The Homing Instinct: Meaning and Mystery in Animal Migration

Although early American woodsmen, whose lives depended almost directly on the knowledge gained by close contact with nature, were beelining devotees who had deduced that honeybees recruit hive mates, it would remain for Karl von Frisch to unravel the marvelous story of how the bees communicate within the hive. He earned the Nobel Prize in Physiology or Medicine for this work. I feel lucky that a Maine neighbor, Floyd Adams, took me beelining when I was eleven years old, and that when I was a teenager, my father gave me an inspiring little book by von Frisch entitled Bees: Their Vision, Chemical Senses, and Language. It explained the experiments that he and colleagues had performed. They were mesmerizing because they connected the practical experience of beelining in the Maine woods with the imaginative power of a scientist who had penetrated into the core of the bees’ world, their hive, their home.

Floyd’s family’s home was the farm four hundred meters down our dirt road. It was populated by chickens, geese, cows, pigs, plus all the other usual and unusual wildlife that lives in a place with a tolerance for disorder. Along with Floyd, my companions were the four Adams boys, Butchy, Billy, Jimmy, and Robert, an in-law of theirs. Floyd, a dark-haired, mustachioed, wounded Marine Corps veteran recently returned from the Pacific, had a bad limp and a thirst for Black Label beer. Leona, his blond, petite wife, appreciated his fondness for honey but less so his taste for beer. He and the “boys,” after a hot day haying, sometimes went fishing on our nearby Pease Pond in the evening, but in August our big draw was always the beelining.

After we found a bee tree, we carved our initials into the bark to proclaim ownership (property lines were irrelevant with regard to bee trees; finders keepers was the rule), and at some convenient time we returned with crosscut saw, axes, wedges, a beehive, and pails and kettles for honey. Getting part of our living from the land was fun, and it meant understanding and using the bees’ homing behavior to find their hollow trees in the forest and resettling them into a new home, which we brought back to the farm and set up at a window in the attic of the house.

Fast-forward to a quarter-century later: My nephew Charlie Sewall and I are in a patch of goldenrod blooming in a pasture where each fall the wild honeybees gather nectar to top off their honey stores for the coming winter. We start by capturing a single bee in our bee box, a simple four-sided wooden box that has a ten-by-fifteen-centimeter piece of honeycomb with sugar syrup filling out the bottom. We dab the box with a drop of anise for scent and capture our bee by holding the box under her after she has landed on a flower and then slapping the box cover over her. At first the captive buzzes in the box trying to escape, but the buzzing stops when she stumbles onto the sugar syrup and starts to tank up, which will take her a minute or two. We then remove the cover and set the open box onto a pole that reaches to just above the tips of the goldenrod. We gently daub her with a spot of paint while she is absorbed in sucking up syrup, as I remembered Floyd doing. We then hunker down into the goldenrod and wait as she continues sucking up her newfound sweets that she will soon share with her hive mates. After about two minutes, her honey stomach is filled. She crawls out onto the edge of the box, stops to wipe her antennae with her front feet, lifts off, and flies back and forth downwind of the box. We duck lower to keep her silhouetted in sight against the sky as she starts flying loops, which become increasingly wider and oriented in one direction. Finally she straightens her flight path and takes off, making a “beeline” into the distance. Knowing that nobody in that direction keeps bees, it’s clear that she is on her way to a bee tree. She will soon be back with others, and we then consult our wristwatches to time her trip. A bee flies about four hundred meters a minute, and it may take her three to six minutes in the hive to regurgitate and unload her honey stomach’s contents into the mouths of begging, receiving bees.

We settle down and wait, and after perhaps ten minutes or less a bee suddenly appears and makes very rapid zigzagging flights just downwind of the box. The sound of her fight has a higher pitch than that of the bees foraging on the nearby goldenrod flowers. This means that she is more motivated and has a higher body temperature because of the rich food she is expecting. She settles into the box and starts imbibing the syrup. More bees will come soon, and when they get near our bee box, they will be guided in by the scent of the anise that marks the spot. After they tank up, we watch their flight directions.

If the food is in the immediate home vicinity, the bee does a “round dance” on the honeycomb when she returns to her home. She repeatedly runs in small circles while shaking her abdomen, and she regurgitates small samples of her find at intervals during her dance. If they become motivated after receiving information about the quality and scent of the food advertised, her hive mates leave the hive and search for the advertised food. If it is beyond a few hundred meters, the bee alters her dance to also contain information concerning location. The distance of the journey to the food is proportional to the duration of the waggle runs, and the angle of the straight runs with respect to the vertical direction informs the bees in what direction to fly when they leave the hive. If the straight run is in the up-direction on the honeycombs (which always hang vertically in the hive), the food source is in the direction toward the sun. If the food location is, for example, at an angle of ten degrees to the right of the vertical, the food direction is ten degrees to the right of the horizontal component of the sun direction when the bee would fly from the hive. Thus, her behavior is a symbolic representation in body movements of the flight to the food.

