[BELL TOLLING] HOWARD C. HOWLAND: --in '78. After teaching at Yale for six years, he worked his way home to Ithaca, Cornell in 1986, where he has been ever since. Like his father, he was a department chair, my chairman in neurobiology and behavior, from 2005 to 2008.
In recognition of his scientific work, he has received the Senior Scientist Prize of the Alexander von Humboldt Foundation, been awarded a Guggenheim Fellowship, and been elected a fellow of the American Academy of Arts and Sciences. His research focuses on the internal organization of honeybee colonies and has been summarized in three books, Honeybee Ecology, 1985, Princeton University Press; The Wisdom of the Hive, 1995, Harvard University Press; and Honeybee Democracy, 2010, Princeton University Press. And you've see copies of the book out there in the atrium.
So he's going to talk to us today on honeybee democracy. Tom, it's a great pleasure to welcome you to our podium.
TOM SEELEY: Well, thank you very much folks for coming out today. I hope you'll enjoy this talk. It's about honeybee democracy. One of the reasons for me it's a pleasure to be here is that this is a story of scientific discovery that was conducted here at Cornell. As Howie said, I started out working with Professor Roger Morse in the entomology department. And the story stretches back to the early 1970s, out at Dyce Laboratory. It then moved on to being involved with Jack Kingsbury at the Shoals Marine Laboratory. And we'll see some slides of that as well today.
So it's a pleasure to share this story with you. And I know many of you probably have served in administrative roles, been department chairs, and so forth. I'll mentioned at the outset that a number of the things that emerged from this story about how bees work together to make a decision, in their case it's a life or death decision, a lot of their methods I applied to our own department, our faculty meetings. And I think the reputation, of course, is that a department faculty meeting can be like a hornet's nest. And I hope Howie that I made ours a little bit more like a beehive. So I'll share with you how that worked towards the end of the talk, some of the things that I learned.
You might be wondering why democracy, why this term democracy? Well, we all know that democracy is a form of government where the power is vested in the people, really in the people, not in some king or other leader. But with a bee colony, everybody knows that a honeybee colony has a queen in it. So why isn't it a honeybee monarchy?
Well, there is a queen. But she's not a ruler of the colony, of the bee colony. She's just the mother of it. She lays the eggs. She's the genetic heart of the colony. But she's not a royal decider.
She doesn't know, for example, when the colony should produce more comb, or whether it should make more brood, or where the foragers should be deployed each day. These are all decisions that are made by the workers. And they make these decisions collectively. And what we're going to be looking at is one of these democratic decisions made by the bees about where their home will be.
We all know that democracy is not perfect. There's the famous Winston Churchill quote here. His quote expresses the idea that it's the best of bad options. And maybe that's true. But I think one of the things I've learned from the bees is that if it's done in the right way, in the right circumstances, democracy is not just making the best of a bad situation. It can be making the best of a good situation.
And that's not just my idea. There was a gentleman, James Surowiecki. That name may be familiar to some of you. He writes the financial page in The New Yorker. He wrote this nice book a few years ago called The Wisdom of Crowds.
And his main point was that with the right organization, democratic groups are remarkably intelligent, often smarter than the smartest individual in them. Why is that? It's because a democratic group with the right organization can take advantage of the knowledge of the members and can winnow through the possibilities and identify the best one. And that's exactly what the bees have to do when they're choosing a home.
So we're going to be looking at this case of democratic decision-making by scout bees in a swarm. And as I mentioned, at the end of the talk, I'll share with you some of what I like to call swarm smarts, things that I've learned from the bees and have used.
Now, where in the bee biology are we going? Well, we're going to be looking at this thing over here on the left. It's a swarm of bees. It contains one queen bee and about 10,000 worker bees.
It's a reproductive propagule of the colony. In the spring or early summer, a colony outgrows its hive. It's strong enough that it can fission and to set out to create a new colony. So the old queen leaves with about 10,000 worker bees.
They fly out of the live in a swarm. It only takes a few minutes. But they don't go straight to their new home. In fact, they haven't chosen their new home. Instead, they form this bivouac, this cluster on a tree branch. And here they will make the choice of where their new home is going to be. And they'll be hanging here for a day or two.
So clearly they're homeless. How do they go about finding a home? Well, they do as maybe at a university you might expect them to do. They form a committee. They form a search committee. And it's a pretty big search committee.
It's a couple hundred bees, scout bees. And these scouts are among the oldest, most experienced bees in the swarm. These are the bees that have had the most experience flying out from the hive when they were foragers. They flew out of the hive, flew out for miles, out to patches of flowers. And have a lot experience finding their way back home.
They're going to have to do the same thing here. They're going to be flying out for miles, looking for tree cavities, the right hollows in the right trees, and then coming back to the swarm. So that's what they're going to do.
So they start out, these scouts, these eldest, most-experienced bees. They fly out, fan out. They will search independently. So they're going to fan out across the countryside. And then they're going to be searching for potential homes.
What are they looking for? Well, this slide shows what-- I like it because it shows a dream home for honeybees. This was a sugar maple that was up in the hills of Caroline before I studied it. Bees were living in a hollow in the tree here, accessed through this knothole.
