a) have suggested that nine of the eton, including nearly complete hindlimbs Mongolian Late Cretaceous multituberculate and pelvis, an incomplete scapulocoracoid, genera are included in a monophyletic group- and some damaged cervical, lumbar, sacral, ing along with Pentacosmodon from the and caudal vertebrae, was collected in asso- North American Paleocene (see ®g. 38); ex- ciation with a skull of Kryptobaatar dash- cluded was Buginbaatar, which was closer zevegi at Bayn Dzak during the 1968 Polish± to the taeniolabidids, and Nessovbaatar, Mongolian expedition. Described in this which had not yet been named. Kielan-Ja- specimen for the ®rst time in multitubercu- worowska and Hurum (1997) proposed a lates were epipubic bones (Kielan-Jaworow- somewhat modi®ed monophyletic Mongo- ska, 1969). Other details of the pelvis were lian assemblage, which they named the Dja- reported by Kielan-Jaworowska (1979) and dochtatheria. Included were the same taxa in of the remaining postcranial elements by the unnamed Mongolian clade of Rougier et Kielan-Jaworowska and Gambaryan (1994). al. (1997a) plus Nessovbaatar and Paraci- Partial humeri, an ulna, and ribs were found mexomys from the North American Creta- and described with the holotype skull of Tu- ceous. Kryptobaatar was assigned by Kielan- grigbaatar saichanensis by Kielan-Jawo- Jaworowska and Hurum (1997) to the Dja- rowska and Dashzeveg (1978), and a com- dochtatheriidae, along with Djadochtathe- plete humerus of Kryptobaatar was de- rium, Catopsbaatar, and Tombaatar. scribed by Kielan-Jaworowska (1998). We Based on additional specimens from Bayn add that Kryptobaatar also provides the ®rst Dzak, some additions and corrections to the evidence of a tarsal spur in multituberculates. original description of Kryptobaatar dash- The bone that Kielan-Jaworowska and Gam- zevegi subsequently were made by Kielan- baryan (1994: ®g. 2A) labeled as a possible Jaworowska and co-authors. For example, fragment of dentary between the distal ends the slender palatal vacuities included in the of the tibia and ®bula is a displaced tarsal original diagnosis were found to be artifac- spur (personal obs.) as occurs in monotremes tual (Kielan-Jaworowska and Dashzeveg, (Grif®ths, 1978), the gobiconodontid Gobi- 1978). Comparisons were made by Kielan- conodon (Jenkins and Schaff, 1988), and the Jaworowska and Dashzeveg (1978) with the symmetrodont Zhangheotherium (Hu et al., skull of a new eucosmodontid, Tugrigbaatar 1997). saichanensis, from Tugrugeen Shireh or Too- The current report documents the anatomy greeg (®g. 1), considered to be contempora- of several well-preserved specimens of 8 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247

Fig. 1. Map of Mongolia (A) with inset of south-central region (B) indicating the major fossil localities visited by the joint expeditions of the Mongolian Academy of Sciences and the American Museum of Natural History. The specimens of Kryptobaatar dashzevegi described here come from Ukhaa Tolgod and Tugrugeen Shireh. 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 9

Kryptobaatar dashzevegi collected over the the skull and lower jaws that can serve as an last few years by the Joint Mongolian±Amer- appropriate model for future comparative ican Museum Expeditions (®g. 2). Two spec- morphological and phylogenetic studies on imens have been ®gured elsewhere (Novacek Multituberculata and on Mammaliaformes in et al., 1994; Dashzeveg et al., 1995: ®g. 3a, general, the latter being the clade comprising misidenti®ed as Chulsanbaatar; Rougier et the common ancestor of the Liassic taxon al., 1996c: ®gs. 1±4; Monastersky, 1996; Morganucodon and Mammalia plus all its Rose and Sues, 1996; Novacek, 1997: ®g. 2), descendants (Rowe, 1988). but thus far only the anatomy of the middle- ear ossicles has been described in detail MATERIALS AND METHODS (Rougier et al., 1996c). Although K. dashzevegi is known from more skulls than any To date, more than 400 multituberculate other Mesozoic mammal, its entire cranial specimens have been collected by the Joint anatomy had yet to be documented in detail, Mongolian±American Museum Expeditions. and therefore we present a bone-by-bone de- The majority are in the size range of Krypto- scription of the skull exclusive of the ear os- baatar dashzevegi, but only a handful have sicles and reconstruct the course of the major been prepared suf®ciently to enable precise cranial nerves and vessels. The dentition of taxonomic determination. Our descriptions K. dashzevegi has been described by Kielan- herein are based on the following specimens Jaworowska (1970) and Kielan-Jaworowska of K. dashzevegi, all of which are catalogued and Dashzeveg (1978), and will not be re- in the Institute of Geology, Ulaan Baatar. The described here, although relevant dental mea- ®rst two specimens comprise the bulk of the surements are presented in table 1. descriptions of external surfaces and the Because postcranial and cranial remains measurements in tables 1 and 2; the remain- (exclusive of teeth) are rarely recovered for ing specimens reveal structures of the endo- multituberculates, the few well-preserved cranium and/or selected external features. taxa that have been described are central to (1) PSS-MAE 101 (®gs. 2, 3): Anterior our understanding of the morphology and relationships of the group. Included among the half of a fully articulated skeleton collected most thoroughly studied multituberculates, in from Ukhaa Tolgod, including a pristine addition to several of the Mongolian Late skull with lower jaws attached, the anterior Cretaceous taxa (e.g., Nemegtbaatar gobien- portion of the vertebral column, and both sis, Chulsanbaatar vulgaris), is the taenio- shoulder girdles and forelimbs. Some exter- labidid Lambdopsalis bulla from the Paleo- nal surfaces of the skull are not accessible; cene of China, whose anatomy has been the matrix has not been removed from part of the subject of several reports (e.g., Miao, 1986; left ear region, the area of the foramen mag- Kielan-Jaworowska and Qi, 1990), including num, and part of the palate in order to pre- monographic treatment of the skull (Miao, serve the stylohyal, atlas and axis, and lower 1988). Lambdopsalis is a highly specialized jaws, respectively. The skull has been ®gured form, perhaps a burrower, with an expanded elsewhere in different views (Novacek et al., vestibular apparatus; it exhibits many apo- 1994; Rougier et al. 1996c: ®g. 2; Dashzeveg morphic cranial features compared with the et al., 1995: ®g. 3a; Monastersky, 1996; No- Mongolian Late Cretaceous taxa (Miao, vacek, 1997: ®g. 2), and the forelimb has 1988; Luo, 1996; Kielan-Jaworowska and been ®gured in dorsolateral view (Dashzeveg Hurum, 1997). Despite its extreme speciali- et al., 1995: ®g. 3a; Rose and Sues, 1996). zations, Lambdopsalis has served as a model (2) PSS-MAE 113 (®gs. 4, 5): Skull and for Multituberculata in some phylogenetic the left lower jaw collected in 1991 from Tu- studies (e.g., Wible, 1991; Meng, 1992; grugeen Shireh. Missing are parts of the cra- Meng and Wyss, 1995) because of the pau- nial vault, the left alisphenoid, part of the city of material and of comprehensive, de- pterygoid, and the left zygomatic arch. The tailed descriptions of other taxa. Therefore, a skull has been ®gured elsewhere in ventral goal of our report on Kryptobaatar dashze- view, as have closeups of the fragmentary vegi is to provide a detailed description of left malleus-ectotympanic complex and the 10 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 11 fragmentary right stapes (Rougier et al., (2) PSS-MAE 128: Partial skull without 1996c: ®gs. 1, 3, 4). lower jaws collected in 1993 from Ukhaa (3) PSS-MAE 123 (®g. 25): Partial skull Tolgod. Missing are the left braincase and with lower jaws collected in 1994 from Uk- zygomatic arch. Little matrix has been re- haa Tolgod (Sugar Mountain). Missing are moved from the exterior of the skull, but the the braincase roof, right braincase wall, oc- interior has been prepared. Exposed is the ciput, right zygomatic arch, and posterior inner surface of the right petrosal and ante- part of the right lower jaw. Preparation has rior lamina, as well as parts of the sphenoid exposed the inner surfaces of the skull base complex. and left braincase. (3) PSS-MAE 134: Natural nasal endocast (4) PSS-MAE 124: Partial skull with low- collected in 1996 from Ukhaa Tolgod (Zo- er jaws collected from Ukhaa Tolgod. Miss- ®a's Hill). ing are the braincase roof and left zygomatic Kielan-Jaworowska and Hurum (1997) arch. Little matrix has been removed from recognized two species of Kryptobaatar, K. the exterior of the skull, but the interior has dashzevegi and K. saichanensis, with the been partially prepared. Exposed is the inner most substantive difference being the cusp surface of part of the sphenoid complex. formula (in the labial and lingual rows) of Specimen is not ®gured. the second lower molar (3:2 in K. dashzevegi (5) PSS-MAE 125 (®g. 26): Partial skull versus 4:2 in K. saichanensis). We have in- with fragments of both lower jaws collected spected the single specimen of K. saichanen- from Ukhaa Tolgod. Missing are the left zy- sis (GI SPS 8-2 PST) and found three cusps gomatic arch, left braincase wall, and most in the labial row on the left second lower of the occiput. Preparation has exposed most molar and three cusps plus an accessory cusp of the left orbital wall, the braincase ¯oor, on the right side. Given that the other differ- and the inner surface of the right braincase ences identi®ed by Kielan-Jaworowska and wall. Hurum (1997) are minor (e.g., the length of (6) PSS-MAE 127: Skull and lower jaws the mandibular ascending ramus), we rec- collected in 1993 from Ukhaa Tolgod. Matrix ognize K. saichanensis as a junior synonym has been removed only from selected exter- of K. dashzevegi. Our specimens of Krypto- nal surfaces. Exposed is the tip of the ros- baatar also exhibit differences. For example, trum, right upper and lower jaws, and the in PSS-MAE 101, the sphenopalatine fora- posterior part of the roof of the braincase and men lies between the frontal and maxilla, zygomatic arch. Specimen is not ®gured. whereas in 113 it is wholly within the max- In addition, three other indeterminate illa. Additionally, in PSS-MAE 101 the jugal skulls resembling Kryptobaatar from the is concealed in lateral view by the maxilla Mongolian±American Museum Expeditions and squamosal, whereas in 113 a thin sliver are included here for comparative purposes, of jugal is exposed along the dorsal edge of but are not ®gured. the zygomatic arch. PSS-MAE 113, collected (1) PSS-MAE 126: Skull without lower in Tugrugeen Shireh, is slightly larger than jaws collected in 1994 from Gilbent Wash 101 from Ukhaa Tolgod (see tables 1 and 2); (locality 20) near Ukhaa Tolgod. Missing are this is in agreement with the purported larger both zygomatic arches and exoccipitals, as size of K. saichanensis also from Tugrugeen well as the tip of the rostrum. Also missing Shireh compared to K. dashzevegi from Bayn is part of the cranial vault, which exposes the Dzak (Kielan-Jaworowska and Dashzeveg, posterior part of the brain endocast and vas- 1978). We believe that these and a few other cular molds. Matrix has been removed from differences in our specimens are minor and the right orbit, but not the from left. we treat the specimens as conspeci®c.

← Fig. 2. Kryptobaatar dashzevegi PSS-MAE 101 in left lateral and two oblique dorsolateral views (from top to bottom), prior to removal of the skull and lower jaws from the block containing the postcranium. 12 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247

Fig. 3. The skull and lower jaws of Kryptobaatar dashzevegi PSS-MAE 101 in (clockwise from upper left) dorsal, ventral, anterior, posterior, right lateral, and left lateral views. 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 13

Fig. 4. The skull of Kryptobaatar dashzevegi PSS-MAE 113 in (clockwise from upper left) dorsal, ventral, anterior, posterior, right lateral, and left lateral views. 14 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247

Fig. 5. The left lower jaw of Kryptobaatar dashzevegi PSS-MAE 113 in lateral and medial views.

Different authors (e.g., Kermack et al., tuberculates have been described by numer- 1981; Novacek, 1986; Miao, 1988) have ous authors (e.g., Broom, 1914; Simpson, used various organizational schemes in de- 1937; Hahn, 1969; Kielan-Jaworowska, scribing the skulls of fossil mammals. Here 1970; Miao, 1988), and the anatomical ter- we provide a bone-by-bone description of the minology used has not always been consis- exterior of the skull of Kryptobaatar. In ad- tent. In general, our usage of terminology dition, the ventral endocranial surfaces visi- agrees closely with that used by Kielan-Ja- ble in several specimens are described as a worowska and co-authors (e.g., Kielan-Ja- single anatomical unit. Also provided are re- worowska, 1970, 1974; Kielan-Jaworowska constructions of the major cranial nerves, ar- et al., 1986; Gambaryan and Kielan-Jawo- teries, and veins. Following the descriptions, rowska, 1995), but for the auditory region comparisons of individual regions of the and cranial vasculature we follow Wible skull are made with those of other multitu- (1987, 1990), Rougier et al. (1992, 1996a, berculates and mammaliaforms. Emphasis is 1996c), Wible and Hopson (1993, 1995), and on characters used in previous phylogenetic Wible et al. (1995). In addition, for many of analyses (e.g., Simmons, 1993; Wible et al., the endocranial structures not described pre- 1995; Rougier et al, 1996a, 1996c). viously in multituberculates, we follow the Over the years, cranial remains of multi- terminology in the dog (Evans and Christen- 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 15 sen, 1979). Our usages for the cranial foram- with some authors (e.g., Meng and Wyss, ina and canals are speci®ed in a glossary at 1995) promoting a sister-group relationship the end. We have also provided an index for for multituberculates and monotremes, and the major page citations for most anatomical others (e.g., Rougier et al., 1996a, 1996c; Hu terms used. For the dentition, we use the ab- et al., 1997; Ji et al., 1999) allying multitu- breviations ``I, P, M'' and ``i, p, m'' to refer berculates more closely with therians. Nev- to upper and lower incisors, premolars, and ertheless, under the terminology proposed by molars, respectively. Regarding the numera- Witmer (1995), the extant phylogenetic tion of the upper premolars, we follow Sim- bracket (minimally, the ®rst two extant out- mons (1993) and designate the ®ve premo- groups) for multituberculates consists of lars present in primitive multituberculates monotremes and therians. Inferences that are (e.g., the Late Jurassic paulchoffatiines, based on soft-tissue structures and osteolog- Hahn, 1993) as P0±P4. The ®rst upper pre- ical correlates occurring in both extant out- molar having been lost, the remaining four groups are considered more decisive than in Kryptobaatar are designated P1±4. How- those occurring in only one. ever, our usage of the term premolars does not necessarily imply strict homology with Institutional Abbreviations the premolars of therians. There is no in- AMNH Department of Vertebrate Paleon- stance in which the multituberculate P4 is tology, American Museum of Nat- known to have a deciduous precursor ural History (Greenwald, 1987; Hahn and Hahn, 1999); a FMNH Department of Geology, Field Mu- deciduous p4 has only been reported for one seum of Natural History, Chicago form, Kuehneodon dietrichi (Hahn, 1978), GI Geological Institute, Mongolian but this was based on a single broken lower Academy of Sciences, Ulaan Baa- jaw that contained no replacement teeth. Giv- tar en that there are no unequivocal examples of IVPP Institute of Vertebrate Paleontolo- replacement of the upper and lower last pre- gy and Paleoanthropology, Chi- molars in multituberculates, it is possible that nese Academy of Sciences, Beijing these teeth are molars, i.e., unreplaced per- MAE Mongolian±American Museum manent teeth (see Clemens and Lillegraven, Expedition 1986; Luckett, 1993). MCZ Museum of Comparative Zoology, Several authors (e.g., Bryant and Russell, Harvard University, Cambridge 1992; Witmer, 1995) have recently proposed PIN Institute of Paleontology, Acade- explicit methods for reconstructing soft tis- my of Sciences, Moscow sues in fossils and for evaluating levels of con®dence in those inferences. In formulat- PSS (SPS)Paleontological and Stratigraphic ing hypotheses about soft-tissue reconstruc- Section of the Geological Institute, tion here, we follow recent phylogenetic Mongolian Academy of Sciences, analyses (Rowe, 1988, 1993; Wible et al., Ulaan Baatar 1995; Rougier et al., 1996a, 1996c; Hu et al., USNM Department of Paleobiology, Na- 1997; Ji et al., 1999) that identify multitu- tional Museum of Natural History, berculates as members of Mammalia (®g. Smithsonian Institution, Washing- 39), the group including the last common an- ton, D.C. cestor of monotremes and therians plus de- V.J. Museum of the Geological Ser- scendants (Rowe, 1988). We acknowledge, vice, Lisbon however, that the relationships of multituber- YPM-PU Yale Peabody Museum, Princeton culates within Mammalia remain controver- University Collection, New Haven sial (e.g., Rougier et al., 1996b; Sereno and ZPAL Institute of Paleobiology, Polish McKenna, 1996; Meng and Wyss, 1996), Academy of Sciences, Warsaw 16 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247

DESCRIPTIONS As accompaniment to the following de- each other. It forms the walls of the anterior scriptions, composite views of the skulls and part of the nasal cavity and bears two inci- lower jaws of PSS-MAE 101 and 113 are sors, a larger I2 and a smaller I3. The pre- shown in ®gure 3 and 4±5, respectively. Ste- maxilla is complete in PSS-MAE 101 (®g. reophotographs of anterior, dorsal, lateral, 6), but its anterior portion is missing in PSS- ventral, and posterior views of these speci- MAE 113 (®g. 7). mens are in ®gures 6±17. Enlargements of In lateral view (®gs. 10±13, 33), the facial particular regions (i.e., orbit, basicranium, component comprises approximately one- zygoma, endocranium, and mandible) are in third of the preorbital skull length. It forms ®gures 18±23, 25±26, 29, and 30. Recon- the anterolateral wall of the nasal cavity and structions of the skull of Kryptobaatar dash- the ventral half of the margin of the external zevegi in various views are in ®gures 27, 31± nares. Facets for the septomaxilla on the ex- 35, 36A, and 37A, with vascular reconstruc- ternal narial aperture are lacking. The ante- tions in ®gures 36B and 37B. Abbreviations rior margin of the facial component is ver- used in the ®gures are listed in table 3. tical, ¯ush superiorly with the anterior tip of the nasal and inferiorly with the anterior edge PREMAXILLA of the I2 alveolus; consequently, the rim of The premaxilla is a large bone with facial the external nares is circular in outline and and palatal components of near equivalent opens directly anteriorly without any lateral size oriented approximately at right angles to exposure. The facial component contacts the maxilla posteriorly and the nasal dorsally. The premaxillary±maxillary suture is vertical near the palatal margin, roughly halfway be- 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 17

tween I2 and P1. Near the nasal, the suture (®gs. 6, 31), the anterior part of the facial curves posteriorly, and a slender tongue of component, which is complete on both sides, the premaxilla, the posterodorsal process, ex- is very narrow dorsoventrally and shows no tends between the maxilla and nasal to the sign of a prenasal process, which, if present, level of I3. In rostral view in PSS-MAE 101 would have formed an internarial bar. Con- 18 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247

Fig. 6. Stereophotograph of the skull and lower jaws of Kryptobaatar dashzevegi PSS-MAE 101 in anterior view, with accompanying line drawing. Gray pattern represents matrix; parallel lines denote breakage. Abbreviations: con (mandibular) condyle; fr frontal; gl glenoid fossa; iof infraorbital foramen; mf mental foramen; mx maxilla; na nasal; pmx premaxilla; sq squamosal.

tained within the anterior part of the facial ris, which raised the upper lip (Evans and component is the alveolus of I2, and extend- Christensen, 1979). ing posteriorly from the anterior part to the In ventral view (®gs. 14, 15, 34), the pre- premaxillary±maxillary suture is a small hor- maxilla occupies roughly one-third of the izontal furrow, perhaps representing an at- palate, with the most conspicuous features tachment for facial musculature (``muf'' in being the alveolus of I2 anteriorly, and the ®g. 14). The most probable occupant of such alveolus of I3 and the incisive foramen pos- a furrow was the musculus incisivus superio- teriorly. The lateral margin is thickened to 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 19

Fig. 7. Stereophotograph of the skull of Kryptobaatar dashzevegi PSS-MAE 113 in anterior view. form a sharp, ventrally projecting ridge. The of equivalent size is the incisive foramen central portion of the palatal component, be- (``inf'' in ®g. 14). This aperture is oval and tween the alveoli of I2 and I3, is very long lies between the premaxilla and maxilla (®g. and concave ventrally. Subdividing this part 34). Separating the left and right incisive fo- of the palatal component into two longitu- ramina and meeting on the midline are the dinal depressions is a cord-shaped ridge, the broad palatal processes of the premaxillae. thickening of the premaxilla (``tpmx'' in ®g. 14) of Kielan-Jaworowska et al. (1986). Of NASAL the two depressions divided by the thickening of the premaxilla, the medial is the deep- The nasals are elongate bones that form er; it lies on the midline and meets its com- nearly the entire skull roof in the preorbital plement of the opposite side. The alveolus of area. They are severely damaged in PSS- I2 is large and although close to the midline, MAE 101, 113, 123, and 127, but the pres- is separated from the I2 alveolus of the op- ervation is suf®cient to restore their sutural posite side by considerable bone. The left relationships. and right I2s are directed ventromedially to- In dorsal view (®gs. 8, 9, 32), the nasal on ward one another and come into contact ven- its lateral margin contacts the premaxilla and trally, leaving a triangular space between maxilla and on its posterior margin, the lac- them and the edge of the premaxilla visible rimal and frontal. The sutural relationships of in rostral view (®gs. 6, 31). A similar ar- the nasal are as follows. Where the nasal rangement with contact between the left and contacts the premaxilla, the lateral margin of right I2s has been illustrated for Sloanbaatar the nasal is subparallel to the medial margin. from the Mongolian Late Cretaceous (Kie- However, where the nasal contacts the max- lan-Jaworowska, 1971: ®g. 9) and for Tae- illa, the nasal is strongly expanded laterally. niolabis taoensis from the North American The lateral three-quarters of the nasal's pos- Paleocene (Granger and Simpson, 1929: ®gs. terior suture are roughly transverse, but in 5, 6). The portion of the premaxilla in front the medial quarter there is a slender tongue of the alveolus of I2 in Kryptobaatar slopes of frontal that projects into a short, broad U- posteroventrally from the narial aperture to shaped notch between the two nasals. The the alveolus. The smaller alveolus for I3 lies frontal is overlapped slightly by the nasal, as roughly halfway between the median suture is evident on the right side of PSS-MAE 113 and the labial margin. The I3 alveolus is where the posterolateral part of the nasal is formed entirely by the premaxilla and is sep- missing and the frontal is exposed (®g. 9). arated from the premaxillary±maxillary su- Nasal foramina (Simpson, 1937) are present ture by bone approximating the diameter of on the dorsal surface (``naf'' in ®gs. 8, 10). the alveolus. Medial to the I3 alveolus and However, because of damage, the number 20 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247

Fig. 8. Stereophotograph of the skull of Kryptobaatar dashzevegi PSS-MAE 101 in dorsal view, with accompanying line drawing. Gray pattern represents matrix; parallel lines denote breakage. Ab- breviations: al anterior lamina; ano nasal notch; con (mandibular) condyle; cor coronoid process; exoc exoccipital; fdac foramen of dorsal ascending canal; fr frontal; frt foramen for ramus temporalis; ju jugal; juf jugal facet; lac lacrimal; mx maxilla; na nasal; naf nasal foramen; pa parietal; pmx premaxilla; sgf supraglenoid foramen; sq squamosal; sup supraoccipital.

and position of the foramina cannot be fully LACRIMAL ascertained. In the anterior third of the nasals in PSS-MAE 101, there are two foramina on The lacrimal occupies the anterior margin the right side and three on the left, and they of the orbit and has orbital and facial pro- are not arranged symmetrically between the cesses, with the latter being the larger of the two. A complete lacrimal is not preserved in two sides (®g. 8). On the basis of more com- either PSS-MAE 101 or 113; the most completely preserved nasals, Kielan-Jaworowska plete is on the left side of the former (®gs. and Hurum (1997) reported two pairs of na- 8, 18), which is missing only the anterior- sal foramina for Kryptobaatar dashzevegi. most part. The anterior edge of the nasal is not fully The orbital exposure is wedge-shaped and preserved in any MAE specimens, but does not extend ventrally far from the orbital enough is present to report the presence of a rim, contributing mostly to the overhanging notch there (``ano'' in ®gs. 8, 12), especially orbital roof (®gs. 16, 18, 33). It contacts the in PSS-MAE 127, which we term the ante- maxilla laterally and inferiorly, and the fron- rior nasal notch following Lillegraven and tal medially. The facial exposure (®gs. 8, 32) Krusat (1991). Additionally, enough is pre- is transversely convex and subrectangular in sent to preclude a substantial nasal overhang shape (see Kielan-Jaworowska and Hurum, of the external narial aperture and to exclude 1997: ®gs. 2B, C). It contacts the maxilla the presence of an internarial bar, as is also anterolaterally, the nasal anteromedially, and indicated by the morphology of the premax- the frontal posteromedially. The part of the illa. lacrimal forming the orbital rim is thickened, 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 21

Fig. 8. Continued. and on its ventrolateral aspect bears a single The horizontal components of the frontals small lacrimal foramen, which is not subdi- (®gs. 8, 9, 32) are ¯at and meet on the dorsal vided (``lacf'' in ®g. 16). The foramen is midline at a slightly irregular suture. They high in the orbit, surrounded by strong crests, roof the posterior part of the nasal cavity and and opens near the contact between the lac- the anterior part of the cranial cavity. Ante- rimal and the facial process of the maxilla. riorly, the frontals contact the lacrimals and send a tongue-shaped process anteromedially FRONTAL into the preorbital area between the nasals. The frontal has a horizontal component in Posteriorly, the frontals narrow, extending to the skull roof and a vertical component in the the level of the postorbital processes on the orbit. Both components are well preserved in parietals and contacting the parietals largely PSS-MAE 113, but the horizontal is dam- through a broad U-shaped suture; lateral to aged in PSS-MAE 101 (®g. 8). the arms of the U is a small prong of the 22 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247

Fig. 9. Stereophotograph of the skull of Kryptobaatar dashzevegi PSS-MAE 113 in dorsal view. frontal that projects posterolaterally, forming there is a low, raised area directed dorsoven- the medial rim of the supraorbital notch trally in the same position as is the well-de- (``son'' in ®gs. 10, 12). Along their suture, ®ned, high orbital ridge of Kielan-Jaworow- the frontal and parietal do not overlap. The ska et al. (1986) that occurs in some other lateral border of the horizontal component of Mongolian Late Cretaceous multitubercu- the frontal forms a thickened ridge contrib- lates (e.g., Nemegtbaatar). Interestingly, the uting to the orbital rim, the supraorbital crest orbital ridge is present in a specimen of of Miao (1988). In the orbital margin, the Kryptobaatar from Tugrugeen Shireh, frontal is widest anteriorly where it contacts GISPS 8-2 PST (Kielan-Jaworowska, per- the lacrimal and tapers posteromedially, be- sonal commun.), originally identi®ed as Tu- ing narrowest where it contacts the anterior grigbaatar saichanensis by Kielan-Jawo- prong of the parietal in front of the postor- rowska and Tro®mov (1980). Anterior to this bital process. At the frontal±parietal contact raised area, the frontal is fairly ¯at in Krypto- in the orbital margin, the frontal is indented baatar, PSS-MAE 101 and 113, whereas by the supraorbital notch (see below). some other Mongolian Late Cretaceous taxa In the orbit (®gs. 10±13, 18, 19, 33), the (e.g., Nemegtbaatar) have a small, rounded vertical component of the frontal forms the fossa, the orbitonasal fossa of Kielan-Jawo- dorsal third of the orbital mosaic, with its rowska (1971). This part of the frontal bone ventralmost projection extending to just is the posterodorsal end of a pocketlike struc- above the sphenopalatine foramen. The fron- ture (orbital pocket or theca orbitalis) occur- tal contacts the lacrimal and the maxilla an- ring in many multituberculates that repre- teroventrally, the orbitosphenoid posteroven- sents the origin for the pars anterior of the trally, the anterior lamina posteriorly, and the medial masseter muscle (Gambaryan and parietal posterodorsally. Within the orbit, at Kielan-Jaworowska, 1995). As noted by a level just in front of the supraorbital notch, these authors, the orbital pocket is roofed 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 23 dorsally and laterally by the frontal, lacrimal, 18, 19, 36A). It lies within the orbit on the and maxilla in Kryptobaatar. Because the or- suture between the frontal and orbitosphe- bital pocket is open ventrally, its roof is vis- noid, at the latter bone's anterodorsal margin. ible in ventral view (®gs. 14, 15). Leading to the ethmoidal foramen from be- Either in the frontal or between it and oth- low is a broad sulcus whose anterior border er bones are several nervous and vascular fo- is formed by a raised ridge along the frontal- ramina and grooves (®gs. 12, 18, 19, 36A). orbitosphenoid contact. At the posterodorsal limit of the orbital plate of the frontal, beneath the postorbital pro- MAXILLA cess, is a well-developed, anteriorly facing foramen (`òtc'' in ®gs. 12, 36A). This is the The maxilla is an enormous bone, extend- anterior opening of the orbitotemporal canal ing from the snout and palate deep into the of Rougier et al. (1992) for the ramus su- orbitotemporal fossa. In lateral view, its praorbitalis of the stapedial artery and ac- length is more than half that of the entire companying veins (postorbital foramen of skull (®g. 33). It bears six cheekteeth, four Kielan-Jaworowska et al., 1986). The bound- premolars (P1±4) and two molars (M1±2), aries of this aperture are the frontal ventro- and it has four distinct processes: facial, zy- medially, the anterior lamina ventrolaterally, gomatic, orbital, and palatal. and the parietal dorsally. In the right side of The facial process (®gs. 10±13, 33) oc- PSS-MAE 101 in which the postorbital pro- cupies the preorbital region behind the pre- cess is missing, the frontal can be seen to maxilla. It is incomplete in both PSS-MAE contribute to the ¯oor of the anteriormost 101 and 113, but its outer margin is well pre- part of the orbitotemporal canal. Running served in the former. The facial process is forward from the anterior opening of the or- tall and strongly convex laterally; dorsally, it bitotemporal canal is a shallow longitudinal bends medially such that its contact with the sulcus in the frontal. This sulcus turns dor- nasal and lacrimal tends toward the horizon- sally ventral to the supraorbital notch and tal (®g. 8). As described by Kielan-Jawo- then continues onto the notch. Where the sul- rowska (1970), the infraorbital canal opens cus turns dorsally, a small foramen opens an- to the rostrum between the embrasures of teromedially into the frontal (``fdv'' in ®gs. P1±P2 or slightly posterior to them (`ìof'' in 18, 36A); this opening may have transmitted ®g. 10). The single infraorbital foramen is the frontal diploic vein, as does a similarly dorsoventrally compressed, and its roof is placed foramen in the frontal of the dog visible in ventral view (®gs. 14, 15). Be- (Evans and Christensen, 1979). The frontal tween the posterior margin of the infraorbital diploic vein has been described in only a foramen and the root of the zygomatic arch handful of placentals and functions either as is a small, horizontal shelf, visible in ventral an emissary or diploic vein between the su- view, that extends laterally from P2, P3, and perior sagittal sinus or veins of the frontal the anterior half of P4 (®gs. 14, 15). This paranasal air sinus and the ophthalmic vein shelf merges laterally with the zygomatic (Thewissen, 1989). Given that the supraor- process of the maxilla and likely provided bital notch is subdivided by a faint ridge, it additional attachment area for the anterior transmitted at least two structures from the part of the super®cial masseter (Gambaryan orbit to the skull roof: one ran anteriorly in and Kielan-Jaworowska, 1995). a sulcus on the dorsal surface of the frontal, The zygomatic process is preserved on and the other ran posteriorly into a foramen both sides in PSS-MAE 101 (®gs. 10, 12, 22) between the parietal and frontal. The likely and on the left in PSS-MAE 113 (®gs. 13, occupants of the notch, sulcus, and foramen 23). It is a robust, posterolaterally trending are branches of the frontal artery, vein, and lamina of bone that contacts the squamosal nerve derived from the ramus supraorbitalis, posteriorly and the feeble jugal medially. It orbital veins, and ophthalmic nerve, respec- is the main constituent of the zygomatic arch tively. The ®nal aperture associated with the and, following the characterization of Kie- frontal is the ethmoidal foramen for the eth- lan-Jaworowska (1970), is described as con- moidal vessels and nerve (`èf'' in ®gs. 12, ¯uent with the snout, that is, there is no ¯are 24 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247

Fig. 10. Stereophotograph of the skull of Kryptobaatar dashzevegi PSS-MAE 101 in right lateral view. Gray pattern represents matrix; parallel lines denote breakage. Abbreviations: al anterior lamina; ali alisphenoid; fbu foramen buccinatorium; fr frontal; iof infraorbital foramen; lac lacrimal; mc masseteric crest; mx maxilla; naf nasal foramen; ocon occipital condyle; ompf orbital opening of minor palatine foramen; opf optic foramen; or orbitosphenoid; pmx premaxilla; son supraorbital notch; spf sphenopalatine foramen; sq squamosal; tg temporal groove. in the zygomatic arch (®gs. 8, 14). As noted squamosal (®gs. 10, 12, 33). On the lateral by Kielan-Jaworowska (1970), the posterior surface of the zygomatic process is a distinct, edge of the root of the zygomatic process elongate, arcuate depression extending from originates opposite the posterior half of P4. the squamosal suture forward to the root of The ventral margin of the zygomatic process the zygomatic arch and being bordered dor- is fairly straight and horizontal, and posteri- sally by a crest. This depression and crest, orly it stops short of the glenoid fossa; the the anterior zygomatic ridge (``azr'' in ®g. dorsal margin does not extend posteriorly as 14), provided attachment area for the anterior far as the ventral margin, producing a diag- part of the super®cial masseter muscle (Gam- onal suture between the maxilla and the baryan and Kielan-Jaworowska, 1995). The 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 25

Fig. 11. Stereophotograph of the skull of Kryptobaatar dashzevegi PSS-MAE 113 in right lateral view. anterior origin of this crest is on the ventral ®ssure (``sphf'' in ®g. 36A), which is the ap- face of the shelf extending between the den- erture that transmitted nerves and vessels tal arcade, the posterior edge of the infraor- from the cavum epiptericum into the orbit. bital foramen, and the root of the zygomatic The posteromedial extent of the maxilla can- process. The anterior origin is marked by a not be con®dently identi®ed, because it is blunt, rounded process and a conspicuous deep within the sphenorbital recess. That re- kidney-shaped depression behind that. On cess is completed by the frontal and orbito- the left zygomatic arch in PSS-MAE 101, the sphenoid dorsomedially and by the alisphe- jugal is lost, exposing an elongate, shallow noid posterolaterally. In the anteroventral facet on the medial surface of the maxilla for corner of the orbital process of the maxilla, this element (``juf'' in ®gs. 8, 22). immediately above the root of the zygomatic The orbital process is complete in PSS- arch, a circular opening into the infraorbital MAE 101 (®g. 18), but it is broken in PSS- canal is present (``oiof'' in ®g. 18). Prepa- MAE 113 (®g. 19). It is very extensive and ration of the infraorbital canal on the right represents the main constituent of the orbital side of PSS-MAE 113 exposed two openings mosaic. In the anterior part of the orbit (®gs. leading medially into the maxilla (either to 18, 19), the maxilla forms the wall and base the maxillary sinus or nasal cavity), which of the roof, which is completed by the frontal may represent entrances into the alveolar ca- and lacrimal; there is no ¯oor to the orbit, nals, which transmitted alveolar branches of which emphasizes the vertical nature of the the infraorbital nerves and vessels (Rougier maxilla here. In the posterior part of the orbit et al., 1997a). The presence of alveolar ca- (®gs. 10, 12), the maxilla is con®ned to the nals is con®rmed by the study of high-reso- inferior one-fourth of the wall and forms the lution CT scans of PSS-MAE 101. In the or- ¯oor to the sphenorbital recess, the space bital process at the level of the anterior part medial to the lateral rim of the sphenorbital of M2 is a large, circular sphenopalatine fo- 26 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247

Fig. 12. Stereophotograph of the skull of Kryptobaatar dashzevegi PSS-MAE 101 in left lateral view, with accompanying line drawing. Gray pattern represents matrix; parallel lines denote breakage. Abbreviations: aar atlas arch; al anterior lamina; ano nasal notch; ef ethmoidal foramen; fr frontal; izr intermediate zygomatic ridge; maf masseteric fossa; mafo masseteric fovea; mf mental foramen; mx maxilla; na nasal; ocon occipital condyle; opf optic foramen; or orbitosphenoid; otc orbitotemporal canal; pa parietal; pmx premaxilla; pop postorbital process (broken); ppr paroccipital process; ptc posttemporal canal; son supraorbital notch; spf sphenopalatine foramen; sq squamosal; sth stylohyal. ramen, the enclosing elements of which dif- foramen transmitted the sphenopalatine fer between the two specimens. In PSS-MAE nerve and vessels, which went to the nasal 113 the foramen is entirely within the max- cavity, and the major palatine nerve and ves- illa and the frontal approaches but does not sels, which went to the hard palate. The dor- contribute to the rim (``spf'' in ®g. 19). On sal aperture into the palatine canal, which the other hand, in PSS-MAE 101, the pos- transmitted the major palatine nerve and ves- terodorsal margin of the foramen is formed sels to the palate, cannot be seen. Posterior by the frontal (®g. 18). The sphenopalatine to the sphenopalatine foramen, a sulcus, the 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 27

Fig. 13. Stereophotograph of the skull of Kryptobaatar dashzevegi PSS-MAE 113 in right lateral view. sphenopalatine groove of Kielan-Jaworow- the level of the anterior half of M1, where it ska et al. (1986), extends posteriorly along meets the transverse suture with the pala- the maxillary±frontal suture toward the sphe- tines. The palatines have an essentially norbital ®ssure (®g. 19). Occupying this square-shaped exposure at the rear of the groove were the nerves and vessels destined hard palate. Only a small splinter of the max- for the sphenopalatine foramen. Near the illa is interposed between the palatine and the posterior limit of the orbital process of the medial alveolar margin. Despite distortion in maxilla is a foramen that runs anteroventrally PSS-MAE 113 (®g. 15), two medium-sized into the maxilla (``ompf'' in ®gs. 10, 36A); foramina (one on each side) are preserved we believe this is the dorsal opening into the that notch the transverse part of the maxil- minor palatine canal that transmitted the mi- lary±palatine suture (see ®g. 34); these are nor palatine nerves and vessels to the hard the major palatine foramina, which transmit- and soft palate. ted the major palatine nerves and vessels. The palatal processes, the largest elements Where the longitudinal part of the maxillary± of the hard palate (®g. 34), are fully acces- palatine suture meets the alisphenoid, lateral sible in PSS-MAE 113, where they are com- to the postpalatine torus, there is a slitlike plete but distorted (®g. 15); in PSS-MAE foramen, called here the minor palatine fo- 101, the lower jaws obscure all but the an- ramen (``mpf'' in ®g. 14), that leads poste- terior- and posteriormost parts of the palatal rodorsally into a canal, the minor palatine ca- processes (®g. 14). When restored to their nal. The alveolar portion of the maxilla pro- natural positions, the palatal processes are jects ventrally below the level of the palate moderately concave ventrally. As noted by and extends posteriorly beyond the posterior Kielan-Jaworowska and Dashzeveg (1978), margin of the hard palate to contact the ali- palatal vacuities are absent. The intermaxil- sphenoid. lary suture extends from the premaxilla to A small triangular portion of the orbital 28 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247

Fig. 14. Stereophotograph of the skull of Kryptobaatar dashzevegi PSS-MAE 101 in ventral view, with accompanying line drawing. Gray pattern represents matrix. Abbreviations: ali alisphenoid; ax axis; azr anterior zygomatic ridge; bo basioccipital; cp crista parotica; fhy fragment of hyoid arch; foi foramen ovale inferium; fv fenestra vestibuli; gl glenoid fossa; i3a alveolus for third upper incisor; inf incisive foramen; iof infraorbital foramen; jf jugular fossa; mpf minor palatine foramen; msy mandibular symphysis; muf muscular facet; mx maxilla; oiof orbital aperture of infraorbital canal; P3 third upper premolar; pal palatine; pat postpalatine torus; pef perilymphatic foramen; pmx premaxilla; ptc posttemporal canal; ptca pterygoid canal; rvnf recess for vascular and nervous foramina (prootic canal, ventral ascending canal, canal for ramus inferior, and facial canal); sth stylohyal; tpmx thickenings of premaxilla; ttf tensor tympani fossa; vo vomer. process of the maxilla is exposed in the lat- pterygoid) inside the choana. A contact with eral wall and roof of the choana. It is limited the vomer within the choana is not preserved anteroventrally by an oblique contact with and was probably absent. the palatine, medially by the pterygoid, and The horizontal portion of the palatine is posteroventrally by the alisphenoid. subrectangular (®g. 34). As seen in PSS- MAE 113 (®g. 15), the anterior limit of the PALATINE palatine, at the level of the P4/M1 embrasure, The palatine lacks any orbital exposure is the transverse suture with the maxilla, and therefore can only be seen in ventral which encloses the major palatine foramen. view, where it is fully exposed in PSS-MAE The lateral limit is the longitudinal suture 113 (®g. 15), but is partially covered by ma- with the maxilla, which encloses the minor trix left on the palate to support the lower palatine foramen, and the posterior limit is jaws in PSS-MAE 101 (®g. 14). The absence formed by the strong postpalatine torus. Par- of the orbital exposure has been con®rmed allel to the sides and close to the suture with in PSS-MAE 101 via the study of high-res- the maxilla are four to ®ve small foramina in olution CT sections through the skull. The the palatine in the position of accessory pal- palatine contacts the maxilla on the palate, atine foramina (®g. 34). No accessory pala- the alisphenoid at the ventrolateral edge of tine foramina are visible in the exposed area the choana, and the maxilla (and likely the of the left palatine in PSS-MAE 101 (®g. 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 29

Fig. 14. Continued.

14); that portion of the right palatine is cov- posteriorly by a wedge-shaped process that ered by matrix. In the midline, the palatine is covered laterally by the alisphenoid. This projects ventrally, forming a strong crest process marks the ventrolateral border of the along the indistinguishable interpalatine su- choana. ture. This crest is higher and thicker poste- The palatine forms the ¯oor and a small riorly and at the border of the choanae merg- part of the ventrolateral wall of the choana. es into the postpalatine torus (``pat'' in ®g. Behind the palatine contribution to the ven- 14). The torus is massive, thick anteropos- trolateral choanal wall is the maxilla. These teriorly, projects strongly ventrally, and ex- two bones are separated by an oblique suture tends laterally to the suture with the maxilla that runs anterodorsally inside the choana and the alisphenoid. As preserved, the torus above the minor palatine foramen. In the extends ventrally to the level of the occlusal midline between the choanae, the palatine surface of M2. The lateral part of the torus, has a strong crest that projects dorsally, par- at the level of the middle of M2, is continued tially subdividing the air passageway. 30 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247

Fig. 15. Stereophotograph of the skull of Kryptobaatar dashzevegi PSS-MAE 113 in ventral view.

PTERYGOID no contact across the midline, with the base of the vomer separating them. The ptery- The pterygoid lies on the skull base, ex- goids converge rostrally inside the choanae, tending from within the choana to the ante- but the full anterior extent cannot be ascer- rior pole of the ear region (®g. 34). It is pre- tained because the interior of the nasal cavity served in both PSS-MAE 101 (®g. 14) and cannot be accessed. 113 (®g. 15), but is most complete on the Each pterygoid bears a tall, longitudinal, right side of the former. The following description is based mostly on PSS-MAE 101, ventrally directed crest, the pterygopalatine where the pterygoid appears to be essentially ridge (``ptr'' in ®g. 37A) of Barghusen complete, although parts of its sutures are not (1986), that delimits a medial and lateral very clear. The medial suture with the pre- pterygopalatine trough (``mpt'' and ``lpt'' in sphenoid and basisphenoid is distinct on ®g. 37A). As there is also a midline crest PSS-MAE 113. ventral to the presphenoid, likely formed by The pterygoid is elongate and underlies the vomer, four troughs extending within the the presphenoid and basisphenoid, which are choanae into the rear of the nasal cavity are fused with the alisphenoid and orbitosphe- present on the mesocranium. The two lateral noid to form the sphenoid complex. The pter- troughs expand both anteriorly in the nasal ygoid contacts the alisphenoid laterally, the cavity and posteriorly on the mesocranium, basisphenoid posteriorly, and the presphe- and their lateral wall likely included a con- noid, vomer, and basisphenoid medially. It tribution from the alisphenoid. However, the may reach as far posteriorly as the anterior suture between the alisphenoid and pterygoid pole of the promontorium of the petrosal and cannot be established. The two medial epitympanic recess, but the sutures are un- troughs are deeper and become only slightly clear. What is visible of the pterygoid shows wider posteriorly. They are con¯uent behind 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 31 the vomer where they are roofed by the sphe- goids through a mostly longitudinal suture noid complex. The pterygopalatine ridge be- that diverges laterally toward the back of the comes taller posteriorly and ends at the level skull; therefore, the basisphenoid is wider of the anterior portion of the epitympanic re- than the presphenoid. As mentioned above in cess in a rounded, posteroventrally directed relation to the pterygoids, the midline crest process. A similar process was described for in the roof of the choanae is likely formed Kamptobaatar by Kielan-Jaworowska exclusively by the vomer. Its anterior termi- (1971). As noted by her, this process is rem- nus is obscured by the sediment still present iniscent of the pterygoid hamulus in modern inside the nasal cavity in both PSS-MAE 101 mammals. and 113. The midline crest becomes taller In PSS-MAE 113 (®g. 15), the pterygoids posteriorly, and at its terminus, approximate- are broken posteriorly on both sides, expos- ly at the level of the anterior margin of the ing a groove for the internal carotid artery in epitympanic recess, the ventral and posterior the petrosal. The pterygoid likely formed the edges of the crest become quite massive. ¯oor to an enclosed carotid canal (see Petro- This enlargement makes the posterior portion sal). of the crest a robust and prominent, ventrally projecting process; it does not, however, ex- SPHENOID COMPLEX tend as far ventrally as the pterygopalatine ridges. Behind the midline crest, the basi- The individual components of the sphe- sphenoid exposure is wedge-shaped and noid complexÐpresphenoid, basisphenoid, slightly convex ventrally. Extending posteri- orbitosphenoid, and alisphenoidÐare not de- orly from each of the pterygopalatine ridges limited by sutures in any specimens of to the suture between the basisphenoid and Kryptobaatar. Although these elements the basioccipital, there is a small, blunt, lon- clearly form a unity in the adult skull, they gitudinal crest. This crest marks the lateral are described here as separate entities that boundary of the basisphenoid, but the correspond with similar structures in living bone(s) forming the crest as well as those mammals for which the individual ossi®ca- immediately lateral to it cannot be con®- tion centers are known. As sutures between dently ascertained. Candidates include the the sphenoid and what we interpret as the pterygoid and the petrosal. In Kamptobaatar vomer are lacking or cannot be identi®ed, we (ZPAL MgM-I/33), this crest is formed by describe the latter bone together with the pre- the pterygoid. sphenoid and basisphenoid. In PSS-MAE 113 (®g. 15), the back part of the right pterygoid is missing, exposing PRESPHENOID/BASISPHENOID/VOMER the overlying basisphenoid. A large antero- We describe this portion of the sphenoid medially directed foramen on the thick dor- complex and vomer as the elements lying on solateral surface of the basisphenoid is found the midline in ventral view between the pos- at the end of a vascular groove traceable terior edge of the choanae and the basioccip- back to the anterior pole of the promonto- ital (®g. 34). Based on the morphology in rium. We interpret this foramen and groove living mammals, the presphenoid comprises as for the internal carotid artery. As stated the anterior part of this complex, the basi- above, the pterygoid likely formed the ¯oor sphenoid the posterior part, with the vomer to an enclosed carotid canal and obscured the lying ventral to both. These three elements carotid foramen in the basisphenoid. are most fully preserved in PSS-MAE 101 (®g. 14); however, additional details are ORBITOSPHENOID shown in PSS-MAE 113, in which the midline crest (vomer) and parts of the pterygoids The orbitosphenoid is the most complicat- are not preserved (®g. 15). ed portion of the sphenoid complex. It forms The presphenoid and basisphenoid form part of the posteromedial wall of the orbito- the midline ¯oor of the braincase in the me- temporal fossa and is continuous posteriorly socranium; their ventral surfaces are essen- with the ossi®ed primary wall of the brain- tially ¯at. Laterally, they contact the ptery- case, the pilae antotica and metoptica (see 32 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247

Fig. 16. Stereophotograph of the skull of Kryptobaatar dashzevegi PSS-MAE 101 in posterior view, with accompanying line drawing. Gray pattern represents matrix; parallel lines denote breakage. Ab- breviations: aar atlas arch; ax axis; con (mandibular) condyle; cor coronoid process; exoc exoccipital; lac lacrimal; lacf lacrimal foramen; mx maxilla; ocon occipital condyle; pc pterygoid crest; pet petrosal; ptc posttemporal canal; sq squamosal; sup supraoccipital.

