Observation Hives Research Articles

The signals initiating the mass exodus of a honeybee swarm from its nest

Animals that travel in groups must synchronize the timing of impending departures to ensure group cohesion. The mechanisms used by a honeybee, Apis mellifera, colony to organize the departure of a swarm from its nest remain a mystery. We examined the signals that trigger a swarm's explosive exodus from the parental nest, and we documented the concurrent changes in bee density and mobility. Using video recordings of swarms exiting observation hives, we analysed how bees in three swarming colonies prepared for and then performed their sudden departures. Over the 60 min before swarm exodus, the production of piping signals gradually increased and ultimately peaked at the start of the swarm departure. Also, during swarm exodus, bee density (number of bees in 100 cm2) dropped markedly, whereas the average bee velocity (mm/s), and the production of buzz-run signals, spiked dramatically. Neither waggle runs nor shaking signals showed increases before or during swarm exodus. The explosive departure of a honeybee swarm from its parental nest shows how animals can use the same communication signals in different contexts; we now know that honeybees use piping and buzz-run signals to initiate both a swarm's departure from its nest and a swarm's take-off from its bivouac site. This study also shows how a small minority of individuals in a social insect colony can operate as an oligarchy to make an important decision, i.e. when a swarm should leave its nest to found a new colony.

East Learns from West: Asiatic Honeybees Can Understand Dance Language of European Honeybees

The honeybee waggle dance, through which foragers advertise the existence and location of a food source to their hive mates, is acknowledged as the only known form of symbolic communication in an invertebrate. However, the suggestion, that different species of honeybee might possess distinct ‘dialects’ of the waggle dance, remains controversial. Furthermore, it remains unclear whether different species of honeybee can learn from and communicate with each other. This study reports experiments using a mixed-species colony that is composed of the Asiatic bee Apis cerana cerana (Acc), and the European bee Apis mellifera ligustica (Aml). Using video recordings made at an observation hive, we first confirm that Acc and Aml have significantly different dance dialects, even when made to forage in identical environments. When reared in the same colony, these two species are able to communicate with each other: Acc foragers could decode the dances of Aml to successfully locate an indicated food source. We believe that this is the first report of successful symbolic communication between two honeybee species; our study hints at the possibility of social learning between the two honeybee species, and at the existence of a learning component in the honeybee dance language.

Direct introduction of mated and virgin queens using smoke: a method that gives almost 100% acceptance when hives have been queenless for 2 days or more

SummaryWe compared the acceptance of virgin and mated queens introduced into queenless hives using either artificial queen cells or direct introduction accompanied by smoke. In Experiment 1, virgin queens aged 3–4 days were introduced into 5-frame hives than had been dequeened 1, 2, 3, 4, 5, or 6 days previously. Acceptance increased significantly with the length of time a colony had been queenless, and direct introduction gave significantly greater success than artificial queen cells (between 31% and 100% acceptance vs. 8% to 92% for direct and cell introduction respectively, depending on the period of queenlessness). In Experiment 2, virgin and mated queens were introduced into 2-frame observation hives that had been dequeened 1, 2, 3 and 4 days previously. The probability of acceptance was significantly higher for mated queens than virgins, for direct introduction versus artificial queen cells, and for longer queenless periods. Accordingly, the probability of a queen being balled by the workers declined significantly with the duration of the queenless period, and was significantly less for mated versus virgin queens. Finally, in Experiment 3, we introduced mated queens into medium-sized hives (10 medium Langstroth frames) that had been queenless for 2 days using both the direct introduction and artificial cells. All queens were accepted.

