It’s not every day that one gets to prove a Nobel Laureate wrong.

I say so with all humility, given the stature of the laureate: Karl von Frisch, the German-Austrian scientist, was a founding giant in the field of ethology—the study of animal behavior in their natural habitats. In the 1920s, von Frisch decoded the “language” of honey bees, which the bees use to tell each other about the direction and distance to some source of nectar, and its quality. He was awarded the Nobel Prize in 1973 for this immense body of work. He had, however, misinterpreted one aspect of the bee’s behavior: how it estimates distances. My team had the good fortune to stumble upon the correct explanation.

A forager honey bee sets out from her hive, performs a meandering search for food over a distance of up to six miles, discovers a good source of nectar, and then makes a literal bee line back home. Upon returning, she regurgitates her bounty to her nestmates, who store the nectar in cells of the honeycomb. Then, in the darkness of the hive and on its vertical surface, the returnee performs a choreographed, information-rich “waggle” dance, shaking her abdomen from side to side, to brief her nestmates about the precise location of the food source.

These female foragers do all the hard work. They go out to collect nectar and pollen. They wear themselves out and die in a month. The male drones sit there doing nothing. They have only one thing on their mind, and it isn’t hard work.

Karl von Frisch deciphered this dance. He noticed that the bee waggles along in a straight line from point A to B inside the hive. Since the bee is moving on the hive’s vertical surface, the line from A to B can itself go straight up, from the bottom to the top, or it can be at some angle to the vertical. Once the bee reaches point B, she turns right and loops back to return to A. She then waggles over to B again, and this time loops left, and returns to A. She repeats this sequence of waggles and right- and left-handed loops numerous times.

This simple dance encodes two important pieces of information about the food source: distance and direction. It turns out that the time the bee spends shaking her abdomen is proportional to the distance she has flown to find food. And the angle that the line from A to B makes to the vertical indicates the angle between the current bearing of the sun along the horizon and the direction to the source of food. For example, if the sun were due east and the food source lay due south, the line from A to B would be at 90 degrees to the right of the vertical, where the vertical represents the sun (east) and the line represents the food source (south). The bee even encodes its ‘attractiveness’—an assessment based on factors such as the abundance and sweetness of the nectar, ease of access and extraction, and its proximity—in the intensity of the dance, by varying the number of loops and their rate of repetition: the more intense the dance, the more attractive the nectar.

How does a bee know how far she’s flown? Karl von Frisch tried to work out the bee’s odometer. Cleverly, he trained bees to go from their hive to a food source that was about 500 yards away. When they came back to the hive, their dance signaled a distance of 500 yards. He then put tiny lead weights on top of the bees to increase their loads while they flew. When they came back, the bees reported a much larger distance, leading von Frisch to conclude that the bees must be using some measure of energy consumption to work out the distance flown—they were consuming more energy because they were carrying excess weight, and more energy equated to more distance.

This was a plausible explanation. But it was wrong.

The Hole in the Wall

Before my research involvement with bees, I was quite wary of them. It was only when I first encountered them in a scientific setting in the early 1980s, in Rudiger Wehner’s lab at the University of Zurich, that I learned that bees are not innately aggressive creatures. They get aggressive only when they perceive a threat. If you go near a hive and poke a finger inside or otherwise disturb it, they come out to defend their home and sting you. But it’s not in their interest to sting you because, when they do, they lose their stinger. It stays inside you, pumping its poison. The bees bleed to death, because the loss of the stinger tears a big hole in their abdomen, allowing all the blood to flow out. They sting you only when it's absolutely necessary.

But when they're coming to get a food reward—some sugar water set out for them, for example—they are as docile as cows. You can actually reach over and stroke them on the back as they're drinking.

I became fascinated by bees because of their ability to be trained—they are incredible learning machines. You don't have to wait for them to do something; you can invite them to come into your lab and they will do almost anything you want, provided you give them a good reward at the end. One of the nice things about working with bees is that you're not harming or torturing these creatures in any way. I used to do insect electrophysiology, which involved cutting up insects and putting electrodes inside them. I stopped doing that. It didn’t appeal to me; I think insects feel pain. With bees, we do mostly behavioral experiments. Once you finish, they are free to pursue their normal life. They are not imprisoned. If they find the experiment dull or boring or not rewarding enough, they can go elsewhere. They come only of their own free will.