Endocranium). Additionally, the orbitosphe- 113 (®gs. 11, 13, 19). However, the speci- noid wholly or partially encloses the exits for mens showing the endocranial surface of the the nerves and vessels that reached the orbit skull provide additional information that en- from the cranial cavity and constitutes the ables a fairly complete reconstruction of the posteroventral edge of the ethmoidal foramen orbitosphenoid's external morphology. (®gs. 33, 36A). The orbitosphenoid is a del- As is visible in lateral view (®gs. 10±13, icate element and is only partially preserved 33, 36A), the orbitosphenoid contacts the in both PSS-MAE 101 (®gs. 10, 12, 18) and frontal anteriorly and dorsally, the maxilla 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 33

Fig. 17. Stereophotograph of the skull of Kryptobaatar dashzevegi PSS-MAE 113 in posterior view. ventrally, and the anterior lamina posterolat- is notched on both sides. Upon comparison erally. The orbitosphenoid is a laminar ele- with the endocranial specimens, it is clear ment with two major components, an oblique that the notches on the left and right sides dorsal portion forming the ¯oor of the cranial constitute the anteroventral portions of the cavity and a subvertical ventral portion con- optic foramen (``opf'' in ®gs. 10, 12), which necting the braincase to the skull base. would have been completely enclosed within Where these two portions meet is a broad the orbitosphenoid (®gs. 33, 36A). On the groove leading into the ethmoidal foramen right side of PSS-MAE 101 dorsomedial to from behind and below. It transmitted nerves the sphenorbital ®ssure is a small, circular and vessels to the ethmoidal foramen and be- foramen anterolaterally directed. A similar comes deeper and broader dorsally as it ap- aperture seems to be present on the right side proaches the foramen. of PSS-MAE 113, but is not as clear. After The dorsal portion of the orbitosphenoid comparison with the specimens showing the has a slightly arched contact with the frontal endocranial surfaces, this aperture is inter- dorsally and extends from the ethmoidal fo- preted as the metoptic foramen for the ocu- ramen in front to the anterior lamina behind. lomotor nerve (``mef'' in ®g. 36A). Ventral to its contact with the frontal, the dorsal portion of the orbitosphenoid bulges anterolaterally, re¯ecting the outer contour of ALISPHENOID the brain. Posteriorly, at the level of the rostral edge of the hypophyseal (pituitary) fossa, The alisphenoid is a small, laminar com- the dorsal portion ¯ares laterally, contribut- ponent of the sphenoid complex with contri- ing at least partially to the dorsomedial edge butions to the orbitotemporal region and the of the sphenorbital ®ssure. This aperture is mesocranium. It is preserved on both sides completed laterally by the anterior lamina. of PSS-MAE 101 (®g. 14) and on the left The ventral portion of the orbitosphenoid side of 113 (®g. 15). extends ventrally from the ethmoidal fora- Despite its small size, the alisphenoid men in front and converges on the midline touches a number of bones on the skull base to contact the ventral portion of the orbito- (®gs. 14, 34, 37A) and in the orbitotemporal sphenoid of the opposite side. Together, in fossa (®g. 10). Its contacts are on the palate PSS-MAE 101, these elements form a mid- the maxilla and palatine anteriorly, on the line structure, a broad, short crest that ven- choana the pterygoid medially, on the me- trally contacts the maxilla as well as the pre- socranium the petrosal posteriorly and me- sphenoid and basisphenoid portions of the dially, and in the orbitotemporal fossa the an- sphenoid complex. Anterodorsally, this crest terior lamina posterodorsally, the remaining 34 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247

Fig. 18. Stereophotograph of the left orbitotemporal region of the skull of Kryptobaatar dashzevegi PSS-MAE 101 in oblique dorsolateral view, with accompanying line drawing. Gray pattern represents matrix; parallel lines denote breakage. Abbreviations: al anterior lamina; ali alisphenoid; fdv foramen for frontal diploic vein; fr frontal; frt foramen for ramus temporalis; lac lacrimal; man mandible; mx maxilla; na nasal; oiof orbital aperture of infraorbital canal; or orbitosphenoid; pa parietal; sgf supraglenoid foramen; spf sphenopalatine foramen; sq squamosal.

portions of the sphenoid complex medially, and the maxilla anteriorly. The alisphenoid has the shape of a portion of a MoÈbius strip, so that its medial margin in the lateral wall of the choana through a continuous line becomes the lateral edge of the epitympanic recess in the ear region. Based on this shape, the alisphenoid can be divided into anterior and posterior portions. The anterior portion is exposed only on its ventral surface; the posterior portion is exposed ventrally and dorsally. The anterior portion (®gs. 14, 34, 37A) contacts the palatine and maxilla in the wall of the choana. Lateral to this contact, there is a deeply recessed, semilunar area on the alisphenoid that probably housed the medial pterygoid muscle (Gambaryan and Kielan-Jaworow- ska, 1995: ®g. 8B). The anterolateral part of this attachment area for the medial pterygoid is completed by the maxilla. From the anteromedial edge of this recess, a slender process of alisphenoid extends forward approximately to the level of M2 and contributes to the minor palatine foramen (®g. 14). Medi- ally, the anterior portion of the alisphenoid meets the pterygoid in the lateral pterygo- Fig. 18. Continued. palatine trough, but as already stated, the su- 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 35

Fig. 19. Stereophotograph of the left orbitotemporal region of the skull of Kryptobaatar dashzevegi PSS-MAE 113 in oblique dorsolateral view, with accompanying line drawing. Gray pattern represents matrix; parallel lines denote breakage. Abbreviations: al anterior lamina; ef ethmoidal foramen; fr frontal; ju jugal; lac lacrimal; mx maxilla; pa parietal; pop postorbital process (broken); spf sphenopalatine foramen; sq squamosal.

ture between these two bones is unclear (®g. 37A). The posterior portion of the alisphenoid can be divided into two parts: one exposed ventrally and the other dorsally. The ventral exposure constitutes the anterior pole of the epitympanic recess (®g. 37A). It also forms the anterior part of the tall crest demarcating the lateral margin of the epitympanic recess; the bulk of this crest is formed by the anterior lamina. Extending medially from the anterior pole of the epitympanic recess is a tongue of alisphenoid that forms the anterolateral edge of the posterior aperture into the carotid canal (see Petrosal). The dorsal exposure of the posterior portion of the alisphenoid lies behind the maxilla and in front of the anterior lamina (®g. 10). This part of the alisphenoid forms the ventrolateral edge of the sphenorbital recess (®gs. 10, 36A), the lateral limit of which is demarcated by a crest that continues anteroventrally in a ridge along the alisphenoid±maxillary suture. In- ferior to this ridge is a concave surface on the dorsal exposure of the alisphenoid that is continuous with the surface described in ven- Fig. 19. Continued. tral view as for the medial pterygoid muscle. 36 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247

This concave surface likely represented an eas of the skull is found only in monotremes. additional attachment area for pterygoid The monotreme anterior lamina forms as the musculature, including the lateral pterygoid lamina obturans, an intramembranous ossi®- (Gambaryan and Kielan-Jaworowska, 1995: cation in the sphenoobturator membrane, that ®g. 8B), which we term the pterygoid fossa fuses with the endochondral petrosal proper of the alisphenoid. in subsequent ontogeny (Kuhn, 1971; Pres- Close to or within the suture between the ley, 1981; Zeller, 1989). Given that the dorsal exposure of the alisphenoid and the monotreme lamina obturans is the only mod- anterior lamina is a small, anteriorly directed el for the anterior lamina in extinct taxa, it foramen (``fbu'' in ®gs. 10, 36A). We think seems likely that the element in multituber- that this foramen communicated with the ca- culates and other mammaliaforms also forms vum epiptericum, based on an as yet undes- intramembranously. cribed skull of a new species of Mongolian We describe the petrosal of Kryptobaatar Late Cretaceous multituberculate (PSS-MAE in three viewsÐventral, lateral, and occipi- 126). Among living mammals, comparably talÐreserving the endocranial surface of the situated foramina (i.e., within the attachment petrosal for a description of that space as a area of the pterygoid musculature and con- single unit (see Endocranium below). In necting to the cavum epiptericum) are re- PSS-MAE 101, both petrosals are preserved ported in some rodents and transmit branches in situ, but the left is partially hidden in ven- of the mandibular nerve (Hill, 1935; Wahlert, tral view by the stylohyal and matrix left to 1974). In fact, four apertures in or along the support the stylohyal (®gs. 14, 20). In PSS- alisphenoid accommodating branches of the MAE 113, the right petrosal is preserved in mandibular nerve are known to occur in ro- life position, but the left is displaced poste- dents: the foramen ovale, the foramen ovale roventromedially and the anterior end of its accessorius, the masticatory foramen, and the epitympanic recess is missing (®gs. 15, 21). buccinator foramen (Wahlert, 1974). Of the Ventral View (®gs. 14, 20, 21, 34, 37A): possible occupants of the anteriorly directed In this view, the petrosal contacts medially opening in the pterygoid fossa of Kryptobaa- the basisphenoid and basioccipital, postero- tar, we think the most plausible was the buc- medially the exoccipital, laterally the squa- cal nerve, which is the most medial branch mosal, and anteriorly the alisphenoid and of the anterior division of the mandibular pterygoid. The most conspicuous feature of nerve and runs forward dorsal to the lateral the petrosal is the promontorium (``pr'' in ®g. pterygoid muscle in the dog (Evans and 37A), the tympanic surface of the cochlear Christensen, 1979). Consequently, we iden- housing, which is elongated, ventrally ¯at- tify this opening as the foramen buccinato- tened, and anteromedially directed. In both rium (®gs. 10, 36A). PSS-MAE 101 and 113, the ventral surface of the promontorium is distinctly marked by PETROSAL grooves, presumably for the internal carotid and stapedial arteries, and it bears ¯anges The petrosal is the most complex element along its medial aspect. However, these sur- of the basicranium; it houses the organs of face features do not mask the essential ®n- hearing and balance, and contributes to the gerlike contour of the promontorium, which braincase wall on the lateral surface, ¯oor, in turn likely re¯ects the shape of the en- and occiput. In living therians, the petrosal is closed cochlear duct, known to be rodlike in generally conceived as comprising two dif- other Late Cretaceous and Paleocene multi- ferent regions: the pars cochlearis (housing tuberculates (Miao, 1988; Luo and Ketten, the cochlea) and the pars canalicularis (hous- 1991; Fox and Meng, 1997; Hurum, 1998b). ing the vestibule and semicircular canals). In The grooves on the promontorium (®g. multituberculates, in addition to these two re- 37A) exhibit a Y-shaped pattern. The short gions, an anterior lamina forms part of the stem of the Y, the medial groove, is oriented lateral wall of the braincase and an extensive in a near transverse plane at a level posterior epitympanic recess. Among living mammals, to the basisphenoid±basioccipital suture; it is an anterior lamina contributing to similar ar- interpreted as having housed the internal ca- 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 37

Fig. 20. Stereophotograph of the basicranium of Kryptobaatar dashzevegi PSS-MAE 101. rotid artery. The anteriorly trending arm was ramen is formed jointly by the alisphenoid for the rostral continuation of the internal ca- and petrosal and is called here the posterior rotid (``gica'' in ®g. 37A), and the postero- aperture of the carotid canal (``pacc'' in ®g. laterally directed arm was for the internal ca- 37A). The lateral position of this aperture de- rotid's main extracranial branch, the stapedial termined a long course for the artery to the artery (``gpsa'' in ®g. 37A). Of these three carotid foramen in the hypophyseal fossa, grooves, the widest is the main stem of the which is known from the specimens showing internal carotid, and the longest is the one the endocranial surface. The exact course of for the stapedial artery, which is subequal in the internal carotid artery from the posterior diameter to the one for the rostral continua- aperture of the carotid canal to the endocra- tion of the internal carotid. This Y-shaped nium is not entirely clear, as it was hidden pattern is very rostrally positioned on the within bone, but the following reconstruction promontorium, which is unusual among liv- provides the best ®t with the evidence avail- ing mammals (Wible, 1987) but is found in able. On the right side of PSS-MAE 113 (®g. some other multituberculates (e.g., Kampto- 21), the back part of the pterygoid is missing baatar, ZPAL MgM-I/33). and the rostral continuation of the carotid From the point of the origin of the stape- groove can be followed as it curves medially dial artery, the internal carotid groove ex- on the anterior pole of the promontorium. tends anterolaterally close to the medial bor- From the promontorium, the carotid groove der of the epitympanic recess. On the right runs anteromedially to a foramen in the dor- side of PSS-MAE 101 (®g. 20), the carotid solateral part of the basisphenoid that pre- groove leads anteriorly into a foramen situ- sumably represents the ventral aperture of ated between the lateral aspect of the anterior the carotid foramen in the hypophyseal fossa. pole of the promontorium and the medial With the pterygoid in place, as on the right margin of the epitympanic recess. This fo- side of PSS-MAE 101 (®g. 20), the carotid 38 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247

Fig. 21. Stereophotograph of the basicranium of Kryptobaatar dashzevegi PSS-MAE 113. groove on the anterior pole of the promon- (``ptca'' in ®g. 37A). Its usual occupant in torium was presumably enclosed in a canal, modern mammals is the nerve of the ptery- which left no visible trace on the ventral ba- goid canal, which is formed by the greater sicranial surface. The carotid canal appears and deep petrosal nerves, carrying parasym- to be have been formed by the petrosal, ali- pathetic and sympathetic ®bers, respectively sphenoid, and pterygoid. The length of the (Evans and Christensen, 1979; Williams et carotid canal is a consequence of the lateral al., 1989). However, in some instances, there position of its posterior aperture and the very is also an accompanying artery off the inter- dorsal placement of the ventral aperture of nal carotid (McDowell, 1958; MacPhee, the carotid foramen, which in turn implies a 1981). The relatively large size of the pre- very thick basisphenoid from the ¯oor of the served groove on the alisphenoid in PSS- hypophyseal fossa to the ventral surface of MAE 113 suggests that the pterygoid canal the skull base. contained both an artery and nerve. The en- The absence of the pterygoid on the right trance of the deep petrosal nerve, a branch side of PSS-MAE 113 exposes another of the internal carotid nerve, into the ptery- groove in the alisphenoid that notches the goid canal must have been from the carotid dorsal surface of the ventral aperture of the groove, but the entrance of the greater petro- carotid canal and leads anterolaterally toward sal nerve is unclear. Given that there is no the orbit (®g. 21). It is uncertain whether this likely tympanic aperture for the greater pe- second groove opened in the posterior aper- trosal nerve, we speculate that it must have ture of the carotid canal or in a separate fo- entered the pterygoid canal through the su- ramen anteromedial to it. The exact position ture (gap) between the petrosal and alisphe- of the orbital opening of this groove is un- noid directly from the cavum epiptericum certain, but was likely in the ¯oor ventral to and/or cavum supracochleare. the foramen buccinatorium. This anterolat- From its origin, the groove for the stape- erally directed groove was likely enclosed in dial artery runs posterolaterally on the pro- a canal by the pterygoid that is subequal to montorium toward the fenestra vestibuli or that for the internal carotid artery. In fact, on oval window (``fv'' in ®gs. 14, 37A). The the right side of PSS-MAE 101, in which the stapedial groove exhibits a slightly different pterygoid is in place, there is a ridge running relationship to the fenestra vestibuli in PSS- forward from the posterior aperture of the MAE 101 and 113. In the former, the groove carotid canal that likely marks the position notches the rim of the fenestra vestibuli, of the enclosed canal (®g. 20). We interpret whereas in the latter it extends slightly pos- this canal as the pterygoid (Vidian) canal terior to the fenestra, with the crest forming 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 39 the ventral edge of the stapedial groove ex- very conspicuous jugular fossa (``jf'' in ®gs. tending posteroventral to the ventral margin 14, 37A), the large depression around the of the oval window. jugular foramen (``jfo'' in ®g. 37A). A sulcus On the medial aspect of the ventral pro- for the cochlear aqueduct, which is present montorial surface, near the contact with the on the petrosal in some other multitubercu- basisphenoid, is a fan-shaped ¯ange trending lates (Rougier et al., 1996c; Fox and Meng, anteroposteriorly. Running transversely on 1997), is lacking. We term this aperture in the surface of the ¯ange is the groove for the Kryptobaatar a perilymphatic foramen, be- main stem of the internal carotid described cause observations on the endocranium of above. Extending rostrally from the sulcus on PSS-MAE 123 have failed to reveal a coch- the ¯ange is a low crest, which we call here lear aqueduct (cochlear canaliculus), a bony the rostral tympanic process of the petrosal. canal that transported the perilymphatic duct. It may mark the contact with either a mem- Consequently, the only possible channel for brane, cartilage, or bone contributing to the the perilymphatic duct from the inner ear to ¯oor of the tympanic cavity, as it does in the jugular foramen was via the perilym- extant mammals (Novacek, 1977; MacPhee, phatic foramen. 1981). However, the composition of this Immediately behind the perilymphatic fo- ¯oor, whether membranous, cartilaginous, or ramen and forming part of its posterodorsal bony, cannot be determined. To date, there is edge is the prominent ventral bulge housing no evidence for a bony bulla in any non- the posterior ampulla of the semicircular ca- therian mammal. On the right side of PSS- nals. Portions of the lateral and posterior MAE 101, the rostral tympanic process of semicircular canals can be traced posteriorly the petrosal is continuous with a low crest from the posterior ampulla. These, along that extends rostrally to the base of the ptery- with the crista interfenestralis, delimit two gopalatine ridge (®g. 20). At least the pos- deep pits on either side of the posterior am- terior part of this crest is petrosal. pulla. The larger pit is the one lateral to the At the posterior end of the promontorium ampulla, which is encircled by the posterior (®g. 37A) are two apertures, the fenestra ves- extension of the crista interfenestralis and the tibuli and the perilymphatic foramen, sepa- posterior extension of the lateral semicircular rated by a narrow bridge of bone, the crista canal. The small medial pit is subtriangular interfenestralis, that reaches posteriorly to and is placed between the posterior exten- the base of the paroccipital process. The peri- sions of the lateral and posterior semicircular lymphatic foramen (``pef'' in ®gs. 14, 37A) canals. These pits likely housed the expanded is roughly circular, and the subequal fenestra middle-ear cavity. vestibuli has an average stapedial ratio (see Positioned medial and very near to the Segall, 1970) of 1.39 in PSS-MAE 113. The perilymphatic foramen is the jugular foramen fenestra vestibuli is oriented in a near vertical (®g. 37A), which is about half the size of the plane and faces anterolaterally with a slight former aperture. The jugular foramen lies on ventral component. Immediately anterior to the suture between the petrosal and exocci- the fenestra vestibuli is a deeply excavated pital; the basioccipital appears to be exclud- pocket (``ttf'' in ®gs. 14, 37A), the fossa for ed, judging by a suture visible on the right the tensor tympani muscle (fossa muscularis side of PSS-MAE 101 (see Exoccipital). major of Kielan-Jaworowska et al., 1986). Also visible on the same side of this speci- The lateral margin of the tensor tympani fos- men is a smaller foramen just posterolateral sa is formed by the ¯aring medial edge of to the jugular foramen completely encircled the epitympanic recess. by the petrosal; the function of this opening The perilymphatic foramen and environs is unknown and its presence in PSS-MAE are best shown on the right side of PSS-MAE 113 cannot be veri®ed. 101 but are preserved on both sides in PSS- Directly in front of the anteromedial edge MAE 113. The foramen is posteriorly di- of the perilymphatic foramen is a niche on rected and only its ventral margin is well de- the promontorium, which is deeply recessed, limited; its roof lacks a de®nitive edge and more so in PSS-MAE 113 than in 101. This is formed by the petrosal's contribution to the niche represents the anteromedial extent of 40 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247 the greatly expanded jugular fossa, the large here that the external acoustic meatus likely depression around the jugular foramen (®gs. entered the middle ear. The anterior part of 14, 37A). A robust, long laminar process the longitudinal component, beginning at the projects posteromedially from the medial level of the posterior margin of the fenestra side of the promontorium and forms a hori- vestibuli, is more prominent. It runs antero- zontal shelf that partially ¯oors this part of medially as a tall subvertical crest that limits the jugular fossa. This shelf is well devel- the medial margin of the epitympanic recess oped in PSS-MAE 113 (®g. 21), but barely and continues forward as the medial edge of has any horizontal contribution in PSS-MAE the infolded lateral ¯ange. 101 (®g. 20). Slightly posterolateral to the level of the As stated above, the crista interfenestralis fenestra vestibuli, the crista parotica shows reaches posteriorly to contact the paroccipital the attachment of a well-developed tympa- process (``ppr'' in ®g. 37A). Lateral to the nohyal (``th'' in ®g. 37A). In ventral view, crista interfenestralis and in front of the par- this pronglike element is posteriorly and only occipital process (and hidden by that struc- slightly medially directed; in medial view, it ture in the ®gures) is a broad depression in has a distinctly triangular outline. The the tympanic roof, the fossa for the stapedius tympanohyal appears completely preserved muscle. The limits of this depression are not on the left side of PSS-MAE 113 (®gs. 21) conspicuous, but its ventral extension on the and has a hooklike pro®le for cradling the lateral face of the crista interfenestralis is hyomandibular branch of the facial nerve. very distinct. Nevertheless, this depression is This nerve left the middle ear via a stylo- at least twice the surface area of the fenestra mastoid notch immediately posterior to the vestibuli. The well-developed, triangular par- tympanohyal. occipital process (®g. 37A) is preserved on The epitympanic recess (``er'' in ®g. 37A), both sides of PSS-MAE 101 (®g. 20) and on following Kielan-Jaworowska et al. (1986), the right side of 113 (®g. 21). It is not a is the fossa above the dorsal margin of the vertical structure, but is slanted somewhat tympanic membrane that accommodated the anteroventrally. From the blunt apex of the body of the malleus and incus and housed paroccipital process two crests arise, one the crus breve of the incus (fossa incudis). In leading medially and the other laterally (®g. Kryptobaatar, the epitympanic recess is a 37A). The medial one, the caudal tympanic process of the petrosal (``ctpp'' in ®g. 37A), deeply excavated, elongated, ellipsoidal fos- is very short and marks the posteroventral sa trending roughly parallel to the crista pa- limit of the conspicuous jugular fossa. The rotica (®gs. 20, 21, 37A). It has contributions three bones encircling the jugular fossa (pe- from three bones: the petrosal forms the bulk trosal, exoccipital, and basioccipital) provide with the alisphenoid at the anteriormost end. laminae that wall and partially ¯oor this Additionally, the squamosal forms the pos- space. As one of these laminae, the caudal terolateral edge of the epitympanic recess tympanic process of the petrosal is coplanar where the fossa incudis is located. The me- with the laminar projections from the other dial wall of the fossa incudis is delimited by bones. the prominent crista parotica and its rostral The lateral crest arising from the parocci- continuation, the lateral ¯ange. This sharp, pital process is the crista parotica (``cp'' in ventrally projecting crest forms a barrier be- ®g. 37A). It is L-shaped with transverse and tween the fossa incudis and the fenestra ves- longitudinal components. The shorter trans- tibuli, and it severely constrained the likely verse component is anteroposteriorly thick positions of the stapes and incus. As in other and smooth and runs from the tip of the par- multituberculates (Rougier et al., 1996a, occipital process to a scooped area housing 1996c), the lateral ¯ange in Kryptobaatar is the external acoustic meatus on the squa- infolded such that its anterior end contacts mosal-petrosal suture. The posterior part of the cochlear housing; in other Mesozoic taxa, the longitudinal component is not well indi- such as Morganucodon and Vincelestes, the vidualized, and sutures indicate that its lat- lateral ¯ange runs parallel to and is separated eral surface contacted the squamosal; it was from the promontorium by a depression 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 41 called the lateral trough (Wible and Hopson, on comparison with isolated petrosals of oth- 1993). er Late Cretaceous multituberculates (see In the anterior part of the epitympanic re- Kielan-Jaworowska et al., 1986; Luo, 1989; cess are several foramina interpreted for Wible and Hopson, 1995). Of the three fo- branches of the mandibular nerve (``foi'' and ramina in the recess in PSS-MAE 113, the ``fma'' in ®g. 37A); Kryptobaatar, as in all posterior two share a common space and are known multituberculates and some rodents higher than the anterior one. The posterior- (Hill, 1935; Wahlert, 1974), has multiple ap- most foramen opens into a posterolaterally ertures for this division of the trigeminal directed channel and likely transmitted one nerve, one of which, the buccinator foramen, of the two end branches of the stapedial ar- was described already (see Alisphenoid). In tery, the ramus superior, into the ventral as- PSS-MAE 101 (®g. 20), two foramina for the cending canal (the pterygoparoccipital fora- mandibular nerve are situated in the epitym- men). The similarly sized middle foramen is panic recess: the foramen ovale inferium posteroventrally directed and likely transmit- opens into the recess and faces anteroven- ted the prootic sinus; therefore, it represents trally, whereas the foramen masticatorium the ventral aperture of the prootic canal. The notches the lateral rim of the recess and faces dorsal aperture of the prootic canal will be ventrolaterally. The foramen ovale inferium described with the endocranial surfaces. The is centrally placed in a deep fossa and dom- morphology of these two foramina, the ven- inates the anterior pole of the epitympanic tral apertures of the ventral ascending and recess; it is traversed by a broad sulcus, prootic canals, accords well with that report- which anteriorly falls just short of the ali- ed in some other multituberculates (Wible sphenoid's contribution to the recess. The fo- and Hopson, 1995: ®g. 7A). The anterior- ramen masticatorium is strongly elliptical most foramen in the recess in PSS-MAE 113, with the anteroposterior axis longer than the the largest, is anteromedially directed, hori- dorsoventral and is placed lateral and slightly zontal, and elliptical in outline. We consider posterior to the foramen ovale inferium. A it to be a joint aperture leading to two canals, similar arrangement is present in PSS-MAE one for the hyomandibular branch of the fa- 113 (®g. 21) with the exception that the fo- cial nerve (the secondary facial foramen) and ramen masticatorium is subdivided into two the second for the other end branch of the foramina on the right side. The division is stapedial artery, the ramus inferior, an ar- accomplished by a narrow bar of bone con- rangement different from that in described tinuous with the lateral edge of the epitym- ptilodontoid and taeniolabidoid multituber- panic recess. This delicate bar was complete, culates (Luo, 1989; Wible and Hopson, but was broken prior to illustration here and 1995). Supporting this interpretation is a dis- in Rougier et al. (1996c: ®g. 3). The left side tinct sulcus leading posteriorly from this ap- of PSS-MAE 113 is damaged, but only one erture to the dorsal rim of the fenestra ves- foramen masticatorium seems to have been tibuli that resembles the sulcus for the facial present. Variation in the number of exits for nerve in other multituberculates (Wible and the mandibular nerve on different sides of Hopson, 1995: ®gs. 7A, 8A). Moreover, a the same skull is known in other multituber- second, well-developed foramen at the ante- culates (e.g., Kamptobaatar, ZPAL MgM-I/ rior edge of the petrosal, lateral to the prom- 33). ontorium, likely held the rostral continuation Medial and dorsal to the crista parotica, of a vessel that initially ran with the facial and anterior to the tympanohyal, is a re- nerve. That this vessel was the ramus inferior cessed area (``rvnf'' in ®gs. 14, 37A). On the is supported by the large size of the groove left side of PSS-MAE 113, three foramina for the stapedial artery on the promontorium are visible in this recess. These foramina lead and the small size of the foramen for the ra- into canals presumably within the petrosal, mus superior, implying that the other primary the exact course of which could not be de- ramus of the stapedial artery was present. termined without damaging the available However, we cannot rule out that a vein, the specimens. Our reconstruction of the occu- post-trigeminal vein, accompanied the ramus pants of these foramina and canals is based inferior in its passage through the petrosal. 42 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247

The positions of the foramina transmitting the deeply pitted articular surface of the pa- the superior and inferior rami and the sta- rietal. pedial groove on the promontorium in both Slightly anteroventral to the anterior open- PSS-MAE 101 and 113 suggest that the sta- ing of the orbitotemporal canal is the suture pedial artery ran across the fenestra vestibuli between the anterior lamina and frontal (®g. and through the presumed bicrurate stapes 12). Along this arched junction, the anterior (Rougier et al., 1996c). lamina lies lateral to the frontal. Farther ven- Lateral View (®gs. 10±13, 33, 36A): The trally, the anterior lamina meets the orbito- most conspicuous feature of the petrosal in sphenoid at a suture that is concave anteri- lateral view is the anterior lamina. It contacts orly (only the dorsal portion of which is pre- dorsally the parietal, anteriorly the frontal, served in PSS-MAE 101). As is apparent orbitosphenoid, and alisphenoid, and poste- from one of the specimens showing the en- riorly it is overlapped by the squamosal; ven- docranial surface (®g. 25; PSS-MAE 123), trally, it is continuous with the tympanic sur- the anterior lamina overlaps the orbitosphe- face of the petrosal through the lateral ¯ange. noid laterally. In the area of the sphenorbital The anterior lamina formed wholly or par- ®ssure, the anterior lamina is sharply in¯ect- tially a number of passageways for nerves ed medially so that a portion of it faces an- and vessels leaving the braincase for the or- teriorly (®gs. 10, 12). This surface has a deep bitotemporal fossa and also provided a major fossa that is continuous ventrally with the area of attachment for several muscles of surface for muscle attachment in the alisphe- mastication. noid, the pterygoid fossa. The likely occu- Dorsally, the anterior lamina overlaps the pant of this area was the lateral pterygoid parietal as shown on the left side of PSS- muscle. The suture between the alisphenoid MAE 101. From back to front, the suture be- and anterior lamina runs obliquely across this tween the anterior lamina and parietal runs muscular fossa from the anteroventral margin of the sphenorbital ®ssure to the margin of slightly dorsally from the triple junction of the foramen masticatorium. Midway along parietal, squamosal, and anterior lamina to its length, the suture is interrupted by an an- the anterior opening of the orbitotemporal teriorly facing foramen described above as canal between the parietal, frontal, and an- having transmitted the buccal branch of the terior lamina. Parallel and medial to this su- mandibular nerve (®gs. 10, 36A). ture is a preserved endocast on both sides of The ventral edge of the anterior lamina PSS-MAE 101 (®gs. 10, 12) and the right (the lateral edge of the lateral ¯ange) forms side of 113 (®g. 11), representing the ®lling a gentle arch from the contact with the ali- of the orbitotemporal canal. The orbitotem- sphenoid in front to the root of the zygomatic poral canal appears to have been bounded arch behind. Anteriorly, this margin is laterally by the parietal only, with the ante- notched by the foramen masticatorium. rior lamina lateral to that. Throughout most Slightly dorsal and posterior to the edge of of its course, there is no distinct medial wall the foramen masticatorium is another small, for the orbitotemporal canal, suggesting that anteriorly directed foramen, the supraglenoid the orbitotemporal vessels ran endocranially foramen for a ramus temporalis of the ramus within a sulcus on the medial surface of the superior (``sgf'' in ®gs. 8, 18, 36A). This fo- parietal. Toward the front, however, the fron- ramen was likely continuous with the ventral tal provides the medial wall of a true orbi- ascending canal. totemporal canal, and the anterior lamina, pa- Posteriorly, the anterior lamina's contact rietal, and frontal complete the anterior open- with the squamosal can be divided into two ing of this canal. Directly lateral to the an- portions: a ventral one that is essentially hor- terior opening of the orbitotemporal canal is izontal, and a dorsal, essentially vertical one. a thickened lateral process of the anterior Forming the horizontal portion is a ¯at ¯ange lamina that ventrally supports the large post- of the anterior lamina that buttresses the front orbital process of the parietal. This process of the root of the zygoma and extends lat- on the anterior lamina is incomplete on the erally toward, but falls short of, the glenoid left side of PSS-MAE 101, exposing dorsally fossa (``gl'' in ®gs. 6, 14, 37A). The hori- 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 43 zontal portion of the suture is gently arched, when the skull is maximally extended (dor- concave posteriorly, and meets the dorsal si¯exed) on the neck. The dorsolateral de- portion in a small foramen, also for a ramus pression is shallow, roughly circular, and is temporalis (``frt'' in ®gs. 8, 36A). The dorsal separated from the atlantal fossa by a vertical portion runs along the course of the dorsal ridge. Opening into the dorsolateral depres- ascending canal, which, because of breakage sion is the posterior opening of the posttem- in PSS-MAE 101, is shown to be formed by poral canal, which is wholly in the petrosal the squamosal and petrosal. On the left side (``ptc'' in ®gs. 14, 16, 36A). An elongated, of this specimen (®g. 12) there is a notch at ventral depression runs parallel to the crista the posterodorsal corner of the anterior lam- parotica and probably represented the site of ina that opens to the dorsal ascending canal attachment for the sternomastoid muscle and probably was enclosed in a foramen by (Evans and Christensen, 1979). the squamosal posteriorly (``fdac'' in ®gs. 8, 36A). This dorsal ascending canal foramen JUGAL also transmitted a ramus temporalis. Similar The jugal is a thin, essentially oval, lami- foramina are found, for example, in Kamp- nar bone on the inner surface of the zygo- tobaatar (ZPAL MgM-I/33) and Lambdop- matic arch. In PSS-MAE 101, both zygo- salis (Miao, 1988). matic arches are preserved, but a nearly com- The external surface of the anterior lamina plete jugal is present only on the right side has a fairly complex topology and, in addi- (®g. 22). In PSS-MAE 113, only the left arch tion to the lateral pterygoid muscle, provided is complete and the jugal preserved (®g. 23). attachment for the temporalis muscle. The On the right side of PSS-MAE 101 (®g. temporalis attachment can be divided into 22A), the jugal is not as tall as the zygoma; two major parts: an anterodorsal one that is it is recessed from both the ventral and dorsal convex and best expressed under the post- margins of the zygomatic arch, but is closer orbital process, and a ventral concave one. to the latter. The anterior extent of the jugal These two parts correspond to the attachment is at the level of the posterior half of M1; areas identi®ed as for the pars anterior and posteriorly, it reaches almost to the same lev- pars posterior of the temporalis, respectively, el as does the ventral margin of the zygo- in Nemegtbaatar by Gambaryan and Kielan- matic process of the maxilla. The zygomatic Jaworowska (1995). arch, and so the jugal, does not lie in a sag- Occipital View (®gs. 16, 17, 35, 37A): ittal plane, but is tilted such that its dorsal Well preserved in both PSS-MAE 101 and margin is slightly lateral to the ventral. On 113, the mastoid exposure of the petrosal the left side of PSS-MAE 101 (®g. 22B), the forms the lateral portion of the occiput and jugal is not preserved, and a shallow, oval comprises approximately half of the occi- facet is exposed. The bulk of the facet is on put's bony surface. The mastoid exposure is the maxilla, with only the posterodorsal ®fth subtriangular with the base medially and on the squamosal. In PSS-MAE 113 (®g. 23), with the apex laterally. It contacts the squa- the jugal conforms in most features to that in mosal laterally on the ventral portion of the PSS-MAE 101, but the dorsal margin of the nuchal crest, the parietal dorsomedially on jugal is exposed in lateral view, forming the the dorsal portion of the nuchal crest, and the dorsal edge of the zygomatic arch (®g. 19). supraoccipital and exoccipital medially. The Moreover, the dorsal margin is slightly thick- ventral edge of the mastoid exposure is er than the portion of the jugal directly me- formed by the paroccipital process and the dial to the maxilla and squamosal. crests arising from it, namely the caudal tympanic process of the petrosal medially and SQUAMOSAL the crista parotica laterally. There are three depressions on the mastoid The squamosal is appressed to the poste- exposure. The dorsomedial one continues rior part of the side wall of the braincase and onto the exoccipital and is subvertical with can arbitrarily be divided into two parts: the its deepest point along the suture between the zygomatic process and the dorsal ¯ange. The petrosal and exoccipital; it lodges the atlas squamosal is well preserved in both PSS- 44 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247

Fig. 22. Pencil drawings of the right and left zygomatic arches (A and B)ofKryptobaatar dashzevegi PSS-MAE 101 in dorsal view, with accompanying line drawings. Gray pattern represents matrix; parallel lines denote breakage. Abbreviations: al anterior lamina; fr frontal; ju jugal; juf jugal facet; lac lacrimal; lacf lacrimal foramen; mx maxilla; pa parietal; son supraorbital notch; sq squamosal. 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 45

Fig. 23. Pencil drawing of the left zygomatic arch of Kryptobaatar dashzevegi PSS-MAE 113 in dorsal view, with accompanying line drawing. Abbreviations: fr frontal; ju jugal; lac lacrimal; mx maxilla; pa parietal; sq squamosal; tr temporal ridge.

MAE 101 and 113, although in the latter the 16, 17, 35, 37A), the dorsal ¯ange contacts zygomatic process is incomplete on the right the mastoid exposure of the petrosal to form side. the inferolateral extent of the nuchal crest, The squamosal's contacts are as follows just lateral to the posterior opening into the (®gs. 8, 12, 32±34): anteriorly, with the max- posttemporal canal. In dorsal view (®gs. 8, illa and the jugal via the zygomatic process, 9, 31), the chief contact is with the petrosal, and with the anterior lamina via the dorsal in front with the anterior lamina and behind ¯ange; medially and posteriorly, with the pe- with the edge of the mastoid exposure form- trosal via the dorsal ¯ange; and dorsally, with ing the nuchal crest. The dorsal ¯ange also the parietal via the dorsal ¯ange. The squa- has a narrow contact with the parietal dor- mosal has been partially lost in several spec- somedially. The suture with the anterior lam- imens (PSS-MAE 113, 124, 125), exposing ina can be divided into two arched portions: an extensive underlying facet on the petrosal. the anterior one is chie¯y horizontal and is From this, it is evident that the squamosal continuous with the front edge of the zygo- has no direct contribution to the side wall of matic process; the posterior one is vertically the braincase. directed (®g. 8). Where these two portions of The dorsal ¯ange of the squamosal is lam- the squamosal±petrosal suture meet is a small inar, tongue-shaped, and appressed to the foramen, probably for a ramus temporalis side wall of the braincase. In ventral view (``frt'' in ®gs. 8, 36A). The posterior edge of (®gs. 20, 21, 34, 37A), the dorsal ¯ange con- the dorsal ¯ange, together with the dorsal tacts the petrosal through an L-shaped suture, edge of the mastoid exposure of the petrosal, with the long arm oriented sagittally and the forms the inferolateral portion of the nuchal more anteriorly located short arm transverse- crest, which is low and sharp. The ventrally. The short arm abuts a laterally directed, most portion of the nuchal crest partially de- wedge-shaped projection from the lateral limits the notch that presumably lodged the ¯ange of the petrosal. In posterior view (®gs. external acoustic meatus. 46 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247

The transition between the dorsal ¯ange PARIETAL and zygomatic process of the squamosal is The parietals are laminar bones that form indicated by a neck projecting anterolaterally the bulk of the roof of the cranial cavity. from the braincase to the glenoid fossa. In They are essentially lacking in PSS-MAE ventral view (®gs. 15, 21), a shallow depres- 101 (®g. 8) and are considerably damaged in sion, likely marking the course of the exter- PSS-MAE 113 (®g. 9); however, enough is nal acoustic meatus, runs along the long axis preserved to provide the major morphologi- of the neck and is continued medially on the cal details. petrosal. The anterior limit of the depression The parietals are limited to the dorsal por- is marked by a blunt, low transverse ridge, tion of the braincase (®gs. 8, 9, 32) and are which extends between the medial margin of only moderately convex. Anteriorly, they the glenoid and the crista parotica on the pe- contact the frontals at a broad U-shaped su- trosal. The glenoid fossa is essentially ¯at, ture; lateral to the arms of the U, a narrow with the anteromedial and posterolateral cor- anterior process of the parietal extends for- ners projecting slightly ventrally and the pos- ward, nearly to the lacrimal in the orbital rim, teromedial corner being elevated. Its outline to form the posterior part of the supraorbital is slightly teardrop-shaped, with the major notch and supraorbital crest. Behind the an- axis running from anterolateral to postero- terior process, level with the posterior border medial. This axis forms an abrupt angle with of the frontals, is the triangular, posterolat- the squamosal portion of the zygomatic arch, erally and ventrally directed postorbital pro- an unusual feature among mammals, which cess (``pop'' in ®gs. 12, 19). As originally imparts a distinctive outline to the skull of preserved, the process was long in PSS-MAE Kryptobaatar and several other Mongolian 113 but was damaged during preparation. It Late Cretaceous multituberculates. In front is completely preserved on both sides in a of the glenoid, the zygomatic process is skull referred to Kryptobaatar (PSS-MAE ¯angelike and has an extensive oblique con- 127). The posterior margin of the postorbital tact with the maxilla; it is considerably short- process is continuous with weakly developed er and weaker than is the maxilla's contri- temporal ridges (``tr'' in ®g. 23). The tem- bution to the zygoma. As described above, poral ridges are not fully preserved in PSS- the squamosal and maxilla support the jugal, MAE 113, but in PSS-MAE 127 the tem- which lies medial to both bones but has the poral ridges do not meet on the midline to greater part of its contact with the maxilla. form a sagittal crest. Instead, the ridges ap- In external view, lateral to the glenoid fossa proximate each other posteriorly, delimiting and extending anterior to it on the zygomatic a broad middorsal ridge. Posteriorly, the pa- process, there is an arched ridge, concave in- rietals contact the supraoccipital at a mostly feriorly, the intermediate zygomatic ridge transverse suture, and together with the su- (``izr'' in ®g. 12) of Gambaryan and Kielan- praoccipital form the dorsal tip of the nuchal Jaworowska (1995). This ridge marks the crest. An interparietal is not present. On its dorsal border of a shallow depression, which ventrolateral surface, the parietal contacts, these authors suggested was for the origin of from front to back, the frontal and the ante- the posterior part of the super®cial masseter rior lamina with an essentially straight su- muscle. This intermediate ridge is con¯uent ture, and the squamosal with a suture that is with the anterior zygomatic ridge on the concave medially. The contact with the fron- maxilla, in contrast with the condition re- tal and squamosal is fairly narrow, whereas ported by Gambaryan and Kielan-Jaworows- that with the anterior lamina is extensive. ka (1995) in Nemegtbaatar, Chulsanbaatar, and Catopsbaatar where the anterior and in- SUPRAOCCIPITAL termediate ridges are not con¯uent. A third, more posterior ridge, the posterior zygomatic The supraoccipital is a laminar bone that ridge, described for Nemegtbaatar, Chulsan- is essentially con®ned to the occiput; it is baatar, and Catopsbaatar (Gambaryan and well preserved in both PSS-MAE 101 (®g. Kielan-Jaworowska, 1995), is not discernible 16) and 113 (®g. 17). in Kryptobaatar. The supraoccipital is a hexagonal element 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 47

