Task Stimulus And Task Threshold

To understand honey bee behavioural genetics, we must first appreciate the genetic structure of colonies. A colony is neither an individual nor a

Drone offspring


Fathering drone 1

Fathering drone 2

Subfamily 1 Subfamily 1 Subfamily 2

Fathering drone n a

Subfamily 1 Subfamily 1 Subfamily 2

Subfamily n

FIG. 5 The social and genetic structure of a honey bee colony. A single queen mates with multiple fathering drones to produce female (worker) offspring. The offspring sired by each drone represent different patrilines among the workers. Behavioural variation among patrilines can indicate that behavioural variation has a genetic basis. The queen will also lay unfertilized eggs parthenogenetically, which give rise to male (drone) offspring.

Subfamily n

FIG. 5 The social and genetic structure of a honey bee colony. A single queen mates with multiple fathering drones to produce female (worker) offspring. The offspring sired by each drone represent different patrilines among the workers. Behavioural variation among patrilines can indicate that behavioural variation has a genetic basis. The queen will also lay unfertilized eggs parthenogenetically, which give rise to male (drone) offspring.

population, but an extended family (Fig. 5). Colony-level behavioural phe-notypes arise primarily from the interactions of workers with each other and their environment (Page and Mitchell, 1998). Because queens mate with about 10-20 haploid males (Palmer and Oldroyd, 2000), the worker population comprises a mixture of full and half sisters (Page and Laidlaw, 1988). Further, because males produce sperm clonally, there is three times more genetic distance between subfamilies (or patrilines; daughters of different males) than there is among workers within them (Laidlaw and Page, 1984). This means that for any trait for which there is genetic variance for the probability that a worker will engage in a particular task (i.e. they differ in their task threshold), workers of different patrilines will behave differently in response to the same task stimulus (Fig. 6), some measure of the colony's need for a task is to be performed (Calderone and Page, 1988; Page et al., 1989a; Robinson and Page, 1989a; Robinson, 1992; Bonabeau et al., 1996, 1998;Theraulaz et al., 1998; Beshers and Fewell, 2001; Fewell, 2003; Myerscough and Oldroyd, 2004). This results in the often-reported phenomenon of task specialization, in which workers of certain subfamilies are more likely to engage in particular tasks than workers of other subfamilies (Table 1), given a particular level of task stimulus.

Despite the widespread occurrence of task specialization, there is no simple relationship between a genetically based task threshold and actually performing that task or behaviour (Jones et al., 2004; Graham et al., 2006). This is because, within the context of a nest, the level of task stimulus experienced by individual workers is influenced by the activities of other workers (Page

FIG. 6 A descriptive model of how differing task thresholds among patrilines translates into the number of workers engaged in a task. The model describes a situation in a small nest in which a few workers are available to engage in the task (i.e. they are of the correct age). (A) There is a genetic basis to task threshold such that patrilines (a,b,c,d,e) differ in their response to the same stimulus. (B) Patrilineal proportions in a sample of workers observed to be engaged in a task when the task stimulus is 'Low' (from A). At this level of stimulus, the task threshold of all members of patriline a is exceeded and all members of this patriline will engage in the task. A small number of patriline b will also be engaged, as may a small number of workers of other patrilines if their local stimulus is sufficiently high. (C) Here the stimulus is high enough that about 2/3 of the members of patriline b engage in the task. As this is a numerically dominant patriline, the proportion of workers of patriline a engaged in the task declines relative to the low stimulus scenario, even though the task threshold for all members of the patriline is exceeded. The task threshold model is subject to negative feedback.

FIG. 6 A descriptive model of how differing task thresholds among patrilines translates into the number of workers engaged in a task. The model describes a situation in a small nest in which a few workers are available to engage in the task (i.e. they are of the correct age). (A) There is a genetic basis to task threshold such that patrilines (a,b,c,d,e) differ in their response to the same stimulus. (B) Patrilineal proportions in a sample of workers observed to be engaged in a task when the task stimulus is 'Low' (from A). At this level of stimulus, the task threshold of all members of patriline a is exceeded and all members of this patriline will engage in the task. A small number of patriline b will also be engaged, as may a small number of workers of other patrilines if their local stimulus is sufficiently high. (C) Here the stimulus is high enough that about 2/3 of the members of patriline b engage in the task. As this is a numerically dominant patriline, the proportion of workers of patriline a engaged in the task declines relative to the low stimulus scenario, even though the task threshold for all members of the patriline is exceeded. The task threshold model is subject to negative feedback.

