Artis's impression of Limusaurus inextricabilis. Credit: Portia Sloan

Birds are generally considered to be the living descendants of dinosaurs, yet differences between bird wings and dinosaur hands have long left palaeontologists struggling to explain how birds would have evolved from their dinosaur ancestors.

Birds' wings are thought to form from the fusion of the second, third and fourth digits on their hands as the embryo develops. Theropods, the predominantly carnivorous dinosaurs that included tyrannosaurids such as Tyrannosaurus rex and dromaeosaurids such as Velociraptor mongoliensis, also only had three long fingers.

But palaeontologists had thought that these were the first, second and third digits because in early theropod fossils, such as that of Dilophosaurus, these three fingers were elongated, with a seemingly semi-vestigial fourth digit and a nearly absent fifth.

Based on this species, it looked very much like theropod dinosaurs lost their fifth digit early on then, around the time of Dilophosaurus, started to lose their fourth digit too.

But this evolutionary explanation left researchers wondering how birds' wings could have developed from digits 2, 3 and 4. The prevailing explanation was that theropods lost their fourth and fifth digits, then birds lost their first digits and regrew their fourth digits.

Mistaken identity

Now, a team led by Xing Xu from the Institute of Vertebrate Paleontology and Paleoanthropology in Beijing and James Clark from The George Washington University in Washington DC is proposing a simpler answer based on a new dinosaur species found in Jurassic rocks formed 156 million to 161 million years ago in the Junggar Basin in western China.

Limusaurus has a vestigial first finger. Credit: James Clark

The species, a ceratosaur named Limusaurus inextricabilis, is a beaked and herbivorous early theropod with two remarkable characteristics: a reduced first digit and a metacarpal (lower finger bone) at the base of the second digit that matches those found at the base of the first digits in tyrannosaurids and dromaeosaurids1. These two features suggest that the first digits in late theropods are in fact the second digits, the researchers report today in Nature. We believe that late theropod dinosaurs "had digits 2, 3 and 4, but that these have long been misidentified as digits 1, 2 and 3", says Xu.

On the basis of this evidence, the team argue that early theropods lost their first and fifth digits once and that these digits remained lost in late theropods, an explanation that vastly simplifies the current convoluted evolutionary story. "When birds are considered to [have digits] 2, 3 and 4 it is far simpler for most theropods to [have] 2, 3 and 4 as well," says Clark.

This fossil "can be viewed as a kind of 'missing link' that shows an intermediate digital morphology between living birds and more primitive theropods", says Paul Barrett, a palaeontologist at the Natural History Museum in London.

Evolutionary oddity

However, some researchers are concerned about how the new fossil is being interpreted. "A lot of the arguments depend on bird wings being formed from the second, third and fourth digits and it is possible that they are not," explains evolutionary geneticist Günter Wagner from Yale University.

During development, digits are identified by what embryonic tissue forms them, where they grow and what genes shape them as they grow.

If the tissue that forms the second digit is bombarded by genes telling it to form in the shape of a first digit, it will appear to be a first digit - but growing in the location where a second digit normally forms. And experimental evidence suggests this is happening inside modern bird wings, Wagner says.

"The ceratosaur fossil may be showing us a species in the midst of a digit identity shift, but whether the digits that we see in later theropods are the actual second, third and fourth digits or the first, second and third digits in the second, third and fourth positions, altered by gene bombardment to look like the second, third and fourth digits, is difficult to determine," he says.

It is also possible that the ceratosaur does not play a part in the larger evolutionary story and evolved its unusual hands in response to a lifestyle that, because it was beaked and herbivorous rather than toothed and carnivorous, was quite different from that of other theropods. "I think it far more likely that this new animal just has an oddly reduced hand," says Kevin Padian, an palaeontologist at the University of California, Berkeley. "It is equally reasonable that we are just dealing with another odd possibility of evolution," he says.