By Geoff Tribe, Karin Sternberg and Jenny Cullinan
Fig. 1. Male Xylocopa caffra carpenter bee at his patch of Pelargonium cucullatum busy challenging the photographer.
Patrolling his patch!
Ever wonder about the antics of the yellow-haired carpenter bee hovering around a flowering shrub (Fig. 1), then darting off, only to return seconds later to hover briefly again? This is the male Xylocopa caffra which is distributed throughout South Africa. The females are unlike the males in that they are black and hairy with two bands of yellow hair (Fig. 2).
Fig. 2. The female Xylocopa caffra carpenter bee does not resemble the male.
What the males are doing is to patrol a patch of flowers to which females are attracted out of necessity whilst foraging in order to mate with them. A patrolling male will aggressively chase away intruding males and will even attempt to discourage you from coming too close by mock diving at you. Stretch out your arm and point your finger at the carpenter bee and it will often approach the finger directly and will follow the finger as it is moved slowly up and down – as if it is an intruder.
Xylocopa caffra males make quick circuits of an area of several square meters and females that are receptive allow themselves to be seized in flight (Watmough, 1974). The receptive female will vibrate her wings and help the male carry her in flight, a successful mating flight ending 50 meters or so up in the air about a kilometre away.
Exposing interstitial membranes
What is most fascinating is that the normally completely yellow male, when hovering around its patch of flowers, lowers its abdomen and exposes two of its interstitial membranes which appear as two black bands on the abdomen (Fig. 3). The most likely explanation for this behaviour is that the male is releasing a sex pheromone, both as a species specific chemical signal and female sexual attraction. When the male momentarily darts away, the membranes are less exposed when compared to when it hovers.
Fig. 3. Male Xylocopa caffra with his interstitial membranes exposed while patrolling his patch of Pelargonium.
For semi-social carpenter bees which nest in tunnels they construct within dead branches of trees or dried Aloe inflorescences, this method of patrolling patches of flowers that females must visit for the collection of pollen and nectar, is obviously suffice for the sexes to meet. The success of these encounters is enhanced by the permeation of a sex pheromone released by the male while patrolling his patch of flowers. Males of the carpenter bee Xylocopa hirsutissima have a different strategy and fly to the top of the mountain where they spread mandibular gland secretion over the ventral surface of their abdomen while hovering in the air awaiting the female (Velthuis and Camargo, 1975).
Drone Congregation Areas
For highly specialized social insects like honeybees, a more intricate and highly efficient behaviour has evolved to bring virgin queens and drones together. Mating in honeybees takes place in the air at considerable distance from the hive. These are called drone congregation areas (DCA) where thousands of drones congregate in certain locations in the air where they await the arrival of a virgin queen. Drones from as many as 240 colonies have been recorded in a DCA at one time. The phenomenon of DCAs is still little understood but certain characteristics pertaining to them are known. Both drones and queens independently find these locations, some of whose locations have been known for hundreds of years and are therefore persistent from year to year (Zmarlicki and Morse 1963). Drones and queens from hives brought in from outside the region are immediately able to find these DCAs.
Thousands of drones milling about in a DCA greatly reduce the chance of a queen being eaten by alpine swifts (Fig. 4a), bee-eaters, redwinged starlings (Fig. 4b) or other avian predators.
Fig. 4a Alpine swift – a major predator of honeybees in flight in southern Africa.
Fig. 4b Fork-tailed Drongo, a successful predator of stinging insects with a bee in its beak.
Swallows in Perth, Western Australia were found to have up to 19 drones in their stomachs (Tribe 1989). Even spider’s webs may prove a hazard, even within a DCA (Fig. 5). A virgin queen always leaves on her mating flight during a worker orientation flight which persists until the queen returns – thus further reducing the chance of predation.
Fig. 5. Drones within a DCA caught in a spider’s web between two tall trees in a pine plantation at Grabouw, Western Cape. The web was eventually shredded by comets of drones attracted to those caught in the web.
