PAGE 2 OF 2
To get his research subjects, Leonardo dons boots and a net and catches nymphs from the outdoor pond at HHMI's Janelia Farm Research Campus. So far, he's collected and raised 10 species of dragonfly.
The flight arena fully occupies a basement room about 14 feet wide, 18 feet long, and 15 feet high. In the beginning, it had white walls, a white ceiling, and a white floor. Put a couple of dragonflies in it, and they wouldn't budge; they just sat still on the bare floor. Being visual animals—they have two compound eyes, made of thousands of lenses, plus three simple eyes—they had no frame of reference, nothing to give them clues as to where they were or where they should fly.
So, Leonardo and his technician, Elliot Imler, started adding high- and low-tech props from floor to ceiling. First they installed a carpet of artificial grass. Dragonflies zipped around the room, but without any other visual references, they flew in circles and into walls. To slow them down, the team put in “speed bumps”—vertical stripes on the walls, which later gave way to wall-to-wall posters of verdant forests and tulip gardens. To make it homier for the water lovers, they added a shallow pond and decorated its edges with plastic flowers and cattails, which serve as perches. Aimed at two of the perches are high-speed video cameras, one per perch.
Mindful of the insects' internal clocks, they re-created dawn and dusk with an array of lights programmed to brighten and dim, from east to west. A humidifier maintains constant moisture and the temperature hovers at 80 degrees; heat lamps directed at flower perches provide the additional warmth that dragonflies prefer.
To prompt the dragonflies to hunt, Leonardo and Imler suspended small plastic trays carrying banana bits crawling with fruit flies. The trays buzz regularly to startle the flies off their feast so the dragonflies can see and pursue them.
What keeps the arena softly humming with a steady stream of four to five dragonflies at a time is the lab's vivarium, where about 10 different species of dragonflies are raised from nymphs netted from Janelia's outdoor pond.
After several months of fine tuning, about half the dragonflies released in the room appeared to behave as if outdoors. Of those, 80 to 90 percent of one particular species, Libellula lydia, acted normally—that is, flying and hunting in the room for at least a couple of days. This fraction is large enough to satisfy Leonardo that he's figured out an effective, basic formula for producing this complex behavior in an experimentally controlled setting.
Now, he is working with engineer Reid Harrison at the University of Utah on a wireless, electronic “backpack” for his tiny subjects. The pack, which will be glued to the belly of the dragonfly, will carry a mini-telemetry system. It will connect to electrodes inserted into the dragonfly's body. The electrodes will detect signals from the neurons Olberg identified, and a transmitter will send the data to a remote computer while an array of high-speed cameras simultaneously measures the dragonflies' flight path.
Together, the videos and the data from neuronal signals—and eventually muscle contractions—may one day yield a complete picture of how dragonflies' neural circuitry makes them such enviable hunters.
Photos: Erik Johnson