The trouble with chasing a bee: radar struggles to track backyard bugs.

Randolf Menzel runs a sophisticated insect-neurobiology lab in Berlin, but he puzzled for years over how to follow a bee. "Running behind a bee doesn't help very much," he says. Racing along, an observer can keep a tiny spot against the sky in sight for a good distance, but sooner or later the person glances away from the bee for a second. "If you have lost it from sight, you will never find it again," says Menzel.

One athletic student could keep up for record distances of some 50 meters, "but he was falling down a lot," Menzel recalls. So, Menzel gave up on runners and instead maximized conditions for stationary observers to track the bee. "We had lots of students lying on the ground, following it with their fingers," he says.

Finger-pointing didn't help much beyond 30 meters, but the students could at least estimate the direction in which the bee was heading when last seen, which bee researchers refer to as "the vanishing bearing." Menzel's group even instrumented this approach, attaching direction detectors to the arms of three pointing students lying on the ground in various spots. The devices linked to a computer that triangulated the bee's location in the sky and drew a flight path.

Recently, Menzel replaced pointing students with radar equipment. For decades, entomologists have used radar to pick out masses of migrating insects at high altitudes, but the technology proved even less effective than the naked eye at tracking a single insect closer to the ground.

Starting in the 1980s, however, several research groups zeroed in on single insects in flight by modifying a commercially available system used for rescue operations. That method indicated only the direction of flight. Then, a team that's now at Rothamsted Research in Harpenden, England, built a larger, more powerful system that tracked an insect with true radar, which reveals both direction and distance.

Instead of tripping over rocks or squinting into the sky, Menzel and his colleagues could then use radar to see a bee as a little blip on a tracking screen. "It was absolutely exciting," he says.

Such tracking studies are now enabling researchers to ask new kinds of questions, and the past several years have seen radar documentation of the travels of butterflies and beetles. Menzel is tackling decades-old questions about bee behavior as he finally learns where those specks go when the finger-pointing is over.

FIGHTING CLUTTER Radar does such a fine job of detecting invisible things--faraway planes or sky-high clouds of insects--that it's hard to believe that a backyard bee would stump the technology.

The problem, says Rothamsted's Alan Smith, comes from the backyard. Basically, radar works by sending out radio waves and listening for any that bounce back. By analyzing the delay and direction of those returning signals, trackers pinpoint the object that sent them back.

Radio waves sent out close to the ground, though, bounce off the ground, trees, houses--just about everything. The result is a chaos of returning waves. A signal returning from a tiny insect gets lost in the clutter.

A strategy to get around this problem comes from harmonic radar. Trackers fit the object that they want to follow with a device called a transponder, which distorts the incoming wave so that the energy is radiated back at a multiple of the original frequency, what a musician would call a harmonic. Trackers listen for that particular responding frequency and ignore the rest of the radio cacophony.

The idea has appealed to people tracking all kinds of things, including large animals. Traffic-safety visionaries have worked with it in attempts to invent technologies to avoid car crashes.

Harmonic radar appealed to entomologists because the transponder doesn't need its own power source: The energy in the incoming radio wave drives the outgoing wave. No longer having to include a battery, designers saw an opportunity to slim down the device to fit an insect.

OVER THERE The first success using direction-finder radar to track insects came from adaptations of equipment sold by the Swedish company RECCO for locating skiers buried in avalanches. The pulses of radio waves penetrate snow and locate any people wearing transponders.

The equipment was already portable because rescue teams had to lug it out onto the slopes. A device that looks like a flattened hair dryer sends the outgoing radio waves. Its detectors pick up return signals and discern the direction of the buried transponder.

Technically, that's not radar, notes Smith. The word radar comes from "radio detection and ranging," and these systems don't give a distance reading. However, the researchers who use these systems still refer to them as a special kind of harmonic radar.

In the 1980s, Henrik Wallin and Daniel Mascanzoni, both at the Swedish University of Agricultural Sciences in Uppsala, modified a skier-rescue system to track individual carabid beetles. The original system had included transponders made of a sturdy metal diode and a pair of antennas. Designed to be sewn into ski clothes or fastened onto boots, the transponders were far too big for a beetle.

