Tyrannosaurus rex, Osborn, 1905

Peter Larson, 2008, Atlas of the skull bones of Tyrannosaurus rex, Tyrannosaurus rex, the tyrant king, lndiana University Press, pp. 233-243 : 233-241

Tyrannosaurus rex was described by Osborn in a series of papers at the beginning of the last century (1905, 1906, 1912, 1916).

His work was primarily based on 3 specimens: the holotype ( AMNH 973, now CM 9379 ), the holotype for Dynamosaurus ( AMNH 5866, now BMNH R7994) , which was synonymized with T. rex in 1906, and AMNH 5027 , which provided much of the description of the skull. Interestingly, the designation of AMNH 5027 has recently come under scrutiny; it may actually represent a second, unnamed species of Tyrannosaurus (P. Larson this volume). Molnar’s (1991) treatise on the skull added greatly to Osborn’s earlier descriptive work, and it also contained a section on arthrology. Molnar studied LACM 23844 , AMNH 5027 , MOR 008 , and SDSM 12047 . In 1998 and in this volume, Molnar increased our understanding of skull mechanics and musculature. This was followed by Brochu’s (2003) description of FMNH PR2081 , including computed tomography of the articulated skull, presented significant new data and interpretations. Several other important contributions to the skull of tyrannosaurids include Carr (1999) on craniofacial ontogeny, Currie (2003) on cranial anatomy of tyrannosaurid dinosaurs (particularly Gorgosaurus and Daspletosaurus ), Currie et al. (2003) on skull structure, Hurum and Sabath (2003) on a comparison of Tarbosaurus and Tyrannosaurus , Carr and Williamson (2004) on tyrannosaurid diversity, and Carr et al. (2005) on a description of a new tyrannosaurid from Alabama.

Although several new Tyrannosaurus rex skulls are now available for study, by far the most significant is the disarticulated, undistorted skull of BHI 3033 (see N. L. Larson this volume). For the first time, the individuals bones of the skull can be described and illustrated for Tyrannosaurus rex . This specimen was discussed by Hurum and Sabath (2003) in their comparison of Tyrannosaurus rex and Tarbosaurus bataar . Smith (2005) also utilized BHI 3033 in his work describing the heterodont dentition of T. rex . This subject was also discussed in some detail in Osborn (1916), Molnar (1991), and Brochu (2003). P. Larson (this volume) has utilized this specimen in his discussion of an avianlike kinetic palate in T. rex .

Here, I present the first in-depth supplemental description of the skull bones and illustrate them in multiple views in the supplemental CD-ROM, based primarily on BHI 3033 , with observations of other specimens.

Dermal Skull Bones

Premaxillae: The premaxillae of BHI 3033 are typical for Tyrannosauridae and for Tyrannosaurus rex , bearing 4 alveoli. Details not discussed by Osborn (1905, 1906, 1912), Molnar (1991), Brochu (2003), and Hurum and Sabath (2003), but visible in BHI 3033 , include a scalloped region of ridges on the articular surface shared by the right and left premaxillae. This scalloped region is centrally located dorsoventrally and is at the anterior border of the suture. Although nearly 80% of this suture is smooth, ligaments attached to the scalloped area could have restricted lateral movement along this surface, in effect bonding the 2 premaxillae into a single structural unit. Below this scalloped area, the premaxillae separate, leaving a cleft in the articulation. The significance of this cleft is unclear, and it is not seen in the articulated FMNH PR2081 ( Brochu 2003), a robust morphotype (P. Larson this volume).

The right premaxilla shows erosion surrounding the anterior portion of the fourth alveolus. This erosion is associated with the deposition of spongy bone. Osteomyelitis is the probable cause of this pathology.

The palatal surfaces of the premaxillae also bear a feature analogous to the interdental plates Osborn 1912) of the maxillae and dentaries. These interdental plates are found at the junctions of the alveoli. They are depressed from the palatal surface of the premaxillae and separated from one another at the base by large nutrient foramina.

