Behavioral Ecology Advance Access originally published online on December 6, 2007
Behavioral Ecology 2008 19(2):255-261; doi:10.1093/beheco/arm117
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Foraging in honeybees—when does it pay to dance?
a Behaviour and Genetics of Social Insects Lab, School of Biological Sciences A12 b Centre for Mathematical Biology, The University of Sydney, Sydney, NSW 2006, Australia
Address correspondence to M. Beekman. E-mail: mbeekman{at}bio.usyd.edu.au.
Received 6 May 2007; revised 16 October 2007; accepted 18 October 2007.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIAL AND METHODSRESULTSDISCUSSIONSUPPLEMENTARY MATERIALFUNDINGREFERENCES |
|---|
Honeybees are unique in that they are the only social insects that are known to recruit nest mates using the waggle dance. This waggle dance is used by successful foragers to convey information about both the direction and distance to food sources. Nest mates can use this spatial information, increasing their chances of locating the food source. But how effective is the bees' dance communication? Previous work has shown that dancing does not benefit a honeybee colony under all foraging conditions and that the benefits of dancing are small. We used an individual-based simulation model to investigate under which foraging conditions it pays to dance. We compared the net nectar intake of 3 types of colonies: 1) colonies that use dance communication; 2) colonies that did dance but could not use the dance's spatial information; and 3) colonies that did not dance. Our results show that dancing is beneficial when the probability of independent discovery of food sources is low. Low independent discovery rates occur when patches are very small or very far away. Under these conditions, dancing is beneficial as only a single individual needs to find a patch for the whole colony to benefit. The main benefit of the honeybee's dance communication, however, seems to be that it enables the colony to forage at the most profitable patches only, ignoring forage patches that are of low quality. Thus, dancing allows the colony to rapidly exploit high-quality patches, thereby preventing both intra- and interspecific competitors from using that same patch.
Key words: Apis, central-place foraging, communication, recruitment, scouting, waggle dance.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIAL AND METHODSRESULTSDISCUSSIONSUPPLEMENTARY MATERIALFUNDINGREFERENCES |
|---|
Living in groups provides many benefits to individual group members. Individuals within a group are less vulnerable to predators (Hamilton 1971
; Sword et al. 2005), their locomotion may be enhanced when moving as a group (Krause and Ruxton 2002), territories can be defended better (Packer et al. 1990), and a group can effectively modify the environment, for example, nest thermoregulation in social insects (Jones and Oldroyd 2007). Another important benefit of being in a group is that an individual can obtain information from other members of the group about their environment (Lachmann et al. 2000). When individual group members have accurate information about predation risk and food availability of a habitat, other group members can use this social information to decide whether to stay or leave their current foraging range (King and Cowlishaw 2007). Similarly, naive Bechstein's bat females obtain information about the location of roosting sites (Kerth and Reckardt 2003), dwarf mongooses are warned of approaching predators (Rasa 1987), and naive fish can respond to information provided by a minority of individuals with pertinent directional information (Reebs 2000; Couzin et al. 2005).
Probably, the best-known examples of information sharing are found in the social insects (the ants and termites and some species of bees and wasps) in which many species have evolved complex signaling mechanisms that are used to recruit group members to food sources (Beekman and Dussutour, forthcoming). Many species of ant mark the route between their nest and discovered food sources with a chemical, thus indirectly leading nest mates to the food (Hölldobler and Wilson 1990) allowing the colony to exploit food sources that are too large to be exploited by a single individual (Beckers et al. 1990; Detrain and Deneubourg 2002).
Honeybees are unique among the social bees because they are the only bees that have evolved a recruitment mechanism that relies on the direct transfer of information: the waggle dance. The honeybee's waggle dance encodes both the direction and the distance to the advertised source. This source can either be a food patch (von Frisch 1967) or a new nest site (Lindauer 1955). During a typical dance, the dancer strides forward while vigorously shaking her body from side to side (Tautz et al. 1996). This is known as the "waggle phase" of the dance. After the waggle phase, she makes an abrupt turn to the left or right, circling back to start the waggle phase again. This is known as the "return phase." At the end of the second waggle, the dancer turns in the opposite direction so that with every second circuit of the dance she will have traced the famous figure-of-8 pattern of the waggle dance (von Frisch 1967). Directional information is contained in the angle of the waggle phase such that the alignment of the dancing bee’s body relative to vertical is similar to the angle of the goal from the sun's current azimuth. Distance information is encoded in the duration of the waggle phase (Esch et al. 2001).
