Novitate MUSEUM
ovitates
PUBLISHED BY THE AMERICAN MUSEUM OF NATURAL HISTORY
CENTRAL PARK WEST AT 79TH STREET, NEW YORK, NY
Number 3247, 16 pp., 2 figures
10024 November 5, 1998
A Partial Ornithomimid Braincase from Ukhaa Tolgod (Upper Cretaceous, Mongolia)
PETER J. MAKOVICKY'! AND MARK A. NORELL?
ABSTRACT
Among the dinosaurian remains recently dis- covered by the Mongolian Academy of Sciences— American Museum of Natural History expeditions at the Ukhaa Tolgod locality (Southeastern Gobi Desert, Mongolia) are a partial braincase and cer- vical vertebrae of an ornithomimid dinosaur (IGM 100/987). This specimen represents the first rec- ord of an ornithomimid from this rich locality, as well as the first discovery of such an animal in Djadokhta-like beds. Although broken and slight- ly distorted, the preserved portion of the braincase reveals new information on the anatomy of orni- thomimids. The middle ear region is enlarged and is connected to three expansive tympanic pneu-
matic systems as in other advanced theropods. The hypoglossal nerve (cranial nerve XII) is di- vided into three branches, a feature otherwise known among nonavialan coelurosaurs only in Troodon formosus. Several autapomorphies of the Ornithomimidae are preserved in IGM 100/987, including expansive pneumatization of the basi- occipital—exoccipital region dorsal to the basal tubera and a large depression of the posterior face of the quadrate shaft. IGM 100/987 displays sub- tle differences from North American ornithom- imid taxa and Gallimimus bullatus, but a more definite taxonomic assessment must await a thor- ough revision of ornithomimid phylogeny.
INTRODUCTION
In the summer of 1993 the Mongolian Academy of Sciences—-American Museum of Natural History expedition collected an as- semblage of small theropod remains as sur- face float at the Upper Cretaceous locality of
Ukhaa Tolgod, Southeastern Gobi Desert, Mongolia (Dashzeveg et al., 1995). This as- semblage comprised the occipital portion of a braincase, several cervical vertebrae, and a cervicodorsal vertebra: of an ornithomimid
‘Graduate Student, Department of Vertebrate Paleontology, American Museum of Natural History. 2Curator and Chairman, Department of Vertebrate Paleontology, American Museum of Natural History.
Copyright © American Museum of Natural History 1998
ISSN 0003-0082 / Price $1.80
2 AMERICAN MUSEUM NOVITATES
dinosaur (IGM 100/987), as well as the re- mains of a new troodontid taxon (Norell et al., in prep.). The remains of these two taxa are easily separated by diagnostic features as well as size; the ornithomimid is significantly larger than the troodontid. The cervicodorsal vertebra from the ornithomimid (probably the 10th or 11th presacral) is 33 mm long, whereas the length of the 1st dorsal vertebra of the troodontid (based on the presence of a large hypapophysis and the parapophysis being partially situated on the neural arch) is only 14.2 mm. Other comparable measure- ments include span of the paroccipital pro- cesses (well in excess of 40 mm in the or- nithomimid and ca. 34 mm in the troodontid) and width across the postzygapophyses of the axis (24.3 mm in the ornithomimid but estimated to be only 12.5 mm in the troo- dontid).
Currently, there are five ornithomimid taxa known from the Cretaceous of Mongolia and Inner Mongolia. These include the primitive forms Garudimimus brevipes (Barsbold, 1981) and Harpymimus okladnikovi (Bars- bold and Perle, 1984), as well as the more advanced ornithomimids Archaeornithomi- mus asiaticus (Gilmore, 1933), Gallimimus bullatus (Osmdlska et al., 1972), and Anser- imimus planynychus (Barsbold, 1988). The first two taxa are thought to be of Early Cre- taceous age, whereas there is some uncer- tainty with regard to the age of the Iren Da- basu beds that yielded Archaeornithomimus asiaticus. Although these deposits have tra- ditionally been interpreted as being of Cen- omanian age (Berkey and Morris, 1927), re- cent reanalysis of the fauna and stratigraphy suggests that they may in fact be Campanian (Currie and Eberth, 1993). Gallimimus bul- latus is known from a number of Upper Ne- megt sites, and Anserimimus planynychus is from the Bugin Tsav locality, which has been considered a temporal equivalent of the Late Campanian to Maastrichtian Nemegt For- mation (Jerzykiewicz and Russell, 1991). Jerzykiewicz and Russell (1991) listed the occurrence of an indeterminate ornithomimid from the Barun Goyot Formation without re- ferral to a specific specimen or locality. Ba- run Goyotian localities listed by these au- thors extend over a considerable area, and their stratigraphic relationship to each other
NO. 3247
and to other Djadokhta-like and Nemegt beds is unresolved. No other ornithomimids have been reported from Djadokhta or Djadokhta- like beds in Central Asia.