Advantages of a Drone Congregation Area
Obviously the DCA must serve an important function, especially if one considers the resources entailed in rearing hundreds of drones in each honeybee colony each year. The amount of drone comb and the number of adult drones present is positively correlated with the number of workers in the hive and represent a maximum of 14% of the total brood at any time. The number of drones kept in an active colony is restricted to about 1000. Because it takes 24 days from egg to emergence (compared with 19 days for an African worker) the larger drones consume abundant pollen. The cessation of the rearing of drone brood is in response to pollen shortage, while the killing or expulsion of drones by workers is mostly the result of a shortage of nectar and may occur at periods throughout the year (Currie 1987). Because the mating flights of drones and virgin queens are not random but concentrated into distinct areas, mating takes place in the shortest possible time. In times of very low drone population density, queens have a reasonable chance of being mated in a DCA. The lifespan of a drone is between 13-59 days.
The main advantage of a DCA appears to be to lessen the chance of a virgin queen mating with a related drone which is detrimental to the colony. With the sometimes thousands of drones drawn from diverse colonies in the area, the possibility of mating with a related drone is much reduced. In addition, queens of European races mate with an average of 8 drones and up to 17 (Woyke, 1962) which further reduces the level of inbreeding if she inadvertently mates with a related drone. By using DNA fingerprinting techniques, Moritz et al. (1996) determined that Apis mellifera capensis queens of the same population will have mated between 24 and 44 times.
Although the spermatozoa from one drone (12.7million) is suffice to fill the spermatheca of the queen, as the spermatozoa move from the oviduct past the spermatheca, fractions of the sperm of each drone is absorbed into it. Only about 370 000 spermatozoa (3%) are retained in the spermatheca. On the eversion of the aedeagus of the drone (like a glove blown inside out) (Fig. 6), the clump of sperm is followed by mucus and following the dissection of a recently mated queen the number of drones with which a queen has mated can thus be calculated.
Fig. 6. Everted genitalia of a drone.
The mated queen determines whether she must lay a fertilized or un-fertilized egg by measuring the cell width with her forelegs before commencing oviposition. At the height of summer a huge number of drones are present in a DCA, the majority of which will never mate – an estimate that only 4% of drones mate naturally. This huge investment in the transfer of genes indicates the importance of DCAs.
Characteristics of Drone Congregation Areas
DCAs vary greatly but most appear to have some form of relief in the form of trees or mountains near which they are established. Usually they occur high above the ground and are discerned from the noise the drones make whilst flying which sounds like a huge swarm of bees about to settle. Years back in Pretoria an irate grounds manager of a rugby club phoned to ask if we could remove a huge swarm of bees that interfered with their rugby matches on Saturday afternoons. It was quickly ascertained that no one had ever been stung, nor had any bees been observed but that the swarm ‘hovered high above the ground’. The rugby field was below a DCA and was flanked by high Eucalyptus trees. A DCA may expand or contract in size throughout the year but there is always a ‘core’ area which persists from year to year.
DCAs in Europe are recorded located at heights of 15-25m above ground (range of 8-40+m) over open ground or on summits, often in areas demarcated by buildings or rows of trees. They were 50m to 5km from the apiary and their individual sizes fluctuated from 30 to 200m in diameter but there was always a ‘steady centre’. Researchers found that drones of European races did not follow queens out of DCAs but always remained within its limits and could not be attracted to ground level with queen pheromone lures. Certain DCAs were more attractive than others and the number of drones visiting a DCA varied between sites and on different days. The sun was not used to orientate to DCAs. And of course there were no dances among the drones to indicate the location of a DCA! Successful matings from 10 to 16km distant have been recorded but the usual distance is from 5-7km. By placing drone-less colonies on islands or in desert areas, it was shown that to prevent mating an isolation of at least 15km from the nearest hive was required.