The scientific study of the honeybee's dance has been controversial (Wenner and Wells 1990). This controversy was not helped by the absence of unambiguous proof of the dance's ability to increase the amount of food collected by a colony. It is only recently that experimental studies have focussed on quantifying the benefits, in terms of resources collected, of the honeybee's dance by comparing colonies that could and could not utilize dance information (Sherman and Visscher 2002; Dornhaus and Chittka 2004). By turning a colony on its side, dance communication is scrambled so that although successful foragers continue to perform dances, dance followers cannot extract spatial information from them. Both studies suggest that the benefits of dance communication depend on the distribution of forage and are most beneficial to a colony when forage is hard to find, variable in richness, and ephemeral. However, how exactly the honeybee's dance communication increases the colony's food intake and how this is affected by the distribution of food patches are not clear.
Seeley and Visscher (1988) proposed 3 means by which a honeybee colony could benefit from information sharing among foragers about the location of profitable food sources. The sharing of information 1) reduces the cost of finding new food sources; 2) increases the quality of the food sources found; or 3) increases the colony's ability to exploit a high-quality food source quickly, thereby outcompeting other colonies as well as other species. Their empirical findings did not support the first hypothesis as they found that recruits (foragers that have followed a dance and then searched for the advertised food source) spent more time searching for a food patch than scouts, those that search independently without using dance information. Seeley and Visscher (1988) did find support for the 2 other hypotheses. The quality of the food sources found by scouts, measured as total nectar or pollen collected, was significantly lower for scouts than for recruits, showing that on average recruits forage at better food sources than scouts. Second, dancing increased the speed at which a colony can exploit new food sources. This suggests that the dance language is most beneficial when the quality of food sources is variable because it allows recruitment toward the most profitable sources.
Here, we explore how the relative success of scouts and recruits, measured as the amount of time it takes to locate a food patch, affects whether or not dance communication increases a colony's nectar intake. We used an individual-based simulation model, mimicking as much as possible the behavioral rules employed by foraging bees (e.g., Seeley 1995). Previous modeling work (Dornhaus et al. 2006) suggested that communication is most beneficial when food patches are of low quality, few, and variable. The model of Dornhaus et al. (2006) also suggested (based on a sensitivity analysis) that the extent to which a colony benefits from communication depends on the relative success of scouts and recruits locating food patches. However, their model did not allow this phenomenon to be explored further. In contrast, our model includes the difference in search time between scouts and recruits based on empirically obtained search times (Seeley and Visscher 1988). At the same time, we need to ensure that the relative number of scouts and recruits is regulated realistically. Thus, in accordance with recent theoretical and empirical work (Beekman et al. 2007), our model bees use the availability of recruitment dances to decide whether to become a scout or a recruit.
We present 3 variations of our model. In 2 variations, successful foragers return to their colony and advertise the location of their patch to nest mates by performing dances. In the first variation, dance followers utilize this information and attempt to locate the advertised patch, whereas in the second variation no spatial information is transferred and dance followers fly out in a random direction and behave as scouts (this is similar to the experimental manipulations by Sherman and Visscher [2002] and Dornhaus and Chittka [2004]). In the third variation, successful foragers do not dance and thus bees do not follow dances. Our 3 model variations allow us to distinguish between the transfer and use of spatial information as well as the costs associated with dancing and dance following.
