by Dave Goulson
At university I once had to write an essay on homing in pigeons, and became intrigued. In my later life as a university lecturer in the late 1990s, I couldn’t resist investigating the navigation abilities of bumblebees. My experiments were very simple, and followed the model used by pigeon fanciers the world over. I set up five buff-tailed bumblebee nests in boxes in my garden. At that stage I was living in Southampton, at almost the opposite end of the country from my current home in Scotland. We had there a very dilapidated, ivy-encrusted structure that might once have been described as a gazebo, in which I placed the row of nests on a bench seat. This I called my ‘bumblebee loft’. When I first opened the doors to the newly installed bumblebee nests, the bees poured out, eager to explore their new environment. The air filled with hovering, circling bees, but within a few minutes they had all disappeared off into the surrounding gardens where they happily began foraging, returning a little later with balls of pollen on their legs and full honey stomachs. Once the bees were experienced, which took some only an hour or two, their behaviour on leaving the nest became obviously different; instead of indecisively hovering about or circling the nest, they whizzed out purposefully and disappeared at high speed.
When the bees had had a week to settle down, I began my homing trials. I caught them as they left their nest to forage and glued a tiny coloured and numbered disc to each of their backs so that I could recognise them again. I then placed them individually in small cylindrical cardboard pots (still known to entomologists as pillboxes, from their original use) and drove them in my car to a random location. At the time I had a rather silly two-seater Toyota MR2 sports car and the stack of cardboard pillboxes would sit on the seat next to me as I roared through the Hampshire countryside. At a random location, predetermined by blindly sticking a pin in a map, I would stop by the roadside and release a batch of ten bees, noting down their numbers. Invariably and understandably they looked rather confused, and would usually circle about in much the same way as they had when they first left their nest in my bumblebee loft. Some would head straight to the nearest flowers for a quick drink after the hot journey in the car. Within a few minutes, all had disappeared into the distance. At this point I would jump back into the car and race for home. Once back at the loft I would sit and wait for the bees to return, eagerly looking at each incoming bee to see if it was one of those I had released.
Some I released within 1 kilometre of my house, and more often than not these bees would beat me back to the nest. Even from 2 or 3 kilometres the bees would often be back within a few minutes, while I got caught up in the usual Southampton traffic (bees don’t have to worry about such things). From further afield, it took rather longer. I released bees up to 15 kilometres from my house and sadly none of these ever returned. I like to think that they had a nice life, freed from the burden of work in the nest, and enjoyed themselves. Perhaps they found a different nest of their species and tried to lay some eggs in it (something that worker bees sometimes do). I am probably kidding myself. Whatever happened to them, I never saw them again.
The record distance over which a bee successfully returned to the nest was 10 kilometres. I was very proud of Blue 36. It took her two days to get home. Between 3 and 10 kilometres the number of bees that made it back steadily declined, and some would take up to three days to return to their nest. White 15 was released next to a nice patch of borage flowers in the grounds of Chilworth Manor, about 3 kilometres from my house. Borage produces particularly large amounts of sweet nectar, which both honeybees and bumblebees love (the flowers also make a colourful addition to salads, and the leaves are considered by some to be an effective cure for premenstrual tension, should you need any other reason to grow some). I happened to be doing some other experiments with bees on this patch of flowers, and was returning to them at regular intervals when not transporting bees around Hampshire or sitting by the loft waiting for them to return. White 15 was one of the clever bees that made her way home successfully and swiftly, but to my amazement on subsequent days she went back to the borage flowers. Every evening she was in her nest in my garden, but she would spend her days repeating the 6-kilometre round trip to Chilworth over and over again to collect nectar.
What can we learn from all this? Firstly, that bumblebees have pretty amazing navigational abilities. Scale the feat up in proportion to size, and Blue 36’s feat is equivalent to a man being taken about 1,600,000 kilometres from home and managing to find his way back again under his own power. This is more than four times the distance to the moon. White 15’s repeated journey to Chilworth to collect food is the equivalent of a man circumnavigating the globe ten times just to get to the shops – and then having to come back again – several times a day. Perhaps these comparisons are a bit silly, but it is hard not to be impressed.