(®g. 35). Its contacts on the occiput are with (``ocon'' in ®gs. 10, 12, 16) with the major the petrosals laterally and with the exocci- axis oriented from dorsolateral to ventrome- pitals ventrolaterally. A small portion of the dial. The articular surface on the condyle can supraoccipital forms the dorsal margin of the be divided into two parts: (1) a dorsal one foramen magnum and, with the parietal, the contributing to the lateral margin of the fo- dorsal part of the nuchal crest. The nuchal ramen magnum and extending anterolaterally crests ¯are out posterolaterally along the with a similar orientation as the occipital contact between the mastoid and squamosal; plate, and (2) a ventral one contributing to the medial continuation of these crests along the ¯oor of the foramen magnum and lying the supraoccipital±parietal suture is less chie¯y in a horizontal plane. The left and prominent and is moderately notched in the right condyles bound approximately half of sagittal plane. The supraoccipital is slightly the foramen magnum (®gs. 16, 17). They do convex dorsoventrally and concave medio- not contact each other in the midline, but are laterally. The occipital plane in PSS-MAE separated by a distinct odontoid notch (``on'' 101 as determined by the supraoccipital in- in ®g. 37A). Despite the fact that the sutures clines anteriorly at 35Њ from the vertical. In cannot be traced, it is likely that the medial- PSS-MAE 113, the supraoccipital is more most portion of the condyles and the odon- concave mediolaterally and a little taller than toid notch are actually formed by the basi- in PSS-MAE 101. occipital as, for example, in monotremes (Kuhn, 1971; Zeller, 1989). EXOCCIPITAL The exoccipital is one of the major components of the jugular fossa, a deep excava- The exoccipital has contributions to the tion of the basicranium posterior to the prom- occiput and basicranium and can arbitrarily ontorium of the petrosal (®gs. 14, 15, 20, 21, be divided into three parts: the occipital 37A). The medial wall of the jugular fossa is plate, the condyle, and the contribution to the formed by the exoccipital and basioccipital; jugular fossa. The sutures delimiting the the posterior wall largely by the exoccipital exoccipital from its neighbors are distinct on and the paroccipital process of the petrosal; the occiput, but are not as discernible in other and the lateral wall entirely by the petrosal, areas. The exoccipital is well preserved in the crista interfenestralis in the posterior part PSS-MAE 101 (®gs. 14, 16) and 113 (®gs. and the promontorium anteriorly. In addition 15, 17), but its contribution to the jugular to contributing to the posterior wall, the an- fossa has not been cleaned of matrix on the terior margin of the condyle projects as a thin left side of the former. lamina that partially ¯oors the jugular fossa. The occipital plate of the exoccipital is a In fact, the jugular fossa is deeply recessed subrectangular bony lamina that forms one- into all the surrounding bones, so that its size third of the rim of the foramen magnum and is considerably larger than what is visible in extends anterolaterally from the edges of that ventral view. Inside the fossa, the exoccipital aperture (®g. 35). Its contacts on the occiput is the major constituent of the roof. It con- are with the supraoccipital dorsomedially and tacts the petrosal through an oblique suture with the petrosal ventrolaterally. The suture running anteriorly from lateral to medial. with the supraoccipital is straight and runs Along this suture is the small jugular fora- from dorsolaterally to ventromedially; the su- men, which occupies the posterolateral cor- ture with the petrosal is arched and extends ner of the fossa, posteromedial to the peri- ventrolaterally from the point where the lymphatic foramen (®g. 37A). A possible su- exoccipital, supraoccipital, and petrosal meet ture between the basioccipital and exoccipital above the level of the condyle to the medial is visible in the roof of the jugular fossa on slope of the paroccipital process. Along the the right side of PSS-MAE 101; it begins just petrosal suture, there is a deep pit that is con- anteromedial to the jugular foramen and ex- tinuous with the atlantal fossa on the petrosal tends posteromedially but cannot be traced that probably lodged part of the atlas when onto the condyle. the skull was maximally extended. On the right side of PSS-MAE 101, deep The condyle is a rounded structure in the jugular fossa and posterolateral to the 48 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247 jugular foramen, is a small, circular foramen curved, with the basioccipital becoming (not visible in the ®gures). An opening in a slightly wider posteriorly along its petrosal similar position may be present on the left contact. The maximum width of the basioc- side of PSS-MAE 113, but it is not as clear. cipital is at the level of the jugular fossa, and The foramen in PSS-MAE 101 is interpreted then it tapers posteriorly to the odontoid as for the hypoglossal nerve. The right jug- notch. As stated above, the suture between ular fossa in PSS-MAE 113 seems to be well the basioccipital and exoccipital in the con- preserved, but in the area where a hypoglos- dylar region is not discernible, but it was sal foramen is expected, no aperture is found. likely oblique, running anterolaterally from It is possible, although unlikely, that the fo- the odontoid notch and the medial aspect of ramina mentioned above are artifacts and the condyle. A possible oblique suture sep- that the true hypoglossal foramina are con- arates these same two bones in the jugular cealed by matrix in the caudalmost portion fossa on the right side of PSS-MAE 101. of the jugular fossa. In the basicranial axis, the anterior part of According to Kielan-Jaworowska et al. the basioccipital is essentially ¯at, whereas (1986), the large size of the jugular fossa in posteriorly there is a concavity on the mid- Mongolian Late Cretaceous multitubercula- line. In the jugular fossa, the basioccipital tes suggests the presence of large ganglia on contributes a fairly vertical wall along the the nerves below the jugular foramen. How- medial aspect and the anteromedial part of ever, large ganglia alone cannot account for the roof. The ventral edges of the basiocci- the immense jugular fossa in Kryptobaatar, pital's contribution to the jugular fossa especially considering the minute size of the strongly project laterally to form a partial jugular foramen. Therefore, we think that the ¯oor for that depression. structural continuity between the jugular fossa and the middle-ear space suggests that the ENDOCRANIUM fossa housed a diverticulum of the cavum tympani (see also Rougier et al., 1996c). Described here are structures on the endocranial surface of the braincase preserved BASIOCCIPITAL in three skulls of Kryptobaatar, PSS-MAE 123 (®g. 25), 124, and 125 (®g. 26). Com- The basioccipital is a long, narrow, lami- parative specimens also employed include nar bone that forms the base of the skull an- two indeterminate skulls resembling Krypto- terior to the foramen magnum and medial to baatar, PSS-MAE 126 and 128. A recon- the ear regions. It can be divided into two struction of the endocranial ¯oor is shown in parts: one on the basicranial axis, and the ®gure 27. As is evident from the exterior, the other contributing to the jugular fossa. The braincase in Kryptobaatar is formed by the basioccipital is best preserved in PSS-MAE exoccipitals, basioccipital, supraoccipital, pe- 101 (®g. 14), having been distorted in PSS- trosals, parietals, frontals, and the sphenoid MAE 113 in which the left petrosal is dis- complex; as stated above, the squamosal placed (®g. 15). does not contribute directly to the braincase The basioccipital contacts the exoccipital wall. With the exception of the petrosals, su- posterolaterally, the petrosal laterally, and the tures delimiting the individual braincase el- basisphenoid anteriorly (®gs. 14, 34). In ad- ements are subject to some degree of uncer- dition, there may even be a contact with the tainty in the specimens showing the endocra- pterygoid along the lateral edge of the basi- nium. Consequently, we base our descrip- occipital in front of the petrosal contact, de- tions on the internal morphology and pending on the composition of the crest ex- attribute features to individual elements in tending posteriorly from the pterygopalatine light of our understanding of similar struc- ridge (see Sphenoid Complex). The suture tures in extant mammals in which the bound- with the basisphenoid is straight and trans- aries are known. The dorsal components of verse, just behind the level of the back of the the braincase (i.e., the frontal and parietal) pterygopalatine ridges. The suture with the were effectively missing in the specimens promontorium of the petrosal is gently described herein, providing easy access to 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 49 the more complex ¯oor and side wall of the endocranium. Because endocrania are seldom preserved, let alone described, we add the following remarks by way of introduction to this anatomical region. The orbitotemporal region of the mammalian skull has a primary braincase wall formed by the chondrocranium and dura mater, and external to that, a secondary braincase wall formed by different patterns of several skeletal elements in different taxa (De Beer, 1937; Moore, 1981; Kuhn and Ze- ller, 1987). The extradural space between the primary and secondary wall is the cavum epiptericum (Gaupp, 1902, 1905), which houses the trigeminal and facial ganglia and is traversed by various cranial nerves and blood vessels. The cranial nerves enter the cavum epiptericum from the brain through speci®c gaps between near vertical bars or pillars of chondrocranial cartilage (®g. 24). The pattern of these gaps (i.e., how many are present and their contents) differs in monotremes, marsupials, and placentals (see below; Kuhn, 1971; Kuhn and Zeller, 1987). There is a general consensus among mor- phologists (e.g., Starck, 1967, 1978; Kuhn, 1971; Moore, 1981) that the chondrocranium in the common ancestor of mammals had three pillars in the orbitotemporal region between the nasal and otic capsules, as occurs in most extant sauropsids (®g. 24A; De Beer, 1926, 1937; Bellairs and Kamal, 1981). From anterior to posterior, these are the pila preoptica, pila metoptica, and pila antotica (®g. 24B). These three pilae provide borders for the apertures that transmitted cranial nerves through the primary braincase wall Fig. 24. Schematic drawings of embryonic (Kuhn and Zeller, 1987; Zeller, 1989). An- chondrocrania in left lateral view. A, generalized teriorly, between the nasal capsule and the sauropsid; B, generalized multituberculate, as re- pila preoptica is the orbitonasal foramen constructed from here; C, generalized monotreme; transmitting the ethmoidal branch of the oph- D, generalized marsupial; E, generalized placen- tal. Abbreviations: ais anterior intercavernous si- thalmic nerve into the nasal cavity. Posterior nus; cev capsuloparietal emissary vein; ea eth- to that, between the pila preoptica and me- moidal artery; en ethmoidal nerve; ev ethmoidal toptica is the optic foramen transmitting the vein; ips inferior petrosal sinus; nc nasal capsule; optic nerve into the orbit. Next, between the oa ophthalmic artery; oc otic capsule; pan pila pilae metoptica and antotica is the metoptic antotica; pis posterior intercavernous sinus; pm foramen transmitting the oculomotor nerve pila metoptica; pp pila preoptica; ps prootic sinus; into the orbit (or into the front of the cavum pv pituito-orbital vein; II optic nerve; III oculo- epiptericum). Finally, the gap between the motor nerve; IV trochlear nerve; V trigeminal pila antotica and the otic capsule, the prootic nerve; VI abducens nerve, VII facial nerve. foramen, transmitted the trochlear, trigeminal, and abducens nerves into the cavum 50 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247 epiptericum. Immediately posterior to that, tic foramen (e.g., Sphenodon, Bellairs and the facial nerve entered the back of the ca- Kamal, 1981). In mammals, a comparable vum epiptericum through a separate opening, vein drains medially from the cavernous si- the primary facial foramen, between the otic nus within the cavum epiptericum, immedi- capsule and a bar of cartilage, the prefacial ately anterior to the pila antotica into the hy- commissure, connected to the front of the pophyseal fossa in the echidna (Gaupp, otic capsule. 1908) and the platypus (personal obs.); in the A reconstruction of the vessels passing latter, the vein exits the skull via the carotid through the primary wall in basal mammals foramen. This is not the only vein draining has not yet been proposed, and we offer the the pituitary, and, following a suggestion of following model here (®g. 24A, B). Judging one of our reviewers, Robert Presley, we re- from the anatomy of extant amniotes, blood fer to it as the pituito-orbital vein in light of vessels ran through all of the apertures in the its pathway. Marsupials and placentals do not primary wall named above except the pri- have a strictly comparable vein, but it is pos- mary facial foramen. Also in the orbitonasal sible that the stem of the anterior intercav- foramen were ethmoidal arteries and veins; ernous sinus of therians (Shindo, 1915) is ho- in the optic foramen, the ophthalmic branch mologous with the pituito-orbital vein of of the internal carotid artery; in the metoptic monotremes and sauropsids. foramen, the pituito-orbital vein; and in the (4) Prootic sinus and inferior petrosal si- prootic foramen, the prootic vein and inferior nus in the prootic foramen: The prootic sinus petrosal sinus. Our justi®cation for this mod- (middle cerebral vein) exits the prootic fo- el is as follows. ramen to join the lateral head vein lateral to (1) Ethmoidal vessels in the orbitonasal the otic capsule in sauropsids, monotremes, foramen: Vessels accompanying the ethmoid- and some marsupials (Shindo, 1915; Wible, al nerve are typically present in both saurop- 1990; Wible and Hopson, 1995). Moreover, sids (Shindo, 1914) and mammals (Tandler, a canal housing this vessel is widely distrib- 1899). uted among the extinct outgroups to mam- (2) Ophthalmic artery in the optic fora- mals (Wible and Hopson, 1995; Rougier et men: The ophthalmic artery has been recon- al., 1996a). Among extant amniotes, an in- structed by Miao (1988), citing De Beer ferior petrosal sinus is known only in mam- (1937), with the oculomotor nerve and the mals, in which it runs posteriorly from the pituitary vein in the metoptic foramen in cavernous sinus within the cavum epipteri- Lambdopsalis. However, we disagree with cum, through the prootic foramen into the this placement of the ophthalmic artery in cranial cavity (Shindo, 1915; Rougier et al., Lambdopsalis and in basal mammals. There 1996a). are some sauropsids in which the ophthalmic As stated above, the pattern of the aper- artery enters the orbit via the metoptic fora- tures in the primary braincase wall of basal men (e.g., Crocodilus, Shiino, 1914) or even mammals is altered in different ways in via a separate foramen that secondarily fuses monotremes, marsupials, and placentals. All with the metoptic foramen (e.g., Chrysemys, three retain the pila preoptica, but that is the Shaner, 1926; Sphenodon, Bellairs and Ka- extent of the resemblance between them. mal, 1981). Yet, other sauropsids (e.g., La- (1) Monotremes (®g. 24C): In addition to certa, Shindo, 1914; Platydactylus, Hafferl, the pila preoptica, a complete pila antotica 1921) and mammals (Tandler, 1899; Wible, forms in the chondrocranium of both the 1984) have the ophthalmic artery and optic echidna (Gaupp, 1908; Kuhn, 1971) and nerve intimately associated. platypus (Zeller, 1989). However, the pila an- (3) Pituito-orbital vein in the metoptic fo- totica regresses during ontogeny, and in the ramen: A pituitary vein running from the pi- adult it is represented by only its slender os- tuitary (hypophysis) to the orbit is broadly si®ed base on the basisphenoid, the middle distributed among sauropsids (Bruner, 1907), clinoid process (``mclp'' in ®g. 28; Kuhn, and is either in the metoptic foramen (e.g., 1971; Zeller, 1989). The pila metoptica fails Lacerta, De Beer, 1937) or in a separate fo- to form, leaving con¯uent the optic and me- ramen that secondarily fuses with the metop- toptic foramina, which together are termed 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 51 the pseudoptic foramen (``psf'' in ®g. 28; marsupials, the prootic sinus involutes in pla- Gaupp, 1908). The pseudoptic foramen centals and is replaced by a secondary vessel, transmits the optic and oculomotor nerves the capsuloparietal emissary vein (Gelderen, (Kuhn, 1971; Zeller, 1989), the pituito-orbit- 1924). al vein (Gaupp, 1908; personal obs.), and in the echidna the ophthalmic artery (Tandler, EXOCCIPITAL 1901); in the platypus, the ophthalmic artery In PSS-MAE 123 (®g. 25B), a small frag- branches off the maxillary artery in the orbit ment of the left exoccipital has been pre- (Wible, 1984). The contents of the prootic served in articulation with the petrosal, pos- foramen do not differ from that inferred for teromedial to the subarcuate fossa. Along the basal mammals. suture between the exoccipital and petrosal, (2) Marsupials (®g. 24D): The pila preop- a broad groove directed ventromedially is in- tica is the only pillar to form in the orbito- terpreted as having housed the sigmoid sinus. temporal region of the primary wall in mar- In PSS-MAE 125 (®g. 26), an even smaller supials (Kuhn and Zeller, 1987; Maier, 1987). portion of the right exoccipital completing Consequently, the optic, metoptic, and pro- the posteromedial margin of the jugular fo- otic foramina are con¯uent. Transmitted ramen has been preserved. The contribution through this large gap are the second through of this bone to the jugular foramen is clearly sixth cranial nerves, the ophthalmic artery indicated by sutures. Along the suture be- (Tandler, 1899), the anterior and posterior in- tween the exoccipital and petrosal, directly tercavernous sinuses, the inferior petrosal si- posterior to the jugular foramen, is an ante- nus, and the prootic sinus (Shindo, 1915). riorly directed, digitiform process of the The prootic sinus is present during early on- exoccipital that impinges on the lumen of the togenetic stages in all marsupials investigat- foramen. Also in this specimen, a longer seg- ed to date, but is retained only in adult di- ment of the sulcus for the sigmoid sinus is delphids, caenolestids, and some dasyurids clearly preserved (``sss'' in ®g. 26). It does (Wible, 1990; Wible and Hopson, 1995). not approach the very small jugular foramen, This single large aperture transmitting all but instead is directed toward the place these structures in the marsupial chondrocra- where the missing foramen magnum would nium has been referred to (e.g., Cords, 1915) be expected. An endocast of the sigmoid si- as the sphenoparietal fenestra. Following nus is preserved on both sides in the undes- most recent authors (e.g., Kuhn and Zeller, cribed skull of multituberculate, n. sp. (PSS- 1987), we reserve the similar term of sphe- MAE 126), and it follows the pattern de- noparietal foramen for the opening in the scribed above in PSS-MAE 125. chondrocranium of eutherians that transmits the third through sixth cranial nerves (see be- BASIOCCIPITAL low). (3) Placentals (®g. 24E): In addition to the In PSS-MAE 123, the rostral portion of pila preoptica, a complete pila metoptica the basioccipital is preserved; a smaller frag- forms in placentals and thus completes a true ment is found in PSS-MAE 125; and the en- optic foramen in the orbitosphenoid of the tire bone is likely present in PSS-MAE 124, adult (De Beer, 1937; Starck, 1967), which but is not accessible for study. As preserved in most forms transmits the ophthalmic ar- in PSS-MAE 123 (®g. 25), the basioccipital tery along with the optic nerve (Tandler, is a ¯at, slightly concave bone that is rather 1899, 1901; Wible, 1984). The pila antotica, featureless. As gleaned from both specimens, however, fails to develop, leaving con¯uent the suture between the basioccipital and ba- the metoptic and prootic foramina, which to- sisphenoid is situated immediately behind the gether are termed the sphenoparietal foramen dorsum sellae and runs transversely. The su- (Voit, 1909). The sphenoparietal foramen ture between the basioccipital and petrosal transmits the third through sixth cranial runs subparallel to the sagittal plane, con- nerves, the anterior and posterior intercav- verging slightly anteriorly. The basioccipital ernous sinuses, and the inferior petrosal sinus forms only a narrow portion of the skull (Shindo, 1915). Unlike monotremes and base. In PSS-MAE 123, the basioccipital is 52 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247

Fig. 25. Stereophotograph of the ¯oor of the endocranium of Kryptobaatar dashzevegi PSS-MAE 123 in dorsal (A) and oblique dorsal (B) views, with accompanying line drawings. Gray pattern represents matrix; parallel lines denote breakage. Abbreviations: al anterior lamina; bo basioccipital; ce cavum epiptericum; ds dorsum sellae; exoc exoccipital; ff facial foramen; ®ca foramen for internal carotid artery; fpv foramen for pituito-orbital vein; fr frontal; fV3 foramen for mandibular nerve; hf hypophyseal fossa; iam internal acoustic meatus; jn jugular notch; jsp jugum sphenoidale; mef metoptic foramen; mx maxilla; na nasal; opf optic foramen; or orbitosphenoid; ow orbital wing (pila preoptica); pan pila antotica; prc prootic canal; pm pila metoptica; sf subarcuate fossa; sphf sphenorbital ®ssure; tus tuberculum sellae. 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 53

PSS-MAE 128, which is not Kryptobaatar but is referred here as an indeterminate multituberculate, conforms to the pattern of the skull base described above. The only substantive exception is that instead of a shallow depression, there is a deep trough on the basioccipital running to the basioccipital-petrosal suture. This unambiguously indicates the endocranial course of the inferior petrosal sinus.

PETROSAL As was the case in our description of the exterior of the petrosal, when describing the interior, the anterior lamina is considered along with the petrosal proper. The following descriptions are based on the three specimens of Kryptobaatar with exposed the endocranial surfaces (i.e., PSS-MAE 123, 124, and 125). The endocranial surface of the petrosal can be subdivided into four quadrants by two in- tersecting crests. One crest is the crista petrosa, which runs along the prefacial commissure (``pfc'' in ®g. 26) and the anterior margin of the subarcuate fossa (``sf'' in ®gs. 25±27); the other crest separates the subarcuate fossa from the internal acoustic meatus (``iam'' in ®gs. 25±27) and continues ante- roventrolaterally as the dorsolateral margin of the cavum epiptericum (``ce'' in ®gs. 25B, 26, 27). The crests intersect dorsolateral to the internal acoustic meatus in a rounded em- inence. The four quadrants so delimited are identi®ed by their most conspicuous features: the subarcuate fossa, the internal acoustic meatus, the cavum epiptericum, and the anterior lamina (®g. 27). Fig. 25. Continued. The subarcuate fossa, which accommodated the para¯occulus of the cerebellum, is a broken across the anteriormost portion of the large, deep, subspherical depression that jugular fossa, and the preserved fragment in- opens into the braincase through an elliptical dicates that the fossa deeply excavates the aperture, the major axis of which is dorso- basioccipital, leaving only a thin lamella sep- ventrally oriented (®gs. 25, 26). The dorsal- arating the cranial cavity from the middle- most margin of the subarcuate fossa is not ear cavity. Left and right jugular fossae also preserved or fully prepared in any specimen. converge medially, being separated in the Considering the presence of the large vas- midline by a bridge of the basioccipital bone cular groove in the exoccipital already de- less than 1 mm in thickness. Along the ba- scribed, it is likely that a groove for the sig- sioccipital±petrosal suture is a shallow de- moid sinus was present on the dorsal edge of pression likely occupied by the inferior pe- the subarcuate fossa. A similar structure is trosal sinus. present in the skull of multituberculate, n. sp. 54 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 55

(PSS-MAE 126). Along the posteroventral ertures. The ventral (``fv3'' in ®g. 25) and edge of the subarcuate fossa, in the crest sep- lateral openings transmitted branches of the arating the subarcuate fossa and the internal mandibular nerve, and the large, subcircular acoustic meatus, there is small foramen, the anterior opening transmitted the trochlear, vestibular aqueduct, with a posterodorsally abducens, ophthalmic, and maxillary nerves, directed groove, interpreted as having housed given that the cavum shows the same rela- the endolymphatic duct. tionship to the pila antotica (see Sphenoid The quadrant housing the internal acoustic Complex below) as in monotremes (Kuhn meatus is the part that contributes to the ¯oor and Zeller, 1987). The anterior opening, the of the braincase, and it does so in an oblique sphenorbital ®ssure (``sphf'' in ®gs. 25A, fashion (®gs. 25±27). Medially, the meatal 27), probably also transmitted the ramus in- quadrant contacts the basioccipital and exoc- fraorbitalis of the stapedial artery and the cipital via a straight parasagittal suture; an- ophthalmic veins (Rougier et al., 1992; Wi- teriorly, it cannot be differentiated from the ble and Hopson, 1995). The cavum epipte- sphenoid complex. Apparently, the sphenoid ricum is longer than wide, and its anterior complex and petrosal are fused endocra- portion is roofed by the laminar, laterally ex- nially, although a distinct morphological dis- panded pilae antotica and metoptica, which continuity that forms the dorsum sellae prob- connect the medial and lateral crests delim- ably indicates the anterior extent of the pe- iting the trigeminal fossa. Based on PSS- trosal. A cochlear aqueduct is not identi®able MAE 123, it is apparent that a separate ca- in any specimen, and hence the aperture on vum supracochleare for the facial ganglion is the back of the promontorium is the perilym- lacking in Kryptobaatar; the primary facial phatic foramen, as occurs in monotremes foramen opens directly into the back of the (Kuhn, 1971; Zeller, 1985, 1989, 1991). The cavum epiptericum where the facial ganglion posteromedial margin of the meatal quadrant would have been located. The secondary exit is notched by the small jugular foramen (®g. of the facial nerve from the cavum epipteri- 26). The internal acoustic meatus is a sub- cum was likely in the posterior part of the circular depression, although its medial wall cavum's ¯oor, but it was inaccessible to prep- is not very tall, and the meatus is continuous aration. medially with the petrosal's contribution to The anterior lamina quadrant is the largest the braincase ¯oor. At the posteromedial and contributes to the side wall of the brain- edge of the meatus is an anteroventrally di- case (®gs. 25±27). It is concave in medial rected foramen, probably for the cochlear view and extends from the subarcuate fossa nerve. Separating this small foramen from forward to contact the primary wall of the the deeper part of the meatus is a low crest braincase represented by the pila antotica/or- that forms the posteromedial margin of a sec- bitosphenoid, which the anterior lamina ond aperture, probably for the facial and ves- overlaps laterally. Anterolateral to the subar- tibular nerves. cuate fossa in the anterior lamina quadrant is Lateral to the meatal quadrant is the ca- a deep, broad, anteroventrally directed sulcus vum epiptericum (®gs. 25±27), the very deep leading into a foramen (``prc'' in ®gs. 25, fossa that lodged the trigeminal or semilunar 27). This is the sulcus and foramen for the ganglion and provided passage to various prootic sinus, which leads into the prootic nerves and vessels. The cavum opens ante- canal. As preserved in the skull of PSS-MAE riorly, ventrally, and laterally via several ap- 123 (®g. 25), the foramen for the prootic si-

← Fig. 26. Stereophotograph of the ¯oor of the endocranium of Kryptobaatar dashzevegi PSS-MAE 125 in oblique dorsal view. Gray pattern represents matrix; parallel lines denote breakage. Abbreviations: al anterior lamina; bs basisphenoid; ce cavum epiptericum; ds dorsum sellae; fr frontal; iam internal acoustic meatus; jfo jugular foramen; mef metoptic foramen; mx maxilla; or orbitosphenoid; pan pila antotica; pet petrosal; pfc prefacial commissure; sf subarcuate fossa; sss sulcus for sigmoid sinus; tsc transverse sinus canal. 56 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247

Fig. 27. Reconstruction of the ¯oor of the endocranium of Kryptobaatar dashzevegi. Parallel lines represent the cut edge of the braincase. Abbreviations: al anterior lamina; bo basioccipital; ce cavum epiptericum; ds dorsum sellae; exoc exoccipital; ®ca foramen for internal carotid artery; fpv foramen for pituito-orbital vein; fr frontal; hf hypophyseal fossa; iam internal acoustic meatus; jfo jugular foramen; jsp jugum sphenoidale; mef metoptic foramen; ocon occipital condyle; opf optic foramen; or orbitosphenoid; pa parietal; pan pila antotica; pet petrosal; pm pila metoptica; pop postorbital process; prc prootic canal; sf subarcuate fossa; son supraorbital notch; sphf sphenorbital ®ssure; sq squamosal; sup supraoccipital. nus opens at a level ventral to the lower mar- ossi®ed pila antotica. Inside the hypophyseal gin of the subarcuate fossa. The anterior lam- fossa, the carotid foramina open in the pos- ina in front of the prootic sulcus is very terolateral corner of the ¯oor (``®ca'' in ®gs. smooth and essentially featureless. 25, 27). Laterally on the anterior wall of the hypophyseal fossa, in front of each carotid SPHENOID COMPLEX foramen, there is a small foramen in PSS- Given that the hypophyseal or pituitary MAE 123 (``fpv'' in ®gs. 25, 27) and 125, fossa is universally lodged in the basisphe- the only specimens preserving the relevant noid among recent mammals (Starck, 1967), area. This foramen connects the hypophyseal we identify the corresponding part of the fossa with the orbitotemporal fossa, and in sphenoid complex in Kryptobaatar as the ba- PSS-MAE 123 it is also connected to the ca- sisphenoid. The hypophyseal fossa (``hf'' in rotid foramen via a distinct sulcus in the lat- ®gs. 25, 27) is a deep, hemispherical depres- eral ¯oor of the hypophyseal fossa. The ex- sion on the skull base that is limited poste- ternal aperture of this foramen is shown on riorly by a prominent dorsum sellae, also on the right side of PSS-MAE 101 (in addition the basisphenoid (``ds'' in ®gs. 25, 26). The to 123 and 125). Based on our observations dorsum sellae forms a steep angle with the of serially sectioned platypuses, as well as ¯oor of the braincase behind it; the dorsum Gaupp's (1908) observations of the echidna, is lower in the midline, and taller and thicker we interpret this foramen and sulcus as for laterally where it becomes con¯uent with the the pituito-orbital vein. 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 57

Fig. 28. The ¯oor of the endocranium of the platypus Ornithorhynchus anatinus in dorsal view (modi®ed from Zeller, 1989: ®g. 6, with the author's permission). Parallel lines represent the cut edge of the braincase. Abbreviations: al anterior lamina; bo basioccipital; fV2 foramen for maxillary nerve; fV3 foramen for mandibular nerve; hf hypophyseal fossa; iam internal acoustic meatus; jfo ؉ hyf con¯uent jugular and hypoglossal foramina; jsp jugum sphenoidale; mlcp middle clinoid process (ossi®ed remnant of pila antotica); onsf orbitonasal foramen; pet petrosal; pp pila preoptica; prc prootic canal; psf pseudoptic foramen (for II, III, IV, V1, VI); ptc posttemporal canal; sf subarcuate fossa; sq squamosal; tus tuberculum sellae; vaq vestibular aqueduct.

The tall laminar walls forming the sides of ossi®ed pila antotica in other multitubercu- the hypophyseal fossa are here considered as lates, Kielan-Jaworowska et al. (1986), Miao the ossi®ed pilae antotica (``pan'' in ®gs. 25± (1988), and Hurum (1998a) have employed 27), a component of the primary wall of the the term taenia clino-orbitalis. This term was braincase (®g. 24A±C). These walls project originally coined by Gaupp (1902, 1908) for anteriorly and laterally as broad, winglike the well-developed pila antotica that he en- structures that connect the primary wall with countered in the chondrocranium of the the dermal elements that form the secondary echidna. Gaupp (1908: ®g. 56) also applied wall; sutures demarcating the lateral extent the term to the remnant of the pila antotica of the pila are clearly visible in PSS-MAE in the adult echidna skull, but the primary 125 (®g. 26). The pila antotica is particularly usage was for a cartilaginous structure. Con- thick posterodorsally, where in conjunction sequently, we continue to use the more gen- with the anterior lamina it roofs the cavum eral term pila antotica rather than the more epiptericum. Running behind the pila anto- restricted taenia clino-orbitalis. tica into the cavum epiptericum in Krypto- Lateral to the hypophyseal fossa, on the baatar were the trochlear, trigeminal, and ab- laminar extension of the pila antotica, there ducens nerves, based on the relationships is a deep trough that leads anteriorly to a that these structures exhibit in monotremes sizable round foramen at the level of the an- (®g. 24C; Kuhn and Zeller, 1987; Zeller, terior margin of the hypophyseal fossa 1989). In describing the similarly situated, (``mef'' in ®g. 25±27) that opens externally 58 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247 into the orbitotemporal fossa (``mef'' in ®g. the pila antotica, and the oculomotor as hav- 36A). This foramen is present in the three ing entered the orbit via a separate metoptic specimens of Kryptobaatar with the endo- foramen. cranium exposed as well as in the indeter- The region anterior to the hypophyseal minate multituberculate (PSS-MAE 128). fossa is damaged in all available specimens, Miao (1988) interpreted a similarly placed but a full restoration (®g. 27) can be offered aperture in Lambdopsalis as the metoptic fo- based on the fragments preserved in two ramen, the main occupant of which was the specimens of Kryptobaatar (PSS-MAE 123 oculomotor nerve. We report here a similar and 125; ®gs. 25, 26) as well as in the in- opening in the right orbit of Kamptobaatar determinate multituberculate (PSS-MAE (ZPAL MgM-I/33). Even though a metoptic 128). Extending forward from the hypophy- foramen is not known for any ontogenetic seal fossa on the midline is a rodlike portion stage in extant mammals, we accept Miao's of the sphenoid complex, the tuberculum sel- interpretation for Lambdopsalis and extend it lae (``tus'' in ®g. 25A). It extends anteriorly here for the same structures in Kryptobaatar, toward paired, laterally directed projections, Kamptobaatar, and PSS-MAE 128. As stated the orbital wings or alae orbitales (``ow'' in above, the metoptic foramen is found in the ®g. 25B). These wings are part of the orbi- chondrocranium of most extant sauropsids tosphenoid and represent the ossi®ed pilae between the pilae antotica and metoptica (®g. preoptica. The edge of the wings, the orbi- 24A; De Beer, 1926, 1937; Bellairs and Ka- tosphenoidal crest, dorsally delimits large, mal, 1981) and is generally considered to paired foramina that are not preserved intact have been present in the chondrocrania of the in any specimen, although they are nearly common ancestor of monotremes and the- preserved in PSS-MAE 123 (``opf'' in ®g. rians (Kuhn, 1971; Kuhn and Zeller, 1987). 25). These anterolaterally directed foramina, Following this, the bone enclosing the an- for the optic nerves and ophthalmic arteries, teromedial aspect of the foramen in Krypto- are situated between the orbital wings and baatar is the ossi®ed pila metoptica (``pm'' the tuberculum sellae, close to the midline. in ®gs. 25B, 27). The portion of the tuberculum sellae directly In amniotes, the two most signi®cant fo- medial to the optic foramina is grooved ramina in the interval between the passage- transversely by a shallow sulcus, the chias- ways for the optic and trigeminal nerves are matic sulcus for the optic chiasm. the metoptic foramen and the foramen for the In front of the optic foramina, the brain- abducens nerve (De Beer, 1926, 1937). Al- case ¯oor between the right and left orbital though the abducens nerve runs through the wings is marked by a midline trough of un- prootic foramen in monotremes and many certain function, which becomes narrower sauropsids (®g. 24A, C), it pierces the base and shallower rostrally. This midline portion of the pila antotica at the level of the dorsum of the sphenoid complex in this region is the sellae in various sauropsids (De Beer, 1926, yoke or jugum sphenoidale (``jsp'' in ®gs. 1937; SaÈve-SoÈderburgh, 1947; Oelrich, 25A, 27). The overall orientation of the ju- 1956), and it has been similarly interpreted gum sphenoidale is anterodorsally. Slightly in various extinct forms, including the prim- rostral to the end of the midline trough, the itive non-mammalian cynodont Thrinaxodon cranial cavity is constricted, demarcating the (Parrington, 1946), the tritylodontid Oli- posterior boundary of the olfactory bulbs. gokyphus (KuÈhne, 1956), and the tritheledon- The rostralmost extent of the cranial cavity tid Diarthrognathus (Crompton, 1958). The housing the olfactory bulbs is not prepared aperture in question in Kryptobaatar, how- or preserved in any available specimens. In ever, is placed too far dorsally and forward, the indeterminate multituberculate (PSS- at the level of the front of the hypophyseal MAE 128), part of the ¯oor in front of the fossa, to be a foramen for the abducens nerve olfactory bulb constriction is preserved on (®gs. 25±27). Therefore, we reconstruct the the left side. The surface is smooth, devoid abducens nerve in Kryptobaatar as having of perforations, and corresponds to the lam- entered the cavum epiptericum with the ina infracribrosa. Laterally, where the lamina trochlear and trigeminal nerves, posterior to infracribrosa contacts the jugum sphenoidale 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 59 behind, there is a foramen from which a siz- viously for Lambdopsalis (Miao, 1988) and able sulcus runs anterodorsally. This foramen Nemegtbaatar (Hurum, 1994), and fragments and sulcus are interpreted as for the endocra- only for Chulsanbaatar (Hurum, 1994). nial portion of the ethmoidal nerve and ves- Among Mesozoic mammaliaforms, turbinals sels. None of the available specimens exhib- have been preserved in the docodontid Hal- its any indication of an ossi®ed cribriform danodon (Lillegraven and Krusat, 1991) and plate. Our observation should be regarded as the zalambdalestid Barunlestes (Kielan-Ja- provisional, because the preservation of the worowska and Tro®mov, 1980). available specimens hindered the investiga- tion of this region of the endocranium. MANDIBLE In the three specimens of Kryptobaatar showing the endocranial surfaces (PSS-MAE As is typical for the lower jaws or man- 123, 124, and 125) and in the indeterminate dible in multituberculates, that of Kryptobaa- multituberculate (PSS-MAE 128), a broad tar has a robust, deep horizontal ramus and canal connects the right and left sphenorbital an ascending ramus that is proportionally recesses, that is, the space medial to the walls smaller. Housed in the lower jaw are ®ve demarcating the lateral rims of the sphenor- teeth: an enlarged, procumbent incisor (i1), bital ®ssures. This canal is low on the brain- two premolars (p3±p4) with the mesial one case ¯oor and traverses the primary wall greatly reduced, and two molars (m1±2). ventral to the tuberculum sellae; a thin wall Both lower jaws are complete and in place separates this canal from the hypophyseal in PSS-MAE 101 (®g. 14). Only the left one fossa (``tsc'' in ®g. 26). Among living mam- is preserved in PSS-MAE 113; it is well pre- mals, the transverse canal is in a similar po- served except that the ventral margin under sition, crossing the midline immediately an- the incisor alveolus is missing (®g. 5). We terior to the hypophyseal fossa, and it trans- describe the lower jaw ®rst in lateral and then mits a vein called the transverse canal vein in medial views; the terminology used fol- in some marsupials (Archer, 1976; Marshall lows that of Gambaryan and Kielan-Jawo- et al., 1990; Marshall and Muizon, 1995) and rowska (1995) unless noted otherwise. some placentals (McDowell, 1958; MacPhee, In lateral view (®gs. 10, 12, 29), the al- 1994). This is the only likely model for the veolus of the incisor is separated from p3 by similarly situated canal in Kryptobaatar (see a long, dorsally concave diastema. In PSS- ®g. 36) and the indeterminate multitubercu- MAE 113, in which the ventral margin of the late. lower jaw is broken, the root of the incisor The occurrence of turbinals or ridges for can be traced posteriorly at least to the mid- the turbinals on the appropriate bones in the dle of m1. The mental foramen (``mf'' in ®g. nasal cavity could not be studied in any of 12) is positioned in front of the lateral bulge the specimens of Kryptobaatar considered in over the roots of the premolars; it lies closer this report. However, the natural nasal en- to the superior than to the inferior margin of docast of an indeterminate multituberculate the lower jaw. In PSS-MAE 101 (®g. 10), it from Ukhaa Tolgod (PSS-MAE 134) shows is even closer to the superior margin and to ridges in the maxillary, nasal, and frontal the lateral bulge over the premolar roots than portions of the nasal cavity (Rougier et al., in PSS-MAE 113 (®g. 29). The mental fo- 1997b). These ridges correlate closely with ramen is quite large and faces anterolaterally. similar structures present in living mammals The area on the horizontal ramus between supporting the turbinals (Paulli, 1900; the lateral bulge over the premolar roots and Moore, 1981; Hillenius, 1994). PSS-MAE the base of the coronoid process is very short 134 con®rms the presence of ethmoturbinals and concave. This is the masseteric fovea and nasoturbinals, and a crest running par- (``mafo'' in ®g. 12), which Gambaryan and allel to the sulcus for the nasolacrimal duct Kielan-Jaworowska (1995) interpreted as for suggests the existence of maxilloturbinals. the pars anterior of the medial masseter mus- Also, fragments of turbinals appear to be pre- cle. As noted by these authors, the masseteric served. Among multituberculates, turbinal fovea in Kryptobaatar is con¯uent posteri- ridges and fragments have been reported pre- orly with the masseteric fossa. The masse- 60 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247

Fig. 29. Stereophotograph of the left lower jaw of Kryptobaatar dashzevegi PSS-MAE 113 in lateral view. teric fossa (``maf'' in ®g. 12) is shallow, masseteric line is nearly straight and is con- broad, and subdivided by a low, oblique tinuous posteriorly with the nearly vertical, ridge running anteroventrally from the con- straight rear edge of the condylar process. dyle, which we call the condyloid crest fol- The coronoid process (``cor'' in ®gs. 8, 16) lowing the terminology used for the dog originates below the alveolar margin and its (Evans and Christensen, 1979). The portion anterior extent lies at the back edge of the of the masseteric fossa anterodorsal to the anterior root of m1 (®g. 29). Between the condyloid crest is larger and deeper than the lateral alveolar margin and the coronoid pro- posteroventral one; these two fossae proba- cess there is a trough running obliquely pos- bly accommodated different parts of the mas- terodorsally, the temporal groove (``tg'' in seter muscle (see Gambaryan and Kielan-Ja- ®g. 10), which is said to be for the pars an- worowska, 1995). The anteroventral margin terior of the temporalis muscle (Gambaryan of the masseteric fossa is formed by the mas- and Kielan-Jaworowska, 1995). The coro- seteric crest (``mc'' in ®g. 10), a well-devel- noid process is subtriangular and only slight- oped ridge that is concave posteriorly and ly higher than the condyle and the molar con¯uent posteriorly with the weaker mas- cusps (®g. 29). The angle of the anterior bor- seteric line, as seen in PSS-MAE 101. The der of the coronoid process relative to the 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 61

Fig. 30. Stereophotograph of the left lower jaw of Kryptobaatar dashzevegi PSS-MAE 113 in medial view. toothrow is approximately 60Њ (contra Kie- is an asymmetrical distribution (Crompton lan-Jaworowska and Hurum, 1997, who re- and Hylander, 1986). port it as less than 45Њ). A sharp crest con- In medial view (®g. 30), the area of the nects the tip of the coronoid process with the symphysis is only visible in the left lower condyle, delimiting a shallow mandibular jaw of PSS-MAE 113. Most of this surface notch. The condyle (``con'' in ®gs. 6, 16) is has been lost, but enough remains to char- robust, lacks a distinct neck, and, as noted by acterize the symphysis as vertical and narrow Kielan-Jaworowska and Hurum (1997), is (``msy'' in ®g. 14). The dorsal surface of the positioned at approximately the same level as diastema forms a roughly horizontal shelf the molars. The articular surface on the con- that narrows posteriorly; the posterior nar- dyle is teardrop-shaped and directed poste- rowing of this shelf is a result of the oblique rodorsally. In posterior view, the articular orientation of the alveoli of the premolars surface is distributed symmetrically with re- and molars in the lower jaw. The most dis- spect to the plane of the dorsal border of the tinctive feature of the medial side of the ascending ramus. This is in contrast to the mandible is the very deep, enlarged ptery- condition in most mammals in which there goid fossa located behind the alveolar pro- 62 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247 cess. This fossa, for the medial pterygoid pterygoid fossa is a concave area below the muscle (Gambaryan and Kielan-Jaworows- condyle, the pterygoid fovea for the lateral ka, 1995), is delimited anteriorly by the pter- pterygoid muscle (Gambaryan and Kielan- ygoid crest (``pc'' in ®g. 16), which extends Jaworowska, 1995). The fovea is both small- posteroventrally as the medial margin of the er and shallower than the pterygoid fossa and prominent pterygoid shelf. The dorsal mar- is delimited posteriorly by the posterior margin of the pterygoid fossa is formed by a gin of the lower jaw. longitudinal, inconspicuous elevation, the Another measure of the lower jaw in mul- transversal elevation, and the posterior mar- tituberculates reported by Kielan-Jaworows- gin is formed by a crest that is the medial ka and Hurum (1997) is the angle between equivalent of the condyloid crest described the lower margin of the mandible and the on the lateral surface. Continuing forward occlusal surface of the molars. In Kryptobaa- from the pterygoid fossa just behind the level tar, this angle is very low, approximately of m2 is the mandibular canal. Behind the 11Њ.

VASCULAR RECONSTRUCTIONS Comprehensive reconstructions of the cra- niolabidid Lambdopsalis (Miao, 1988), a nial vascular system in multituberculates generalized ``taeniolabidoid'' (Rougier et al., were ®rst attempted by Kielan-Jaworowska 1992), and ``Catopsalis'' joyneri and Mesod- et al. (1984, 1986). These authors surveyed ma thompsoni (Wible and Hopson, 1995). the relevant osteology in the then currently The vascular reconstructions proposed known Mongolian Late Cretaceous taxa, but here for Kryptobaatar dashzevegi (®gs. 36B, they relied most heavily on cranial remains, 37B) are indebted to those already published, endocasts, and sectioned skulls of two forms, because the channels that previous authors Nemegtbaatar and Chulsanbaatar, along identi®ed in various multituberculates are, with isolated petrosals tentatively referred to with minor amendments, repeated in Krypto- the Late Cretaceous North American ``Ca- baatar. The basis for the reconstructions and topsalis'' joyneri and Mesodma thompsoni the terminology employed is our own pub- (see Wible and Hopson, 1995). Kielan-Ja- lished and unpublished studies of the cranial worowska et al. (1984, 1986) discovered that vessels occurring in extant mammals (Wible, most of the presumed vascular channels in 1984, 1986, 1987, 1989, 1990; Rougier et al., the multituberculates also occurred in other 1992; Wible and Hopson, 1995; and refer- Mesozoic mammaliamorphs, with Mamma- ences cited therein). liamorpha being the clade comprising the last common ancestor of Tritylodontidae and VEINS Mammalia plus all descendants (Rowe, 1988). Given that remnants of these channels In extant mammals, the veins of the head transport vessels in extant mammals, Kielan- can be divided into the intracranial dural si- Jaworowska et al. (1984, 1986) were able to nus system and the extracranial veins drain- offer reconstructions of the major cranial aring the dural sinuses and super®cial struc- teries and veins, including the dural sinuses, tures. Although the components of the dural in the multituberculates. In general, these re- sinus system are conservative among extant constructions have been followed by subse- mammals, the extracranial veins draining it quent researchers and applied to other Me- vary considerably (Gelderen, 1924; Wible, sozoic mammaliamorphs, although some re- 1990; Wible and Hopson, 1995). In mono- ®nements have been proposed by Wible tremes, the principal drainage is through the (1989) and co-workers (Rougier et al., 1992; prootic canal and foramen magnum (Hochs- Hopson and Wible, 1993; Wible and Hopson, tetter, 1896), whereas it is through the post- 1995). With these re®nements, vascular re- glenoid foramen and foramen magnum in constructions have been offered for the fol- marsupials (Gelderen, 1924; Archer, 1976) lowing multituberculates: the Paleocene tae- and through the postglenoid and internal jug- 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 63

Fig. 31. Reconstruction of the skull of Kryptobaatar dashzevegi in anterior view. Abbreviations: fr frontal; lac lacrimal; man mandible; mx maxilla; na nasal; pmx premaxilla; sq squamosal. ular foramina and foramen magnum in most showing the endocranial surfaces. The pro- placentals (Gelderen, 1924; Butler, 1967). otic sinus is indicated by a groove and foramen into the prootic canal in the anterior DURAL SINUS SYSTEM lamina quadrant (®gs. 25, 27), resembling those structures in monotremes (Wible and The dural sinus system reconstructed for Kryptobaatar resembles in most regards that Hopson, 1995). The sigmoid sinus is indi- proposed for Nemegtbaatar, Chulsanbaatar cated by grooves on the petrosal and exoc- (Kielan-Jaworowska, 1986; Kielan-Jawo- cipital (®g. 26), and it is apparent that the rowska et al., 1986), and Lambdopsalis sinus left the cranial cavity via the foramen (Miao, 1988). Included are superior (dorsal) magnum based on the extremely reduced size sagittal, transverse, sigmoid, and prootic si- of the jugular foramen and the sulcus pre- nuses; the ®rst three are nearly ubiquitous in served in PSS-MAE 125 (®g. 26). This pat- their incidence among extant mammals (Gel- tern is also clearly shown on the endocast of deren, 1924), whereas the prootic sinus is an undescribed skull of a new species of known only for adult monotremes and some Mongolian Late Cretaceous multituberculate adult marsupials (Wible, 1990; Wible and (PSS-MAE 126). Kryptobaatar shows no in- Hopson, 1995). The evidence for these dural dication of a tentorial sinus, described as sinuses in Kryptobaatar is in the form of draining into the prootic sinus from in front grooves on the endocranial surfaces and in Chulsanbaatar and Nemegtbaatar (Kie- molds of vascular channels on the endocasts lan-Jaworowska et al., 1986). that resemble those occurring in extant mam- Other major components of the dural sinus mals. The superior sagittal and transverse si- system that are ubiquitous among extant nuses are visible in dorsal view in the ex- mammals are the cavernous sinus, the major posed endocast of PSS-MAE 113; the supe- venous pool in the cavum epiptericum, and rior sagittal sinus lies between the right and its two principal conduits, the inferior (ven- left cerebral hemispheres, and the paired tral) petrosal sinus (petrobasilar sinus) and transverse sinuses are between the cerebral the ophthalmic veins (Shindo, 1915; Gelde- hemispheres and the vermis (central lobe) of ren, 1924). The cavernous sinus and oph- the cerebellum. Evidence for the sigmoid and thalmic veins are not indicated in Krypto- prootic sinuses, the end distributaries of the baatar, but were likely present, with the oph- transverse sinus, is found on the specimens thalmic veins entering the cavernous sinus 64 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247

Fig. 32. Reconstruction of the skull of Kryptobaatar dashzevegi in dorsal view. Abbreviations: al anterior lamina; exoc exoccipital; fr frontal; ju jugal; lac lacrimal; mx maxilla; na nasal; pa parietal; pmx premaxilla; sq squamosal; sup supraoccipital. via the sphenorbital ®ssure. The course of the (``vv'' in ®g. 37B), as occurs in both mono- inferior petrosal sinus is indicated by a shal- tremes (Hochstetter, 1896). A transverse low sulcus on the endocranial surface of the groove on the back of the dorsum sellae in basioccipital±petrosal suture in PSS-MAE PSS-MAE 123 (®g. 25) and 124 likely held 123 (®g. 25) and by a trough in the same the posterior intercavernous sinus, as in the place in the indeterminate multituberculate dog (Evans and Christensen, 1979) and var- PSS-MAE 128. Similarly situated sulci for ious marsupials (Archer, 1976); this groove the inferior petrosal sinus are found in many opens into the area of the sulcus for the in- extant mammals, including both monotremes ferior petrosal sinus along the basioccipital± (Hochstetter, 1896; Kuhn, 1971; Zeller, petrosal suture. 1989). Given the small size of the jugular foramen in Kryptobaatar, we think that the LATERAL HEAD VEIN inferior petrosal sinus left the cranial cavity In addition to the venous drainage from via the foramen magnum as vertebral veins the sigmoid and inferior petrosal sinuses ex- 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 65

Fig. 33. Reconstruction of the skull of Kryptobaatar dashzevegi in lateral view. Abbreviations: al anterior lamina; exoc exoccipital; fr frontal; lac lacrimal; man mandible; mx maxilla; na nasal; or orbitosphenoid; pa parietal; pet petrosal; pmx premaxilla; sq squamosal; sth stylohyal. iting the skull via the foramen magnum, the prootic sinus, the post-trigeminal vein is the other major egress for the dural sinuses in other major tributary of the lateral head vein Kryptobaatar was via the prootic canal. As in the echidna, draining posteriorly from the in monotremes and some marsupials (Wible, cavernous sinus (Wible and Hopson, 1995); 1990; Wible and Hopson, 1995), the prootic we have observed a small vein in a similar sinus in Kryptobaatar (``ps'' in ®g. 37B) position accompanying the ramus inferior of drained into the lateral head vein (``lhv'' in the stapedial artery in the juvenile platypus ®g. 37B) in the middle-ear cavity through the described by Zeller (1989). We think that the prootic canal, which in extant taxa forms in ramus inferior was present in Kryptobaatar the posterior aspect of the prootic foramen in (see below), and a post-trigeminal vein has the primary braincase wall (®g. 24A±C). been reconstructed accompanying that artery Judging from the size of the endocranial ap- in other multituberculates by Wible and Hop- erture of the prootic canal (®gs. 25, 27), the son (1995). However, given the absence of lateral head vein was well developed in this vein in therians, without speci®c osteo- Kryptobaatar; the tympanic aperture of the logical evidence for it, the presence of the prootic canal is partially hidden within a re- post-trigeminal vein in Kryptobaatar and cess. The intratympanic course of the lateral other multituberculates is equivocal. head vein (®g. 37B) was posteromedially, dorsal to the facial nerve. The vein then continued ventrally into the neck as the internal TRANSVERSE CANAL VEIN jugular vein. The lateral head vein probably did not leave the tympanic cavity with the A canal in front of the hypophyseal fossa facial nerve through the stylomastoid notch, connecting the right and left orbitotemporal contra Kielan-Jaworowska et al. (1986) and regions is occupied among living mammals Miao (1988), but had a separate, more medial by a large vein known as the transverse canal exit (Rougier et al., 1992; Wible and Hopson, vein (McDowell, 1958; Archer, 1976). Given 1995). Whether a post-trigeminal vein was that a similar canal is present in specimens present in Kryptobaatar is uncertain, al- of Kryptobaatar (®gs. 26, 36A) and in PSS- though we have reconstructed one here MAE 128, we restore that vessel here (``tcv'' (``ptv'' in ®gs. 36B, 37B). Along with the in ®g. 36B). Judging from the size of the 66 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247

Fig. 34. Reconstruction of the skull of Kryptobaatar dashzevegi in ventral view. Abbreviations: al anterior lamina; ali alisphenoid; bo basioccipital; bs basisphenoid; exoc exoccipital; fr frontal; mx maxilla; pal palatine; pet petrosal; pmx premaxilla; pt pterygoid; sq squamosal; vo vomer. transverse canal, the vein in Kryptobaatar bitotemporal fossa, which in one specimen was a substantial one. (PSS-MAE 123) leads into a sulcus running posteriorly to the carotid foramen (®g. 25). PITUITO-ORBITAL VEIN This arrangement resembles that which we have observed in the platypus. Consequently, A vein to the pituitary gland from the or- we restore the pituito-orbital vein in this fo- bital sinus, passing through the primary ramen in Kryptobaatar. braincase wall, is known for various sauropsids (Bruner, 1907; Bellairs and Kamal, VEINS OF ASCENDING,ORBITOTEMPORAL, AND 1981). Among mammals, monotremes have POSTTEMPORAL CANALS a similar vein (Gaupp, 1908), which in the platypus exits the hypophyseal fossa via the Veins likely also accompanied at least carotid foramen (personal obs.). Kryptobaa- some of the major arteries passing through tar has a foramen in the anterolateral wall of the lateral braincase wall in Kryptobaatar the hypophyseal fossa connecting to the or- (®g. 36B). Following a model proposed by 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 67

Fig. 35. Reconstruction of the skull of Kryptobaatar dashzevegi in posterior view. Abbreviations: bo basioccipital; exoc exoccipital; fr frontal; lac lacrimal; man mandible; pa parietal; pet petrosal; sq squamosal; sup supraoccipital.