The articular surface of the premaxillary-maxillary suture is smooth, interrupted (in the center of the lateral aspect) only by the opening for the subnarialis foramen ( Carr 1999; Brochu 2003), whose border is shared by the 2 elements. The premaxillary-nasal articulation is also quite smooth, allowing dorsoventral movement while restricting anteroposterior movement. It is probable that the premaxillae unit was at least passively kinetic, movable relative to the maxillae and nasals, providing shock-absorbing benefits for the premaxillary teeth.

Maxillae: The maxillae of BHI 3033 bear 11 alveoli. This number is shared with MOR 980 , although to date, all other Tyrannosaurus rex specimens have 12 (P. Larson this volume). The 2 maxillae of BHI 3033 articulate near the front of the palate in a series of overlapping ridges not mentioned by Molnar (1991) or Brochu (2003). This is similar to the condition in Tarbosaurus bataar ( Hurum and Sabath 2003).

In Tyrannosaurus , the antorbital fossa is a very deep depression in even the smallest individuals (i.e., BHI 4100 and LACM 2345 ). The ventral border of the antorbital fossa and the antorbital fenestra are coincident along much of the ventral border of the antorbital fenestra.

In most tyrannosaurs, like Gorgosaurus (TCM 2001.89.1; Carr 1999), Daspletosaurus (RTMP 91.36.500; Currie 2003), Appalachiosaurus (RMM 6670; Carr et al. 2005), and Nanotyrannus (CMNH 7541, BMRP 2001.4.1; Bakker et al. 1988; Larson in press), the antorbital fossa is shallower than T. rex , and a thin ridge of bone (the ventral antorbital maxillary ridge) rises along the dorsal margin of the posteroventral extension of the maxilla. It extends well past the last alveolus and passes under the jugal where it articulates to the maxilla. In Gorgosaurus , Daspletosaurus , Appalachiosaurus, and Nanotyrannus , the ventral border of the antorbital fenestra and the antorbital fossa are not coincident. The constriction at the base of the ventral antorbital maxillary ridge forms the ventral border of the antorbital fossa, and the top of the ridge is the ventral border of the antorbital fenestra (Larson in press)

The maxillary fenestra (the second antorbital fenestra of Osborn 1912) in Tyrannosaurus BHI 3033 ) and Tarbosaurus ( Hurum and Sabath 2003) contacts the anterior and posterior border of the antorbital fossa. An additional opening between the maxillary and antorbital fenestrae is found on both maxillae of BHI 3033 , but is not seen in other specimens of T. rex . A similar opening is seen in a Tarbosaurus specimen, ZPAL MgD-I/4, described by Hurum and Sabath (2003). This opening may be the result of predepositional weathering and breakage of an extremely thin area of bone.

A long, narrow, and deep abrasion is incised into the lateral aspect of the left maxilla. This gouge is oriented dorsoventrally and is directly above the third maxillary tooth near the center of the mass of the maxilla. The abrasion measures 12 cm long by 2 cm wide and is approximately 1 cm deep at the center of the scar. There is evidence of new bone growth, especially near the ventral end of the gouge, demonstrating that this mark was made some days or weeks before death.

Several perforations mar the anterior palatal shelf of the right maxilla. It could be that these are a result of postmortem weathering or a fungal infection. They do not seem to have been caused by osteomyelitis because cancellous bone is not evident.

The nasal-maxillary suture is highly scalloped on BHI 3033 . Interlocking fingers of bone prevent any anterioposterior movement along this articulation. There is, however, the possibility of lateral movement, with the nasal-maxillary suture acting as a hinge, allowing the ventral portion of the maxilla to swing outward. If such movement were to occur, the interlocking ridges of the maxilla-maxilla palatal contact would have spread apart. Such kinetic movement would enhance the shock-absorbing abilities of the maxillae and teeth.