And what made it a dream home was several properties. The most important ones I've listed here. The entrance was high off the ground, which makes it safe from predators, like mice and raccoons, and these days, bears. They're just not easily found by predators this high.
They like the entrance to be small, only a few square inches in area. This is very important again, keeping predators out. It makes an easily defended location of the nest. And in the wintertime, when the bees are keeping warm inside their nest, it means there's not a lot of air, cold air blowing into the nest.
Another feature of this tree cavity-- and this log here is this section of the tree split open, exposing the nest. It's quite spacious. This has about 10 gallons of space or 40 liters. And the bees need that. They need that much space because honeybees survive the winter, not by letting their bodies cool down, like all other insects do. They fight the cold. They form a cluster inside their nest.
And it's a well-insulated cluster, about the size of a soccer ball. And they will be producing heat in that ball all winter long. It's like there's a little 40-watt light bulb burning inside the nest all winter long. And because the cluster is well insulated, they can survive.
In fact, they keep the temperature about 95 degrees Fahrenheit in the core of that cluster. It's amazing. But it's expensive. And they burn through about 40 pounds of honey. That's their winter heating fuel. So they need a lot of space inside that cavity, nest cavity, for them to survive.
And the last thing I've listed here they like these nests to be in a sturdy, live tree. So the thing doesn't blow down.
Now when you think about it, this is probably not an easy combination of traits to find. You've got to find a live tree, that has a big, rotted space inside it. That large space has access just through a little opening. And to top it all off, that opening is high off the ground. And there are not a lot of cavities to fulfill those requirements. And that's probably why a swarm allocates several hundred bees, scout bees, going out, searching for possibilities.
So we've got these several hundred bees that have gone out. And if you watch in the woods in the spring, you might see a honeybee, a scout bee, going up and down trees, looking for crevasses, dark openings, things like that. If a bee finds one, what does it do?
Well, it first makes a private evaluation of the site. It'll spend about 45 minutes at the site, making visit after visit inside the tree cavity. This is a diagram that shows the movements of a scout bee that discovered one of my experimental nest boxes. And I tracked her movements.
And I'll just show a representation of movements for four of the 25 visits she made. You can see that initially, she came in, walked around, and then quickly came back to the entrance. The next visit, she walked more deeply into the cavity, or visit eight that is. Visit 17, by that 17th visit, she's moving all around inside the cavity. Her last visit, she's roaming throughout the cavity quite extensively.
And in ways that aren't entirely clear, she's acquiring information about the properties of that site. She's measuring its volume, the size of the entrance, the distance of the entrance off the ground, and so forth. The one thing we do know is that if you make a bee walk, if you make an experimental nest cavity where the bee has to-- it basically puts bee on a treadmill, a rotating cylinder. And you make her walk more or you make her walk less, she has the impression that cavity is either large, very large, or very small. So we know there's something about that walking that is involved in measuring the volume of the cavity.
So she measures all of those things, over this approximately 45-minute period. And she's been working entirely on her own because only one bee finds a site at a time. And then she comes back to the swarm. And she will now make a public report at the swarm of what she's found.
And this is a remarkable thing. These are little insects. And yet on the side of the swarm, on the vertical surface of the swarm, the bee will do a dance. I'll show you a movie of it momentarily. And in doing this dance, she shares with the other scouts, the ones that did not happen to find a possible home site, information about the direction, and the distance, and the goodness of what she found.
Let's look at what this dance looks like. This is what the surface of a swarm looks like. You're going to see that most of the bees in the image here are sitting quietly. Those are all the non-scouts. They're sitting quietly, conserving the honey that-- all the bees, when they left their hive, they left with a stomach full of honey, stuffed with honey.
They're conserving that honey supply. Only the scout bees will be running around on the surface. So we'll see one bee dancing, advertising the home site. And we'll see other scout bees tripping along behind her, getting the information.
TOM SEELEY: OK, here's our scout, waggly, waggly. Note, she's quite consistent in the direction her body points. She's pointing to the left every time she does that waggling. And she's very consistent in how long each one of those waggle phases lasts. It's about two seconds, 1,001, 1,002; 1,001, 1,002. And you see the other bees tagging along behind her.
So that is a scout bee that found something. And she's sharing that news with her fellow scouts. Now, how does the information expressed in the dance?
Well, this was worked out by a German scientist, Karl von Frisch, in 1946, working at the University of Munich, right at the end of World War II. What he discovered-- and it was a discovery that really shook up our view of animal behavior, that a creature as simple a bee could do this.
But he was studying it in the context of bees advertising food sources, patches of flowers. And what he discovered is that to indicate the direction to the flowers, the direction to fly from a home out to the flowers, the bee takes note of the angle to the flight to the flowers relative to the direction of the Sun, or technically the Sun's azimuth. In this case, the flowers are 40 degrees to the right of the Sun's azimuth, the Sun's direction.
That means that when she gets back to the nest and performs a dance on the vertical comb, she copies that angle. Where she does her waggles, back and forth, back and forth, moving forward, she does that at the same angle, 40 degrees to the right of straight up on the comb. So outside the hive, the Sun is the reference to direction. Inside the hive, straight up is the reference direction.