Watson (1911, 1920), various authors (e.g., other vascular canal between the subarcuate Kermack et al., 1981) have reconstructed fossa and the posttemporal canal, which in veins as the sole occupants of the ascending, the embryos of some extant bats and ele- orbitotemporal, and posttemporal canals in phant shrews transmits a venous connection Mesozoic mammaliaforms. However, Kie- between the dural sinuses and the mastoid lan-Jaworowska et al. (1984, 1986) argued emissary system. These authors also reported effectively, on the basis of the anatomy of a minute fenestration of the subarcuate fossa extant mammals, that arteries and not veins in ``Catopsalis'' joyneri. Luo (1989) found a were the primary occupants of these canals large fenestration in Ptilodus montanus and in multituberculates and other Mesozoic syn- Mesodma, and he (1996) proposed this con- apsids (see also Wible, 1989). Following the dition as a synapomorphy of ptilodontids and model of Kielan-Jaworowska et al. (1984, eucosmodontids. In Kryptobaatar, the subar- 1986), we reconstruct in Kryptobaatar venae cuate fossa is very deep, hindering identi®- commitantes with the ramus superior, ramus cation of a subarcuate fenestration; neverthe- supraorbitalis, and arteria diploeÈtica magna less, as far as we can tell, such an opening in the ascending, orbitotemporal, and post- is lacking. temporal canals, respectively (®g. 36B). These veins drained from the orbit to the ARTERIES middle ear and occiput, and they likely were substantial, as evidenced by the preserved In extant mammals, three major paired ar- endocast of the orbitotemporal canal showing teries supply blood to the head: the internal two subequal vessels in the undescribed skull carotid, the external carotid, and the verte- of a new species of Mongolian Late Creta- bral. In most forms, the internal carotid and ceous multituberculate (PSS-MAE 126). vertebral supply the brain, whereas the external carotid along with the stapedial artery, the principal extracranial branch of the inter- SUBARCUATE FENESTRATION nal carotid, supply the meninges and super- In Nemegtbaatar and Chulsanbaatar, Kie- ®cial structures (Tandler, 1899, 1901). How- lan-Jaworowska et al. (1986) identi®ed an- ever, in the echidna, marsupials, and certain 68 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247 placentals (e.g., xenarthrans, carnivorans, un- is present in extant monotremes, marsupials, gulates), the stapedial artery is lacking and and the vast majority of placentals (Hochs- most or all of its end branches are annexed tetter, 1896; Tandler, 1899, 1901; Gillilan, to the external carotid (Bugge, 1978, 1979; 1972), with cetaceans being an exception Wible, 1984, 1987). (McFarland et al., 1979).

INTERNAL CAROTID ARTERY STAPEDIAL ARTERY In extant mammals, the internal carotid ar- During development in all but one of the tery has an extracranial course ventral to the extant mammals investigated, a well-devel- basicranium en route to the carotid foramen, oped stapedial artery arises from the extra- which opens endocranially posterolateral to cranial portion of the internal carotid and the hypophysis (pituitary). Both the position sends branches with the trigeminal nerve of the extracranial course and the foramen (Tandler, 1902; Wible, 1984, 1987); the one vary. The artery's course is either on the exception is the echidna, in which the sta- promontorium (transpromontorial), through pedial artery apparently does not form and the substance of the tympanic wall (perbul- the end branches accompanying the trigem- lar), or medial to the tympanic wall (extra- inal nerve are supplied through the external bullar); additionally, the extracranial course carotid (Hochstetter, 1896). In placentals, of the artery is not always indicated by a there are two major branches of the embry- groove or canal (Wible, 1986). The carotid onic stapedial system: the ramus superior, foramen is either within the basisphenoid or supplying a ramus supraorbitalis which ac- between the basisphenoid and petrosal (De companies the ophthalmic nerve, and the ra- Beer, 1937), and its endocranial aperture is mus inferior, supplying a ramus infraorbitalis either within the hypophyseal fossa or pos- and mandibularis in company with the max- terolateral to it. illary and mandibular nerves, respectively In Kryptobaatar (``ica'' in ®g. 37B), the (Tandler, 1902; Wible, 1984, 1987). The ra- occupant of the well-developed groove on mus superior fails to form in marsupials and the anterior half of the promontorium, the ca- is a late embryological addition in the platy- nal between the petrosal, alisphenoid, and pus that does not extend rostrally as far as pterygoid, and the endocranial foramen with- the ophthalmic nerve (Wible, 1984, 1987). in the hypophyseal fossa was the internal ca- The embryonic stapedial system is retained rotid artery, vein, and nerve. Among extant in few adult mammals, and in most, at least mammals, similar transpromontorial grooves one of the end branches of the stapedial ar- are found only in certain placentals (e.g., li- tery is annexed to a branch of the external potyphlans, aardvarks) and, with but one ex- carotid (Tandler, 1899, 1901; Bugge, 1974; ception, are occupied by the internal carotid Wible, 1984, 1987). In fact, the main stem artery along with its companion vein and of the stapedial (proximal stapedial) is re- sympathetic nerve (Wible, 1986). The excep- tained only in the adult platypus and certain tion occurs in some strepsirhine primates, in adult placentals (e.g., lipotyphlans, scanden- which the artery is lost and the nerve is the tians) (Tandler, 1899; Bugge, 1974; Wible, principal occupant of the transpromontorial 1987). The two principal sites of anastomosis groove (Conroy and Wible, 1978). However, between the external carotid and stapedial given that this is an unusual occurrence systems are the maxillary artery, which con- among extant mammals, we deem it more nects to the ramus inferior of the stapedial likely that both the internal carotid artery and below the mandibular nerve exit, and the ar- nerve were present in Kryptobaatar. Addi- teria diploeÈtica magna, which connects to the tionally, the connections and sizes of the var- ramus superior of the stapedial through the ious sulci on the promontorium in Krypto- posttemporal canal (Wible, 1987). baatar strongly support a vascular interpre- Kryptobaatar has grooves, canals, and fo- tation. The internal carotid artery was likely ramina that indicate the presence of a well- not the sole supplier of blood to the brain in developed stapedial system, including the Kryptobaatar, given that the vertebral artery proximal stapedial artery, ramus inferior, ra- 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 69

Fig. 36. Reconstruction of the skull of Kryptobaatar dashzevegi in left lateral view (A) with the major cranial arteries (white outline) and veins (black) added (B). Zygomatic arch is cut to show vessels in the ¯oor of the orbit. Vessels in shadow pass through intramural canals, with the exception of the part of the ramus superior immediately rostral to the dorsalmost ramus temporalis branch which is endocranial. Abbreviations: adm arteria diploeÈtica magna (and accompanying vein); ea ethmoidal artery (and accompanying vein); ef ethmoidal foramen; fbu foramen buccinatorium; fdac foramen of dorsal ascending canal; fdv foramen for frontal diploic vein; fma foramen masticatorium; frt foramen for ramus temporalis; ioa infraorbital artery (and accompanying vein); iof infraorbital foramen; lhv lateral head vein; mef metoptic foramen (which transmitted pituito-orbital vein); oa ophthalmic artery; oca occipital artery (and accompanying vein); ompf orbital opening of minor palatine foramen; opf optic foramen; otc orbitotemporal canal; psa proximal stapedial artery; ptc posttemporal canal; ptv post- trigeminal vein; ri ramus inferior; rio ramus infraorbitalis (and accompanying vein); rm ramus mandibularis (and accompanying vein); rs ramus superior (and accompanying vein); rso ramus supraorbitalis (and accompanying vein); rt ramus temporalis (and accompanying vein); sgf supraglenoid foramen; son supraorbital notch; spa sphenopalatine artery (and accompanying vein); spf sphenopalatine foramen; sphf sphenorbital ®ssure; tc transverse canal; tcv transverse canal vein. 70 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247 mus superior, and arteria diploeÈtica magna in question, because the large size of the (®gs. 36, 37). The proximal stapedial artery groove for the proximal stapedial on the occupied the groove on the promontorium promontorium and the small size of the fo- that runs posterolaterally from the groove for ramen for the ramus superior suggest that the the internal carotid artery toward the fenestra other major end branch of the proximal sta- vestibuli (``psa'' in ®g. 37B). In extant mam- pedial, the ramus inferior, was present. The mals, grooves on the promontorium directed evidence for the occurrence of the post-tri- at the fenestra vestibuli invariably contain geminal vein in Kryptobaatar has been dis- the proximal stapedial artery, but such cussed above. grooves are only found in some placentals The course of the ramus inferior rostral to retaining that artery in the adult (e.g., li- the petrosal bone in Kryptobaatar was pre- potyphlans; Wible, 1987). From the rim of sumably through the cavum epiptericum and the fenestra vestibuli, the proximal stapedial then the sphenorbital ®ssure (®g. 36B), be- artery in Kryptobaatar ran through what ap- cause other suitable foramina of exit in the pears to be a bicrurate stapes (Rougier et al., mesocranium are lacking. An endocranial 1996c); this is the artery's course in the pla- course for the ramus inferior (or its rostral centals that retain it, but in the platypus the continuation, the ramus infraorbitalis) is un- artery runs posterodorsal to the columnar sta- usual among extant mammals, but is known pes (Wible, 1987). for some chiropterans (Kallen, 1977; Wible Two canals lateral to the fenestra vestibuli and Davis, 2000) and dipodoid rodents (Bug- in Kryptobaatar indicate that beyond the stage, 1971). Within the orbit, the ramus infe- pes the proximal stapedial divided into the rior, now the ramus infraorbitalis (``rio'' in ramus superior and the ramus inferior (``rs'' ®g. 36B), ran forward in the sphenopalatine and ``ri'' in ®gs. 36B, 37B), as occurs in the groove along the suture between the maxilla platypus and certain placentals (Wible, and frontal. En route, it likely supplied the 1987). The canal for the ramus inferior in minor palatine artery into a foramen in the Kryptobaatar is directed anteriorly and lies orbital process of the maxilla, the major pal- dorsal to the epitympanic recess on the in- atine and sphenopalatine artery in the sphe- folded lateral ¯ange. Posteriorly, this canal is nopalatine foramen (``spa'' in ®g. 36B), and con¯uent with the canal for the facial nerve, its terminal branch, the infraorbital artery and anteriorly it presumably opens into the into the infraorbital canal (``ioa'' in ®g. 36B). cavum epiptericum. In extant mammals, sim- It also likely provided a ramus orbitalis con- ilarly situated canals are found in the echidna necting with the rostral continuation of the and certain placentals (e.g., scandentians, ramus superior (see below). macroscelidids). In the placentals, the main Isolated petrosals of Late Cretaceous mul- occupant is arterial, with the ramus inferior tituberculates, in particular ``Catopsalis'' in front and sometimes the proximal stape- joyneri, have been critical for reconstructing dial behind; the course of the facial nerve is the ramus superior in Kryptobaatar, as much very close to the arterial canal, and, in fact, of this vessel is hidden within the lateral wall in macroscelidids, the facial nerve and prox- of the braincase. In ``C.'' joyneri, en route imal stapedial share a canal (MacPhee, from the middle ear to the orbit, the ramus 1981). In the echidna, in which the ramus superior passed through a complex canal, inferior is lacking, the main occupant is ve- which for descriptive purposes is treated as nous, the post-trigeminal in front and the lat- three successive canals: the ventral ascend- eral head vein behind, with the posterior part ing, dorsal ascending, and orbitotemporal of the canal also transmitting the facial nerve (Rougier et al., 1992; Wible and Hopson, (Wible and Hopson, 1995). The only form in 1995). The ventral ascending canal begins in which the ramus inferior and post-trigeminal the middle ear, runs back posteriorly in a hor- vein co-occur is the platypus, and these izontal plane, and then turns dorsally; open- structures travel together in an open sulcus ing into it from the temporal fossa is the su- (Wible and Hopson, 1995; personal obs.). We praglenoid foramen, which transmitted a ra- think that in Kryptobaatar the ramus inferior mus temporalis to the temporalis muscle. The was the major vascular occupant of the canal dorsal ascending canal is essentially vertical, 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 71

curving forward into the anteriorly directed orbitotemporal canal. Opening from behind and marking the limits of the ventral and dorsal ascending canals is the posttemporal canal, which transmitted the arteria diploeÈtica magna from the occiput. The ventral ascending canal is largely within the anterior lamina, the dorsal ascending canal is between the petrosal and presumably the squamosal, and the posttemporal canal is entirely in the petrosal. The reconstruction of arteries into these various canals is based largely on the pattern in monotremes; between the platypus and echidna all the canals present in ``C.'' joyneri can be found (Rougier et al., 1992; Wible and Hopson, 1995). The ramus superior in Kryptobaatar exhibited essentially the same pattern as in ``Ca- topsalis'' joyneri, as reconstructed from the various surface foramina into this system of canals (®gs. 36A, 37A). The ventral ascending canal in Kryptobaatar is indicated by its tympanic aperture and by two foramina in the anterior lamina on the lateral braincase wall, which transmitted rami temporales from the canal (``rt'' in ®g. 36B). The position of these three foramina suggests that the ventral ascending canal runs posteriorly in a largely horizontal plane. Breakage in one specimen, PSS-MAE 101, provides access to the near vertical dorsal ascending canal between the petrosal and squamosal and shows a notch that was probably enclosed to complete a foramen, which transmitted a third ramus temporalis from the canal (®g. 36). The anteriorly directed orbitotemporal canal is indicated by a horizontal endocast representing its ®lling and by its anterior aperture in the orbit. The

← Fig. 37. Reconstruction of the right basicranium of Kryptobaatar dashzevegi in ventral view mpt medial pterygopalatine trough; oca occipital (A) with the major cranial arteries (white outline) artery (and accompanying vein); ocon occipital and veins (black) added (B). Abbreviations: adm condyle; on odontoid notch; pacc posterior aper- arteria diploeÈtica magna (and accompanying ture of carotid canal; pef perilymphatic foramen; vein); ali alisphenoid; aptca artery of pterygoid ppr paroccipital process; pr promontorium of pe- canal; cp crista parotica; ctpp caudal tympanic trosal; ps prootic sinus; psa proximal stapedial ar- process of petrosal; er epitympanic recess; fma tery; ptc posttemporal canal; ptca pterygoid ca- foramen masticatorium; foi foramen ovale infer- nal; ptr pterygopalatine ridge; ptv post-trigeminal ium; fv fenestra vestibuli; gica groove for internal vein; ri ramus inferior; rs ramus superior; rvnf carotid artery; gl glenoid fossa; gpsa groove for recess for vascular and nervous foramina (prootic proximal stapedial artery; ica internal carotid ar- canal, ventral ascending canal, canal for ramus in- tery; jf jugular fossa; jfo jugular foramen; lhv lat- ferior, and facial canal); th tympanohyal; ttf ten- eral head vein; lpt lateral pterygopalatine trough; sor tympani fossa; vo vomer; vv vertebral vein. 72 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247 endocast reveals that the anterior part of the few instances (e.g., some lipotyphlans, Mc- orbitotemporal canal was enclosed, but that Dowell, 1958; MacPhee, 1981; Novacek, the posterior part had no medial wall and 1986) is the artery derived from the internal therefore is an endocranial orbitotemporal carotid, running anteriorly into the canal groove. Finally, the posttemporal canal is in- from behind. Given the large size of the sul- dicated by its posterior aperture on the occi- cus running forward from the carotid canal put, entirely within the petrosal (®gs. 16, 17). visible in PSS-MAE 113 (®g. 21), it seems Judging from the size of this aperture, the ar- likely that Kryptobaatar exhibited the pattern teria diploeÈtica magna was of substantial size as some lipotyphlans (`àptca'' in ®g. 37B). (`àdm'' in ®gs. 36B, 37B). However, in light of the rare occurrence of The rostral continuation of the ramus su- this origin for the artery of the pterygoid ca- perior within the orbit is the ramus supraor- nal among extant mammals, we are less con- bitalis (``rso'' in ®g. 36B). In the extant mam- ®dent in this particular aspect of the vascular mals retaining this vessel (i.e., monotremes reconstruction. Moreover, note that the major and some placentals), it provides up to four occupant of the pterygoid canal in extant sorts of branches: the lacrimal, frontal, and mammals is the nerve (Wible, 1984). ethmoidal arteries running with the ophthalmic nerve branches of the same name, and OPHTHALMIC ARTERY the ramus orbitalis anastomosing with the ra- An ophthalmic artery derived from the in- mus infraorbitalis (Tandler, 1899, 1902; Bug- ternal carotid (circulus arteriosus) and ac- ge, 1974; Wible, 1987). Given that all four companying the optic nerve is widely dis- branches are present in both the platypus and tributed among extant mammals, including echidna and in diverse placentals (e.g., ar- the echidna, all marsupials investigated, and madillos, lemurs), it is probable that Krypto- most placentals (Tandler, 1899, 1902; Bugge, baatar exhibited the same condition. 1974; Archer, 1976; Wible, 1984). In light of this distribution, an ophthalmic artery prob- ARTERY OF THE PTERYGOID CANAL ably ran with the optic nerve in the optic An artery accompanies the nerve of the foramen in Kryptobaatar as well (`òa'' in pterygoid canal in many extant mammals, ®g. 36B). It may have anastomosed with but in most it is derived from the external branches of the ramus supraorbitalis within carotid system and runs posteriorly into the the orbit via a ramus orbitalis (not shown in canal from the orbit (Wible, 1984). In only a ®g. 36B).

COMPARISONS The following section is devoted to high- As a cautionary note, the taxonomy of the lighting particular characters or character Late Jurassic Paulchoffatiidae has been re- systems of the multituberculate skull to vised recently by Hahn (1993), who has been which our descriptions of Kryptobaatar are publishing on this group for more than a relevant or have prompted reevaluation of quarter century. Hahn (1993) has created current evidence. An ongoing project beyond several new genera for previously described the scope of this report is our reevaluation of specimens and reassigned other specimens to the phylogenetic interrelationships within previously named genera. We follow Hahn's Multituberculata. In conducting this reeval- (1993) revision in our taxonomic usages. uation, we have observed many of the relevant multituberculate specimens, and some SNOUT AND PALATE of our observations are included here. To place these taxa in a phylogenetic context, PRENASAL PROCESS OF THE PREMAXILLA we have reproduced a recent hypothesis of multituberculate interrelationships (®g. 38) The tip of the rostrum is rarely preserved and of the higher-level relationships of mul- in multituberculates. PSS-MAE 101 (®g. 6) tituberculates (®g. 39). is an exception and clearly shows that a pre- 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 73

Fig. 38. Cladogram of phylogenetic relationships within Multituberculata (a), based on the analysis in Rougier et al. (1997), with an enlargement of the tree section containing Kryptobaatar and its nearest relatives, all of which are Mongolian Late Cretaceous taxa with the exception of Pentacosmodon from the North American Paleocene (b). A similar Mongolian Late Cretaceous grouping was called Dja- dochtatheria by Kielan-Jaworowska and Hurum (1997). In their phylogeny, Kryptobaatar is in a clade, Djadochtatheriidae, with Djadochtatherium, Catopsbaatar, and Tombaatar. nasal process of the premaxilla (internarial ton and Luo, 1993), but is lacking in extant bar) is lacking in Kryptobaatar. A prenasal mammals and the prototribosphenidan Vin- process dividing the external nares is found celestes (Rougier, 1993). in the outgroups to Mammalia, such as Hal- To date, an internarial bar has been recon- danodon (Lillegraven and Krusat, 1991), structed for only one multituberculate, the Morganucodon (Miao, 1988; Wible, 1991) Paleocene taeniolabidid Lambdopsalis. Miao and Sinoconodon (Wible et al., 1990; Cromp- (1988) based this on the presence of a wedge 74 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247 of bone between the anterior ends of the na- 1930; Hill and De Beer, 1949; Zeller, 1989). sals, which he interpreted to be the broken This element does not come into contact with dorsal tip of the prenasal process. Miao the nasals. In the echidna, the os carunculae (1988) speculated that the prenasal process forms independent of, but in contact with, the in Lambdopsalis either was an evolutionary prenasal process of the premaxilla (Gaupp, reversal or that the process was more wide- 1908; Kuhn, 1971). Among marsupials, an spread in multituberculates than indicated in independent ossi®cation in the cleft between previously described specimens. In support the anterior cupulae of the nasal capsule, also of the latter, Miao (1988: 9) suggested that called the os carunculae, has been reported the left nasal in stereophotographs of Chul- in embryos of the didelphid Caluromys phi- sanbaatar (ZPAL MgM-I/89) published in lander (Denison and Terry, 1921). In addi- Kielan-Jaworowska et al. (1986: ®g. 16) ap- tion, condensed connective tissue in the same peared ``indicative of a sutural facet at the region has been reported and homologized anterior end,'' presumably for the prenasal with the prenasal process of the premaxilla process. However, we interpret the left nasal in Trichosurus vulpecula (Broom, 1909), Di- in this specimen as broken, in light of other delphis marsupialis (Toeplitz, 1920), Didel- specimens of Chulsanbaatar (e.g., ZPAL phis aurita, and Perameles nasuta (Hill and MgM-I/139) that show a more complete na- De Beer, 1949). Remnants of an egg-tooth sal with no sutural facets at the anterior end papilla beneath the premaxilla have also been (see Kielan-Jaworowska, 1974: pl. XI). discovered in Trichosurus and Phascolarctos Therefore, we ®nd no evidence of an inter- cinereus (De Beer, 1937; Hill and De Beer, narial bar in Chulsanbaatar or other multi- 1949). Finally, in placentals, remnants of the tuberculates described to date. The presence os carunculae, egg-tooth, and internarial bar of an internarial bar in Lambdopsalis is con- are unknown (De Beer, 1937). In light of the troversial (see below), and the absence of this distribution of the characters associated with structure primitively in multituberculates is the os carunculae, egg-tooth, and internarial corroborated by our observations of snouts bar, as well as the phylogenetic position of of Late Jurassic paulchoffatiids, including multituberculates (®g. 39; Rowe, 1988, 1993; two species of Pseudobolodon (V.J. 111-155, Rougier et al., 1996a, 1996c; Hu et al., 1997; V.J. 450-155) and one of Kuehneodon (V.J. Ji et al., 1999), it is likely that an os carun- 112-155). culae and egg-tooth, or remnants thereof, The element in question in Lambdopsalis were at least temporarily present in multitu- is illustrated in two rostra (Miao, 1988; ®gs. berculates at some point in ontogeny. The 1, 2), but it exhibits differences in size and implications of these features for assessing position between the two, as well as between whether multituberculates were oviparous or the right and left sides of the same individ- viviparous are inconclusive, because differ- ual. Given these differences, the purported ent conditions for the os carunculae, egg- prenasal process of the premaxilla may be a tooth, and internarial bar are present in the broken part of the nasal. Despite our doubt extant phylogenetic bracket for multituber- about the element in Lambdopsalis,itis culates (i.e., monotremes and therians). largely irrelevant as a systematic feature, be- Based on the reconstructed small size of the cause it is not shared with any other multi- pelvic outlet in Kryptobaatar (ZPAL MgM- tuberculate and was absent primitively in the I/41) relative to that in monotremes, Kielan- group. Jaworowska (1979) proposed that multitu- Although the internarial bar is lacking in berculates were viviparous with extremely adult extant mammals, remnants of it occur small neonates. Lillegraven et al. (1987) ar- during development in monotremes and mar- gued that the same condition was found in supials. In the platypus, it forms prenatally the common ancestor of marsupials and pla- as a dorsal outgrowth from the rostral part of centals. the premaxilla (os carunculae of Wilson, 1900), which supports the egg-tooth; it per- SEPTOMAXILLA sists in young individuals after the egg-tooth In multituberculates, the premaxilla has a is shed and is subsequently resorbed (Green, large dorsal process that forms the lateral 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 75

Fig. 39. Cladogram of phylogenetic relationships within Mammaliaformes based on analyses by Rougier et al. (1996a) and Hu et al. (1997). margin of the external nares (®g. 33), as in 1997; Ji et al., 1999), the presence of a sep- most therians. In contrast, in monotremes tomaxilla in the group would be expected. In and most groups of Mesozoic mammals, this fact, such a bone has been recently described position is occupied by a separate bone, the for the Late Jurassic paulchoffatiid Pseudo- septomaxilla. Based on recent phylogenetic bolodon krebsi (Hahn and Hahn, 1994). Its hypotheses of multituberculate relationships purported septomaxilla is cuneiform and lies (e.g., Rougier et al., 1996a, 1996c; Hu et al., between the dorsal process of the premaxilla, 76 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247 the nasal, and the maxilla; that is, it is ex- tained, at least in Kryptobaatar (``ano'' in cluded entirely from the external nares. This ®gs. 8, 12), Kamptobaatar (Kielan-Jawo- would be a unique position for the septo- rowska, 1969: pl. XVI), and Catopsbaatar maxilla among synapsids, which otherwise (Kielan-Jaworowska, 1974: pl. XVIII). have the septomaxilla in the margin of the external nares (Wible et al., 1990). Our ob- NASAL FORAMINA servations of the specimen of Pseudobolodon krebsi (V.J. 451-155) described by Hahn and Once considered to be highly unusual Hahn (1994) suggest that the bone in ques- among mammals (Kielan-Jaworowska, 1971; tion is likely part of the premaxilla and that Miao, 1988), nasal foramina are now known what is illustrated by these authors (1994: ta- to be distributed widely among Mammalia- bles 1, 2) as a suture between the purported formes, including Sinoconodon (Crompton septomaxilla and premaxilla is merely a and Luo, 1993), Haldanodon (Lillegraven crack within the premaxilla. Therefore, pend- and Krusat, 1991), Vincelestes (Rougier, ing new information, we consider the septo- 1993), and Ornithorhynchus (Zeller, 1989). maxilla to be absent in multituberculates. It Nasal foramina have also been described is also absent in marsupials and placentals widely among Late Cretaceous and Cenozoic (Wible et al., 1990), although the homologies multituberculates. However, until recently of the purported septomaxilla of some xe- they had not been reported for any Late Ju- narthrans remain controversial (Zeller et al., rassic paulchoffatiid. In 1994, Hahn and 1993). Hahn reported a single pair of small nasal foramina in Pseudobolodon krebsi (V.J. 451- NASAL OVERHANG AND ANTERIOR NASAL 155), and we have seen the same condition NOTCH in another species of the same genus, P. oreas (V.J. 461-155). However, we report In primitive mammaliaforms (e.g., Hal- here that nasal foramina are not ubiquitous danodon, Lillegraven and Krusat, 1991; Si- among paulchoffatiids, as they are wholly noconodon, Crompton and Luo, 1993), the lacking in Kuehneodon dryas (V.J. 454-155). anterior end of the nasal extends rostrally be- Another feature of the nasal foramina that yond the septomaxilla and the posterodorsal varies considerably among multituberculates process of the premaxilla, dorsal to the ex- is the number of pairs present (Kielan-Ja- ternal narial aperture. We call this condition worowska and Hurum, 1997), from a single the nasal overhang of the nares. Additionally, pair in Pseudobolodon to seven or eight pairs at least in Haldanodon, there is a deep an- in Lambdopsalis (Miao, 1988). Kielan-Ja- terior nasal notch in the rostral margin that, worowska and Hurum (1997) reported two according to Lillegraven and Krusat (1991), pairs of nasal foramina in Kryptobaatar and may have transmitted ethmoidal vessels. in Chulsanbaatar, Catopsbaatar, Tombaa- Among multituberculates, a nasal over- tar, and the ptilodontoid Ptilodus. However, hang of the nares is illustrated for the Late we found only one pair in Catopsbaatar Jurassic paulchoffatiid Pseudobolodon krebsi (ZPAL MgM-I/78), and in Ptilodus,we (Hahn and Hahn, 1994: ®g. 2b), and we ob- found three pairs in P. gracilis (USNM served it in Kuehneodon simpsoni (V.J. 112- 6076) and two pairs, asymmetrically placed 155) and Henkelodon nais (V.J. 401-155); on the right and left sides, in P. montanus these three taxa also have an anterior nasal (AMNH 35490). Finally, in Kryptobaatar notch. In contrast, a nasal overhang is absent PSS-MAE 101, we found two foramina on or greatly reduced in all younger taxa pre- the right and three on the left (``naf'' in ®gs. serving the anterior tip of the nasal: Krypto- 6, 8). baatar (®g. 33), Kamptobaatar (Kielan-Ja- Foramina in the nasals were ®rst described worowska, 1971: pl. III), Nemegtbaatar in multituberculates by Simpson (1925) in (Kielan-Jaworowska, 1974: pl. IX), Chulsan- Djadochtatherium matthewi. However, it was baatar (ibid.: pl. XII), Catopsbaatar (ibid.: not until 1937 that Simpson applied the term pl. XIX), and Lambdopsalis (Miao, 1988: ®g. nasal foramina to these apertures in his de- 4). However, the anterior nasal notch is re- scription of Ptilodus. Miao (1988) cited 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 77

Simpson (1937) as employing two different (1997a), the discrepancies noted by Miao terms for these foramina: either vascular fo- (1988) have no bearing when considered in ramina or nasal foramina. From our reading a wider phylogenetic framework, because the of Simpson (1937), we think that the term he variations occur in only a few very disparate intended for these apertures was nasal foram- groups. We add here that after review of the ina, and that vascular foramina was meant as major collections of Mongolian Late Creta- a descriptor of their function and not an al- ceous multituberculates (MAE and ZPAL), ternative term. Rather than vascular, a ner- we have found no instances of more than a vous function was suggested for the nasal fo- single infraorbital foramen. ramina by Kielan-Jaworowska (1971). Another feature of the infraorbital system preserved in Kryptobaatar is several small INFRAORBITAL FORAMINA foramina linking the infraorbital canal with either the maxillary sinus or nasal cavity. Some Jurassic multituberculates have two Rougier et al. (1997a) reported similar ap- infraorbital foramina on the facial process of ertures in another form from Ukhaa Tolgod, the maxilla, which vary in size from sube- Tombaatar, and speculated that they may qual to the anterior one being considerably lead into alveolar canals and have carried the larger (Hahn, 1985). The latter condition nerves and vessels to the upper teeth. Con- is also found in Early Cretaceous Arginbaa- ®rmation of the presence of alveolar canals tar from Khoobur in Mongolia (Kielan-Ja- for Kryptobaatar is provided by high-reso- worowska et al., 1987). The smaller posterior lution CT scans of PSS-MAE 101. foramen is absent in Kryptobaatar (®g. 33) and other Mongolian Late Cretaceous mul- PALATINE FORAMINA tituberculates, and in Cimolodonta (sensu Simmons, 1993; i.e., ptilodontoids and tae- There is considerable confusion about the niolabidoids), except for Meniscoessus ro- foramina in and around the palatine's contri- bustus (Archibald, 1982; Simmons, 1993). bution to the palate because different terms Multiple facial exits for the infraorbital ca- have been used by various authors. As our nal system are present in basal mammalia- point of reference, we have used the dog morphs (Kermack et al., 1981; Lillegraven (Evans and Christensen, 1979), which has and Krusat, 1991; Crompton and Luo, 1993), three sets of palatine foramina: a large an- monotremes (Kuhn, 1971; Zeller, 1989), and terior one transmitting the major palatine in members of the therian lineage, including nerve and vessels, a small one behind it paurodontids (personal obs. of an undescrib- transmitting the accessory palatine nerve and ed new genus from the Morrison Formation) artery, and at the back of the palate a notch and Vincelestes (Rougier, 1993). A foramen or large foramen for the minor palatine nerve for the lacrimal branch of the infraorbital and vessels (see ®gs. 15, 34). Additionally, system is present in most forms with multiple we have named these foramina based on their exits; it opens between the maxilla, the lac- occupants (contra Evans and Christensen, rimal, and sometimes the jugal. Some Juras- 1979): that is, major, accessory, and minor sic multituberculates and monotremes have palatine foramina, respectively. Similarly multiple exits, but lack the foramen for the placed foramina are widely distributed lacrimal branch. among mammaliaforms, but all three do not Miao (1988) questioned the utility of char- necessarily occur together. For example, acters of the infraorbital foramina in phylo- among primitive mammaliaforms, Sinoco- genetic analyses, because there are instances nodon (Crompton and Luo, 1993) has all of individual variability. For example, on one three, but Morganucodon (Kermack et al., side of a maxilla of Lambdopsalis of the 13 1981) and Haldanodon (Lillegraven and maxillae reported by Miao (1988), two rather Krusat, 1991) have only the major and minor than a single infraorbital foramina were pre- palatine foramina. sent, though the smaller posterior foramen Among multituberculates, paulchoffatiids was distinguished by only a tiny bridge. appear to have only the major and minor fo- However, as pointed out by Rougier et al. ramina (Hahn, 1987). Most of the Mongolian 78 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247

Late Cretaceous taxa with well-preserved multituberculates. We offer the following palates (Kryptobaatar; Kamptobaatar, Kie- amendments to Kielan-Jaworowska and Hu- lan-Jaworowska, 1971; Chulsanbaatar, Ne- rum's scoring. megtbaatar, Kielan-Jaworowska et al., 1986) Among Djadochtatheria, whereas the typ- have all three (®g. 15), although the number ical djadochtatherian facial process is said to of accessory foramina varies. Another vari- be present in Sloanbaatar, Catopsbaatar, ant is presented by Ptilodus montanus, which and Bulganbaatar (Kielan-Jaworwoska and seems to have only the major and accessory Hurum, 1997), we were not able to unequiv- palatine foramina (Simpson, 1937; AMNH ocally identify a lacrimal in the specimens of 34590). This is the same condition reported these taxa housed in Warsaw. According to for Lambdopsalis, although all three fora- Kielan-Jaworowska (personal commun.), the mina are visible in stereophotographs and re- typical djadochtatherian facial process is pre- constructions in Miao (1988: ®gs. 3, 14, 18). sent on the Catopsbaatar specimen housed According to Kielan-Jaworowska et al. in Moscow (PIN 4537-5). (1986), the opening we call the minor pala- Among Taeniolabididae, Kielan-Jawo- tine foramen does not penetrate bone fully in rowska and Hurum (1997) reported a small, Nemegtbaatar as shown from their study of arcuate facial process in Lambdopsalis and a sectioned specimen. To distinguish this Taeniolabis. However, the lacrimal is wholly opening from a true foramen, Kielan-Jawo- lacking in Lambdopsalis (Miao, 1988), and rowska et al. (1986) employed the term pa- if one was present in Taeniolabis, it had no latonasal notch. Additionally, they general- facial exposure (Broom, 1914). ized from their observations of Nemegtbaa- Among Ptilodontoidea, for the observation tar, suggesting that the purported foramen of a small facial process of the lacrimal in Ptilodus, Kielan-Jaworowska and Hurum present in other multituberculates is also (1997) cited Simpson (1937). Although Gid- merely a palatonasal notch. We cannot say ley (1909) originally identi®ed a lacrimal in unequivocally that the minor palatine fora- P. gracilis, Simpson (1937: 740), in reana- men in Kryptobaatar penetrates bone, be- lyzing the same specimen and other speci- cause full preparation of this passageway is mens of P. montanus, stated that ``the ante- essentially impossible. However, Kryptobaa- rior orbital rim is not perfectly preserved in tar has a foramen in the orbital process of any case and certainty is impossible, but the maxilla (®g. 12) that closely resembles there is no suggestion of facial exposure of the one leading into the minor palatine canal the lacrimal and the maxilla probably forms in various therians (Novacek, 1986). Conse- this rim.'' In our observations of the speci- quently, we are fairly con®dent that the mi- mens of P. montanus and P. gracilis de- nor palatine foramen on the palate transmit- scribed by Simpson (1937), we were unable ted the same named nerves and vessels in to either deny or con®rm the presence of a Kryptobaatar. We also think that a true mi- lacrimal. This is also the case in the skull of nor palatine foramen is more widespread the ptilodontoid Ectypodus tardus (YPM-PU among multituberculates, but justi®cation 14724), which was originally described by awaits better preserved orbital processes of Sloan (1979) as having no lacrimal. the maxilla. Among Paulchoffatiidae, a small, arcuate facial process contacting the posteriorly ex- FACIAL PROCESS OF LACRIMAL panded nasals has been reported for Pseu- dobolodon krebsi (Hahn, 1969; Hahn and In their phylogenetic analysis of selected Hahn, 1994), but we have also seen it in P. multituberculate taxa, Kielan-Jaworowska oreas (V.J. 460-155), Kuehneodon dryas and Hurum (1997) reported a large, roughly (V.J. 454-155), and Meketichoffatia krausei rectangular facial process of the lacrimal sep- (V.J. 446-155). arating the frontal and the maxilla as characteristic of Djadochtatheria, the monophy- ORBITOTEMPORAL REGION letic clade that includes 10 Mongolian Late ORBITAL MOSAIC Cretaceous genera. This is in contrast to the The sutural pattern of the bony elements small, arcuate facial process of some other contributing to the orbital mosaic has been 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 79 reported for fewer than a handful of multi- The primitive condition of the palatine for tuberculates. The two with the most evident multituberculates is still controversial, but sutures are Kryptobaatar (this report; ®gs. Miao's (1988, 1993) suggestion that the ab- 18, 19, 33) and Lambdopsalis (Miao, 1988). sence of an orbital process may be primitive In both taxa, the palatine is wholly excluded is still viable. In addition to Kryptobaatar, from the orbit, and the major contributing el- an orbital process of the palatine is lacking ements are the frontal and the maxilla, with in Catopsbaatar (ZPAL MgM-I/78) and the latter forming the anterior part of the or- Kamptobaatar (ZPAL MgM-I/33) (contra bital roof and extending posteriorly through Kielan-Jaworowska, 1971, and Hurum, the orbit to ¯oor the sphenorbital recess. 1994), but it is said to be present in a sec- However, the proportions of several elements tioned skull of Nemegtbaatar (Hurum, 1994, are strikingly different between the two taxa. 1998a) and is illustrated in Ectypodus (Sloan, In Lambdopsalis, there is no lacrimal; the or- 1979: ®g. 1). We have not had the opportu- bitosphenoid has a restricted orbital expo- nity to study the specimen of Nemegtbaatar sure; and the alisphenoid is expanded dor- described by Hurum (1994, 1998a); however, sally, contacting the frontal and parietal and we were not able to identify sutures distin- excluding the anterior lamina from the orbit- guishing any orbital elements in Ectypodus al mosaic. In contrast, in Kryptobaatar (®gs. tardus (YPM-PU 14724). If Hurum's (1994, 18, 19, 33), there is a lacrimal with a small 1998a) interpretations of the orbital sutures orbital process, the orbitosphenoid has a in Nemegtbaatar are correct, then the Mon- more extensive orbital exposure; the anterior golian Late Cretaceous taxa would be poly- lamina contacts the frontal, and the alisphe- morphic regarding the presence of the pala- noid is con®ned to the sphenorbital recess. tine in the orbit. Because an orbital process How widespread these patterns are among has been reported in Nemegtbaatar and Ecty- multituberculates is not entirely clear be- podus, we carefully scrutinized the condition cause of preservational problems. A lacrimal in Kryptobaatar. On the left side of PSS- was primitively present in multituberculates, MAE 101 (®g. 12) and possibly also 113, given its distribution reported above in paul- there is a line along the ventral edge of the choffatiids and Mongolian Late Cretaceous part of the orbit bulging over the frontal lobe taxa, but the extent of the orbital contribution that we interpret as a fracture within the or- is uncertain. A maxilla resembling that in bitosphenoid. The bone anteroventral to this Kryptobaatar and Lambdopsalis is also re- line is the only orbital element that conceiv- ported in Chulsanbaatar and Nemegtbaatar ably could be palatine, but the high-resolu- (Hurum, 1994, 1998a), but the maxilla does tion CT sections of PSS-MAE 101 do not not form a bony roof over the anterior orbital support this interpretation. Interestingly, this space in the paulchoffatiids Pseudobolodon is not the position the palatine is reported to (V.J. 447-155, 451-155, 460-155) and Me- exhibit in Nemegtbaatar, which is in the an- ketichoffatia (V.J. 446-155) or in the ptilo- teroventral corner of the orbit (Hurum, dontoids Ectypodus (Sloan, 1979) and Pti- 1998a). In Kryptobaatar, the corresponding lodus (Simpson, 1937). The orbital exposure part of the orbit is most assuredly maxilla. of the orbitosphenoid in Pseudobolodon One feature of the orbit that clearly varies (originally described as Kuehneodon in within multituberculates is the orbital pocket Hahn, 1977), Kamptobaatar (Kielan-Jawo- of Gambaryan and Kielan-Jaworowska rowska, 1971), Chulsanbaatar, and Nemegt- (1995), the depression in the orbital process baatar (Hurum, 1994, 1998a) appears to be of the frontal for the pars anterior of the me- more substantial than that in Kryptobaatar. dial masseter muscle. Among Mongolian Finally, as in Kryptobaatar, the alisphenoid Late Cretaceous taxa, an orbital pocket is re- is con®ned to the sphenorbital recess in ported for Kamptobaatar (Kielan-Jaworows- Kamptobaatar and Nemegtbaatar (see be- ka, 1971), Catopsbaatar (Gambaryan and low), and the anterior lamina sends a process Kielan-Jaworowska, 1995), Tombaatar forward that contacts the frontal high in the (Rougier et al., 1997a), Nemegtbaatar, Chul- orbit in Nemegtbaatar and Chulsanbaatar sanbaatar, Sloanbaatar (Kielan-Jaworowska (Hurum, 1998a). et al., 1986), and Kryptobaatar (®gs. 10±14). 80 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247