The ornithomimid braincase was only marginally treated by Parks (1928) and Rus- sell (1972) but was treated in more detail by Osmolska et al. (1972) and Barsbold (1983). Currie and Zhao (1994a) revised several of the interpretations of braincase foramina pre- sented in Osmdlska et al. (1972). Descrip- tions of the ornithomimid IGM 100/987 will be supplemented with descriptions of an oc- cipital portion of a braincase of Struthiomi- mus altus (AMNH 5355) when relevant to interpretation of structures in the former and also when assaying structural differences in ornithomimid braincases. Specimen 5355 was collected by an AMNH expedition in 1914 to the Dinosaur Park Formation (Cam- panian) of Alberta, Canada, but it has only been briefly discussed and figured by Russell (1972).
INSTITUTIONAL ABBREVIATIONS: AMNH— American Museum of Natural History (Ver- tebrate Paleontology Collection), New York; IGM—Institute of Geology, Mongolia, Ulaanbaatar, ROM—Royal Ontario Muse- um, Toronto; and TMP—Royal Tyrrell Mu- seum of Palaeontology, Drumheller.
REFERRED MATERIAL: Derived characters of Ornithomimidae present in IGM 100/987 include a posteriorly excavated quadrate shaft (personal obs.), large subcondylar re- cesses (Witmer, 1997), a deep excavation be- low the ligament scar of the anterior face of the neural spine on the cervical vertebrae (Makovicky, 1995), and an expanded flange on the anterior face of the last cervical rib (personal obs.).
DESCRIPTION
Only the occipital region of the braincase is preserved in IGM 100/987, comprising the supraoccipital, both exoccipital—opisthotics, the basioccipital, the posteroventral part of the basisphenoid, the posterior part of the right prootic, and the shaft of the right quad- rate with a sliver of the quadratojugal adher- ing to it (fig. 1A). The braincase is slightly distorted with the left exoccipital displaced dorsally relative to the right one. The basi-
1998
sphenoid is disarticulated from, but in close association with, the basioccipital. Erosion of the anteromedial portion of the prootic ex- poses the inside of the floccular recess (fossa auricularis) and the internal aspect of the ves- tibular pyramid.
The supraoccipital forms the median dor- sal border of the foramen magnum (fig. 1A). At the midline, the supraoccipital forms a transversely convex bulge that rises antero- dorsally to meet the parietals. Unlike the su- praoccipital of Struthiomimus altus (AMNH 5355) (fig. 2A), which bears a low median crest, the posterior surface of the supraoccip- ital is smooth in IGM 100/987 as it is in Gal- limimus bullatus (IGM 100/12). As in Gal- limimus bullatus IGM 100/11; IGM 100/12) and Dromiceiomimus samueli (ROM 840), the dorsal border of the median bulge is hor- izontal, reflecting the flat morphology of the skull roof. On either side of the bulge, the supraoccipital levels out laterally and forms an expanded, but dorsally concave, shelf that sutures to the mediodorsal part of the exoc- cipital. Medially, a vertical pillar of bone forms the internal wall of the braincase dor- sal to the prootic and anterior to the exoccip- ital. The supraoccipital contributes to the dorsal part of the vestibular pyramid. A large trough coursing posteroventrally along the base of the preserved part of the bone rep- resents the dorsal wall of the floccular recess (figs. 1B, G). The anterior and lateral walls of the recess are lost; its ventral extremity is preserved on the prootic. The exact shape of the floccular recess cannot be determined due to relative displacement between the bones that form it.
As in other dinosaurs, the exoccipital and opisthotic form a single compound bone with no discernible border between centers of os- sification (Currie, 1997). The exoccipital— opisthotics make up most of the foramen magnum border (fig. 1A). Ventrally, each ex- occipital—opisthotic bears a posteroventrally directed process that rests upon the basioc- cipital and contributes approximately 20% of the occipital condyle. Above the pedicel that reaches the occipital condyle, the exoccipital boundary of the foramen magnum is emar- ginated into a thin edge. Adjacent to the fo- ramen magnum, the exoccipital—supraoccip- ital suture is sinuous; a small convexity on
MAKOVICKY AND NORELL: UKHAA TOLGOD ORNITHOMIMID 3
the exoccipital fits into a small concavity of the supraoccipital margin. Lateral to the fo- ramen magnum, the posterior face of each exoccipital—opisthotic is shallowly concave but becomes convex farther laterally on the paroccipital process. The left paroccipital process is broken at its base, whereas the right one is missing only the distal tip. The latter fracture reveals that the interior of the paroccipital process is hollow. Its ventral sur- face is perforated by a large pneumatic fo- ramen, which opens into the middle ear re- gion and marks the entry of the posterior tympanic recess (fig. 1E). Posteromedial to this fossa, the exoccipital-opisthotic forms the metotic strut, which descends toward the basioccipital and forms the posterior wall of the middle ear chamber. Two large and three smaller foramina perforate the right exoccip- ital in the triangular region between the me- totic strut, the suture with the basioccipital, and the occipital condyle on the posterior cranial surface (figs. 1A, F). The two larger foramina are lateral to the smaller ones. The dorsal of the two larger foramina passes through the metotic strut into the middle ear region and represents the vagus foramen (cranial nerve X and XI). The ventral fora- men does not enter the braincase and is an accessory pneumatic feature termed the sub- condylar recess by Witmer (1997). The hy- poglossal nerve exits the braincase through the vertically oriented series of three smaller apertures near the base of the occipital con- dyle (fig. 1F).