A scientific note on the directional accuracy of the waggle dance over the course of a day

Honey bees communicate information about distance and direction to a food source with a waggle dance (von Frisch, 1993). During the waggle phase see (Tautz et al., 1996), a bee walks forward an angle with respect to vertical while waggling her abdomen. This angle matches, on average, the angle between food source and sun azimuth. For a given food location, there is variation both within dances of the same bee and between different bees. Lindauer (1961) reported that when the sun is at the zenith (which only occurs in the tropics), dancing either ceases or is disoriented 10 minutes before and after solar noon. No directional information can be communicated when the sun is directly overhead. In temperate regions, colony foraging activity declines around noon, a phenomenon von Frisch (1993) called “noontime laziness”. He argued that this effect could not be explained in terms of food profitability, and attributed it to a diurnal periodicity in activity, although this was not directly tested. An alternative explanation of noontime laziness links these two phenomena: perhaps foragers experience greater difficulty orienting their dances when the sun is higher in the sky, and are less inclined to dance, which in turn reduces foraging activity. This study tests one component of this hypothesis: the effect of the sun’s position on the accuracy of the information within the dance. Three unrelated colonies, A, B, and C, were housed in observation hives (see Seeley, 1995, Chap. 4). Colonies were housed at two locations (A

The Use of the Vibration Signal and Worker Piping to Influence Queen Behavior during Swarming in Honey Bees, Apis mellifera

AbstractWe investigated worker regulation of queen activity during reproductive swarming by examining the rates at which workers performed vibration signals and piping on queens during the different stages of the swarming process. Worker–queen interactions were first examined inside observation hives during the 2–3 wk that preceded the issue of the swarm (pre‐swarming period) and then inside the swarm clusters during the period that preceded liftoff and relocation to a new nest site (post‐swarming period). Queen court size did not differ between the pre‐ and post‐swarming periods, but workers fed the queens less inside the swarm clusters. Workers performed vibration signals on the queens at increasing rates throughout the pre‐swarming period inside the natal nest, but rarely or never vibrated the queen inside the swarm. Piping was performed on the queens during both the pre‐ and post‐swarming periods and always reached a peak immediately before queen flight. During the final 2–4 h before swarm liftoff, queens were increasingly contacted by waggle dancers for nest sites, some of which piped the queen. The vibration signal may operate in a modulatory manner to gradually prepare the queen for flight from the natal nest, and the cumulative effects of the signal during the pre‐swarming period may make further vibrations on the queen unnecessary when inside the swarm cluster. In contrast, worker piping may function in a more immediate manner to trigger queen takeoff during both the pre‐ and post‐swarming periods. Workers that vibrate and pipe the queen tend to be older, foraging‐age bees. The regulation of queen activity during colony reproduction may therefore be controlled largely by workers that normally have little contact with queens, but help to formulate colony reproductive and movement decisions.

Improved technique for introducing four-day old virgin queens to mating hives that uses artificial and natural queen cells for introduction

SummaryWe compared the acceptance of 4-day old virgin queens introduced into mating nucleus hives using natural and artificial queen cells versus a wooden 3-hole mailing cage, a standard introduction method. The queen cell methods gave high acceptance (95% and 93% for natural and artificial, respectively) even though the queen was released from the queen cell approximately 10 minutes after being introduced into the mating hive. By contrast, success using mailing cages was significantly lower (47% and 73%) when the queen was released from her cage after 1 hour or 48 hours, respectively. The equal success rates of the reused and artificial queen cells suggests that high success is not due to chemicals present in natural queen cells transferring to the queens. To further investigate why queen cells give higher introduction success than cages, we introduced virgin queens into queenless observation hives. Workers attacked only 1 of 12 queens leaving a queen cell whereas 5 out of 6 queens leaving a cage were attacked.

Do honey bees tune error in their dances in nectar-foraging and house-hunting?

The "tuned-error" hypothesis states that natural selection has tuned the divergence angle in the dances of the honey bee to produce an optimal scatter of recruits across a resource. Weidenmuller and Seeley (Behav Ecol Sociobiol 46:190–199, 1999) supported this hypothesis by finding smaller divergence angles in dances indicating potential home sites, which are always point sources, than in dances indicating food sources, which often occur in patches. This study tested for the same effect, but controlled for variables, e.g., substrate and context, that may have confounded those results. When performed on the same substrate, divergence angle does not differ between dances for the two resources. Furthermore, dances performed for food within an observation hive exhibit significantly greater divergence angle when performed on comb (as Weidenmuller and Seeley measured food dances) than on hardware cloth (as they measured home-site dances on a swarm). These findings suggest that the angular variance in direction indication in dances is more likely an artifact of physical constraints, rather than an adaptive modification of a behavior that a bee could perform more precisely.

