"Flies would see that light as movement, like lights turning on and off at a disco," he observes.
A blowfly, tethered by a noose made of thread, is watching. "That light is cycling at 60 hertz," Dickinson explains. Mere humans, with a "flicker fusion" perception of 30 hertz, or 30 cycles of light per second, see the light as a solid glow. Flies, on the other hand, don't see solid light until 300 hertz.
Dickinson, an associate professor of integrative biology and a self-trained aerodynamicist, has studied the physical oddities of flies for half a decade. Eternal vigilance is the pact flies have made with evolution. Over their average monthlong life spans, they are always on the lookout for threats -- a fly swatter here, a bird there.
And evolution has held up its end of the deal, giving flies the fastest eyesight of any insect and extraordinary flight capabilities. These flight capabilities are what most interest Dickinson -- and potentially, NASA engineers and the U.S. military.
When the common blowfly, Calliphora vicina, isn't landing on ceilings or walls, it's most commonly found working the garbage can and dog excrement circuit. Still, biologists have been attracted to flies for "literally centuries," Dickinson says. For instance, neurobiologists use flies to study nervous system development.
What scientists haven't been able to piece together, though, has been the puzzle of flies' flight control system. The most popular hypothesis was that flies were able to operate so quickly because neural messages shot directly from their hyperfast eyes to their wings.
But when Dickinson began sticking electrodes into flies' wing muscles in 1994, he found there was no such direct connection.
Something else had to account for flies' ability to make flight adjustments inside of 30 milliseconds -- something Dickinson was sure flight control designers would dearly love insight into.
But though Dickinson had studied flies at the Max Planck Institute in Germany -- also known in the biology world as "Fly Central" -- and was teaching biology at the University of Chicago, he was as stumped as anyone by the mystery of fly flight control. His only comfort: No one else had the answers either.
Still, he reasoned, "Evolution didn't build a fly from scratch." We watch a blowfly in super slo-mo on a video monitor. "It built it from something that was less than a fly. Organisms are just clunked together after all."
What Dickinson has discovered is that, millennia before flies looked the way they do now, they possessed an extra set of wings. These extra wings have persisted, as vestigial organs called halteres. Enlarged on the video monitor, they look like tiny clubs. Biologists dismissed halteres as an evolutionary hangover. But Dickinson suspected there was more to them than that.
In fact, the halteres perform the same function as a human's semicircular canals. They are the fly's balance organs. Dickinson found this out with a very simple experiment: When he removed a fly's halteres, the insect lost its equilibrium and crashed.
As we watch, the blowfly's natural wing speed of 150 flaps a second is slowed to one-hundredth of that speed on the video screen. The insect's halteres beat in anti-phase just under its wings. This image is transmitted to us from the wraparound "flight arena" in which Dickinson puts the fly through its paces. Glass electrodes, two microns across, probe the muscles of the fly's halteres; they are connected by multicolored wires to a computer, which records the muscles' electrical activity as the fly alters its pitch, roll, and yaw in response to a series of striped patterns projected across the flight arena's screen. To the fly, the effect must be like being lost inside a barber pole.
Dickinson looks back at the ceiling lights. "There's no aerodynamic mechanism that keeps the fly stable," he says. "This is a classic example of extremism in animal design."
Aircraft designers have to work within a fixed set of aerodynamic principles: thrust, lift, weight, and drag. But no matter how they tweak the variables, they always face the deadly problem of stall, a state where a wing's lift gives out and the aircraft plunges earthward.
But Dickinson says designers have never taken a serious look at flies' flight control system, which he argues could serve as a model for controlling high-speed aircraft, such as NASA's Mach 10 Hyper X. "It's only been recently that the military and industry have been interested in flight at all. They thought, 'We know how to build airplanes. We don't need entomologists or zoologists.' "
How flies fly, he says, has profound implications for aerodynamics, interplanetary travel, and Earthly rescue operations -- assuming that technology catches up with scientific theory.
The hope that it will is what animates Dickinson's face as we talk -- that, and the hope that soon he'll get to play creator of an army of robotic flies.
Robotic flies? "There are a lot of applications that even liberals could get behind," says Dickinson. (The subtext being, there are many more applications that the military could get behind.) Earthquake salvage, for example, where rescue dogs are limited by their size. Dickinson says it would be easy to program robotic flies with a "find anything emitting CO2" command and set them loose in a pancaked structure. "They're incredibly good at finding things with very simple algorithms."
The other big project: planetary exploration. Dickinson claims that a colony of robotic flies could be sent into space to work collectively at assessing territory -- much as ants do -- and that this would be less risky than putting our space dollars into one big expensive project, in which if anything goes wrong "you're totally hosed."
Ronald Fearing, an associate professor of electrical engineering at UC Berkeley, is talking with Dickinson about creating a robotic fly prototype, which would be a half-inch long and weigh less than a dime. Fearing says it will be at least four years before the polymer fly will be able to take off without a tether. And it'll be another decade before it will be reliable enough to use for earthquake rescue efforts, much less space exploration.
The technological challenges -- miniaturized computers and high-powered/low-weight batteries -- of creating a robotic fly are immense. But, as wild as some of the applications seem, technology often catches up with theory. As we talk, Dickinson fields a phone call from an engineer at Cal Tech who's designed fly eyes on a computer chip and wants to work halteres into the package. And Dickinson's study, recently published in Science, has been reported in newspapers in Western Europe.
But it says much about Dickinson's research subject that all of these stories open with the general titter: Biologist explains why people have such a difficult time swatting flies. "I don't mind them trivializing it," he says. "I just wish they'd trivialize it accurately.