Albertosaurus sarcophagus, Osborn, 1905

Currie, Philip J., 1998, Possible evidence of gregarious behavior in Tyrannosaurids, Gaia 15, pp. 271-277 : 271-276

On August 10, 1910, Barnum Brown and Peter Kaisen discovered what they at first determined to be an articulated skeleton of the tyrannosaurid AIbertosaurus sarcophagus . They had been floating by boat down the Red Deer River of the province of Alberta in Canada, and had been prospecting in the badlands near what is now Dry Island Buffalo Jump Provincial Park, across the river from the mouth of Big Valley Creek. Both men were seasoned collectors from the American Museum of Natural History (New York), and were aware of the Albertosaurus fossils that had been collected in the same region more than twenty years earlier ( OSBORN, 1905). The bones were found in a very hard sandstone about twelve meters above river level. The hard rock separated cleanly from the bones, which are well preserved. After working on the site for less than a day, they realized that the articulated limb bones they were collecting represented several individuals. The three square meter excavation continued until September 1, during which time they excavated some skull bones, two dentaries with teeth, vertebrae, and many front and hind limb bones (unpublished information in Barnum Brown's "Annual Report for the Year 1910: Expedition to the Laramie Cretaceous of Montana and Alberta", which is on file in the Department of Vertebrate Paleontolog y, American Museum of Natural History). The excavation took two weeks to complete (a maximum of 29 man-days were invested by Brown, Kaisen and Davenport), and eight boxes of fossils were shipped back to New York.

Brown recognized that they had collected five partial or complete legs of Albertosaurus , plus limb bones that he identified as ornithomimid, and two hadrosaur phalanges. Preparation of the bones initially classified as ornithomimids revealed that they were actually from juvenile Albertosaurus . Each of the articulated specimens was given its own catalogue number ( TABLE I). Most of the unassociated tyrannosaur bones were catalogued as AMNH 5218, and this collection consisted of two nearly complete dentaries, 14 vertebrae, two chevrons, one scapula, one coracoid, two humeri, a pair of pubes, two femora, three tibiae, half a fibula, two astragali, one calcaneum, a pair of associated metatarsals (II-III), six isolated metatarsals, 42 phalanges, and seven unguals. The two hadrosaur phalanges were also given this number. Although no quarry map was made, no detailed notes were written, and only one photograph of the quarry was taken from a distance, it is clear that both articulated and disarticulated specimens were being recovered. Articulated metatarsals with associated limb bones were the most common elements excavated. It is not known whether Brown and Kaisen were collecting everything they found, orwere collecting only those elements that could be used to determine how many individuals were represented in the quarry. If the former, then there must have been some kind of taphonomic filter that favoured the preservation of hind limb bones over all else. It is impossible to know what such a filter may have been. Alternatively, it is possible that the party was collecting only the most useful bones. Peter Kaisen estimated in his field notes of August 13, 1910, two weeks before they finished the excavation, that it would take three weeks to finish the work. This suggests all of the bones in the quarry were not collected.

Brown occasionally referred to this bonebed ( BROWN, 1914), but he never published anything on it. RUSSELL & CHAMNEY (1967) commented on the association, but were unable to relocate the quarry (Russell, pers. comm., 1997). Because of the height of the bonebed above the river, they knew itwas from Member B of what was then known as the Edmonton Formation. This formation has now been elevated to group status, and the relevent rocks are from the upper part of the Horseshoe Canyon Formation (GIB­ SON, 1977). The site was also briefly mentioned by FARLOW (1976), along with other evidence for packing behavior in theropods.

In 1996, I decided to investigate further by examining specimens, field notes, field photographs, and letters that Brown wrote while hewas in the field. The site was relocated in 1997 by ajoint expedition of the Royal Tyrrell Museum of Palaeontology and the Dinamation International Society. Examination of the spoil piles around the original quarry confirmed that Brown was being selective in his choice of specimens collected. The bonebed, which extends deep into the hill, is now being quarried by staff of the Tyrrell Museum.

