Jurassic abelisauroids in England? Why (and how) to deal with fragmentary specimens

Vertebrate paleontologists often have to deal with fragmentary remains; complete skeletons, such as that of Sciurumimus, are extremely rare. The vast majority of specimens in vertebrate palaeontology consist of isolated teeth or bones, or often even only parts of these. Thus, palaeontologists are often faced with the challenge of interpreting these fragments, identifying their identity and evaluating their evolutionary or biogeographic significance. Many of us have published on such remains and sometimes have drawn far-reaching conclusions on the basis of a few isolated bones or teeth. I myself am certainly guilty of this, having announced the first dromaesaurid (sickle-clawed dinosaur, a relative of the famous Velociraptor) from the southern hemisphere on the basis of a tooth and a few foot bones, or having described the first Jurassic tyrannosauroid from Europe on the basis of a single ilium. Of course, claims on the basis of such material are often contested. A good example for this are dinosaur remains from the Cretaceous of Australia, which almost all consist of isolated bones, and which have been interpreted very differently by different scientists.

Recently, my Argentinean colleagues Martín Ezcurra and Federico Agnolín re-interpreted a distal end of a tibia (the shin bone) of a small predatory dinosaur from the Middle Jurassic of England as one of the oldest representatives of the abelisauroids, a group of peculiar predatory dinosaurs which are mainly known from the southern hemisphere. This specimen would thus not only be one of the oldest representative of abelisauroids, but also the oldest record of this group from the northern hemisphere. This interpretation obviously had far-reaching implications for the evolution and biogeography of this dinosaur group, which were clearly lined out by Ezcurra & Agnolín.

Example of an abelisauroid: Carotaurus sastrei (by Nobu Tamura)

Since I am currently working on a Middle Jurassic abelisaurid from Argentina - which Diego Pol and me recently published as Eoabelisaurus mefi - I was, of course, very interested in this re-interpretation of the small theropod (the predatory dinosaurs) specimen from England. However, reading the article, I was unconvinced of the interpretation presented by Ezcurra and Agnolín. Thus, I got out my photographs of predatory dinosaurs that I took in many different collections over the years, asked a few colleagues for images of other specimens that I thought might be of interest (which many of them provided; special thanks here to Roger Benson, Lindsay Zanno, Jacques Gauthier and Brooks Britt), and then wrote a short article arguing against the interpretation presented by Ezcurra & Agnolín. In this article, I could show (I hope) that the characters used by Ezcurra & Agnolín to refer the specimen to abelisauroids have a wider distribution within theropods, and that the bone in question can thus only be identified as an indeterminate predatory dinosaur.

The bone of contention: distal end of a shin bone of a small predatory dinosaur.

So, having thus publicly disagreed with my colleague Martín Ezcurra, does that mean that I think that he is a poor anatomist? Certainly not; on the contrary, I value Martín's detailed anatomical observations very highly and generally think that he does excellent work! Indeed, a few months ago, Martín started working as a PhD student here at Ludwig Maximilian University in Munich with my friend and colleague Richard Butler, and in all things theropod, I often consult with him. The article by him and Federico Agnolín is also not an example of poor science. The two presented testable characters for the referral of the element in question to abelisauroids, which only made it possible for me to criticise their work. Thus, this is proper scientific procedure: You present your data and the hypothesis derived from it. This hypothesis is then out there to be tested in the light of new data or other observations, which is exactly what I did in this case: I presented observations that showed that their data is not sufficient to support their conclusions. Only by putting the data and your ideas on it out there can science progress; being wrong sometimes is not a failure, but a necessary prerequisite of scientific progress! (Which, however, does not mean that any bogus idea is beneficial to science. It should be backed by hard data, repeatable observations, and logical deduction. See Darren Naish's blog "tetrapod zoology" for an example of what is not very helpful for science.)

So does that mean that I think we should not deal with fragmentary material any more? Also not, since that would mean that we have to ignore some 80 % of the vertebrate fossil record, and would thus loose an enormous amount of potentially important information. In the interpretation of such materials, we have made enormous progress in the past twenty to thirty years: Previously, most referrals of such fragmentary specimens was based on rather vague notions of "overall similarity", but, with the development of large data sets of anatomical characters that are typical for certain groups (so-called apomorphic characters), we can be much more specific on what basis we refer fragmentary remains to a specific clade. Of course, with incomplete material, there is always the possibility that more complete remains might prove us wrong, or that new finds or observations on other materials might show that these characters are more widely distributed than previously thought, but with this data we can formulate specific hypotheses that are then testable by such new data. There is also the danger of overinterpreting minute differences that might not be of systematic significance (for example due to individual variation or just abnormalities), but who can say when we reach this level in animals that are extinct?

As scientists, we should always remain critical. It is often said that "extraordinary claims need extraordinarily good evidence", and we are wise in remembering this when reading about the implications of interpretations of fragmentary specimens. Thus, I am not convinced that there were tyrannosaurs in the Cretaceous of the southern hemisphere, as claimed by Roger Benson and colleagues based on a single pubis. However, I am also not entirely convinced that there were carcharodontosaurids in the Late Jurassic of Tanzania, a claim made by myself on the basis of some teeth and a few vertebrae. Nevertheless, at our current state of knowledge, these are the best hypotheses to explain the available data. New finds, or new observations might prove us wrong, but such is the nature of scientific investigation.