After four years in Zurich, I came to the Australian National University in Canberra in 1985, and continued my experiments with bees. It was during one of these sessions that a chance observation of bee behavior helped us figure out how bees measure distance to a source and even how they land ever-so-gently on a flower. In doing so, we showed that Karl von Frisch had wrongly interpreted the results of his experiments with lead-bearing bees. It began with a hole in the wall.

I don’t remember the experiment we were conducting, but we were trying to get bees to come into our lab to feed on sugar water. The bees had to take a detour and fly in through the window. There happened to be a hole in a wall which provided a shorter route. The bees began taking that route. We noticed that they were flying right through the center of the hole. To do so, they would have to accurately gauge the distances to the hole’s edges and fly such that they were equidistant from both. How were they doing that?

A marked bee entering an experimental chamber through a hole in the outside wall.Photo: Shaowu Zhang.

One way would be to use stereoscopic vision to discern depth and distance. As humans, we can do that. Our interocular separation—the distance between our two eyes—can range from five to eight centimeters. Our brains use the parallax induced by our eyes, which is the apparent difference in the positions of objects as seen by each eye, to create our 3D vision. But insects don’t have stereo vision because their interocular separation is so small—it's almost like they have just one eye. Also, when they're flying through a hole, both eyes are not seeing both edges: the right eye sees only the right edge, the left eye sees only the left edge. So, they could not have used stereo vision, even if they had it. In this particular case, it had to be something else.

A bee’s eye has thousands of tiny, independent photoreceptive units that can detect motion very well. Could it be that the bees were sensing the apparent speeds of the left-and right-hand edges as they flew through the hole and were balancing the two speeds to steer a course through the middle? So, if one rim appeared to be moving faster, the bee should veer away from it because it was nearer, and vice-versa.

To test this idea, we trained honey bees to fly through a narrow tunnel to reach a sugar water feeder at the far end. The walls of the tunnel were lined with vertical stripes to provide visual texture, and the bees were filmed from above as they flew through. The tunnel was only about five feet long.

The videos revealed that the bees indeed flew close to the midline of the tunnel. We hypothesized that the bees were measuring optic flow—the rate at which an image moves across the eyes—and ensuring that both eyes were experiencing equal image speeds.

We then moved one of the walls of the tunnel to see what happened. When the wall began moving in the direction of a flying bee (thanks to a conveyor belt arrangement, placed behind a transparent shield to block air currents), the bee flew in a path that was closer to the moving wall. Evidently, when the wall was moving with the bee, the eye viewing that wall experienced a lower image speed, compared to the opposite eye. The bee perceived the moving wall as being further away, and began flying closer to it to compensate. Conversely, when a wall was moving against the bee’s direction of flight, the bee veered away from it.

On a lark, we then trained bees to fly through a 20 foot-long tunnel. They continued to fly through the middle when the walls were stationary. Curiously, though, when they returned to their hive, they danced to signal that they had flown almost a thousand feet. They were hugely overestimating the distance they had traveled. We were baffled.

It turned out that, because the walls of the tunnel and the floor were very close to the bee, even a small amount of forward motion caused a large amount of optic flow. Even though they had flown a short distance, they thought they had flown a lot. Think of it like this: if you were to fly from Berkeley to Brooklyn, and you looked down at the ground beneath you, it wouldn’t move much at all. You wouldn’t think you had traveled much. But if you were to drive from Berkeley to Brooklyn, since the ground is so close, you would get a large amount of image motion and the impression that you've gone a long distance. The bees were similarly measuring optic flow and gauging the distance flown.

To test this further, we made the stripes on the walls of the tunnel appear parallel to the direction of the bee’s flight. Flying along the stripes, the bee now saw no patterned motion. When it came back, even though it had flown the same distance, it danced to signal that it had flown zero distance.

More experiments followed. We used this simple apparatus to analyze the flight speeds of bees in tunnels of various widths and with varying wall speeds, and we showed that the bees maintain the optic flow they experience at a value of about 300 degrees per second. Such findings, combined with additional evidence, revealed that bees gauge the distance flown by measuring the total amount by which the image of the environment has moved in the eyes during the journey from the hive to the food source. The honey bee’s ‘odometer’ is visually driven: the longer the journey, the more the image movement. Compared to other potential ways of estimating travel distance—such as measuring energy consumption, flight duration, or counting wingbeats—the advantage of a visual odometer is that its readings aren’t affected by headwinds, tailwinds, or flight speed.