And she codes the distance. It's indicated by the duration of each waggle. And so the greater the distance to the flowers, the food source, the longer that each waggle run lasts. They can last from a fraction of a second to several seconds.
And so the other bees that are following the dancer, they can measure that angle. They can measure the duration of the waggle runs. And they can fly out and find these food sources.
Now, von Frisch always studied it in the context of foraging. And it was one of his students, Martin Lindauer, who really made the discovery that bees are quite versatile in the use of this waggle dance. They use it not just to advertise food sources. They also use it to advertise home sites, nest sites.
And Lindauer made this discovery in 1949. He was studying with Karl von Frisch. He came out of the Zoological Institute at Munich one afternoon in the spring. And he saw a swarm of bees, like this one here, hanging in a bush near the front door of the institute. They had hives of bees for their bee work at the institute. And one of them swarmed.
Lindauer looked at the swarm. And he noticed bees dancing on the surface of the swarm. And he at first thought, oh, those must be bees that have gone out, found food. And they're bringing still more food to the swarm. So everybody would be even thoroughly loaded with food.
But Lindauer was a very good observer. And he noticed a couple of things that didn't seem quite right. One is that he noticed that the bees doing the dances never unloaded any food to other bees. He also noticed that a lot of the bees were kind of dirty. Some of the bees looked like they had red brick dust on them or dirt. Some even looked like they had soot on them.
And he told me that-- you can pluck a bee, grab a bee its wings. He told me he plucked one of these sooty looking bees off the surface of the swarm, one of these dancing, sooty bees. And he sniffed it.
And he said it smelled like a chimney sweep. And that was this very good indication to him that this bee was definitely not coming back from a patch of flowers, but had been poking around in a chimney somewheres. Now, what's going on? Why a chimney?
Well, you have to remember, 1949, Munich. Munich was, and still is, an industrial city. It was heavily bombed during the Second World War from bombers coming up from Italy. And at the time, there was a great deal of rubble. Lindauer said that as a student, one of the jobs they had to do was just spend an hour a day cleaning rubble in the city.
What these bees were doing, these house-hunting scout bees were doing, is they were going out in the city. They were finding cracks in walls, broken chimneys, and so forth, had discovered those as potential home sites, and were coming back and reporting on them. But Lindauer didn't leave it just there. He said, OK, I want to learn more about this. I want to really be sure that these bees are nest site scouts.
So what did he do? He realized that he knew how to read a dance of a scout on a swarm. He assumed that it was a scout bee. And by timing the dance, and measuring the angle of the dance, and so forth, he could determine for each dancing scout bee the location in the city that she was advertising. And he did this. And he followed more swarms. He followed the dancers on swarms. And he made some very interesting observations.
He discovered that initially the bees performing the dances on a swarm, the different bees will be advertising different sites, multiple sites around the city. But before the swarm flies away, to wherever it flies away to its new home, all of the dances are indicating just one location. And he was able to determine in addition that the swarm flies to this dance consensus site and moves in.
Now, how did he know that? Well, when a swarm of bees takes off, it's a ball of bees about this big. It'll launch into flight. It'll explode into flight once the decision has been made. It only takes about 60 seconds. But then they stay fairly tight.
A cloud of a swarm would fill about maybe a third of this open area in the front of the auditorium here. And it doesn't fly very high. The bottom of this cloud of bees is right above my head height, so about two or three meters up. And it only flies about a maximum of about 6 miles per hour.
And so Lindauer could run along beneath the swarm when it was flying away. And he could track some of them to their new home. For example, here's a map of downtown Munich, where the main train station is. The Zoological Institute was one block away.
Here's where the swarms were taking off. And the first one, he tracked one, two, three, four, 4 and 1/2 blocks to the north and found that the swarm-- which was the site that the dances were consistent for right before that swarm took off. And the swarm moved into a broken wall there.
And so he could conclude from that that these dances on the swarm really do indicate possible home sites. And he had the very good insight-- he didn't quite say it in terms of democracy. He said it was a plebiscite. The scout bees conduct a plebiscite to choose their new home.
But that's where the story ended. And he finished his work in 1954, published it in his Habilitationsschrift in 1955. And that's where the story lay quiet until I picked it up in the early '70s. Didn't make great progress on it then. Picked it up in the late '90s and took it forward from there.
Now, why did this story sit quietly for many years? Everybody wanted to know how does democracy work in a honeybee swarm? Clearly, there was something interesting going on here. These bees go out, find things, come back, report on them. It looks like they have a kind of debate, come to an agreement. How does that all work?
And the reason Lindauer couldn't investigate it more thoroughly is he explained to me the tools he had at the time were very simple. This was in the late '40s, early '50s. The German economy was just starting to rebuild itself. Here are the tools he had.
He had a chair. He had a notebook. He had a pencil. He had a little paint set for labeling some of the bees.
Did I say notebook? Yeah. He had a notebook. He had a wristwatch. And he had a stopwatch. That's what he had.