So how do bumblebees find their way home and, more generally, how do they navigate? This is hard to study – after all, they are small and move fast, so it isn’t possible to watch. Some clues may be gleaned from observing how a worker bee behaves when she first leaves her nest – the behaviour is very distinctive. She usually flies out just 20 or 30 centimetres, and then turns to face back towards the nest entrance. She hovers from side to side, and sometimes flies a few small loops around the nest, not going more than 2 or 3 metres away. She is probably memorising the entrance to her nest, fixing the relative location of any obvious landmarks in her tiny brain (sticks, tussocks of grass or whatever). If she flies out and cannot find her way back in, then she can never bring food to the nest – and given that her role in life is to help her mother rear more of her siblings, this would be a disaster. After a few moments, she ventures off and is lost from sight, but she usually returns soon afterwards, and repeats this several times before disappearing off for longer periods. After a little while she then settles down into foraging, flying purposefully from the nest, and reappearing at regular intervals with a full honey stomach, pollen in her pollen baskets, or both.
My notion that the bees memorise the landmarks around the nest is not just idle conjecture. If you find a bumblebee nest in your garden, try putting a novel object nearby – anything will do, a tennis ball, a plastic bucket, a garden gnome. This will cause immediate but brief consternation. Instead of flying straight into the nest entrance, returning foragers will pause, fly small loops, and hover for a few moments, just as they do when they first leave the nest. If it is a large nest a swarm of them may quickly congregate. You can almost hear the cogs in their brains turning: ‘Hmmm, what’s going on here? I’ve not seen that before…’ They usually land somewhere near the entrance and explore the last few centimetres on foot. Once they get very close, they are able to smell that they are in the correct place and will enter the nest. On future foraging trips they will then pay no heed whatsoever to the object – that is until you remove it again, at which point they will once more rememorise their nest entrance.
This behaviour makes perfect sense. The sudden appearance of a garden gnome may not be a regular occurrence in the wild, but landmarks do occasionally change, and if bees were unable to cope with this then they would be in trouble. I was once recording the traffic from a buff-tailed bumblebee nest which was down a hole near the edge of a meadow when an inquisitive cow came to investigate. I had climbed halfway into the adjacent hedge so as not to disturb the bees, but the cow was less obliging and stood more or less on top of the nest. Within a minute or two the unconcerned cow was surrounded by confused bees, both incoming foragers with pollen on their legs, unable to locate their nest entrance, and outgoing bees busily circling to accommodate this new landmark in their memories. They didn’t sting her, and were probably entirely aware that she was an animate object – within a minute or two the bees had readjusted and got back to work. Of course the confusion briefly resurfaced when the cow ambled away, oblivious to the effect she was having.
When a bee is released from a box in a strange location, she behaves in much the same way that she did when she first left her nest. Sh
e hovers, performs small loops, and repeatedly returns to the release site. She is clearly confused, and understandably so. After a few moments, she starts making longer loops, disappearing from sight, and soon she is not seen again. What is much harder to do is to work out what happens between her disappearing from sight, and her arrival back at the nest (assuming she makes it).
There is something to be learned from the time it took my bees to return. As I said, from up to 3 kilometres some bees returned very quickly, in just a few minutes, but other bees took much longer, and most of the bees taken more than 3 kilometres away took days, or didn’t return at all. They clearly do not have the pigeon’s inbuilt GPS system. Even from 10 kilometres it would take a bee less than an hour to fly home if she knew the way. This, and the fact that many bees don’t get home at all from lesser distances, suggests that something less precise is going on.