The first steps in the evolution of recruitment likely involved simple alerting signals in or at the nest entrance before takeoff. Other bees could have followed those signaling bees, probably by scent, for at least a short distance in flight. Through a few million years, the alerting likely became modified to take on an ever-greater leading function by bees flying in an ever more conspicuous manner in the direction of the food, so that followers could start off flying with ever greater accuracy in the right direction. These flights, later in the evolutionary progression, were eventually restricted to a buzz run directly on the top of the combs, but still in the food direction. We can infer this, because such “primitive” recruitment is still found in some tropical honeybee species that have their combs in the open air, where this mechanism makes sense. But open-air homes, though convenient for such communication of food location, were vulnerable to predators and also precluded the bees from living in huge areas of the globe, those with cold climates.

Homes in hollow trees allowed the bees to live in areas where they would otherwise be excluded because of cold and/or nest predators. But in such safer homes the combs hung from the roof of a cavity and left no horizontal dancing platform, and additionally the “dance floor” was now in darkness, so bees could not point directly toward the food. Even if they could, they would not be seen. But a breakthrough for indicating horizontal directions on vertical surfaces became possible after some bees started using the hanging flat surfaces of combs as their dancing platform while indicating the sun’s location as the up or “toward” direction in their dance. Additionally, tactile rather than visual orientation became predominant for recruits in reading the code within the nest.

It is amazing enough for an animal to be able to navigate to a location it has never been to before. But some ants do something even more amazing. In North Africa, desert ants live in underground homes where they are protected from the heat. But they must venture out onto the searing surface periodically to forage by scavenging on heat-killed prey. The ants are fast runners that have evolved a very high tolerance for heat. Still, at times it is a matter of life and death even for them to make it back to their cool underground home; they cannot afford to wander on the sand surface for an extended time without access to their shelter to cool down and replenish body fluids. This is where their homing ability comes in; they may have zigzagged in all directions to find a heat-killed insect, but after finding one they must make a straight “ant line” directly back home. This begs the question, Since they are often on a featureless plain and have not kept a steady course, how do they know in what direction to head home?

If one captures bees in one pasture and releases them in another, they usually depart in the direction they would have flown from the original field. That is, they act as one would expect if they do not realize that they have been moved to a new location. Rüdiger Wehner and his colleagues at the University of Zurich came to the same conclusion about desert ant homing in their lifelong experimental studies. The ants use the sun as a compass, but a compass is not enough; the ants, when released from a point they had not themselves traveled to, like the bees caught in one pasture and released in another, apparently got lost.

For homing you must know where you are on “the map” before you head off in the correct direction. The desert ants can return home, but only if they walk to where they find themselves. Wehner concluded that the ants’ homing mechanism involves somehow calculating where they are at all times, probably in measuring distance by keeping a kind of count of their steps, and also keeping track of the angles of their direction from their home relative to the sun’s location. These were not mere speculations, but a hypothesis tested in painstaking experiments that entailed altering the ants’ perception of the sun (holding filters over them that varied the direction of polarized light that they, like bees, use in orientation) and altering their stride length (altering their leg length by gluing on extensions) to find out what information they valued and how they used it. Presumably bees could also have a similar “map sense,” and Randolf Menzel, a neurobiologist in Berlin, was trying to find out how it might work.

Menzel runs the large and active Institute of Neurobiology at the Free University of Berlin, and one of his projects was the burning question of how honeybees seem to find out where they are in order to be able to go where they want to be. Honeybees are suitable animals with which to study this problem because, like ants, you can count on their motivation to return home after they are loaded with food.

We can’t look into a bee’s brain and determine what it knows and what it wants. However, clever experiments based on the bee’s natural history permit inferences. We can determine, for instance, where a bee perceives herself to be relative to her hive. If a bee regularly visits a feeding place, she knows where she is, because she always flies off in a straight line from it back to her hive. If we then remove either the hive or the feeding spot, she circles in the area where her target had been. We know what she is looking for, because when we provide the hive and/or the feeding station within the area where she circles, she quickly finds it. But suppose we capture our bee at the usual feeding station after she tanks up on honey or syrup, put her into a dark box, and then carry her “blind” to a place she has never been. As mentioned, most bees will then make a beeline in the same direction they had normally flown to return to the hive. They will fly as far as before but find no hive there. Yet, they usually eventually do make it back home. How do they find their way? What do they do until they reach home? Until Menzel’s experiments, it had not been possible to track them in flight when they were out of sight out in the field. Menzel had a tool — radar but with a unique twist — whereby he could trace bees’ actual flight paths over a kilometer away by radar and record them on a computer. And he invited me to come see the work in progress.