A scientific note on the lack of nepotism in queen larval feeding during emergency queen rearing in a naturally mated honey bee colony

Honey bee (Apis mellifera L.) colonies consist of many patrilines (Estoup et al., 1994) and workers can potentially increase their inclusive fitness during queen replacement if they can favour a full-sister offspring queen (Visscher, 1998). There is evidence of preferential rearing of some patrilines in emergency queen rearing (Tilley and Oldroyd, 1997; Osborne and Oldroyd, 1999; Châline et al., 2003), although the reason for this is unclear. Some patrilines could act nepotistically or a “royalty allele” could make some larvae more attractive to workers (Osborne and Oldroyd, 1999). Early studies on behavioural interactions and larval preferences during queen rearing used colonies with limited numbers of patrilines or cross-fostering of larvae between unrelated colonies with naturally mated queens. They showed weak or no evidence of nepotism (Breed et al., 1984; Noonan, 1986; Page et al., 1989). To avoid the problem of unnaturally low number of patrilines and the use of artificial insemination, we used microsatellites to analyse the kinship of feeding workers and queen larvae during emergency queen rearing, for the first time in a colony with a naturally mated queen. The experiment was carried out in double blind in a queenless observation hive containing individually marked bees. This was done by marking bees emerging from two brood combs of the test colony over 6 consecutive days. Before the experimental colony set-up, the marked bees were returned to their mother colony to ensure a normal ontogeny. Twentyfour hours before the set-up, 50 queen cells were initiated in a queen rearing colony by grafting larvae from the test colony. Thirteen of these were then attached to a brood comb from the test colony and transferred in an observation hive with 2000 of the marked workers. During the experiment, three additional queen cells were built naturally. Each queen cell was observed individually for 1/2 h every day until capping. The total observation time was 2130 min. The bees feeding each queen larvae were recorded, as this seemed the best behaviour for the

Representing Honey Bee Behavior for Recognition Using Human Trainable Models

Identifying and recording subject movements is a critical, but time-consuming step in animal behavior research. The task is especially onerous in studies involving social insects because of the number of animals that must be observed simultaneously. To address this, we present a system that can automatically analyze animal movements, and label them, by creating a behavioral model from examples provided by a human expert. Further, in conjunction with identifying movements, our system also recognizes the behaviors made up of these movements. Thus, with only a small training set of hand labeled data, the system automatically completes the entire behavioral modeling and labeling process. For our experiments, activity in an observation hive is recorded on video, that video is converted into location information for each animal by a vision-based tracker, and then numerical features such as velocity and heading change are extracted. The features are used in turn to label the sequence of movements for each observed animal, according to the model. Our approach uses a combination of kernel regression classification and hidden Markov model (HMM) techniques. The system was evaluated on several hundred honey bee trajectories extracted from a 15 minute video of activity in an observation hive.

Confinement of small hive beetles (Aethina tumida) by Cape honeybees (Apis mellifera capensis)

In this study we quantify small hive beetle (Aethina tumida Murray) and Cape honeybee (A.m. capensis Esch., an African subspecies) behaviours that are associated with beetle confinement in an effort to understand why Cape bees can withstand large beetle infestations. Four observation hives were each inoculated with 25 beetles and were observed for 11-17 days. Data collected included guard bee (worker bees who guard beetle confinement sites) and confined beetle behaviour. There was considerable colony variation for many bee and beetle behaviours. Overall, there were more beetle guards during evening, which was likely an effort to keep increasingly active beetles contained. About one-fifth of the beetles were found at the comb periphery although the colonies suffered no ill effects. Although beetles reached the combs, the bees were able to prevent beetle reproduction. Our data suggest that confinement is more likely an initial defence against invading beetles and not the sole reason African subspecies of honeybees are usually immune to beetle infestations while European bees are not. Aethina tumida / Apis mellifera / confinement / Cape honeybee / defense behavior

