Allosaurus was a large bipedal predator with a large skull, equipped with dozens of large, sharp teeth. It averaged 8.5 meters (30 ft) in length, though fragmentary remains suggest it could have reached over 12 meters (39 ft). Relative to its large and powerful hind legs its three-fingered forearms were small but longer than that of the Tyrannosaurus Rex. They were also large enough to hold its prey while clamping down on the body with its massive jaws and pushing away with one foot in order to tear away chunks of flesh. The body was balanced by a long, heavy tail. It is classified as an allosaurid, a type of carnosaurian theropod dinosaur. The genus has a complicated taxonomy, and includes an uncertain number of valid species, the best known of which is A. fragilis. The bulk of Allosaurus remains have come from North America's Morrison Formation, with material also from Portugal and possibly Tanzania. It was known for over half of the 20th century as Antrodemus, but study of the copious remains from the Cleveland Lloyd Dinosaur Quarry brought the name Allosaurus back to prominence, and established it as one of the best-known dinosaurs.
As the prominent large predator in the Morrison Formation, Allosaurus was at the top of the food chain, probably preying on contemporaneous large herbivorous dinosaurs. Potential prey included ornithopods, stegosaurids, and sauropods. While it is often thought of as preying on sauropod dinosaurs in groups, there is little evidence for cooperative social behavior in this genus, and individuals may have been aggressive toward each other instead. It may have attacked large prey by ambush, using its upper jaws like a hatchet.
Allosaurus was a typical large theropod, having a massive skull on a short neck, a long tail and reduced forelimbs. Allosaurus fragilis, the best-known species, had an average length of 8.5 meters (28 ft), with the largest definitive Allosaurus specimen (AMNH 680) estimated at 9.7 meters long (32 ft), and an estimated weight of 2.3 metric tons (2.5 short tons). In his 1976 monograph on Allosaurus, James Madsen mentioned a range of bone sizes which he interpreted to show a maximum length of 12 to 13 meters (40 to 43 ft). As with dinosaurs in general, weight estimates are debatable, and since 1980 have ranged between 1500 kilograms (3300 lb), 1000 to 4000 kilograms (2200 to 8800 lb), and 1010 kilograms (2230 lb) for modal adult weight (not maximum). John Foster, a specialist on the Morrison Formation, suggests that 1000 kg (2200 lb) is reasonable for large adults of A. fragilis, but that 700 kg (1500 lb) is a closer estimate for individuals represented by the average-sized thigh bones he has measured.
Several gigantic specimens have been attributed to Allosaurus, but may in fact belong to other genera. The closely related genus Saurophaganax (OMNH 1708) reached perhaps 10.9 meters (36 ft) in length, and its single species has sometimes been included in the genus Allosaurus as Allosaurus maximus, though recent studies support it as a separate genus. Another potential specimen of Allosaurus, once assigned to the genus Epanterias (AMNH 5767), may have measured 12.1 meters in length (40 ft). A more recent discovery is a partial skeleton from the Peterson Quarry in Morrison rocks of New Mexico; this large allosaurid may be another individual of Saurophaganx.
The skull had a pair of horns above and in front of the eyes. These horns were composed of extensions of the lacrimal bones, and varied in shape and size. There were also lower paired ridges running along the top edges of the nasal bones that led into the horns. The horns were probably covered in a keratin sheath and may have had a variety of functions, including acting as sunshades for the eye, being used for display, and being used in combat against other members of the same species (although they were fragile). There was a ridge along the back of the skull roof for muscle attachment, as is also seen in tyrannosaurids.
Inside the lacrimal bones were depressions that may have held glands, such as salt glands. Within the maxillae were sinuses that were better developed than those of more basal theropods such as Ceratosaurus and Marshosaurus; they may have been related to the sense of smell, perhaps holding something like Jacobson's organ. The roof of the braincase was thin, perhaps to improve thermoregulation for the brain. The skull and lower jaws had joints that permitted motion within these units. In the lower jaws, the bones of the front and back halves loosely articulated, permitting the jaws to bow outward and increasing the animal's gape. The braincase and frontals may also have had a joint.