The exoccipital—opisthotic forms a small portion of the interior braincase wall, poste- rior to the prootic and posteroventral to the supraoccipital. A tall, elongate depression is situated adjacent and parallel to the sharp edge bordering the foramen magnum (fig. 1F). A small depression is present in a sim- ilar position in Dromiceiomimus samueli (ROM 840), Dromaeosaurus albertensis (Currie, 1995), Troodon formosus (Currie and Zhao, 1994a), and Velociraptor mongo- liensis (IGM 100/976). In Struthiomimus al- tus (AMNH 5355) this pit is situated on the border of the foramen magnum and opens posteriorly (fig. 2A). Currie (1995) identified this structure as the opening of the endolym- phatic duct in Dromaeosaurus albertensis, but its position posterior to the metotic strut
AMERICAN MUSEUM NOVITATES
1998
MAKOVICKY AND NORELL: UKHAA TOLGOD ORNITHOMIMID 5
Fig. 1.
(Continued)
icm
clearly shows that it represents some other anatomical feature perhaps associated with the posterior cerebellar venous sinus (Sues, 1997). Anterior to this depression in IGM 100/987, the exoccipital forms the postero- ventral portion of the vestibular pyramid. Erosion precludes positive tracing of the semicircular canals through this bone.
Just over half of the occipital condyle is formed by the basioccipital. On each side of the condyle, the suture with the exoccipital descends ventrolaterally. The condyle is dor- soventrally flattened and not spherical as it is in troodontids (figs. 1A, 2A). A distinct neck between the condyle and braincase, as is present in troodontids, is not present in either IGM 100/987 or AMNH 5355 (figs. 1E, 2E). As in a majority of theropods, the dorsal sur- face of the condyle is transversely concave. Sues (1997) suggested that this concavity ar- ticulated with the dens or atlas centrum, but this is doubtful given the relative proportions of the dens and the occipital condyle in most nonavialan theropods and the negative al- lometry of the occipital condyle to foramen magnum ratio within the group (Currie and Zhao, 1994b). Unlike the small tablike dens in Aves (Baumel and Witmer 1993), the atlas centrum is large and subspherical in orni-
_thomimids (TMP 93.62.1). Rather, the con-
<_
Fig. 1. Braincase of ornithomimid indet. IGM 100/987 in (A) posterior, (B) anterior, (C) lateral views. Abbreviations: VII facial foramen; VIIIc exit for cochlear branch of vestibulocochlear nerve; VIIIv exit for vestibular branch of vesti- bulocochlear nerve; X vagus foramen; XII foram- ina for hypoglossal nerve; a? ?atlas fragment; ac accessory pneumatic connections; af floccular re- cess; bo basioccipital; br basisphenoid recess; bs basisphenoid; bt basal tuber; ci crista interfenes- tralis; d depression at suture between exoccipital and supraoccipital; dtg groove leading to dorsal tympanic recess; dtr dorsal tympanic recess; dv vestibule; eo exoccipital; fm foramen magnum; hf hypophyseal fossa; mf metotic fissure; mo middle ear opening; ms metotic strut; oc occipital con- dyle; pp paroccipital process; pr prootic; ps prootic recess; ptr posterior tympanic recess; q quadrate; qf quadrate fossa; qj quadratojugal; so supraoccipital; sr subcondylar recess; ss subotic recess; t tubercle.
AMERICAN MUSEUM NOVITATES NO. 3247
Fig. 1 (continued). Braincase of ornithomimid indet. IGM 100/987 in (D) dorsal, (E) ventral, (F) posteroventrolateral oblique, (G) anterolateral oblique views. See previous caption for abbreviations.
1998
MAKOVICKY AND NORELL: UKHAA TOLGOD ORNITHOMIMID 7
Fig. 1.
(Continued)
cave surface of the condyle may have formed a floor for the medulla oblongata. The slight- ly rugose articular surface of the occipital condyle wraps onto the ventral surface of the condyle in IGM 100/987. A similar feature is developed in Gallimimus bullatus (IGM 100/11) and Struthiomimus altus (AMNH 5355) and indicates that the occipital condyle articulated posteroventrally with the atlas in- tercentrum. The basal tubera of IGM 100/987 are modest in size and ventrally rounded. A shallow depression for a ligament separates the two tubera. A small tubercle is present beneath this depression (fig. 1A) between the ventral extremities of the two tubera. Currie and Zhao (1994a) interpreted a small pit in this region in Troodon formosus as marking the attachment point for a ligament from the neck musculature. Dorsolateral to each basal tuber, the basioccipital is excavated by a large pneumatic pocket from which several foramina contribute to the subcondylar recess (Witmer, 1997).