To understand what that is, and to study bee navigation in detail, we would ideally follow their tracks. This is easy enough to do for larger creatures such as birds or mammals, by attaching a radio transmitter which allows their every move to be followed. Even with the wonders of modern technology, however, it is not quite possible to obtain a radio transmitter small enough to attach to a worker bumblebee (at least not if you seriously expect it to take off and behave in anything like a normal way). The smallest transmitters are roughly the weight of the largest bumblebees. To overcome this problem, Juliet Osborne and Joe Riley, scientists at Rothamsted Research in Harpenden, have developed a system using harmonic radar to track bees and other insects. The weight of a traditional radar transmitter is largely determined by the battery. Harmonic radar doesn’t use a battery at all, but instead relies upon attaching a transponder aerial to the bee. The aerial is very light, weighing in at about 12 milligrams, which is roughly 6 per cent of the bee’s weight. Given that bumblebees in flight can carry up to 90 per cent of their own body weight in food, the aerial shouldn’t be a major problem. That said, it is roughly 3 centimetres long (considerably longer than the bee itself), and has to be mounted vertically on the bee’s back, so it looks rather cumbersome. Halfway along the aerial is a tiny electronic gadget. My understanding of the physics is rudimentary, but essentially it works as follows: the aerial absorbs any incoming radar signal, and uses the energy in this signal to generate a new signal at a slightly different frequency. A large vehicle-mounted transmitter unit with a revolving radar dish on the roof then sends out a signal which the transponder picks up and bounces back. Inside the vehicle, a bewildering and intimidating array of electronic gadgetry controls the dish and monitors any returning radar signal. The end result is a luminous green dot on a screen – the bee – which is tracked as it flies.
This system is not perfect. The aerial has to be fitted to the bee when it leaves the nest and be taken off before the bee can get back in. The aerial is bound to make it difficult for the bee to visit flowers. Also, harmonic radar only works on direct line of sight so if the bee goes behind a hedge or tree then it disappears from the screen. Moreover, the system seems to work only up to about 1 kilometre, and it costs millions of pounds to build. Nonetheless, this is still a very cool and exciting way to watch how bees explore, and Juliet’s team were able to track the exact movements of foraging bumblebees on their experimental farm. Novice bees that had never previously left the nest had very characteristic flight patterns. To start with they flew erratically and stayed close to the nest, just as I had observed in my garden. They then started short exploratory flights, looping out 30 or 40 metres from the nest before returning. Each loop would go in a different direction, so that when the flight paths were plotted on a map they made a pattern very like that of a flower, with the nest at the centre and the loops resembling a circle of petals.
After exploring their local area, the bees would then range further afield, flying hundreds of metres, and stopping to explore patches of flowers. At this point many bees started to exceed the limits of detectability and disappeared off the radar screen’s edge altogether. The looping flights became replaced with fast, linear flights to particular patches of flowers, which would be repeated over and over again by individual bees. They had moved from exploring the landscape to foraging. The radar work showed that foragers travelled at about 25 km/h (even with the aerial) and could maintain their linear flight paths between the nest and flower patches even in strong crosswinds, which the bees handle by flying at an angle to their intended direction; most impressive. On one occasion a bee fitted with an aerial was spotted on the radar screen but seemed to be giving a very faint signal. When it arrived back at the nest it was found to have a foxglove flower impaled on the aerial; it had obviously tried to visit the flower and the tubular petals had become stuck and come away with the bee, giving it the appearance of wearing a giant purple hat.
From this work, it was clear that bumblebee workers regularly travel more than 1 kilometre to find and bring back food. The question then is just how far? The answer is crucial to conserving bumblebees, for the distance the workers can travel for food determines whether a particular nest will thrive or die, depending on where it is relative to patches of flowers. Too far away, and they are doomed. We might plant patches of flowers in the landscape to provide food for bees, but how many patches do we need, and how far apart should they be? The answers are determined in part by the distance over which bees can forage.