So he could only watch one bee at a time. He could only read one dance at a time. He couldn't get a synoptic view of the whole process. That had to wait until the advent of video recording technology. And even the video recording technology wasn't really good enough until the late '80s and early '90s.
So I picked up on this project with video technology. And I started my work where Lindauer left off. And that is to say, I started by making a detailed eavesdropping on the scout bees' activities on a swarm.
And when I say detailed, I mean detailed. I really wanted to know everything that is going on by each individual. And I wanted to know at the level of what different individuals were doing.
To do this, I had to make swarms, where bee in the swarm is labeled for individual identification. This is a matter of gluing a little-- it's like a little license plate, a little tag, on the back of the thorax of the bee. And then putting a colored paint mark on the abdomen.
I've got 500 different tags, color and number combinations, m 500 combinations of tags. I've got lots of different paint colors for the abdomen. So I could make swarms with 4,000 bees labeled for individual identification.
And a lot of times people ask me how long does it actually take to label 4,000 bees? And I have to say, it only takes about a day if you get three or four hardworking Cornell undergraduates to help you because then you can take an assembly line, where everybody does their part of the process. And I tell the students this is going to be thoughtful work. I don't explain to them that that means they're going to have a lot of time to think while they're doing this work.
So make a swarm with all 4,000 bees labeled in it. And, of course, it has a queen bee in it as well. And then we can take a swarm, one of these swarms. And the worker bees, the labeled worker bees, will cluster wherever the queen bee is. And so you always put the queen bee in a little cage.
So we put the queen bee on this swarm stand here, with some feeder bottles to keep bees well fed. The workers will cluster on this flat board. And so we've got a nice flat surface on which the scout bees will do their dances. The scouts always do their dances on the outside surface of the swarm. That's very convenient.
And then with a video camera you can record every dance. You know the time. You can transcribe from the video recordings the dance angle, the dance duration, and so forth. And you can record who did the dance.
And what I want to share with you now-- this is the most important slide in the talk-- is a record of one of these debates by scout bees. This unfolded over three days, 16 hours, which is typical for choosing a home. They take their time. In this particular debate, 11 possible home sites were considered and 149 scout bees were involved in the discussion.
And I want to remind us at the outset, this is a process. We're going to look at a process. This is being done by insects, not people. This diagram might look like something people would generate. But it's produced by insects.
Well, let's go through this. In this diagram, there are these eight panels. Each panel records the pattern of dancing done during a two-hour time period. So this first one is the first day. It's the 20th of July, from 11:00 to 1:00 PM.
This pattern of dancing was observed. Now, the circle in the center represents the swarm. Each arrow represents a record of bees performing dances for a certain location, for a certain site.
This pink arrow indicates dancing was done for a site to the east, as indicated by the arrow's direction. The length of the arrow indicates the distance to the site. So it was about two kilometers or about 1 and 1/4 miles away. And the width of the arrow indicates how many different scout bees during the two-hour time period performed a dance advocating that site.
So here's our scale of number of bees down here. About 10 bees during that first two hours danced for this site two kilometers to the east. And, in addition, there were other sites, one to the south, southeast, another southeast site, southwest site, north site. So six different sites, where it had been discovered by scout bees and were reported by them during the first two-hour time period.
Clearly, no agreement at that point. Second two-hour time period, 1:00 to 3:00 in the afternoon. Again, various sites were being advertised by dancing scout bees, all different directions, different distances.
Third, two-hour time period, 3:00 to 5:00 PM, getting to the end of the day, end of the afternoon. Various sites were being advertised still.
Here's the last record of dancing that first day, 5:00 to 7:00 PM. A variety of sites are being advertised. But just two of them are getting a lot of attention, the site to the south and the site to the southwest.
And that's where the dancing ended that day. Clearly, they hadn't reached an agreement. But it looked like those two sites were starting to gain a lead in the debate.
The next morning, we arrived. We got our equipment set up early so we could catch the first dancing. The bees only started dancing around a little after 7:00. So we got all the dances. And again, they picked up where they left off the previous day, with the site to the south and the southwest being the most popular.
And that process continued during the morning. And the sight of the southwest, as you can see, started to gain a lead over the site to the south. Got a strong lead by the end of the morning. Almost all of the bees were dancing for the southwest site. But then it started raining at the very end of the morning. It rained all afternoon. And the process ended at 11:54 for that day.
That night, it continued to rain. But in the morning, it was no longer raining. And by 9:00 o'clock, the bees were dancing again. And as you can see, now every scout bee that did a dance that next morning danced for the site to the southwest. And at the end of the morning, at 11:58, the swarm took off and flew to the direction to the southwest, to that consensus site.
Now, the reason I emphasize this slide is it shows the overall pattern. You can see that these scouts have gone out. They've found various options. They've put the options on the table. And there were 11 different options. And then somehow they went around all of the options except for one. Only one is being advertised at the end.
And it's a remarkable process. It still amazes me that insects can conduct such a well-organized discussion and come to an agreement. But what's even more remarkable is the agreement, the consensus site, is the best site. It's the best site.
So they're not just making a choice. They're not just building an agreement. But they're making a good choice. Now, how do I know it's a good decision?