The basisphenoid of IGM 100/987 is dis- articulated and shifted a short distance away from the basal tubera. It is also broken along the median axis through the middle of the ventral recess. Although not fully preserved, this recess apparently was proportionately deep as in Gallimimus bullatus (IGM 100/ 11; Osmdlska et al., 1972) and Dromiceiom- imus samueli (ROM 840). The basisphenoid appears to be partially divided by a low, lon- gitudinal ridge on the roof of the recess in Struthiomimus altus (AMNH 5355), but this character cannot be ascertained in IGM 100/ 987. The basisphenoid forms a large part of the braincase floor anterior to the basioccip- ital and exoccipitals. Three foramina perfo- rate a lateral wing of the basisphenoid below the middle ear recess (fig. 1A, E), demon- strating that the series of pneumatic spaces in the basisphenoid (subotic recess of Wit- mer, 1997) is connected to the anterior pneu- matic spaces of the anterior tympanic recess, as in Troodon formosus (Currie and Zhao, 1994a).
Internally, the prootic forms the laterov- entral portion of the braincase wall (figs. 1B, 2D). Opposite the exoccipital—opisthotic, the posterior edge of the prootic defines the an- terior rim of the middle ear. A small, finger- like process, which is broken dorsally and
AMERICAN MUSEUM NOVITATES
NO. 3247
Fig. 2. Braincase of Struthiomimus altus (AMNH 5355) in (A) posterior, (B) anterior, (C) lateral views. See fig. 1 for abbreviations.
crosses horizontally across the top of the middle ear, may represent the dorsal part of the crista interfenestralis or promontorium (fig. 1C, F). The exit of the cochlear branch of cranial nerve VIII is found anterior and medial to this process at the preserved an-
terior rim of the metotic fissure (fig. 2D). Ab- sence of a perilymphatic foramen indicates that the anterior rim is not fully preserved. In the troodontid (IGM 100/983) collected together with IGM 100/987, the perilym- phatic opening is located on the anterior rim
1998
Fig. 2.
(Continued)
of the metotic fissure and opens toward the posterior. The opening into the vestibule lies anterolateral to the cochlear exit of the ves- tibulocochlear nerve, and the vestibule itself courses anterodorsally beneath the vestibular pyramid. A circular cross section of the ves- tibule is exposed in anterior view by the frac-
MAKOVICKY AND NORELL: UKHAA TOLGOD ORNITHOMIMID 9
ture that runs through the prootic (fig. 1B). This cross section is visible just beneath the ventral border of the floccular recess. A thin lamina of bone forms the anterior wall of the floccular recess. The space anterior to this lamina may have been a pneumatic recess, as seen in Troodon formosus (Currie and Zhao, 1994a: fig. 2). A large prootic recess (Wit- mer, 1997), which is part of the anterior tym- panic system, is present on the lateral face of the prootic anteroventral to the acoustic fossa (fig. 1F). A similar depression is found in Velociraptor mongoliensis (IGM _ 100/976), Gallimimus bullatus (IGM 100/11), Stru- thiomimus altus (AMNH 5355), and in troo- dontids (Currie and Zhao, 1994a).
A shallow groove that begins near the opening of the inner ear courses dorsolater- ally away from the middle ear and approach- es the base of the paroccipital process (fig. 1F). It leads to the dorsal tympanic recess and marks a pneumatic connection between the middle ear and a pneumatic fossa anter- odorsal to the paroccipital process. Although not preserved in IGM 100/987, such a de- pression is well developed in Dromiceiomi- mus samueli (ROM 840), Struthiomimus al- tus (AMNH 5355), and Gallimimus bullatus (IGM 100/10; 100/11; 100/12) and extends from the dorsal part of the prootic above the facial foramen on to the opisthotic, where it excavates the base of the paroccipital process anteriorly (fig. 2B, G). Dorsal tympanic re- cesses are also encountered on the prootic and opisthotic of some other advanced the- ropods (Witmer, 1997), including ovirapto- rosaurs (IGM 100/42), avialans (Walker, 1985), and some, but not all, troodontids (present in IGM 100/983 but absent in Troo- don formosus [Witmer, 1997]).
The posterior wall of the exit for the facial nerve (cranial nerve VII) is visible on the preserved anterior rim of the right prootic (fig. 1B). By comparison with AMNH 5355, two foramina on the internal face of the acoustic fossa of IGM 100/987 are here in- terpreted as the internal exits of the cochlear and vestibular branches of cranial nerve VIII (N. vestibulocochlearis). The two foramina are almost level with each other, as in AMNH 5355, and are situated posterior and slightly dorsal to the opening for the facial nerve. Damage to this region in IGM 100/
AMERICAN MUSEUM NOVITATES NO. 3247
Fig. 2 (continued). Braincase of Struthiomimus altus (AMNH 5355) in (D) dorsal, (E) ventral, (F) posteroventrolateral oblique, (G) anterolateral oblique views. See fig. 1 for abbreviations.
1998 MAKOVICKY AND NORELL: UKHAA TOLGOD ORNITHOMIMID
D
Xll
Fig. 2. (Continued)
11
12 AMERICAN MUSEUM NOVITATES
987 obscures much of the anatomy, but in AMNH 5355 the openings for the facial and vestibulocochlear nerves lie within a shallow acoustic fossa on the inner wall of the brain- case (fig. 2G). The exit for the facial nerve is set in a deep fossa on the lateral braincase wall in AMNH 5355, as in other theropod taxa.