RESU LTS

The specimens ( Fig. 1 ) collected from the Albertosaurus quarry include at least seven articulated sets of metatarsals, several of which are associated with phalanges and other limb bones (TABLE I). Additional metatarsals catalogued as AMNH 5218 and AMNH 5230 had not been located by the time this paper was written, and could therefore not be considered in this census. These may represent the same individuals, or might have come from one or two additional animals. Of the seven groups of articulated metatarsals, three represent animals of about the same size. Because each of these three includes a right metatarsal IV, they clearly came from three different animals. The remaining sets of metatarsals are different enough in size and propor- tions to demonstrate the presence of four more individuals.

All other elements collected from the quarry were sorted by size and were assigned whenever possible to'the seven individuals (TABLE I) using data from articulated tyrannosaur skeletons. One pedal phalanx (111-3) was found to indicate the presence of an Albertosaurus smaller than the first seven, whereas another pedal 111-3 was clearly from a larger, more massive individual. Therefore, the quarry has yielded the remains of at least nine tyrannosaur specimens. Comparison of the bones with articulated skeletons of Albertosaurus sarcophagus (ROM 807, TMP 81.10.1, TMP 85.64.1, TMP 85.98.1 and TMP 86.205.1) suggests that all individuals, with one possible exception, should be assigned to this species. The exception is the large pedal phalanx 111- 3, which is massive enough to possibly be Oaspletosaurus RUSSELL, 1970, the only other large tyrannosaurid presently recognised in the Horseshoe Canyon Formation.

The smaller sets of metatarsals (individuals 1 and 2) are slender and more elongate than the larger ones ( Fig. 1 ). Because oftheir size and proportions, Brown originally believed that they were from ornithomimids. However, anatomical details led him to re-identify them as tyrannosaur once they were prepared. Regression analysis of hind limb elements clearly supports this result. In tyrannosaurids, the tibia shows negative allometry (compared with the femur) during growth ( RUSSELL, 1970; GATES Y, 1991; HOLTZ, 1994). but not to the same degree as the metatarsus ( Fig. 2 ).

DISCUSSION AND CONCLUSIONS

The discovery of nine or more tyrannosaurids, at least eight of which are the same species, in a single. quarry can be interpreted in many ways. Given the fact that twenty species ofdinosaurs are known from the Horseshoe Canyon Formation (APPENDIX I), and that the number of individuals of Albertosaurus sarcophagus made up less than ten percent of the dinosaur fauna, it is highlY,unlikely that the nine individuals ended up in the same place as a result of chance. The bonebed was apparently not very thick, and the similarity in degree of disarticulation of most of the skeletons indicates that all of the animals died around the same time and shared the same taphonomic history of burial. The almost complete lack of bones of herbivorous dinosaurs, and the absence of tooth marks or other evidence of predation suggests that the site was probably not a predator trap. The most parsimonious interpretation is that the tyrannosaurids were part of a group that died together. Because it is difficult to imagine why they would have collected into a group immediately before death, they were probably living together for a period 6f time before they died. There is not enough information available about the excavation to speculate what caused their death, but disease, drought, drowning and suffocation are some of the ca uses proposed for similar catastrophic death assemblages of herbivorous dinosaurs (CURRIE & DODSON, 1984; VARRICCHIO & HORNER, 1992; CORIA, 1994; SAMPSON, 1995).

Brown's Albertosaurus bonebed is not the only report of theropod bonebed concentrations. During Late Triassic and Early Jurassic times, Coelophysis COPE. 1889 ( COLBERT, 1989) and Syntarsus RAATH, 1969 (RAATH. 1990) are two examples of theropods that died en masse, whereas trackway sites (OS­ TROM, 1972) suggest that these an imals may have actually been moving in packs. The Coelophysis bone bed has most recently been interpreted as species-selectivity in drought conditions (SCHWARTZ & GILLETTE, 1994). The Upper Jurassic Cleveland-Lloyd quarry, which has produced evidence of more than seventy specimens of Allosaurus fragilis, has been interpreted as a predator trap ( MADSEN, 1976; MILLER, HORROCKS & MADSEN. 1996) for both individuals and packs of this spe cies ( RICHMOND & MORRIS, 1996). Skeletal evidence points towards pack hunting behavior in the small dromaeosaurid theropod Deinonychus OSTROM, 1969 ( OSTROM, 1969; MAXWE LL & OSTROM, 1995). There are many examples of Jurassic and Cretaceous trackway sites that show theropods moving in groups ( LOCKLEY, 1991; MOSSMAN & SARJEANT, 1983). The Late Cretaceous theropod Troodon LEIDY, 1856 may have been a pack animal (VARRIC­ CHIO & CURRIE, 1991). For tyrannosaurids, however. only one record of multiple individuals has been reported to date. When the large Tyrannosaurus rex OSBORN, 1905 known as "Sue" was excavated, parts of three other tyrannosaur skeletons (an adult, a juvenile, and a "baby") were collected from the same quarry (LARSON. 1995).