How, then, does one explain Karl von Frisch’s finding that his lead weight-carrying bees were overestimating distances? We think that the lead was weighing the bees down, causing them to fly closer to the ground, thus increasing their optic flow and making them think they had flown a longer distance than they had.

A Biological Autopilot, But Not an Automaton

One upshot of this work is the realization that insects only see the world in 3D when they’re moving. When they are stationary, the world around them collapses into two dimensions. There is no depth information at all. An insect’s life principle might be described as, “I move, therefore I see.”  It's a very active way of perceiving the world. If you are an insect, you can't just sit in one place and see the whole world in 3D.

Seeing the world in 3D is of the utmost importance during one of the most challenging maneuvers for a flying creature: landing. How do bees orchestrate smooth landings? To explore this, we trained bees to visit a feeder placed on a horizontal surface and filmed them as they landed to collect their bounty. When we reconstructed the 3D landing trajectories, we saw several interesting features. Landing bees reduce their flight speed progressively as they approach the ground: the lower the altitude, the slower the speed. When a bee is near touchdown, its speed is negligible, ensuring a soft landing. But how does it decelerate smoothly? It turns out that the speed of flight is strictly proportional to the altitude. This critical observation reveals that a bee decelerates by adjusting its flight speed in such a way as to hold constant the optic flow (in this case the speed of the apparent motion of the ground) during the landing process. Thus, the bee slows down progressively and automatically as it approaches the ground. This elegantly simple landing strategy does not require the bee to know about or measure the instantaneous height above the ground, or even the speed of approach. The bee only needs to adjust its flight speed, from moment to moment, to ensure that the speed of the image of the ground moving past its eyes is held constant. A wonderful biological autopilot!

A bee feeding at a sugar water droplet. Bees are marked with color codes for individual identification. The proboscis extends into the droplet, acting as a straw to suck the solution. You may also notice one of the feet touching one of the droplets. Yes, bees taste with their feet!Photo: Torill Kornfeldt, Marie Dacke.

Keeping the optical flow constant while flying implies that bees automatically fly faster in an open environment, where objects are far away, than in a cluttered environment, where objects are in close proximity. That is, they fly at a speed that is safe and appropriate for their immediate environment.

Bees, of course, do much more than estimate distances and regulate flight speeds. They also figure out their direction of flight by using the patterns of polarization of sunlight—which humans cannot see—as a compass (something Rudiger Wehner studied, both in bees and in African desert ants). Bees can learn to associate food with colors, shapes of objects, and patterns (even abstract properties of patterns, such as their orientation). They can be trained to detect camouflaged objects and to navigate mazes and labyrinths. Just as Marcel Proust vividly described the memory evoked by the taste of a madeleine in Remembrance of Things Past, bees can be triggered to fly to an exact location just by the hint of a smell which they associate with that place.

Bees use the same waggle dance to signal the site for a new home. Scout bees seem to know the size of their colony and can gauge whether it’ll fit in the hollow of some new tree. Various scouts do the waggle to report the locations of prospective sites. Other bees fly out to inspect those locations, and those that like what they see come back and dance to signal approval. A democratic process unfolds, as more and more bees dance and ‘vote’ in favor of the most attractive location. Eventually, the entire colony flies out in a swarm to the new home. This is a striking example of ‘colony intelligence.’ Apparently, honey bees invented democracy millions of years ago, well before the Greeks.

Are bees doing all this as preprogrammed automatons, with no higher cognition or consciousness? I think not. A colleague of mine found that when a bee has visited a flower and is wounded, say, by a lurking spider, and when it comes back to its hive and finds another bee dancing to signal a food source in the exact direction and at the exact distance where danger lurks, the spider-scarred bee will head-butt the dancing bee, to stop it from advertising the dangerous food source. The head-butting is very specific: bees do it only when a dance points to a potentially hazardous location. It is hard to imagine behavior like this being purely reflexive. Bees may not be as conscious as humans, but there's a great deal more there than meets the eye.

It’s humbling that the remarkable faculties of bees—their individual and collective intelligence—emanate from a brain no larger than a sesame seed. Our disregard for ‘lowly’ creatures such as insects reflects our own ignorance, rather than theirs. Karl von Frisch said, “The bee's life is like a magic well: the more you draw from it, the more it fills with water.” I am lucky to have been privy to some of this magic. ♦