Pterosaur fossils launch over flight.

Long before birds made their debut, a group of winged reptiles called pterosaurs took to the skies as the first vertebrates endowed with the power of flight. Over the last two centuries, paleontologists have debated what kind of wings carried these "dragons of the air" and how they managed to take off.

A new analysis of pterosaur fossils from Kazakhstan in central Asia raises questions about the prevailing image depicting pterosaurs launching themselves by running quickly along the ground. Instead, the exquisitely preserved Asian specimens suggest that at least one species had wings that made walking an awkward affair, reports David M. Unwin and Natasha N. Bakhurina of the University of Bristol in England. They describe their research in the Sept. 1 NATURE.

The Asian species bears the name Sordes pilosus, or "hairy devil," because parts of its wings had small wavy fibers, which Soviet paleontologists interpreted as fur when they first described the animal in 1971. The pigeon-sized flier dates from the late Jurassic period, roughly 150 million years ago.

Judging from impressions in the rock next to the Sordes skeleton, Unwin and Bakhurina conclude that the animal's wings attached along the body from the front limbs to the hind feet, much like the wings of a bat. They also found evidence of a smaller wing connecting both legs.

When the animal was grounded, "the attachment of [the wing membrane] to the legs and feet must have severely impeded movement," the researchers say. They suggest that Sordes took to the air by climbing to some height and then launching itself. With its relatively large wings, Sordes would have flown slowly and maneuvered well.

This interpretation of wing design challenges the prevailing theory about pterosaurs, put forward over a decade ago by paleontologist Kevin Padian of the University of California, Berkeley. Padian argued that pterosaurs flew more like birds, with narrow wings attached to the trunk of the animal. With their legs unencumbered, pterosaurs could have run rather than crawled along the ground.

Padian remains convinced that pterosaur wings did not reach back to the hind limbs. He dismisses the conclusions of Unwin and Bakhurina because the specimens of Sordes they describe are flattened, making it difficult to distinguish impressions of the wings from squashed parts of the body.

"Unwin has not tested the hypothesis that a lot of what he's seeing is postmortem damage. It's very hard to tell what is actually wing," Padian says.

He argues further that the structure of the ankle, knee, and hip joints in pterosaurs precludes them from flying as bats do. They could not have brought their legs up in the same plane as their wings, Padian says.

Finding more specimens, preserved in different positions and various environments, would provide a means of testing whether the wings actually extended to the legs or not.

Unwin and Bakhurina report that seven specimens of Sordes exist, but Padian contends that other researchers have not had an opportunity to study the fossils. "Access to these specimens by other investigators has not been granted, so really what anyone says about them is up for grabs," says Padian.