Allosaurus had nine vertebrae in the neck, fourteen in the back, and five in the sacrum supporting the hips. The number of tail vertebrae is unknown and varied with individual size; James Madsen estimated about 50, while Gregory S. Paul considered that to be too many and suggested 45 or less. There were hollow spaces in the neck and anterior back vertebrae. Such spaces, which are also found in modern theropods (that is, the birds), are interpreted as having held air sacs used in respiration. The rib cage was broad, giving it a barrel chest, especially in comparison to less derived theropods like Ceratosaurus. Allosaurus had gastralia (belly ribs), but these are not common findings, and they may have ossified poorly. In one published case, the gastralia show evidence of injury during life. A furcula (wishbone) was also present, but has only been recognized since 1996; in some cases furculae were confused with gastralia. The ilium, the main hip bone, was massive, and the pubic bone had a prominent foot that may have been used for both muscle attachment and as a prop for resting the body on the ground. Madsen noted that in about half of the individuals from the Cleveland Lloyd Dinosaur Quarry, independent of size, the pubes had not fused to each other at their foot ends. He suggested that this was a sexual characteristic, with females lacking fused bones to make egg-laying easier. This proposal has not attracted further attention, however.
The forelimbs of Allosaurus were short in comparison to the hindlimbs (only about 35% the length of the hindlimbs in adults) and had three fingers per hand, tipped with large, strongly curved and pointed claws. The arms were powerful, and the forearm was somewhat shorter than the upper arm (1:1.2 humerus/ulna ratio). The wrist had a version of the semilunate carpal also found in more derived theropods like maniraptorans. Of the three fingers, the innermost (or thumb) was the largest, and diverged from the others. The legs were not as long or suited for speed as those of tyrannosaurids, and the claws of the toes were less developed and more hoof-like than those of earlier theropods. Each foot had three weight-bearing toes and an inner dewclaw, which Madsen suggested could have been used for grasping in juveniles. There was also what is interpreted as the splint-like remnant of a fifth (outermost) metatarsal, perhaps used as a lever between the Achilles tendon and foot.
Allosaurus was an allosaurid, a member of a family of large theropods within the larger group Carnosauria. The family name Allosauridae was created for this genus in 1878 by Othniel Charles Marsh, but the term was largely unused until the 1970s in favor of Megalosauridae, another family of large theropods that eventually became a wastebasket taxon. This, along with the use of Antrodemus for Allosaurus during the same period, is a point that needs to be remembered when searching for information on Allosaurus in publications that predate James Madsen's 1976 monograph. Major publications using the name Megalosauridae instead of Allosauridae include Gilmore, 1920, von Huene, 1926, Romer, 1956 and 1966, Steel, 1970, and Walker, 1964.
Following the publication of Madsen's influential monograph, Allosauridae became the preferred family assignment, but it too was not strongly defined. Semi-technical works used Allosauridae for a variety of large theropods, usually those that were larger and better-known than megalosaurids. Typical theropods that were thought to be related to Allosaurus included Indosaurus, Piatnitzkysaurus, Piveteausaurus, Yangchuanosaurus, Acrocanthosaurus, Chilantaisaurus, Compsosuchus, Stokesosaurus, and Szechuanosaurus. Given modern knowledge of theropod diversity and the advent of cladistic study of evolutionary relationships, none of these theropods is now recognized as an allosaurid, although several, like Acrocanthosaurus and Yangchuanosaurus, are members of related families.