Bee flight paths. A. A bee’s first trip from a flower patch or bee box back to its bee tree (hive) begins with an orientation flight. B. Later trips are more direct. C. After a bee has been transferred while “blindfolded” to a new spot, she acts as though she perceives herself to be still at the same place as before.

The problem of tracking small objects such as insects from a long distance by radar had always been that radar would “see” too much. You could not isolate and then plot a single specific bee out of all the extraneous noise of echoes bouncing off all objects. The new insect-tracking radar technique started in 1999, when Joe Riley, a British researcher, applied a radar system able to track very small objects over long distances by attaching to the insect a small device that, after receiving the energy of an electromagnetic sound pulse, would respond with a frequency other than that of the transmitted ultrasound. The receiver is then tuned to amplify only that frequency. In this way, it became possible to track the flight paths of individual preselected bees equipped with the appropriate transponders because the echoes from all other objects were filtered out.

The Menzel group’s electronics technician, Uwe Greggers, adopted the Riley system in 1999 and 2001 and got interesting results, but then ran into software problems. Nevertheless, given the promise from the data they did get, the scientists contacted a radar specialist at Emden (north Germany) who agreed to develop the system. The Menzel group then needed to find the right site in which to use it. They needed to locate the experiments at a large flat area devoid of trees in order to be able to record the complete flight paths without interference such as the bees’ trying to avoid objects or being attracted to them. The closest suitable area was an expanse of marshy meadow about a two-hour drive from Berlin. The large, idyllic farmstead near the village of Klein Lübben and land associated with it had accommodations for seven or more helpers, making this site amenable.

One Menzel group experiment in the works when I visited involved training individual bees to expect food at two widely separated feeding stations, but only one station at a time was open to them. I had no idea what to expect, and on my day with the team I was eager not only to watch the bees but also to see the experiment in action.

It was early in the morning when Menzel picked up Greggers and me for our trip to the experiment site in the Brandenburg countryside. We loaded a large, heavy printer that would be used to handle the large-scale printouts of flight paths, and then we were off down the Autobahn. Two hours later we arrived at Klein Lübben, a quiet village of farmsteads that at least in outward appearance has changed little since medieval times. The fields were several kilometers square, flat, and moist — perfect also for frogs, and hence storks which nest there in baskets attached to the tops of red-tiled house roofs. Swarms of starlings swirled through the air, and a pair of white swans paddled serenely down a canal along a dirt road, followed by a line of five still-downy gray cygnets.

At one end of the study field stood a steadily turning radar apparatus with a large round antenna for sending out the signal. A smaller dish antenna mounted directly above it would receive the transformed signal bouncing off the transponder on an airborne bee in the field. On the field sat two blue triangular tents and three yellow ones. They were experimental landmarks for bees that could be made available to them, to find out if they used them, and manipulated for experiments by changing their locations. In the distance sat a beehive, and I noticed a man running from it. He was wildly slapping himself, in an obviously defensive mode. He had been assigned to provide food for the bees close to the hive and then was to gradually move the feeder into the field so that a population of bees from that hive would be available for us to study when we arrived at midmorning. He had come too close to the hive, and at that moment it was he who was getting dispersed over the field, not the bees. Also, as we soon found out, there were no bees coming to the two feeder stations, as they were supposed to have been by now; the student had apparently overslept or been otherwise distracted from his assigned job of luring bees.

The experiment we wanted to do was in doubt. This was serious. Two hundred thousand Euros had already been spent on this study, and the boss was intolerant of negligence. Luckily, bees from hives used previously for another experiment were still coming into the field to search for feeders. He could let some of those bees find the feeders and then train them to come back to specific locations.

For our experiment we needed to establish two feeding stations, A and B, separated by about three hundred meters. Certain bees were already keyed into the routine. When I walked across the field, one bee started following me. It looked most extraordinary: it had a lot of blue and green color, not just the usual plain brown honeybee attire. As soon as Menzel’s helper and I set up our feeder, this specific bee landed on it and immediately started to suck up the rich sugar solution. Now I could examine her more closely: the green was a plastic tag with the number 29 on it that had been glued to her thorax. The blue was a slash of paint that had been daubed onto her abdomen.

Within a few minutes an assembly of several differently color-coded bees was lined up around the edge of the syrup dish. All were sucking up syrup. Some had green on the thorax, some had blue, and still others had yellow tags on their thoraxes, with additional daubs of white, blue, or yellow paint on their abdomens. Uwe Greggers and the unfortunate helper immediately started logging a list of the bees that had shown up in a notebook.

