Insects aren't airplanes. If they flew like airplanes, they would fall right out of the sky because their wings are much too small. But since they flap their wings, they can generate a `leading-edge vortex', a rotating element of fluid along the front of the wing that dramatically increases lift briefly and helps keep them aloft. A recent debate has focused on how and when the leading-edge vortex forms and what it does once formed. Gregory Lewin and Hossein Haj-Hariri have added more complications to the argument in their computational study of small flapping wings, published in J. Fluid Mech. They describe how the leading-edge vortex can interact with other wing vortices to produce forces that differ on the up- and downstroke, even when the wing movements are symmetrical.

Normally, each time a wing changes direction when it is flapping, it sheds a vortex off the back edge, the `trailing-edge vortex', which then drifts backwards to form the wake. Sometimes, it also produces a leading-edge vortex at the front of the wing, which can remain attached at the front edge or drift back to interact with the trailing-edge vortex.

Using a 2-D computational model of a wing moving up and down in steady viscous flow, Lewin and Haj-Hariri found that the leading-edge vortex can interact with the trailing-edge vortex in either a completely symmetrical way,a completely asymmetrical way or can produce a type of interference pattern with the trailing-edge vortex.

In the symmetrical mode, which happens at low frequencies and amplitudes,the wake consists of alternating vortices with a backward-pointing jet, like in ordinary flapping propulsion. By contrast, in the asymmetrical mode, which can happen at high frequencies and amplitudes, the wake deflects to one side for many beats but occasionally flips sides. The same effect has been seen with flapping airfoils in wind tunnels, but this is the first numerical study to reproduce it. For insects, it would probably be detrimental, since the lift force would increase or decrease unpredictably. At intermediate frequencies,Lewin and Haj-Hariri observed something like an interference pattern between the leadingedge and trailing-edge vortices, with dramatic changes in the output power over four or five beats. They showed that the leading-edge vortex usually moves backwards, interacting with the trailingedge vortex, but sometimes, because of subtly different background conditions, the leading-edge vortex from one stroke wraps around the front of the wing to interact with the leading-edge vortex that forms on the next stroke, producing a dramatic drop in the power output.

What do these results mean for our understanding of insect flight? First,it isn't clear whether these strange effects even happen during insect flight. Insects might avoid these effects by tuning how they flap their wings. Nonetheless, Lewin and Haj-Hariri show that small changes in the background flow, perhaps caused by gusts of wind or turbulence, can produce large qualitative changes in the wake and ultimately in efficiency and power output. Specifically, the wing's propulsive efficiency depends critically on when the leading-edge vortex separates. Efficiency is highest when the leading-edge vortex stays attached to the wing for a whole stroke and merges with the next trailing-edge vortex. But Lewin and Haj-Hariri have shown that when the leading-edge vortex remains attached for longer, or even wraps around the front edge of the wing, flight efficiency decreases. So, although the leading-edge vortex can increase efficiency in some circumstances, it isn't beneficial in all. With these results, the picture of the leading-edge vortex is growing even more convoluted.

Lewin, G. C. and Haj-Hariri, H. (
). Modelling thrust generation of a twodimensional heaving airfoil in a viscous flow.
J. Fluid Mech.