Another fact revealed by the metatarsals from Brown's Albertosaurus quarry ( Fig. 1 ), and by comparative analysis of tyrannosaur measurements, is that juvenile tyrannosaurids have hind limb proportions ( HOLTZ, 1994) similar to those of ornithomimids (TABLE II). For individuals with a femur length between 500 and 700 mm, the average ratio of tibia tofemuris 1.05 in tyrannosaurids, and 1.06 in ornithomimids. The same comparison for metatarsal III is 0.70 in tyrannosaurids and 0.77 in ornithomimids. The tibia/femur and metatarsus/femur ratios for carnosaurs and ceratosaurs in this size range are 0.95 and 0.50 respectively, and these were clearly much slower animals. It is widely accepted that limb proportions indicate ornithomimids were capable of running rapidly ( RUSSELL, 1972). In fact, they are often compared with ostriches, which can attain speeds of up to 70 km.h·'. Although proportions are similar in the hind legs of an ornithomimid and an ostrich, the femur is held in a more horizontal position in the extant animal, and it lacks the long tail. The biomechanical effects of these differences on speed have not been analyzed. Nevertheless, the relative proportions of limb elements and their slender proportions suggest thatjuvenile tyrannosaurids were almost as fast as equivalent sized ornithomimids, which in turn were probably faster than any other known dinosaurs.

.

The implications of this are extremely interesting. First, it is apparent that juvenile tyrannosaurids were probably faster runners than both their own parents and most other theropods of the same general size. In part this probably increased their chances of survival. Furthermore, juvenile tyrannosaurids, like ornithomimids and troodontids, were unquestionably faster than any of the herbivores that lived during Late Cretaceous times, regardless of what their absolute speeds were. Limb proportions also suggest that even the largest tyrannosaurid was capable of moving faster than any equivalent-sized, contemporary herbivore. If tyrannosaurids were packing animals, it is also possible that they practiced a division of labor, just as lions, wolves, or many other modern carnivores do. The faster, more agile juveniles may have been responsible for driving potential prey towards the larger, more powerful adult tyrannosaurids.

The evidence for packs oftyrannosaurids argues against these animals being obligatory scavengers ( HORNER & LESSEM, 1993). They were clearly opportunists that took advantage of eating the carcasses of other animals when they were available ( CURRIE & DODSON, 1984). However, it is doubtful that a single tyrannosaur could have found enough meatto scavenge to keep it alive. For a pack oftyrannosaurids to find enough dead animals is even less likely. Any mature tyrannosaur would have been large enough, fast enough, and strong enough to bring down any contemporary herbivore. Packing behavior would have improved this advantage, re- ducing the chance of injury to any individual tyrannosaur.

Figures

Fig. 1 - Albertosaurus sarcophagus OSBORN 1905 metatarsals (right) collected by Barnum Brown in the Horseshoe Canyon Formation ofAlberta, Canada. From left to right, AMNH 5229 (individual 2 oITABLE I), AMNH 5233 (individual 3), AMNH 5234 (individual 4), and AMNH 5232 (individual 8). See TABLE I for measurements.

Fig. 2 - Bivariate comparisons of tibial versus femoral lengths (upper graph), and metatarsal III versus femoral lengths (lower chart) in tyrannosaurids. A; Alberlosaurus, D; Daspletosaurus, G; Gorgosaurus, t; Tarbosaurus, T = Tyrannosaurus.