Each bee tanked up quickly, flew off directly toward her hive at the other end of the field, and then came right back to take a next load. Newly recruited (unmarked) individuals were also coming every minute to our site A. At the second feeder (site B) there was a similar flurry of activity, except it involved different individuals.

Menzel then instructed us to move our food station A one hundred meters closer to the second one, B. Bee numbers 29 and 30 green, both with blue tails, number 2 yellow with white tail, and number 39 green with red tail (who had all been present at A) then almost immediately started showing up at B, the new location. When crowds of bees had done the same, we removed one station and put the remaining one into the middle, between the two original sites. Next we moved our feeder to site B. Most of the bees, such as 30 green with blue and 39 green with red, who had been at the previous site, showed up. That is, we had trained bees who had been at one site to come to the second site, so we knew they now knew two sites and could potentially use either as a reference site to return home.

For the planned experiment it was important that the bees forget the intermediate sites that had been instrumental in getting them to go to the two widely separated sites. So, for the rest of the day, we alternately fed the bees, first at site A, then at site B, and monitored which individuals were showing up at both sites (most of the individuals continued to forage at either one or the other site).

We were now, near the end of the day, finally ready to move from training to trials. The experimental plan was to select one of the bees who knew both sites. This bee would, after feeding at one site and getting ready to leave, be captured in a dark box and thus “blindfolded” and then brought to a third feeding site where she had never been before. Here she would be released after being equipped with a radar-tracking transponder. We presumed that she would do at least one of three things: she might recognize where she was and fly straight home; she might instead fly off in her original (now wrong) direction; or she might immediately know that she was at an unknown location and search until she found one of her two feeding sites and from there take a direct beeline home. Knowing her exact flight path would allow us to distinguish among the alternatives, which would be essential to ultimately decoding her homing mechanism. Setting up this experiment had taken a long time, but I would now, possibly, be treated to an exciting demonstration of bee homing, one I could never have imagined possible.

Menzel picked up his walkie-talkie to call the radar station: “Mike, we’re now going to put a transponder on a bee — are you ready?” Mike had spent some years in the army where he was trained on radar, and he was now working part-time while getting a university degree in computer technology. He replied yes, he was ready. Menzel then took me to feeding station A, where a whole lineup of bees was coming and going.

“Which one do you want?” Menzel asked me. I wanted a bee that I had gotten to know over the course of the day, so I chose 39 green with red-tipped abdomen. We waited for her to arrive and let her feed for a while. As planned, Menzel then held a glass vial over her while she was distracted sucking up syrup. When she was full, she walked up into the vial, and Menzel corked it shut and darkened it by wrapping his hand around it. We then took her to a site distant from both feeders, a place where she had not previously fed and from where we would now release her.

The vial holding “39 green” had a plunger at the bottom with a wide-mesh screen at the top. Menzel gently pushed the plunger in and forced the bee up against the screen, held her there, and picked up a tiny transponder (a wire holding a diode with a sticky pad at one end). With fine tweezers, he deftly removed the protective paper from the sticky pad and glued the transponder onto the top of the bee’s thorax. “Ready?” he radioed Mike.

“OK.”

Menzel removed the plunger and held the vial with the open end up, for the bee to crawl up. She hesitated at the lip of the glass, groomed her antennae, and then lifted off. She showed no strain in flight. (The transponder’s weight is twenty milligrams, and a bee can fly with double her body weight, carrying a hundred-milligram load of nectar in her honey stomach plus two pollen packets on her hind legs.) However, she flew only two or three meters before dropping down into the grass, stopping to preen herself some more. But a couple of minutes later, she finally took off again. Mike, who was now monitoring her flight, radioed us. At intervals we heard: “She is heading south-north-east-north-west-south.” Then, finally, Mike continued: “Now her path is straightening out — now she is heading directly for her hive!”

She had suddenly oriented correctly. This was the crucial point: she had apparently recognized something that had “placed her on the map,” so that she then “knew” in what direction to fly to reach home. Assuming she had taken a path she had never taken before, did her successful homing suggest a “map sense”?

I ran over to the radar tent where Mike showed me the radar screen and the dots where the three-second successive readings traced the bee’s path. A computer screen, where software had converted the time and directions of the bee’s flight path into different-colored images for easy reading, showed that the bee’s original flight direction was toward where the hive would have been had we not moved her from her feeding spot. In other words, she acted as though she didn’t know where she was when we released her. As expected, however, after she reached the area where her hive would have been, she flew loops in several directions. Then, after she had flown ever-farther away from both her real and “would-have-been” hive locations, she suddenly seemed to orient and flew directly toward the hive. Amazingly, it was along a route that had not been her normal foraging route from her two feeding sites. Had she perhaps seen a blue or a yellow tent and, having learned their relationship to each other during previous orientation flights, transposed that information to fix her new location? Only more bees could tell.