Bactrosaurus: an early relative of the duck-billed dinosaurs

Bactrosaurus johnsoni was a plant-eating ornithopod from Late Cretaceous Inner Mongolia. Its fossils come chiefly from the Iren Dabasu Formation near Erenhot, but the age assigned to those beds has shifted between studies. A broad Late Cretaceous age is secure; a single precise date is not.

The genus is useful for studying the evolutionary grade close to the ancestry of hadrosaurids, the familiar duck-billed dinosaurs. Its jaws already carried a sophisticated tooth-replacement system, while its skeleton retained a combination of earlier and more derived features. Bones from individuals of different sizes also reveal growth. They do not, however, add up to one complete articulated animal.

Quick facts

Name, discovery and the type specimen

Charles Whitney Gilmore named Bactrosaurus in 1933. The name combines Greek elements usually rendered as “club” or “staff” and “lizard”. Gilmore referred to the widened, club-like tips on some vertebral spines; he did not mean that the animal had a clubbed tail or a weapon. The species name johnsoni honours Albert Johnson, a member of the American Museum of Natural History's Central Asiatic expeditions.

The Johnson Quarry lies about thirteen kilometres east of the former Iren Dabasu telegraph station. Roy Chapman Andrews led the 1922 expedition, with field palaeontologist Walter Granger among those working at the site. The quarry yielded many separated ornithopod bones, including remains from animals at different growth stages. The name-bearing specimen is AMNH 6553. Its early registration history is complicated: a single catalogue number may include bones from more than one individual.

Similar hadrosauroids lived in the same deposits. In particular, Gilmoreosaurus mongoliensis was recognised from material that had once been mixed with or compared to Bactrosaurus. Later collecting and study changed the picture. A Sino-Belgian expedition found a bone concentration in 1995; its 1998 description provided a more informative skeletal reconstruction. A 2011 study compared juvenile and adult material and documented changes through growth.

Classification and changing interpretations

Bactrosaurus is an ornithischian dinosaur, an ornithopod and a derived iguanodontian. It is usually placed among Hadrosauroidea or the nearby Hadrosauromorpha, outside Hadrosauridae. The exact name and boundary used for the broader branch differ between phylogenetic analyses, but the useful distinction is stable: Bactrosaurus is close to duck-billed hadrosaurids without being a typical member of that family.

Its position among early hadrosauroids depends on which taxa and anatomical characters an analysis includes. Some results place it on a relatively early branch; others recover it among a set of close forms that can include Gilmoreosaurus and Gobihadros. That makes it part of a complex radiation, not a demonstrated direct ancestor of later hadrosaurids.

Older classifications placed Bactrosaurus among lambeosaurines, the hadrosaurids known for hollow cranial crests. The incomplete early skull made it possible to suppose that a crest had simply gone missing. Better cranial bones did not support that explanation. Its apparently “primitive crestless lambeosaurine” status was a taxonomic error, not an anatomical discovery.

Species and names once assigned to it

Bactrosaurus johnsoni is the only species generally accepted as secure. The name B. prynadai was applied to jaw fragments of young Central Asian hadrosauroids, but those remains lack a distinctive combination of characters. It is usually treated as a doubtful name rather than confidently included in this genus.

Bactrosaurus kysylkumensis arose when fragmentary Uzbek fossils originally called Cionodon kysylkumensis were moved into Bactrosaurus. Isolated vertebrae and other non-diagnostic bones cannot establish that relationship securely. Those proposed species should not be counted as equivalent to the well-supported B. johnsoni.

What the fossils preserve

The known sample includes premaxillae, maxillae, lower-jaw bones, teeth and pieces from the back of the skull. The postcranial record includes many vertebrae and ribs, shoulder and pelvic elements, upper and lower limb bones, and parts of the hands and feet. Together the material represents a substantial share of the skeleton, but the parts come from different individuals and are not joined in a single complete skeleton.

Some elements preserve growth stages. A young animal's jaw depth, spine proportions and bone robustness could differ from adult proportions. Overlapping specimens allow researchers to separate some growth changes from differences between species, though an isolated bone can still be difficult to assign. The skull was never found complete. Several specimens together support a long, low head without the hollow crest once added to older restorations.