Movement is also possible at the jugal-maxillary suture. Here the bones are joined by a tongue (jugal) and groove (maxilla) joint. This joint allows anterioposterior slippage.

Lachrymals: One of the characters uniting Tyrannosaurus with Tarbosaurus , but separating it from other tyrannosaurs, such as Daspletosaurus , Gorgosaurus , and Albertosaurus , is the absence of a corneal process on the lachrymal ( Carr 1999; Carr and Williamson 2004). This character also separates Tyrannosaurus from Nanotyrannus (Larson in press). An additional character is the shape of the lachrymal: it is an inverted L shape in Tyrannosaurus , as it is in Tarbosaurus , and T shaped in Albertosaurus , Gorgosaurus , Daspletosaurus ( Carr et al. 2005), and Nanotyrannus (Larson in press).

In addition to the pneumatic “lateral foramen” mentioned by Molnar (1991), the lachrymals of BHI 3033 possess 3 additional pneumatic openings. The first of these pneumatopores lies anterior to the lateral foramen, near the front of the lachrymal. It is on the dorsal margin of the antorbital fenestra and opens ventrally. In medial view, the lachrymals of BHI 3033 expose 2 more pneumatic foramina. The medial lachrymal pneumatopore (Larson in press) opens anterior to and at the dorsal margin of a thin ridge that descends diagonally across the vertical ramus. This ridge terminates at the ventral, anterior border of the vertical ramus. Finally, a fourth pneumatopore (also medial) may be found at the front of the horizontal ramus. This pneumatopore opens anteriorly.

The lachrymals barely touch the ascending ramus of the maxilla, thus allowing movement. However, the nasal-lachrymal articulation locks quite firmly, with a somewhat scalloped surface on the medial aspect of the horizontal ramus of the lachrymal and a cleft on the lateral aspect of the horizontal ramus that receives the lachrymal process of the nasal ( Hurum and Sabath 2003). The posterior surface at the junction of the horizontal and vertical rami is somewhat ball shaped. This ball inserts into a socket in the anterior surface of the frontals, suggesting possible prokinesis, or lifting of the muzzle ( Larson and Donnan 2002).

Nasals: As is the case with all tyrannosaurids, the nasals of BHI 3033 are fused. This fusion, as well as lateral arching and the consistent thickness of the bone (2 cm or more over nearly the entire length of the nasals), provides an extremely strong structure for the dissemination of stress developed during feeding (see Rayfield 2004). The nasal-frontal suture is a tongue (frontal) and groove (nasal) joint that allows fore-and-aft movement and would not preclude prokinesis.

Postorbitals: The medial aspects of the postorbitals show a shallowcentral depression, or fossa. Other Tyrannosaurus specimens ( BHI 4812 , MOR 555 , and MOR 980 ) also show this feature. Much deeper fossae are seen in Gorgosaurus (TCM 2001.89.1), Daspletosaurus ( Currie 2003) , and Nanotyrannus (Larson in press).

Two paired, isolated osteoderms were found associated with the skull of BHI 3033 ( Larson et al. 1998). These dermal elements are analogous to the “postorbital rugosity” noted as fused to the postorbitals of some robust Tyrannosaurus specimens by Molnar (1991), Larson (1994), and Brochu (2003). Tyrannosaurus specimens with fused postorbital rugosities include MOR 008 , FMNHPR2081 , BHI 4182 , and UWGM 181 (NS 1565.26) . On BHI 3033 , these loose postorbital rugosities articulate with the postorbitals on the lateral surface directly above the orbit.

The anterior surface of the postorbitals (above the orbit) and the posterior surface of the horizontal ramus of the lachrymals articulate with pyramidally shaped osteoderms, or horns. These horns were found as isolated elements in BHI 3033 . They were noted by Brochu (2003) as fused to the postorbitals in FMNH PR2081 and were thought to be part of the postorbital rugosity of Molnar (1991). It is clear from BHI 3033 (a set of isolated horns was also found with MOR 1125 ) that they are independent dermal elements. These horns bridge the gap between the postorbital and the lachrymal.