CLIMATE CONTROL, BEE STYLE

Whatever the weather, no self-respecting hive lets the temperature drop inside; the future of the next generation depends on it. If the hive's temperature falls a few degrees below 35°C, the incubating young's development is seriously threatened. Marco Kleinhenz explains that much of the hive's heat derives from general hive-keeping activity, but workers also help raise the temperature by congregating at the brood and vibrating their thoracic muscles to warm the incubating young. But how each individual worker contributes to the heating sum wasn't clear. Brigitte Bujok had noticed that some of the workers appear to position themselves over an incubating brood cell and press their thoraxes against the sealing caps `leaving a hot spot behind' explains Kleinhenz. But not much more was known about the brood's central heating strategy until Kleinhenz and Bujok teamed up with Stefan Fuchs and Jürgen Tautz to take a closer look at the heating bee's activity (p. 4217).Setting up a series of hives where they could monitor the bee's brood heating antics, Kleinhenz and his colleauges recorded individual's body temperatures with an infrared camera, and watched their behaviour. Some of the workers were happy to stay completely motionless on a brood cap for several minutes, pressing their thoraxes against the cap to warm the young within. But many of the bees preferred to find an empty cell amongst sealed brood cells,and take up residence, sometimes for over an hour. Kleinhenz explains that this apparently lazy behaviour had been noticed before, so he suspected that the bees were simply taking a rest; until he looked at their temperature. They were warming themselves before they vanished inside, and some bees emerged even warmer than when they went in! Were the bees really taking a nap, or were they taking up residence in an empty cell to warm the adjacent pupae?The team built a close observation hive where they could monitor bees'thoracic temperatures as they hunkered down in vacant cells. This time it was clear. Far from easing off and taking a rest, closeted bees were vibrating their thoracic muscles and reaching temperatures up to 41°C. Far from being idle, the workers were warming the brood.But how effective is the bee's novel heating strategy? Fitting a small resistor inside a bee's thorax, the team placed the simulated bee heater inside a cell, and recorded the temperature in adjacent cells. After 30 minutes, the heater bee's effect could be felt three chambers away. Kleinhenz also realised that once in place, a hot worker makes the most of its heat,warming all six of the surrounding cells, while usually a single occupant is the only beneficiary from a worker perched above it on the comb. What is more,the heater bees aren't profligate with their valuable resource, staying cool when surrounded by honey or vacant brood cells that don't require warmth, or when the brood is already warm enough.Kleinhenz also noticed that this dedicated band of bees also take great care not to over-heat their precious charges. They monitor the brood cells'temperatures and ensure a comfortable clime for all, by never letting the hive rise above 35.9°C.

Autogrooming by resistant honey bees challenged with individual tracheal mites

Autogrooming responses of resistant and susceptible strains of honey bees were measured when bees were challenged by placing adult female tracheal mites on their thoraces. Marked, young adult workers of the two strains of bees were added to colonies in observation hives. We transferred a single, live, adult, female mite onto the mesoscutum of a marked bee, monitored the bee for seven minutes and then removed it and searched for the mite. Greater proportions of resistant bees autogroomed, and resistant bees made more grooming attempts. Bees of both strains had equal apparent grooming effectiveness; grooming bees lost approx. 75% of mites. Control-group bees (those only stroked with the brush used to transfer mites) of the two bee strains did not differ in any response parameter. Resistant bees may have a lower threshold for responding by autogrooming when stimulated by mites on their body. Apis mellifera / honey bees / Acarapis woodi / tracheal mites / genetic resistance / grooming behavior

Absence of nepotism in the harassment of duelling queens by honeybee workers.

Nepotism shapes interactions among the members of almost every animal society. However, clear evidence of nepotism within highly cooperative insect societies, such as ant, wasp and honeybee colonies, is rare. Recent empirical findings suggest that nepotism occurs within honeybee colonies where kin-selection theory most strongly predicts its existence: during the lethal queen-queen duels that determine which of several young queens will become the colony's next queen. In this study, I test whether worker bees act nepotistically by hindering duelling queens that are distantly related to themselves. I accomplished this by observing labelled workers harassing duelling queen bees in observation hives and subsequently by determining worker-queen relatedness using DNA microsatellites. I show that the workers that harassed duelling queens were neither more-closely nor more-distantly related to them than were workers selected randomly from the colony. Thus, workers did not behave nepotistically by hindering half-sister queens more than full-sister queens. These results demonstrate that under certain conditions, natural selection limits the evolution of nepotism within animal societies despite strong theoretical predictions for its existence.