Neither the head nor the condyle of the quadrate are preserved. The posterior margin of the preserved section of the shaft is gently bowed anteriorly (fig. 1C). It is transversely narrow and slender in comparison with the quadrates of adult specimens of Gallimimus bullatus (IGM 100/11; IGM 100/12), al- though a juvenile specimen referred to the same taxon has a transversely compressed quadrate shaft (IGM 100/10). The propor- tions of the shaft and condyles of the orni- thomimid quadrate are tall and narrow in comparison with those of dromaeosaurids (AMNH 5616) and oviraptorids (Maryariska and Osmélska, 1997). A short section of the pterygoid wing is preserved in IGM 100/987. In cross section, the quadrate shaft is sub- triangular with the anterior apex formed by the pterygoid wing. Posteriorly, a large tear- drop-shaped depression is situated near the distal end of the quadrate (fig. 1C, F). Broken surfaces reveal that the quadrate is hollow, and a slitlike foramen near the dorsal end of the depression may represent a pneumatic connection into the quadrate interior. As is typical in archosaurs, the quadrate apparently formed a portion of the anterolateral wall of the external auditory meatus in ornithom- imids. A small sliver of bone adhering to the lateral face of the quadrate shaft is the only preserved piece of the quadratojugal.
An elongate piece of bone adhering to the left, disarticulated side of the occipital con- dyle may be part of the proatlas or atlas (fig. 1A). It has no distinct features that allow positive identification.
DISCUSSION
IGM 100/987 represents the first evidence of an ornithomimid in the Ukhaa Tolgod as- semblage and adds yet another taxon to an already astounding diversity of theropod di- nosaurs from this locality (Norell et al., 1996). In some localities such as Iren Da-
NO. 3247
basu, ornithomimids are numerous and occur in bone beds. Elsewhere, as in the Dinosaur Park Formation, ornithomimids are usually found individually, but their skeletal remains are still more numerous than those of other small theropods. This appears not to be the case at Ukhaa Tolgod, where discoveries of Oviraptorids, dromaeosaurids, avialans, and troodontids outnumber the single ornithom- imid occurrence represented by IGM 100/ 987. The presence of an omithomimid in Djadokhta-like beds at Ukhaa Tolgod dem- onstrates once again that faunal differences between Djadokhta and Nemegt beds are more a matter of frequency than composi- tion.
Several characters of IGM 100/987 have relevance to the study of other advanced the- ropods. Three tympanic recesses connect to the middle ear in IGM 100/987, as in avi- alans (Whetstone, 1983; Walker, 1985; Wit- mer, 1990) and dromaeosaurids (Norell et al., 1992). Unlike avialans and troodontids (Chiappe et al., 1996), however, the posterior tympanic recess invades the base of the par- occipital process through a large foramen (fig. 1E) instead of being confined to the col- umellar recess. This condition is primitive for coelurosaurs (Witmer, 1997) and is en- countered in tyrannosaurids, dromaeosaurids (Chiappe et al., 1996), oviraptorosaurs (IGM 100/42), and Erlicosaurus andrewsi (Clark et al., 1994) as well as ornithomimids. AIl- though the part of the prootic that bears the depression for the dorsal tympanic recess is not preserved in IGM 100/987, a groove leads from the tympanic region toward this region in IGM 100/987 and Struthiomimus altus (AMNH 5355). Furthermore, a deep depression is present on the anterodorsal part of the paroccipital process in Gallimimus bullatus (IGM 100/11), Struthiomimus altus (AMNH 5355), and Dromiceiomimus samue- li (ROM 840). The dorsal tympanic recess of ornithomimids is proportionately large in comparison with the same structure in troo- dontids (IGM _ 100/983), Velociraptor mon- goliensis (IGM 100/976), and Archaeopteryx lithographica (London specimen; Walker, 1985).
The anterior tympanic recess is also well developed in ornithomimids. Structures re- lating to it include the prootic recess ventral
1998
to the foramen for the facial nerve (figs. 1E 2C, G), a feature encountered widely among coelurosaurs. In dromaeosaurs and ornithom- imids the prootic recess of the anterior tym- panic recess is positioned directly ventral to the exit of the facial nerve (Witmer, 1997). Oviraptorids have a disproportionately elon- gate braincase, and the prootic recess lies an- terior, rather than ventral, to the exit for the facial foramen (IGM 100/42). In troodontids this pocket is greatly expanded and confluent with the subotic recess and other air spaces within the basisphenoid (Currie and Zhao, 1994a: fig. 4; Norell et al., in prep.). The invasion of the basisphenoid (subotic recess) and basioccipital by air spaces in ornithom- imids is similar to, but less expressed than, the condition found in troodontids (Currie and Zhao, 1994a). The anterior tympanic re- cess enters the interior of the para-basisphe- noid mainly by way of the basisphenoidal re- cess or median pharyngeal system (Witmer, 1997) and to a lesser degree via lateral con- nections from the pneumatic space within the basisphenoid (Currie and Zhao, 1994a), al- though such connections are evident in IGM 100/987. In troodontids, the basisphenoid re- cess is lost, or ventrally enclosed as in or- nithurine birds, and the parasphenoid bulla is connected exclusively with the anterior tym- panic recess (Currie and Zhao, 1994a).