|
MATERIAL AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIAL AND METHODSRESULTSDISCUSSIONSUPPLEMENTARY MATERIALFUNDINGREFERENCES |
|---|
Model description
We developed an individual-based simulation model of foraging honeybees using Simulink in Matlab 7.3.1. Simulations were run in discrete time steps (t) in which 1 simulation time equals 10 s real time. Total simulation time was 2880 time steps representing 8 h. We calculated the net nectar intake of the colony at the end of the simulation. Bees only consumed nectar while in flight at a rate of 1 mg per kilometer flown. Bees are known to fly at a speed of 7 m/s (Beekman et al. 2006
); hence, a bee that spends 16 min on a trip will have flown a total distance of 6.72 km during which she will have consumed 6.72 mg of nectar. Similarly, a bee visiting a patch at a distance of 4 km from the colony flies a total distance of 8 km per trip thereby consuming 8 mg. Successful foragers bring back to the colony 40 mg of nectar (note that the net amount of nectar will depend on the distance to the patch, as the further a patch, the more nectar a bee consumes in flight). We simulated a total of 260 bees, all of which were either active or potential foragers. The number 260 is sufficient to ensure that there are always bees available to be recruited to the different tasks associated with foraging. We used 4 equal food patches per experiment. We varied the likelihood that patches were discovered by independent search (scouting) by changing the distance of the patch and the size of the patch. Patches were located at 2, 4, 6, 8, or 10 km from the hive. Patch size was measured in degrees: 5, 15, 30, 45, and 60. We did not include crowding, so that even at "small" patches all bees that had found the patch could collect nectar. Patches were depleted as bees collect nectar at a rate of P0KB, where P0 is the patch's initial profitability, K is a constant set at 0.999, and B is the total number of foragers that have visited that patch. Thus, the profitability of a patch decreases as the number of bees visiting that patch increases.
We used 3 variations on the model: 1) bees perform recruitment dances when foraging at a profitable patch, and dance followers can use this spatial information to help them locate the advertised site (costs and benefits of dancing); 2) bees perform recruitment dances, but dance followers are unable to use this information and instead search for food patches independently (costs but no benefits); and 3) bees do not perform recruitment dances (no costs and no benefits). All 3 variations are illustrated in Figure 1. Bees active in foraging can be in one of the 6 following states.
- Novices: Bees without previous foraging experience; all bees start their foraging day as novices. Novices will go to the dance floor to attempt to find a dance to follow.
- Dancing bees: Bees that judge the profitability of their patch high enough to recruit to that patch.
- Recruits: Bees that have followed a dance and attempt to locate the advertised patch.
- Scouts: Bees that have failed to find a dance. Scouts fly out in a random direction. (Note that in the 2 model variations in which dance information cannot be used or is not available, all foragers leave the colony as scouts.)
- Employed foragers: Bees that have located a profitable food patch, either as scouts or as recruits.
- Retired bees: Bees that no longer forage at their patch because the patch's profitability has declined.
- Dancing bees: Bees that judge the profitability of their patch high enough to recruit to that patch.
|
Novices
Novices entered the system one at a time every 6 time steps so that all bees were in the system at t = 1560. Bees were staggered to mimic the natural situation where not all potential foragers will leave the colony in the morning together. This also avoided them being treated by the model as a group instead of individual bees. Novices first attempt to locate a dance. In the absence of dancing (see Dancing bees below), novices directly entered the scout subsystem.
Dancing bees
Dancing bees convey information about the location of food patches by performing dances on an area in the colony called the "dance floor." As the distance to a patch is coded in the duration of the waggle phase (von Frisch 1967), dances for farther patches took longer to finish (assuming equal profitability). The number of time steps needed to perform 10 waggle phases was set at 1, 2, 3, 4, and 5 for patches at distances 2, 4, 6, 8, and 10 km, respectively.
Only dances for one particular patch could be witnessed by dance followers at any one time step. To decide which dance was witnessed, a random number between 1 and 100 was assigned to each patch in each time step. This number was then multiplied by the total number of bees dancing for that patch, and a dance for the patch with the highest number was read. In doing so, we ensured that the probability of a particular patch being witnessed increased as more bees were dancing for that patch without completely excluding patches for which only a small number of bees were dancing.
