When a male horsefly (hybomitra hinei wrighti) spies a suitable female, he chases after her, catches her in midair, and they fall to the ground copulating. The behavior is not surprising, but the speed is. Jerry Butler, an entomologist at the University of Florida, once got a male Hybomitra to chase a plastic pellet fired from an air rifle. "Which it caught in midair and dropped," Butler says. From the speed of the pellet, he calculated the fly was going at least 90 miles per hour.
Scientists are far from fully understanding such biomechanical marvels, but in the past few years they have made an excellent start. For one thing, they now have a plausible story for how insect flight evolved. And if they can't quite say how a bug manages to catch a speeding bullet, they at least have an aerodynamic explanation for why it can stay aloft. In both cases, the key insight is this: Insects are not like airplanes.
If an aeronautical engineer had been in charge of the evolution of flying insects, he would have gotten them started with the simplest form of flight: gliding. He would have had them grow rigid wings on either side of their skeletons, crawl up a tree, and then jump off. Once they mastered gliding, the insects could graduate to flapping, hovering, and buzzing annoyingly in our ears at night. That, at least, was the prevailing evolutionary theory until the 1980s, when Jarmila Kukalova-Peck, a paleontologist at Carleton University in Ottawa, came along. "That's the theory I hope I murdered," she says.
Kukalova-Peck pointed out that the gliding theory not only lacked practicality--until their wings grew long enough to glide, bugs who jumped out of trees would have fallen on their mouthparts--but there was no fossil evidence to support it. The first fossils of flying insects date from the Carboniferous Period, roughly 360 million years ago, and they show jointed wings capable of flapping, not rigid ones designed for gliding. The earliest insects with winglike appendages lived in an aquatic environment. They shared a common ancestor with crustaceans, such as shrimp, which also have jointed legs and jointed exoskeletons.
Crustaceans swim by furiously beating their segmented legs, aided by branches that grow from their joints. In insects, Kukalova-Peck thinks--and recent genetic comparisons of crustaceans and insects have pretty much cinched the case--the flattened branches on the first segment of one pair of legs evolved into wings. Even at an early, stubby stage, those proto-wings, while useless for gliding, might have allowed an early insect to escape spiders or scorpions that lived alongside it on floating mats of vegetation. Extending those wings to catch the wind, the insect might have found it could sail across the water. Flapping them clumsily like a chicken, it might have discovered that a system adapted to locomotion through water also worked tolerably well in the air.
While Kukalova-Peck was rethinking flight evolution, a zoologist named Charles Ellington was starting to rethink the biomechanics of insect flight. The dominant view then, put forward by Ellington's adviser at Cambridge University, Torkel Weis-Fogh, was that most insect wings work like airplane wings. Sure, they move up and down, but that motion generates lift the same way a fixed airplane wing does: When air flows over the wing, it moves down to follow the wing's slightly tilted surface. That downward flow lowers the air pressure above the wing, lifting it just enough to keep the airplane--or bug, Weis-Fogh believed--aloft.
"He thought he'd pretty well sewn up the aerodynamics of insect flight," Ellington says. When Ellington and other researchers put a variety of bugs in wind tunnels, however, the conventional lift they measured was never more than half of what was required to lift the weight of the animal, and often less than a third. Apparently, insects had some secret that aeronautical engineers lacked.
When an insect wing flaps, its front edge goes down and forward and then up and backward, tracing a flattened figure eight. That motion leads insect wings to attack the oncoming air at a high angle. An airplane wing, on the other hand, has a relatively small angle of attack, almost parallel to the direction of travel. Increase the angle (by tilting the leading edge upward) and you increase the lift, but only up to a point. Tilt it more than 18 degrees, on a typical commercial airplane, and the airflow pulls away from the upper surface and the lift vanishes. This is called stalling. The secret of flapping insect wings, Ellington discovered, is that they are always on the edge of stalling, the point of maximum lift.
In the mid-1990s, Ellington took a hawkmoth (wingspan: four inches) and tethered it to the end of a wind tunnel. He then blasted it with strobe-lit "smoke" created from oil droplets and filmed it as it flapped for its life. (To give it a rest, he let it grab onto a Kleenex.) Stereo photographs revealed a crucial feature in the flow of smoke over the moth's wings: a large vortex nestled on top of the leading edge. That vortex, Ellington says, generates 70 percent of the moth's lift. "A vortex is like a tornado," he says. "It sucks the wing up."
But why doesn't the vortex pull away from the wing, as it would on a stalling airplane? To see that kind of detail, Ellington needed an insect bigger than a hawkmoth. He needed the Flapper. The Flapper is a robotic hawkmoth (wingspan: 40 inches) that Ellington built himself. Its wings have four rotation axes that allow them to turn nice, slow figure eights. "It takes about three seconds to do one wing beat," says Ellington. "It's wonderful--you just pull up a chair and sit and look at it."
When smoke hits the leading edge of the Flapper's wing, you do see a vortex form, but that vortex immediately takes a sharp turn and spirals out along the wing toward the tip. By draining air outward, that spiral keeps the vortex from growing so big it must detach from the wing and create a stall. The source of the spiral flow is simple: As the wing pivots, the tip travels farther than the base and so moves through the air faster. "Because the outer parts of the wing are moving faster, they're generating more lift, and there's more suction out there," says Ellington. "So there's a pressure gradient sucking air to the tip."
A stable leading-edge vortex, Ellington thinks, is what keeps most insects aloft. True, it doesn't explain their aerial acrobatics, but another phenomenon, recently discovered by Michael Dickinson at the University of California at Berkeley, might help. At the end of each downstroke and upstroke--as it enters the turns of the figure eight--an insect wing rotates, shedding the leading-edge vortex. For an instant, the rotation accelerates the airflow over the top of the wing, thus generating a burst of even greater lift. By controlling the timing of those wing flips, the insect can steer the direction of the lift. That may have something to do with how H. hinei wrighti manages to turn in mid-flight and pursue a passing female at 90 mph.
With military funding, Ellington and Dickinson are both now trying to build flying microvehicles. To be useful for surveillance, Ellington says, the vehicles would need to generate enough lift to carry an instrument payload and ideally be able to hover. Insects are masters at high-lift hovering. But building a flapping robot with a six-inch wingspan that can be steered and won't shake itself to bits may still be beyond the skill of human engineers. It's possible, says Ellington, that a microhelicopter, with propeller blades that imitate insect wings--short, wide blades with a high angle of attack--will turn out to be an easier solution.
On the other hand, a microflapper would certainly be more fun. "I've been working on flapping flight all my career," says Ellington. "It's always been my ultimate goal to have a little radio-controlled flying insect. I mean, that's what it's all about! The day that can be done, we'll have really understood flapping insect flight. We're starting to get close to it now. But we're not quite there."
Michael Dickinson's official insect flight laboratory Web site can be found at socrates.berkeley.edu/~flymanmd.