The relative positions of the foramina for
cranial nerves VII and VIII on the internal |
aspect of the braincase appear to vary among coelurosaurs. The internal foramen of the fa- cial nerve lies anteroventral to those of the vestibulocochlear nerves in AMNH 5355 and IGM 100/987. The vestibular branch of cra- nial nerve VIII is positioned anterior and only slightly ventral to the exit of the coch- lear branch. By contrast, the facial nerve ex- its the braincase directly ventral to the ves- tibular exit of cranial nerve VIII in the ho- lotype of Dromaeosaurus albertensis (AMNH 5516; Currie, 1995; fig. 6) and in Itemirus medullaris (Kurzanov, 1976; Currie, 1995). A unique condition is present in Troo- don formosus, where the vestibular branch of cranial nerve VIII does not appear to have a separate exit but passes through the braincase together with the facial nerve (P. Currie, per- sonal commun.). In the Ukhaa Tolgod troo- dontid, however, the two branches of the ves-
MAKOVICKY AND NORELL: UKHAA TOLGOD ORNITHOMIMID 13
tibulocochlear are separate from the facial nerve, as in ornithomimids. Tyrannosaurus rex also displays three separate openings in the acoustic fossa (Osborn, 1912), but, when the braincase is oriented with the foramen magnum facing posteriorly, the configuration is that of a triangle where the exit of the fa- cial nerve lies anterior to the openings for cranial nerve VIII, with the vestibular and cochlear branches of the latter nerve exiting dorsoposterior and ventroposterior, respec- tively, to the facial nerve. The two openings for cranial nerve VIII are situated posterior to the exit for cranial nerve VII in Chiros- tenotes pergracilis (Sues, 1997).
Citing a personal communication from S. Chatterjee, Currie and Zhao (1994a) stated that the vagus foramen is directed anterolat- erally in ornithomimids, but it is directed an- teromedially in both IGM 100/987 and Stru- thiomimus altus (AMNH 5355), as it is in troodontids and Dromaeosaurus albertensis (AMNH 5516). IGM 100/987 displays a third small foramen for a branch of the hy- poglossal. Such a foramen is present in some extant birds but has not been reported widely in other theropods, although it is present in Troodon formosus (Currie and Zhao, 1994a). The number of hypoglossal exits is indeter- minate in Struthiomimus altus (AMNH 5355). The number of hypoglossal foramina is highly variable in extant archosaurs, dis- playing even interspecific variability.
In many respects, ornithomimid braincase anatomy is similar to that of troodontids. Common derived features include a bulbous parabasisphenoid (not preserved in IGM 100/ 987), expansion of the anterior tympanic re- cess to invade the basisphenoid (also in Ve- lociraptor mongoliensis [IGM 100/967]), the relative positions of the facial and vestibu- locochlear nerve (also in Avialae [Witmer, 1990] and Velociraptor mongoliensis [IGM 100/967] but not Dromaeosaurus alberten- sis), and three exits for the hypoglossal (also in some extant avialans). These similarities contrast with substantial differences in the postcranial skeleton, where troodontids dis- play many maniraptoran apomorphies not present in ornithomimids.
A number of features of the ornithomimid braincase region are potentially autapo- morphic, including the presence of pneumat-
14 AMERICAN MUSEUM NOVITATES
ic subcondylar recesses that open posteriorly on the basioccipital and exoccipital, just dor- sal and lateral to the basal tubera. The pres- ence of a large depression on the posterior face of the quadrate shaft is also unique to ornithomimids among coelurosaurian dino- saurs. The quadrate is hollow, and a slitlike foramen in the depression may have con- nected the interior to the tympanic pneumatic systems, as in birds (Witmer, 1990) and some nonavialan theropods such as_ troodontids (Currie and Zhao, 1994a), tyrannosaurids (Molnar, 1985), and oviraptorids (Maryanska and Osmdlska, 1997). Examination of other ornithomimid skulls (ROM 851; ROM 840) has not revealed such a connection to the ex- terior, however, and this feature may be ei- ther an artifact or a derived character of IGM 100/987.
IGM 100/987 differs from preserved cra- nial material of other advanced ornithom- imids in a number of subtle features. The ba- sal tubera of IGM 100/987 are shallow and widely separated with a small tubercle oc- cupying the center of the cleft between them. By contrast, the basal tubera of Struthiomi- mus altus (AMNH 5355) are deeper, and sep- arated by only a narrow cleft. In Gallimimus bullatus (IGM 100/12) the basal tubera are also proportionately deep, but preservation does not permit a determination of the extent of their separation. Another difference be- tween Struthiomimus altus and IGM 100/987 is the absence of a midline ridge on the su- praoccipital of the latter taxon, a similarity it shares with Gallimimus bullatus (Osm6lska et al., 1972; IGM 100/12). IGM 100/987 dif- fers from the holotype of Gallimimus bulla- tus (IGM 100/11) in having a proportionately narrower quadrate shaft with a more devel- oped posterior fossa. The quadrate fossa is well developed in Dromiceiomimus samueli (ROM 840) and Ornithomimus edmontonicus (ROM 851). These differences clearly indi- cate that IGM 100/987 cannot confidently be referred to any known ornithomimid taxon.