Well, I know it because I was able to take advantage of the Shoals Marine Laboratory's very special location on the island, on Appledore Island off the coast of Maine. How many of you have been to this island? Yeah. Oh, wow. OK, great.
So you know that this is an island that's north of Boston. It's about six miles out. So it's far enough from the mainland that if I took a swarm of bees out of the island, they can't fly back to the mainland to get back to the trees back to the mainland.
And the future of this island is that though it has strong winds, which are not good for bee work, it has poison ivy, which is not good for bee researchers, and likewise for the seagulls, it has no trees on the island, or no large trees, no large trees with large cavities in them. So that means I can bring a swarm of bees out to the island. I could put out nest boxes on the island. And the poor bees then were stuck. They had to choose among the options I gave them.
And I was basically giving them a multiple choice test. I would take a swarm. Here's one of the swarms mounted on the swarm mount. And it's mounted on the porch of the old Coast Guard building on the island. These orange boards are windbreaks.
Put the swarm at the center of the island. And then I'm going to put out five nest boxes. So it really is a multiple choice problem, five nest boxes. And only one is the right answer. One of them is going to be an excellent home site and the other four are going to be wrong, going to be not great home sites.
They're are going to be OK. But they're not as good as the really good ones. I can do this experiment over and over. And the question is, could swarms consistently choose the good site, the correct answer?
And you might want to know, well, how do you vary the quality of the nest boxes? Well, it goes back to those properties I talked about earlier that matter to the bees, volume, entrance, size, and so forth. For this experiment, I had everything right in the next boxes, everything the bees liked, except they were low to the ground.
And some of the nest boxes were set at 15 liters of space. The bees like 40 liters. If I take out this inner wall, they get all of the 40-liter space in the box. But if I put a wall in at this point, they only have 15 liters of nesting space. And that makes it less than ideal.
So the choice is going to be four boxes are going to be set at 15 liters. One's going to be set at 40 liters. And the question is, can the bees choose the good box? And here's what these boxes actually look like, set up. Each one is in a little hut to protect it from the wind, and the Sun, and so forth.
And I won't go through this in detail. But this is an example of what we would find that the bees would do. Here's the first trial. There's five boxes, one, two, three, four, five. One of them, way at the end of the array, was given the excellent, was the 40-meter box. The other four boxes were set at 15 liters.
And here we have time of day on the horizontal. The number of scout bees visible at the nest box, just buzzing around the entrance at each box, is on the vertical axis. And you see that they actually discovered the four mediocre sites a little after 8:00 o'clock in the morning. They discovered all four of them.
And they started building up interest at these boxes. Scout bees were checking them out. They didn't discover the excellent box still a little after noon. But once they discovered the excellent box, the number of scout bees at that box built up strongly. And by the end of the afternoon, the bees had chosen that box and took off and tried to move to that box.
In these experiments, I don't let the swarm actually move to the new home. I keep the queen caged. So they take off. They try to move to the home. They discover they don't have their queen. They come back to their swarm mount. And that makes it easy for me to take down the swarm and do another trial.
So this swarm made the right decision. It chose the right box. The next swarm did the same sort of thing. It found two of the mediocre boxes and got interested in those two. And then finally it found the excellent box. And then that one prevailed. The strongest number of scout bees built up at that box. And they took off and tried to move to it as well.
And we've done this experiment. It's over 10 times now. And we found they made a mistake just once. So they're about 95% accurate in their decision-making. And the one time they messed up was a time when the box, the excellent box, was discovered just so late in the process. They had such a strong build up of interest in one of the mediocre boxes, they couldn't quite recover from that problem.
Now, I want to share with you a little bit of a feel for what it's like to do these experiments. What I've shown you so far, you might get the sense that it's like clockwork. You set up a swarm on the island. You put out your boxes. The scouts go out. They find the boxes and they choose.
Well, it actually didn't initially at least, didn't work quite that smoothly. This is a shot back at the beginning of some of this experimental work. I had my nest box set up. I set it up by this little bit of ledge. So I had a nice little bench there. I could sit and wait for the scout bees.
I set up my swarm. I set up my nest boxes. I went down to the nest boxes, to wait for scout bees to arrived. And I waited and I waited for an hour or more, a couple of hours actually. And no scout bees came.
And I was going from box to box, thinking, well, they'll find this one or find this one. Finally, I decided something's not going right here. And so I went back to the swarm because I knew that if the scouts had found something, they would be doing dances for it on the swarm. So I'm wondering where are these bees?
I go back to the swarm. I immediately see the scout bees are the dancing excitedly. But I quickly read the dances. The dances are not pointing towards my array. They're pointing over here, to this particular building.
And I have to tell you folks this was not good news. Why was this bad news? It was because when I had come to the island a few days before-- this when I was starting the work-- Jack Kingsbury told me, Tom, you can work all over this island, wherever you want, except you can't go over here.
These are private dwellings. And we want to respect their privacy. But most importantly, don't go to that house. And what was that house? Well, that was the house of Rodney Sullivan. He was a reclusive lobster fisherman, who very much valued his privacy and enforced it with a loaded shotgun behind his door.