et al., 1989a; Calderone and Page, 1991, 1992; Myerscough and Oldroyd, 2004) and by unequal numerical strength or efficiency of patrilines. In Fig. 6 we assume a five-patriline colony in which the patrilines have a different task threshold for a particular task, like ventilating the nest entrance by wing fanning, for example. When the stimulus for fanning (e.g. concentration of CO2 or nest temperature) is low, only the patriline with a low task threshold (a in Fig. 6) will engage in the task. When the task stimulus increases (e.g. when the ambient temperature is increased to 30 °C), workers of additional patrilines (b and c) will engage in the task. If we sampled and genotyped the workers engaged in fanning over a range of temperatures, our model predicts that the patriline proportions in our sample would change. Importantly, however, there will not necessarily be a linear increase in the number of patrilines among fanning workers as the temperature is increased and the patriline with the lowest task threshold will not necessarily be the most represented at all levels of task stimulus. The reasons why this is so are complex and interacting. When a numerically large patriline starts to engage in a task, its efforts should reduce the level of stimulus within the colony. Thus the efforts of the numerically dominant subfamily will swamp a numerically smaller patriline of lower or equal task threshold. Furthermore, all workers have a diverse behavioural repertoire where each behaviour has its own threshold (O'Donnell and Foster, 2001; Weidenmuller, 2004) and workers may abandon one task should their threshold for another be exceeded.


A related phenomenon to task specialization is behavioural overdominance (Moritz and Southwick, 1987; Hillesheim et al., 1989;Fuchs and Moritz, 1998). Behavioural overdominance describes the situation where a colony-level phenotype is strongly influenced by a relatively small proportion of workers. There are two main ways in which behavioural overdominace is likely to occur. First, if a colony's need for a task can be taken care of by a relatively small number of workers, who have a low threshold for that task, then the colony will have a colony-level phenotype that reflects that of the minority low-threshold workers (Fuchs and Moritz, 1998). Another way of looking at this is to note that a colony comprising a mixture of both low-and high-threshold workers for a particular task would have the same or similar colony-level phenotype as a colony comprised solely of low-threshold individuals (Arathi and Spivak, 2001). Second, behavioural overdom-inance can occur if the activities of a small number of workers lowers the task threshold of other workers or increases the task stimulus that nestmates perceive. Any colony-level phenotype that is pheromonally orchestrated is likely to show a behavioural overdominance, because low-threshold workers will release pheromones leading to a response in high-threshold workers that would not otherwise have responded to the stimulus (Guzman-Novoa and Page, 1994; Hunt et al., 2003).


The primary determinant of a worker's task threshold is its age (Lindauer, 1971; Seeley and Kolmes, 1991). Young workers perform tasks within the nest, but soon graduate to duties like guarding on the nest periphery (Calderone, 1998). At about three weeks of age, workers begin foraging tasks that they undertake for the rest of their lives. This age-based task ontogeny is accompanied by fundamental changes in the exocrine and endocrine systems so that workers of a particular age are equipped with the appropriate glandular secretions that allow them to undertake tasks typical for that age. Thus, in young nurse bees the hypopharangeal glands are active, producing royal jelly for brood feeding. Later, the mandibular glands regress and the wax glands become active, allowing the worker to engage in comb building. At the onset of foraging, the wax glands will have regressed and the venom sack of the sting will be full. At this stage, the worker is motivated via neurochemical changes in its brain to forage and to learn the scents of flowers and their location (Bozic and Woodring, 1998; Schulz et al, 2002a).