Among ptilodontoids, an orbital pocket is re- however, we found a lacrimal foramen in the ported for Ectypodus by Sloan (1979), but it orbital process of the isolated maxillae of P. has not been noted in Ptilodus (Simpson, montanus (USNM 9735). 1937; Wall and Krause, 1992). Among tae- Sphenopalatine Foramen: A sphenopala- niolabidids, Sloan (1981) noted that the large tine foramen in the anteroventral portion of orbital pocket in Taeniolabis taoensis had the orbital wall is widely distributed among been misidenti®ed as the orbit proper by multituberculates. What may differ among Broom (1914) and Granger and Simpson taxa are the enclosing bony elements. Ac- (1929). In Lambdopsalis, an orbital pocket is cording to Kielan-Jaworowska et al. (1986), not apparent in the reconstructions in Miao the foramen is at the junction of the maxilla, (1988), but a reduced one was said to be pre- frontal, orbitosphenoid, and palatine in tae- sent in a personal communication by that au- niolabidoids, based presumably on Kampto- thor to Gambaryan and Kielan-Jaworowska baatar (Kielan-Jaworowska, 1971) and Ne- (1995: 65) dated December 1994. Finally, an megtbaatar (see also Hurum, 1994), whereas orbital pocket is absent in the paulchoffatiids it lies in or near the frontomaxillary suture Meketichoffatia (V.J. 443-155) and Pseudo- in Lambdopsalis (Miao, 1988) and Krypto- bolodon (V.J. 447-155, 451-155, 460-155). baatar (``spf'' in ®gs. 10, 12, 18, 19), the two forms whose orbital mosaics are known ORBITAL FORAMINA best. A true characterization of the borders of the sphenopalatine foramen among mul- Given that few orbits are known in detail tituberculates awaits better preserved orbits. for multituberculates, it is not surprising that Foramen for Frontal Diploic Vein: An ap- information concerning orbital foramina is erture in the frontal near the dorsal orbital also limited. The reported distributions of rim resembling that described here in various orbital foramina in multituberculates Kryptobaatar for the frontal diploic vein are treated below. (``fdv'' in ®gs. 18, 36A) is also found in Ca- Lacrimal Foramen: A lacrimal foramen topsbaatar (ZPAL MgM-I/78), Chulsanbaa- has been noted thus far for only the paul- tar (ZPAL MgM-I/168), Ptilodus (AMNH choffatiids Meketichoffatia (originally de- 35490, USNM 6076), and the paulchoffatiids scribed as Paulchoffatia by Hahn, 1969) and Kuehneodon (V.J. 454-155), Meketichoffatia Pseudobolodon (Hahn and Hahn, 1994), (V.J. 446-155), and Pseudobolodon (V.J. Lambdopsalis (1988), Nemegtbaatar (Hu- 447-155, 451-155, 458-155, 460-155). Miao rum, 1994), and Kryptobaatar (``lacf'' in ®g. (1988) did not report one in Lambdopsalis, 16). In the mammaliaforms Morganucodon and we did not see any in the specimens of (Kermack et al., 1981), Haldanodon (Lille- Kamptobaatar and Nemegtbaatar available graven and Krusat, 1991), and Vincelestes for study. Among primitive mammaliaforms, (Rougier, 1993), there are two lacrimal fo- there are two foramina in the frontal in Mor- ramina in the orbital process of the lacrimal, ganucodon (Kermack et al., 1981) and one although Sinoconodon has only one (Cromp- in Haldanodon (Lillegraven and Krusat, ton and Luo, 1993). The paulchoffatiids have 1991) that may have transmitted the frontal a single opening in the ventral part of the diploic vein. orbital process (Hahn and Hahn, 1994), Anterior Opening of Orbitotemporal Ca- whereas in Nemegtbaatar (Hurum, 1994) and nal: Kamptobaatar (ZPAL MgM-I/33) and Kryptobaatar the foramen is in the ventral Nemegtbaatar (Kielan-Jaworowska et al., part of the lacrimal's contribution to the or- 1986) have a foramen below the postorbital bital rim. Lambdopsalis, in which the lacri- process at the anterior end of the orbitotem- mal is absent, has a single lacrimal foramen poral canal, which transmitted the orbitotem- in the orbital process of the maxilla (Miao, poral vessels, resembling that reported here 1988). The lacrimal is also lacking in mono- in Kryptobaatar. According to Kielan-Ja- tremes, and their lacrimal foramen is be- worowska et al. (1986), this aperture (their tween the maxilla and frontal (Kuhn, 1971; postorbital foramen) is in the parietal in Ne- Zeller, 1989). As reported above, the pres- megtbaatar; it is between the parietal, fron- ence of the lacrimal is uncertain in Ptilodus; tal, and anterior lamina in Kryptobaatar 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 81

(`òtc'' in ®g. 12). Lambdopsalis (Miao, 1986) connects to the upper part of the ca- 1988) also has an opening at the anterior end vum epiptericum rather than to the orbit. Be- of the orbitotemporal canal, but it lies well yond excluding these openings as optic fo- posterior to the reduced postorbital process, ramina, we are uncertain of their function. between the parietal, frontal, and alisphe- Among cynodonts, an optic foramen occurs noid. Hahn and Hahn (1994) reported a post- in Probainognathus (MCZ 4274), Sinoco- orbital foramen in Pseudobolodon krebsi, but nodon (Crompton and Luo, 1993), Adeloba- the relevant specimen (V.J. 451-155) shows sileus (Lucas and Luo, 1993), and eutherians no communication between this aperture and (Novacek et al., 1997), and was likely more an orbitotemporal canal. If this is a vascular widespread among multituberculates and foramen, it served a nutritive function only. mammaliaforms than is currently known. The only basal mammaliamorph in which an Metoptic Foramen: Among mammalia- anterior opening of the orbitotemporal canal forms, a metoptic foramen has thus far been has been reported is Adelobasileus, in which identi®ed in only ®ve taxa, all of them mul- the aperture is positioned in the parietal well tituberculates. In Kryptobaatar, the metoptic posterior to the orbit (Lucas and Luo, 1993). foramen is anteromedial to the sphenorbital In addition, although the foramen has not ®ssure (``mef'' in ®g. 36A), and on the en- been described, orbitotemporal vessels enter- docranial surface is anterolateral to the hy- ing the orbit have been restored for Morga- pophyseal fossa, with a deep sulcus on the nucodon (Kermack et al., 1981; Rougier et ossi®ed pila antotica running into the fora- al., 1992). men from behind (``mef'' in ®gs. 25±27). In Ethmoidal Foramen: Simmons (1993) em- Lambdopsalis (Miao, 1988), the metoptic fo- ployed the presence/absence of an ethmoidal ramen is also anterolateral to the hypophy- foramen in her phylogenetic analysis of Mul- seal fossa, but is posterior to the sphenorbital tituberculata, noting its absence in only two ®ssure. The metoptic foramen in Kampto- taxa, the ptilodontoids Ptilodus and Ectypo- baatar (ZPAL MgM-I/33) is only known dus. This result is surprising, given that an from the orbital surface; as in Kryptobaatar, ethmoidal foramen to our knowledge is uni- it is anteromedial to the sphenorbital ®ssure. versally present among living mammals, rep- Finally, the metoptic foramen in Nemegtbaa- resenting the only pathway for branches of tar and Chulsanbaatar has only been de- the ophthalmic nerve to reach the nasal cav- scribed from the endocranial surface, and it ity from the orbit. However, our inspection lies posterolateral to the hypophyseal fossa of the relevant specimens reveals an eth- (Hurum, 1998a). Miao (1988) also suggested moidal foramen to be present in both P. mon- that tiny pores in the anterior ¯oor of the tanus (AMNH 35490; see also Simpson, hypophyseal fossa reported in Meketichoffa- 1937) and E. tardus (YPM-PU 14724). tia (originally described as Pseudobolodon Therefore, an ethmoidal foramen is now by Hahn, 1981) may represent metoptic fo- known for all multituberculates preserving ramina. In our study of this specimen (V.J. the relevant portion of the orbit, including 443-155), we think these pores are artifac- Kryptobaatar (`èf'' in ®gs. 12, 18, 19, 36A). tual. A metoptic foramen is a primitive am- Optic Foramen: Lambdopsalis (Miao, niote feature that was likely more widespread 1988), Kamptobaatar (ZPAL MgM-I/33), among multituberculates and other mamma- and Kryptobaatar (`òpf'' in ®gs. 10, 12, liaforms than is currently known, although 36A) have a foramen in the orbitosphenoid the ossi®cation of the pila metoptica or lack that, given its position, likely transmitted the thereof may be a variable feature. optic nerve. An optic foramen has been re- Sphenorbital Fissure: The opening of the ported for two other multituberculates, but cavum epiptericum into the orbit, the sphe- these do not resemble the openings in the norbital ®ssure, reported for Lambdopsalis three taxa named above. The supposed optic (Miao, 1988), Ptilodus (Simpson, 1937), foramen in Meketichoffatia (originally de- Ectypodus (Sloan, 1979), and other Mongo- scribed as Pseudobolodon by Hahn, 1981) is lian Late Cretaceous multituberculates (Kie- too small and too far forward, whereas that lan-Jaworowska et al., 1986), resembles that in Nemegtbaatar (Kielan-Jaworowska et al., described here for Kryptobaatar (``sphf'' in 82 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247

®g. 36A). It is not directly visible in lateral short, and on the parietal and long. An in- view, but is concealed behind the alisphenoid conspicuous postorbital process on the fron- and/or anterior lamina (®gs. 10, 12, 36A). tal is found in paulchoffatiids (Meketichof- One potential difference is the occupants of fatia krausei, V.J. 446-155; Pseudobolodon the sphenorbital ®ssure. As mentioned above, oreas, V.J. 460-155; Pseudobolodon krebsi, Lambdopsalis (Miao, 1988) and Kryptobaa- V.J. 451-155), Ptilodus (Simpson, 1937), and tar have separate optic and metoptic fora- Ectypodus tardus (YPM-PU 14724). In con- mina; if these apertures are truly lacking in trast, in those Mongolian Late Cretaceous other multituberculates, the optic and oculo- taxa preserving the postorbital process, it is motor nerves and the ophthalmic artery on the parietal. It is short in Chulsanbaatar, probably left the cranial cavity through the Kamptobaatar, and Nemegtbaatar (Kielan- cavum epiptericum via the sphenorbital ®s- Jaworowska and Hurum, 1997) and is long sure. in Catopsbaatar (Kielan-Jaworowska and Miao (1988) noted that the multitubercu- Sloan, 1979) and Kryptobaatar. Kielan-Ja- late sphenorbital ®ssure is unusual in that it worowska and Hurum (1997) characterized is concealed in lateral view. This differs from the postorbital process of Kryptobaatar as the condition in basal mammaliaforms, such short, but PSS-MAE 113 had long ones prior as Sinoconodon (Crompton and Luo, 1993) to preparation damage (see reconstruction in and Haldanodon (Lillegraven and Krusat, ®g. 33). The only non-Mongolian multitu- 1991), and in modern mammals in which the berculate with a postorbital process on the foramen is generally visible from the side. parietal, and a short one at that, is Taenio- Miao (1988: 54) speculated that ``this is due labis taoensis from the Paleocene of North to the retention of a medially bony walled America (Broom, 1914). We are uncertain of cavum epiptericum in multituberculates, the condition in Lambdopsalis, which has among early mammals, which prevents the been said to have a short but distinct post- orbitosphenoid from articulating with the ali- orbital process on the frontal (Miao, 1988) sphenoid in most parts.'' As an alternative and on the parietal (Meng, personal commun. explanation, this concealment may result cited in Gambaryan and Kielan-Jaworowska, from the apparent foreshortening of the tem- 1995). poral region in the multituberculates for Miao (1988) suggested that a true postor- which the position of the sphenorbital ®ssure bital process is lacking in Mongolian Late is known. Cretaceous multituberculates and that the process on the parietal in these forms does POSTORBITAL PROCESS not delimit the back of the orbit. In support, he noted that the postorbital constriction is The postorbital process on the frontal bone on the posterior side of the parietal process, is a widespread feature of the mammalian whereas it is on the anterior side of a true skull (Novacek, 1986). Presumably, as in the postorbital process on the frontal. Based on dog (Evans and Christensen, 1979), it serves Kryptobaatar, however, we think that the el- as an attachment for the orbital ligament, de- ement on the parietal represents a true post- limiting the back of the orbit, and marks the orbital process that has shifted onto a differ- anterior limit of the temporalis muscle at- ent skull roof bone. In the dog (Evans and tachment. Christensen, 1979) and the platypus (Zeller, In multituberculates, the postorbital pro- 1989; personal obs.), anterior to the postor- cess has been said by Miao (1988, 1993) to bital process, are the foramen for the frontal be reduced or absent. Gambaryan and Kie- diploic vein and the passageway by which lan-Jaworowska (1995) pointed out, contra the supraorbital nerve and vessels leave the Miao (1988, 1993), that a parietal postorbital orbit. The same structures are found imme- process is characteristic of taeniolabidoids. In diately in front of the process in question on our preliminary comparisons among multi- the parietal in Kryptobaatar (®g. 18). Addi- tuberculates, we have identi®ed three condi- tionally, Miao (1988) has misidenti®ed the tions for the postorbital process: on the fron- position of the postorbital constriction in the tal and inconspicuous, on the parietal and Mongolian Late Cretaceous taxa; it does lie 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 83 posterior to the process on the parietal, in the ported displaced, broken jugals in Ptilodus very foreshortened temporal fossa (®g. 32). montanus, which contacted the medial sur- Incidentally, the postorbital process is not in- face of both the maxilla and the squamosal variably on the frontal in extant mammals; in the zygomatic arch and extended forward in hyracoids, for example, both the frontal into the anteroventral part of the orbit. They and the parietal contribute to the formation also reported an incomplete jugal in Ne- of this process (Fischer, 1986: ®g. 27; megtbaatar, which contacted both the max- MacPhee, 1994: ®g. 6). illa and the squamosal, as well as small fragments of a jugal in Chulsanbaatar. The ju- NASOPARIETAL CONTACT gals described here in Kryptobaatar (PSS- MAE 101 and 113) are the ®rst complete One of the more unusual features of the ones reported for multituberculates (®gs. 22, skull roof above the orbit in the taeniolabi- 23). They conform with the jugals in Ptilo- dids Taeniolabis (Broom, 1914; Simpson, dus montanus, except that they do not extend 1937) and Lambdopsalis (Miao, 1988) is the as far forward into the anteroventral part of long anterior process of the parietal that the orbit. overlies the frontal and extends forward to Outside of multituberculates, monotremes contact the nasal. The absence/presence of are the only other mammaliaform group in this feature has been used in recent phylo- which replacement of the jugal in the zygo- genetic analyses within Multituberculata by matic arch by an elongate maxillary process Simmons (1993), Rougier et al. (1997a), and extending nearly to the squamosal glenoid is Kielan-Jaworowska and Hurum (1997). We likely primitive (Hopson et al., 1989). How- restudied the incidence of nasoparietal con- ever, rather than internal to the maxilla and tact in paulchoffatiids, following a sugges- squamosal as in multituberculates, the re- tion from J. A. Hopson, and offer the follow- duced jugal is dorsal to them in the platypus ing amendments to the scoring of this fea- (Zeller, 1989). The jugal is wholly absent in ture. Whereas nasoparietal contact has been the echidna (Kuhn, 1971). widely held to be a unique feature of taeniolabidids, we have also observed it in sev- BASICRANIUM AND LATERAL eral paulchoffatiids: Kuehneodon dryas (V.J. BRAINCASE WALL 454-155), Pseudobolodon krebsi (V.J. 447- 155, 451-155), and Pseudobolodon,n.sp. PTERYGOPALATINE RIDGES AND TROUGHS (V.J. 400-155, 450-155). The condition in In the mesocranium, Kryptobaatar has Meketichoffatia krausei (V.J. 445-155, 446- paired medial and lateral pterygopalatine 155) is uncertain. troughs (``mpt'' and ``lpt'' in ®g 37A) separated by the paired pterygopalatine ridges JUGAL (``ptr'' in ®g. 37A) and on the midline by the Until recently, the jugal was widely con- vomer (®gs. 14, 15, 20, 21, 34, 37A). A sim- sidered (e.g., Clemens and Kielan-Jaworows- ilar arrangement is known for other Mon- ka, 1979; Hahn, 1983) to be wholly lacking golian Late Cretaceous multituberculates in multituberculates. J. A. Hopson (personal (e.g., Kamptobaatar, Kielan-Jaworowska, commun. cited in Kielan-Jaworowska et al., 1971; Nemegtbaatar, Kielan-Jaworowska et 1986) noted the presence of a jugal or a facet al., 1986) and for Late Jurassic paulchoffa- for it on the medial side of the zygomatic tiids (e.g., Pseudobolodon oreas, Hahn, arch in several previously described speci- 1981), except in the latter the lateral trough mens. From Hopson's observations, Hahn is signi®cantly narrower and shallower than (1987) reported the presence of a jugal facet the medial. We have seen a similar arrange- on the medial side of the zygomatic process ment in Ptilodus (USNM 6076). Medial and of the maxilla in several paulchoffatiids, lateral pterygopalatine troughs are also found Kuehneodon dryas, Pseudobolodon oreas, in primitive mammaliaforms (e.g., Morgan- and ?Pseudobolodon, sp. indet., and we have ucodon, Kermack et al., 1981; Sinoconodon, observed this facet in Meketichoffatia krau- Crompton and Luo, 1993). In contrast, in sei (V.J. 446-155). Hopson et al. (1989) re- Lambdopsalis, pterygopalatine ridges are 84 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247 lacking, and a single trough bordered by a ilar function, directing the pull of the tendon very low, midline crest enters the choana of the tensor veli palatini. However, she (Miao, 1988). doubted its homologies with the therian ham- Regarding the occupants of the medial and ulus because the two are formed from dif- lateral troughs, the former held the air pas- ferent parts of the pterygoid. We agree with sageway, whereas there are two hypotheses Kielan-Jaworowska (1971), noting that the for the occupants of the lateral in multitu- therian hamulus is formed from the postero- berculates. According to Barghusen (1986), lateral part of the pterygoid, whereas the pro- the auditory (Eustachian) tube connecting the cess in Kryptobaatar and Kamptobaatar is nasopharynx and the cavum tympani occu- off the posteromedial part of the bone. pied the lateral trough. Other authors have reconstructed muscle there: tensor veli pala- ECTOPTERYGOID tini by Kielan-Jaworowska (1971) and Hahn (1981), and medial pterygoid by Gambaryan A separate ectopterygoid bone in the pos- and Kielan-Jaworowska (1995). We think terolateral corner of the palate has been de- that the morphology in Kryptobaatar elimi- scribed for one specimen of Morganucodon nates the auditory tube as a possible occu- oehleri (Kermack et al., 1981) and in the me- pant; the pterygopalatine ridge extends far socranium of various non-mammalian cyno- rostrally into the choana, an arrangement that donts (e.g., Thrinaxodon, Fourie, 1974). An required the auditory tube to empty into the ectopterygoid has also been said (e.g., Par- nasal cavity and not into the nasopharynx as rington and Westoll, 1940; Presley and Steel, it does in extant mammals (Starck, 1995). 1978) to be present in monotremes, but the The more likely course for the auditory tube homologies of this element have been ques- was one posterior to the pterygopalatine tioned (Wible, 1991). At one time, an ecto- ridge to join the nasopharynx in the medial pterygoid was also thought to be present in trough. On the other hand, the morphology the lateral wall of the choana in certain Mon- of the lateral trough in Kryptobaatar, in par- golian Late Cretaceous multituberculates ticular the fossa at its anterior end, appears (Kielan-Jaworowska, 1971; Clemens and appropriate for muscle attachment. However, Kielan-Jaworowska, 1979), although the su- we are uncertain what the muscular occupant ture delimiting it from the alisphenoid behind could have been. The tensor veli palatini was said to be indistinct (Kielan-Jaworowska arises more posteriorly in living therians (see et al., 1986). A separate ectopterygoid has Evans and Christensen, 1979), and the me- also been reconstructed in the early Eocene dial pterygoid would seem to have been at a ptilodontoid Ectypodus (Sloan, 1979). Miao mechanical disadvantage within the lateral (1988, 1993) suggested that the purported trough, with the medially in¯ected, lateral multituberculate ectopterygoid may merely choanal wall constraining the direction of the be part of the alisphenoid, which is how it muscle's pull. As a third hypothesis, we pro- was restored in Chulsanbaatar by Rougier et pose that a parallel channel to the main air al. (1992: ®g. 8C) and in subsequent recon- passageway may have occupied the lateral structions by Kielan-Jaworowska and co- trough. Either such a channel and/or muscle workers (e.g., Gambaryan and Kielan-Jawo- seem to be the only viable options. Perhaps rowska, 1995; Kielan-Jaworowska and Hu- evaluating these options in light of the mor- rum, 1997; Hurum, 1998a). This view is sup- phology of other mammaliaforms and mam- ported by the morphology of the alisphenoid maliaform outgroups, which is beyond the in Kryptobaatar and our observations of the scope of this report, may resolve this prob- multituberculates in the ZPAL collections. lem. Miao (1988) also suggested that although The pterygopalatine ridges in Kryptobaa- the ectopterygoid was not present as a sep- tar and Kamptobaatar (ZPAL MgM-I/33) arate element in multituberculates, it was end as a rounded, posteroventrally directed fused with the pterygoid to form the equiv- process that is reminiscent of the pterygoid alent of the therian hamulus. The basis for hamulus of living therians. Kielan-Jawo- this hypothesis is that the ventrolateral por- rowska (1971) suggested that it served a sim- tion of the therian pterygoid, which forms 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 85 from cartilage separate from the intramem- tures delimiting the alisphenoid from the an- branous dorsomedial portion of the ptery- terior lamina are very distinct and show goid, is homologous with the ectopterygoid without doubt that the alisphenoid is a small bone of other amniotes (Presley and Steel, element. Moreover, the suture between the 1978). This is not the only view on the ho- alisphenoid and anterior lamina in Krypto- mologies of these elements; Kuhn (1971) and baatar closely resembles that tentatively Zeller (1989) treated the pterygoid cartilage identi®ed for Kamptobaatar by Kielan-Ja- as a derivative of the palatoquadrate. This worowska (1971), with which we concur controversy is beyond the scope of our re- based on ZPAL MgM/I-33, and for Nemegt- port, but we have stated above that a hamulus baatar by Kielan-Jaworowska et al. (1986). as occurs in therians is not present in Kryp- Hurum (1998a) has provided a new recon- tobaatar or other multituberculates. struction for Nemegtbaatar based on a sectioned specimen in which the contribution of ALISPHENOID the alisphenoid to the braincase is more substantial than that reconstructed by Kielan-Ja- The alisphenoid in multituberculates has worowska et al. (1986). Because we have not been said (e.g., Kermack and Kielan-Jawo- rowska, 1971; Kielan-Jaworowska et al., seen the sectioned skull of Nemegtbaatar (or 1986) to be a small element with little con- Chulsanbaatar), we are not able to evaluate tribution to the side wall of the braincase. Hurum's reconstruction. Regarding Lamb- This is in contrast to the usual condition in dopsalis, we have not had access to the entire mammaliaforms in which the alisphenoid is collection of specimens studied by Miao expanded dorsally and contacts the frontal (1988), but we identi®ed a possible suture in (Novacek, 1986; Wible and Hopson, 1993; the anterior portion of the epitympanic recess Hopson and Rougier, 1993). Miao (1988, of IVPP V7151.80 (see Miao and Lillegrav- 1993) has questioned the purported small en, 1986; Rougier et al., 1996c). This pos- size of the alisphenoid in multituberculates, sible suture is in the same place that the ali- noting that there are few specimens that un- sphenoid contacts the petrosal in Kryptobaa- ambiguously detail the extent of the alisphe- tar, suggesting that Lambdopsalis also might noid in the braincase wall; one is Lambdop- have a small alisphenoid. This single obser- salis, which has an alisphenoid extending vation does not refute Miao's reconstruction dorsally to meet the parietal in the skull roof, for Lambdopsalis, but it does raise doubts according to Miao (but see below). In addi- about it, in particular in light of the frag- tion, Miao (1993: 69) included under his cat- mentary nature of the material. In summary, egory of multituberculate apomorphies an we are skeptical about the extreme variability ``elastic ossi®cation pattern in the orbitotem- reported for the multituberculate alisphenoid, poral region as re¯ected by the variability in and we think that Miao's (1993) inclusion of extent of the alisphenoid and the so-called an elastic ossi®cation pattern in the orbito- `anterior lamina of the petrosal'.'' temporal region as a multituberculate apo- A review of recent reconstructions of the morphy seems unwarranted. multituberculate braincase would seem to Kielan-Jaworowska et al. (1986) claimed support Miao's (1993) observation of ali- that the alisphenoid in multituberculates is sphenoid variability. At one extreme is Lambdopsalis, in which the alisphenoid is not pierced by any foramina. However, Miao said to extend to the skull roof (Miao, 1988); (1988) described in Lambdopsalis a foramen at the other extreme is Kryptobaatar,in ovale inferium and foramen masticatorium which the alisphenoid has very little contri- wholly within the alisphenoid and an ali- bution to the braincase (®g. 10). Intermediate sphenoid canal between the alisphenoid and between these two extremes, according to anterior lamina. Additionally, we report here Hurum (1998a), is the alisphenoid in Ne- for Kryptobaatar and various other Mongo- megtbaatar, with Chulsanbaatar showing an lian Late Cretaceous forms, a buccinator fo- element intermediate between Nemegtbaatar ramen between the alisphenoid and anterior and Kryptobaatar.InKryptobaatar, the su- lamina (see below; ``fbu'' in ®gs. 10, 36A). 86 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247

FORAMINA FOR BRANCHES OF THE tablishment of morphological criteria to rec- MANDIBULAR NERVE ognize the different branches of the trigeminal system based on osteology. Most Mesozoic mammaliaforms, includ- All multituberculates known to date with ing multituberculates, have two foramina in one exception have at least two apertures in the anterior lamina and/or petrosal that are the anterior lamina and/or petrosal that have interpreted for branches of the trigeminal been interpreted as for branches of the man- nerve (Wible and Hopson, 1993; Luo, 1994). dibular division of the trigeminal nerve: the Recently, we (Rougier et al., 1996a) have foramen ovale inferium and the foramen raised doubts about which branches of the masticatorium. The exception is Lambdop- trigeminal nerve occupied these various ap- salis in which the two foramina are said to ertures. To date, two models of nerve resto- be in the alisphenoid (Miao, 1988). In our ration have been proposed. First, Simpson preliminary comparisons, we have noted that (1937) restored branches of the mandibular the positions of the foramina in the petrosal nerve in the two foramina in Ptilodus mon- vary with regard to each other and to the tanus, citing Hill's (1935) observation of two epitympanic recess. In Late Jurassic paul- foramina for the mandibular nerve in the ali- choffatiids (e.g., Kuehneodon dryas, V.J. sphenoid of some extant rodents. Second, 454-155; Pseudobolodon, n. sp., V.J. 450- Patterson and Olson (1961) restored the max- 155), the foramen ovale inferium is at a level illary and mandibular nerves respectively in posterior to the foramen masticatorium, and the two foramina in the anterior lamina of both foramina are lateral to the epitympanic Sinoconodon. This model has been followed recess. In ptilodontoids (e.g., Ptilodus mon- by Kermack (1963, 1967) in reconstructing tanus, AMNH 35490; Mesodma thompsoni, nerves in Morganucodon and by most sub- Wible and Hopson, 1995), the two foramina sequent authors considering Mesozoic mam- are at the same level and the foramen ovale maliaforms other than multituberculates. In inferium traverses the epitympanic recess. contrast, authors researching multitubercu- Among Mongolian Late Cretaceous multitu- lates (e.g., Kielan-Jaworowska, 1971; Sloan, berculates, the foramen ovale inferium tra- 1979) invariably have followed the model verses the epitympanic recess in Sloanbaatar proposed by Simpson (1937). In describing (ZPAL MgM-I/20), whereas both foramina Kryptobaatar, we too have followed Simp- traverse it in Kryptobaatar (®g. 37A), Kamp- son's model. However, we have accepted it tobaatar (ZPAL MgM-I/33), Nemegtbaatar because the pattern of the foramina in Kryp- (ZPAL MgM-I/81), Chulsanbaatar (ZPAL tobaatar ®ts the arrangement of the mandib- MgM-I/168), and Catopsbaatar (ZPAL ular nerve foramina in some living rodents MgM-I/78). The foramina are roughly at the (Hill, 1935; Wahlert, 1974). Speci®cally, the same level in Kryptobaatar, Kamptobaatar, two foramina in Kryptobaatar are roughly at and Nemegtbaatar, but the foramen ovale in- the same level, with the medial one ante- ferium is more posterior in Sloanbaatar, roventrally directed and the lateral one ven- Chulsanbaatar, and Catopsbaatar. Finally, trolaterally directed (``foi'' and ``fma'' in ®g. in Lambdopsalis, the two foramina are at the 37A); this does not correspond to the posi- same level, lateral to the epitympanic recess, tions of the maxillary and mandibular nerve but supposedly enclosed in the alisphenoid exits in any extant mammals. As detailed be- (Miao, 1988). Another variant is found in low, there is some variation in the position Kamptobaatar (ZPAL MgM-I/33) in which of the two foramina in question in other mul- there are more than two foramina in the epi- tituberculates; however, the variation is not tympanic recess (®ve on the right side and extreme enough in any taxon to warrant an four on the left) that likely held branches of alternative nerve restoration. The implica- the mandibular nerve (Kielan-Jaworowska, tions of this for non-multituberculate mam- 1971). maliaforms are beyond the scope of this re- We have identi®ed in Kryptobaatar anoth- port, and what is required is careful compar- er foramen in the pterygoid fossa of the ali- ison of the relevant extinct taxa with the sphenoid between that bone and the anterior morphology of extant mammals and the es- lamina as for the buccal branch of the man- 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 87 dibular nerve (``fbu'' in ®gs. 10, 36A). This ited through the jugular foramen in Lamb- foramen is not unique to Kryptobaatar, but dopsalis (Miao, 1988) and through both the also occurs in Catopsbaatar (ZPAL MgM-I/ jugular foramen and foramen magnum in 78), Chulsanbaatar (ZPAL MgM-I/168), paulchoffatiids (Lillegraven and Hahn, Kamptobaatar (ZPAL MgM-I/33), Nemegt- 1993). baatar (ZPAL MgM-I/81), and Sloanbaatar Miao (1988) questioned the taxonomic (ZPAL MgM-I/20). Kielan-Jaworowska et al. signi®cance of differences in the sigmoid si- (1986) described for Chulsanbaatar (ZPAL nus's egress in multituberculates, because MgM-I/168; see also Hurum, 1998a: ZPAL this vessel exhibits considerable diversity MgM-I/84, the sectioned skull) and Nemegt- among extant and extinct mammals. How- baatar (ZPAL MgM-I/82 and 76, the sec- ever, at least some of this diversity falls along tioned skull) another more dorsally situated taxonomic lines; for example, the principal foramen wholly within the anterior lamina, exit of the sigmoid sinus in all extant mar- as possibly for the deep temporal branch of supials is via the foramen magnum (Archer, the mandibular nerve. We have not yet stud- 1976; Wible, 1990). Consequently, this char- ied the sectioned skulls, and we were not acter should not be eliminated a priori from able to con®rm the presence of this dorsal phylogenetic analyses. In fact, Rougier et al. foramen of the anterior lamina in the other (1996a, 1996c) have used the egress of the cited specimens. sigmoid sinus in their analyses of mammaliaform interrelationships. They observed that VENOUS SYSTEM in addition to paulchoffatiids, the sulcus for the sigmoid sinus extends to the jugular fo- Our reconstruction of the venous system ramen in Morganucodon and Haldanodon. has considered only a few vessels. We treat In contrast, in the other taxa for which the these below, with the exception of the frontal sigmoid sulcus is known, it does not reach diploic vein, which has already been dis- the jugular foramen. cussed above (see Orbital Foramina). Inferior Petrosal Sinus: Among extant Superior Sagittal and Transverse Sinuses: mammals, the inferior petrosal sinus follows A superior sagittal sinus connecting with a course along the basioccipital-petrosal su- paired transverse sinuses is nearly ubiquitous ture that may be largely either endocranial, among extant mammals (Gelderen, 1924), intramural, or extracranial (Wible, 1983). We and the multituberculates for which the rel- reconstruct the inferior petrosal sinus in an evant anatomy is known are no exception. endocranial course in Kryptobaatar in part Endocasts of Kryptobaatar, Ptilodus (Simp- because there is a shallow sulcus in the ap- son, 1937; Krause and Kielan-Jaworowska, propriate location in PSS-MAE 123 (®g. 25) 1993), Lambdopsalis (Miao, 1988), Nemegt- and in part because there is no bony pas- baatar, Chulsanbaatar (Kielan-Jaworowska sageway appropriate for an extracranial or et al., 1986), and an undescribed new species intramural route. Moreover, given the small of Mongolian Late Cretaceous multituber- size of the jugular foramen (®gs. 26, 37A), culate (PSS-MAE 126) show evidence that it seems likely that this vessel exited the cra- this dural sinus pattern was present. nial cavity via the foramen magnum in the Sigmoid Sinus: In Kryptobaatar, the sig- vertebral veins (``vv'' in ®g. 37B). We be- moid sinus does not exit the skull via the lieve that Nemegtbaatar and Chulsanbaatar small jugular foramen, but likely left through likely exhibited the same pattern; there is no the foramen magnum in the vertebral veins likely route other than an endocranial one, (``vv'' in ®g. 37B). The same arrangement and the jugular foramen in these forms is has been reported for Chulsanbaatar and Ne- also small (Kielan-Jaworowska et al., 1986). megtbaatar by Kielan-Jaworowska et al. The inferior petrosal sinus was not recon- (1986) based on clear evidence from endo- structed for Lambdopsalis by Miao (1988), cranial casts (contra Miao, 1988). A foramen but we think it likely occupied a deep sulcus magnum exit for the sigmoid sinus is not the on the endocranial surface between the pe- universal pattern in multituberculates. trosal and the basioccipital (Miao, 1988: ®g. Grooves indicate that the sigmoid sinus ex- 23). In contrast to the Late Cretaceous taxa, 88 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247 the jugular foramen in Lambdopsalis is large otic canal lies between the cavum epipteri- (Miao, 1988) and may have transmitted the cum and the subarcuate fossa (see also Lil- inferior petrosal sinus, as occurs in most ex- legraven and Hahn, 1993: ®gs. 1, 6). Also, tant therians (Wible, 1983). As noted by the tympanic aperture is not con¯uent with Rougier et al. (1996a), canals in the petrosal the pterygoparoccipital foramen in Chulsan- of some paulchoffatiids indicate that the in- baatar (ZPAL MgM-I/168) and in an isolat- ferior petrosal sinus had an intrapetrosal ed petrosal tentatively assigned to Menis- course to the jugular foramen as occurs in all coessus (Luo, 1989; Wible and Hopson, mammaliaforms with the exception of extant 1995). Additionally, the condition in paul- mammals. We have also observed such ca- choffatiids and Nemegtbaatar is uncertain. nals in Ptilodus montanus (AMNH 35490). Post-trigeminal Vein: In multituberculates, The small size of the jugular foramen in the vascular canal in the petrosal above the Kryptobaatar, Nemegtbaatar, and Chulsan- epitympanic recess and lateral ¯ange has baatar suggests an unusual pattern for the been called variously the canal for the ?max- drainage of the dural sinuses. As stated illary artery (Kielan-Jaworowska et al., above, it is likely that neither the inferior pe- 1986), the post-trigeminal canal (Rougier et trosal sinus nor the sigmoid sinus exited al., 1996a, 1996c), and the canal for the ra- through the jugular foramen. Consequently, mus inferior (Miao, 1988; Wible and Hop- the lateral head vein was the major constit- son, 1995). In ``Catopsalis'' joyneri and Me- uent of the internal jugular vein. The only sodma thompsoni, the canal is wide enough extant mammal exhibiting this same pattern that it likely had two occupants, the ramus is the echidna (Hochstetter, 1896). The platy- inferior and the post-trigeminal vein (Wible pus has a minor variant in that a part of the and Hopson, 1995). We are con®dent that a inferior petrosal sinus drains into the internal ramus inferior occupied the canal in question jugular vein through the enormous metotic in Kryptobaatar, given the size of the groove ®ssure, which is formed by the con¯uent jug- for the stapedial artery and the ventral as- ular and hypoglossal foramina (®g. 28), and cending canal; however, we cannot exclude part is into the vertebral veins via the fora- the presence of a post-trigeminal vein as well men magnum (Rougier et al., 1992). In con- (``ptv'' in ®gs. 36B, 37B). A canal for the trast, in marsupials and placentals, the major ramus inferior is also present in Lambdop- contributor to the internal jugular vein (the salis, but the incidence of this structure inferior petrosal and/or sigmoid sinuses) ex- among other multituberculates is uncertain its the cranial cavity via the jugular foramen. due to preservational problems. The only Lateral Head Vein and Prootic Sinus: A other mammaliaform with a post-trigeminal well-developed prootic canal that transmitted vein enclosed in a canal is the echidna (Wi- the prootic sinus to the lateral head vein is ble and Hopson, 1995). known for all multituberculates preserving Pituito-orbital Vein: In Kryptobaatar,we the relevant portion of the petrosal bone. have identi®ed a separate foramen in the an- Several characters of this vessel have been terolateral wall of the hypophyseal fossa for employed in phylogenetic analyses among the pituito-orbital vein (``fpv'' in ®gs. 25, 27, mammaliaforms: most recently, the position 36A). Miao (1988), in his description of of the cranial and tympanic apertures of the Lambdopsalis, is the only other author to dis- prootic canal (Rougier et al., 1996a, 1996c). cuss this vein (his pituitary vein) in Meso- In multituberculates, the cranial aperture is zoic mammals. He concluded that this vein said by Rougier et al. (1996a, 1996c) to lie shared the metoptic foramen with the ocu- at the anterodorsal margin of the subarcuate lomotor nerve, based on the co-occurrence of fossa, whereas the tympanic aperture is con- these structures in the chondrocranium of ¯uent with the pterygoparoccipital foramen some living sauropsids (e.g., Lacerta,De (ventral ascending canal). The prootic canal Beer, 1937). This restoration was the only in Kryptobaatar conforms with these obser- logical one for Miao, because a separate fo- vations (®gs. 27, 37A). However, we add ramen for the vein, as occurs in the chondro- here that in some paulchoffatiids (V.J. 73- cranium of some other sauropsids (e.g., 155, 82-155) the cranial aperture of the pro- Sphenodon, Bellairs and Kamal, 1981), was 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 89 not identi®ed in Lambdopsalis (although he mammaliaforms as Morganucodon and Vin- did report a nutrient foramen in the midline celestes (Rougier et al., 1992). In fact, one of the anterior wall of the hypophyseal fos- speci®c feature of the arterial pattern, a hor- sa). In the case of Kryptobaatar, there are izontal ventral ascending canal, supports the two distinct foramina, and in our restoration grouping of multituberculates with tricono- we follow the model provided by those sau- dontids and prototribosphenidans (Rougier et ropsids where the nerve and vein occupy al., 1996a). separate foramina. A likely candidate for a Internal Carotid Artery: Kryptobaatar has separate foramen for the pituito-orbital vein provided information about certain arteries does not appear to be present in Meketichof- not included in other vascular reconstructions fatia (V.J. 446-155), although it should be of multituberculates. Probably the most note- noted that this region is not in pristine con- worthy of these is the basicranial portion of dition. the internal carotid artery, which has proven Transverse Canal Vein: Here is reported to be of systematic value among some groups for the ®rst time in multituberculates the of extant mammals (Wible, 1986). The presence of a transverse canal vein. This vein groove on the promontorium in Kryptobaa- in Kryptobaatar (``tcv'' in ®g. 36B) is an un- tar clearly indicates that the internal carotid expected occurrence among Mammaliafor- followed a transpromontorial course en route mes, because a similar canal is only known to the cranial cavity (``ica'' in ®g. 37B). for some metatherians (Marshall et al., 1990; Among multituberculates, a similar course Marshall and Muizon, 1995) and placentals has been described previously from sulci (McDowell, 1958; MacPhee, 1994). The ab- only in Ectypodus (Sloan, 1979), and outside sence of this canal in other multituberculates of multituberculates, only in Vincelestes may be an artifact of preservation or related (Rougier et al., 1992) and certain eutherians to the relative inaccessibility of the medial (MacIntyre, 1972; Wible, 1986). However, wall of the sphenorbital recess for prepara- the basicranial course of the internal carotid tion. In fact, two of the sections of Nemegt- in Kryptobaatar is atypical in two regards: baatar published by Hurum (1998a: ®gs. 2, ®rst, it does not run the length of the prom- 4) show gaps across the midline dorsal to the ontorium but is con®ned to the rostral half horizontal portion of the basisphenoid in the (also in Ectypodus), and second, its rostral- skull base that are reminiscent of the canal most part is enclosed in a long canal between in Kryptobaatar. The meaning of these gaps the petrosal, alisphenoid, and pterygoid that awaits further consideration. The phyloge- leads to the carotid foramen in the basisphe- netic implications of the transverse canal noid. Both of these derived conditions occur vein are unclear, and we are uncertain about only rarely among certain placentals the homology of the transverse canal of mul- (MacPhee, 1981; Wible, 1984; MacPhee and tituberculates with that of later therians. Cartmill, 1986). Among multituberculates, a transpromontorial course con®ned to the ros- ARTERIAL SYSTEM tral half of the promontorium like that in Kryptobaatar and Ectypodus is also indicated The arterial reconstructions presented here by promontorial grooves in Meketichoffatia for Kryptobaatar do not differ in principle (V.J. 446-155), Ptilodus (AMNH 35490), from that offered previously for Lambdop- Chulsanbaatar (ZPAL MgM-I/168), Kamp- salis (Miao, 1988), ``Catopsalis'' joyneri tobaatar (ZPAL MgM-I/33), Nemegtbaatar (Rougier et al., 1992; Wible and Hopson, (ZPAL MgM-I/82), and Sloanbaatar (ZPAL 1995), and a generalized taeniolabidoid MgM-I/20). A carotid groove is lacking in (Rougier et al., 1992). The most striking dif- Pseudobolodon oreas (V.J. 460-155), Kueh- ferences among these reconstructions con- neodon dryas (V.J. 454-155), Lambdopsalis cern the positions of particular osseous pas- (Miao, 1988), and the isolated petrosals re- sageways (e.g., supraglenoid foramen, ven- ferred to ``Catopsalis'' joyneri and Mesodma tral ascending canal). The pattern recon- thompsoni (Wible and Hopson, 1995). Given structed here for Kryptobaatar also generally that carotid foramina are putatively lacking agrees with that interpreted for such diverse in Lambdopsalis (Miao, 1988), its internal 90 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247 carotid artery apparently did not reach the morphology of the carotid canal in Krypto- cranial cavity, and the vertebral arteries must baatar. Among other mammaliaforms, the have been the principal supplier of blood to ventral aperture is generally within the ba- the brain. Noted, however, that given the vas- sisphenoid (Archer, 1976; Rougier et al., cular pattern reconstructed for Lambdopsalis 1992; Luo, 1994) except in some placentals by Miao (1988), including the presence of a (De Beer, 1937; Wible, 1984). well-developed stapedial artery, the absence Stapedial Artery: A stapedial groove run- of a carotid foramen is surprising. ning posterolaterally on the promontorium to In Kryptobaatar, the carotid foramen is the oval window like that in Kryptobaatar entirely within what we think is the basi- (``gpsa'' in ®g. 37A) has been reported presphenoid (``pacc'' in ®g. 37A); its ventral ap- viously for multituberculates in Ectypodus erture is not on the basicranial surface but is (Sloan, 1979) and in various isolated petro- recessed somewhat dorsally at the end of the sals, including those referred to ``Catopsal- carotid canal (®g. 14), and its dorsal aperture is'' joyneri (Kielan-Jaworowska et al., 1986; is in the posterolateral part of the hypophy- Wible and Hopson, 1995) and Meniscoessus seal fossa (``®ca'' in ®gs. 25, 27). Among (Luo, 1989), as well as unreferred specimens multituberculates, the location of the dorsal from the Bug Creek Anthills (Fox and Meng, aperture of the carotid foramen has been re- 1997). We have also observed such a stape- ported only in Meketichoffatia krausei dial groove in Meketichoffatia (V.J. 446- (Hahn, 1981), Nemegtbaatar (Kielan-Jawo- 155), Ptilodus (AMNH 35490), Chulsanbaa- rowska et al., 1986; Hurum, 1998a), and tar (ZPAL MgM-I/168), Kamptobaatar Chulsanbaatar (Hurum, 1998a). As in Kryp- (ZPAL MgM-I/33), Nemegtbaatar (ZPAL tobaatar, the carotid foramen is in the pos- MgM-I/82), and Mesodma thompsoni terolateral part of the hypophyseal fossa in (FMNH 53904). In light of the position of M. krausei and Nemegtbaatar. However, in the groove in these taxa, it seems likely that Chulsanbaatar, it is illustrated (Hurum, their stapedial artery ran through a bicrurate 1998a: ®g. 12) in the anterolateral part of the stapes, as reconstructed for Kryptobaatar.A hypophyseal fossa, although no basis for this groove for the stapedial artery is also present position is provided. Among other mamma- in Lambdopsalis, but it differs in that it runs liaforms, the dorsal aperture of the carotid wholly posterior to the oval window; there- foramen is in the anterolateral part of the hy- fore, the artery did not perforate the stapes pophyseal fossa in Triconodon (Kermack, but ran behind it (Miao, 1988), a fact sub- 1963) and Morganucodon (Kermack et al., sequently con®rmed by Meng's (1992) de- 1981) and in the posterolateral part in mono- scription of a columnar stapes in Lambdop- tremes (Zeller, 1989). In contrast, the dorsal salis. A stapedial groove is not ubiquitous aperture appears to be wholly lateral to the among multituberculates, being absent for hypophyseal fossa in marsupials (see Archer, example in paulchoffatiids (Lillegraven and 1976) and placentals (see Ellenberger and Hahn, 1993) other than Meketichoffatia. Baum, 1908; Cooper and Schiller, 1975; Among other mammaliaforms, a stapedial Evans and Christensen, 1979; Novacek, groove is known only for Vincelestes (Roug- 1986). Among multituberculates, the ventral ier et al., 1992) and certain eutherians (Wi- aperture of the carotid foramen is on the ba- ble, 1987). However, among these taxa, the sicranial surface within the basisphenoid in artery's course to the oval window is usually Meketichoffatia krausei (Hahn, 1981) and a lateral or anterolateral one; a posterolateral Ptilodus (Simpson, 1937), and between the course is known only for the tarsier (Mac- basisphenoid and the petrosal in Ectypodus Phee and Cartmill, 1986). (Sloan, 1979), Kamptobaatar (ZPAL MgM- All multituberculates for which the rele- I/33), Sloanbaatar (ZPAL MgM-I/20), Ne- vant anatomy is preserved have a lateral megtbaatar, and Chulsanbaatar (Kielan-Ja- ¯ange that is bent medially to contact the worowska et al., 1986). Noted, however, that promontorium (Rougier et al., 1996a), except the position of the carotid foramen in the for Lambdopsalis in which, according to Mongolian taxa implies a long intraosseous Miao (1988), the alisphenoid accomplishes course for the artery, which approximates the this. It is the infolded lateral ¯ange that 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 91 forms the ¯oor of the canal for the ramus inferior in other multituberculates, and we in- inferior (post-trigeminal canal). Because of terpret them as homologous. The homoge- preservational problems, the canal for the ra- neity of this region across Multituberculata, mus inferior has been described in only a few invariably formed by the petrosal, casts ad- multituberculates; in addition to Kryptobaa- ditional doubts on Miao's (1988) interpreta- tar, these include Nemegtbaatar (Luo, 1989), tion of the alisphenoid in Lambdopsalis (see Lambdopsalis (Miao, 1988), and various iso- above). What Miao (1988) terms the alisphe- lated petrosals, including those referred to noid canal, we think merely re¯ects the ``Catopsalis'' joyneri (Kielan-Jaworowska et course of the ramus inferior through the ca- al., 1986), Mesodma thompsoni (Wible and vum epiptericum en route to the sphenorbital Hopson, 1995), and Meniscoessus (Luo, ®ssure. 1989). To date, we have found only one mul- A pattern for the canals associated with the tituberculate in which the canal for the ramus ramus superior and arteria diploeÈtica magna inferior is without a doubt absent (Ptilodus resembling that described here for ``Catop- montanus, AMNH 35490), although the ab- salis'' joyneri and Kryptobaatar (®g. 36A) sence of the canal does not necessarily imply has been described for other multitubercu- the absence of the artery, which could have lates, including Chulsanbaatar, Nemegtbaa- run ventral to the tympanic roof. Among oth- tar (Kielan-Jaworowska et al., 1986; Rougier er mammaliaforms, similar canals in the tym- et al., 1992), Lambdopsalis (Miao, 1988), panic roof are known only for the echidna Ptilodus (Wible and Hopson, 1995), and iso- and certain placentals (e.g., scandentians, lated petrosals referred to Meniscoessus macroscelidids), where they are formed by (Luo, 1989) and Mesodma thompsoni (Wible appositional bone growth from the petrosal and Hopson, 1995). Differences among taxa (Kuhn, 1971; MacPhee, 1981). In the echid- concern the positions of foramina, the com- na, the major occupant is venous (Wible and munications between canals, and the enclos- Hopson, 1995), whereas in placentals, it is ing bony elements. For example, the tym- the ramus inferior of the stapedial artery panic aperture of the ventral ascending canal (MacPhee, 1981). (pterygoparoccipital foramen of non-mam- Miao's (1988) interpretation of the osseous malian cynodonts) for the ramus superior is structures associated with the course of the con¯uent with that for the prootic canal in ramus inferior in Lambdopsalis differs sub- most taxa including Kryptobaatar (®g. 37), stantially from that presented here for other but is posterior to the prootic canal in Chul- multituberculates. He proposed that the ra- sanbaatar (ZPAL MgM-I/168) and Menis- mus inferior en route to the cavum epipteri- coessus (Luo, 1989). The ventral ascending cum passed through an alisphenoid canal, in- canal has a communication with the prootic dicated by a groove on the medial side of the canal intramurally in Nemegtbaatar (Kielan- alisphenoid and an opening anterior to the Jaworowska et al., 1986) and Lambdopsalis foramen ovale inferium. Additionally, the ra- (Miao, 1988), but remains separate in ``Ca- mus inferior was exposed laterally through topsalis'' joyneri and Mesodma thompsoni an elongated fenestra slightly larger than the (Wible and Hopson, 1995); the condition in foramen ovale inferium. Behind this fenestra, Kryptobaatar is unknown. The supraglenoid the ramus inferior was enclosed in a second foramen for a ramus temporalis opens in the canal that Miao terms ``canal of maxillary anterior lamina at a level anterior to the fe- artery,'' following Kielan-Jawarowska et al. nestra vestibuli in most taxa, including Kryp- (1986). Although we have not seen all the tobaatar and the paulchoffatiids Meketichof- relevant specimens of Lambdopsalis,we fatia (V.J. 446-155), Kuehneodon (V.J. 454- speculate that the fenestra is an artifact and 155), and Pseudobolodon (V.J. 450-155), but that the ramus inferior was fully enclosed in is level with the fenestra vestibuli in Chul- an osseous canal as in other multitubercula- sanbaatar (ZPAL MgM-I/168), Sloanbaatar tes, except Ptilodus. The relationships of the (ZPAL MgM-I/20), Lambdopsalis (Miao, osseous canal in Lambdopsalis (i.e., to the 1988), and Mesodma thompsoni (Wible and facial nerve, prootic canal, and stapedial sys- Hopson, 1995). As in most mammaliaforms tem) are the same as the canal for the ramus (Luo, 1994), the posterior opening into the 92 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247