The postorbital-frontal suture is interfingered and allows no movement. The joint between the postorbital and the jugal is smooth and planar, allowing diagonal slippage. The postorbital-squamosal joint is an expanding tongue and groove, which allows fore-and-aft separation, although kinesis at this point is difficult to reconstruct.

SQUAMOSALS: The central portions of the anterior aspect of the squamosals are perforated by a large pneumatopore (most clearly seen in BHI 4100 ). This pneumatopore is also present in Tarbosaurus ( Hurum and Sabath 2003), but it is absent in Gorgosaurus , Nanotyrannus (Larson in press), and Daspletosaurus ( Currie 2003) . Just anterior of center, the lateral ventral borders of the V-shaped squamosals are marked by a deep fold. This fold corresponds to a deep concavity on the squamosal process noted by Molnar (1991) on the quadratojugal. These 2 surfaces could have been the origin and insertion for a ligament connecting the 2 elements (see Quadratojugals, below). The articulation of the squamosal with the exoccipital is fairly flat and may have allowed limited sliding motion. The squamosal-quadrate joint is well developed and is a double ball-in-socket (quadrate-in-squamosal) joint. The 2 balls are side by side and are connected by a saddle that is reversed in the squamosal. This joint gives great flexibility fore and aft, but limits lateral movement.

JUGALS: Both jugals have pathological, healed puncture wounds (Larson 2001). These may have been the result of face-biting behavior described by Tanke and Currie (1998). The injury to the left jugal is a circular hole, 3.5 cm by 2.5 cm, that penetrates at a point near the origin of the cleft that receives the quadratojugal (here the jugal is approximately 4 cm thick). In medial aspect, the perforation emerges, leaving an opening 2.7 cm by 1.5 cm. The injury to the right jugal is located just anterior of center of the ascending ramus, at the ventral termination of the postorbital-jugal joint. The lateral opening measures 2 cm by 2.5 cm. The medial exit is larger (3.5 cm by 6.5 cm) and somewhat triangular in shape. Molnar (1991) showed that there is pneumatic sinus in this area (also seen in BHI 3033 ) that separates the lateral layer from the medial layer of bone. The larger exit hole may be explained by the thinness of that medial layer.

The jugal-ectopterygoid joint is smooth, allowing palatal kinesis (see Larsson this volume). The articulation of the jugal and the quadratojugal is also smooth, marked by the cleft mentioned above. A horizontal ridge on the lateral aspect of the jugal process of the quadratojugal fits into this cleft. Fore-and-aft movement of the quadratojugal along this joint, with the firmly attached quadrate (see below) rocking on the double ball-and-socket squamosal-quadrate joint described above, would produce streptostyly.

Quadratojugals: The left quadratojugal is isolated, and the right one is firmly attached to the quadrate. Although Molnar (1991, p. 161) presumed that “no articulation with the squamosal existed,” there might have been a ligamental attachment (see Squamosals, above) that joined the 2 elements. This ligament could have acted as a tensile spring, returning the quadrate-quadratojugal to rest position after a streptostylic extension. The quadratojugal-quadrate articulation is somewhat edentate, creating a nonmovable joint. This suture also accounts for the firm attachment of the right quadratojugal to the right quadrate in BHI 3033 and other disarticulated skulls (i.e., MOR 1125 ).

Quadrates: The quadrates of BHI 3033 are both preserved with a small notch (1 cm in diameter) on the dorsal edge of the squamosal process. This notch lies just anterior to the articulation with the quadrate. The notch probably corresponds to a much larger notch (several centimeters in diameter) found in Gorgosaurus , Nanotyrannus (Larson in press), Albertosaurus , and Daspletosaurus ( Currie 2003) .