Novices, retired bees, and failed recruits (termed unemployed bees) attempt to witness dances. If at least one dance was available on the dance floor, the unemployed bee would follow it and enter the recruit subsystem. If no dance information was available, she entered the scout subsystem. Successful witnesses spent 1 time step reading the dance before becoming a recruit.
In the model variation in which dances do take place but dance followers cannot utilize this information, dance followers were sent to the scout subsystem after finishing witnessing a dance (Figure 1B). In the absence of dancing, all novices entered the scout subsystem without delay (Figure 1C).
Recruits
Recruited bees are known to make 4.1 ± 2.8 failed trips before locating the advertised patch (Seeley and Visscher 1988). Hence, the number of failed trips, x, was generated using a normal distribution with a mean of 4.1 and a variance of 2.8 (when the value returned was negative, we set this at zero). Each failed trip took 96 time steps (16 min real time that is similar to what has been found in real bees [Seeley and Visscher 1988]). Approximately 60% of recruits successfully locate the advertised patch after their failed trips (Seeley and Visscher 1988). We therefore assigned a random number between 0 and 1, y, to each recruit after she has made her allocated number of failed trips. If y > 0.4, she successfully found the advertised patch; otherwise, she returned to the dance floor.
Scouts
Scouts are bees that failed to find a dance (Beekman et al. 2007) and search for food independently. In our model, the success of a single scouting trip was proportional to the linear direction (in degrees) and distance (in kilometers) that the scout had flown on leaving the colony. (Note that the assumption of a linear flight direction is a simplification.) Each scout exited the colony in a random direction, and this direction remained the same throughout her flight. She flew a random distance sampled from a normal distribution with mean and variance of 4. This random flight distance had to be at least equal to the distance of the patch from the colony for the scout to discover that patch. Therefore, the closer the patch was to the colony, the more likely the scout would encounter the patch. As a result, patches that were distant from the colony could still be discovered but with a lower probability than those closer to the colony. In addition, the larger the patch (measured in degrees), the more likely the scout's flight would intercept that patch. If the scout did not find a patch, she returned to the colony and attempted another scouting flight.
Employed foragers
Successful recruits and scouts as well as retired bees that revisit a patch they have previously foraged at (see further) become employed foragers if their patch was of sufficient profitability. (The profitability of a patch decreases when bees visit that patch because they remove food from the patch.) If the patch's profitability is too low, the probability that she abandoned that patch (become retired) increased (see Supplementary Figure S1). If she decided to continue foraging at that patch, she next had to decide whether or not to dance for that patch. The decision to dance depended on the patch's profitability and on the bees' influx rate b (number of employed bees that returned to the colony in the last 6 time steps) (e.g., Seeley 1994). The returning forager first decided whether or not to dance depending on patch profitability. When patch profitability was high, she would dance irrespective of the bees' influx rate. However, when patch profitability was not very high, she would only perform a dance when the bees' influx rate was low (Supplementary Figure S2).
Employed foragers always carried the maximum amount of nectar back to the colony (40 mg). Once a bee had become an employed forager, she continued foraging at that patch (provided patch profitability remained high); hence, she remembered the location of the patch.
Retired bees
Retired bees are employed foragers that abandoned their patch because its profitability has decreased. A retired bee inspects her patch to determine if its profitability has increased as often occurs under natural conditions where nectar and pollen are renewed. The probability to become a reactivated forager (a previously retired bee that resumed foraging when its patch is again profitable [Biesmeijer and de Vries 2001]) depended on the patch's profitability (Supplementary Figure S3). When patch profitability remained low, the bee completely abandoned the patch (Supplementary Figure S1) and returned to the dance floor. Even though our patches never became profitable again (as we only simulate one foraging day), we needed to include retired bees for 2 reasons. Firstly, retired bees consume nectar when inspecting their patch. Secondly, when performing inspection flights, retired bees do not follow dances and can therefore not be recruited to other food patches. Only when a retired bee has completely abandoned its patch, will she start following dances again and be either recruited to a new patch or become a scout.