And he was known to shoot it off over the heads of students if they got too close to his house. Well, where were my scout bees going? They were going into his chimney.
I didn't know that at first. All I knew was the bees were going into Rodney Sullivan's house or someplace near it.
So I go to Jack, Jack Kingsbury, and say, Jack, what are we going to do? And Jack says, well, we can't just walk over to his house because he'll think we're sneaking up on him if come up from behind. We got to go by boat.
So we get a boat, go over. Call up to Rodney's house, hoping he's home. And we call up and say, hey, Rodney, you home? And he says, yeah, I'm home. Can we come up? Yeah. Come on up.
I've got a problem. I got bees.
And he was quite terrified by these bees. He had never seen anything like this before because at this point the site was very popular among the scouts. There were probably a hundred scout bees buzzing around the top of his chimney, flying around his house. And he had wondered whether they'd been blown out by a storm and whatnot.
And I don't think I quite revealed to him that I had brought the bees out. But I did say, oh, just by chance, I can help you with this. The solution was simple.
He had a ladder. He had some old screen. He had some duct tape. So I built a fire in his stove, which is what the chimney was for. And made a fire, smoked the scout bees out. Taped a screen over the top of the chimney. And that prevented the scouts from getting back in.
So at this point, the alternatives to the nest boxes were pretty well eliminated. And the scout bees then went on to do the experiment that I described.
So we've seen that a swarm as a whole is able to bring in information about possible home sites, work through a decision. And then eventually, once they've reached agreement, they will move on to their new home. But how do they actually do this? What actually happens among the scout bees at the individual level?
Well, there's two phenomenon we really would like to understand here. One is we've seen the bees interest grows and grows for the best site. This is the winning, the best site.
It started out on day one. It only had 13% of the dancing here. It garnered 33% of the dance activity here, at the end of the first day. The middle of the second day, it had risen to 62% of the attention. And, of course, the final day it had 100% of the attention of the scout bees. How did that happen? Why did it grow and grow?
And the other thing is somehow the scout bees interest faded for all the poorer sites. How did that happen? All these bees, they were so excited about these other sites, especially these bees advertising the site with the blue arrow to the south. Somehow, by the end, they stop their dancing. How did that happen?
And I think we all know, when we have discussions, one of the more difficult things is getting somebody to give up on a losing idea. But with the scout bees, it works. The interest faded for the poorer sites. How did that happen?
Well, let's start with the first question. Well, one way to think of this-- and it is an accurate way to do it-- is to think that the scout bees' decision-making process is like a democratic election. That is to say, you've got different sites, different candidate sites.
You've got supporters for each of these sites, the scouts that are supporting site A, for example, or the scouts supporting site B. And they produce ads, these waggle dances, to try to recruit uncommitted individual, uncommitted scout bees from this pool here of being uncommitted, over to being a supporter for site A or site B.
And, of course, apathy can set in to scout bees, as we'll see. And so they can drop out of being a supporter for one of the sites and go back into the pool of uncommitted scouts. So you've got this competition among the supporters for different candidates, different candidates sites, for individuals that are not yet committed. So that's like a democratic election. So it's a competition.
Now, one feature of honeybee elections which is maybe a little curious or different is it turns there's completely honest advertising of the candidates. That's a key feature of why they make good decisions. What do I mean by honest advertising?
I mean that site A, for example, will give strong advertisements, that is to say strong dancing by the scout bees for that site. It will only get strong ads if it's a really good site. And, likewise, if this site isn't as good, it will get only weak ads, weaker ads. So the bees are honestly adjusting the strength of their advocacy of a site based on the goodness of the site. And I think you can see that that's going to help the best site win in this election process.
Now, how do I know and what exactly do I mean by honest advertising? How do we know that they honestly advertise their sites? Well, to do this, I went back to Appledore. We put up a swarm or two, and one swarm at a time.
Put up two nest boxes, now just two nest boxes. One's going to be excellent. One's going to be mediocre. And in this situation, I'm going to have scout bees coming back from two sites, two sites, back to the same swarm, at the same time, under the same weather conditions. Everything's the same. Except that one bunch of scouts is coming back from an excellent site. One is coming back from the mediocre site.
And I'm going to watch closely at how those bees dance differently. And in particular, what I wanted to do is I needed to count how many of these dance circuits a bee does depending on the goodness of her site. A circuit consists of a waggle phase. And then the bee stops wagging and comes back around. It does another waggle, comes around the other way.
Each one of these things at the waggle phase, combined with the return, is what we call one dance circuit. And the question is, do bees coming back from better sites do more of these dance circuits than bees coming back from mediocre sites? And why were we interested in that? It's because we knew from studies of foraging that if a bee does more dance circuits, she creates more opportunities for bees to get her message. And so she gets more recruits. So that's why we were looking at the number of dance circuits performed.
And what does it actually look like? Well, this is the nest boxes. So we'd have one helper stationed at a nest box. That was the fixer-upper site, the 15-liter. Over here is the dream home. There's Another person there.