The most fundamental change in a worker's adult life is the transition from nest-bound duties to foraging and the mechanisms behind this transition have been extensively studied. It is possible to manipulate age demography so that a colony is comprised mainly of nurse bees or of foragers (Robinson et al., 1989). As colonies require both nurses and foragers, manipulated colonies rapidly adapt by either accelerating or reversing the behavioural development of some of the workers, so that a more typical ratio of foragers and nurses is restored. The major conclusion of such experiments is that the rate at which a worker progresses from hive to foraging is modulated by social interactions with other workers (Robinson et al., 1992; Huang and Robinson, 1992, 1996, 1999; Giray et al., 1999; Jassim et al., 2000; Beshers et al., 2001). Foragers secrete a pheromone, ethyl oleate, which is transmitted via trophallaxis, and it is this pheromone that retards the development of nurse workers (Leoncini et al., 2004). In the absence of foragers, some workers become precocious foragers. Nurse workers have lower juvenile hormone (JH) titres than foragers, and progression to foraging tasks can be accelerated by treatment with JH (Fahrbach and Robinson, 1996; Giray and Robinson, 1996; Giray et al., 1999; Bozic and Woodring, 2000; Jassim et al., 2000; Sullivan et al., 2000, 2003; Bloch et al., 2001; Schulz et al., 2002a,b). Although workers that have had their corpora allata (sole source of JH) surgically removed still develop into foragers, the rate at which they do so is delayed (Sullivan et al., 2003). This suggests that JH is an important factor in regulating the behavioural development of workers.

Despite the important effects of the social environment on the rate of behavioural development, there is clear evidence that the age at first foraging (AFF) is genetically influenced (Giray et al., 1999; Rueppell et al., 2004a). Although we are unaware of any explicit demonstration, it seems likely that genetically based variance in the rate of progress through the normal age-based behavioural ontogeny would result in task specialization (Calderone and Page, 1988; Sullivan et al., 2000; Page and Peng, 2001; Page and Erber, 2002; Rueppell et al., 2004a). Patrilines with retarded development are more likely to be found engaged in nest-based tasks than patrilines with rapid behavioural development. Thus, we suspect that it will be eventually shown that many instances of task specialization (Table 1) are actually emergent properties of genetically-based variance in rates of behavioural ontogeny. However, this is by no means the whole story as genetically-based variance in perception of the environment, and the degree of response to it are known to underpin most task specialization in honey bees.

In the next section we discuss examples where genetically-based differences in perception lead to differences in worker behavioural phenotypes.

4.3.1 Hygienic behaviour

Hygienic behaviour can be quantified as a colony-level phenotype by the rate at which a colony cleans out brood cells containing dead larvae and dead pupae (Spivak and Gilliam, 1998a,b). Hygienic behaviour can be experimentally assayed by providing a colony with a small piece of dead brood, usually 100 brood cells. Colonies that remove all the dead pupae within 48 h are said to be 'hygienic'. Colonies that retain dead brood for more than 48 h are, by contrast, said to be 'non-hygienic' (Spivak and Downey, 1998). Hygienic behaviour is an important commercial trait because hygienic colonies are resistant to brood diseases and parasites (Spivak and Gilliam, 1998a,b; Spivak and Reuter, 1998, 2001). Genetic markers that could identify strongly hygienic colonies in the field would be very useful as a tool for selecting choice colonies for breeding and avoiding colonies that do not carry the hygienic marker.

For a colony to express the hygienic phenotype, a proportion of its middle-aged (2-3 weeks old) workers must sense the dead brood, uncap the brood cell and remove the infected pupae for disposal outside the nest. In his now classic study of behavioural genetics, Rothenbuhler (1964) proposed that the two steps of this process, uncapping and removal, are under separate genetic control, each behaviour controlled by a single unlinked Mendelian locus. This hypothesis has been very useful for understanding the behaviour and promoting new tests. The two-locus model is, however, probably an over-simplification. Moritz (1988) suggested that Rothenbuhler's data was more suggestive of three loci than of two, and Lapidge et al. (2002) suggested that the trait showed a quantitative pattern of inheritance, potentially involving many loci. Nonetheless, all authors agree that there is a strong genetic component to hygienic behaviour and that variation in this two-step task is controlled by a small number of loci affecting the hygienic threshold of workers.

Hygienic behaviour is based, in part, on the stimulus threshold of reaction to the odours of disease-killed brood. In a line bred for increased hygienic behaviour, workers can discriminate between, and respond to, the odours of healthy and disease-killed brood at lower concentrations than can workers from a line bred for poor hygienic behaviour (Masterman et al., 2001). When exposed to the same level of odour stimulus, hygienic-strain bees experience a stronger electrical signal in the antennal lobes of their brain, generating a higher level of the neuromodulator octopa-mine than do non-hygienic bees exposed to the same level of stimulus (Masterman et al., 2001; Spivak et al., 2003). High levels of octopamine appear to be a crucial factor behind a wide range of behaviour in honey bees, including hygienic behaviour, possibly because it is necessary for the formation of olfactory memory (Farooqui et al., 2003).