Experiments
In each experiment, we determined the net nectar intake of our 3 model variations for the different patch distances and patch sizes (4 distances, 5 patch sizes: 20 experiments per model variation). Patch profitability was set at 1 (maximum) for all patches at the beginning of the simulation. To specifically study the effect of patch variability on the relative success of dance communication, we also performed experiments in which we used 4 patches of size 15° located at 4 km that differed in their initial profitability (1, 0.5, 0.2, and 0.2 in the first experiment; 1, 0.4, 0.1, and 0.1 in the second experiment). The aim of these experiments was to determine how the different model variations allocate their foraging force over patches that differ in their initial profitability.
For all experiments, we performed 20 runs per model variation. We compared the net nectar intake among our 3 different model variations using 1-way analysis of variance. To determine which model variations were significantly different, we performed post hoc least square difference tests on the observed mean nectar intakes.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIAL AND METHODSRESULTSDISCUSSIONSUPPLEMENTARY MATERIALFUNDINGREFERENCES |
|---|
Our results clearly show that dance communication benefits the colony, measured in net nectar intake, when food patches are hard to find independently. When we offered small patches (5°), colonies that used dance communication always collected more nectar than colonies that could not communicate the location of forage (Figure 2; see figure and legend for details of the statistical tests). Similarly, when patches were hard to find independently because they were far away (8 km), dance communication increased the colony's net nectar intake relative to colonies that could not communicate (Figure 2). Conversely, as patch size increased (and food was thus easier to find), dance communication no longer benefited the colony when patches were relatively close to the colony (2 and 4 km). In fact, when patch size was greater than 15°, using dance information actually disadvantaged colonies when the patches were found at a distance of 2 or 4 km. This is because bees are sent to search in a random direction (and hence become scouts) in the model variations that cannot use dance communication, and this facilitated the finding of relatively large and nearby patches. The chance that dance followers successfully find the advertised site is lower compared with randomly searching scouts in real honeybee colonies (Seeley and Visscher 1988
), something that we explicitly incorporated into our model.
|
= 0.05 (post hoc least square difference).When patches were relatively small (15°) and neither very far nor very close (4 km), all 3 model variations collected a similar amount of nectar (Figure 2). We therefore used these foraging conditions to see if dance communication does benefit the colony when patches differ in their initial profitability. And indeed, when 1 patch is of superb quality, 1 of mediocre, and 2 of poor quality, the ability to communicate significantly increased the colony's net nectar intake (Figure 3). This is because colonies that made use of dance communication allocated a larger number of foragers to the most profitable patch whereas mostly ignoring poor patches (Figure 4). Even though colonies that could not use dance information still allocated more foragers to the best patch, this difference was not as pronounced as when bees were able to communicate.
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIAL AND METHODSRESULTSDISCUSSIONSUPPLEMENTARY MATERIALFUNDINGREFERENCES |
|---|
Our results show that when all patches are of equal profitability, dance communication benefits the colony as soon as the average success of scouts in locating food patches falls below the average success of recruits. Thus, when food patches are easily discovered by independent search, a colony does best to send its foragers out on a random search instead of attempting to recruit them to a particular site, whereas recruitment is beneficial when patches are difficult to locate, for example, when they are far away (Figure 2). In fact, only in the presence of dance communication does it pay for scouts to search over long distances, as otherwise the potential benefits of locating a forage site far away will never outweigh the costs of long-distance scouting (Beekman and Ratnieks 2000
). When food patches were easy to find, net nectar intake was higher in the absence of dancing compared with the nectar intake of colonies in which the bees did dance but could not use this information (Figure 2). This is because no time was wasted on dancing and dance following so that a larger number of bees were sent out as scouts. However, the difference between the model variations with and without such "wasted" time was not great (Figure 2).