In this particular experiment, we didn't have to label all of the scout bees in the swarm. We just had to label the individuals that were scouting our two nest boxes. And the way we did that is when a scout bee would come to a nest box, discover a nest box, we'd just put an insect net over the entrance of the nest box. So when the scout bee came out after her first visit, we could catch her in the folds of the insect net and put a paint mark on her.
And remarkably, the bees are not disturbed by this handling. We handle them gently. But even so, we thought they'd be disturbed. But no, you let the bee out of the insect net. What does she do? She takes off and just goes back in the box.
It's kind of abduction by aliens. And as we all know, when you're abducted by aliens, you have no disturbance, no memory even of it.
So this honest advertising, this is what I mean. The bees, if they come back from a dream home, these are the distribution, numbers of bees doing different numbers of dance circuits. And if they come back from the dream home, most of the bees will do more than a hundred. Many of the bees will do more than a hundred dance circuits to advertise that site. And on average, they do about 90. If they come back from the fixer-upper, of them do less than a hundred. And the average is only about 30.
So the higher a bee judges a site's quality, the longer she dances. And this biases things in favor of the better site. And curiously, the scouts don't compare sites. Each scout just goes to one site, checks it out, and then comes back and reports on it, which tells us that these bees know innately what makes a good home.
It's like they go to a site, check it's volume, entrance size, entrance height, and so forth. And then give it a ranking, say on a scale of 1 to 10. And then they adjust their dancing accordingly to that ranking. To me, as I say, it's stunning that this is all being done by a little insect.
So the best candidate becomes the most popular through this honest and accurate campaigning. And I show it schematically here. Let's say it's the simpler situation of choosing between two sites. The one on the right here is the better side by virtue of a smaller entrance opening. Let's say both sites are discovered exactly the same time. So you have a scout bee coming back from this site and this site, both exactly at 10:00 AM in the morning.
But the bee coming back from this site, represented by this blue dancer here, let's say she does 90 dance circuits, whereas the bee from the poor site only does 30 dance circuits to advertise her site. Well, if you came back three hours later, you're going to have three times as many bees at this site now because the dance was three times as strong. So you got more bees at the better site and you have more dancers at the better site.
And now, the process is going to snowball in favor of the better site because each of these blue bees is still dancing three times as strong as the red bee. Moreover, now there are three times as many blue bees as the red bees. So it's a case of kind of the rich getting richer, or positive feedback, which carries it. If you came back three hours later, almost everybody is dancing for this better site to the right. And there's a lot of bees at this site. So this is how they can build an agreement. At least, it's part of the story.
So we've talked about it as an election. How does a site, site A or site B, how does a site actually win an election in this contest? Well, one possibility-- we thought this for many years-- was that the consensus choice, the site that becomes the consensus of the dancing bees on the swarm, was the winning site. But one thing that made us wonder about that is for that to work, it would imply that the scout bees are walking around on the surface of the swarm, polling the dancers to know whether a site has become the agreed-upon site or not. And we never actually saw scout bees polling the dancers. So that cast doubt on that idea.
The other possibility was that maybe what really matters is a site wins if it gains enough popularity in the sense that enough bees build up at the site, a quorum of bees, of scout bees, builds up at the site. I should explain that. A scout bee finds a site, checks it out, comes back and dances on the swarm. And then she goes back to the site. In a sense, you could think of the bee voting for a site both by doing a dance for the site, but also by going to the site and spending time at the site.
And it turns out-- well, just to exemplify what these alternatives are, the question is, are the bees paying attention to this build-up of consensus at the swarm or are they paying attention to the build-up of bees at the winning site? And it turns out the answer is, it's this. Reaching a quorum is how the scouts know that one site has won the contest. And we'll go through the experimental details of that. But that's a curious thing.
They use a quorum. And so you can think of it as a bee voting in favor of her site by going back to the site and spending time there. And the more dances there are for the site, the more bees that we'll be spending time at the site. So that's a curious thing. They use a quorum system. It's like people voting, not by raising their hands, but by saying I vote for that option by going and standing on that side of the room or this side of the room.
We probably should have known this way back when we were doing the multiple choice experiments because we did see that every time the number of scout bees visible outside the nest box built up to about 15, the swarm would then make preparations to take off and move to that site, here, here, and the other. So we don't know how the bees sense a quorum. But we do know that they're paying attention to that build-up at the site.
There is this second question. I won't go through it in great detail. Bees' interest fades for all the poorer sites. How does that happen?
Well, in a nutshell, what happens is that bees have a built in sense-- there's two parts to the answer to that. One is the bees have a built-in sense of knowing that they shouldn't persistently dance for a site. In fact, they shouldn't even persistently keep visiting a site. So a scout bee typically either discovers a site or gets recruited to a site. She'll check it out.
She'll come back to the swarm. She'll advertise it with her dance. And then she'll go back to the site. But after a while, she stops going back to the site. She stops spending time at the site. And when she comes back to the swarm, she doesn't even dance for it.
It's almost as if she said to herself, OK, I've made my contribution to the debate. I'm going to pass it on to the next generation. I've done my thing. I've made my assessment. I've made my vote. I've expressed my opinion. But now I'm going to go quiet. I'll let the next generation of scout bees take the process forward.