Mounted skeletons therefore need to be read as composites. Real bones from more than one animal may be arranged together, while gaps are filled by comparison with related ornithopods. The result can communicate a plausible overall body plan, but it is not a single skeleton excavated intact. The same care applies to the boundary between Bactrosaurus and Gilmoreosaurus: the two occur in the same formation and some bones have changed assignments over time.

Size and feeding anatomy

The largest studied skull has been estimated at roughly 50–55 centimetres long. Scaling it against better-known relatives suggests an animal about 6–6.5 metres long; the largest femur is around 81 centimetres. These are estimates assembled from separate bones, not measurements of one complete individual. There is no reliable whole-skeleton mass estimate.

The front of the mouth ended in a toothless beak. Behind it, upper and lower dental batteries carried successive replacement teeth. Fewer teeth contributed to the active grinding surface at one time than in the later hadrosaurid Edmontosaurus, but replacement still maintained a working edge as crowns wore down. Deep jaws and a strong lower-jaw coronoid process provided attachment for substantial muscles.

Jaw shape and tooth wear support repeated processing of plant matter rather than simply snipping leaves. The detailed path of jaw movement in hadrosauroids is still reconstructed from bone geometry and mechanical models. Describing it as identical to mammalian chewing would go beyond the evidence. The feeding apparatus was effective, but it does not identify one favourite plant.

Vertebrae, limbs and movement

Several neural spines widened near their tips. That feature inspired the genus name and likely provided attachment area for ligaments and muscles supporting the back or tail. It does not by itself demonstrate a tall hump, fat store or external sail. Those soft-tissue structures leave no direct evidence in the vertebral tips.

The hind limbs were long and powerful. The forelimbs were shorter but robust enough to bear weight. Comparisons with related ornithopods make both quadrupedal walking and bipedal movement plausible: four limbs could support slow travel or feeding, while the hind limbs likely supplied most propulsion during faster movement. No trackway can be assigned confidently to Bactrosaurus, so its exact gait and speed are unknown.

Floodplain habitat and behaviour

The Iren Dabasu deposits record river channels, a broad floodplain, smaller streams, temporary water bodies and soils that formed between floods. Plant remains and sediments fit a vegetated landscape capable of supporting large herbivores better than a bare, waterless desert. Climate reconstructions vary, while the river-and-floodplain setting is comparatively well supported.

A beak could crop leaves and shoots, and the dental batteries could grind them. A low head and the ability to use four limbs fit feeding at low or moderate height. No gut contents, securely linked coprolites or distinctive plant bite marks identify the diet more precisely. The animal should not be assigned a single preferred plant on the basis of habitat alone.

Several individuals in a deposit have sometimes been taken to suggest herding. Flowing water can also transport and mix remains of animals that died at different times. The assemblage does not establish a family group, a migration, parental care or even that all the individuals were together in life. These behaviours remain possible hypotheses, not observations from the bones.

Old reconstructions and limits

The best-known outdated image gave Bactrosaurus a hollow lambeosaurine crest. New skull material supports a lower, flatter profile and phylogenetic studies place the genus outside Hadrosauridae. A tall hollow crest copied from Corythosaurus is therefore not justified.

The “club lizard” translation can also mislead: only the tips of some vertebral spines were club-shaped. Neither a tail club nor a high dorsal hump follows from that name. Similarly, a single precise age is inappropriate. Vertebrate fossils, microfossils, regional correlations and geochronology have produced different age estimates for Iren Dabasu, ranging across parts of the Late Cretaceous.

Some abnormal vertebrae of hadrosauroids, including material associated with Bactrosaurus, have been interpreted as possible tumours. Radiological diagnosis of fossil bone is valuable, but it does not establish a disease outbreak, unusual susceptibility in the genus or a particular environmental cause. The evidence supports a plant-eating hadrosauroid with a substantial, mixed fossil sample; it leaves its exact age, some specimen assignments and much of its soft-tissue appearance uncertain. Compare it with ornithopods and hadrosaurids in the dinosaur catalogue.