Drone flight in Pretoria during the summer months began at about 12h00 and ended after 17h00 with peak flight times between 14h30-15h45. The average number of flights undertaken by flight mature A. m. scutellata drones in Pretoria was 3.7 flights/day (n=23) with a range of 3-5 flights. Individual marked drones made multiple flights in an afternoon with a mean duration of 21min 30sec (range 12min 15sec to 33min 53sec). Using 8m/s as the flight speed, an A. m. scutellata drone is able to travel a distance of 10.3km in 21.5min with a maximum of 16.4km. When visiting a DCA no further than 1 to 1.5km distant (in ±3min), a drone could be cruising within the DCA for 45min on each flight. Between flights, drones would spend an average of 4 minutes feeding on uncapped honey. With three flights in an afternoon, including feeding intervals between flights, drones on average would spend 72min/day on mating.
From limited data it appeared that the older the drone, the longer the duration of the mating flight. The African bee produces and maintains larger numbers of drones even in resource-scarce conditions than do European races, which accounts for the huge numbers found in DCAs.
The first orientation flight of an Apis mellifera scutellata queen would take place on the third day after emergence. Usually there were orientation flights on two successive days before the virgin queen departed on a mating flight. The queen would only leave the hive during a worker orientation flight and such orientation flights would persist until the queen returned to the hive. The queen would fly once or several times in the same afternoon or on successive days until she was fully mated.
Gary (1971) recorded queen flight at 340m/min (5.6m/s), enabling a queen to take 2min 56sec to fly 1 000m. On Sylt Island (Germany) where the wind hardly ever stops blowing, matings took place at windspeeds of up to 5m/s (Tiesler 1972). The mean duration of a queen flight in Pretoria was 13 minutes 17 seconds (range 6min 11s – 21min 58s) (Fletcher and Tribe, 1977). Thus the A.m. scutellata queen could fly 4 515m in 13.28min or a return journey of 2 257m. Two further queen flights in November 1979 gave a duration of 19min 49sec and 26min 34sec which gives a maximum range of 9km. Drones evert within 1-6 seconds on mounting the queen. Upon eversion, the drone falls backwards and the resulting pressure within the aedeagus causes the non-chitinized section to rupture, the drone falling to the ground while the queen flies off with the ‘mating sign’.
Flight experienced workers appear to determine if and when the queen departs on her mating flight. They orientate in front of the hive in ever widening circles, possibly gauging if weather conditions are suitable for flight. In an observation hive, a virgin queen primed for a flight can be seen preening herself on the comb. Ever more workers vacate the hive on their orientation flight, and the cue for the queen to leave happens after several of the orientating workers return to the hive and run ‘randomly’ around on the comb. The queen, while trying to find the exit (because the cover is off the observation hive), invariably ‘jumps’ onto the glass window and then quickly follows the bees out of the entrance.
Behaviour within the Drone Congregation Area
Fig. 7. Honeybee drones attracted to a strip of cloth attached to fishing gut and a helium filled balloon which had been lowered to near ground level within a drone congregation area in Pretoria.
Drones form comets within the DCA of several hundred individuals which can dissipate as fast as they may be formed (Fig.7). Comets may be formed even in the absence of a queen when they orientate to any object that may enter the space such as butterflies, or even to another drone. Despite the literature recording that drones of European races of honeybees never descend near the ground (Gary 1962), in October 1986 in Floreat Park, Perth, Western Australia drone comets were observed flying at head height in an open park surrounded by trees (Tribe 1989b). Within the DCA of about 100m x 50m, individual drones were flying at knee height.
The queen is approached from below where she is silhouetted – hence the function of the abnormally large eyes of the drones. Because of its small visual acuity, a drone must fly within 1m of an object the size of a queen in order to see it (Butler and Fairey 1964). Queens may return fully mated on their first flight or may undertake subsequent mating flights on the same day or subsequent days.
Types of DCAs
How do drones and queens independently locate DCAs? Olfaction, especially pheromones, plays a crucial role in the life of the honeybee and no doubt this extends also to the formation and function of DCAs. Pioneering research into DCAs in Pretoria showed what appeared to be three types of DCAs which were used interchangeably in response to varying weather conditions: ‘barrier’, ‘sheltered’ and ‘convectional’. The most regularly visited DCA was situated downwind from a ridge which formed a barrier (e.g. A in figure 9, the most consistent DCA). After consulting weather bureau personnel, it was determined that a breeze passing over the ridge on an otherwise relatively still day would cause an eddy downwind. The height of the ridge and the velocity of the wind would determine the strength of the eddy and its distance from the ridge. Drones fly in the afternoons on hot days and ground that bakes in the sun during the day results in late afternoon breezes as cooler air is drawn in as the hotter air rises which often results in showers of rain.