posttemporal canal for the arteria diploeÈtica PTERYGOID CANAL magna is between the squamosal and petrosal in paulchoffatiids (Lillegraven and Hahn, Arising from the carotid groove on the an- 1993), Nemegtbaatar (Hurum, 1998a), and terolateral pole of the promontorium in Kryp- Kamptobaatar (ZPAL MgM-I/33), but is tobaatar is a canal of subequal diameter that within the petrosal in other multituberculates is directed anterolaterally toward the orbit; including Kryptobaatar (®gs. 35, 36A). Also, we have interpreted this as the pterygoid ca- the posterior opening is located low on the nal (``ptca'' in ®g. 37A) based on the resem- occiput in most taxa, including Kryptobaatar blance to that, for example, in lipotyphlans and Pseudobolodon (V.J. 460-155), but is (McDowell, 1958; MacPhee, 1981; Novacek, dorsally situated in Ptilodus (AMNH 35490) 1986). We think that some other Mongolian and Lambdopsalis (Miao, 1988); the position Late Cretaceous taxa have a similar arrange- of the posterior opening into the posttempo- ment, with the only difference being that the ral canal dictates the length of ventral as- posterior aperture of the pterygoid canal is cending canal. A foramen of the dorsal as- farther away from the posterior aperture of cending canal for a ramus temporalis is the carotid canal, and these two openings are found, in addition to Kryptobaatar,inKamp- connected by a groove for the contents of the tobaatar (ZPAL MgM-I/33), Chulsanbaatar pterygoid canal. The morphology in question (ZPAL MgM-I/168), Ptilodus (AMNH was described for Chulsanbaatar and Ne- 35490), and Lambdopsalis (Miao, 1988), but megtbaatar by Kielan-Jaworowska et al. is lacking in specimens of Nemegtbaatar. Fi- (1986). In these forms, the carotid groove nally, as discussed above, an orbitotemporal leads to the posterior aperture of the carotid channel that transmitted the ramus superior canal situated more posteriorly than in Kryp- to the orbit has yet to be identi®ed in paul- tobaatar. Originating at the posterior aper- choffatiids, but it was likely present given ture of the carotid canal is a narrower groove the widespread presence of this structure. that curves anteromedially across the lateral The general pattern for the ramus superior and anterior surface of the promontorium's and arteria diploeÈtica magna that occurs in anterior pole and ends at the pterygoid canal multituberculates is repeated in most regards between the petrosal and basisphenoid. Kie- in other mammaliamorphs (Rougier et al., lan-Jaworowska et al. (1986) interpreted 1992, 1996a; Wible and Hopson, 1995). The what we think to be the posterior aperture of most striking differences among taxa concern the carotid canal as the `hiatus Fallopii' the course of the ramus superior and its ros- (quotes in original) and the pterygoid canal tral continuation, the ramus supraorbitalis, as the carotid foramen. That the `hiatus Fal- which may have been wholly or partially ex- lopii' of Kielan-Jaworowska et al. (1986) is tracranial, intramural, or endocranial. The not homologous with that of other mammals taxon that most closely resembles the mul- was recognized by these authors and has tituberculate pattern is the prototribospheni- been discussed by others (e.g., Miao, 1988; dan Vincelestes (Rougier et al., 1992, 1996a). Luo, 1989; Wible and Hopson, 1995), but Like multituberculates, Vincelestes has an in- there is no consensus on alternative occu- tramural ventral ascending canal within the pants for this aperture in Chulsanbaatar and anterior lamina that initially runs posteriorly Nemegtbaatar. Given the remarkable posi- in a horizontal plane; it also has multiple fo- tional similarity with the posterior aperture ramina for rami temporales, including a fo- of the carotid canal in Kryptobaatar,we ramen of the dorsal ascending canal on the think that it served the same function in suture between the anterior lamina, squa- Chulsanbaatar and Nemegtbaatar. We have mosal, and parietal (Rougier et al., 1992). seen the same arrangement, that is, a ptery- The only other taxa with a horizontal course goid canal arising from the carotid groove, for the ventral ascending canal are tricono- in Sloanbaatar (ZPAL MgM-I/20) and dontids (Rougier et al., 1996a, 1996c), but Kamptobaatar (ZPAL MgM-I/33). Given the their canal differs in that it is intramural be- size of the aperture into the pterygoid canal tween the petrosal and squamosal (Rougier in these forms, its occupants likely included et al., 1996a). an artery and nerve (i.e., the deep petrosal 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 93 nerve). As with Kryptobaatar, the greater pe- rowska et al., 1986; Luo, 1989, 1996; Lille- trosal nerve must have entered the canal di- graven and Hahn, 1993; Wible and Hopson, rectly from the cavum epiptericum and/or ca- 1995; Kielan-Jaworowska and Hurum, vum supracochleare. 1997). Some of these differences are discussed here. PETROSAL Jugular Fossa: Kielan-Jaworowska and Hurum (1997) used the size and depth of the During the last few years, we have pub- jugular fossa in their phylogenetic analysis of lished several phylogenetic analyses of Me- selected multituberculate taxa. They reported sozoic mammaliamorphs relying principally it to be ``small and shallow'' in the hypo- on features of the petrosal bone (Wible and thetical ancestor, based on Late Jurassic Pla- Hopson, 1993; Wible et al., 1995; Rougier et giaulacoidea and Ptilodus, and ``large and al., 1996a, 1996c). Among multituberculates, deep'' in Kryptobaatar (``jf'' in ®g. 37A), the petrosal has been described in some de- Nemegtbaatar, Chulsanbaatar, Catopsbaa- tail in various paulchoffatiids (Hahn, 1988; tar, Kamptobaatar, and Lambdopsalis.To Lillegraven and Hahn, 1993), Ptilodus their list we add Sloanbaatar (ZPAL MgM- (Simpson, 1937; Luo, 1989; Wible and Hop- I/20). The only plagiaulacoid for which we son, 1995), Chulsanbaatar, Nemegtbaatar could con®rm a small, shallow jugular fossa (Kielan-Jaworowska et al., 1986; Hurum, was Pseudobolodon (V.J. 450-155, 460-155). 1998a, 1998b), Lambdopsalis (Miao, 1988; In Ptilodus gracilis (USNM 6076) and Ec- Meng and Wyss, 1995), and from an array typodus (YPM-PU 14724), the jugular fossa of isolated elements from the Late Creta- is shallow but not small. Finally, we could ceous of North America (Kielan-Jaworowska not con®rm the size and depth of the jugular et al., 1986; Luo, 1989, 1996; Luo and Ket- fossa in Catopsbaatar. ten, 1991; Fox and Meng, 1997), including Channel for Perilymphatic Duct: Prior to those referred to ``Catopsalis'' joyneri and our description of Kryptobaatar, two char- Mesodma thompsoni (Wible and Hopson, acter states of the channel for the perilym- 1993, 1995). A striking result of these studies phatic duct in multituberculates had been has been how uniform the petrosal morphol- identi®ed by us (Rougier et al., 1996a, ogy is across Multituberculata, a conclusion 1996c): (1) an open sulcus, and (2) a sulcus reinforced by our descriptions of Kryptobaa- partially enclosed by bony lappets. The for- tar. However, the only derived petrosal fea- mer state occurs in Ptilodus and Mesodma ture shared by all multituberculates known to thompsoni, the latter in ``Catopsalis'' joy- date is an infolded bony ¯ange that contacts neri. We have subsequently seen bony lap- the promontorium and housed the epitym- pets on the perilymphatic channel in Meke- panic recess (Rougier et al., 1996a, 1996c). tichoffatia krausei (V.J. 446-155) and Pseu- This bony ¯ange is continuous with the an- dobolodon oreas (V.J. 460-155). Kryptobaa- terior lamina and crista parotica in all mul- tar exhibits a third state: no indication of the tituberculates with the sole exception of perilymphatic channel whatsoever. Lambdopsalis, following Miao's (1988) in- Recessus Scalae Tympani: Ornithorhyn- terpretation (but see our comments in the chus possesses a distinct depression imme- Alisphenoid section above). Some of the pe- diately external to the perilymphatic fora- trosal features discussed below may be di- men, the recessus scalae tympani (perilym- agnostic of Multituberculata but are depen- phatic recess of Fox and Meng, 1997), where dent on the phylogenetic positions of the ex- the perilymphatic duct contacts the cavum ceptions. Other petrosal features have been tympani (Zeller, 1991, 1993). We have ob- discussed already under the Venous System served a similar depression in the tricono- and Arterial System sections. dontid Priacodon (Rougier et al., 1996a). We Despite the general uniformity/similarity also report a conspicuous recessus scalae in petrosal morphology within Multituber- tympani in the following multituberculates culata, our comparisons of Kryptobaatar (see also Fox and Meng, 1997): Ptilodus with other multituberculates have also high- montanus (AMNH 35490), Ectypodus lighted differences (see also Kielan-Jawo- (YPM-PU 14724), Mesodma thompsoni 94 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247

(FMNH 53904), ``Catopsalis'' joyneri To the list of multituberculates with a caudal (AMNH 119445), and Kamptobaatar (ZPAL tympanic process, we now add Meketichof- MgM-I/33). In contrast, the depression is in- fatia krausei (V.J. 446-155), Chulsanbaatar conspicuous in Kryptobaatar, Pseudobolo- (ZPAL MgM-I/145), Nemegtbaatar (contra don (V.J. 460-155), Meketichoffatia krausei Rougier et al., 1996a; ZPAL MgM-I/82), and (V.J. 446-155), Nemegtbaatar (ZPAL MgM- Ptilodus gracilis (USNM 6076). On the list I/81), Sloanbaatar (ZPAL MgM-I/20), and of those without this process are Pseudobol- Lambdopsalis (Miao, 1988). odon (V.J. 450-155, 460-155) and Ectypodus Post-promontorial Tympanic Recess: Our (YPM-PU 14724). analyses (e.g., Wible and Hopson, 1993; Paroccipital Process: The form of the par- Rougier et al., 1996a, 1996c) have indicated occipital process varies among multituber- that the tall, horizontal crista interfenestralis culates. We have described it above as tri- marks the posterior boundary of the cavum angular in Kryptobaatar (``ppr'' in ®g. 37A); tympani in multituberculates and that a post- this shape also occurs in Meketichoffatia promontorial tympanic recess behind the krausei (V.J. 446-155), Chulsanbaatar, Ne- perilymphatic foramen was lacking. How- megtbaatar, Sloanbaatar, and Kamptobaatar ever, the crista interfenestralis is not very tall (see Kielan-Jaworowska and Hurum, 1997: in any multituberculates, and at least in a few ®g. 11). The paroccipital process is a ®nger- forms (e.g., Pseudobolodon, V.J. 450-155, like projection in Ectypodus (YPM-PU 460-155) it seems likely that the cavum tym- 14724), and is expanded posterolaterally in pani extended into a post-promontorial re- Pseudobolodon (V.J. 450-155) and Ptilodus cess. Moreover, if the depression identi®ed montanus (AMNH 35490). Two of these above as the recessus scalae tympani had the conditions appear to characterize the paroc- same function as in Ornithorhynchus (i.e., cipital process of Lambdopsalis, where it is housing a diverticulum of the perilymphatic both triangular and posterolaterally expanded duct that contacted the cavum tympani, Zell- (Miao, 1988). er, 1991), then the crista interfenestralis did Medial Wall of Epitympanic Recess: not prevent the cavum tympani from reach- Rougier et al. (1996c) reported that the me- ing the perilymphatic duct in those multitu- dial wall of the epitympanic recess in Kryp- berculates with a conspicuous recessus scalae tobaatar (®g. 21; PSS-MAE 113) is formed tympani. A post-promontorial recess has by a sharp, ventrally projecting crest com- heretofore been known only for prototribo- posed of the crista parotica posteriorly (``cp'' sphenidans (Wible, 1990) and isolated petro- in ®g. 37A) and the lateral ¯ange anteriorly. sals from the Early Cretaceous Khoobur lo- They also noted that the vertical nature of cality (Wible et al., 1995). In light of the this wall would have severely constrained the morphology in multituberculates, this char- possible positions of the stapes and incus. acters needs to be reevaluated before use in Hurum et al. (1996) observed a similar ar- future phylogenetic studies. rangement to the medial wall of the epitym- Caudal Tympanic Process: Until recently, panic recess in Kamptobaatar and Chulsan- our analyses (e.g., Wible et al., 1995; Roug- baatar and suggested ``that there must have ier et al., 1996a) had shown that a crest me- been a signi®cant vertical component in the dial to the paroccipital process forming the plane of the manubrium and therefore of the posterior wall of the tympanic cavity was a tympanic membrane'' (p. 269). We add here synapomorphy of prototribosphenidans plus that a tall medial wall of the epitympanic re- isolated petrosals from the Early Cretaceous cess also occurs in Meketichoffatia (V.J. 110- Khoobur locality. In addition, a caudal tym- 155), Ptilodus (AMNH 35490), Mesodma panic process has been reported in the sym- thompsoni (FMNH 53904), Nemegtbaatar metrodont Zhangheotherium (Hu et al., (ZPAL MgM-I/82), Catopsbaatar (ZPAL 1997). A caudal tympanic process had been MgM-I/78), and Lambdopsalis (Miao, 1988). unknown for multituberculates, but Rougier In fact, the only multituberculate preserving et al. (1996a, 1996c) reported one in Kryp- the relevant anatomy that has a low medial tobaatar (``ctpp'' in ®g. 37A; PSS-MAE wall is Ectypodus (YPM-PU 14724). We 113) and Kamptobaatar (ZPAL MgM-I/33). have not seen similarly developed medial 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 95 walls of the epitympanic recess in other Me- ganglion, continuous with the cavum epi- sozoic mammaliaforms. ptericum. As with the facial ganglion ¯oor, Lateral Wall of Epitympanic Recess: We there are few taxa in which this morphology found two different conditions for the lateral can be veri®ed, most notably paulchoffatiids wall of the epitympanic recess in multituber- (Lillegraven and Hahn, 1993), ``Catopsalis'' culates: (1) either it is a low crest and the joyneri (AMNH 119445), Mesodma thomp- epitympanic recess is not sharply demarcat- soni (FMNH 53904), and Kryptobaatar ed, so that the lateral braincase wall and epi- (PSS-MAE 123). Luo (1996: 49A) reported tympanic recess are con¯uent, or (2) promi- that a ``geniculate ganglion completely en- nent ridges separate the epitympanic recess closed in the cavum supracochleare'' is a from the lateral braincase wall. The former synapomorphy of Cimolomyidae and Tae- condition is found in Meketichoffatia krausei niolabididae. Given that this statement was (V.J. 446-155), Kuehneodon dryas (V.J. 454- taken from a published abstract, we cannot 155), Ectypodus (YPM-PU 14724), Mesod- evaluate the supporting evidence. However, ma thompsoni (FMNH 53904), and Kamp- Miao (1988, 1993) clearly has described the tobaatar (ZPAL MgM-I/33); the latter con- cavum supracochleare continuous with the dition is found in Kryptobaatar (®gs. 20, 21), cavum epiptericum in the taeniolabidid Pseudobolodon (V.J. 450-155), Ptilodus Lambdopsalis. The only cimolomyid repre- montanus (AMNH 35490), Chulsanbaatar sented by a petrosal is Meniscoessus (Luo, (ZPAL MgM-I/168), Nemegtbaatar (ZPAL 1989), and the pattern of its cavum supra- MgM-I/76), Sloanbaatar (ZPAL MgM-I/20), cochleare has not yet been described. In our Catopsbaatar (ZPAL MgM-I/78), ``Catop- comparisons, we have discovered one mul- salis'' joyneri (AMNH 119445), and Lamb- tituberculate in which the cavum supracoch- dopsalis (Miao, 1988). leare is partially walled off from the cavum Tensor Tympani Fossa: Our analyses (e.g., epiptericum and enclosed within the petrosal Wible and Hopson, 1993; Rougier et al., (i.e., Ptilodus montanus, AMNH 35490). A 1996a, 1996c) have indicated that the tensor partially enclosed cavum supracochleare has tympani muscle occupied a deep recess in been reported previously for Vincelestes, iso- multituberculates (``ttf'' in ®gs. 14, 37A). lated petrosals from the Early Cretaceous of This remains true for most taxa, but a few Mongolia, and a few marsupials, and a com- (e.g., Pseudobolodon, V.J. 450-155, 460- pletely enclosed one was reported only in tri- 155) have a shallow tensor tympani fossa. conodontids, Tachyglossus, and most theri- Facial Ganglion Floor: In Rougier et al. ans (Wible et al., 1995; Rougier et al., (1996a, 1996c), we characterized the facial 1996a). ganglion ¯oor in multituberculates as a nar- Hypertrophied Vestibule: Miao (1988) re- row petrosal bridge beneath the primary fa- ported a greatly expanded vestibular appa- cial foramen. There are only a few forms in ratus in the Paleocene taxon Lambdopsalis which this morphology can be veri®ed, most and suggested it represented an adaptation to notably various isolated petrosals from the the perception of low-frequency vibrations, North American Late Cretaceous studied by as has been posited for certain extant verte- Luo (1989) and Wible and Hopson (1993, brates with enlarged vestibules. Luo and Ket- 1995). However, Kryptobaatar exhibits a dif- ten (1991) examined isolated petrosals from ferent morphology; the apparent con¯uence the North American Late Cretaceous Menis- of the facial canal with the canal for the ra- coessus and ``Catopsalis'' joyneri with CT mus inferior produces a broad petrosal bridge and reported a more than 10-fold expansion beneath the facial ganglion. It is uncertain in vestibular size over the condition in opos- how widespread this morphology is within sums and humans. They proposed that this Multituberculata. extraordinary in¯ation was a diagnostic fea- Cavum Supracochleare: Our analyses ture of known Cretaceous and Tertiary mul- (e.g., Wible and Hopson, 1993; Rougier et tituberculates and a possible synapomorphy al., 1996a, 1996c) have characterized multi- of Multituberculata. The latter is unlikely in tuberculates as having a cavum supracoch- light of Lillegraven and Hahn's (1993; see leare, which housed the facial (geniculate) also Hurum et al., 1996; Fox and Meng, 96 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247

1997) observation that the vestibule is not ENDOCRANIUM enlarged in the Late Jurassic paulchoffatiids. CAVUM EPIPTERICUM AND PRIMARY Moreover, we have seen no evidence of ves- BRAINCASE WALL tibular in¯ation from the outer contour of the petrosal in Kryptobaatar. Hurum et al. Among multituberculates, the structure of (1996) and Hurum (1998b) reported from the cavum epiptericum has been considered analyses of serial sections that the vestibule in detail in Kryptobaatar, Meketichoffatia in Nemegtbaatar and Chulsanbaatar is not krausei (Hahn, 1981), Lambdopsalis (Miao, much expanded, and Hurum (1998b: 90) 1988), Nemegtbaatar and Chulsanbaatar stated that the dimensions of the inner ear (Hurum, 1998a). In Kryptobaatar, the cavum ``are within the limits expected for an extant, epiptericum is an oval endocranial space out- small mammal.'' Therefore, the hypertro- side the primary wall of the braincase, pos- phied vestibule seen in Lambdopsalis (Miao, terolateral to the hypophyseal fossa (``ce'' in 1988), Meniscoessus, and ``Catopsalis'' joy- ®gs. 25B, 26, 27). Contributing to the walls neri (Luo and Ketten, 1991) most likely is a of the cavum epiptericum in Kryptobaatar unique adaptation of these taxa for low-fre- are: posteriorly, the petrosal; laterally, the an- quency hearing, as would be anticipated, for terior lamina and alisphenoid; medially, the orbitosphenoid, presphenoid, and basisphe- example, in burrowing forms. noid, as well as the ossi®ed pilae antotica and metoptica; dorsally, the orbitosphenoid and ossi®ed pilae antotica and metoptica; and HYPOGLOSSAL FORAMEN ventrally, the petrosal, alisphenoid, and perhaps maxilla. As reconstructed here, the ca- Most Mesozoic mammaliaforms and near vum epiptericum in Kryptobaatar contained outgroups have a single hypoglossal fora- the trigeminal and facial ganglia and the cav- men, although two are present in Adeloba- ernous sinus; traversing the cavum were the sileus (Luo, 1994). Hahn (1969) reported two divisions of the trigeminal nerve, the troch- small openings behind the jugular foramen lear and abducens nerves, the greater petrosal in Meketichoffatia krausei, both of which he nerve, the ramus infraorbitalis of the stape- termed condyloid foramina. He (see also dial artery, and the ophthalmic veins. Hahn, 1988) reconstructed the spinal acces- The cavum epiptericum in Lambdopsalis, sory nerve in the more anterior opening and Nemegtbaatar, and Chulsanbaatar resembles the hypoglossal nerve in the posterior one. that in Kryptobaatar in most regards. Differ- We are aware of no instance among extant ences exhibited by Lambdopsalis include the mammals where the spinal accessory nerve presence of a conspicuous recess for the fa- has such a course, and in forms with two cial ganglion in the posterior part of the cavum epiptericum; the large alisphenoid in the foramina (e.g., Didelphis, Wible, 1990) both lateral wall of the cavum epiptericum, en- transmit components of the hypoglossal larged at the expense of the anterior lamina; nerve. Among other multituberculates, a sin- and the exclusion from the cavum epipteri- gle hypoglossal foramen has been described cum of the ramus infraorbitalis at least in part for Ptilodus (Simpson, 1937) and Lambdop- by its course through an alisphenoid canal salis (Miao, 1988) and was illustrated for Ec- (but see above). Also, given the more pos- typodus (Sloan, 1979: ®g. 1). Among Mon- terior position of the metoptic foramen, the golian Late Cretaceous taxa, the hypoglossal oculomotor nerve in Lambdopsalis would foramen is reported to be either absent, con- have run through the front part of the cavum ¯uent with the jugular foramen (as in mono- epiptericum. tremes; ``jfo ϩ hyf'' in ®g. 28), or indiscern- Hahn (1981) illustrated and described the ible (Kielan-Jaworowska et al., 1986). Kryp- cavum epiptericum in a specimen of Pseu- tobaatar (PSS-MAE 101) is the exception, dobolodon oreas, which he later identi®ed with a small aperture in the appropriate place (Hahn, 1993) as Meketichoffatia krausei. Its that we have tentatively identi®ed as a hy- cavum epiptericum was shown to be a long, poglossal foramen. narrow, roo¯ess space lateral to the hypo- 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 97 physeal fossa and tuberculum sellae that end- though as discussed below this may be a ed anteriorly at the sphenorbital canal. As de- preservational artifact and is likely primitive- scribed by Hahn (1981), the absence of a ly present in Mammalia. Finally, perhaps the roof distinguishes the cavum of Meketichof- most striking difference is the presence of a fatia from that of Kryptobaatar, Lambdop- massive roof over the cavum epiptericum salis, Nemegtbaatar, and Chulsanbaatar in formed by the primary braincase in multitu- which an extensive roof is formed by the os- berculates, and its absence in all other mam- si®ed primary wall of the braincase. We have maliaforms known thus far, as noted by Kie- had the opportunity to study the basisphenoid lan-Jaworowska et al. (1986). According to of Meketichoffatia (V.J. 443-155) in light of Lucas and Luo (1993), only a very small that in Kryptobaatar. We think that a roof remnant of the primary wall, the slender os- formed by the primary wall was present, al- si®ed base of the pila antotica, occurs in Ade- though not as thick and extensive as in Kryp- lobasileus, Sinoconodon, Megazostrodon, tobaatar, Lambdopsalis, Nemegtbaatar, and Morganucodon, and monotremes. In the ®rst Chulsanbaatar; however, the roof has been three taxa, the pila antotica has only been damaged in Meketichoffatia through com- described from the ventral view, where it is pression. In fact, there is evidence of breaks visible through the open ¯oor of the cavum in what we interpret to be the medial edge epiptericum, and the pila may have been of the roof at the level of the anterior extent more substantial than thus far reported. How- of the hypophyseal fossa. Additionally, what ever, the massive ossi®ed pila antotica of Hahn (1981) described as the sphenorbital Kryptobaatar (``pan'' in ®gs. 25±27), Lamb- canal is really just the anterior part of the dopsalis, Nemegtbaatar, and Chulsanbaatar roofed cavum epiptericum. Also distinguish- exceeds any other ossi®ed pila reported thus ing the cavum of Meketichoffatia is a small far among cynodonts, suggesting that this is longitudinal canal running through the ¯oor, a secondary specialization of these multitu- within which Hahn (1981) placed a blood berculates. This thickening and extension of vessel and which Miao (1988) subsequently the pila antotica might be a derived feature identi®ed as an alisphenoid canal resembling for Multituberculata, of which the condition that purported for Lambdopsalis. We are un- in Meketichoffatia might represent an early certain of the nature of this canal in Meke- stage. tichoffatia. However, because of the canal's Considering that the common ancestor of length, extreme medial position, and small all mammals is thought (e.g., Starck, 1967, diameter, it is unlikely an alisphenoid canal. 1978; Kuhn, 1971; Zeller, 1989) to have had Finally, a metoptic foramen is not indicated a fully developed primary wall (the pilae in Meketichoffatia, but may have been dam- preoptica, metoptica, and antotica; ®g. 24A, aged along with the bulk of the roof of the B), the presence of these structures is ex- cavum epiptericum. pected at least in all pre-mammalian synap- The structure of the cavum epiptericum in sids if these elements are homologous across other mammaliaforms and near outgroups Amniota. However, the term pila refers to a differs in various ways from that in multi- chondrocranial structure that need not be os- tuberculates. First, there is only a partial ¯oor si®ed in the adult, as occurs in most adult below the cavum epiptericum in Adelobasi- sauropsids (De Beer, 1937; Bellairs and Ka- leus, Sinoconodon, Dinnetherium, Megazos- mal, 1981). In Mammaliaformes, the orbital trodon, and Morganucodon (Luo, 1994). region both externally and internally is rarely Second, the cavum supracochleare is sepa- preserved or accessible for preparation, ham- rated wholly or partially from the cavum pering the recognition of the presence of the epiptericum in triconodontids, isolated petro- primary wall and its foramina. Based on the sals from the Early Cretaceous Khoobur lo- presence of an optic foramen in the orbito- cality, Vincelestes, Tachyglossus, and theri- sphenoid of Probainognathus, Sinoconodon, ans (Rougier et al., 1996a). Third, a course and Adelobasileus, we expect the optic fo- for the oculomotor nerve separate from the ramina to have been present in the orbito- other contents of the cavum epiptericum is sphenoid of the mammaliaforms occupying a not known for any other mammaliaforms, al- phylogenetic position between these Triassic/ 98 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247

Early Jurassic forms and Mammalia. This in teriorly placed in the hypophyseal fossa in turn implies an independent loss of an ossi- Chulsanbaatar (Hurum, 1998a) and are de- ®ed pila metoptica in monotremes, marsupi- scribed by Miao (1988) as wholly lacking in als, and Vincelestes, or the absence of the Lambdopsalis. pila metoptica in the ancestry of multituber- Among Mesozoic mammaliaforms, the culates and therians, with the reacquisition of structure of the hypophyseal fossa is known a neomorphic pila metoptica in eutherians as only in Morganucodon and Triconodon.In suggested by Maier (1987) and in multitu- the former, as reconstructed from several in- berculates. Accepting the phylogenetic posi- complete basisphenoids by Kermack et al. tion for multituberculates (®g. 39) proposed (1981), the hypophyseal fossa is bordered by by Rowe (1988, 1993), Sereno and McKenna tall anterior and lateral walls, with the dor- (1995), Rougier et al. (1996a, 1996c), Hu et sum sellae being very low; the carotid fora- al. (1997), and Ji et al. (1999), the latter sce- mina enter the anteriormost aspect of the fos- nario is favored. The loss of the pila antotica sa's ¯oor. A single basisphenoid is known for in marsupials and eutherians can be consid- Triconodon (Kermack, 1963; Kermack et al., ered a synapomorphy (Rowe, 1988). The 1981); it shows a fairly shallow hypophyseal problem of the presence/absence of a metop- fossa with anteriorly placed carotid foramina tic foramen is more complex because even if in the ¯oor, but this specimen has been com- perfectly preserved in most cases, it will be pressed, hindering evaluation. Among extant con®ned deep within the cavum epiptericum mammals, the hypophyseal fossa is generally and hidden in lateral view by the secondary shallow, with the dorsum sellae being the braincase wall. As noted by Miao (1988), most prominent wall (Kuhn, 1971; Archer, given the presence of a metoptic foramen in 1976; Novacek, 1986; Zeller, 1989); the ca- multituberculates, it is likely that this fora- rotid foramina enter the posterior aspect of men was also present in pre-multituberculate the hypophyseal fossa in monotremes but mammaliaforms. tend to lie wholly lateral to the fossa in therians. HYPOPHYSEAL FOSSA,TUBERCULUM SELLAE, In addition to Kryptobaatar, the endocra- AND JUGUM SPHENOIDALE nial surface anterior to the hypophyseal fossa, that is, the tuberculum sellae and the ju- The entire structure of the hypophyseal gum sphenoidale (``tus'' and ``jsp'' in ®g. fossa is known for only ®ve multitubercu- 25A), has been described in detail only in lates: Kryptobaatar (``hf'' in ®gs. 25, 27), Meketichoffatia by Hahn (1981). The major Meketichoffatia krausei (originally described difference between Kryptobaatar and Meke- as Pseudobolodon oreas by Hahn, 1981), tichoffatia is the proportional shortening of Lambdopsalis (Miao, 1988), Nemegtbaatar, the mesocranium in the former. In Meketi- and Chulsanbaatar (Hurum, 1998a). In all choffatia, the tuberculum sellae, hypophyseal taxa, the hypophyseal fossa is a very deep fossa, and cavum epiptericum are long, excavation in the dorsal surface of the brain- whereas in Kryptobaatar these structures are case ¯oor, bordered laterally by tall walls shorter and broader (®g. 27). Damage to the formed by the primary wall of the braincase. area labeled as the fossa hypochiasmatica in The dorsum sellae, which forms the posterior Meketichoffatia (Hahn, 1981: ®g. 4) obscures wall, is also very prominent in Kryptobaatar the morphology of the rostral portion of the (``ds'' in ®gs. 25, 26), Meketichoffatia, Ne- tuberculum sellae; crushing has collapsed the megtbaatar, and Chulsanbaatar, but is low orbital wings over this region. This is the in Lambdopsalis. Well-developed carotid fo- area where the optic foramina and chiasmatic ramina enter the posterolateral aspect of the sulcus would be expected. Consequently, hypophyseal fossa in Kryptobaatar (``®ca'' crushing may account for the apparent ab- in ®gs. 25, 27), Nemegtbaatar, and, as inter- sence of structures in Meketichoffatia that we preted here, Meketichoffatia; Hahn (1981) in- have found in Kryptobaatar. correctly identi®ed the endocranial entrance The jugum sphenoidale in Meketichoffatia of the carotid as in the dorsum sellae in Mek- has a broad, shallow sulcus on the midline, etichoffatia. Carotid foramina are more an- whereas the jugum is narrow in Kryptobaa- 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 99 tar and even narrower in PSS-MAE 128. The been reported in a serially sectioned Ne- orbital wings in Meketichoffatia are broad megtbaatar (Kielan-Jaworowska et al., 1986; and continue anterodorsally into the lamina Hurum, 1994) based on tiny bone fragments infracribrosa, much as in Kryptobaatar. in the appropriate region. Kryptobaatar and Hahn (1981) describes the cranial aperture of Lambdopsalis (Miao, 1988), which show the ethmoidal foramen in Meketichoffatia, pristine preservation of delicate endocranial which seems to occupy a position similar to structures, have failed to provide any evi- that in Kryptobaatar. dence of an ossi®ed cribriform plate. How- ever, given the extant phylogenetic bracket CRIBRIFORM PLATE for multituberculates, the presence of a cartilaginous or ossi®ed cribriform plate is ex- An ossi®ed cribriform plate is ubiquitous pected. In fact, the presence of at least a car- in living mammals except for the platypus, tilaginous cribriform plate has been suggest- which is known to have a cartilaginous one ed in Diarthrognathus and Thrinaxodon that resorbs during ontogeny (Zeller, 1988). based on differential matrix preservation in The cribriform plate is seldom preserved in the nasal and cranial cavities (Crompton, fossils, because of its delicate nature and its 1958). Consequently, the presence of a crib- inaccessibility to preparation. Among multi- riform plate cartilaginous and/or osseous is tuberculates, a purported cribriform plate has likely an ancient trait in cynodont history.

CONCLUSIONS The cranial anatomy of Mongolian Late nasal fossa to enable description. The speci- Cretaceous multituberculates, in particular mens of K. dashzevegi studied by us provid- Nemegtbaatar and Chulsanbaatar, has pre- ed little new information on brain structure viously been evaluated in some detail in a to address the questions raised by Kielan-Ja- series of papers by Kielan-Jaworowska (e.g., worowska (1997) concerning neglected char- 1974) and various co-workers (e.g., Kielan- acters of the multituberculate brain. Jaworowska et al., 1986; Gambaryan and The exquisite state of preservation of Kielan-Jaworowska, 1995; Hurum et al., specimens of Kryptobaatar dashzevegi has 1996). In the tradition of these studies, our offered us a rare glimpse of details of cranial descriptions of Kryptobaatar dashzevegi pre- anatomy that are seldom preserved in an un- sent the ®rst bone-by-bone treatment of cra- distorted manner in Mesozoic mammali- nial anatomy in the most abundant Mongo- aforms, with Morganucodon (Kermack et al., lian Late Cretaceous multituberculate, which 1981) and Vincelestes (Rougier et al., 1992; is derived from specimens revealing most of Rougier, 1993; Hopson and Rougier, 1993) the osseous sutural relationships as well as being among the more noteworthy excep- most of the grooves, canals, and foramina tions. The anatomical details uncovered by relevant to the cranial nervous, arterial, and our study impact knowledge of the skull of venous systems. The only signi®cant com- multituberculates, and of Mesozoic mam- ponents of the skull of Kryptobaatar lacking maliaforms in general, in various ways. from our description are the nasal fossa and (1) We have been able to describe osseous paranasal sinuses. Yet, without recourse to elements poorly known or not known at all serial sectioning, which allowed the descrip- for other multituberculates or other Mesozoic tion of this region in Nemegtbaatar and mammaliaforms. Elsewhere (Rougier et al., Chulsanbaatar (Hurum, 1994), or high-res- 1996c), we have provided descriptions of the olution CT scanning, which has revealed de- stylohyal and partial stapes, malleus, and ec- tails of internal anatomy in the skull of, for totympanic, the ®rst for some of these ele- example, the primitive non-mammalian cy- ments among Multituberculata and for the nodont Thrinaxodon (Rowe et al., 1995), stylohyal among Mesozoic Mammaliafor- none of the available specimens of Krypto- mes. baatar exhibited suf®cient exposure of the Although partial jugals have been reported 100 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247 for Ptilodus, Nemegtbaatar, and Chulsan- have been posited for multituberculates (see, baatar (Hopson et al., 1989), those in Kryp- respectively, Kielan-Jaworowska, 1971; tobaatar are the ®rst complete jugals known Miao, 1988; Hahn and Hahn, 1994). These for multituberculates. In all instances, the ju- elements now appear to be ubiquitously ab- gal is a reduced element lying internal to the sent in multituberculates or, at the minimum maxilla and the squamosal in the zygomatic in the case of the prenasal process, not pre- arch. Among other mammaliaforms, mono- sent primitively in the group. Multitubercu- tremes resemble multituberculates in that the lates share the absence of an ectopterygoid jugal is replaced in the zygomatic arch by the with other mammaliaforms except apparently maxilla and the squamosal (Hopson et al., Morganucodon (Kermack et al., 1981), the 1989). However, the pattern of replacement absence of a prenasal process with other differs somewhat; in the platypus, the re- mammals (Wible, 1991), and the absence of duced jugal is situated dorsal to the maxilla a septomaxilla with marsupials and placen- and the squamosal (Zeller, 1989), and in the tals (Wible et al., 1990), with the possible echidna, the jugal is wholly absent (Kuhn, exception of some xenarthrans (Zeller et al., 1971). 1993). The orbital mosaic and foramina and the Kryptobaatar also clearly lacks an orbital endocranium were largely unknown in mul- exposure of the palatine, which may repre- tituberculates prior to our study, with the ma- sent a synapomorphy of multituberculates as jor exceptions being Lambdopsalis (Miao, suggested by Miao (1988, 1993). Hurum 1988), Nemegtbaatar, and Chulsanbaatar (1994, 1998a) argued against this interpre- (Hurum, 1994, 1998a), and were also poorly tation, in light of his description of the pal- known in other Mesozoic mammaliaforms. atine in the orbit of Nemegtbaatar. However, We have con®rmed Miao's (1988) observa- it appears that Hurum's interpretation of Ne- tion in Lambdopsalis of an optic foramen, a megtbaatar, if correct, is unusual among metoptic foramen, and an extensive, thick os- multituberculates. si®ed primary wall of the braincase including (3) We have been able to con®rm the pres- the pilae metoptica and antotica; the latter ence of several elements in the multituber- two also occur in Nemegtbaatar and Chul- culate skull whose incidence has been con- sanbaatar (Hurum, 1998a). The presence of troversial. Miao (1988, 1993) questioned these structures in these taxa, as well as em- whether any multituberculate really has an bryological evidence, suggests that the optic alisphenoid with a reduced contribution to foramen, the metoptic foramen, and an ex- the lateral wall of the braincase, as initially tensive ossi®ed primary wall may be more described by Kermack and Kielan-Jawo- widespread among Multituberculata and rowska (1971) and Kielan-Jaworowska Mammaliaformes. The robust nature of the (1971). Kryptobaatar clearly has such a re- pila antotica in Kryptobaatar, Lambdopsalis, duced alisphenoid, and its presence con®rms Nemegtbaatar, Chulsanbaatar, and perhaps the tentative observations of such an element Meketichoffatia may be a derived feature of in other Mongolian Late Cretaceous taxa Multituberculata, because the ossi®ed rem- (Kielan-Jaworowska, 1971; Kielan-Jawo- nants of the pila in the outgroups are not as rowska et al., 1986). Kryptobaatar also clear- substantive. We have also described for the ly shows that the anterior lamina is expanded ®rst time in multituberculates a foramen for forward dorsal to the alisphenoid, again con- the pituito-orbital vein, unknown in other ®rming previous observations in other Mon- mammaliaforms, and a transverse canal, oc- golian Late Cretaceous taxa (Kielan-Jawo- curring only in some metatherians and pla- rowska, 1971; Kielan-Jaworowska et al., centals (McDowell, 1958; Archer, 1976). 1986; Hurum, 1998a). Both of these ele- (2) We have been able to con®rm the ab- ments, a reduced alisphenoid and expanded sence of several elements from the multitu- anterior lamina, are shared by Mongolian berculate skull whose incidence has been Late Cretaceous multituberculates and mono- controversial. Kryptobaatar clearly lacks an tremes (Wible and Hopson, 1993). Moreover, ectopterygoid, a prenasal process of the pre- these characters have been used to support a maxilla, and a septomaxilla, all of which monotreme-multituberculate clade (Wible 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 101 and Hopson, 1993; Meng and Wyss, 1995), canal in therians (Novacek, 1986). Therefore, which more recent analyses, some of which we deem it likely that the palatal aperture in include this feature, do not support (®g. 39; question in Kryptobaatar is a true foramen Rougier, 1996a, 1996c; Hu et al., 1997; Ji et that transmitted the minor palatine nerve and al., 1999). vessels, as occurs in most mammaliaforms (4) We have been able to resolve compet- including other multituberculates (except ing anatomical hypotheses for several ele- possibly Nemegtbaatar). ments of the multituberculate and mamma- Most Mesozoic mammaliaforms, includ- liaform skull. The troughs lateral to the pter- ing multituberculates, have multiple aper- ygopalatine ridges that occur in Kryptobaa- tures in and around the anterior lamina that tar and most Mesozoic mammaliaforms have have been interpreted as having transmitted been said to have accommodated the audi- branches of the trigeminal nerve (Luo, tory tube (Barghusen, 1986) or to have pro- 1994). Two models for the nerves in these vided muscle attachment (Kielan-Jaworow- apertures have been proposed: branches of ska, 1971; Hahn, 1981; Gambaryan and Kie- the mandibular nerve transmitted in multi- lan-Jaworowska, 1995). The morphology in tuberculates (Simpson, 1937; Kielan-Jawo- Kryptobaatar eliminates the auditory tube as rowska, 1971), and the maxillary and man- a possible occupant, because the troughs ex- dibular nerves transmitted as in other mam- tend too far anteriorly into the nasal cavity maliaforms (Patterson and Olson, 1961; Ker- rather than ending in the nasopharynx, the mack et al., 1981). The morphology in site of origin for the auditory tube in extant Kryptobaatar and the vast majority of mul- mammals (Starck, 1995). We think that the tituberculates support the former restoration lateral troughs provided muscle attachment for multituberculates, because the foramina and/or an accessory route for air in Krypto- in question are at the same anteroposterior baatar and other Mesozoic mammaliaforms. level, which is not the pattern that the max- Miao (1988) has argued that the element illary and mandibular nerve exits exhibit in identi®ed (e.g., Kielan-Jaworowska et al., any extant mammal. The implications of this 1986) as a postorbital process on the parietal for the reconstruction of nerves in other Me- in Mongolian Late Cretaceous multitubercu- sozoic mammaliaforms are yet to be ex- lates is not a true postorbital process, which, plored, but at a minimum legitimate concerns he suggested, is reduced or absent in Multi- have been raised (see also Rougier et al., tuberculata. However, the process on the pa- 1996a). rietal in question in Kryptobaatar has the Kielan-Jaworowska et al. (1986) suggested same positional relationship to the frontal di- that the enormous jugular fossa in Mongolian ploic vein and the supraorbital nerves and Late Cretaceous multituberculates housed vessels, as does the true postorbital process large ganglia on the cranial nerves below the on the frontal in extant mammals. Conse- jugular foramen. The exceedingly large di- quently, the postorbital process, which is in- mensions of the jugular fossa and its con- conspicuous and on the frontal in other mul- nection with the middle-ear cavity in Kryp- tituberculates, has shifted onto the parietal tobaatar suggest that this recess housed a di- and become enlarged in Mongolian Late Cre- verticulum of the cavum tympani (Rougier et taceous taxa. al., 1996c). We think that this interpretation Kielan-Jaworowska et al. (1986) have ar- can be extended to other multituberculates gued, on the basis of serial sections of Ne- with large jugular fossae. megtbaatar, that the aperture at the back of (5) We have been able to reconstruct the the palate identi®ed (e.g., Kielan-Jaworows- major components of both the cranial arterial ka, 1971) as a minor palatine (palatonasal) and venous patterns in Kryptobaatar. The foramen in multituberculates is merely a cranial vascular system of Kryptobaatar in notch that does not fully penetrate bone. general resembled that restored previously Along with a minor palatine foramen, we for multituberculates (e.g., Kielan-Jaworow- have found a second aperture in the orbital ska et al., 1986; Miao, 1988; Wible and Hop- process of the maxilla in Kryptobaatar re- son, 1995) or for other Mesozoic mammali- sembling that leading into the minor palatine aforms (e.g., Rougier et al., 1992). However, 102 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247 the specimens of Kryptobaatar have provid- terolaterally to reach the fenestra vestibuli, ed insight into the course of several vessels whereas it had a lateral or anterolateral previously unreconstructed for multituber- course in the remaining taxa. culates. Kryptobaatar clearly had a trans- The cranial venous system of Kryptobaa- promontorial internal carotid artery, and tar and other multituberculates resembled based upon this observation, we have dis- that in monotremes, with the primary exits covered transpromontorial carotid grooves in from the cranial cavity being the prootic ca- a variety of other multituberculates. A nal and the foramen magnum. However, groove for a transpromontorial internal ca- Kryptobaatar is unique among mammali- rotid is a feature that multituberculates share aforms known to date in having a separate with Vincelestes (Rougier et al., 1992) and foramen for the pituito-orbital vein and is certain eutherians (Wible, 1986), but it is ab- unique among non-therians in having a trans- sent in basal mammaliaforms, symmetro- verse canal vein. donts, triconodontids, monotremes, and mar- Lambdopsalis, whose cranial anatomy has supials. Consequently, a transpromontorial been described in detail (Miao, 1988), has sulcus has been acquired convergently in served as a model for Multituberculata in multituberculates and prototribosphenidans. several phylogenetic analyses (e.g., Wible, The endocranial aperture of the carotid fo- 1991; Meng, 1992; Meng and Wyss, 1995), ramen is in the posterolateral aspect of the despite this taxon's extreme specializations deep hypophyseal fossa in Kryptobaatar, for burrowing and numerous apomorphic Pseudobolodon (Hahn, 1981), and Nemegt- cranial features compared with Mongolian baatar (Kielan-Jaworowska et al., 1986; Hu- Late Cretaceous taxa (Miao, 1988; Luo, rum, 1998a), whereas it is more anteriorly 1996). A goal of our report on Kryptobaatar placed in Chulsanbaatar (Hurum, 1998a). has been to provide a detailed description of Our comparisons with other Mesozoic mam- the skull and lower jaws of a less specialized maliaforms have led us to an additional ob- form that can serve as an appropriate model servation of potential phylogenetic impact: for future comparative morphological and the endocranial aperture of the carotid fora- phylogenetic studies on Multituberculata and men is in the hypophyseal fossa in most taxa on Mammaliaformes in general. In most as- with the exception of marsupials and placen- pects of its cranial anatomy, Kryptobaatar, tals, in which it is wholly lateral to the fossa. and not Lambdopsalis, resembles the vast The positions of the groove for the stape- majority of other multituberculates. Cranial dial artery and the foramina for its end features that Kryptobaatar shares with most branches, along with the partial bicrurate sta- multituberculates, but which differ in Lamb- pes described for Kryptobaatar elsewhere dopsalis, include: the absence of the prenasal (Rougier et al., 1996c), support a course for process of the premaxilla; an anterior nasal the artery through the stapes. We have found notch; fewer nasal foramina; the presence of a similar arrangement of sulci and foramina a lacrimal; a more extensive orbital exposure in a wide variety of other multituberculates, of the orbitosphenoid; a foramen for the fron- with the exception being Lambdopsalis,in tal diploic vein; the anterior opening of the which the artery ran posterodorsal to the oval orbitotemporal canal below the postorbital window (Miao, 1988), as in the platypus process; the presence of a jugal; the presence (Wible, 1987). Consequently, the vast major- of pterygopalatine ridges; a reduced alisphe- ity of multituberculates exhibited the same noid; mandibular nerve foramina in the an- relationship between the stapedial artery and terior lamina; a transpromontorial internal stapes as in all extant eutherians in which the carotid artery; carotid foramina in the pos- artery is present in the adult (Wible, 1987). terolateral hypophyseal fossa; a stapedial ar- The same pathway has been inferred for Vin- tery that ran through a bicrurate stapes; a su- celestes (Rougier et al., 1992). Multituber- praglenoid foramen opening at a level ante- culates differ, however, from Vincelestes and rior to the oval window; the posterior open- most eutherians in the course of the stapedial ing of the posttemporal canal low on the artery beneath the promontorium of the pe- occiput; an epitympanic recess on the lateral trosal. The multituberculate vessel ran pos- ¯ange; a crista parotica continuous with the 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 103 lateral ¯ange; a caudal tympanic process of in other multituberculates). A large roughly the petrosal; a triangular paroccipital process; rectangular facial surface of the lacrimal, ex- a nonhypertrophied vestibule; and a tall dor- posed on the cranial roof, separating the sum sellae. In contrast, there are only a few frontal from the maxilla is also characteristic features in which Lambdopsalis, and not for Djadochtatheria, but this may be a ple- Kryptobaatar, exhibits the likely primitive siomorphy.'' Of these proposed synapomor- multituberculate condition. These include: an phies, only two are found exclusively in dja- inconspicuous postorbital process on the dochtatherians: the frontoparietal suture is V- frontal; the absence of a buccinator foramen shaped in non-djadochtatherians, including between the anterior lamina and alisphenoid; the paulchoffatiid Pseudobolodon (V.J. the absence of a transverse canal foramen; 447.155, 451-1255, 460-155), Ptilodus and a large jugular foramen that transmitted (Simpson, 1937), and taeniolabidids (Grang- the sigmoid sinus. er and Simpson, 1929; Miao, 1988); and the In our descriptions of Kryptobaatar and facial exposure of the lacrimal is arcuate in comparisons with other multituberculates, a non-djadochtatherians, including the paul- number of features of the cranial anatomy choffatiids Kuehneodon (V.J. 454-155), Mek- have been identi®ed that are shared at vari- etichoffatia (V.J. 446-155), and Pseudobolo- ous levels within Multituberculata. Evaluat- don (Hahn and Hahn, 1994; V.J. 460-155). ing the phylogenetic signi®cance of these However, given that a V-shaped process of features both within Multituberculata and the frontals also contacts the nasals in the within Mammaliaformes is beyond the scope paulchoffatiids Kuehneodon (V.J. 454-155), of the current report and awaits an in-depth Meketichoffatia (V.J. 446-155), and Pseudo- analysis of multituberculate interrelation- bolodon (V.J. 447.155, 451-1255, 460-155), ships. this purported djadochtatherian feature may In lieu of a such an analysis, we compare be plesiomorphous for Multituberculata. In our descriptions of Kryptobaatar to the re- addition to djadochtatherians, the sharp edge cent diagnoses of Djadochtatheria, Djadoch- between the lateral and palatal walls of the tatheriidae, and Kryptobaatar published by premaxilla occurs in Lambdopsalis (Miao, Kielan-Jaworowska and Hurum (1997), 1988). based on their cladistic analysis of 43 char- For Djadochtatheriidae, Kielan-Jaworow- acters, (18 dental and 25 cranial). Djadoch- ska and Hurum (1997: 208) cited the follow- tatheria includes 10 of the 11 Mongolian ing: ``. . . differs from all other multituber- Late Cretaceous multituberculate genera culates (and all other mammals) in having a (with the exception being Buginbaatar) and subtrapezoidal snout in dorsal view, with the tentatively 2 North American genera (Para- anterior and lateral margins con¯uent with cimexomys and Pentacosmodon). Djadoch- the zygomatic arches rather than incurved in tatheriidae includes Kryptobaatar, Djadoch- front of the arches . . . [differs] from other tatherium, Catopsbaatar, and Tombaatar. members of Djadochtatheria in having the Rougier et al. (1997a) have also placed Kryp- snout extending for 50% or more of the skull tobaatar within a monophyletic Mongolian length, anterior part of promontorium (sensu Late Cretaceous clade, which also includes Hurum et al., 1996) irregular, with incurva- Pentacosmodon (®g. 38). However, the char- tures on both sides rather than oval (a characters supporting the interrelationships of acter shared with Lambdopsalis).'' As noted these taxa are largely from the dentition and, by Kielan-Jaworowska and Hurum (1997), therefore, less relevant to our consideration the features of the snout distinguish Krypto- of the cranial anatomy of Kryptobaatar. baatar, Djadochtatherium, and Catopsbaatar For Djadochtatheria, Kielan-Jaworowska from other multituberculates, but the prom- and Hurum (1997: 206) cited the following ontorial feature is known only for Krypto- cranial synapomorphies: ``large frontals, baatar and Catopsbaatar. None of these dja- pointed anteriorly in the middle, deeply in- dochtatheriid characters has yet been de- serted between the nasals; U-shaped fronto- scribed for Tombaatar. parietal suture; a sharp edge between the lat- For Kryptobaatar, Kielan-Jaworowska and eral and palatal walls of premaxilla (rounded Hurum (1997: 208) provided the following 104 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247 revised diagnosis: ``The smallest member of rior opening into the posttemporal canal is the Djadochtatheriidae nov. (skull length 25± within the petrosal. Differs from Chulsan- 32 mm), most similar to Djadochtatherium, baatar and Sloanbaatar in that the supragle- with which it shares an arcuate (rather than noid foramen is at a level anterior to the fe- trapezoidal as in Catopsbaatar) p4. The p4 nestra vestibuli. Resembles Sloanbaatar in in Kryptobaatar is relatively longer than in having I2s directed ventromedially and in Djadochtatherium and has eight serrations contact with each other, leaving a triangular (number unknown in Djadochtatherium). open space between them and the edge of the Differs from Catopsbaatar in having smaller premaxilla. Resembles Djadochtatherium facial surface of the lacrimal; differs from and Catopsbaatar in having a long postor- Djadochtatherium and Catopsbaatar in hav- bital process on the parietal (contra Kielan- ing a shorter postorbital process. Shares with Jaworowska and Hurum, 1997). Resembles MLCM [Mongolian Late Cretaceous multi- Nemegtbaatar and Sloanbaatar in the ab- tuberculates] except Tombaatar the alveolus sence of a conspicuous recessus scalae tym- for I3 formed by premaxilla, rather than by pani. Resembles Nemegtbaatar and Chulsan- both premaxilla and maxilla (autapomorphy baatar in having a sigmoid sinus that exits of Tombaatar). Differs from Catopsbaatar via the foramen magnum and not the jugular and Tombaatar in having four upper pre- foramen. Resembles Kamptobaatar and molars. Differs from Catopsbaatar in having Chulsanbaatar in having a foramen of the an inner row of cusps in M1 extending for dorsal ascending canal. Resembles Kampto- less than a half, or a half of the tooth length, baatar, Nemegtbaatar, Chulsanbaatar, and rather than a long row; shares short inner row Catopsbaatar in having both the foramen of cusps in M1 with Tombaatar and the ovale inferium and foramen masticatorium North American Paracimexomys (Lillegrav- traverse the epitympanic recess. en 1969; Archibald 1982). Differs from Chulsanbaatar and Sloanbaatar in having ACKNOWLEDGMENTS cusps on the inner ridge in M1, rather than a smooth ridge. Differs from Tombaatar in For the great privilege of studying the ex- having M1 cusp formula 4±5:4:3±5 rather quisite MAE skulls of Kryptobaatar, we are than 5:5:2. Differs from Djadochtatherium grateful to Dr. Demberelyin Dashzeveg of and Catopsbaatar in having a shorter snout the Geological Institute of the Mongolian (but longer than in non-djadochtatheriid dja- Academy of Sciences, Ulaan Baatar, Mon- dochtatherians), a relatively less robust lower golia, and Dr. Michael J. Novacek of the incisor, and less prominent masseteric and American Museum of Natural History. For parietal crests.'' the high-resolution CT scans of Kryptobaa- Our additions and amendments to their re- tar, we thank Dr. Timothy Rowe of the Uni- vised diagnosis follows: Differs from other versity of Texas, Austin. For access to com- djadochtatherians in having either con¯uent parative materials, we thank G. Hahn, Institut carotid and pterygoid canals or a very short fuÈr Geologie und PalaÈontologie, Phillipps- separation between the two, a separate hy- UniversitaÈt, Marburg; J. A. Hopson, Univer- poglossal foramen, a foramen for the pituito- sity of Chicago, Chicago; Z. Kielan-Jawo- orbital vein, and a transverse canal. Differs rowska, Institute of Paleobiology, Polish from Nemegtbaatar in that the anterior open- Academy of Sciences, Warsaw; B. Krebs and ing of the posttemporal canal is between the T. Martin, Institut fuÈr PalaÈontologie, Freie parietal, frontal, and anterior lamina. Differs UniversitaÈt, Berlin; R.D.E. MacPhee and from Nemegtbaatar, Chulsanbaatar, and N.B. Simmons, Department of Mammalogy, Kamptobaatar in having a smaller orbital ex- AMNH; W. Maier, Lehrstuhl fuÈr Spezielle posure of the orbitosphenoid. Differs from Zoologie, Karl-UniversitaÈt, TuÈbingen; M.J. Nemegtbaatar, Chulsanbaatar, and Catops- Novacek and M.C. McKenna, Department of baatar in that the anterior and intermediate Vertebrate Paleontology, AMNH; and U. zygomatic ridges are con¯uent and the pos- Zeller, Museum fuÈr Naturkunde, Berlin. For terior ridge is absent. Differs from Nemegt- amassing the vast collection of Kryptobaatar baatar and Kamptobaatar in that the poste- specimens described above, we thank all the 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 105 members of the MAE ®eld crews. The de- version of this manuscript, we are extremely scription of Kryptobaatar was made possible grateful to E.F. Allin, J.A. Hopson, Z. Kie- through the skilled preparation completed by lan-Jaworowska, J. Meng, and R. Presley. Amy Davidson at the AMNH. Also at the This research was supported by NSF Grants AMNH, Edward Heck produced ®gures 1± BSR 91-19212, DEB 93-0070, DEB 94- 23 and 29±37, Lorraine Meeker produced 0799, DEB 95-27811, DEB 96-25431, DEB ®gures 25±26, and G.W.R. produced the re- 99-96051, and DEB 99-96172. GRW's re- mainder. Discussions over the years with E.F. search has also been supported by a Ralph Allin, G. Hahn, J.A. Hopson, J. Hurum, Z. E. Powe Junior Faculty Award from Oak Kielan-Jaworowska, Z. Luo, J. Meng, D. Ridge Associated Universities. Finally, we Miao, M.J. Novacek, and R. Presley have en- acknowledge the extreme patience and un- hanced our comprehension of the multitu- derstanding of our spouses, Stephanie Davis berculate skull. For comments on an earlier and Adela Reale.