Notwithstanding the benefit of dancing when scouting success is low, our results show that the dance's main benefit is the ability to rapidly exploit the most profitable patches when patch quality is variable. When we offered our colonies patches that varied in their initial profitability, colonies that utilized dance communication collected more nectar than those that could not use such information (Figure 3). The ability to communicate the location of food patches to nest mates leads to a rapid increase in the number of individuals that forage at that site (Figure 4) (Seeley and Visscher 1988) so that the discovery of a profitable food source by just one scout is, in theory, sufficient to allow a colony to rapidly exploit that food source. Without the use of dance communication, scouts will continue to find food patches of low profitability and the forager force will be more evenly spread over the available patches. Thus, contrary to earlier studies which concluded that the honeybee's dance communication evolved as an adaptation to boom and bust foraging conditions typical of the tropics (Sherman and Visscher 2002; Dornhaus and Chittka 2004; Dornhaus et al. 2006), it seems more likely that it is competition for food sources that has driven the evolution of information transfer in social insects (Nieh 1999, 2004).
One potential role for dancing that we did not include in our model is the possibility that dances serve as a signal aimed at increasing the number of bees departing to search for food without communicating spatial information. Returning foragers can either arouse unemployed individuals ("there is something out there") or transfer floral scents ("look for something that smells like this"). Such arousal signals are indeed used by bumblebees (Dornhaus and Chittka 1999), several species of stingless bees (Nieh 2004), and by honeybees when food is found close to the colony (von Frisch 1967; Weidenmüller and Seeley 1999; Beekman et al. 2005). Floral scents play an important role in reactivating honeybee foragers (Lindauer 1963; Reinhard et al. 2004; Beekman 2005).
It is interesting that honeybees are the only social bees that have evolved an intranidal spatially explicit recruitment mechanism. In many species of social insects, recruitment toward profitable food sources occurs outside the nest, for example, via odor trails or the marking of food sources. This allows other species as well as conspecifics to "eavesdrop" on those signals (Nieh et al. 2004) thus severely reducing the benefits of recruitment. Such eavesdropping can be avoided when recruitment takes place inside the nest (Nieh 1999; Nieh 2004). Hence, the honeybee's dance communication prevents eavesdropping, whereas the transfer of spatial information allows the colony to rapidly monopolize highly profitable food sources thus making the honeybee a superior competitor.
The stingless bees (Meliponini) share many traits with honeybees such as relatively large colonies and food storage within the nest. They also evolved in a similar environment as honeybees, the tropics (Michener 2000; Oldroyd and Wongsiri 2006). Could it therefore be that honeybees suffer more intra- and interspecific competition compared with stingless bees and that this has driven the evolution of the dance communication? This seems unlikely given the apparent competition between stingless bee species (Hubbel and Johnson 1978; Nieh et al. 2004) and between stingless bees and honeybees (Koeniger and Vorwohl 1979), as well as the high densities in which stingless bees are often found (Eltz et al. 2002; Eltz et al. 2003).
A possible alternative hypothesis for the evolution of the honeybees' dance communication is an original function in communicating the location of new nest sites after reproductive swarming. Once the dance communication had evolved in the context of nest site selection (as the location of a potential nest site is also encoded in the dance [Lindauer 1955], a situation unique to honeybees), it could then be adapted to foraging, thus giving the honeybee an extra competitive advantage. Obviously, this hypothesis remains speculative.
|
SUPPLEMENTARY MATERIAL |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIAL AND METHODSRESULTSDISCUSSIONSUPPLEMENTARY MATERIALFUNDINGREFERENCES |
|---|
Supplementary Figures 1–3 can be found at http://www.beheco.oxfordjournals.org/.
|
FUNDING |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIAL AND METHODSRESULTSDISCUSSIONSUPPLEMENTARY MATERIALFUNDINGREFERENCES |
|---|
Australian Research Council (DP0345360 to M.B.); University of Sydney (to M.B.).
