Tyrannosauridae, Osborn, 1906

Holtz, TR jr., 2004, Tyrannosauroidea, The Dinosauria, University of California Press, pp. 111-136 : 9-3

Tyrannosauridae ,

originally proposed by Osborn (1906), is defined as those theropods more closely related to Tyrannosaurus rex than to Eotyrannus lengi . Although the discovery of basal tyrannosauroids has reduced the morphological gap between primitive coelurosaurs and Tyrannosauridae , the latter clade remains characterized by numerous synapomorphies, including the following: the surface of the nasals contacting the maxillae facing ventrally; the supratemporal fossa occupying the caudolateral half of the frontal; the frontal edge notched in the region of the lacrimal suture; the supratemporal fenestrae confluent over the parietals, forming a sagittal parietal crest; a squamosal recess (also in ornithomimosaurs); the jugal pneumatized by a foramen in the caudal rim of the jugal antorbital fossa; the squamosal-quadratojugal flange constricting the infratemporal fenestra; prominent muscular fossae on the dorsal surface of the palatines; a palatine foramen on the dorsal surface of the palatine recess; a transverse nuchal crest at least twice as tall as the foramen magnum, rising well above the parietal crest and having a transverse width more than twice the height; a pair of tablike processes on the supraoccipital wedge; the caudal tympanic recess extending into the opisthotic caudodorsal to the fenestra ovalis, confluent with this fenestra; a deep basisphenoid sphenoidal sinus with large foramina; the basisphenoid having a well-developed subcondylar recess (also in ornithomimosaurs); the supraoccipital excluded from the foramen magnum; a tripartite supraoccipital sinus; the coronoid and supradentary fused (Hurum and Currie 2000; although this apparent synapomorphy may be due to the better preservation of tyrannosaurid skulls compared with those of most other toothed theropods, as this condition is also found in allosauroids and dromaeosaurids [Holtz et al., this vol.]); the retroarticular process further reduced until it is simply a broad, shallow, concave semicircular plate; the vertical ridge on the distal surface of the premaxillary teeth strongly developed; the maxillary and dentary incrassate (the cross section greater than 60% as wide labiolingually as long mesiodistally in adults); the carninae of the dentary teeth asymmetrical with the long axis of the dentary (also in allosaurids and Acrocanthosaurus ); small ventral projections on the ventral sides of the cranial cervical centra; the dorsal series subequal to the femur length; the midcaudal vertebrae cranial zygapophyses extending more than one-half but less than one centrum length (also in ornithomimosaurs and coelophysids); the distal caudal spinous processes axially elongate; the cranialmost gastralia fused into a platelike mass (although this may be more widespread among large-bodied tetanurans [see Holtz et al., this vol.]); the scapula contributing more than one-half to the glenoid; the coracoid dorsoventral length five times or more the coracoid glenoid diameter; the femur-humerus ratio greater than 2.6; the distal articular surface of the ulna expanded transversely; the distal end of the radius flattened craniocaudally; metacarpal III lacking phalanges; the ilium craniocaudal length subequal to or greater than the femur length; the pubic tubercle crestlike and pointing cranially; the ischial tuberosity expanded into a proximal dorsal ischial process; the most proximal point of the cranial trochanter extending above the proximal margin of the femoral head; a pronounced, rugose, cranially facing proximal tuberosity on metatarsal V; metatarsals II and IV contacting at midshaft on the plantar surface; an arctometatarsus; and a short contact between metatarsals II and IV on the dorsal surface of the proximal end of the metatarsals (also in Ornithomimidae ).

Most previous studies of the phylogenetic relationships of tyrannosaurid taxa (Matthew and Brown 1922; Russell 1970a; Bakker et al. 1988; Paul 1988a; Carpenter 1992b) have not employed explicit numerical cladistic analyses. Holtz (2001b) presented a preliminary cladistic study. Subsequent to that analysis, new data on some of these taxa strongly support them as juveniles of other tyrannosaurid taxa (Currie and Dong 2001b; Carr and Williamson, in press) and/or revealed additional data on those individuals (Carr and Williamson 2000). Carr (1999), Currie and Dong (2001b), Carr and Williamson (in press), and Currie (2003a) further observed some of the difficulties in distinguishing taxonomic from ontogenetic features in tyrannosaurids.Their work suggests that many of the differences in bone and landmarks shape, orientation, and proportions in juvenile and subadult Gorgosaurus libratus relative to adult forms are the same differences used by various previous authors to diagnose smaller individuals as different taxa ( Maleevosaurus , Shanshanosaurus , and Nanotyrannus ) from larger sympatric taxa ( Tarbosaurus , Tarbosaurus , and Tyrannosaurus , respectively). Among these changes are development of cranial ornamentation, dorsoventral deepening of the skull relative to its length, and a proportional increase in tooth size and labiolingual diameter. Carr (1999) and Brochu (2002) considered the last trait to result in an ontogenetic decrease in tooth number, as the distalmost tooth positions have been lost; Currie (2003a) disagrees. Holtz (1995b) and Currie (2000, 2003a) demonstrated that in tyrannosaurids (and other nonavian theropods) the proportions and gracility of the distal limb elements decrease as body size increases.Thus, larger tyrannosaurids have shorter and stouter tibiae and metatarsi than do smaller individuals, both juveniles of the same species and adults of smaller species. Hindlimb element proportions and gracility remain a problematic source of taxonomic information, although Currie (2003a) has established some general patterns. Complicating the question of gracility is the problem of sexual dimorphism.The presence of robust and gracile morphs within the species Tyrannosaurus rex has been described by Carpenter (1990a) and Larson (1994, 1997), with the robust morph tentatively identified as the female. Pending future morphometric studies, these suppositions should be treated cautiously; nevertheless, they do suggest a degree of variability in the limb and pelvic proportions within a single tyrannosaurid species. Furthermore, Carpenter (1990a) has demonstrated that the degree of cranial ornamentation within T. rex can vary, with, for example, some forms lacking highly rugose nasals and postorbital bosses, others having one ornamentation but not the other, and still others having both.

Interestingly, although the basal synapomorphies for Tyrannosauridae as a whole are divided between the skull and the postcranial skeleton, almost all of the potential synapomorphies within the clade are cranial. Similar situations exist for other dinosaur taxa ( Ceratopsidae , Hadrosauridae , etc.), where the skulls may be distinctive but the postcranial skeleton was constant within the clade. However, there has been little direct study of the variation in tyrannosaurid postcrania, and future analysis may indeed reveal more diagnostic characters outside of the skull.

Like Currie (2003a), this study supports a split in Tyrannosauridae between a gracile Albertosaurinae and the more robust Tyrannosaurinae . Paul (1988a) recommended a similar division but used the generic terms Albertosaurus and Tyrannosaurus for these clades. Two Asian tyrannosauroids have a problematic distribution: Alectrosaurus , which is not well known; and Alioramus , known from a partial skull and a fragmentary postcranium. In this analysis Alectrosaurus can equally parsimoniously assume nearly any phylogenetic position near the base of Tyrannosauridae—as the sister taxon to this clade, as a basal albertosaurine, or as a basal tyrannosaurine ( fig. 5.22); Alioramus falls either as the sister taxon to Tyrannosauridae or as a tyrannosaurine closer to Tarbosaurus and Tyrannosaurus than to Daspletosaurus .