Fig. 8. Map of the situation of drone congregation Areas around the apiary on the University of Pretoria Experiment Farm.
Eight DCAs were located around the apiary situated on the University of Pretoria Experimental Farm in 1975 with an average distance from the apiary of one kilometre (Fig. 8). By monitoring the wind speed at the entrance to the hive, it was determined that queens would not fly if the wind speed ≥5m/s, and drones at ≥7m/s. On extremely windy and usually overcast days, a DCA which normally attracted a smaller number of drones would supplant the barrier DCA in numbers. It occurred in the sheltered lee of a kopje where the air within was relatively calm compared to the gusts of wind all around it (D in figure 9). On Lion’s Rump in Cape Town there is such a DCA which can be looked down upon where the drones can be observed flying. Here again there is relatively calm air while all about the easterly rages.
At the height of summer in Pretoria with its oppressive heat with nary a breeze, the air is filled with the sound of drones. Huge convection currents were initiated by the air above the hot soil attracting in cooler air with the DCAs expanding considerably in size in response to this.
When comparing DCAs in Europe with those in Africa, it must be remembered that in Europe most honeybee colonies are hived and relatively few are feral, whereas in Africa the opposite is true and drones within a DCA are mostly from wild swarms in natural nesting sites. The late Beowulf Cooper when visiting the Pretoria DCAs related that DCAs in the United Kingdom were perceptibly warmer within them than without and regarded this as a distinctive feature for their formation in Britain. Cooper (1977) proposed a thermal vortex of warmer air within the DCA contributing to its formation.
Fig. 9. Capturing drones attracted to pieces of cloth on fishing gut after the helium filled balloon had been lowered and tethered within a drone congregation area in Pretoria. Only the uppermost piece of cloth below the balloon had queen pheromone on it.
This is what appears to be happening. The drones exiting their hives on mating flights appear to fly mostly upwind or alternately downwind until they reach an area of mild turbulence where they congregate. Here they fly around randomly within the limits of the eddy. Drones from comets caught in nets in DCAs (Fig. 9) were transferred to plastic bottles where they were marked by dabbing them with ‘tippex’ of different colours according to their DCA location. As they emerged between the fingers of the hand placed over the mouth of the bottle, a distinct musty smell could be detected (which incidentally attracted the occasional worker bee which would not leave). This odour could be a pheromone which not only assists the queen in finding the DCA but helps create the comets. Guard bees have been recorded in prolonged lapping over much of the dead body of a drone, especially near its thorax, apparently to salvage some highly desirable substance (Australian Beekeeper 1979, 80(8): 174). Crushed drones are highly attractive to worker bees and may produce a substance which is attractive to workers in summer, enabling a drone to drift to other colonies with immunity (Holmes and Henniker 1972). Another indication of a substance produced by drones is when drones and workers were returned to their queenless colony they were accepted, but the workers when confined to a cage with their queen and drones and returned to their colony are killed – but not the queen or drones (Orὅssi-Pál 1959). Drones could be attracted in DCAs with drone extracts as efficiently as with queen extracts – head extracts being the most successful (Ruttner H. 1972). In the presence of queen pheromone the drones will form comets as one drone follows the other. They will attempt to mate with any object silhouetted above them, a butterfly or stone thrown into the air.