|
ACKNOWLEDGEMENTS |
|---|
We would like to thank Ben Oldroyd, Sharoni Shafir, Anne Houde, Andrew King, and an anonymous reviewer for helpful comments on an earlier version of the manuscript.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIAL AND METHODSRESULTSDISCUSSIONSUPPLEMENTARY MATERIALFUNDINGREFERENCES |
|---|
Beckers R, Deneubourg JL, Gross S, Pasteels JM. Collective decision making through food recruitment. Insectes Soc (1990) 37:258.[CrossRef]
Beekman M. How long will honey bees (Apis mellifera L.) be stimulated by scent to revisit past-profitable forage sites? J Comp Physiol A (2005) 191:1115–1120.[CrossRef][Web of Science][Medline]
Beekman M, Doyen L, Oldroyd BP. Increase in dance imprecision with decreasing foraging distance in the honey-bee Apis mellifera L. is partly explained by physical constraints. J Comp Physiol A (2005) 191:1107–1113.[CrossRef][Web of Science][Medline]
Beekman M, Dussutour A. Forthcoming. How to tell your mates—costs and benefits of different recruitment mechanisms. In: Food exploitation by social insects: an ecological, behavioral, and theoretical approach—Jarau S, Hrncir M, eds. Boca Raton (FL): CRC Press.
Beekman M, Fathke RL, Seeley TD. How does an informed minority of scouts guide a honey bee swarm as it flies to its new home? Anim Behav (2006) 71:161–171.[CrossRef][Web of Science]
Beekman M, Gilchrist AL, Duncan M, Sumpter DJT. What makes a honeybee scout? Behav Ecol Sociobiol (2007) 61:985–995.[CrossRef][Web of Science]
Beekman M, Oldroyd BP, Myerscough MR. Sticking to their choice—honey bee subfamilies abandon declining food sources at a slow but uniform rate. Ecol Entomol (2003) 28:233–236.[CrossRef]
Beekman M, Ratnieks FLW. Long range foraging by the honeybee Apis mellifera L. Funct Ecol (2000) 14:490–496.[CrossRef]
Biesmeijer JC, de Vries H. Exploration and exploitation of food sources by social insect colonies: a revision of the scout-recruit concept. Behav Ecol Sociobiol (2001) 49:89–99.[CrossRef][Web of Science]
Couzin ID, Krause J, Franks NR, Levin SA. Effective leadership and decision making in animal groups on the move. Nature (2005) 455:513–516.
Detrain C, Deneubourg JL. Complexity of environment and parsimony of decision rules in insect societies. Biol Bull (2002) 202:268–274.
Dornhaus A, Chittka L. Evolutionary origins of bee dances. Nature (1999) 401:38.[CrossRef][Web of Science]
Dornhaus A, Chittka L. Why do honey bees dance? Behav Ecol Sociobiol (2004) 55:395–401.[CrossRef][Web of Science]
Dornhaus A, Klügl F, Oechslein C, Puppe F, Chittka L. Benefits of recruitment in honey bees: effects of ecology and colony size in an individual-based simulation model. Behav Ecol (2006) 17:336–344.
Eltz T, Brühl CA, Imiyabir Z, Linsenmair KE. Nesting and nest trees of stingless bees (Apidae: Meliponini) in lowland diptocarp forests in Sabah, Malaysia, with implications for forest management. For Ecol Manage (2003) 172:301–313.[CrossRef]
Eltz T, Brühl CA, van der Kaars S, Linsenmair KE. Determinants of stingless bee nest density in lowland dipterocarp forests of Sabah, Malaysia. Oecologia (2002) 131:27–34.[CrossRef][Web of Science]
Esch HE, Zhang S, Srinivasan MV, Tautz J. Honeybee dances communicate distances measured by optic flow. Nature (2001) 411:581–583.[CrossRef][Medline]
Hamilton WD. Geometry of the selfish herd. J Theor Biol (1971) 31:295–311.[CrossRef][Web of Science][Medline]
Hölldobler B, Wilson EO. The ants. (1990) Cambridge (MA): Belknap Press.
Hubbel SP, Johnson LK. Comparative foraging behavior of six stingless bee species exploiting a standardized resource. Ecology (1978) 59:1123–1136.[CrossRef][Web of Science]
Jones JC, Oldroyd BP. Nest thermoregulation in social insects. Adv Insect Physiol (2007) 33:153–191.