Allosauridae is one of three families in Carnosauria; the other two are Carcharodontosauridae and Sinraptoridae. Allosauridae has at times been proposed as ancestral to the Tyrannosauridae (which would make it paraphyletic), one recent example in Gregory S. Paul's Predatory Dinosaurs of the World, but this has been rejected, with tyrannosaurids identified as members of a separate branch of theropods, the Coelurosauria. Allosauridae is the smallest of the carnosaur families, with only Saurophaganax and a currently unnamed French allosauroid accepted as possible valid genera besides Allosaurus in the most recent review. Another genus, Epanterias, is a potential valid member, but it and Saurophaganax may turn out to be large examples of Allosaurus. Recent reviews have kept the genus Saurophaganax and included Epanterias with Allosaurus.
Allosaurus itself is based on YPM 1930, a small collection of fragmentary bones including parts of three vertebrae, a rib fragment, a tooth, a toe bone, and, most useful for later discussions, the shaft of the right humerus (upper arm). Othniel Charles Marsh gave these remains the formal name Allosaurus fragilis in 1877. Allosaurus comes from the Greek allos/αλλος, meaning "strange" or "different" and saurus/σαυρος, meaning "lizard" or "reptile". It was named 'different lizard' because its vertebrae were different from those of other dinosaurs known at the time of its discovery. The species epithet fragilis is Latin for "fragile", referring to lightening features in the vertebrae. The bones were collected from the Morrison Formation of Garden Park, north of Cañon City. Marsh and Edward Drinker Cope, who were in scientific competition, went on to coin several other genera based on similarly sparse material that would later figure in the taxonomy of Allosaurus. These include Marsh's Creosaurus and Labrosaurus, and Cope's Epanterias.
In their haste, Cope and Marsh did not always follow up on their discoveries (or, more commonly, those made by their subordinates). For example, after the discovery by Benjamin Mudge of the type specimen of Allosaurus in Colorado, Marsh elected to concentrate work in Wyoming; when work resumed at Garden Park in 1883, M. P. Felch found an almost complete Allosaurus and several partial skeletons. In addition, one of Cope's collectors, H. F. Hubbell, found a specimen in the Como Bluff area of Wyoming in 1879, but apparently did not mention its completeness, and Cope never unpacked it. Upon unpacking in 1903 (several years after Cope had died), it was found to be one of the most complete theropod specimens then known, and in 1908 the skeleton, now cataloged as AMNH 5753, was put on public view. This is the well-known mount poised over a partial Apatosaurus skeleton as if scavenging it, illustrated as such by Charles R. Knight. Although notable as the first free-standing mount of a theropod dinosaur, and often illustrated and photographed, it has never been scientifically described. The multiplicity of early names complicated later research, with the situation compounded by the terse descriptions provided by Marsh and Cope. Even at the time, authors such as Samuel Wendell Williston suggested that too many names had been coined. For example, Williston pointed out in 1901 that Marsh had never been able to adequately distinguish Allosaurus from Creosaurus. The most influential early attempt to sort out the convoluted situation was produced by Charles W. Gilmore in 1920. He came to the conclusion that the tail vertebra dubbed Antrodemus by Leidy was indistinguishable from those of Allosaurus, and Antrodemus thus should be the preferred name because as the older name it had priority. Antrodemus became the accepted name for this familiar genus for over fifty years, until James Madsen published on the Cleveland-Lloyd specimens and concluded that Allosaurus should be used because Antrodemus was based on material with poor, if any, diagnostic features and locality information (for example, the geological formation that the single bone of Antrodemus came from is unknown). "Antrodemus" has been used informally for convenience when distinguishing between the skull Gilmore restored and the composite skull restored by Madsen.
The period since Madsen's monograph has been marked by a great expansion in studies dealing with topics concerning Allosaurus in life (paleobiological and paleoecological topics). Such studies have covered topics including skeletal variation, growth, skull construction, hunting methods, the brain, and the possibility of gregarious living and parental care. Reanalysis of old material (particularly of large 'allosaur' specimens), new discoveries in Portugal, and several very complete new specimens have also contributed to the growing knowledge base.