A queen flying into a DCA releases queen pheromone (9-oxodecenoic acid) which results in the immediate formation of comets which zigzag after her, following her pheromone plume from at least 420m downwind of her. Drones are attracted by 9-oxodecenoic acid from distances of 50m and more while the queen tergite gland’s pheromone is effective only within short distances (less than 30cm) and increases copulation activity (Renner & Vierling 1977). A pheromone is also produced from the abdominal tergites in virgin queens 8 days or older and in mated queens, probably in the glands described by Renner and Baumann (1964) which are most active at the time of mating (Butler 1971). Production of a chemical from the abdominal sternites shows that it is restricted to nubile virgin queens because ‘wipes’ of newly emerged and very young laying queens had none of the perfume (Boch, Shearer and Young 1975). Several comets may be seen simultaneously within the DCA, dissipating as they lose the scent and reforming at will. They approach from below the queen and mate in the air, each drone in turn removing the previous ‘mating sign’ before inturn mating with her. Thus the reason why DCAs form in certain areas appears to be to facilitate through pheromones, the fast and efficient mating of the queen by following the pheromone plume.
Of interest is the question of why the multitudes of drones within a hive do not attempt to mate with their own mated or virgin queen who is continually secreting queen pheromone? Presumably the pheromone ‘bouquet’ released by a virgin as opposed to that of a mated queen is perceptibly different. Yet if mating within the hive were to occur, it would negate the major function of a DCA to prevent in-breeding. The difference between an inactive drone within the hive and that of a drone in a DCA is one of flight. Thus presumably flight changes the physiology of a drone and makes it responsive to queen pheromone. This could be a result of the accumulation of carbon dioxide in their tissues caused by exercise.
Could sound play an integral part in the formation of a DCA? Quietly sitting in a meadow, one of the most harmonious sounds is that of honeybees going about their business. Yet we are all familiar with the raised, high-pitched sound of an angry bee. Sound plays almost as important a role within the hive as that of pheromones. The piping of queens, the moaning of bees besieged by banded bee pirates (Palarus latifrons), the distress calls of worker bees in a queenless A. m. capensis colony, and many more sounds are familiar to the beekeeper. In addition, the male semi-social carpenter bee Xylocopa hirsutissima is known to raise its flight tone when it perceives the female approaching the top of the mountain that it is patrolling (Velthuis and Camargo, 1975).
Sounds have frequencies which are distinct, and in some instances the frequency or vibration may be more important than the sound that we hear. For example, following the departure of the old queen in a reproductive swarm, several virgin queens cut their way out of their queen cells. They soon detect the presence of the other queens and are further agitated by the workers which goad them on by harassing them. To shake off the harassing workers, the queen pipes with her wings. This high pitched sound causes the workers around her to freeze, and she scoots away temporarily before they again catch up with her. In an observation hive it is possible to follow the path of the queen as she pipes on the opposite side of the comb by observing the bees ‘freezing’ in the absence of the queen on this side. The sounds/ vibrations are carried through the comb.
The sound of drones returning to their hive after a mating flight is distinctive and cannot be confused with that of a worker bee. The screeching sound of a Stuka fighter aircraft in a bombing dive was deliberately built into the plane to terrify the enemy on the ground. Surely then, the loud and distinctive sound of a drone in flight also has an important purpose? The sound (and vibrations in the air?) of several thousand drones in a DCA could serve to make the locating of the DCA easier for both drones and virgin queens. Sound could also assist, besides sight and pheromones, in the creation of a DCA and the formation of drone comets and also to help distinguish a drone from a queen within the DCA. While marking drones in a DCA several would accidentally have ‘tippex’ splattered on their wings. This would result in a different sound being produced as it flew off which immediately attracted other drones to it. A major function of a DCA is to facilitate the complete mating of a virgin queen in the shortest possible time and a combination of sound, sight and pheromones may serve to accomplish this.
Evolution of male bee pheromone glands for mating?
While monitoring the number of times marked drones flew each afternoon and the duration of their flights, it was observed on many occasions that returning drones had the two interstitial abdominal membranes exposed – in exactly the same position as that of Xylocopa caffra males. Although dissections of drones were made to discover any underlying pheromone glands, they were so small that the equipment used could not detect them. If this is the source of a mating pheromone, it appears that it may be present in many bee species from solitary, sub-social and social species. Besides different flight times, the comparison of the composition of this pheromone could possibly help explain the sexual isolation of the different species of Apis and their phylogenetic linkages.