REFERENCES Archer, M. 1914. On the structure and af®nities of the 1976. The basicranial region of marsupicar- Multituberculata. Bull. Am. Mus. Nat. nivores (Marsupialia), interrelation- Hist. 33: 115±134. ships of carnivorous marsupials, and Bruner, H. L. af®nities of the insectivorous perame- 1907. On the cephalic veins and sinuses of lids. Zool. J. Linn. Soc. 59: 217±322. reptiles, with description of a mecha- Archibald, J. D. nism for raising the venous blood-pres- 1982. A study of Mammalia and geology sure in the head. Am. J. Anat. 7: 1±117. across the Cretaceous±Tertiary bound- Bryant, H. N., and A. P. Russell ary in Gar®eld County, Montana. Univ. 1992. The role of phylogenetic analysis in the California Publ. Geol. Sci. 122: 1±286. inference of unpreserved attributes of Averianov, A. O. extinct taxa. Philos. Trans. R. Soc. Lon- 1997. New Late Cretaceous mammals of don B337: 405±418. southern Kazakhstan. Acta Palaeontol. Bugge, J. Pol. 42: 243±256. 1971. The cephalic arterial system in mole- Barghusen, H. R. rats (Spalacidae), bamboo rats (Rhizo- 1986. On the evolutionary origin of the ther- myidae), jumping mice and jerboas ian tensor veli palatini and tensor tym- (Dipodoidea) and dormice (Gliroidea) pani muscles. In N. Hotton III, P. D. with special reference to the systematic MacLean, J. J. Roth, and E. C. Roth classi®cation of rodents. Acta Anat. 79: (eds.), The ecology and biology of 165±180. mammal-like reptiles: 253±262. Wash- 1974. The cephalic arterial system in insecti- ington, D.C.: Smithsonian Institution vores, primates, rodents and lago- Press. morphs, with special reference to the Bellairs, A. d'A., and A. M. Kamal systematic classi®cation. Ibid. 87 1981. The chondrocranium and the develop- (suppl. 62): 1±159. ment of the skull in recent reptiles. In 1978. The cephalic arterial system in carni- C. Gans and T. S. Parsons (eds.), Bi- vores, with special reference to the sys- ology of the Reptilia: 1±263. London: tematic classi®cation. Ibid. 101: 45±61. Academic Press. 1979. Cephalic arterial pattern in New World Bonaparte, J. F. edentates and Old World pangolins 1986. Sobre Mesungulatum houssayi y nue- with special reference to their phylo- vos mamõÂferos CretaÂcicos de Patagon- genetic relationships and taxonomy. ia. Actas IV Congr. Argent. Paleontol. Ibid. 105: 37±46. 2: 48±61. Butler, H. Broom, R. 1967. The development of the mammalian 1909. Observations on the development of dural venous sinuses with especial ref- the marsupial skull. Proc. Linn. Soc. erence to the post-glenoid vein. J. Anat. N.S.W. 34: 195±214. 102: 33±56. 106 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247

Clemens, W. A., and Z. Kielan-Jaworowska vertebrate assemblage from the Late 1979. Multituberculata. In J. A. Lillegraven, Cretaceous of Mongolia. Nature 374: Z. Kielan-Jaworowska, and W. A. 446±449. Clemens (eds.), Mesozoic mammals, De Beer, G. R. the ®rst two-thirds of mammalian his- 1926. Studies on the vertebrate head. II. The tory: 99±149. Berkeley: Univ. Califor- orbitotemporal region of the skull. Q. J. nia Press. Microsc. Sci. 70: 263±370. Clemens, W. A., and J. A. Lillegraven 1937. The development of the vertebrate 1986. New Late Cretaceous, North American skull. Oxford: Clarendon Press. advanced therian mammals that ®t nei- Denison, W., and R. J. Terry ther the marsupial nor eutherian molds. 1921. The chondrocranium of Caluromys. In K. M. Flanagan and J. A. Lillegrav- Washington Univ. Stud. 8: 161±181. en (eds.), Vertebrates, phylogeny, and Eisthen, H. L. philosophy. Contrib. Geol. Univ. Wy- 1992. Phylogeny of the vomeronasal system oming Spec. Pap. 3: 55±85. and of receptor cell types in the olfac- Conroy, G. C, and J. R. Wible tory and vomeronasal epithelia of ver- 1978. Middle ear morphology of Lemur var- tebrates. Microsc. Res. Tech. 23: 1±21. iegatus: implications for primate pale- Ellenberger, W., and H. Baum ontology. Folia Primatol. 29: 81±85. Cooper, G, and A. L. Schiller 1908. Handbuch der vergleichenden Anato- 1975. Anatomy of the guinea pig. Cambridge, mie der Haustiere. Berlin: Verlag Von MA: Harvard Univ. Press. August Hirschwald. Cope, E. D. Evans, H. E., and G. C. Christensen 1880. On the foramina perforating the poste- 1979. Anatomy of the dog. Philadelphia: W. rior part of the squamosal bone of the B. Saunders. Mammalia. Proc. Am. Philos. Soc. 18: Fischer, M. S. 452±461. 1986. Die Stellung der Schliefer (Hyracoidea) Cords, E. im phylogenestischen System der Eu- 1915. UÈ ber das Primordialcranium von Per- theria. Cour. Forschungsinst. Senckenb. ameles spec.? unter BeruÈcksichtigung 84: 1±132. der Deckknochen. Anat. Hefte 52: 1± Fourie, S. 84. 1974. The cranial morphology of Thrinaxo- Crompton, A. W. don liorhinus Seeley. Ann. S. Afr. Mus. 1958. The cranial morphology of a new genus 65: 337±400. and species of ictidosaurian. Proc. Fox, R. C., and J. Meng Zool. Soc. London 130: 183±216. 1997. An X-radiographic and SEM study of Crompton, A. W., and W. L. Hylander the osseous inner ear of multitubercu- 1986. Changes in mandibular function fol- lates and monotremes (Mammalia): im- lowing the acquisition of a dentary- plications for mammalian phylogeny squamosal jaw articulation. In N. Hot- and evolution of hearing. Zool. J. Linn. ton III, P. D. MacLean, J. J. Roth, and Soc. 121: 249±291. E. C. Roth (eds.), The ecology and bi- Freeman, E. F. ology of mammal-like reptiles: 263± 1976. A mammalian fossil from the Forest 282. Washington, D.C.: Smithsonian Marble (Middle Jurassic) of Dorset. Institution Press. Proc. Geol. Assoc. 87: 231±235. Crompton, A. W., and Z. Luo 1979. A Middle Jurassic mammal bed from 1993. Relationships of the Liassic mammals Sinoconodon, Morganucodon oehleri, Oxfordshire. Palaeontology 22: 135± and Dinnetherium. In F. S. Szalay, M. 166. J. Novacek, and M. C. McKenna (eds.), Gambaryan, P. P., and Z. Kielan-Jaworowska Mammal phylogeny: Mesozoic differ- 1995. Masticatory musculature of Asian tae- entiation, multituberculates, mono- niolabidoid multituberculate mammals. tremes, early therians, and marsupials: Acta Palaeontol. Pol. 40: 45±108. 30±44. New York: Springer. Gaupp, E. Dashzeveg, D., M. J. Novacek, M. A. Norell, J. 1902. UÈ ber die Ala temporalis des SaÈugetier- M. Clark, L. M. Chiappe, A. Davidson, schaÈdels und die Regio orbitalis einiger M. C. McKenna, L. Dingus, C. Swish- anderer WirbeltierschaÈdels. Anat. Hefte er, and A. Perle 19: 155±230. 1995. Extraordinary preservation in a new 1905. Neue Deutungen auf dem Gebiete der 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 107

Lehre vom SaÈugetierschaÈdel. Anat. 684. Reimpression 1967. Amsterdam: Anz. 27: 273±310. Ascher. 1908. Zur Entwicklungsgeschichte und ver- Hahn, G. gleichenden Morphologie des SchaÈdels 1969. BeitraÈge zur Fauna der Grube Guimar- von Echidna aculeata var. typica. Se- ota Nr. 3. Die Multituberculata. Pa- mon Zool. Forschungsreisen in Aus- laeontogr. Abt. A 133: 1±100. tralien 6: 539±788. 1977. Neue SchaÈdel-Reste von Multitubercu- Gauthier, J. A., A. G. Kluge, and T. Rowe laten (Mamm.) aus dem Malm Portu- 1988. Amniote phylogeny and the importance gals. Geol. Palaeontol. 11: 161±186. of fossils. Cladistics 4: 105±208. 1978. Milch-Bezahnungen von Paulchoffati- Gelderen, C. van idae (Multituberculata; Ober-Jura). 1924. Die Morphologie der Sinus durae ma- Neues Jahrb. Geol. PalaÈontol. Mh. tris. Zweiter Teil. Die vergleichende 1978: 25±34. Ontogenie der neurokraniellen Venen 1981. Zum Bau der SchaÈdel-Basis bei den der VoÈgel und SaÈugetiere. Z. Anat.- Paulchoffatiidae (Multituberculata; Entwickelungsgesch. 74: 432±508. Ober-Jura). Senckenb. Lethaea 61: Gidley, J. W. 227±245. 1909. Notes on the fossil mammalian genus 1983. UÈ berblick uÈber die Erforschungs-Ges- Ptilodus, with descriptions of a new chichte der Multituberculaten (Mam- species. Proc. U.S. Natl. Mus. 36: 611± malia). Schriftenr. Geol. Wiss. 19/20: 626. 217±246. Gillilan, L. A. 1985. Zum Bau des Infraorbital-Foramens bei 1972. Blood supply to primitive mammalian den Paulchoffatiidae (Multituberculata, brains. J. Comp. Neurol. 145: 209±222. Ober Jura). Berl. Geowiss. Abh. 60: 5± Godefroit, P. 27. 1997. Reptilian, therapsid and mammalian 1987. Neue Beobachtungen zum SchaÈdel- teeth from the Upper Triassic of Var- angeÂville (northeastern France). Bull. und Gebiss-Bau der Paulchoffatiidae Inst. R. Sci. Nat. Belg. Sci. Terre 67: (Multituberculata, Ober-Jura). Palaeo- 83±102. vertebr. Montpellier 17: 155±196. Granger, W., and G. G. Simpson 1988. Die Ohr-region der Paulchoffatiidae 1929. A revision of the Tertiary Multituber- (Multituberculata, Ober-Jura). Ibid. 18: culata. Bull. Am. Mus. Nat. Hist. 56: 155±185. 601±676. 1993. The systematic arrangement of the Green, H. L. Paulchoffatiidae (Multituberculata) re- 1930. A description of the egg-tooth of Or- visited. Geol. Palaeontol. 27: 201±214. nithorhynchus, together with some Hahn, G., and R. Hahn notes on the development of the pala- 1994. Nachweis des Septomaxillare bei Pseu- tine processes of the premaxillae. J. dobolodon krebsi n. sp. (Multitubercu- Anat. 64: 512±522. lata) aus dem Malm Portugals. Berl. Greenwald, N. S. Geowiss. Abh. E13: 9±29. 1987. Patterns of tooth eruption and replace- 1999. Neue Beobachtungen an Plagiaulacoi- ment in multituberculate mammals. J. dea (Multituberculata) des Ober-Juras. Vertebr. Paleontol. 8: 265±277. 1. Zum Zahn-Wechsel bei Kielanodon. Gregory, W. K. Ibid. E28: 1±7. 1910. The orders of mammals. Bull. Am. Hahn, G., J. C. Lepage, and G. Wouters Mus. Nat. Hist. 27: 1±524. 1987. Ein Multituberculaten-Zahn aus der Grif®ths, M. Ober-Trias von Gaume (S-Belgien). 1978. The biology of monotremes. New Bull. Soc. Belg. GeÂol. 96: 39±47. York: Academic Press. Hill, J. E. Hafferl, A. 1935. The cranial foramina in rodents. J. 1921. Zur Entwickelungsgeschichte der Mammal. 16: 121±129. KopfgefaÈsse des Gecko. (Platydactylus Hill, J. P., and G. R. De Beer annularis.). Anat. Hefte 59: 1±42. 1949. Development of the Monotremata. Part 1933. Das Arteriensystem. In L. Bolk, E. VII. The development and structure of GoÈppert, E. Kallius, and W. Lubosch the egg-tooth and caruncle in the (eds.), Handbuch der vergleichenden monotremes and on the occurrence of Anatomie der Wirbeltiere, vol. 6: 563± vestiges of the egg-tooth and caruncle 108 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247

in marsupials. Trans. Zool. Soc. Lon- Jenkins, F. A., Jr., and C. R. Schaff don 26: 503±544. 1988. The Early Cretaceous mammal Gobi- Hillenius, W. J. conodon (Mammalia, Triconodonta) 1994. Turbinates in therapsids: evidence for from the Cloverly Formation in Mon- Late Permian origins of mammalian en- tana. J. Vertebr. Paleontol. 8: 1±24. dothermy. Evolution 48: 207±229. Jerzykiewicz, T., and D. A. Russell Hochstetter, F. 1991. Late Mesozoic stratigraphy and verte- 1896. BeitraÈge zur Anatomie und Entwicke- brates of the Gobi Basin. Cretaceous lungsgeschichte des BlutgefaÈsssystems Res. 12: 345±377. der Monotremen. Semon Zool. For- Jerzykiewicz, T., P. J. Currie, D. A. Eberth, P. A. schungsreisen in Australien 5: 189± Johnston, E. H. Koster, and J.-J. Zheng 243. 1993. Djadokhta Formation correlative strata Hopson, J. A., Z. Kielan-Jaworowska, and E. F. in Chinese Inner Mongolia: an over- Allin view of the stratigraphy, sedimentary 1989. The cryptic jugal of multituberculates. geology, and paleontology and compar- J. Vertebr. Paleontol. 9: 201±209. isons with the type locality in the pre- Hopson, J. A., and G. W. Rougier Altai Gobi. Can. J. Earth Sci. 30: 1993. Braincase structure in the oldest known 2180±2195. skull of a therian mammal: implications Ji, Q., Z. Luo, and S. Ji for mammalian systematics and cranial 1999. A Chinese triconodont mammal and evolution. In P. Dodson and P. Ginger- mosaic evolution of the mammalian ich (eds.), Functional morphology and skeleton. Nature 398: 326±330. evolution. Am. J. Sci. 293-A: 268±299. Kallen, F. C. Hopson, J. A., and J. R. Wible 1977. The cardiovascular system of bats: 1993. Cranial vascular patterns in Multituber- structure and function. In W. A. Wim- culata: the postglenoid foramen that satt (ed.), Biology of bats, vol. 3: 289± wasn't there. J. Vertebr. Paleontol. 13 483. New York: Academic Press. (suppl. 3): 42A. Kermack, K. A. Hu, Y., Y. Wang, Z. Luo, and C. Li 1963. 1997. A new symmetrodont mammal from The cranial structure of the tricono- China and its implications for mam- donts. Philos. Trans. R. Soc. London malian evolution. Nature 390: 137± B246: 83±102. 142. 1967. The interrelations of early mammals. Hurum, J. H. Zool. J. Linn. Soc. 47: 241±249. 1994. Snout and orbit of Cretaceous Asian Kermack, K. A., and Z. Kielan-Jaworowska multituberculates studied by serial sec- 1971. Therian and non-therian mammals. In tions. Acta Palaeontol. Pol. 39: 181± D. M. Kermack and K. A. Kermack 221. (eds.), Early mammals. Zool. J. Linn. 1998a. The braincase of two Late Cretaceous Soc. 50 (suppl. 1): 103±115. Asian multituberculates studied by se- Kermack, K. A., F. Mussett, and H. W. Rigney rial sections. Ibid. 43: 21±52. 1981. The skull of Morganucodon. Zool. J. 1998b. The inner ear in two Late Cretaceous Linn. Soc. 71: 1±158. multituberculate mammals, and its im- Kielan-Jaworowska, Z. plications for multituberculate hearing. 1969. Discovery of a multituberculate mar- J. Mamm. Evol. 5: 65±93. supial bone. Nature 222: 1091±1092. Hurum, J. H., R. Presley, and Z. Kielan-Jawo- 1970. New Upper Cretaceous multitubercu- rowska late genera from Bayn Dzak, Gobi De- 1995. Multituberculate ear ossicles. In A. Sun sert. Palaeontol. Pol. 21: 35±49. and Y. Wang (eds.), Sixth symposium 1971. Skull structure and af®nities of the on Mesozoic terrestrial ecosystems and Multituberculata. Ibid. 25: 4±41. biota, short papers: 243±246. Beijing: 1974. Multituberculate succession in the Late China Ocean Press. Cretaceous of the Gobi Desert (Mon- 1996. The middle ear in multituberculate golia). Ibid. 30: 25±44. mammals. Acta Palaeontol. Pol. 41: 1979. Pelvic structure and nature of reproduc- 253±275. tion in Multituberculata. Nature 277: Jenkins, F. A., Jr. 402±403. 1990. Monotremes and the biology of Meso- 1980. Absence of ptilodontoidean multituber- zoic mammals. Neth. J. Zool. 40: 5±31. culates from Asia and its palaeogeo- 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 109

graphic implications. Lethaia 13: 169± eutherian mammal Barunlestes. Ibid. 173. 25: 167±185. 1983. Multituberculate endocranial casts. Pa- Kielan-Jaworowska, Z., C. Poplin, R. Presley, and laeovertebr. Montpellier 13: 1±12. A. de RicqleÁs 1986. Brain evolution in Mesozoic mammals. 1984. Preliminary note on multituberculate In K. M. Flanagan and J. A. Lillegrav- cranial anatomy studied by serial sec- en (eds.), Vertebrates, phylogeny, and tions. In W.-E. Reif and F. Westphal philosophy. Contrib. Geol. Univ. Wy- (eds.), Third symposium on Mesozoic oming Spec. Pap. 3: 21±34. terrestrial ecosystems: 123±128. TuÈ- 1992. Interrelationships of Mesozoic mam- bingen: Attempto Verlag. mals. Hist. Biol. 6: 185±202. Kielan-Jaworowska, Z., R. Presley, and C. Poplin 1994. A new generic name for the multitu- 1986. The cranial vascular system in taenio- berculate mammal `Djadochtatherium' labidoid multituberculate mammals. catopsaloides. Acta Palaeontol. Pol. 39: Philos. Trans. R. Soc. London B313: 134±136. 525±602. 1997. Characters of multituberculates neg- Kielan-Jaworowska, Z., D. Dashzeveg, and B. A. lected in phylogenetic analyses of early Tro®mov mammals. Lethaia 29: 249±266. 1987. Early Cretaceous multituberculates 1998. Humeral torsion in multituberculate from Mongolia and a comparison with mammals. Acta Palaeontol. Pol. 43: Late Jurassic forms. Acta Palaeontol. 131±134. Pol. 32: 3±47. Kielan-Jaworowska, Z., and J. F. Bonaparte Kielan-Jaworowska, Z., M. J. Novacek, B. A. Tro- 1996. Partial dentary of a multituberculate ®mov, and D. Dashzeveg mammal from the Late Cretaceous of 2000. Mammals from the Mesozoic of Mon- Argentina and its taxonomic implica- golia. In M. J. Benton, E. N. Kuro- tions. Mus. Argent. Cienc. Nat. ``Ber- chkin, M. A. Shishkin, and D. M. Un- nardino Rivadavia'' Inst. Nac. Invest. win (eds.), The age of dinosaurs in Cienc. Nat., n. ser. 145: 1±9. Russia and Mongolia. London: Cam- Kielan-Jaworowska, Z., and D. Dashzeveg bridge Univ. Press. 1978. New Late Cretaceous mammal locality Krause, D. W., and F. E. Grine in Mongolia and a description of a new 1996. The ®rst multituberculates from Mad- multituberculate. Acta Palaeontol. Pol. agascar: implications for Cretaceous 23: 115±130. biogeography. J. Vertebr. Paleontol. 16 Kielan-Jaworowska, Z., and P. P. Gambaryan (suppl. 3): 46A. 1994. Postcranial anatomy and habits of Krause, D. W., and Z. Kielan-Jaworowska Asian multituberculate mammals. Fos- 1993. The endocranial cast and encephaliza- sils Strata 36: 92 pp. tion quotient of Ptilodus (Multituber- Kielan-Jaworowska, Z., and J. H. Hurum culata, Mammalia). Palaeovertebr. 1997. Djadochtatheria - a new suborder of Montpellier 22: 99±112. multituberculate mammals. Acta Pa- Krause, D. W., Z. Kielan-Jaworowska, and J. F. laeontol. Pol. 42: 201±242. Bonaparte Kielan-Jaworowska, Z., and T. Qi 1989. Ferugliotherium Bonaparte, the ®rst 1990. Fossorial adaptations of a taeniolabi- known multituberculate from South doid multituberculate mammal from America. J. Vertebr. Paleontol. 12: the Eocene of China. Vertebr. PalAsiat. 351±376. 28: 81±94. Krause, D. W., G. V. Prasad, W. von Koenigswald, Kielan-Jaworowska, Z., and R. E. Sloan A. Sahni, and F. E. Grine 1979. Catopsalis (Multituberculata) from 1997. Cosmopolitanism among Gondwanan Asia and North America and the prob- Late Cretaceous mammals. Nature 390: lem of taeniolabidoid dispersal in the 504±507. Late Cretaceous. Acta Palaeontol. Pol. Kristhalka, L., R. J. Emry, J. E. Storer, and J. F. 24: 187±197. Sutton Kielan-Jaworowska, Z., and A. V. Sochava 1982. Oligocene multituberculates (Mamma- 1969. The ®rst multituberculate from the Up- lia: Allotheria): youngest known re- permost Cretaceous of the Gobi Desert cord. J. Paleontol. 56: 791±794. (Mongolia). Ibid. 14: 356±367. Kuhn, H.-J. Kielan-Jaworowska, Z., and B. A. Tro®mov 1971. Die Entwicklung und Morphologie des 1980. Cranial morphology of the Cretaceous SchaÈdels von Tachyglossus aculeatus. 110 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247

Abh. Senckenb. Naturforsch. Ges. 528: Luo, Z., and D. R. Ketten 1±192. 1991. CT scanning and computerized recon- Kuhn, H.-J., and U. Zeller structions of the inner ear of multitu- 1987. The cavum epiptericum in monotremes berculate mammals. J. Vertebr. Paleon- and therian mammals. In H.-J. Kuhn tol. 11: 220±228. and U. Zeller (eds.), Morphogenesis of MacIntyre, G. T. the mammalian skull. Mamm. Depicta 1972. The trisulcate petrosal pattern of mam- 13: 50±70. mals. In T. Dobzhansky, M. K. Hecht, KuÈhne, W. G. and W. C. Steere (eds.), Evolutionary 1956. The Liassic therapsid Oligokyphus. biology, 6: 275±303. New York: Ap- London: Brit. Mus. pleton-Century-Crofts. Lillegraven, J. A., and G. Hahn MacPhee, R.D.E. 1993. Evolutionary analysis of the middle 1981. Auditory region of primates and and inner ear of Late Jurassic multitu- eutherian insectivores. Contrib. Prima- berculates. J. Mamm. Evol. 1: 47±74. tol. 18: 282 pp. Lillegraven, J. A., and G. Krusat 1994. Morphology, adaptations, and relation- 1991. Cranio-mandibular anatomy of Haldan- ships of Plesiorycteropus, and a diag- odon exspectatus (Docodonta; Mam- nosis of a new order of eutherian mammalia) from the Late Jurassic of Por- mals. Bull. Am. Mus. Nat. Hist. 220: tugal and its implications to the evolu- 214 pp. tion of mammalian characters. Contrib. MacPhee, R.D.E., and M. Cartmill Geol. Univ. Wyoming 28: 39±138. 1986. Basicranial structures and primate sys- Lillegraven, J. A., S. D. Thompson, B. K. McNab, tematics. In D. R. Swindler and J. Er- and J. L. Patton win (eds.), Comparative primate biol- 1987. The origin of eutherian mammals. Biol. ogy, vol. 1. Systematics, evolution, and J. Linn. Soc. 32: 281±336. anatomy: 219±275. New York: Alan R. Lucas, S. G., and Z. Luo Liss. 1993. Adelobasileus from the Upper Triassic Maier, W. of west Texas: the oldest mammal. J. 1987. The ontogenetic development of the or- Vertebr. Paleontol. 13: 309±334. bitotemporal region in the skull of Luckett, W. P. Monodelphis domestica (Didelphidae, 1993. An ontogenetic assessment of dental Marsupialia), and the problem of the homologies in therian mammals. In F. mammalian alisphenoid. In H.-J. Kuhn S. Szalay, M. J. Novacek, and M. C. and U. Zeller (eds.), Morphogenesis of McKenna (eds.), Mammal phylogeny: the mammalian skull. Mamm. Depicta Mesozoic differentiation, multituber- 13: 71±90. culates, monotremes, early therians, Marshall, L. G., J. A. Case, and M. O. Woodburne and marsupials: 182±204. New York: 1990. Springer. Phylogenetic relationships of the fami- Luo, Z. lies of marsupials. Curr. Mamm. 2: 1989. Structure of the petrosals of Multitu- 433±502. berculata (Mammalia) and morphology Marshall, L. G., and C. de Muizon of the molars of early arctocyonids 1995. Part II: The skull. In C. de Muizon (Condylarthra, Mammalia). Ph.D. diss., (ed.), Pucadelphys andinus (Marsupia- Univ. California, Berkeley, 426 pp. lia, Mammalia) from the early Paleo- 1994. Sister-group relationships of mammals cene of Bolivia. MeÂm. Mus. Natl. Hist. and transformations of diagnostic Nat. 165: 21±90. mammalian characters. In N. C. Fraser Martinson, G. G. and H.-D. Sues (eds.), In the shadow of 1982. Pozdnemelovye molliuski Mongolii the dinosaurs Ð early Mesozoic tetra- (The Upper Cretaceous mollusks of pods: 98±128. Cambridge: Cambridge Mongolia). Joint Soviet-Mongolian Pa- Univ. Press. leontol. Exped. Trans. 17: 5±76. [in 1996. Phylogenetic implications of the petro- Russian]. sal features on the relationships of the McDowell, S. B., Jr. multituberculate families of the Late 1958. The Greater Antillean insectivores. Cretaceous. J. Vertebr. Paleontol. 16 Bull. Am. Mus. Nat. Hist. 115: 113± (suppl. 3): 49A. 214. 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 111

McFarland, W. L., M. S. Jacobs, and P. J. Mor- igin and evolution of the eutherian au- gane ditory bulla. Mammal Rev. 7: 131±149. 1979. Blood supply to the brain of the dol- 1986. The skull of leptictid insectivorans and phin, Tursiops truncatus, with compar- the classi®cation of eutherian mam- ative observations on special aspects of mals. Bull. Am. Mus. Nat. Hist. 183: the cerebrovascular supply of other ver- 1±112. tebrates. Neurosci. Biobehav. Rev. 3 1993. Patterns of skull diversity in the mam- (suppl. 1): 1±93. malian skull. In J. Hanken and B. K. McKenna, M. C., and S. K. Bell Hall (eds.), The skull, vol. 2, Patterns 1997. Classi®cation of mammals above the of structural and systematic diversity: species level. New York: Columbia 438±545. Chicago: Univ. Chicago Univ. Press. Press. Meng, J. 1997. Mammalian evolution: an early record 1992. The stapes of Lambdopsalis bulla bristling with evidence. Curr. Bio. 7: (Multituberculata) and transformational R489±R491. analyses on some stapedial features in Novacek, M. J., M. Norell, M. C. McKenna, and Mammaliaformes. J. Vertebr. Paleontol. J. Clark 12: 459±471. 1994. Fossils of the Flaming Cliffs. Sci. Am. Meng, J., and A. R. Wyss 271: 60±69. 1995. Monotreme af®nities and low frequen- Novacek, M. J., M. A. Norell, L. Dingus, and D. cy hearing suggested by multitubercu- Dashzeveg late ear. Nature 377: 141±144. 1996. Dinosaurs, mammals, birds, and lizards 1996. Multituberculate phylogeny. Ibid. 379: from the Late Cretaceous Ukhaa Tol- 407. god fauna, Mongolia. J. Vertebr. Pa- Miao, D. leontol. 19 (suppl. 3): 56A. 1986. Dental anatomy and ontogeny of Lamb- Novacek, M. J., G. W. Rougier, J. R. Wible, M. dopsalis bulla (Mammalia, Multituber- C. McKenna, D. Dashzeveg, and I. Ho- culata). Contrib. Geol. Univ. Wyoming rovitz 24: 65±76. 1997. Epipubic bones in eutherian mammals 1988. Skull morphology of Lambdopsalis from the Late Cretaceous of Mongolia. bulla (Mammalia, Multituberculata) Nature 389: 483±486. and its implications to mammalian evo- Oelrich, T. M. lution. Contrib. Geol. Univ. Wyoming 1956. The anatomy of the head of Ctenosaura Spec. Pap. 4: 1±104. pectinata (Iguanidae). Misc. Publ. Mus. 1993. Cranial morphology and multitubercu- Zool. Univ. Michigan 94: 122 pp. late relationships. In F. S. Szalay, M. J. Parrington, F. R. Novacek, and M. C. McKenna (eds.), 1946. On the cranial anatomy of cynodonts. Mammal phylogeny: Mesozoic differ- Proc. Zool. Soc. London 116: 181±197. entiation, multituberculates, mono- Parrington, F. R., and T. S. Westoll tremes, early therians, and marsupials: 1940. The evolution of the mammalian pal- 63±74. New York: Springer. ate. Proc. R. Soc. London B230: 305± Miao, D., and J. A. Lillegraven 355. 1986. Discovery of three ear ossicles in a Pascual, R., F. J. Goin, D. W. Krause, E. Ortiz- multituberculate mammal. Natl. Geogr. Jaureguizar, and A. A. Carlini Res. 2: 500±507. 1999. The ®rst gnathic remains of Sudameri- Monastersky, R. ca: implications for gondwanathere re- 1996. The lost tribe of mammals. Sci. News lationships. J. Vertebr. Paleontol. 19: 150: 378±379. 373±382. Moore, W. J. Patterson, B., and E. C. Olson 1981. The mammalian skull. Cambridge: 1961. A triconodontid mammal from the Tri- Cambridge Univ. Press. assic of Yunnan. In G. Vandebroek Nomina Anatomica, 5th ed. (ed.), International colloquium on the 1983. Baltimore: Williams & Wilkins. evolution of lower and specialized Nomina Anatomica Veterinaria, 2nd ed. mammals: 129±191. Brussels: K. 1973. Vienna: Adolf Holzhausen's Succes- Vlaamse Acad. Wettensh., Lett. Schone sors. Kunsten Belgie. Novacek, M. J. Paulli, S. 1977. Aspects of the problem of variation, or- 1900. UÈ ber die PneumaticitaÈt des SchaÈdels 112 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247

bei den SaÈugetieren. I. UÈ ber den Bau berculate interrelationships. Am. Mus. des Siebbeins. UÈ ber die Morphologie Novitates 3191: 26 pp. des Siebbeins und die der PneumaticitaÈt Rougier, G. W., J. R. Wible, and M. J. Novacek bei den Monotremen und den Marsu- 1997b. Nasal and endocranial morphology of pialiern. Gegenbaurs Morphol. Jahrb. Late Cretaceous multituberculates from 28: 147±178. Ukhaa Tolgod, Mongolia. J. Vertebr. Presley, R. Paleontol. 13 (suppl. 3): 72A. 1981. Alisphenoid equivalents in placentals, Rowe, T. marsupials, monotremes and fossils. 1988. De®nition, diagnosis and origin of Nature 294: 668±670. Mammalia. J. Vertebr. Paleontol. 8: Presley, R., and F.L.D. Steel 241±264. 1978. The pterygoid and ectopterygoid in 1993. Phylogenetic systematics and the early mammals. Anat. Embryol. 154: 95± history of mammals. In F. S. Szalay, M. 110. J. Novacek, and M. C. McKenna (eds.), Prothero, D. R., and C. Swisher III Mammal phylogeny: Mesozoic differ- 1992. Magnetostratigraphy and geochronolo- entiation, multituberculates, mono- gy of the terrestrial Eocene±Oligocene tremes, early therians, and marsupials: transition in North America. In D. R. 129±145. New York: Springer. Prothero and W. A. Beggren (eds.), Eo- Rowe, T., W. D. Carlson, and W. W. Bottorff cene±Oligocene climatic and biotic 1995. Thrinaxodon: Digital atlas of the skull evolution: 46±73. Princeton: Princeton (CD-ROM). Austin: Univ. Texas Press. SaÈve-SoÈderbergh, G. Univ. Press. 1947. Notes on the brain-case in Sphenodon Rose, K. D., and H.-D. Sues and certain Lacertilia. Zool. Bidr. Upps. 1996. Vertebrate paleontology. Geotimes 25: 489±516. (Feb.): 31±34. Segall, W. Rougier, G. W. 1970. Morphological parallelisms of the bulla 1993. Vincelestes neuquenianus Bonaparte and auditory ossicles in some insecti- (Mammalia, Theria) un primitivo mam- vores and marsupials. Fieldiana Zool. Âferoõ del CretaÂcico Inferior de la Cuen- 51: 169±205. ca Neuquina. Ph.D. diss., Univ. Buenos Sereno, P. C., and M. C. McKenna Aires, 720 pp. 1995. Cretaceous multituberculate skeleton Rougier, G. W., J. R. Wible, and J. A. Hopson and the early evolution of the mam- 1992. Reconstruction of the cranial vessels in malian shoulder girdle. Nature 377: the Early Cretaceous mammal Vince- 144±147. lestes neuquenianus: implications for 1996. Multituberculate phylogeny. Ibid. 379: the evolution of the mammalian cranial 406±407. vascular system. J. Vertebr. Paleontol. Shaner, R. F. 12: 188±216. 1926. The development of the skull of the tur- 1996a. Basicranial anatomy of Priacodon frui- tle, with remarks on fossil reptile taensis (Triconodontidae, Mammalia) skulls. Anat. Rec. 32: 343±367. from the Late Jurassic of Colorado, and Shiino, K. a reappraisal of mammaliaform inter- 1914. Studien zur Kenntnis des Wirbeltier- relationships. Am. Mus. Novitates kopfes. I. Das Chondrocranium von 3183: 38 pp. Crocodilus mit BeruÈcksichtigung der Rougier, G. W., J. R. Wible, and M. J. Novacek Gehirnnerven und der KopfgefaÈsse. 1996b. Multituberculate phylogeny. Nature Anat. Hefte 50: 253±382. 379: 406. Shindo, T. 1996c. Middle-ear ossicles of Kryptobaatar 1914. Zur vergleichenden Anatomie der ar- dashzevegi (Mammalia, Multitubercu- teriellen KopfgefaÈsse der Reptilien. lata): implications for mammaliamorph Anat. Hefte 51: 267±356. relationships and the evolution of the 1915. UÈ ber die Bedeutung des Sinus caver- auditory apparatus. Am. Mus. Novita- nosus der SaÈuger mit vergleichenden tes 3187: 43 pp. anatomischer BeruÈcksichtigung anderer Rougier, G. W., M. J. Novacek, and D. Dashzeveg Kopfvenen. Ibid. 52: 319±495. 1997a. A new multituberculate from the Late Sigogneau-Russell, D. Cretaceous locality Ukhaa Tolgod, 1991. First evidence of Multituberculata Mongolia. Considerations on multitu- (Mammalia) in the Mesozoic of Africa. 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 113