Phylogenetic assessments are further com- plicated by possible ontogenetic changes in cranial characters. In a juvenile specimen re- ferred to Gallimimus bullatus (IGM 100/10) the quadrate fossa is situated more proxi- mally than it is in other known specimens,
NO. 3247
and the quadrate shaft appears relatively nar- row. It is currently not possible to assess whether these quadrate features display on- togenetic changes. Another difference be- tween the juvenile specimen IGM 100/10 and other ornithomimid skulls is the presence of a large foramen on the posterior face of the paroccipital process. Such a foramen is not widely distributed among advanced the- ropods but is encountered in some, although not all, dromaeosaur specimens (IGM 100/ 967).
Advanced ornithomimid taxa are currently diagnosed mainly by different proportions between limb bones (Russell, 1972; Barsbold and Osmdélska, 1990) and the form of the manual unguals, elements not recovered for the Ukhaa Tolgod ornithomimid. Further- more, braincase anatomy remains unde- scribed or unknown in a majority of named ornithomimid taxa, so it is possible to refer IGM 100/987 only to Ornithomimidae inde- terminate. A more exact taxonomic place- ment of IGM 100/987 must await description of new material and, above all, a reevaluation of ornithomimid phylogeny based on discrete osteological characters.
ACKNOWLEDGMENTS
We are grateful to Amy Davidson for spec- imen preparation and Mick Ellison for devel- opment of the figures. Many thanks to Larry Witmer for additional preparation of AMNH 5355, which aided in comparisons between specimens. Peter Makovicky’s work is sup- ported by the Danish Research Academy (Grant 1996-154-0020); this work was also supported by NSF DEB 9407999 and the Frick Laboratory Endowment (to MAN). Members of the 1993 Mongolian Academy of Sciences— American Museum of Natural History expe- dition are thanked for help in the field. We are grateful to Hans-Dieter Sues and Phil Currie for providing access to specimens in the ROM and TMP collections, respectively. We thank Asger Hgeg and Frank Olsen of Experimen- tariet, Copenhagen, for arranging access to IGM specimens on display, and we thank Drs. Barsbold and Tsogtbataar for their permission to study these specimens. Reviews by Larry Witmer and Hans-Dieter Sues significantly im- proved this paper.
1998 MAKOVICKY AND NORELL: UKHAA TOLGOD ORNITHOMIMID 15 REFERENCES Barsbold R. golia, People’s Republic of China. Cre- 1981. Toothless dinosaurs of Mongolia. Joint taceous Res. 14: 127-144.
Soviet-Mongolian Paleont. Trans. 15: 28—39.[In Russian].
Exped.,
1983. Carnivorous dinosaurs from the Creta- ceous of Mongolia Trans. Joint Soviet- Mongolian Paleontol. Exped. 19: 5—- 119.{In Russian].
1988. A new Late Cretaceous ornithomimid
from the Mongolian People’s Republic. Paleontol. J. 1988 (1): 124-127. Barsbold R., and H. Osmélska
1990. Ornithomimosauria; Jn Weishampel, D. B., P. Dodson, and H. Osmédlska, (eds.) The Dinosauria: 225-244. San Diego: Univ. of California Press.
Barsbold R., and Perle A.
1984. The first record of a primitive orni- thomimosaur from the Cretaceous of Mongolia. Paleontol. J. 1984(2): 118- 120
Baumel, J. J., and L. M. Witmer
1993. Osteologia. Jn J. J. Baumel, A. S. King, A. M. Lucas, J. E. Breazile, and H. E. Evans (eds.), Handbook of avian anat- omy: 45-132. Cambridge: Nuttall Or- nithol. Soc.
Berkey, C. P, and F K. Morris
1927. Geology of Mongolia. Jn Natural His- tory of Central Asia, 2: 475pp. New York; Am. Mus. Nat. Hist.
Chiappe, L. M., M. A. Norell, and J. M. Clark
1996. Phylogenetic position of Mononykus (Aves: Alvarezsauridae) from the Late Cretaceous of the Gobi Desert. Mem. Queensland Mus. 39(3): 557-582.
Clark, J. M., Perle A., and M. A. Norell
1994. The skull of Erlicosaurus andrewsi, a Late Cretaceous “‘segnosaur’’ (Thero- poda: Therizinosauridae) from Mon- golia. Am. Mus. Novitates 3115: 39 pp.
Currie, P. J.
1995. New information on the anatomy and relationships of Dromaeosaurus alber- tensis (Dinosauria: Theropoda). J. Ver- tebr. Paleontol. 15: 576-591.