It is as yet unclear how the differing stimulus threshold between hygienic and non-hygienic strains is set. Is it that hygienic strain workers have more odorant receptors than non-hygienic strains, or do the same number of receptors result in a stronger signal that results in greater release of oc-topamine? We do not yet know the mechanism.

4.3.2 Foraging Specialization

Honey bee foragers show genetically-based specialization for water, nectar, or pollen collection, as evidenced by some patrilines that are more likely to collect pollen, nectar or water than others (Table 1). There is a negative correlation between the size of nectar loads and pollen loads carried by workers (Hunt et al., 1995; Page et al., 2000; Rueppell et al., 2004b), and this suggests that foragers are constrained by a maximum loading such that they cannot fly efficiently with both a fully-laden crop and corbicula (Feuerbacher et al., 2003) - hence the tendency for specialization on food types when on foraging trips.

The tendency for a worker to forage for water, nectar or pollen is strongly predicted by its sucrose threshold - the concentration of sucrose that a worker can distinguish from water in a proboscis extension response test (PER test, see Box 1) (Page et al., 1998; Scheiner et al., 2004). Water foragers sampled at the nest entrance have the lowest sucrose threshold. Those returning with pollen have the next highest, and those returning with nectar have the next highest response (Page et al., 1998; Pankiw and Page, 2000; Scheiner et al. , 2001, 2003a). Dreller (1998) found a strong genetic component to scouting behaviour. Hence, because individuals with the highest sucrose threshold are the most likely to return without any load, it may be that these workers are the ones scouting for very high-quality nectar sources.

Day-old workers that have not yet been exposed to environmental cues of the colony's nutritional needs show markedly different sucrose response thresholds, and these thresholds strongly predict whether they will become water, pollen, or nectar foragers later in life (Pankiw and Page, 1999, 2000). Thus it appears that a worker's innate sucrose response threshold has a strong genetic component, and this apparently leads to frequently reported specialization in foraging tasks (Table 1).

Box 1 - PER Test showing individual worker bees being presented with a droplet of sucrose. (Photo Courtesy of R. Maleszka)

Proboscis Extension Response Honey Bee

In the proboscis extension response test, a bee is confined in a narrow tube with its head emerging from one end. To test the bee's response to sucrose, its antenna is touched with a small drop of sucrose in solution. If the bee extends its tongue, it means that it has responded (Kuwabara, 1957). The PER test can be used in a variety of contexts for behavioural research. First, it can be used to test the sucrose concentration that a worker can distinguish from water (Page et al., 1998). Presumably, this is a measure of the individual's sucrose threshold. Second, it can be used as an assay for a worker's ability to learn (Bitterman, 1996; Chandra et al., 2000, 2001; Ferguson et al., 2001). The bee is first exposed to a conditioning stimulus, usually an odour, which is immediately followed by touching a small drop of sucrose to the antenna. If the bee extends its tongue, then it is allowed to feed from the sucrose. Most bees will learn to associate the odour with the reward after just one entrainment. Third, the test can be used to determine the relative ease with which workers can learn to distinguish odours (Masterman et al., 2000).

The central role of sucrose threshold response in determining specialized foraging is not surprising, given that sucrose is the major carbohydrate source for honey bees, requiring that they be able to assess the concentration of sucrose solutions. The antennae and proboscis are replete with sucrose receptors that have direct connections to areas of the brain associated with memory formation and reward mediated via the release of octopamine (Menzel and Muller, 1996; Hammer and Menzel, 1998). Sucrose rewards are apparently important, if not essential, in the formation of memory about foraging tasks (Menzel, 1999; Scheiner et al., 2003b) and bees with a low sucrose threshold can learn more easily in response to a sucrose reward than individuals with a high sucrose threshold (Scheiner et al., 2001). Interestingly, three QTLs for the ability to associate an odour with a sucrose reward have been identified (Chandra et al., 2001), but we do not know how this phenomenon relates to sucrose response threshold.