Honeybee foraging in differentially structured landscapes.

Honeybees communicate the distance and location of resource patches by bee dances, but this spatial information has rarely been used to study their foraging ecology. We analysed, for the first time to the best of the authors' knowledge, foraging distances and dance activities of honeybees in relation to landscape structure, season and colony using a replicated experimental approach on a landscape scale. We compared three structurally simple landscapes characterized by a high proportion of arable land and large patches, with three complex landscapes with a high proportion of semi-natural perennial habitats and low mean patch size. Four observation hives were placed in the centre of the landscapes and switched at regular intervals between the six landscapes from the beginning of May to the end of July. A total of 1137 bee dances were observed and decoded. Overall mean foraging distance was 1526.1 +/- 37.2 m, the median 1181.5 m and range 62.1-10037.1 m. Mean foraging distances of all bees and foraging distances of nectar-collecting bees did not significantly differ between simple and complex landscapes, but varied between month and colonies. Foraging distances of pollen-collecting bees were significantly larger in simple (1743 +/- 95.6 m) than in complex landscapes (1543.4 +/- 71 m) and highest in June when resources were scarce. Dancing activity, i.e. the number of observed bee dances per unit time, was significantly higher in complex than in simple landscapes, presumably because of larger spatial and temporal variability of resource patches in complex landscapes. The results facilitate an understanding of how human landscape modification may change the evolutionary significance of bee dances and ecological interactions, such as pollination and competition between honeybees and other bee species.

“Spraying” Behavior During Queen Competition in Honey Bees

In honey bees (Apis mellifera), virgin queens may eject a liquid substance from their abdomens while they are engaged in fatal combat. We investigated the functional significance of “spraying” behavior by staging queen “duels” within colonies housed in observation hives. Spraying occurred in 39.7% of all interactions between rival queens and was recorded in 12 of 15 duels. Workers were highly attracted to the surfaces and individuals contaminated by spraying, forming tight clusters with hundreds of bees in which movement was severely restricted. One or both queens in a pair became immobilized by the workers in 37.5 and 29.2% of all spraying events, respectively, but the queens were never killed by the workers. Conversely, a mobile queen penetrated the worker aggregation and stung an immobilized queen in one-third of the observed bouts. Thus spraying may serve as a fighting tactic by virgin queens to increase their chances of winning by temporarily immobilizing their rivals.

Cape (Apis mellifera capensis) and European (Apis mellifera) honey bee guard age and duration of guarding small hive beetles (Aethina tumida)

SUMMARYThe guard age and duration of European (Apis meilifera) and Cape (A m. capensis) honey bees guarding small hive beetles (Aethina tumida) were determined using 3–frame observation hives, noting the commencement and termination of beetle guarding by individually labelled honey bees. European honey bees in the USA began guarding small hive beetles significantly earlier (beginning age 18.55 ± 0.52 days; mean ± s.e.), guarded beetles significantly longer (duration 2.36 ± 0.31 days), and stopped guarding beetles significantly sooner (ending age 19.91 ± 0.57 days) than Cape honey bees in South Africa (beginning age 20.61 ± 0.38 days; duration 1.43 ± 0.12 days; and ending age 21.04 ± 0.37 days). Although the timing of beetle guarding behaviour between the two subspecies is significantly different, it does not explain the differential damage to European and Cape honey bee colonies caused by small hive beetles.

Hirsutella thompsonii and Metarhizium anisopliae as potential microbial control agents of Varroa destructor, a honey bee parasite.