1997. Braincase anatomy. Jn P. J. Currie and
K. Padian (eds.), Encyclopedia of di- nosaurs: 81-85. San Diego, CA: Aca- demic Press. Currie, P. J., and D. A. Eberth 1993. Palaeontology, sedimentology and palaeoecology of the Iren Dabasu For- mation (Upper Cretaceous), Inner Mon-
Currie, P. J., and Zhao, X.-J. 1994a. A new troodontid (Dinosauria, Thero- poda) braincase from the Dinosaur Park Formation (Campanian) of Alberta. Can. J. Earth Sci. 30: 2231-2247. 1994b. A new carnosaur (Dinosauria, Thero- poda) from the Jurassic of Xinjiang, People’s Republic of China. I[bid.: 2037-2081. Dashzeveg D., M. J. Novacek, M. A. Norell, J. M. Clark, L. M. Chiappe, A. Davidson, M. C. McKenna, L. Dingus, C. Swish- er, and A. Perle
1995. Unusual preservation in a new verte- brate assemblage from the Late Creta- ceous of Mongolia. Nature 374: 446—
449.
Gilmore, C. W. 1933. On the dinosaurian fauna of the Iren
Dabasu Formation. Bull. Am. Mus. Nat. Hist. 67: 23-78. Jerzykiewicz, T., and D. A. Russell
1991. Late Mesozoic stratigraphy and verte- brates of the Gobi Basin. Cretaceous Res. 1991(12): 345-347.
Kurzanov, S.
1976. Braincase structure in the carnosaur Ite- mirus n. gen. and some aspects of the cranial anatomy of dinosaurs. Paleon- tol. J. 1976(3): 127-137.
Makovicky, P. J.
1995. Phylogenetic aspects of the vertebral morphology of Ceolurosauria (Dino- sauria: Theropoda). Unpubl. M. S. diss., Copenhagen Univ., Denmark.
Maryaniska, T., and H. Osmdlska
1997. The quadrate of oviraptorid dinosaurs.
Acta Palaeontol. Pol. 42(3): 361-371. Molnar, R. E.
1985. Alternatives to Archaeopteryx: a sur- vey of proposed early or ancestral birds. In M. K. Hecht, J. H. Ostrom, G. Viohl, and P. Wellnhofer (eds.), The be- ginnings of birds: 209-217. Eichstatt: Freunde des Jura-Museums.
Norell, M. A., J. M. Clark, and A. Perle
1992. New dromaeosaur material from the Late Cretaceous of Mongolia. J. Ver- tebr. Paleontol. 12(suppl. 3): 45A.
Norell, M. A., J. M. Clark, and L. M. Chiappe
1996. Djadokhta series theropods: a summary
16 AMERICAN MUSEUM NOVITATES
review. Dinofest Int. Symp.: Prog. and Abstr.: 86. Norell, M. A., P. J. Makovicky, and J. M. Clark In prep. A new troodontid from Ukhaa Tolgod (Late Cretaceous: Mongolia), associat- ed remains and the diagnosis and phy- logeny of troodontid dinosaurs (Ther- opoda: Maniraptora). Osborn, H. F 1912. Crania of Tyrannosaurus and Allosau- rus. Mem. Am. Mus. Nat. Hist. 1, pt. 1: 30 pp. Osmolska, H., E. Roniewicz, and R. Barsbold
1972. A new dinosaur, Gallimimus bullatus n. gen., n. sp. (Ornithomimidae) from the Upper Cretaceous of Mongolia. Pa- laeontol. Pol. 27: 103-143.
Parks, W. A.
1928. Struthiomimus samueli, a new species of Omithomimidae from the Belly Riv- er Formation of Alberta. Univ. Toronto Stud. (Geol. Ser.) 26: 1-24.
Raath, M. A. 1985. The theropod Syntarsus and its bearing
on the origin of birds. Jn Hecht, M. K.., J. H. Ostrom, G. Viohl, and P. Well- nhofer (eds.), The beginnings of birds: 219-227. Eichstatt: Freunde des Jura- Museums.
NO. 3247
Russell, D. A.
1972. Ostrich dinosaurs from the Late Creta- ceous of western Canada. Can. J. Earth Sci. 9: 375-402.
Sues, H.-D.
1997. On Chirostenotes, an oviraptorosaur (Dinosauria: Theropoda) from the Up- per Cretaceous of western North Amer- ica. J. Vertebr. Paleontol. 17(4): 698— 716.
Walker, A.
1985. The braincase of Archaeopteryx. In Hecht, M. K., J. H. Ostrom, G. Viohl, and P. Wellnhofer (eds.). The begin- nings of birds: 123-134. Eichstatt: Freunde des Jura-Museums.
Whetstone, K. N.
1983. Braincase of mesozoic birds: I. New preparation of the “‘London’”’ Archae- opteryx. J. WVertebr. Paleontol. 2(4): 439-452.
Witmer, L. M.
1990. The craniofacial air-sac system of Me-
sozoic birds (Aves). Zool. J. Linn. Soc. 100: 327-378. Craniofacial air sinus systems. Jn Cur- rie, P. J. and K. Padian (eds.) Encyclo- pedia of dinosaurs: 151-159. San Die- go, CA: Academic Press.
1997.
Recent issues of the Novitates may be purchased from the Museum. Lists of back issues of the Novitates and Bulletin published during the last five years are available at World Wide Web site
http://nimidi.amnh.org. Or address mail orders to: American Museum of Natural History Library, Central Park West at 79th St., New York, NY 10024. TEL: (212) 769-5545. FAX: (212) 769- 5009. E-MAIL: scipubs @ amnh.org
This paper meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper).