Concavenator

Fernando Escaso, Francisco Ortega, and José Luis Sanz working on Concavenator

Concavenator had structures resembling quill knobs on its ulna, a feature known only in birds and other feathered theropods, such as Velociraptor. Quill knobs are created by ligaments which attach to the feather follicle and, since scales do not form from follicles, the authors ruled out the possibility that they could indicate the presence of long display scales on the arm. Instead, the knobs have been thought to probably anchor simple, hollow, quill-like structures. Such structures are known both in coelurosaurs, such as Dilong, and in some ornithischians, like Tianyulong and Psittacosaurus. If the ornithischian quills are homologous with bird feathers, their presence in Concavenator and other allosauroids would be expected.[7] However, if ornithischian quills are not related to feathers, the presence of these structures in Concavenator would show that feathers had begun to appear in earlier, more primitive forms than coelurosaurs.

Conventional restoration of Concavenator with scales, a sail, and a small amount of quills

Feathers or related structures would then likely be present in the first members of the clade Neotetanurae, which lived in the Middle Jurassic. No impressions of any kind of integument were found near the arm, although extensive scale impressions were preserved on other portions of the body, including broad, rectangular scales on the underside of the tail, bird-like scutes on the feet, and plantar pads on the undersides of the feet.[7]

Speculative restoration of Concavenator with a hump as a hypothetical thermoregulatory device

Restoration of the crushed skull

However, the significance of the 'quill knobs' remains controversial, as some amount of skepticism has been raised among experts on the validity of the interpretation that the ulnar bumps represent quill knobs.[8] Christian Foth and colleagues noted that the quill knobs of Concavenator were on the anterolateral side of the ulna. They suggest they were intermuscular lines that acted as tendon attachments.[9] The hypothesis that the bumps along the ulna represented muscular insertion points or ridges was subsequently examined and the results were presented at the 2015 meeting of the Society of Vertebrate Paleontology. Elena Cuesta Fidalgo, along with two of the researchers who initially described Concavenator (Ortega and Sanz), attempted to reconstruct its forearm musculature to determine if the ulnar bumps would be explained as an inter-muscular ridge. They identified the insertion point for the major arm muscles and determined that the row of bumps could not have been located between any of them. They found that the only possibility was that the bumps could be an attachment scar for the M. anconeus muscle. However, this is unlikely because this muscle normally attaches to a smooth surface without marks or bumps on the underlying bone. They argued that the most likely explanation for the bumps was their initial interpretation as feather quill knobs. The authors admitted that it was unusual for quill knobs to form along the posterolateral surface of the bone, but also noted that the same arrangement is found in some modern birds, like the Moorhen.[10]

The dig site in 2002 before excavation

In 2018, Cuesta Fidalgo published her doctorate thesis on the anatomy of Concavenator, which argued that the ulna was preserved in lateral view. This means that the ulnar bumps were positioned posterolaterally instead of anterolaterally as Cau and Mortimer claimed. Cuesta Fidalgo noted that the proximal part of the ulna is affected by fracturing and abrasion, with certain features that would have shifted when compared to their position in the bone while the animal was alive. For example, in the fossil, the lateral process of the ulna is positioned further posteriorly than the ulnar bumps. In Allosaurus and Acrocanthosaurus, the lateral process is on the lateral (rather than posterior) part of the bone, which would seem to support the ulnar bumps being anterolateral in position if the lateral process was truly preserved in lateral orientation in Concavenator. However, Cuesta Fidalgo described how the lateral process was distorted posteriorly when compared to the bumps and was not valid evidence for the claim that the ulna had shifted into anterior view. The ulna's distortion (as well as genus-specific proportions) means that precise comparisons to Allosaurus and Acrocanthosaurus would be misleading.[11] As Cuesta Fidalgo and her colleagues explained in 2015, the ulnar bumps could not be an intermuscular line if the bone is preserved in lateral view.[10] Cuesta Fidalgo and her colleagues pointed out that these bumps on the ulna are posterolateral, which is unlike that of interosseous ligaments.[12]