The potential of Hirsutella thompsonii Fisher and Metarhizium anisopliae (Metschinkoff) as biological control agents of the parasitic mite, Varroa destructor Anderson and Trueman was evaluated in the laboratory and in observation hives. In the laboratory, time required for 90% cumulative mortality of mites (LT(90)) was 4.16 (3.98-4.42) days for H. thompsonii and 5.85 (5.48-7.43) days for M. anisopliae at 1.1 x 10(3) conidia mm(-2). At a temperature (34+/-1 degrees C) similar to that of the broodnest in a honey bee colony, Apis mellifera L., H. thompsonii [LC(90)=9.90 x 10(1) (5.86-19.35) conidia mm(-2) at Day 7] and M. anisopliae [LC(90)=7.13 x 10(3) (2.80-23.45) conidia mm(-2) at Day 7] both showed significant virulence against V. destructor. The applications of H. thompsonii to observation hives resulted in significant mortality of mites, and reduction of the number of mites per bee 21 and 42 days post-treatments. The treatments did not significantly affect the mite population in sealed brood. However, the fungus must have persisted because infected mites were still observed [82.97+/-(0.6)%] 42 days post-treatment. In addition, the fungus was found to sporulate on the host. A small percentage [2.86+/-(0.2)%] of dead mites found in the control hives also showed fungal infection, suggesting that adult bees drifted between hives and disseminated the fungus. H. thompsonii was harmless to the honey bees at the concentrations applied and did not have any deleterious effects on the fecundity of the queens. Microbial control with fungal pathogens provides promising new avenues for control of V. destructor and could be a useful component of an integrated pest management program for the honey bee industry.

Policing of adult honey bees with activated ovaries is error prone

Individually-marked day-old anarchistic (from a line where workers lay eggs at high frequency) and wild-type worker honey bees (Apis mellifera) were introduced to queenless sections of anarchistic or wildtype host colonies housed in observation hives. After 14 days, some introduced workers had activated ovaries, and we then removed the screens separating the queenless from the queenright sections of the observation hives. We then observed all instances of aggression against marked workers for 3 hours. The colonies were then killed and all marked bees retrieved and scored for ovary activation. About 10–40% of workers had activated ovaries. In 3 of 6 colonies, wild-type workers were attacked more often if they had activated ovaries than workers without activated ovaries. Anarchistic workers were more likely to be attacked if they had activated ovaries in one anarchistic host colony only, but in one wild-type host they were more likely to be attacked if they had inactive ovaries. In all colonies there was no significant difference in attack rates between anarchistic and wild-type workers with activated ovaries. This indicates that, like wild-type workers, anarchistic workers are unable to mask their ovary activation from other workers. This study supports the hypothesis some policing occurs via attacks on individuals with activated ovaries, but that this system is much less accurate and acute than the policing system based on egg eating.

Calorimetric investigations on metabolic rates and thermoregulation of sleeping honeybees ( Apis mellifera carnica)

Heat production rates of sleeping honeybees were determined at ambient temperatures between 20 and 35 °C. They increased with temperature from 4.7 mW g −1 (S.D. 2.2 mW g −1) at 20 °C ( n=18) up to a value of 12.3 mW g −1 (S.D. 7.6 mW g −1 n=12) at 35 °C. This indicates that honeybees behave ectothermicly during sleep. The preferred ambient temperatures for sleep were investigated in a temperature choice experiment. The highest number of sleeping bees were found at 28 °C. Evaluation of sleep behaviour in an observation hive revealed that bees prefer the same ambient temperature of about 28 °C under natural conditions. Honeybees save energy during sleep with an ectothermic behaviour, but do not reduce their metabolic rates as much as possible by choosing places in the beehive with the lowest temperature. Instead, they prefer places with moderate intermediate temperatures, probably in order to promote regenerative processes during sleep.

Reconstruction of honeybee behavior within the observation hive

It is known that the behavior of worker honeybees inside the hive is affected by many factors, such as temperature, age, and number of workers. That is, brain signals for controlling their behavior depend not only on the inner body condition, but also on the environment. In order to investigate the behavior within the hive, the single-individual trailing (SIT) method has been applied to record the time-varying behavior of honeybees. In this study, behavior of worker honeybees was represented as a Markov process from the SIT data. Various behaviors with same statistical properties could be generated by this mathematical description. This model could be applied not only for reproduction of worker behavior in the hive, but also for study about the neural basis behavioral control.