And that's all we thought the process was for losing interest, for many, many years. But in the summer of 2010, I asked myself, hmm, is that the full story or is there some sort of encouragement of these bees to give up on a site? Is there some kind of negative campaigning? The waggle dance is clearly a positive campaign. And it's saying come on the my site, come onto my site. But is there some negative process going on as well?
And in the summer of 2010, I discovered that there is. And it's a new signal. And it's a kind of negative campaigning. And the way it works is the following. I'll show you what it looks like.
While a bee is doing a waggle dance to advertise her site, and she's really promoting her site, every once in a while a bee from another site will come over to her and ram her with her head, headbutt against the bee, the dancing bee. And when she's headbutting against her, she's making a little sound. It's a very faint sound. That's why I didn't know about it. You actually have to hold a microphone, a small microphone, right by the dancing bee to know if she's getting hit by one of these other bees.
We call this the stop signal or the beep signal. We call it the beep signal because it's a short thing. Each hit, each headbutt, is about 150 milliseconds long. There's a little sound associated with it. The bee that's doing butting, she activates her flight muscles, her wing muscles. So she makes a sound.
And that's probably emphasizing that this isn't just an incidental bump, buster. You're actually being told to stop dancing. Any one beep doesn't shut off a dancer cold. But it's a dose of inhibition. And after a dancer gets hit with about 10 of those, she will stop dancing.
What does it look like? Well, I'm going to show you a movie here. What we're going to see is this little bee here, with the blue and yellow paint mark on her thorax, she's going to be merrily dancing along. And another bee, with a pink paint mark, from a different nest site, nest box on the island, is going to come over and hit her three times, beep her three times.
TOM SEELEY: Whoops. Hold on. I got to get the microphone down here.
Let's redo this.
Could you hear that beat at all?
TOM SEELEY: A little bit. I'll do it one more time.
TOM SEELEY: First, second, third.
So that's that signal. As I say, it's a cross-inhibition signal. And it sharpens up their decision-making. Because the winning site, it not only gets ahead by virtue of stimulation, but also stronger inhibition.
So what's this right organization, to wrap things up, that makes a democratic group remarkably intelligent? What is the right organization? Well, I'll share with you what I learned from the bees and I'll put out these caveats.
This works really well for groups in which the members have a unity of purpose. And it's where the individuals differ in their information, but not in their preferences. And that might seem very special.
These things don't work where you've got bedrock differences in the preferences of the individuals in the group. But a lot of times, humans might come into a group, and sometimes even in a faculty meeting, where they do have a unity of purpose. They want to hire the best person. They might have differences in some parts of their preferences. But deep down, they would like to hire the best person. But they might have different pieces of information about that decision.
So what are the bees doing, to do it right? Well, these are the things that I extract and try to apply. First of all, if you can create a situation where the group members share a decision-making goal, or you stress that, or you remind that among the members of the group-- for example, we do want to hire the best person or we want the best allocate these resources. Secondly, try to create a situation where the group members possess diverse knowledge. If you're assembling a committee, do as the scout bees do. Try to get individuals that have gone out, can bring in their different pieces of information, diverse information.
Set up a situation where the individuals feel free to share their knowledge. This might involve saying, OK, let's go around the room and see what people are thinking about this. I want to hear what everybody is saying. And certainly that goes on with the scout bees. They can come back if they find something. They can freely report it, where they're dancing.
Then have the group members express their judgments, express their votes independently. And this might involve a secret ballot, rather than raising hands. So nobody feels peer pressure.
And then, finally, aggregate these votes, the members' votes, in a fair way. In the case of the bees, it's one vote per bee. Each bee has the same weight. And they look at it as a quorum-sensing process. Whichever site wins, reaches the quorum, the threshold level of evidence first, is the winner. Of course, we often do majority votes and so forth.
So these are the things I've learned from the bees. And ask Howie if he thinks those actually worked in our faculty meetings? I have to say my colleagues-- I didn't tell everybody at first that I was doing this. But they were amused to find out bees were inspiring them.
And I see our current chairman sometimes says, well, when we get into a sticky situation, I'll say, well, Tom, what would the bees do? And I will share with my what I would say. Sometimes it's just a matter, well, we need to argue some more.
And I'm not the only person to claim that there's wisdom to be gained from the bees. You can see William Shakespeare grasped this back in 1599, as well.
Well, thank you very much folks for your attention. I hope you've enjoyed this story of the bees.
HOWARD C. HOWLAND: Tom, thank you so much.
TOM SEELEY: You're welcome.
HOWARD C. HOWLAND: This is just a token.
TOM SEELEY: Oh, wow. Thanks very much, Howie.
HOWARD C. HOWLAND: Thanks again. Thanks everybody for attending the meeting.
We've received your request
You will be notified by email when the transcript and captions are available. The process may take up to 5 business days. Please contact email@example.com if you have any questions about this request.
Tom Seeley, author of "Honeybee Democracy," and professor of neurobiology and behavior, reviews the history of behavioral studies of foraging honeybees and explains the process by which swarming honeybees choose a new home in his November 17, 2011 lecture to Cornell Association of Professors Emeriti (CAPE).