Dinosaur Supertree

The press release (30th April, 2002)

Dinosaurrific! The most comprehensive picture yet of how dinosaurs evolved has been produced by a team at Bristol University.

More than 1,000 species of dinosaurs have been named since the first skeletons were dug up in the 19th century, and unravelling their patterns of evolution has been a major area of research.

Since 1980, over 150 evolutionary trees of dinosaurs have been published, most of them looking at small groups of species. The Bristol researchers have put all of these smaller trees together to produce a supertree of 277 species of dinosaur, to be reported in the Proceedings of the Royal Society, Series B next week on May 7.

The research team includes Davide Pisani, an expert on supertree methods and computing; Dr Adam Yates, an expert on plant-eating dinosaurs; Dr Max Langer, an expert on the earliest dinosaurs; and Mike Benton, Head of the Department of Earth Sciences, Professor of Vertebrate Palaeontology and an expert on the origin of dinosaurs and evolutionary tree construction.

Professor Benton said: "It's not complete, but it's the most detailed and comprehensive single evolutionary tree produced for dinosaurs, and indeed for almost any other group.

"We hope the supertree will represent a solid framework for future study of dinosaur evolution and will stimulate and direct further studies towards the less well understood areas of dinosaur classification that are highlighted as branches in the supertree."

The team

Davide Pisani: doctoral student, from Italy, completing his PhD in Bristol; expert on supertrees.

Adam Yates: postdoc., currently working on early plant-eating dinosaurs, as part of the Bristol Dinosaur Project

Max Langer: former doctoral student in the Department, now a junior lecturer at the University of São Paolo, Brazil

Mike Benton: Head of Department of Earth Sciences; team leader


What is the importance of this discovery?

Two things. We have made one of the biggest supertrees yet attempted - with 277 genera. The more species you include, the more possible trees might exist - for 277, the number is huge, more than the number of atoms in the Universe. A year or two ago, this wouldn't have been possible, but now, with ever-faster and larger computers, problems like this can be tackled. Second, we have produced the most comprehensive evolutionary tree of dinosaurs yet available.

What is a supertree?

A supertree is a tree made up from many 'source' trees, the original published trees. In this case, we tracked down 126 published trees, some of which covered only a family, others a larger group. These are then combined to produce the best fitting single solution. Of course, the source trees, published by dozens of palaeontologists around the world, include all sorts of conflicts and disagreements. So the supertree methods have to find ways to find the 'majority' or strongest verdict.

Does this use monster computers?

Yes - the work should be done on a supercomputer ideally. We used the biggest desktop machine we had - a Macintosh G4 - and it ground away for 2 weeks, day and night, trying out all sorts of combinations. We stopped it then, since it was threatening to explode, and it had come up with 34,900 possible tree solutions. Our result then is a consensus of all those. Out of the 34,900 versions, most branches in the tree were agreed, and the variations occurred in two main areas, the Jurassic sauropods and Jurassic theropods.

Is there any point in making big trees?

Supertrees are a key new way to summarise huge amounts of information and to give us a picture of current opinion - a kind of Domesday book. The methods are not just used to sort out dinosaurs, but any other groups of plants, animals or microbes. It's all part of the 'Tree of Life' programme, a large international effort to document all species alive and dead, and to find the one great evolutionary tree that links them all together.

Does it show us anything new about dinosaurs?

No. Supertree methods are unlikely to come up with anything new since they are meant to summarise current knowledge. The supertree does resolve problems, and it will pick out the best-supported solution out of a number that may be available. So it can help settle arguments.

What kind of arguments?

Well, for example, there is a running debate about the place of birds in the scheme of things. Our supertree shows once and for all (as if there was still any doubt) that birds are dinosaurus, very close to the dromaeosaurids like Deinonychus and Velociraptor, and to the troodontids like Saurornithoides. It would take a huge amount of counter-evidence to break up this part of the tree, so the anti-dinosaur-bird people have clearly lost the debate.

What does the tree look like?

Here it is:

Email for more media information: public-relations@bristol.ac.uk

Full reference: Pisani, D., Yates, A. M., Langer, M., and Benton, M. J. 2002. A genus-level supertree of the Dinosauria. Proceedings of the Royal Society, Series B, 269, 915-921.

Download a pdf pre-print of the paper.

Download a pdf of the supertree to see the full detail.

Back to Palaeontology home page