Rewriting the dinosaur family tree

Model challenges 130-year-old theory.

Dinosaurs have traditionally been divided into two major groups, one with a ‘bird-hipped’ pelvis called the Ornithschia and the other with a reptile-like pelvis called the Saurischia. Carnivorous theropods like the Tyrannosaurus rex and giant sauropods such as Diplodocus fell into the Saurischia group. Ornithopods, like the Iguanodon, and armored dinosaurs such as Triceratops and Stegosaurus were catergorized as Ornithschia. However, a new Nature study proposes a radical shift in this grouping. Matthew Baron, from the University of Cambridge told us more.

RG: What motivated this study?

Matthew Baron: After a year of studying ornithischian dinosaurs for my PhD I had started to notice that a lot of ornithischian anatomy was understudied and underrepresented in the literature. There appeared to be a lot of strange similarities between ornithischians and theropods that had gone largely unnoticed or that had simply been considered by others to be coincidences. I also noticed that features in other early dinosaurs that I had worked with didn’t really make any sense if the old hypothesis about dinosaur relationships was accurate. It got me questioning how sure we could really be about the validity of a 130-year-old idea that was essentially based around a single observation on the hips of these animals. I took the idea to my supervisors, and they suggested we expand my PhD topic to be about all early dinosaurs, not just ornithischians, and we then decided that the best direction to go in would be to build a large and robust new dataset of all early dinosaurs and close dinosaur relatives. Our aim was to test objectively the key dinosaur relationships using a very large and very broad sample of dinosaur species. Remarkably, this had not been done before.

RG: Can you tell us what you found?

Baron: What we found was that the old scheme of classification, that all dinosaurs are either ornithischian or saurischian, is completely wrong. Our data suggests that some dinosaurs that were previously considered to be saurischians were actually more closely related to ornithischians. Our trees place the theropods (meat eaters and birds) closer to Ornithischia than either group is to the Sauropodomorpha (long-necked tree browsers like Diplodocus and Brontosaurus). This fundamentally differed from the old scheme. Removing the theropods from Saurischia and placing them next to Ornithischia in a new group that we called Ornithoscelida was something that has never been suggested before, but it is what our data very strongly supports, rather than the previously held ideas about where each group belongs. This result forced the sauropodomorphs out of the dinosaur group because the old definition of dinosaur was based around the common ancestor of Ornithischia and Theropoda and did not account for the possibility that the dinosaur family tree may be arranged in a different way to the way it is in the old system. So we are now offering a new, more inclusive and stable definition for Dinosauria (technical term for the dinosaur group), to make sure that, whether our hypothesis is accepted or not, the same animals, including all the long-necked tree browser types like Diplodocus, will remain in the one dinosaur group forever.

Trees simplified. Credit: Matthew Baron.

RG: How did you come to this conclusion?

Baron: We looked at a lot of early dinosaurs and compiled a large list of anatomical features that could potentially unite and divide them in to groups. We then explored various evolutionary scenarios using new computer technology. Using the extremely large dataset that we had built we were able to compare a large number of species (75) for a huge number of anatomical features (457) and, in minutes, produce a tree that best explained all of the observed anatomical features in the early dinosaurs that we had looked at. This tree was radically different to previous ideas, so we tested and re-tested our results to make sure they were robust. In the end, they all held up, so we concluded that we had found a real result that was worth publishing.

RG: What does this mean for our understanding of dinosaurs?

Baron: It means a lot of things. Firstly, that we need to rethink how they first diversified and evolved. Secondly, it forces us to re-evaluate when they began (we have a new estimate of 247 million years ago). Thirdly, it forces us to re-evaluate where they began (Northern Hemisphere species appear to be just as important as Southern Hemisphere species, if you look at our results). Fourthly, it provides good evidence that the very first dinosaur was an omnivore, not specialist in any feeding strategy, but rather eating anything it could get its hand on, plant or animal. Fifthly, it suggests that feathers may have been a feature of our new group only (only theropods and ornithischian have feathers, meaning it could be an ancestral condition of Ornithoscelida, not Dinosauria). Finally, it suggests that having large grasping hands and being able to run quickly on two legs probably served as the main advantages for the earliest dinosaurs during the very first stages of dinosaur evolution. Perhaps this is why they outcompeted all other reptiles of the time and came to dominate the planet in the way that they did.

RG: What has the scientific community’s reaction been like so far?

Baron: We have actually kept this a pretty closely guarded secret until this week. Only the referees for the paper knew what we had found until it was given to the press. We are hoping that it will start a debate in our field and cause us to all re-evaluate more of the ‘old truths’ in new ways. Obviously, we are going to upset a few people, but this is just science in action. Ideas get put forward and then we test the ideas and the evidence as best as we can and then we form new ideas.

RG: Will this help with future discoveries?

Baron: I hope so. First of all, it points us to the Northern Hemisphere, which is a new direction for field work. The Northern Hemisphere has been largely overlooked so far, at least in terms of early dinosaur evolution. Secondly, all newly discovered species can now be added to our dataset to investigate what they might be and where they might fit in to the dinosaur tree and the bigger picture. We are, after all, also presenting the world with a brand-new tool for assessing newly discovered or previously unidentifiable dinosaur specimens. The dataset also has many other potential uses and there are years worth of work that can be done with it. It’s a good thing too, as it took me over three years to put it all together.

RG: What should the public take from your study?

Baron: Never ever accept something as true just because it has been said for a long time! Always be open to the idea that hypotheses can be wrong and should always be tested and challenged by inquisitive minds. We should also take note of the fact that, despite this being a potentially revolutionary discovery, no new specimens had to be found for this new hypothesis to be made; collections of specimens were revisited and new views were made about old bones, highlighting the importance of historic museum collections and the need to look after what we have already unearthed. These things are treasures and can always yield new information if only someone is willing to give them the attention that they require.

Featured image courtesy of James St. John.