Kerth G, Reckardt K. Information transfer about roosts in female Bechstein's bats: an experimental field study. Proc R Soc Lond B Biol Sci (2003) 270:511–515.[Medline]
King AJ, Cowlishaw G. When to use social information: the advantage of large group size in individual decision making. Biol Lett (2007) 3:137–139.
Koeniger N, Vorwohl G. Competition for food among four sympatric species of Apini in Sri Lanka: Apis dorsata, Apis cerana, Apis florea and Trigona iridipenis. J Apic Res (1979) 18:95–109.
Krause J, Ruxton GD. Living in groups. (2002) Oxford: Oxford University Press.
Lachmann M, Sella G, Jablonka E. On the advantages of information sharing. Proc R Soc Lond B Biol Sci (2000) 267:1287–1293.[Medline]
Lindauer M. Schwarmbienen auf Wohnungssuche. Zeitschr vergl Physiol (1955) 37:263–324.[CrossRef]
Lindauer M. Allgemeine Sinnesphysiologie. Orientierung im Raum. Fortschr Zool (1963) 16:58–140.[Medline]
Michener CD. The bees of the world. (2000) Baltimore (MD): Johns Hopkins University Press.
Nieh JC. Stingless-bee communication. Am Sci (1999) 87:428–435.[CrossRef][Web of Science]
Nieh JC. Recruitment communication in stingless bees (Hymenoptera, Apidae, Meliponini). Apidologie (2004) 35:159–182.[CrossRef]
Nieh JC, Barreto LS, Contrera FAL, Imperatriz-Fonseca VL. Olfactory eavesdropping by a competitively foraging stingless bee, Trigone spinipes. Proc R Soc Lond B Biol Sci (2004) 271:1633–1640.[Medline]
Oldroyd BP, Wongsiri S. Asian honey bees. Biology, conservation and human interactions. (2006) Cambridge (MA): Harvard University Press.
Packer C, Scheel D, Pusey AE. Why lions form groups: food is not enough. Am Nat (1990) 36:1–19.[CrossRef]
Rasa OAE. The dwarf mongoose—a study of behavior and social structure in relation to ecology in a small desert carnivore. Adv Study Behav (1987) 17:121–163.[CrossRef]
Reebs SG. Can a minority of informed leaders determine the foraging movements of a fish shoal? Anim Behav (2000) 59:403–409.[CrossRef][Web of Science][Medline]
Reinhard J, Srinivasan MV, Zhang S. Scent-triggered navigation in honeybees. Nature (2004) 427:411–.[Medline]
Seeley TD. Honey bee foragers as sensory units of their colonies. Behav Ecol Sociobiol (1994) 34:51–62.[CrossRef][Web of Science]
Seeley TD. The wisdom of the hive. (1995) Cambridge (MA): Harvard University Press.
Seeley TD, Visscher PK. Assessing the benefits of cooperation in honeybee foraging: search costs, forage quality, and competitive ability. Behav Ecol Sociobiol (1988) 22:229–237.[CrossRef][Web of Science]
Sherman G, Visscher PK. Honeybee colonies achieve fitness through dancing. Nature (2002) 419:920–922.[CrossRef][Medline]
Sword GA, Lorch PD, Gwynne DT. Migratory bands give crickets protection. Nature (2005) 433:703.[CrossRef][Medline]
Tautz J, Rohrseitz K, Sandeman DC. One-strided waggle dance in bees. Nature (1996) 382:32.[Medline]
von Frisch K. The dance language and orientation of bees. (1967) Cambridge (MA): Harvard University Press.
Weidenmüller A, Seeley TD. Imprecision in waggle dances of the honeybee (Apis mellifera) for nearby food sources: error or adaptation? Behav Ecol Sociobiol (1999) 46:190–199.[CrossRef][Web of Science]
Wenner AM, Wells PH. Anatomy of a controversy. The question of a "language" among bees. (1990) New York: Columbia University Press.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  A. Dussutour, S. C. Nicolis, G. Shephard, M. Beekman, and D. J. T. Sumpter The role of multiple pheromones in food recruitment by ants J. Exp. Biol., August 1, 2009; 212(15): 2337 - 2348. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||