The completeness, preservation, and scientific importance of this skeleton gave "Big Al" its name; the individual itself was below the average size for Allosaurus fragilis, and was a subadult estimated at only 87% grown. The specimen was described by Breithaupt in 1996. Nineteen of its bones were broken or showed signs of infection, which may have contributed to "Big Al"'s death. Pathologic bones included five ribs, five vertebrae, and four bones of the feet; several damaged bones showed osteomyelitis, a bone infection. A particular problem for the living animal was infection and trauma to the right foot that probably affected movement and may have also predisposed the other foot to injury because of a change in gait.
It is unclear how many species of Allosaurus there were. Seven species have been considered potentially valid since 1988 (A. amplexus, A. atrox, A. europaeus, the type species A. fragilis, the as-yet not formally described "A. jimmadseni", A. maximus, and A. tendagurensis), although only a fraction are usually considered valid at any given time. Additionally, there are at least ten dubious or undescribed species that have been assigned to Allosaurus over the years, along with the species belonging to genera now sunk into Allosaurus. In the most recent review of basal tetanuran theropods, only A. fragilis (including A. amplexus and A. atrox as synonyms), "A. jimmadseni" (as an unnamed species), and A. tendagurensis were accepted as potentially valid species, with A. europaeus not yet proposed and A. maximus assigned to Saurophaganax.
A. amplexus, A. atrox, A. fragilis, "A. jimmadseni", and A. maximus are all known from remains discovered in the Kimmeridgian–Tithonian Upper Jurassic-age Morrison Formation of the United States, spread across the states of Colorado, Montana, New Mexico, Oklahoma, South Dakota, Utah, and Wyoming. A. fragilis is regarded as the most common, known from the remains of at least sixty individuals. Debate has gone on since the 1980s regarding the possibility that there are two common Morrison Formation species of Allosaurus, with the second known as A. atrox; recent work has followed a "one species" interpretation, with the differences seen in the Morrison Formation material attributed to individual variation. A. europaeus was found in the Kimmeridgian-age Porto Novo Member of the Lourinhã Formation, but may be the same as A. fragilis. A. tendagurensis was found in Kimmeridgian-age rocks of Tendaguru, in Mtwara, Tanzania. Although the most recent review tentatively accepted it as a valid species of Allosaurus, it may be a more basal tetanuran, or simply a dubious theropod. Although obscure, it was a large theropod, possibly around 10 meters long (33 ft) and 2.5 metric tons (2.8 short tons) in weight.
Allosaurus is regarded as a probable synonym of the genera Antrodemus, Creosaurus, Epanterias, and Labrosaurus. Most of the species that are regarded as synonyms of A. fragilis, or that were misassigned to the genus, are obscure and were based on scrappy remains. One exception is Labrosaurus ferox, named in 1884 by Marsh for an oddly formed partial lower jaw, with a prominent gap in the tooth row at the tip of the jaw, and a rear section greatly expanded and turned down. Later researchers suggested that the bone was pathologic, showing an injury to the living animal, and that part of the unusual form of the rear of the bone was due to plaster reconstruction. It is now regarded as an example of A. fragilis. Other remains thought to pertain to Allosaurus have come from across the world, including Australia, Siberia, and Switzerland, but these fossils have been reassessed as belonging to other dinosaurs.
Allosaurus was the most common large theropod in the vast tract of Western American fossil-bearing rock known as the Morrison Formation, accounting for 70 to 75% of theropod specimens, and as such was at the top trophic level of the Morrison food web. The Morrison Formation is interpreted as a semiarid environment with distinct wet and dry seasons, and flat floodplains. Vegetation varied from river-lining forests of conifers, tree ferns, and ferns, to fern savannas with rare trees.
