Over the last two decades there has been a revolution in the study of dinosaurs after it was discovered that some of these extinct animals were feathered.
Exactly how many species had feathers has been contentious, but a new study has shown that feathers were likely restricted to just a small proportion of the non-bird dinosaurs.
When the first perfectly preserved specimens of feathered dinosaurs were found in China in the 1990s, it was proved beyond doubt that these ancient animals were the ancestors of modern-day birds.
Since then, more and more species of dinosaur have been revealed to have been covered in feather-like structures. But how common these structures were, and how many different groups were feathered, is still being debated today.
Prof Paul Barrett, a dinosaur researcher at the Museum, has conducted an analysis of all the known specimens of dinosaur skin and mapped them onto evolutionary trees to see how they relate. The study was carried out with Nicolás Campione and David Evans and published as a contribution to a new book The Evolution of Feathers.
‘To date, most examples of dinosaur feathers have been found in the meat-eating dinosaurs, known as theropods, which is the group that also includes birds,’ explains Paul. ‘So that is not too much of a surprise.
‘But there’s been speculation as to how far back feathers appear in meat-eating dinosaur evolution, and whether feathers might also have been seen in all other dinosaurs.’
This is because there are a couple of examples of other dinosaurs from completely unrelated groups with feather-like coverings, most notably the herbivorous dinosaurs Kulindadromeus, Psittacosaurus and Tianyulong. In addition, it is also thought that some pterosaurs, which are the next closest relatives to dinosaurs, may also have been covered in feather-like structures.
This has led to speculation that feathers were not just concentrated in the meat-eaters, but that many other groups, like the horned ceratopsians such as Triceratops, may also have had a smattering of feathers.
But the analysis by Paul and his colleagues shows that this was unlikely, and it supports the idea that true feathers were concentrated only in the group closest to living birds. The few other specimens with feather-like features may instead be examples of convergent evolution.
What is a feather?
One of the biggest challenges when it comes to determining whether or not a dinosaur had feathers is the definition of a feather itself.
Skin can do lots of strange things. Crocodiles have ossified parts of their skin into armour plates, mammals developed fur from theirs, while tortoises evolved beak sheaths. Feathers are just another example of what animals have done with the structure of skin.
To be a true feather, there are characteristics that need to be fulfilled. The structures must be made from a protein called beta-keratin, they must be branched, and finally they must originate from a follicle.
While these are obviously easy characteristics to ascertain in living species, when it comes to the fossil record things start to get a little trickier. Aspects like their physical structure might be discernible from well preserved specimens but figuring out what protein they were made from and whether or not they originated from a follicle is far less straightforward.
‘Most of the occurrences of feathers that we know about in the fossil record are all very heavily concentrated in the meat-eating dinosaurs that are closely related to birds,’ explains Paul. ‘The dinosaurs that are furthest away from birds that all scientists agree had feathers are actually animals like tyrannosaurs and comsognathids, which although they look very different from birds are not actually that distantly related.’
We can’t be certain that the feather-like structures seen on theropods like the tyrannosaurs originated from a follicle in the skin because these microscopic structures are not preserved. But the fact that these animals are so closely related to birds, on the basis of numerous features seen throughout their skeletons, means these structures are considered to be true feathers.
In fact, most dinosaurs with strong evidence of feathers come from within a very select group of theropods known as the Coelurosauria. This includes not only tyrannosaurs and birds, but also the ornithomimosaurs, therizinosaurs and compsognathids.
Going further back in time, things get rapidly murkier. ‘We have very little evidence of feathers in earlier meat-eating dinosaurs,’ says Paul. ‘The further down the theropod dinosaur family tree we go, the evidence for feathers gets thinner and thinner.’
This could be for one of two reasons: either the animals simply did not have feathers, or these earlier dinosaurs have been fossilised in rocks that are not conducive for the preservation of soft tissues.
For those 77 dinosaur species where skin has been preserved, Paul and his colleagues were able to map them onto evolutionary trees to see how feathers were distributed across dinosaurs and their close relatives.
‘We have really strong evidence that animals like the duck-billed dinosaurs, horned dinosaurs and armoured dinosaurs did not have feathers because we have lots of skin impressions of these animals that clearly show they had scaly coverings,’ says Paul. ‘We also have zero evidence of any feather like structures in the long-necked dinosaurs, the sauropodomorphs.
‘If we look at the evidence that we do have, and we combine that with evolutionary trees, what we find is that there is no evidence for the first dinosaurs being feathered.’
Instead, it seems to indicate that feathers were an important part of the theropod story but not necessarily so for dinosaurs as a whole. Similarly, it might suggest that some feather-like structures could have appeared in some other dinosaur groups once or twice independently.
The major implication of this research is that the earliest dinosaurs were probably primitively scaly like other reptiles. Even when feathered pterosaurs are added to the evolutionary trees, it doesn’t alter these conclusions.
This could have an impact on why we think feathers evolved, as this means that although they were were almost certainly giving some benefits to meat-eating dinosaurs, maybe in helping them to stay warm or in display, these were seemingly less important in the other dinosaur groups.
It also suggests that perhaps dinosaurs had the underlying genetic ability to produce feathers from the start, but that they didn’t always develop them, or maybe some groups even lost them.
‘Some people might be disappointed that we don’t think we should be putting feathers on dinosaurs other than the meat-eaters,’ says Paul. ‘But this work raises a lot of other interesting questions. If all dinosaurs did possess the possibility of making feathers, then why didn’t they?
‘That’s a whole other set of questions.’