Although born with an innate sucrose threshold, a worker's threshold is modulated throughout life by environmental factors related to the colony's nutritional needs. This is necessary because the sucrose threshold of the colony members is tuned such that they respond appropriately to the colony's need (for pollen or nectar) and to the floral sources available. If colonies are foraging on nectar sources of high sucrose concentration, the sucrose threshold of workers is raised (Pankiw et al., 2004), reducing the probability of bees foraging at low sucrose concentration forage patches or dance for them (Seeley, 1986; Seeley et al., 1991, 2000). Presumably this shifts some workers of intermediate threshold towards pollen foraging. Furthermore, if colonies have a large number of larvae to feed, pherom-ones produced by the brood lower the sucrose threshold, increasing the probability that foragers will collect pollen thereby reducing the average age of workers at onset of foraging (Eckert et al., 1994; Pankiw et al., 1998, 2002; Dreller et al., 1999; Pankiw and Page, 2001, 2003).

Foragers can also directly detect the amount of pollen stored in their colony (Dreller et al., 1999; Dreller and Tarpy, 2000; Vaughan and Calderone, 2002). Depletion of stored pollen results in an increase in colony-level foraging, arising from an increase in the proportion of workers engaged in pollen foraging, the number of trips individuals make, and the size of the loads of pollen that they carry (Fewell and Winston, 1992; Fewell and Bertram, 1999; Janmaat and Winston, 2000; Rotjan et al., 2002). As yet, there has been no demonstration that the sucrose thresholds of foragers are raised in pollen-deprived colonies, but this is a clear prediction if sucrose response threshold is the prime mechanism promoting foraging specialization. The sucrose threshold increases with age, meaning that younger foragers are more likely to collect pollen than older ones (Rani and Jain, 1997; Pankiw and Page, 1999).

The above remarks indicate that sucrose response threshold is a significant causal factor influencing foraging task specialization (Fig. 7).

Water forager

Pollen forager

Nectar forager

Water forager

Pollen forager

Nectar forager

Sucrose concentration of field nectar


Amount of brood

FIG. 7 Hypothetical relationship between factors affecting sucrose threshold, which in turn affects the tendency of individual workers to perform certain tasks related to foraging. Individuals with a low threshold will tend to forage for water, while those with a medium threshold will have a tendency to forage for pollen, and individuals with a high sucrose threshold response tend to forage for nectar.

Physiological mechanisms behind the genetic influence on sucrose threshold are easy to envisage, and could include the number of sucrose receptors on the mouthparts and antennae, and the extent of octopamine release after stimulus by sucrose (Barron et al., 2002; Schulz et al., 2002a).

Four significant QTLs influencing foraging specialization have been identified, each of which is genetically linked to loci that affect foraging behaviour in individual workers. Three of these loci, plnl, pln2, and pln3, were identified via linkage mapping of workers segregating in backcrosses between lines selected at the colony level for high and low pollen hoarding (as in Fig. 2) using neutral molecular markers (Hunt et al., 1995). The pln loci have been identified in different crosses and populations, suggesting that they may be a general attribute of honey bees, potentially present in all populations (Hunt et al., 1995; Page et al., 2000; Rueppell et al., 2004b).

Other loci, such as Amfor, the Apis orthologue of Drosophila's forager gene and encoding cGMP-dependent protein kinase (Ben-Shahar et al., 2003), and Ammvl, the Apis orthologue of Drosophila's malvolio (Ben-Shahar et al., 2004), either directly influence the tendency to forage for pollen or nectar, or are linked to such a locus (Rueppell et al., 2004b). Owing to the possibility that Amfor and Ammvl are not directly involved in foraging specialization, it is best to provisionally regard Amfor and Ammvl as candidate QTLs for foraging specialization rather than as genes causing behavioural variation.