The quadrate-pterygoid joint surfaces are smooth, allowing streptostyly (hinted at by Molnar 1991).The proximal end has a double articular surface (see Squamosals, above). This is similar to the situation in Gorgosaurus (TCM 2001.89.1) and Daspletosaurus ( Currie 2003) , whereas Albertosaurus (BHI 6234) and Nanotyrannus have only a single ball-in-socket articulation. Although all tyrannosaurs exhibit the capacity for streptostyly, it seems that it was accomplished in different ways.

The double condylar surface of the quadrates at the quadrate-articular joint is oriented in such a way that it forms a kind of screw on the surface of the main joint for the opening and closing of the jaws. As the jaws opened, this screw would, in effect, widen the gape of the jaws by forcing the articulars away from each other. Likewise, this screw would bring the rear of the jaws back together again as the jaws closed ( Molnar 1991).

Palatal Complex

Pterygoids: Both pterygoids are complete. The articulation with the quadrate is discussed above. The rear of the palatal plate wraps over the basisphenoid, limiting posterior movement. Although the dorsal surface of the palatal plate of the pterygoid is fairly smooth, little movement probably occurred at the ectopterygoid-ptervgoid joint, and the folding of the quadrate process of the pterygoid blocked the ectopterygoid from moving posteriorly relative to the pterygoid. The folded posterior pterygoid-palatine joint allowed fore-and-aft movement but restricted lateral movement. The smooth anterior pterygoid-palatine joints allow fore-and-aft movement. The vomerine process of the pterygoid is a smooth vertical blade that touches the palatine on its lateral surface and the vomer on its medial surface The pterygoid-vomer suture is bounded by lateral extensions of the bifurcate stem. These 1-cm-deep lips limit forward movement of the pterygoid relative to the vomer. See Larsson (this volume) for a discussion of palatal kinesis.

Epipterygoids: The epipterygoids articulate on the anterior surface of the vertical quadrate process of the pterygoids. The epipterygoids of BHI 3033 are similar to those described for Daspletosaurus ( Currie 2003) . The angle of the inverted V in BHI 3033 is, however, slightly greater than that of Dasplatosaurus, and the concavity at the base, mentioned by Currie (2003), is much deeper in Tyrannosaurus rex to such an extent that the 2 arms of the V are actually bifurcated. The more medial wing of the epipterygoid butts against the anterior medial edge of the quadrate process of the pterygoid, continuing the curve, established by the medial edge of the quadrate process. This butt joint restricts movement (other than folding) at the epipterygoid-pterygoid joint. The more dorsal epipterygoid-laterosphenoid articulations are smooth and would allow movement during palatal kinesis.

Ectopterygoids: The hook-shaped ectopterygoids of BHI 3033 are perforated with a large pneumatopore typical of theropods ( Molnar 1991). Carr et al. (2005) noted that the thick lip that bounds the pneumatopore was a character for Tyrannosaurus . The anterior limb of the ectopterygoid contacts the ventral medial surface near the center of the jugal. The smooth ectopterygoid-jugal joint is critical for palatal kinesis (see Larsson this volume).

Palatines: The palatine was incompletely described by Molnar (1991). Although both palatines are complete in BHI 3033 , they were somewhat weathered before burial, making a complete assessment difficult. BHI 4100 , however, does include an exceptionally well-preserved right palatine that aids in the interpretation of BHI 3033 . This palatine has a large pneumatopore on the central lateral surface near the articulation with the maxilla, which opens into a vast chamber. There is also a deep pneumatic fossa directly anterior to this pneumatopore and of near-equal size. The palatines of Tyrannosaurus are much more inflated than those of Daspletosaurus ( Currie 2003; Carr etal. 2005), Albertosaurus (Carr etal. 2005), Gorgosaurus (TCM 2001.89.1), and Nanotyrannus (BMRP 2002.4.1). However, their similarity to Tarbosaurus ( Hurum and Sabath 2003) could indicate that this inflation is a function of size.