Many scientists have tackled this question. The most obvious approach is somehow to mark the bees in a nest, such as with a dab of paint, and then look to see where the marked bees are foraging. This has been tried a number of times in North America and Europe, but no matter how many bees they have marked, and how hard they have then looked for them on flowers, the scientists have invariably seen very few. The vast majority seem to disappear, reappearing every little while with food. The essential problem is that the landscape is big, and bees are small. Suppose the limit of foraging range is 1 kilometre, and that a nest sends out fifty workers to gather food in this area. A circle of radius 1 kilometre spans 3.1 square kilometres; this equates to an average of one marked bee per 6.2 hectares. At this kind of density, it is perhaps not surprising that these bees are hard to find.
In addition to her radar work, Juliet Osborne tried a much lower-tech approach with her colleague Andrew Martin. This involved planting a whole field with borage, thereby creating a huge food magnet, and then placing buff-tailed bumblebee nests at different distances away. Andrew (who bears a striking resemblance to Baldrick from Blackadder) came up with the cunning plan of mass-marking the bees as they left their nests by making them crawl through a box filled with fluorescent powder (using a different colour for each nest). Andrew and Juliet then watched the bees in the borage field to see which, if any, coloured powder they had on their fur. They also caught bees as they arrived back at their nests, stole the pollen in their baskets, and then analysed it to see which bees were carrying borage pollen.9 The experimental nests were up to 1.5 kilometres from the borage field, but even at this distance foragers were regularly seen on it and 17 per cent of the pollen arriving back at the nest was borage pollen.
This study was part of a larger, government-funded project in which I was also involved, attempting to improve our understanding of how bumblebees survive in arable landscapes. So while Juliet and Andrew were doing their borage experiment, my postdoc Mairi Knight and I were trying a different approach on the same Rothamsted farm. Our idea was to use DNA fingerprinting to look at the spread of sister bees. Mairi, a Scottish lass who had previously worked for a number of years on cichlid fish in Lake Malawi, is a whizz in a genetics lab – while I am absolutely hopeless. Although she knew nothing about bumblebees when she started working with me, she was perfect for this project since bee DNA is much like fish DNA once it is in a tube. The basic principle of DNA fingerprinting is very simple. Sisters are closely related, and by examining their DNA it is fairly easy to identify which worker bees are sisters. Hence we set out to catch workers of a range of co
mmon bee species along a transect across the farm. Each bee was fingerprinted, and her sisters identified. We were then able to look at how far apart sister bees were along the transect, without ever knowing where the nest was that they were coming from. For example if we found two sisters 1 kilometre apart, we could say that the foraging range for this species was at least 0.5 kilometre (if the nest was exactly halfway between the two places where we caught them), but probably more. This approach suggested that some bumblebee species seem to fly much further than others; we estimated that common carder bumblebee workers flew up to 450 metres from their nests but that buff-tailed bumblebees flew at least 750 metres (and we already knew that the buff-tails could go further still if the mood took them). If bumblebee species do differ in how far they can fly to gather food, this could mean that some species are more susceptible to a lack of flowers in the countryside than others; those that cannot fly far will need patches of flowers close to their nest if they are to survive, whereas the hardy and adventurous buff-tailed bumblebee is clearly able to cover long distances in its search for food.
So to return, finally, to my homing experiments, how did Blue 36 make it back from 10 kilometres away? Our best guess is as follows. We think that in their exploratory flights, bees learn the relative positions of landmarks, just as they memorise objects close to the entrance to their nest, mentally logging the locations of tall trees, buildings, fence posts and so on. Researchers have shown that honeybees navigate better when there are obvious landmarks, using the sun as a compass to tell them which direction their nest is in relation to them. Ants are much easier to follow than bees, and ant researchers have found that displaced ants walk in long loops from their place of release (reminiscent of the paths of novice bees when they first leave their nest). It seems likely that when I released bumblebees in strange locations they did the same thing; they looped out from their place of release in search of familiar landmarks. If they were lucky enough to find one, they would quickly be able to get home. Buff-tails obviously forage quite a way from their nests, so these bees probably had a mental map of the landmarks for a kilometre or two around my house. Those bees released many kilometres from home would have had a relatively low chance of ever stumbling upon a familiar landmark; hence many never made it back, and those that did took days to do so.