Amphicoelias: a rare sauropod and the story of a lost giant

Amphicoelias altus was a long-necked plant-eating sauropod from the Upper Jurassic Morrison Formation of Colorado. The genus is known from one incomplete skeleton: two dorsal vertebrae, part of a pubis and a right femur. A 2021 redescription supported its distinction from other sauropods, although its exact position within Diplodocoidea remains unstable. It appears in the dinosaur catalogue as a valid but sparsely known genus.

For decades the name was also associated with a supposed giant dinosaur, Amphicoelias fragillimus. That was a different specimen, known from one enormous vertebra that was later lost. In 2018 it was placed in the separate genus Maraapunisaurus. Claims that the real Amphicoelias altus was 50–60 metres long therefore confuse two animals.

Quick facts

Discovery at Garden Park

Aaron Ripley discovered the bones in 1877 at Cope Quarry XII in Garden Park, roughly eight miles north or northeast of Cañon City. Oramel Lucas sent the material to palaeontologist Edward Drinker Cope in Philadelphia. The find came during the early phase of the famous rivalry between Cope and Othniel Marsh, but its locality is documented and its principal bones survive in a museum collection.

Cope named Amphicoelias altus late in 1877. The genus name combines Greek roots meaning “on both sides” and “hollow”, referring to the form of the vertebral centra; altus means “high”. After Cope’s death, the American Museum of Natural History acquired his collection. The holotype was catalogued as AMNH FARB 5764 and remains in New York. It must not be confused with the lost AMNH FARB 5777, the specimen behind the giant fragillimus story.

What belongs to the holotype?

A modern revision restricts the type material to two middle or posterior dorsal vertebrae, an incomplete right pubis and a right femur. They came from the same quarry, include no duplicate bones, and are consistent in size and preservation, so they are treated as one individual.

Other fossils were later referred to Amphicoelias, including a tooth, scapula, coracoid, ulna and additional elements. Some resemble Camarasaurus more closely, and their connection to the holotype is uncertain. The 2021 study excluded them from the secure material so that the diagnosis would not depend on a mixed assemblage.

This makes Amphicoelias one of the rarest sauropods in the Morrison Formation. In more than 140 years, no other skeleton has been confidently assigned to A. altus. Without new material, the skull, neck, tail and forelimbs remain unknown.

The vertebrae

The two vertebrae come from the middle or rear of the back. Their centra are opisthocoelous: the front joint surface is convex and the rear one concave. Internal air spaces and thin bony partitions helped lighten the large backbone.

The genus name is sometimes simplified as “double-hollow”, but that is not a precise description of these bones. Cope emphasised the shape of the cavities and the thin-walled construction; the joint ends were not symmetrically concave at both sides.

The redescription identified a diagnostic combination of laminae and fossae on the neural arches. Individual features occur in other diplodocoids, but their arrangement supports a distinct species. The vertebrae are damaged and partly restored with plaster, so not all older measurements are equally reliable.

Femur, pelvis and body size

The right femur survives in three joined pieces. It is about 1.7 metres long, although part of its lower end has been reconstructed and the shaft was compressed and slightly twisted after burial. It is large, but not unique among sauropods.

The bone is relatively slender and nearly round in cross-section. Its tubular shape was once treated as a defining difference from Diplodocus. Broader comparisons show similar proportions in some diplodocids, so slenderness alone cannot diagnose the genus. The incomplete pubis lacks the acetabular margin, obturator opening and joint with the ischium. It confirms a large sauropod but contributes fewer diagnostic traits than the vertebrae.

There is no complete spine from which to measure total length. Estimates are scaled from related sauropods. A femur around 1.7 metres long indicates a large diplodocoid, not a demonstrated 50-metre record-holder. Length estimates of roughly 18–25 metres depend on which relative and body proportions are chosen. Mass is even less certain because the width of the torso and the full set of limbs are unknown.

Even the femur’s length needs qualification: one end is partly restored and the bone is deformed. A single exact mass figure without a three-dimensional model and error range would imply more precision than the fossil allows.

Classification and the other species

Cope first compared Amphicoelias with Camarasaurus. In 1921 Henry Fairfield Osborn and Charles Mook argued that it was closer to Diplodocus and Barosaurus. Later phylogenetic analyses generally place it within Diplodocoidea, but disagree about its position inside the group. Some recover it near the base of Diplodocidae or among apatosaurines; others place it outside Diplodocidae and Dicraeosauridae. The 2021 revision found several possible placements and concluded that A. altus is a valid diplodocoid whose exact branch is unresolved.

The uncertainty reflects the small number of bones. Skull, neck and tail vertebrae contain many of the features normally used to distinguish diplodocoid branches, and none is known for this genus.

Cope named another species, Amphicoelias latus, from tail vertebrae and a femur. Those fossils are now assigned to Camarasaurus supremus, making the name a junior synonym. In 1878 Cope described A. fragillimus from part of the neural arch of an immense dorsal vertebra. AMNH FARB 5777 later disappeared, probably during collection moves. Only Cope’s illustration and measurements remain, so the specimen cannot be checked with modern imaging or direct comparison.

Kenneth Carpenter reinterpreted the drawing in 2018 as a vertebra of an early rebbachisaurid and proposed the new genus Maraapunisaurus. Under that reconstruction, the complete vertebra would have been about 2.4 metres high and the animal perhaps 30–32 metres long. This remains highly uncertain because it depends on a lost, incomplete bone. The accepted type species of Amphicoelias remains A. altus, represented by the surviving AMNH FARB 5764.

Age, habitat and limits of reconstruction

Cope Quarry XII lies in the upper Morrison Formation, in an interval dated to about 150.44–149.21 million years ago, early Tithonian. The sediments formed in river and floodplain systems with marked seasonality. Nearby quarries preserve Camarasaurus, other sauropods and a diverse land-vertebrate fauna. Their proximity helped early collectors build rich museum collections, but it also encouraged later confusion when bones were combined without precise field numbers.

No skull or teeth are securely associated with the holotype. A herbivorous diet follows from its identification as a sauropod; this skeleton alone cannot reveal specific plants or browsing height. Direct evidence comprises two dorsal vertebrae, part of the pelvis and a femur. A long neck, small head and tail are reconstructed from other diplodocoids, but their exact proportions in Amphicoelias are unknown.

No skin impressions or soft tissues are known. Colour, lips, spines and any special ornament in artwork are artistic choices. The animal should not be illustrated as the old 60-metre A. fragillimus silhouette, because that name refers to a different and now separately classified fossil.