The palatine-maxilla joint is slightly grooved, with corresponding low tongues on the opposing bones. This joint would allow limited movement fore and aft and would not preclude lateral movement of the maxillae. The palatines also contact the anterior medial surface of the jugals and the ventral medial surface of the lachrymals in a smooth joint that would not restrict palatal kinesis or lateral movement.

Vomer: The vomer of BHI 3033 matches Molnar’s (1991) reconstructed description. The anterior rhomboid plate in BHI 3033 is convex in dorsal aspect and concave ventrally. This rhomboid plate also sports a small healed puncture wound near the base ( Larson and Donnan 2002).

Braincase, Including Skull Roof

BHI 3033 has a relatively complete braincase that is missing only the ventral portion of the basicranium (as a result of predepositional weathering). As in FMNH PR2081 , the prefrontals are firmly affixed to the frontals ( Brochu 2003). The subadult MOR 1125 has loose prefrontals, and a mature individual of T. rex A (see N. L. Larson this volume) preserves both frontals as disarticulated elements. A healed puncture wound is located near the center of the supraoccipital crest of the left parietal. This 3-cm-diameter hole perforates what was originally 3 cm of dense bone. Immediately above the perforation, a 12 cm by 5 cm fragment of bone is missing from the edge of the supraoccipital crest, presumably the result of the same incident causing the perforation (Larson 2001; Larson and Donnan 2002).

Mandibles

Articular: Only the left articular was recovered from BHI 3033 ; it was tightly articulated to the left surangular. The articular surface of the exposed right articular-surangular suture shows a great deal of interfingering, eliminating the possibility of movement at this suture. Likewise, the nature of the articulation with the prearticulars reveals that the articular was firmly locked into position in life.

239

Prearticulars: The prearticulars thin anteriorly to approximately 2 mm in lateral dimension for the final third of their length, although they measure an average of 8 cm in width. Both of the prearticulars show small (several millimeters), possibly pathologic perforations near their anterior terminations.

Angulars: The angulars of BHI 3033 match Molnar’s (1991) description. The left angular shows some minor remodeling on the anterior most surface, corresponding to a puncture wound (where it articulates) on the left surangular. The spatulate posterior medial surface articulates with the surangular dorsally and the prearticular ventrally. Anteriorly, the angular thins dorsoventrally and is sandwiched between the dentary (laterally) and the splenial (medially). All these surfaces are smooth and allow fore-and-aft movement.

SURANGULARS: Both clam-shaped surangulars show healed puncture wounds. The right surangular has a large, remodeled perforation just anterior to and the same size as the surangular fenestra; a second perforation is on the anterior edge of this surangular. The left surangular is even more pathologic, with 4 large perforations. All show evidence of healing. They are rounded and thickened at the edges by remodeling. This is not unusual for Tyrannosaurus , and perforations of this large, thin element are found in nearly every preserved surangular. FMNH PR2081 (Larson 2001), LACM 23844 ( Molnar 1991), AMNH 5027 (Osborn 1916), MOR 555 ( Horner and Lessem 1993), MOR 008 , MOR 980 , BHI 4812 , and BHI 6230 all sport extra perforations, thought to be evidence of face biting ( Tanke and Currie 1998; Larson 2001; Rothschild and Molnar this volume).

Dentaries: The dentaries each have 13 alveoli. They also both exhibit pathologies. The right dentary has a healed puncture, and a healed puncture and tear near the posterior margin. There is also a 4.5 cm by 1.5 cm erosion located 5 cm below the fifth alveolus. This erosion on the right dentary shows no healing or spongy bone and may be evidence of a fungal infection. The left dentary shows a 4 cm by 1 cm abrasion located 5 cm below the 13th alveolus. This abrasion does show7 evidence of new bone growth.