Wallin, now at SIK in Uppsala, and Mascanzoni created a daintier transponder by substituting a diode weighing less that 0.1 gram and fashioning an antenna from just a few centimeters of fine copper wire. The beetles had the reputation of sticking close to home and walking wherever they needed to go, and they walked readily enough with their new decorations in place.

However, the first field test, with night-active beetles, was "a total failure," says Wallin. "Daniel and I came up with this brilliant idea that we could release a whole bunch of them at a convenient time before sunset and then detect the positions next morning--after breakfast, of course." But when they checked, the beetles were nowhere to be found. That clue led to tests in which the researchers discovered that these carabids move considerably farther and faster than anyone had suspected.

Getting the transponders right for the beetles can be a considerable trick, according to Matt O'Neal of Iowa State University in Ames. He and his colleagues worked with the RECCO system to modify transponders for several beneficial beetles that crawl around corn and soybean fields. The transponders that the research team developed tended to snag on weeds and plants, the group reported in 2004.

David W. Williams, however, had more success tracking Asian long-horned beetles. This tree-boring species first showed up in the United States in New York in 1996. Entomologists wanted better information about beetle movements so that they could make decisions about quarantine zones for cutting trees. Williams, now at the U.S. Department of Agriculture's Animal and Plant Health Inspection Service lab on Cape Cod in Massachusetts, says that he heard about harmonic radar for insects through the professional grapevine in the late 1990s and eventually contacted Jens Roland of the University of Alberta in Edmonton.

Williams faced a new design challenge because Asian long-horned beetles do more flying than the carabid beetles did in Wallin and Mascanzoni's earlier studies. The transponders used for carabid beetles were too heavy for flight. However, Roland was working on an even tougher challenge: to get transponders light enough to work on butterflies.

Roland had commissioned a specialty-electronics firm to fabricate gossamer transponders, and the firm sent several samples to Williams. The maker had indeed shrunk the transponders to "hair-fine wires and a diode like a grain of sand," says Williams. In the treacherous world of insect tracking, though, this marvel of weight reduction didn't help Williams. Those transponders worked on dainty butterflies, but for roughhousing beetles, they were too fragile. "You look at them cross-eyed, and they break," Williams says.

Finally, Williams found a USDA lab in Texas with the equipment for making a transponder with a diode just 1 millimeter square-substantial, but not too heavy for beetle flight.

Williams then hit an unexpected hitch. Asian long-horned beetles have such smooth bodies that glues didn't hold well, and the transponders soon fell off. Even superglue didn't stick long.

"I lay awake tossing and turning a few nights worrying about whether the transponders would ever stay on," Williams says. Then, he remembered a trick from his graduate student days. Dental floss is the entomologist's version of duct tape, he says.

Sure enough, Williams found that he could tie the transponder onto a beetle "as a necklace," he says. The beetles even have hornlike projections near their necks that keep the floss from slipping over their bodies. That was "one of the few convenient things about them," says Williams.

He practiced knots on dead specimens but wanted to try the transponders on live beetles before he carried them halfway around the world to perform an experiment on the insects in their native habitat. He took his floss to a quarantine facility in Newark, Del., that had live beetles. His method worked, but not easily. Getting the necklace on a beetle, he says, "takes two full-grown men."

Once Williams reached China, entomologists Guohong Li and Ruitong Gao of the Chinese Academy of Forestry in Beijing took him out to a country road lined with miles of regularly spaced willows. Because China already had plenty of Asian long-horned beetles, the researchers could release 55 necklace-wearing ones over several days without disrupting an ecosystem.

The radar permitted a new twist on an old method of studying dispersal. In the traditional procedure called mark-recapture, researchers mark and release many individuals and later try to recapture them. Researchers extrapolate the fate of the large group on the basis of the usually small percentage that gets recaptured.