The Morrison Formation has been a rich fossil hunting ground, holding fossils of green alage, fungi, mosses, horsetails, ferns, cycads, ginkgoes, and several families of conifers. Other fossils discovered include bivalves, snails, ray-finned fishes, frogs, salamanders, turtles, sphenodonts, lizards, terrestrial and aquatic crocodylomorphans, several species of pterosaur, numerous dinosaur species, and early mammals such as docodonts, multituberculates, symmetrodonts, and triconodonts. Such dinosaurs as the theropods Ceratosaurus, Ornitholestes, and Torvosaurus, the sauropods Apatosaurus, Brachiosaurus, Camarasaurus, and Diplodocus, and the ornithischians Camptosaurus, Dryosaurus, and Stegosaurus are known from the Morrison. The Late Jurassic formations of Portugal where Allosaurus is present are interpreted as having been similar to the Morrison but with a stronger marine influence. Many of the dinosaurs of the Morrison Formation are the same genera as those seen in Portuguese rocks (mainly Allosaurus, Ceratosaurus, Torvosaurus, and Apatosaurus), or have a close counterpart (Brachiosaurus and Lusotitan, Camptosaurus and Draconyx).
Allosaurus coexisted with fellow large theropods Ceratosaurus and Torvosaurus in both the United States and Portugal, The three appear to have had different ecological niches, based on anatomy and the location of fossils. Ceratosaurs and torvosaurs may have preferred to be active around waterways, and had lower, thinner bodies that would have given them an advantage in forest and underbrush terrains, whereas allosaurs were more compact, with longer legs, faster but less maneuverable, and seem to have preferred dry floodplains. Ceratosaurus, better known than Torvosaurus, differed noticeably from Allosaurus in functional anatomy by having a taller, narrower skull with large, broad teeth. Allosaurus was itself a potential food item to other carnivores, as illustrated by an Allosaurus pubic foot marked by the teeth of another theropod, probably Ceratosaurus or Torvosaurus. The location of the bone in the body (along the bottom margin of the torso and partially shielded by the legs), and the fact that it was among the most massive in the skeleton, indicates that the Allosaurus was being scavenged.
Medullary bone tissue, also found in dinosaurs as diverse as Tyrannosaurus and Tenontosaurus, has been found in at least one Allosaurus specimen, a shin bone from the Cleveland-Lloyd Quarry. Today, this bone tissue is only formed in female birds that are laying eggs, as it is used to supply calcium to shells. Its presence in the Allosaurus individual establishes sex and shows she had reached reproductive age. By counting growth lines, it was shown that she was 10 years old at death, so sexual maturity in Allosaurus was attained well before maximum growth and size.
The discovery of a juvenile specimen with a nearly complete hindlimb shows that the legs were relatively longer in juveniles, and the lower segments of the leg (shin and foot) were relatively longer than the thigh. These differences suggest that younger Allosaurus were faster and had different hunting strategies than adults, perhaps chasing small prey as juveniles, then becoming ambush hunters of large prey upon adulthood. The thigh bone became thicker and wider during growth, and the cross-section less circular, as muscle attachments shifted, muscles became shorter, and the growth of the leg slowed. These changes imply that juvenile legs has less predictable stresses compared with adults, which would have moved with more regular forward progression.
Similar conclusions were drawn by another study using finite element analysis on an Allosaurus skull. According to their biomechanical analysis, the skull was very strong but had a relatively small bite force. By using jaw muscles only, it could produce a bite force of 805 to 2,148 N, less than the values for alligators (13,000 N), lions (4,167 N), and leopards (2,268 N), but the skull could withstand nearly 55,500 N of vertical force against the tooth row. The authors suggested that Allosaurus used its skull like a hatchet against prey, attacking open-mouthed, slashing flesh with its teeth, and tearing it away without splintering bones, unlike Tyrannosaurus, which is thought to have been capable of damaging bones. They also suggested that the architecture of the skull could have permitted the use of different strategies against different prey; the skull was light enough to allow attacks on smaller and more agile ornithopods, but strong enough for high-impact ambush attacks against larger prey like stegosaurids and sauropods. Their interpretations were challenged by other researchers, who found no modern analogues to a hatchet attack and considered it more likely that the skull was strong to compensate for its open construction when absorbing the stresses from struggling prey. The original authors noted that Allosaurus itself has no modern equivalent, that the tooth row is well-suited to such an attack, and that articulations in the skull cited by their detractors as problematic actually helped protect the palate and lessen stress. Another possibility for handling large prey is that theropods like Allosaurus were "flesh grazers" which could take bites of flesh out of living sauropods that were sufficient to sustain the predator so it would not have needed to expend the effort to kill the prey outright. This strategy would also potentially have allowed the prey to recover and be fed upon in a similar way later. An additional suggestion notes that ornithopods were the most common available dinosaurian prey, and that allosaurs may have subdued them by using an attack similar to that of modern big cats: grasping the prey with their forelimbs, and then making multiple bites on the throat to crush the trachea. This is compatible with other evidence that the forelimbs were strong and capable of restraining prey.