Neues Jahrb. Geol. PalaÈontol. Mh. human foramen caecum. Anat. Rec. 1991: 119±125. 223: 242±244. Sigogneau-Russell, D., R. Frank, and J. Hemmerle Toeplitz, C. 1986. A new family of mammals from the 1920. Bau und Entwicklung des Knorpel- lower part of the French Rhaetic. In K. schaÈdels von Didelphis marsupialis. Padian (ed.), The beginning of the age Zoologica (Stuttg.) 27: 1±84. of dinosaurs, faunal change across the Tro®mov, B. A. Triassic±Jurassic boundary: 99±108. 1975. New data on Buginbaatar Kielan-Ja- Cambridge: Cambridge Univ. Press. worowska et Sochava, 1969 (Mamma- Simmons, N. B. lia, Multituberculata) from Mongolia. 1993. Phylogeny of Multituberculata. In F. S . Joint Soviet-Mongolian Paleontol. Ex- Szalay, M. J. Novacek, and M. C. Mc- ped. Trans. 2: 7±13. [In Russian]. Kenna (eds.), Mammal phylogeny: Me- Voit, M. sozoic differentiation, multitubercula- 1909. Das Primordialcranium des Kaninchens tes, monotremes, early therians, and unter BeruÈcksichtigung der Deck- marsupials: 146±164. New York: knochen. Anat. Hefte 38: 425±616. Springer. Wahlert, J. H. Simpson, G. G. 1974. The cranial foramina of protrogomor- 1925. A Mesozoic mammal skull from Mon- phous rodents; an anatomical and phy- golia. Am. Mus. Novitates 201: 12 pp. logenetic study. Bull. Mus. Comp. 1937. Skull structure of the Multituberculata. Zool. 146: 363±410. Bull. Am. Mus. Nat. Hist. 73: 727±763. Wall, C. E., and D. W. Krause Sloan, R. E. 1992. A biomechanical analysis of the mas- 1979. Multituberculata. In R. W. Fairbridge ticatory apparatus of Ptilodus (Multi- and D. Jablonski (eds.), The encyclo- tuberculata). J. Vertebr. Paleontol. 12: pedia of paleontology: 492±498. 172±187. Stroudsberg, PA: Dowden, Hutchinson Watson, D.M.S. & Ross. 1911. The skull of Diademodon, with notes 1981. Systematics of Paleocene multituber- on those of some other cynodonts. Ann. culates from the San Juan Basin, New Mag. Nat. Hist. 8: 293±330. Mexico. In S. G. Lucas, J. K. Rigby, 1920. On the Cynodontia. Ibid. 9: 506±524. Jr., and B. S. Kues (eds.), Advances in Wible, J. R. San Juan Basin paleontology: 127±160. 1983. The internal carotid artery in early eu- Albuquerque, Univ. New Mexico therians. Acta Palaeontol. Pol. 28: 281± Press. 293. Starck, D. 1984. The ontogeny and phylogeny of the 1967. Le craÃne des mammifeÁres. In P.-P. Gras- mammalian cranial arterial pattern. seÂ (ed.), TraiteÂ de Zoologie, vol. 16, pt. Ph.D. diss., Duke Univ., Durham, NC, 1: 405±549, 1095±1102. Paris: Masson. 705 pp. 1978. Das evolutive Plateau SaÈugetier. Son- 1986. Transformations in the extracranial derb. Naturwiss. Ver. Hamb. 3: 7±33. course of the internal carotid artery in 1995. Lehrbuch der speziellen Zoologie. mammalian phylogeny. J. Vertebr. Pa- Band II: Wirbeltiere. Teil 5/1: SaÈuge- leontol. 6: 313±325. tiere. Jena: Gustav Fischer. 1987. The eutherian stapedial artery: charac- Tandler, J. ter analysis and implications for super- 1899. Zur vergleichenden Anatomie der ordinal relationships. Zool. J. Linn. Kopfarterien bei den Mammalia. Soc. 91: 107±135. Denkschr. K. Akad. Wiss. Wien Math.- 1989. Vessels on the side wall of the brain- Natwiss. Kl. 67: 677±784. case in cynodonts and primitive mam- 1901. Zur vergleichenden Anatomie der mals. In H. Splechtna and H. Hilgers Kopfarterien bei den Mammalia. Anat. (eds.), Trends in vertebrate morpholo- Hefte 18: 327±368. gy. Fortschr. Zool. 35: 406±408. 1902. Zur Entwicklungsgeschichte der Kopf- 1990. Late Cretaceous marsupial petrosal arterien bei den Mammalia. Gegen- bones from North America and a cla- baurs Morphol. Jahrb. 30: 275±373. distic analysis of the petrosal in therian Thewissen, J. G. M. mammals. J. Vertebr. Paleontol. 10: 1989. Mammalian frontal diploic vein and the 183±205. 114 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247

1991. Origin of Mammalia: the craniodental Witmer, L. M. evidence reexamined. Ibid. 11: 1±28. 1995. The extant phylogenetic bracket and Wible, J. R., and D. L. Davis the importance of reconstructing soft 2000. Ontogeny of the chiropteran basicra- tissues in fossils. In J. Thomason (ed.), nium, with reference to the Indian false Functional morphology in vertebrate vampire bat Megaderma lyra. In R. A. paleontology: 19±33. New York: Cam- Adams and S. C. Pedersen (eds.), On- bridge Univ. Press. togeny, functional ecology and evolu- Zeller, U. tion of bats: 214±246. New York: Cam- 1985. Die Ontogenese und Morphologie der bridge Univ. Press. Fenestra rotunda und des Aquaeductus Wible, J. R., and J. A. Hopson cochleae von Tupaia und anderen 1993. Basicranial evidence for early mammal SaÈugern. Gegenbaurs Morphol. Jahrb. phylogeny. In F. S. Szalay, M. J. No- vacek, and M. C. McKenna (eds.), 131: 179±204. Mammal phylogeny: Mesozoic differ- 1988. The lamina cribrosa of Ornithorhyn- entiation, multituberculates, mono- chus (Monotremata, Mammalia). Anat. tremes, early therians, and marsupials: Embryol. 178: 513±519. 45±62. New York: Springer. 1989. Die Entwicklung und Morphologie des 1995. Homologies of the prootic canal in SchaÈdels von Ornithorhynchus anatin- mammals and non-mammalian cyno- us (Mammalia: Prototheria: Mono- donts. J. Vertebr. Paleontol. 15: 331± tremata). Abh. Senckenb. Naturforsch. 356. Ges. 545: 1±188. Wible, J. R., D. Miao, and J. A. Hopson 1991. Foramen perilymphaticum und Reces- 1990. The septomaxilla of fossil and Recent sus scalae tympani von Ornithorhyn- synapsids and the problem of the sep- chus anatinus (Monotremata) und an- tomaxilla of monotremes and armadil- deren SaÈugern. Verh. Anat. Ges. 84: los. Zool. J. Linn. Soc. 98: 203±228. 441±443. Wible, J. R., G. W. Rougier, M. J. Novacek, M. 1993. Ontogenetic evidence for cranial ho- C., McKenna, and D. Dashzeveg mologies in monotremes and therians. 1995. A mammalian petrosal from the Early In F. S. Szalay, M. J. Novacek, and M. Cretaceous of Mongolia: implications C. McKenna (eds.), Mammal phyloge- for the evolution of the ear region and ny: Mesozoic differentiation, multitu- mammaliamorph interrelationships. berculates, monotremes, early therians, Am. Mus. Novitates 3149: 19 pp. Williams, P. L., R. Warwick, M. Dyson, and L. H. and marsupials: 95±128. New York: Bannister (eds.) Springer. 1989. Gray's anatomy, 37th ed. Edinburgh: Zeller, U., J. R. Wible, and M. Elsner Churchill-Livingstone. 1993. New ontogenetic evidence on the sep- Wilson, J. T. tomaxilla of Tamandua and Choloepus 1900. On the skeleton of the snout and os ca- (Mammalia, Xenarthra), with a reeval- runculae of the mammary foetus of the uation of the problem of the mamma- monotremes. Proc. Linn. Soc. N.S.W. lian septomaxilla. J. Mamm. Evol. 1: 25: 58±59. 31±46.

GLOSSARY Given that usages of anatomical terms vary numerous plesiomorphies retained by monotremes among researchers, we include a glossary speci- (Zeller, 1989, 1993; Jenkins, 1990) and multitu- fying our usages for the cranial foramina and ca- berculates (Kielan-Jaworowska, 1992; Lillegraven nals. The English form is given, followed by the and Hahn, 1993; Kielan-Jaworowska and Gam- Latin equivalent in parentheses where appropriate baryan, 1994), the former might represent a suit- based on the Nomina Anatomica (1983) and/or able model for evaluating multituberculate mor- Nomina Anatomica Veterinaria (1973). For many phology; however, the details of monotreme cra- terms, we have used the morphology of the dog nial anatomy, although well known (e.g., Kuhn, Canis familiaris as the basis for our discussion, 1971; Zeller, 1989), are not described and illus- relying on the detailed treatment of dog anatomy trated to the same degree as for the dog. by Evans and Christensen (1979). Considering the ACCESSORY PALATINE FORAMEN (FORAMEN PALA- 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 115

TINUM ACCESSORIUS). In the dog (Evans and (1986) and the post-trigeminal canal of Rougier Christensen, 1979), the accessory palatine nerve et al. (1996a, 1996c). and artery (branches of the major palatine nerve CAROTID CANAL (CANALIS CAROTICUM). In hu- and artery, respectively) travel in an accessory mans (Williams et al., 1989), the internal carotid palatine canal to supply the caudal portion of the artery and nerves run from the skull base to the mucosa on the hard palate; oddly enough, their cranial cavity through a perbullar carotid canal foramen of exit on the palate is called the minor (Wible, 1986) in the petrosal (petrous temporal). palatine foramen. In Kryptobaatar, there are four In Kryptobaatar, the internal carotid artery and or ®ve small foramina in the lateral margin of the nerves ®rst ran through the middle ear in a groove palatine (minor palatine foramina of Kielan-Ja- on the promontorium of the petrosal and then en- worowska, 1970), behind the major palatine fo- tered a carotid canal via a foramen between the ramen, that we interpret as for the accessory pal- petrosal and alisphenoid at the anterolateral pole atine nerves and arteries. Unlike the terminology of the promontorium. From this aperture, the ca- in the dog, we call the apertures that transmitted rotid canal runs anteromedially and dorsally these nerves and arteries the accessory palatine across the anterior pole of the promontorium be- foramina and reserve the term minor palatine fo- tween the petrosal, alisphenoid, and probably the ramen for the more posterior aperture that trans- pterygoid. The canal ends at the carotid foramen mitted the minor palatine nerve and artery. in the basisphenoid, which opens endocranially in the posterolateral part of the hypophyseal fossa. ASCENDING CANAL (CANALIS ASCENDENS). First used by Kielan-Jaworowska et al. (1986) for the Branching off the ventromedial part of the carotid intramural canal in multituberculates within the canal is the pterygoid canal. suture between the anterior lamina and the squa- DORSAL ASCENDING CANAL (CANALIS ASCENDENS mosal. These authors suggested that this canal DORSALIS). Following Rougier et al. (1992), the segment of the ascending canal dorsal to its com- was more widespread among Mesozoic and extant munication with the posttemporal canal, trans- taxa, and a more detailed evaluation of its ho- porting the ramus superior of the stapedial artery mologies was completed by Rougier et al. (1992). to the orbitotemporal channel. In Kryptobaatar, The major occupants were the ramus superior of the dorsal ascending canal is an intramural pas- the stapedial artery and accompanying veins (Wi- sageway between the squamosal and petrosal. ble, 1989; Rougier et al., 1992; Wible and Hop- However, the equivalent channel can be endocra- son, 1995). The canal was further subdivided into nial, as in lipotyphlans, or an extracranial sulcus, dorsal and ventral ascending canals by Rougier et as in the non-mammalian cynodont Probaino- al. (1992), based on the positional relationship to gnathus (Rougier et al., 1992). the posttemporal canal. ETHMOIDAL FORAMEN (FORAMEN ETHMOIDA- BUCCINATOR FORAMEN (FORAMEN BUCCINATOR- LE). In the dog (Evans and Christensen, 1979), IUM). Aperture in the alisphenoid, on the side two ethmoidal foramina connect the orbit and cra- wall of the braincase, transmitting the buccal nial cavity; the smaller, ventral one is in the suture branch of the mandibular nerve in some rodents between the frontal and orbitosphenoid, and the (Hill, 1935; Wahlert, 1974). Despite the implica- other is entirely in the frontal. The ventral fora- tion of the name, the buccal branch of the man- men transmits the ethmoidal nerve, a branch of dibular nerve is not the motor nerve to the buc- the ophthalmic, and the dorsal transmits the ex- cinator muscle, but is sensory to the skin and mu- ternal ethmoidal artery, a branch of the maxillary, cosa associated with that muscle (Evans and and an accompanying vein. Humans (Williams et Christensen, 1979; Williams et al., 1989). Used in al., 1989) have anterior and posterior ethmoidal Kryptobaatar for the anteriorly directed aperture foramina, which transmit the anterior and poste- in the pterygoid fossa of the alisphenoid between rior ethmoidal nerves and vessels, respectively, that bone and the anterior lamina. with the former nerve being equivalent to the eth- CANAL FOR RAMUS INFERIOR (CANALIS RAMUS IN- moidal nerve of the dog. The foramina in humans FERIUS). Following Wible and Hopson (1995), are either in the frontoethmoidal suture or in the the horizontal canal dorsal to the epitympanic re- frontal. In Kryptobaatar, there is a single eth- cess and lateral ¯ange that connects the middle moidal foramen between the frontal and orbito- ear and cavum epiptericum in multituberculates. sphenoid, which accommodated the ethmoidal Transmitted the ramus inferior of the stapedial ar- nerve and vessels. tery and perhaps the post-trigeminal vein. In FENESTRA VESTIBULI (FENESTRA VESTIBU- Kryptobaatar, the canal for the ramus inferior has LI). Term in the Nomina Anatomica (1983) for a common tympanic aperture with the secondary the oval window, the opening in the petrosal (pe- facial foramen. Equivalent to the canal for the trous temporal) that accommodates the footplate ?maxillary artery of Kielan-Jaworowska et al. of the stapes. 116 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247

FORAMEN FOR FRONTAL DIPLOIC VEIN (FORAMEN FERIUM). In Ptilodus, Simpson (1937) reported VENA DIPLOEÈ TICA FRONTALIS). In the dog (Evans two openings in the area where the foramen ovale and Christensen, 1979), a small unnamed foramen was expected. He interpreted both openings as for in the orbital surface of the postorbital (zygomat- the mandibular nerve and named them the fora- ic) process of the frontal transmits the frontal di- men masticator and foramen ovale inferius. As a ploic vein, an emissary vein from the diploeÈof model he offered the condition in some extant ro- the frontal to the ophthalmic vein. In the few eu- dents where as many as three separate apertures therians for which this vein has been described, it transmitting branches of the mandibular nerve connects either the superior (dorsal) sagittal sinus were reported by Hill (1935). Following Kielan- or the frontal paranasal air sinus and the frontal Jaworowska et al. (1986), we employ the term diploeÈ with the ophthalmic vein (Thewissen, foramen ovale inferium for the more medial of the 1989). Zeller (1989: ®g. 55) reported a frontal di- apertures that transmitted the mandibular nerve ploic vein in Ornithorhynchus, and his illustration through the anterior lamina in multituberculates. of the adult skull shows an unnamed foramen in In Kryptobaatar, the foramen ovale inferium lies the frontal in the same position as that transmit- in the anterior part of the epitympanic recess. ting the frontal diploic vein in the dog. A similarly HYPOGLOSSAL CANAL (CANALIS HYPOGLOSSI). In placed foramen in the frontal in Tachyglossus ac- the dog (Evans and Christensen, 1979), the open- commodates diploic vessels according to Kuhn ing in the exoccipital that transmits the hypoglos- (1971) and, as interpreted here, in Kryptobaatar. sal nerve and the vein of the hypoglossal canal, To facilitate description, we designate the aperture which connects the condyloid and vertebral veins. in monotremes and Kryptobaatar the foramen for In Kryptobaatar, only one hypoglossal foramen the frontal diploic vein. seems to be present; it is deep within the jugular FORAMEN FOR RAMUS TEMPORALIS (FORAMEN RA- fossa and completely encircled by the exoccipital. MUS TEMPORALIS). Openings in the squamosal INCISIVE FORAMEN (FORAMEN INCISIVUM). In the and parietal of various extant therians that trans- dog (Evans and Christensen, 1979), a large aper- mit temporal rami of the ramus superior of the ture (palatine ®ssure) in the anterior palate be- stapedial artery and accompanying veins into the tween the premaxilla and maxilla transmits the temporal fossa (Wible, 1987). Similar foramina septal branch of the caudal nasal nerve, the rostral are found in or along the anterior lamina of Kryp- septal branch of the major palatine artery, and the tobaatar and are presumed to have had the same nasopalatine (incisive) duct, which connects the function. Included are the supraglenoid foramen nasal and oral cavities with the vomeronasal (Ja- in the anterior lamina in front of the root of the cobson's) organ. We call the equivalent aperture zygoma, the foramen of the dorsal ascending ca- in Kryptobaatar the incisive foramen following nal between the anterior lamina and the squamo- the Nomina Anatomica (1983). It likely transmit- sal at their superior limits, and an unnamed fo- ted the same structures as the palatine ®ssure in ramen between the anterior lamina and squamosal the dog; there is no bony evidence other than the above the root of the zygoma. incisive foramen for the nasopalatine duct and FORAMEN FOR PITUITO-ORBITAL VEIN (FORAMEN vomeronasal organ, but their presence is inferred VENA PITUITO-ORBITALIS). In Kryptobaatar, the for Kryptobaatar from the widespread distribution foramen in the anterolateral wall of the hypophy- of these structures among extant amniotes (Gau- seal fossa connecting to the orbitotemporal fossa, thier et al., 1988; Eisthen, 1992). which transmitted a vein between the pituitary INFRAORBITAL FORAMEN (FORAMEN INFRAORBITA- (hypophysis) and the orbit. LE). In the dog (Evans and Christensen, 1979), FORAMEN OF DORSAL ASCENDING CANAL (FORA- the opening in the maxilla dorsal to the fourth MEN CANALIS ASCENDENS DORSALIS). Miao (1988) premolar at the anterior end of the infraorbital ca- coined the term ascending canal foramen for the nal that connects the orbit and snout and accom- opening into the dorsal ascending canal between modates the infraorbital nerve, artery, and vein. the anterior lamina, squamosal, and parietal in In the ¯oor of the infraorbital canal are several Lambdopsalis. This aperture was interpreted by small alveolar canals, and in the medial wall is Rougier et al. (1992) as having housed a temporal the incisivomaxillary canal; alveolar branches of branch of the ramus superior of the stapedial ar- the infraorbital nerve and vessels are transmitted tery. Miao's (1988) term is modi®ed here to fo- to the premolars by the former and to the canine ramen of the dorsal ascending canal to distinguish and incisors by the latter. In Kryptobaatar, the it from other foramina on the ventral ascending infraorbital foramen is dorsoventrally compressed, canal. In Kryptobaatar, the foramen of the dorsal visible in ventral view, and above the embrasures ascending canal is between the anterior lamina of the ®rst and second premolars or slightly pos- and squamosal at their superior limits. terior to them. In the right side of PSS-MAE 113, FORAMEN OVALE INFERIUM (FORAMEN OVALE IN- two openings leading medially into the maxilla 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 117

(either to the maxillary sinus or nasal cavity) are (1937), the term is used for the more lateral of visible. As noted by Rougier et al. (1997a), these the two foramina in the anterior lamina that trans- may be apertures to the alveolar canals. mitted branches of the mandibular nerve in mul- INTERNAL ACOUSTIC MEATUS (MEATUS ACUSTICUS tituberculates (see foramen ovale inferium). On INTERNUS). In the dog (Evans and Christensen, one side of one specimen of Kryptobaatar (PSS- 1979), the recess on the endocranial surface of the MAE 113), the masticatory foramen is bifurcated petrosal accommodating the facial and vestibulo- by the lateral ¯ange, which produces a ventral cochlear nerves as well as the labyrinthine artery opening into the epitympanic recess and a lateral off the basilar. A transverse crest divides the me- opening into the temporal fossa. atus into a dorsal part for the facial nerve and part MAJOR PALATINE FORAMEN (FORAMEN PALATINUM of the vestibular nerve, and a ventral part for the MAJOR). In the dog (Evans and Christensen, rest of the vestibular nerve and for the cochlear 1979), the major palatine nerve and artery nerve. In Kryptobaatar, the meatus is divided by (branches of the maxillary nerve and artery, re- a low transverse crest with a single aperture on spectively) leave the orbit in the palatine canal in either side. the maxilla that opens at the major palatine fora- JUGULAR FORAMEN (FORAMEN JUGULARE). In men on the palate between the maxilla and pala- the dog (Evans and Christensen, 1979), the open- tine. In Kryptobaatar, the major palatine nerve ing between the petrosal, basioccipital, and ex- and artery left the orbit via the sphenopalatine fo- occipital that connects the cranial cavity and skull ramen and appeared on the hard palate via the base and transmits the glossopharyngeal, vagus, major palatine foramen in the anteromedial corner and accessory nerves along with the sigmoid si- of the suture between the maxilla and palatine at nus. In Kryptobaatar, the jugular foramen lies be- the level of the P4/M1 embrasure; the presence of tween the petrosal and exoccipital, and in light of the dorsal aperture of the palatine canal cannot be its small size likely transmitted nerves only; the determined. sigmoid sinus and perhaps the inferior petrosal si- MENTAL FORAMEN (FORAMEN MENTALE). In the nus exited the cranial cavity via the foramen mag- dog (Evans and Christensen, 1979), there are usu- num as, for example, in Tachyglossus (Wible and ally three mental foramina, each of which trans- Hopson, 1995). mits nervous and vascular branches of the inferior LACRIMAL FORAMEN (FORAMEN LACRIMALE). In alveolar nerves and vessels from the mandible to the dog (Evans and Christensen, 1979), the open- the chin. In Kryptobaatar, the single mental fo- ing into the nasolacrimal (lacrimal) canal lies ramen is anterolaterally directed and in the dia- within a depression in the orbital surface of the stema between the incisor and third premolar, lacrimal bone called the fossa for the lacrimal sac. closer to the superior than the inferior margin of Accompanying the nasolacrimal duct into the ca- the bone. nal is a delicate branch of the malar artery, off METOPTIC FORAMEN (FORAMEN METOPTI- the infraorbital. ``In about 50 per cent of dogs, the CUM). The aperture occurring in most extant nasolacrimal duct has two openings (Evans and sauropsids between the pilae metoptica and an- Christensen, 1979: 1105).'' In Kryptobaatar, the totica that transmits the oculomotor nerve and in small opening into the nasolacrimal canal, which some instances also the trochlear nerve, ophthal- we call the lacrimal foramen following, for ex- mic artery, and/or pituito-orbital vein (De Beer, ample, Novacek (1986), is in the lacrimal high on 1926, 1937; Bellairs and Kamal, 1981). A metop- the orbital rim, near the suture with the facial pro- tic foramen is not present in the chondrocranium cess of the maxilla. or bony skull of any extant mammal, because the MANDIBULAR CANAL (CANALIS MANDIBULAR- pilae metoptica and antotica do not co-occur; IS). In the dog (Evans and Christensen, 1979), however, one is generally held to have been pre- the canal opening posteriorly in the ramus of the sent in the chondrocranium of the common an- mandible and anteriorly via the mental foramen cestor of Mammalia (Kuhn, 1971; Kuhn and Zell- that accommodates the inferior alveolar nerve off er, 1987; Zeller, 1989). In Lambdopsalis, Miao the mandibular nerve and accompanying vessels (1988) described an aperture at the anterolateral from the maxillary artery and vein. In Kryptobaa- corner of the hypophyseal fossa as the metoptic tar, the canal for the inferior alveolar nerve and foramen and placed within it the oculomotor vessels opens posteriorly just behind the level of nerve, ophthalmic artery, and pituitary vein (pi- the second molar. tuito-orbital vein), citing De Beer (1937). Follow- MASTICATORY FORAMEN (FORAMEN MASTICATOR- ing Miao, and in light of the morphology in Tach- IUM). Aperture in the alisphenoid, on the side yglossus (see Gaupp, 1908) and Ornithorhynchus wall of the braincase, transmitting the masticatory (personal obs.), we restore the oculomotor nerve branch of the mandibular nerve in some rodents in the similarly placed foramen in Kryptobaatar, (Hill, 1935; Wahlert, 1974). Following Simpson but not the ophthalmic artery and pituito-orbital 118 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247 vein. Although the ophthalmic artery accompa- ing the dorsal ascending canal with the orbit and nies the oculomotor nerve in some sauropsids transmitting the ramus superior of the stapedial (Shiino, 1914; Bellairs and Kamal, 1981), in oth- artery, its continuation the ramus supraorbitalis, ers and in mammals, the ophthalmic artery runs and accompanying veins. In Kryptobaatar, this with the optic nerve (Tandler, 1899; Hafferl, 1933; channel is largely endocranial, bounded laterally Wible, 1984). The pituito-orbital vein in Krypto- by the parietal, except anteriorly where there is a baatar occupied its own foramen. short intramural course between the parietal and MINOR PALATINE FORAMEN (FORAMEN PALATINUM frontal; the anterior opening of the orbitotemporal MINUS). In the dog (Evans and Christensen, canal is beneath the postorbital process, between 1979), the minor palatine nerve and artery the parietal, frontal, and anterior lamina. The or- (branches of the maxillary nerve and artery, re- bitotemporal canal, also called the sinus canal spectively) leave the orbit and reach the soft pal- (Watson, 1911) and internal parietal groove (Kie- ate via an unnamed notch, rarely closed to form lan-Jaworowska et al., 1986), may be wholly ex- a foramen, in the back edge of the maxilla and tracranial as, for example, in the non-mammalian palatine. In Kryptobaatar, the minor palatine cynodont Probainognathus (Rougier et al., 1992). nerve and artery left the orbit via a foramen in PERILYMPHATIC FORAMEN (FORAMEN PERILYMPHA- the posterior limit of the orbital process of the TICUM). Following Kuhn (1971) and Zeller maxilla and traversed the minor palatine canal, (1989, 1991), the aperture in monotremes in the which opens on the back edge of the palate be- back of the promontorium of the petrosal that tween the palatine, maxilla, and alisphenoid. We transmits the perilymphatic duct into the inner ear designate the opening on the palate the minor pal- and provides support for the secondary tympanic atine foramen. membrane. In therians, the perilymphatic duct be- NASAL FORAMEN (FORAMEN NASALE). First used comes enclosed in a separate canal, the cochlear by Simpson (1937) for the two pairs of foramina aqueduct, by the processus recessus; the enclosure visible on the dorsum of the nasals in Ptilodus, of the perilymphatic duct leaves a separate aper- suggested to have been vascular in function. Kie- ture for the secondary tympanic membrane, the lan-Jaworowska (1971) proposed the lateral eth- fenestra cochleae (Zeller, 1985). In Kryptobaatar, moidal nerve as an alternative occupant. Both ner- a cochlear aqueduct is lacking and, therefore, the vous and vascular functions seem likely. Similar aperture on the back of the promontorium is a foramina in the nasal of the iguanid Ctenosaura perilymphatic foramen. pectinata accommodate the dorsal cutaneous POSTTEMPORAL CANAL (CANALIS POSTTEMPORAL- branch of the lateral ethmoidal nerve, as pointed IS). Following Rougier et al. (1992), the vascular out by Kielan-Jaworowska (1971), and venous canal between the petrosal and squamosal and/or tributaries to the orbital sinus (Oelrich, 1956). In parietal connecting the ascending canal with the the dog (Evans and Christensen, 1979), the exter- occiput and transmitting the arteria and vena dinal nasal nerve, off the ethmoidal nerve, and ploeÈtica magna. In Kryptobaatar, the posterior branches of the external ethmoidal artery run ven- opening is at the lateral margin of the occiput, tral to the nasal bone and emerge at that bone's entirely within the mastoid exposure of the petro- anterior limit to supply the muzzle; these seem sal. likely occupants for the nasal foramina in multi- PRIMARY FACIAL FORAMEN (FORAMEN FACIALIS tuberculates. In the specimens of Kryptobaatar PRIMARIUM). Following Wible (1990) and Wible described here, the number of nasal foramina is and Hopson (1993), the opening for the facial uncertain because of damage; there are two fo- nerve in the endocranial surface of the petrosal ramina arranged asymmetrically between the two that connects the internal acoustic meatus with the sides in PSS-MAE 101. cavum supracochleare or cavum epiptericum. In OPTIC FORAMEN (FORAMEN OPTICUM). In the Kryptobaatar, the primary facial foramen shares dog (Evans and Christensen, 1979), the opening a common aperture in the internal acoustic meatus in the orbitosphenoid that connects the cranial with that for the vestibular nerve, and it opens cavity and orbit and transmits the optic nerve and into the back of the cavum epiptericum. the ophthalmic branch of the internal carotid ar- PROOTIC CANAL (CANALIS PROOTICUS). First tery. A similar opening occurs in the orbitosphe- used by Gaupp (1908) for the canal in the petrosal noid of Kryptobaatar. The embryological precur- of the echidna Tachyglossus aculeatus, transmit- sor of the optic foramen is in the chondrocranium ting the prootic sinus from the cranial cavity to between the pilae preoptica and metoptica (Starck, the middle ear where it joins the post-trigeminal 1967; Moore, 1981). vein to form the lateral head vein. In Kryptobaa- ORBITOTEMPORAL CANAL (CANALIS ORBITOTEM- tar, the endocranial aperture of the prootic canal PORALIS). Following Rougier et al. (1992), the is ventral to the lower margin of the subarcuate channel in the side wall of the braincase connect- fossa; the ventral aperture is in a recess medial to 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 119 the crista parotica along with the foramen for the panic surface of the petrosal by which the hyo- ramus superior. mandibular branch of the facial nerve enters the PROOTIC FORAMEN (FORAMEN PROOTICUM). The middle ear from the cavum supracochleare or ca- opening in the orbitotemporal region of the chon- vum epiptericum. In Kryptobaatar, the secondary drocranium between the pila antotica and the otic facial foramen appears to connect with the cavum capsule. In extant mammals, a complete pila an- epiptericum and has a common tympanic aperture totica and, therefore, the prootic foramen are with the canal for the ramus inferior of the sta- found only in the chondrocrania of monotremes pedial artery. (Kuhn, 1971; Zeller, 1989). Transmitted are the SPHENORBITAL FISSURE (FISSURA SPHENORBITAL- fourth through sixth cranial nerves, the prootic IS). The term sphenorbital ®ssure (orbital ®ssure vein (Wible and Hopson, 1995) and inferior pe- or anterior lacerate foramen) has been used for the trosal sinus (Shindo, 1915) between the cavum gap walled medially by the orbitosphenoid and epiptericum and the cranial cavity. This foramen laterally by the alisphenoid or anterior lamina that is altered in adult monotremes by the resorption transmits the contents of the cavum epiptericum of all but the base of the pila antotica (Kuhn and anteriorly into the orbit (Simpson, 1937; McDow- Zeller, 1987) and the enclosure of the prootic si- ell, 1958; Kielan-Jaworowska et al., 1986). The nus into a separate prootic canal (Wible and Hop- nervous and vascular contents of the gap so de- son, 1995). The comparable region of the endo- ®ned vary dramatically among extant mammals cranium in Kryptobaatar is posterior and lateral and may include some combination of the follow- to the ossi®ed pila antotica, medial to the anterior ing: the optic, oculomotor, trochlear, ophthalmic, lamina, and anterior to the petrosal. Transmitted maxillary, and abducens nerves; the ramus in- through this gap were the fourth through sixth cra- fraorbitalis, arteria anastomotica, and ophthalmic nial nerves and the inferior petrosal sinus. The artery; and the ophthalmic veins. In Kryptobaatar, prootic sinus was enclosed in a separate canal in the sphenorbital ®ssure lies between the orbito- the anterior lamina and petrosal. sphenoid ventromedially, the frontal dorsomedi- PTERYGOID CANAL (CANALIS PTERYGOIDEUS). In ally, the anterior lamina laterally, and probably the dog (Evans and Christensen, 1979), the pter- the maxilla ventrally; it transmitted the trochlear, ygoid canal is in the suture between the pterygoid ophthalmic, and maxillary nerves, the ophthalmic and basisphenoid and transmits the nerve and ar- veins, and probably the ramus infraorbitalis. tery of the pterygoid canal from the basicranium SPHENOPALATINE FORAMEN (FORAMEN SPHENOPA- to the posteroinferior ¯oor of the orbit. The nerve LATINUM). In the dog (Evans and Christensen, enters the basicranial end of the canal and is 1979), the opening in the palatine that connects formed by sympathetic ®bers from the internal ca- the orbit and the nasal fossa and transmits the cau- rotid (deep petrosal) nerve and parasympathetic dal nasal nerve and the sphenopalatine artery and ®bers from the greater (major) petrosal nerve; the vein, branches of the maxillary nerve, artery, and artery is a branch of the maxillary that enters the vein, respectively. The sphenopalatine foramen orbital end of the canal. In Kryptobaatar, a canal lies immediately dorsal to the caudal palatine fo- resembling the pterygoid canal of the dog origi- ramen, by which the major palatine nerve and ar- nates from the carotid groove or canal and is en- tery enter the palatine canal. In Kryptobaatar, the closed between the pterygoid and alisphenoid; the sphenopalatine and caudal palatine foramina that location of the orbital aperture could not be found. occur in the dog are fused into a single opening Given the size of this pterygoid canal and its in- either within the maxilla or between the maxilla timate relationship to the carotid canal, the artery and frontal. We call this common opening the of the pterygoid canal was present, probably as a sphenopalatine foramen, following, for example, branch of the internal carotid as occurs in some Kielan-Jaworowska et al. (1986). lipotyphlans (McDowell, 1958; MacPhee, 1981). STYLOMASTOID FORAMEN (FORAMEN STYLOMAS- Whereas both the deep petrosal nerve and the ar- TOIDEUM). In the dog (Evans and Christensen, tery of the pterygoid canal entered the pterygoid 1979), the opening between the petrosal, the os- canal from the carotid canal in Kryptobaatar, the seous bulla, and the tympanohyal (tympanohyoid) greater petrosal nerve must have entered from the cartilage by which the facial nerve (hyomandib- cavum epiptericum; there are no tympanic aper- ular branch) leaves the middle ear and the stylo- tures for the greater petrosal nerve that position it mastoid artery off the posterior (caudal) auricular to enter the carotid canal. Consequently, the pter- enters. In Kryptobaatar, only a stylomastoid notch ygoid canal of Kryptobaatar is not strictly ho- formed by the petrosal, including the ossi®ed mologous with that of the dog. hooklike tympanohyal, is evident. In addition to SECONDARY FACIAL FORAMEN (FORAMEN FACIALIS the facial nerve, the notch probably accommodat- SECUNDARIUM). Following Wible (1990) and Wi- ed a stylomastoid artery, because that vessel is ble and Hopson (1993), the opening in the tym- widely distributed in extant mammals (Wible, 120 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247

1987). However, whether that artery was derived TRANSVERSE CANAL (CANALIS TRANSVER- from the stapedial system as a ramus posterior or SUS). Term used by Gregory (1910) for the ve- from the external carotid system is uncertain. nous canal in the basisphenoid crossing the mid- SUPRAGLENOID FORAMEN (FORAMEN SUPRAGLE- line in some extant therians. In Kryptobaatar, the NOIDALE). First used by Cope (1880) for the pre- transverse canal is a large channel in the sphenoid sumed venous aperture perforating the base of the complex immediately in front of the hypophyseal zygomatic process of the squamosal from above fossa, connecting the left and right sphenorbital that occurs in a variety of extant therians. Simp- recesses. son (1937) applied the name to the aperture in VENTRAL ASCENDING CANAL (CANALIS ASCENDENS Ptilodus in front of the base of the zygomatic pro- VENTRALIS). Following Rougier et al. (1992), the cess of the squamosal that faces anterolaterally ventral part of the ascending canal between its into the temporal fossa; however, he noted this tympanic opening (equivalent to the pterygoparoccipital foramen, see Rougier et al., 1992; Wible opening does not correspond exactly with the su- and Hopson, 1995) and the posttemporal canal, praglenoid foramen of therians and questioned the transporting the ramus superior of the stapedial homology of the two. In Kryptobaatar, the supra- artery. In Kryptobaatar, the ventral ascending ca- glenoid foramen is in the anterior lamina slightly nal is horizontal, posterolaterally directed, and posterior and dorsal to the foramen masticatorium. presumably entirely within the anterior lamina; its Following Rougier et al. (1992) and Wible and tympanic opening shares a common recess with Hopson (1995), we reconstruct as the principal the prootic canal. The equivalent channel can be vessel in it an arterial branch to the temporalis more vertical and either endocranial, as in lipo- muscle off the ramus superior of the stapedial, typhlans, or extracranial, as in Morganucodon known as a ramus temporalis. (Rougier et al., 1992). 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 121

INDEX OF ANATOMICAL TERMS aqueducts premaxilla, 16±19, 72±74 cochlear, 39, 55, 118 presphenoid, 31 vestibular, 55 pterygoid, 30±31, 84 arteries septomaxilla, 16, 74±76, 100 accessory palatine, 77, 114±115 squamosal, 43±46 diploeÈtica magna, 67, 71, 91±92, 118 stapes, 40, 70, 90, 94, 99, 102 ethmoidal, 50, 72, 115, 118 supraoccipital, 46±47 external carotid, 67±68 vomer, 31, 83 frontal, 23, 72 infraorbital, 70, 116 canals internal carotid, 31, 36±37, 68, 89±90, 102, 115 accessory palatine, 114 lacrimal, 72 alisphenoid, 85, 91, 97 major palatine, 70, 117, 119 alveolar, 25, 77, 116 maxillary, 68 ascending, 67, 115 minor palatine, 70, 115, 118 carotid, 31, 37±38, 68, 89±90, 92, 115 ophthalmic, 50±51, 58, 72, 119 dorsal ascending, 43, 70±71, 115 of pterygoid canal, 38, 72, 92, 119 facial, 70 ramus inferior, 41, 65, 68±70, 88, 91, 115 hypoglossal, 116 ramus infraorbitalis, 55, 68±70, 96, 119 infraorbital, 23, 25, 70, 116 ramus mandibularis, 68 mandibular, 62, 117 ramus orbitalis, 72 of ?maxillary artery, 88, 115 ramus superior, 41, 67±72, 91±92, 115 minor palatine, 27, 78, 101, 118 ramus supraorbitalis, 23, 67±72, 92, 118 orbitotemporal, 23, 42, 67, 70±71, 101, 118 ramus temporalis, 42±43, 45, 71, 91±92, 116, palatine, 26, 117, 119 120 posttemporal, 43, 67, 70±72, 91±92, 118 sphenopalatine, 119 post-trigeminal, 88, 90, 115 stapedial, 37, 38±39, 42, 67, 70, 90, 102 prootic, 41, 55, 65, 88, 91, 102, 118 vertebral, 67, 68, 90 pterygoid, 38, 72, 92±93, 119 auditory tube, 84, 101 for ramus inferior, 70, 88, 90, 115 semicircular, 36, 39 bones sinus, 118 alisphenoid, 33±36, 85, 100 sphenoparietal, 97 anterior lamina, 36, 42±43, 55±56, 79, 100 transverse, 59, 65±66, 89, 100, 120 basioccipital, 48, 51±53 ventral ascending, 41, 70±71, 91, 115, 120 basisphenoid, 31, 56, 98 cava ectopterygoid, 84±85, 100 epiptericum, 49, 53, 55, 70, 95, 96±98 ectotympanic, 99 supracochleare, 55, 95 exoccipital, 47±48, 51 cerebellum frontal, 21±23, 79±80 para¯occulus, 53 incus, 40, 94 vermis, 63 interparietal, 46 cerebrum, 63 jugal, 43, 83, 99±100 choana, 28, 29, 30, 31, 34 lacrimal, 20±21, 78±79 cochlea, 36 malleus, 40, 99 condyles mandible, 59±62 of mandible, 61 maxilla, 23±28, 79±80 occipital, 47 nasal, 19±20, 76±77, 83 crests orbitosphenoid, 31±33, 58, 79, 81 condyloid, 60 palatine, 28±29, 79, 100 masseteric, 60 parietal, 46, 82±83, 101 nuchal, 45, 46, 47 petrosal, 36±43, 53±56, 93±96 orbitosphenoidal, 58 122 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247

pterygoid, 62 pterygoparoccipital, 41, 91, 120 sagittal, 46 for ramus temporalis, 43, 45, 46, 71, 92, 116 supraorbital, 22, 46 secondary facial, 41, 118 cribriform plate, 59, 99 sphenopalatine, 25±26, 70, 80, 119 cristae sphenoparietal, 51 interfenestralis, 39, 94 stylomastoid, 119 parotica, 40, 93, 94 supraglenoid, 42, 70, 91, 119±120 petrosa, 53 for vestibular nerve, 55 fossae dorsum sellae, 56, 64, 98 atlantal, 43, 47 ducts glenoid, 46 endolymphatic, 55 hypochiasmatica, 98 nasolacrimal, 59, 117 hypophyseal, 38, 56±57, 98, 102 perilymphatic, 39, 55, 93, 118 incudis, 40 jugular, 39, 40, 47±48, 53, 93, 101 egg-tooth, 74 masseteric, 59±60 orbitonasal, 22 fenestrae pterygoid, of alisphenoid, 35±36, 86±87 cochleae, 118 pterygoid, of mandible, 61±62 sphenoparietal, 51 stapedius, 40 vestibuli, 39, 70, 115 subarcuate, 53 foramina tensor tympani, 39, 95 for abducens nerve, 58 trigeminal, 55 accessory palatine, 28, 77±78, 114±115 foveae buccinator, 36, 85, 115 masseteric, 59 carotid, 37, 56, 68, 89±90, 102, 115 pterygoid, 62 for cochlear nerve, 55, 98 of dorsal ascending canal, 43, 92, 116 ganglia ethmoidal, 23, 33, 58±59, 81, 115 facial, 49, 55, 95, 96 for frontal diploic vein, 23, 80, 82, 115±116 trigeminal, 49, 55, 96 hypoglossal, 47±48, 96, 116 grooves incisive, 19, 116 ethmoidal, 33 infraorbital, 23, 77, 116 internal carotid, 36±38, 68, 89, 92, 102, 115 jugular, 33, 47, 51, 55, 63, 87±88, 117 internal parietal, 118 lacrimal, 21, 80, 117 sphenopalatine, 26±27, 70 magnum, 47, 48, 63, 102 stapedial, 37, 38, 70, 90, 102 major palatine, 27, 77±78, 117 temporal, 60 masticatorium, 36, 41, 86, 117 mental, 59, 117 hamulus, of pterygoid, 31, 84 metoptic, 33, 49, 50±51, 57±58, 81, 88, 96±97, hiatus Fallopii, 92 100, 117 minor palatine, 27, 34, 70, 77±78, 101, 118 jugum sphenoidale, 58, 98±99 nasal, 19±20, 76±77, 118 optic, 33, 49, 50±51, 58, 72, 81, 97, 100, 118 laminae orbitonasal, 49, 50 infracribrosa, 58, 99 ovale, 36 obturans, 36 ovale accessorius, 36 lateral ¯ange, 40, 70, 90, 94 ovale inferium, 41, 86, 116 perilymphatic, 39, 118 meatuses for pituito-orbital vein, 56, 88, 100, 102, 116 external acoustic, 40, 45 primary facial, 50, 55, 95, 118 internal acoustic, 53, 55, 116±117 prootic, 49, 50±51, 65, 118±119 muscles pseudoptic, 50±51 incisivus superioris, 18 2000 WIBLE AND ROUGIER: CRANIAL ANATOMY OF KRYPTOBAATAR DASHZEVEGI 123

lateral pterygoid, 36, 42, 62 pilae medial masseter, pars anterior, 22, 59, 79 antotica, 49, 50±51, 57±58, 97±98, 100 medial pterygoid, 39, 62, 84 metoptica, 49, 50±51, 58, 97±98, 100 sternomastoid, 43 preoptica, 49, 50±51, 58, 97 super®cial masseter, pars anterior, 23, 24 postpalatine torus, 29 super®cial masseter, pars posterior, 46 prefacial commissure, 50, 53 temporalis, 43, 60, 70, 82 processes tensor tympani, 39, 95 caudal tympanic, of petrosal, 40, 94 tensor veli palatini, 84 coronoid, 60±61 facial, of lacrimal, 20, 78 nasal overhang, 20, 76 orbital, of palatine, 28, 79, 100 nerves paroccipital, of petrosal, 40, 94 abducens (VI), 49, 51, 58, 96 posterodorsal, of premaxilla, 17 accessory palatine, 77, 114 postorbital, of frontal, 82±83, 101 buccal, of V, 36, 86±87, 115 postorbital, of parietal, 46, 82±83, 101 caudal nasal, 119 prenasal, of premaxilla, 17, 72±74, 100 cochlear, 55, 117 rostral tympanic, of petrosal, 39 deep petrosal, 38, 92±93, 119 promontorium, of petrosal, 36 deep temporal, 86 ethmoidal, 23, 33, 49, 59, 115, 118 recesses facial (VII), 40, 41, 49, 55, 117, 118, 119 epitympanic, 35, 40±41, 86, 94±95 frontal, 23 post-promontorial tympanic, 94 glossopharyngeal (IX), 117 scala tympani, 93±94 greater petrosal, 38, 93, 96, 119 sphenorbital, 25, 35 inferior alveolar, 117 ridges infraorbital, 25, 116 anterior zygomatic, 24±25 internal carotid, 38, 68, 115 intermediate zygomatic, 46 major palatine, 26±27, 77, 117, 119 orbital, 22 mandibular (V ), 36, 41, 68, 86, 116, 117 3 posterior zygomatic, 46 maxillary (V ), 55, 68 2 pterygopalatine, 30±31, 83±84, 101 minor palatine, 27, 77, 118 oculomotor (III), 49, 51, 55, 58, 88, 97, 117 temporal, 46

ophthalmic (V1), 55, 68 optic (II), 49, 51, 58, 72, 118 sinuses of pterygoid canal, 38, 72, 119 anterior intercavernous, 50, 51 spinal accessory (XI), 48, 117 cavernous, 63 trigeminal (V), 49, 51, 96 inferior petrosal, 50±51, 53, 63±64, 87±88, 117 trochlear (IV), 49, 51, 55, 96 maxillary, 25, 77, 116 vagus (X), 117 posterior intercavernous, 51, 64 vestibular, 55, 117 prootic, 41, 50±51, 55, 63, 65, 88, 118 notches sigmoid, 51, 53, 63, 87, 117 anterior nasal, 20, 76 superior sagittal, 63, 87 odontoid, 47, 48 tentorial, 63 palatonasal, 78 transverse, 63, 86 stylomastoid, 40, 119 sphenoobturator membrane, 36 supraorbital, 22, 23, 46 sphenorbital ®ssure, 25, 33, 55, 70, 81±82, 119 stylohyal, 36, 99 occipital plane, 47 subarcuate fenestration, 67 olfactory bulbs, 58 sulci orbital pocket, 22±23, 79±80 chiasmatic, 58, 98 orbital wing, 58 facial, 41 os carunculae, 74 for inferior petrosal sinus, 53, 64, 87 124 BULLETIN AMERICAN MUSEUM OF NATURAL HISTORY NO. 247

for nasolacrimal duct, 59 veins for pituito-orbital vein, 56, 66 capsuloparietal emissary, 51 for prootic sinus, 55 frontal diploic, 23, 115±116 for sigmoid sinus, 51, 63 internal jugular, 65, 88 symphysis, mandibular, 61 lateral head, 50, 64±65, 88, 118 middle cerebral, 50 taenia clino-orbitalis, 57 ophthalmic, 55, 63, 96 thickening, of premaxilla, 19 pituitary, 50, 117 transversal elevation, 62 pituito-orbital, 50, 56, 66, 88±89, 116, 117±118 troughs post-trigeminal, 41, 65, 70, 88, 115, 118 lateral, of petrosal, 40±41 prootic, 50 pterygopalatine, 30±31, 34, 83±84, 101 transverse canal, 59, 65±66, 89, 120 tuberculum sellae, 58, 98 vertebral, 64, 87 turbinals, 59 vestibule, 36, 95±96 tympanohyal, 40, 119 vomeronasal organ, 116