RALEIGH, N.C.—How did the ancestors of birds evolve the ability to fly? That birds are descended from small, meat-eating dinosaurs is established. Exactly how the creatures conquered the air remains a mystery, however. Now the authors of a new study of a controversial feathered dinosaur say they have resolved a key aspect of the problem—namely, how the animals controlled their flight once they became airborne.

Two theories have dominated the long-running debate over how bird flight evolved. In the so-called cursorial scenario, the ability to fly emerged in terrestrial dinosaurs that raced across the ground with their arms outstretched and leaped into the air after prey or out of harm’s way, their wing feathers providing lift. The arboreal scenario, in contrast, supposes that flight arose in tree-dwelling dinosaurs that were built for gliding and started flapping their arms in order to stay aloft longer.

In 2003 a feathered dinosaur fossil came to light that was purported to elucidate the question of how flight evolved. The roughly 125-million-year-old specimen exhibited evidence of feathers on its hind limbs in addition to its forelimbs, prompting researchers to describe the crow-size animal, Microraptor gui, as a four-winged dinosaur. A startling artist’s reconstruction accompanied the description of the fossil remains, showing the bird flying with its hindlimbs spread out to the side, as if doing a split. The authors argued that the feathered hindlimbs, together with the forelimb wings, acted as an airfoil to help the animal glide. Critics begged to differ.

The new work paints a different picture of how Microraptor’s enigmatic hindlimbs functioned. In two presentations given on October 20 at the annual meeting of the Society of Vertebrate Paleontology (SVP) in Raleigh, N.C., Michael Habib and Justin Hall of the University of Southern California argued that the hindlimbs would have been generally held under the body during steady flight and then deployed to produce rotation movement (roll) or left-right movement (yaw) during unsteady maneuvers such as turning. The team reported that its mathematical modeling indicates that Microraptor’s hindwings would have enabled it to turn twice as fast as a two-winged animal—handy for dodging trees in its cluttered environment. Complimenting the hindlimb’s role in turning and braking, the tail of Microraptor controlled up-down movement (pitch), the researchers say. “A combination of pitch control by the tail, roll generation by the ‘hindwings’ and multi-purpose control by the main wings would have made Microraptor a highly maneuverable animal,” Habib noted.

“This study provides a plausible mechanism by which dinosaurs that otherwise have strongly Velociraptor-like bodies could take to the air and control themselves while in flight,” Hall remarked in a statement to the press. “Obviously crashing is bad for the long-term health of the animal, but until now we had little idea how the earliest flying dinosaurs avoided such catastrophes given their relatively simple wing structure.” Habib added that this so-called distributed control system may have been an independent experiment in flight that had no bearing on the evolution of bird flight, or that it could represent an intermediate phase in the evolution of bird flight, after which most control function shifted to the forelimbs. The presentations were co-authored by David Hone of the Queen Mary, University of London, and Luis Chiappe of the Natural History Museum of Los Angeles County.

Not everyone is convinced by the team’s arguments. Kevin Padian of the University of California at Berkeley, an expert on bird evolution, observed that the presentations focused on the effect of the hindlimb on a gliding animal instead of one that flapped its wings. Last year at the SVP meeting he presented evidence that gliders and flyers are completely unrelated to each other. He says that “there is not a shred of evidence that says gliding is involved in the evolution of flapping flight.” He questioned why the team's model would focus on gliding parameters when the forelimb shape was consistent with flapping, not gliding, and the hindlimb would have generated so much drag.