The skull of Tyrannosaurus (and other theropods) are often reconstructed with the symphysis of the lower jaws widely separated, sometimes by as much as 30 cm (i.e., LACM 23844 ). This seems to be an attempt to force the teeth of the lower jaws to occlude with those of the upper. Of course, theropods are not crocodiles, and their teeth did not occlude in life. Rather, the teeth of the lower jaws pass medial to those of the upper jaws as the jaws close. Evidence for this is presented by Molnar (1991, p. 143), who noted, “The medial face of the maxilla bears a number of shallow depressions... assumed to have accommodated the tooth crowns [of the dentary] when the mouth was closed, as in alligatoroids.” Thus, the symphysis of the lower jaws were joined by a ligamentous attachment that allowed only limited movement for the purpose of absorbing the shock of teeth hitting bone.

Splenials: Both splenials of BHI 3033 are preserved as separate elements showing the lateral surface which articulates against the dentary. A ridge on the dorsal anterior surface of the splenial fits into and becomes part of the Meckelian groove of Osborn (1912). This seems to indicate that there is no movement between the dentary and splenial. The dorsalmost portion of the splenial covers the coronoid, just posterior to the last alveolus of the dentary.

CORONOIDS: Only the right coronoid was recovered, but it was preserved in its entirety. Molnar (1991, p. 155) accepted Osborn’s (1912) interpretation of the coronoid as “a small triangular plate laying at the antidorsal angle of the Meckelian fossa.” However, because the coronoid passes behind the splenial in AMNH 5027 (and in other articulated Tyrannosaurus lower jaws), Osborn’s interpretation was incorrect. The coronoid actually continues behind the splenial and overlaps the interdental plates of the dentary in what Osborn (1912) called the supradentary. Currie (2003), noting that the splenial and supradentary were joined, and called the entire structure a “coronoid-superdentary” in his description of Daspletosaurus . Hurum and Sabath (2003) referred to it as a “superdentary/coronoid.”

Brochu (2003) provides a long discussion about whether this element is a single bone (the coronoid) or 2 separate bones (the coronoid and supradentary) that fused during ontogeny. His argument that no immature tyrannosaur individuals display these as separate elements seems to indicate that this is a single element and there is no supradentary. Certainly in Tyrannosaurus ( BHI 3033 , BHI 4812 , MOR 1125 , FMNH PR2081 , etc.), Daspletosaurus ( Currie 2003) , Gorgosaurus (TCM 2001.89.1), and Tarbosaurus (ZPAL MgDI/4; Hurum and Sabath 2003) the “coronoid-supradentary” is a single bone. The same is true for other theropods like Acrocanthosaurus (NCSM 14345) and the allosauroid Sinraptor ( Currie and Zhao 1993, p. 2055) where this element is illustrated as a “ceratobranchial.” Even in the prosauropod Plateosaurus it has been interpreted by Galton (1990) as a single element: the coronoid. Likewise, Romer illustrates the medial aspect of the lower jaws of Labidosaurus (1956a) , Kotlassia , Diadectes , Bradysaurus , Python , Peloneustes , Dimetrodon , and Edaphosaurus (1956b) as all possessing a coronoid with a thinning anterior projection that passes over the medial aspect of much of the tooth row—the same situation seen in theropods. It is for these reasons that the “coronoid-supradentary” should simply be referred to as the coronoid. The coronoid for BHI 3033 closely resembles that of Tarbosaurus , described by Hurum and Sabath (2003).

Conclusion

The disarticulated skull ofTyrannosaurus rex specimen known as BHI 3033 preserves many details not seen in other specimens. A close examination reveals details of the air sac system that invade or leave its impressions on many of the bones of the skull. Pathologies—evidence of disease or healed injuries—found on some of the cranial elements provide insight into behavior, such as intraspecific combat (face biting) and the commonality of certain pathogens such as osteomyelitis. And finally, the exquisite preservation of the cranial joint surfaces on BHI 3033 opens the door to a more complete analysis of the capacity for cranial kinesis in Tyrannosaurus rex Limited discussions of some aspects cranial kinesis in T. rex may be found in Larson and Donnan (2002) and Larsson (this volume).