The authors at work:
Boch, R., Shearer, D.A. and Young, J.C. 1975. Honey bee pheromones: field tests of natural and artificial queen substance. Journal of Chemical Ecology 1(1): 133-148.
Brὔckner, D. 1979. Effects of inbreeding on worker honeybees. Bee World 60(3): 137-140.
Butler, C.G. 1971. The mating behaviour of the honeybee. Journal of Entomology 46: 1-11.
Butler, C.G. and Fairey, E.M. 1964. Pheromones of the honeybee: biological studies of the mandibular gland secretion of the queen. Journal of Apicultural Research 3: 65-76.
Cooper, B.A. 1977. Have you heard a drone assembly? British Isles Bee Breeders’ Association, 9pp.
Currie, R.W. 1987. The biology and behaviour of drones. Bee World 68(3): 129-143.
Fletcher, D.J.C. and Tribe, G.D. 1977. Natural emergency queen rearing by the African bee, A. m. adansonii, and its relevance for successful queen production by beekeepers. In African Bees: Taxonomy, Biology and Economic Use, ed. D.J.C. Fletcher. Pretoria: Apimondia. Part I, pp. 132-140; II pp. 161-168.
Gary, N.E. 1962. Chemical mating attractants in the queen honey bee. Science NY, 136: 773-774.
Gary, N.E. 1971. Observations on flight-experienced queen honeybees following extra-apiary release Journal of Apicultural Research 10(1): 3-9.
Holmes, F.O. and Henniker, N.H. 1972. Attractiveness of drones to worker honeybees. Gleanings in Bee Culture 100(10): 297.
Johannsmeier, M.F. (Ed.) 2001. Beekeeping in South Africa. Third Edition, Revised. Plant Protection Research Institute Handbook No 14, Agricultural Research Council, Pretoria. 288pp.
Moritz R.F.A., Kryger P. and Allsopp, M.H. 1996. Competition for royalty in bees. Nature 384: 31.
Orὅssi-Pál, Z. 1959. The behaviour and nutrition of drones. Bee World 40(6): 141-146.
Renner, M. and Baumann, M. 1964. Ueber Komplexe von subepidermalen Drὔsenzellen (Duftdrὔsen) der Bienenkὅnigen. Naturwissenschaften 51: 68-69.
Renner, M. and Vierling, G. 1977. Die Rolle des Taschendrὔsenpheromons beim Hochzeitsflug der Bienenkὅnigin. Behavioral Ecology and Sociobiology 2: 329-338.
Ruttner, F. 1968. The life and flight activity of drones. Australia Beekeeping 69(4): 279-284.
Ruttner, H. 1972. Neue Versuche ὔber die Flugbahnen der Drohnen ὔber den Alpen. Apimondia Science Bulletin, Bucharest, pp 79-81.
Tiesler, F.K. 1972. Mating stations in the islands north of Germany. Apimondia Science Bulletin, Bucharest, pp. 91-95.
Tribe, G. 1982. Drone mating assemblies. South African Bee Journal 54(5): 99-100, 103-112.
Tribe, G.D. 1989a. Drones caught in a spider’s web within a drone congregation area. South African Bee Journal 61(5): 110-111.
Tribe, G.D. 1989b. A drone congregation area in Perth, Western Australia. South African Bee Journal 61(4): 83-86.
Velthuis, H.H.W. and Camargo, J.M.F. de 1975. Further observations on the function of male territories in the carpenter bee Xylocopa (Neoxylocopa) hirsutissima Maidl (Anthophoridae, Hymenoptera. Netherlands Journal of Zoology 25(4): 516-528.
Watmough, R.H. 1974. Biology and behaviour of carpenter bees in southern Africa. The Journal of the Entomological Society of Southern Africa. 37(2): 261-281.
Woyke, J. 1962. The hatchability of ‘lethal’ eggs in a two sex-allele fraternity of honeybees. Journal of Apicultural Research 1: 6-13.
Zmarlicki, C and Morse, R.A. 1963. Drone congregation areas. Journal of Apicultural Research 2(1): 64-66.