Williams and his colleagues, however, adopted an individualized mark-recapture strategy to chart the beetles' paths. The scientists tied transponders onto a few-dozen beetles and then looked for them every day both by eye and by radar. By the end of the 2-week study, the researchers had recaptured almost two-thirds of the beetles. This method found a somewhat slower rate of beetle spread than an earlier study had, Williams and his colleagues reported in the June 2004 Environmental Entomology.

MAKING TRACKS To follow an insect, a radar instrument must measure distance as well as direction. Development of such a device began during the 1970s by scientists funded by British foreign aid.

While studying large migrations of pests, such as swarms of locusts sweeping through West Africa, the researchers turned their attention to the tsetse fly. However, a tsetse fly doesn't get even as big as a housefly. "It was a pretty daunting task," says Smith.

He began by modifying commercial radar systems made for ship navigation. They rely on relatively short wavelengths, which unfortunately can't travel through objects such as trees and houses. However, transponders that respond to short wavelengths can stay correspondingly small, increasing the chances of developing something light enough for a small flying insect. The receivers, though, would be much heavier than the handheld units used in the avalanche-rescue system.

Working on the project intermittently for about 4 years, Smith and his colleagues tried out various designs. By the time that the research team got the system going, funding had dried up. Yet other scientists expressed interest in trying out the technology--on bees, for example.

In the mid-1990s, the radar crew joined the Rothamsted Research team in Harpenden to conduct test flights.

In the early trials, people mimicked flying insects by bicycling in beelines and loops around a military base while carrying transponders mounted on poles. The radar tracked the cyclists and later picked up traces from real bumblebees carrying transponders. Smith and his colleagues announced the success of radar tracking for flying insects in 1996, and the group has since been taking a look at "flight patterns that we could never see before," says Rothamsted-team researcher Juliet Osborne.

For example, the group recently managed to track butterflies for the first time with the true-radar transponders. In the April 22, 2005 Proceedings of the Royal Society B, they reported two distinct flight patterns: great, loopy paths that an insect might use to get oriented and relatively linear flights that look more like beelines than the popular stereotype of butterfly meanderings.

With the same method, the researchers have also found that bumblebees are anything but bumbling. They can compensate for strong crosswinds without being blown off course, and they can fly an average of 7 meters per second, "much faster than anyone really expected," says Osborne.

Since Menzel began studying the tracks captured by radar, he's had to rethink some of his opinions about bees' mental powers. In 1986, when Princeton University's James Gould proposed that honeybees have a cognitive map and can figure out shortcuts from it, Menzel says, "I was one of those who rejected it."

However, he's now coming around to a similar idea of a maplike spatial memory in the insects. He tracked three groups of honeybees, all of which had flown from a hive to perform meandering survey missions over a flat, grassy field. Once the bees had this chance to learn about their turf, the researchers put bees from each group into dark boxes and moved them to a new area of the field. The team then used radar to track bees' flight paths.

At first, the insects searched around in different ways, but eventually they all set off on relatively straight routes either to their home or to a feeder, Menzel's group reported in the Feb. 22, 2005 Proceedings of the National Academy of Sciences. The bees were working out where they and their destinations were, by using a maplike memory of the locale, Menzel proposes.

While doing their experiments on bees flying to and from feeders, Menzel and his colleagues tested the communication power of the waggle dance, one of the most famous bits of honeybee science. Austrian biologist Karl von Frisch won a Nobel prize in 1973 for showing that a forager buzzes back into her hive and, through a dance, communicates the location of the food she's found.

Menzel and his colleagues let some bees discover a feeder and then return to dance the news to the rest of the hive. When newly informed bees started leaving the hive, the researchers captured them, affixed transponders, moved the recruits to a novel spot, and let them fly off. The bees flew the right distance in the right direction to where the feeder would have been had they started from the hive where they observed the original forager's dance, the researchers reported in the May 12, 2005 Nature.

"It was the first time one could directly see how [bees] used information," says Menzel.

Insect tracking has come a long way since his students used to bumble about on bee chases.