Andrew Biewener's group has long been fascinated by what keeps birds aloft. How birds transfer the power generated by their flight muscles to air and propel themselves along, without any buoyant support, is an intriguing question. Bret Tobalske and Tyson Hedrick wanted to know how birds that have similar builds but differ in weight change their wing beats as they accelerate. Heavier birds have to work harder to keep themselves off the ground, so Biewener's team expected that heavy birds would need to fly using a gait that generated more power until they reached a higher speed than a light bird. But to the Harvard group's surprise, when they compared flying birds with others that weighed half as much, both species switched gait at the same speed (p. 1389)!

As birds accelerate, they change gait, in much the same way as terrestrial animals. At low speeds, the birds fly with a gait that generates a lot of power, but once they've reached a high enough speed they can shift to a less powerful gait. Big birds need more power to stay aloft, so Biewener's team assumed that the heavier turtle dove would keep flying with the most powerful gait for longer than the lighter cockatiel.

But unlike animals, which visibly change the way they move to switch gait,bird's change gait by modifying the vortex patterns they trail in their wake. Biewener's team knew that they would need to analyse the birds' vortex trails if they were going to understand how birds power flight at different speeds.

But Biewener's team faced a major technical challenge. A bird's flight speed can only be controlled if it's flying in a wind tunnel, but visualising air currents in a wind tunnel is tricky. The team decided to estimate the wake patterns that the birds generated. Hedrick and Tobalske filmed cockatiels and ringed turtle-doves with four cameras as the birds flew in a specially designed wind tunnel. Using the four different views of the birds' flight,Hedrick reconstructed each wing beat in three-dimensions by painstakingly digitising almost one million points on thousands of images. Only then could he reconstruct the instantaneous forces that the wing exerted on the air to generate the vortices that would tell him which gaits the doves and cockatiels used as they flew faster.

Hedrick's patience was eventually rewarded. The vortex patterns showed that both birds set out at low speeds creating closed vortex rings in the bird's wake. At higher speeds, the birds switched to the continuous vortex gait. When he looked for the transition speed for both birds, he was surprised that they both switched up to the continuous vortex gait at 7 m s-1, even thought he'd expected the lighter cockatiels to switch sooner. Hedrick was also surprised when he analysed different components of the bird's wing beats expecting to find a dramatic change in at least one that signalled the sudden gait transition. But this never happened. Somehow the gently increasing flight components combined to suddenly switch from one vortex to the other.

Hedrick is hoping that recent developments in flow visualisation will soon allow him to see the vortex switch directly. But he thinks he'll stick to antipodean birds in future. `The turtle-doves weren't enthusiastic fliers', he remembers. `It's difficult enough doing these experiments without having to overcome bird psychology too!'