Tyrannosaurus rex, Osborn, 1905

Carpenter, Kenneth, 1990, Variation in Tyrannosaurus rex, Dinosaur Systematics: Perspectives and Approaches, Cambridge: Cambridge University Press, pp. 141-145 : 141-144

Skulls

Six partial and complete skulls are now known for Tyrannosaurus rex ,

including the holotype CM 9380 (formerly AMNH 973) , AMNH 5027 , LACM 23844 , MOR 009 , SDSM 12047 , and TMP 81.6.1 . Two additional specimens, AMNH 5029 and AMNH 5117 , are braincases ( Osborn 1912). Five of the skulls are shown in Figure 10.1 . As may be seen, there is a considerable amount of variation in the size and shape of all the cranial openings (e.g., orbit and lateral temporal fenestra), as well as all the individual elements, including the lacrimal, postorbital, quadratojugal, jugal, and surangular. In fact, no two specimens are identical.

Seven maxillae were available for comparison. Five of these were overlain in Figure 10.2A using the anterior-most margins of the maxilla and maxillary fenestra as standard lengths. Most of the maxillae resemble one another, except for TMM 41436-1 , the specimen reported from the Tornillo Formation by Lawson (1976).

Variation can be seen in the depth of the maxilla, size and shape of the maxillary and antorbital fenestrae, position of the lacrimal processes, and the position and shape of the jugal process ( Fig. 10.2A ). Differences in the depth of the maxilla affect the position of the lacrimal process, which in tum influences the height and shape of the antorbital fenestra. There is also a considerable amount of variation in position of the jugal process that forms the lower rim of the antorbital fenestra, but this does not seem to correlate with the size of the antorbital process. The shape of the maxillary fenestra ranges from almost oval ( CM 9380 ) to subtriangular ( LACM 23844 and TMP 81.6.1 ) to almost square ( AMNH 5012 ). There is little or no difference in the position of the largest teeth and therefore this feature is not affected by the depth of the maxilla or the shape of the antorbital fenestra. Even the left and right maxilla of the same skull show variation ( Figs. 10.2B,C ).

The maxilla from the Tornillo Formation lacks the anterior-most margin and was scaled to the other maxilla using height and the anterior margin of the maxillary fenestra ( Fig. 10.2A ). The nasal, or dorsal, margin of the maxilla arcs sharply down suggesting a face considerably shorter than in T. rex . Other differences include a slightly larger maxillary fenestra in proportion to the size of the maxilla and a much deeper jugal process. The antorbital fenestra appears to be smaller, but the lacrimal process is too incomplete to be certain. These differences are great enough to suggest that TMM 41436-1 falls outside the range of individual variation for Tyrannosaurus rex . It may belong to a new genus, but no name should be proposed until additional material is available. Because it is unusual, it is excluded from discussions on variation.

Variation in the dentary is demonstrated by six specimens ( Fig. 10.3 ). The distance between the anterior tip and the highest part of the dentary were used for the standard length. The greatest difference is the depth of the posteroventral margin of the dentary. It is relatively shallowest in SDSM 12047 and deepest in AMNH 5027 ( Fig. 10.3 ). However, AMNH 5027 also has a tooth row that curves more ventrally than the others and this may have had an influence on the posteroventral margin. The posterior margin of the dentary is thin and easily damaged, although three of the dentaries have complete margins ( CM 9380 , AMNH 5027 , and TMP 81.6.1 ). The shape of the anterior portion of the external mandibular fossa (expressed as a notch at the lower edge of the posterior margin, see Fig. 10.1 ) is variable. The teeth seem to show more variation in positional size, but this may be due to how the dentaries were overlain.

Cervicals

Complete cervical series are known for only two specimens, AMNH 5027 and BM(NH) R7994 . There is a considerable amount of variation in the shape of the neural spines. Similar variation appears to be present in the tyrannosaur Albertosaurus libratus , so the use of neural spine shape in the diagnosis of the tyrannosaur Daspletosaurus torosus ( Russell 1970) is suspect.

The neck is more robust in BM(NH) R7994 than in AMNH 5027 ( Fig. 10.4 ). This is especially evident in the atlas and its intercentrum, and in the neural spines of cervicals two and three.

Asymmetrical co-ossification of the last cervical and first dorsal in AMNH 5027 has not been observed in any other specimen, and is probably pathological.

Ischiae

Three ischia were available for comparison. Length was standardized and the articular surface of the iliac peduncle was used to determine orientation in Fig. 10.5 . Points of variation include the relative size of the iliac peduncle, the relative size and position of the pubic peduncle, the position and size of the obturator process, and the development of the insertion scar for the M. flexor tibialis internus part 3. This scar is best developed in CM 9380 where it forms a prominent ridge.

Of the two types of ischia, one ( Fig. 10.5A ) is oriented more ventrally from the horizontal articular surface of the iliac peduncle. I suspect that CM 9380 and TMP 81.6.1 are females because the greater angle between the sacral vertebrae and distal end of the ischium would permit the passage of eggs (or live young) more readily than that of AMNH 5027 . The more divergent ischia are associated with robust skeletons, whereas the less divergent ischium is from a gracile skeleton.

Figures

Figure 10.1. A comparison of four Tyrannosaurus rex skulls. A, CM 9380 (formerly AMNH 973); B, AMNH 5027; C, LACM 23844; D, TMP 81.6.1; E, SDSM 12047. The reconstruction of CM 9380 differs from that given by Osborn (1906) in the orbital and postorbital regions. Osborn had erroneously used the right ectoptygoid as a right postorbital in the reconstructed skull at the Carnegie Museum. As may be seen in the figure, the squamosal no longer extends into the orbit. TMP 81.6.1 is more complete than shown, but preparation was not complete at the time this figure was made. SDSM 12047 is crushed dorsolaterally, which has greatly affected the snout, no attempt has been made to compensate for this. (Scale bar = lOcm.)

Figure 10.2. A, an overlay of maxillae: CM 9380, AMNH 5027, LACM 23844, TMP 81.6.1, and TMM 41436-1. Teeth not shown. B, right (reversed) and left maxillae of LACM 23844 and C, AMNH 5027.

Figure 10.3. An overlay of the dentaries of CM 9380, AMNH 5027, BM(NH) R7994 (formerly AMNH 5866), LACM 23844, SDSM 12047, and TMP 81.6.1. Teeth not shown.

Figure 10.4. A comparison of the cervicals of A, AMNH 5027 and B, BM(NH) R7994. Scale = 10 cm.

Figure 10.5. An overlay of the ischia of A, CM 9380; B, TMP 81.61; and C, AMNH 5027.