Other aspects of feeding include the eyes, arms, and legs. The shape of the skull of Allosaurus limited potential binocular vision to 20° of width, slightly less than that of modern crocodilians. As with crocodilians, this may have been enough to judge prey distance and time attacks. The arms, compared with those of other theropods, were suited for both grasping prey at a distance or clutching it close, and the articulation of the claws suggests that they could have been used to hook things. Finally, the top speed of Allosaurus has been estimated at 30 to 55 kilometers per hour (19 to 34 miles per hour).
Although Allosaurus may have hunted in packs, recent research suggests that Allosaurus and other theropods were like other diapsids and had largely aggressive instead of cooperative interactions with other members of their own species. One study noted cooperative hunting of prey much larger than an individual predator, as is commonly inferred for theropod dinosaurs, is rare among vertebrates in general, and modern diapsids (including lizards, crocodiles, and birds) very rarely cooperate to hunt in such a way. Many modern diapsid predators are territorial and will kill and cannibalize intruders of the same species, and will also do the same to smaller individuals that attempt to eat before them when aggregated at feeding sites. This suggests that, for example, the Cleveland-Lloyd quarry shows Allosaurus individuals were drawn together to feed on other disabled or dead allosaurs, and sometimes were killed in the process, thus accumulating. This could explain the high proportion of juvenile and subadult allosaurs present, as juveniles and subadults are disproportionally killed at modern group feeding sites of animals like crocodiles and komodo dragons. The same interpretation applies to Bakker's lair sites. There is some evidence for cannibalism in Allosaurus, including Allosaurus shed teeth found among rib fragments, possible tooth marks on a shoulder blade, and cannibalized allosaur skeletons among the bones at Bakker's lair sites.
Along with Tyrannosaurus, Allosaurus has come to represent the quintessential large, carnivorous dinosaur in popular culture. It is a common dinosaur in museums, due in particular to the excavations at the Cleveland Lloyd Dinosaur Quarry; by 1976, as a result of cooperative operations, 38 museums in eight countries on three continents had Cleveland-Lloyd allosaur material or casts. Allosaurus is the official state fossil of Utah.
Allosaurus has been depicted in popular culture since the early years of the 20th century. It is top predator in both Arthur Conan Doyle's 1912 novel, The Lost World, and the 1925 film adaptation, the first full-length motion picture to feature dinosaurs (it is not to be confused with Tyrannosaurus, which also appears in the film). It later became the starring dinosaur of the 1956 film The Beast of Hollow Mountain, and the 1969 film The Valley of Gwangi, two genre combinations of living dinosaurs with Westerns. In The Valley of Gwangi, Gwangi is billed as an Allosaurus, although Ray Harryhausen based his model for the creature on Charles R. Knight's depiction of a Tyrannosaurus. Harryhausen sometimes confuses the two, stating in a DVD interview "They're both meat eaters, they're both tyrants... one was just a bit larger than the other. In nonfictional presentations, Allosaurus appears in the second and fifth episodes of the BBC television series Walking with Dinosaurs, and the Walking with Dinosaurs special The Ballad of Big Al chronicles the life of the Allosaurus specimen nicknamed "Big Al".