The four foraging specialization loci, i.e. pln1-3 and Amfor (pln4), interact epistatically to produce an individual's composite genotype that strongly influences the kind of forage that a worker is likely to collect. The actual genes associated with loci are yet to be identified, and so we do not know if or how these genes causally influence the sucrose threshold of workers. However, because these loci were first mapped in a cross between strains that show low and high sucrose thresholds (Page et al., 1998; Pankiw and Page, 1999), such a relationship is possible. It would be interesting to see if these loci influence sucrose threshold, and thus establish a clear pathway from gene, through physiology (sucrose threshold) to behaviour (foraging specialism) (Fig. 7). Rueppell et al. (2006) recently identified an additional QTL for the sucrose response threshold itself. Amfor. One of the most clear-cut examples of a single gene effect on natural variance in behaviour is the forager gene of Drosophila (Pereira and Sokolowski, 1993; Sokolowski, 2001). Two alleles exist in natural populations. The forR allele has a frequency of about 70% (Sokolowski et al., 1997). Individuals carrying the forR allele are of the "rover" pheno-type and forage over a larger area than individuals that are homozygous for the "sitter" allele forS (at 30% frequency), which are more sedentary (Debelle and Sokolowski, 1987). Forager encodes a cyclic guanosine monophosphate (cGMP) dependent protein kinase (PKG), and rovers have a much higher expression of the gene and PKG in their brains (Osborne et al., 1997). Because of its effects on foraging behaviour, forager was investigated as a candidate gene influencing foraging behaviour in honey bees (Ben-Shahar et al., 2002; Ben-Shahar, 2005; Fitzpatrick et al., 2005).

The honey bee orthologue, Amfor, shows increased expression, and PKG activity levels are much higher in workers engaged in tasks outside the nest (Ben-Shahar et al., 2003). Treatment of young workers with cGMP leads to precocious foraging and an increase in phototaxis, though the treatment itself may be partially causal. This led to the interpretation that in honey bees, cGMP initiates phototaxis, drawing middle-aged bees to the nest entrance where they are stimulated to forage by the smells and dances of older foragers (Ben-Shahar et al., 2003). However, this indirect mechanism seems an unlikely hypothesis because many species of honey bees nest in the open, where all the workers are exposed to constant light (Oldroyd and Wongsiri, 2006). Rather, it may be that cGMP stimulates foraging behaviour directly, possibly even to the level of foraging specialization (Rueppell et al., 2004b).

4.3.3 Nest defence

Honey bee nest defence is an extremely complex colony-level character (Collins et al., 1980; Breed et al., 2004). Most authors recognize two distinct components of colony defence: guarding and stinging. In any colony, some (5-10) or many (>100) workers will stand near the entrance and act as guards (Breed et al., 1992). The guards adopt a characteristic posture, with their front legs held off the substrate, their mandibles apart, and wings slightly spread (Butler and Free, 1952; Ghent and Gary, 1962). The guards often approach returning foragers, antennating them, and assessing whether they are non-nestmates. Should a non-nestmate attempt to enter the colony, the guard may grasp the intruder and eject it (Ribbands, 1954; Breed et al., 1992; Downs and Ratnieks, 1999).

Stinging behaviour occurs when a colony is disturbed by a predator and workers leave the nest to sting it. The guards are important to the initiation and coordination of stinging. If a predator approaches a colony, the guards will fly out to attack it, buzzing around the intruder, and possibly biting, and stinging. If the intruder does not withdraw, the guards at the nest entrance release alarm pheromones by exposing their stings. These phe-romones, primarily isopentyl acetate (Boch et al., 1962), alert nestmates, many of which will join the fray (Arechavaleta-Velasco and Hunt, 2003).

Stinging behaviour is generally quantified by measuring the time to first sting after a stimulus, the number of stings in a target provided near the colony under test, and some measure of the duration of attack after the stimulus is removed (Stort, 1974; Collins and Kubasek, 1982; Breed, 1991). Sting stimuli include physical disturbances, exposure to alarm pheromone, or movement - workers tend to sting things that move. Honey bee colonies vary greatly in their stinging behaviour (deGrandi-Hoffman et al., 1998) and defensiveness has a strong genetic component (Collins et al., 1984; Breed and Rogers, 1991; Stort and Goncalves, 1991; Guzman-Novoa et al., 1999). In particular, the tropically-adapted African A. m. scutellata and its New World derivatives are much more defensive than typical eco-types of European origin (Collins et al., 1982; Guzman-Novoa and Page, 1993; deGrandi-Hoffman et al., 1998). Africanized bees react to a visual stimulus 20-times faster than European bees, and deposit eight-times as many stings in experimental targets (Collins et al., 1982).

Guarding behaviour is a good example of task specialization - within a colony, guard bees are drawn from a non-random set of patrilines (Robinson and Page, 1988). The basis of this specialization appears to arise not from differing probabilities of engaging in guarding behaviour, but from the number of days in which individuals engage in the behaviour (Moore et al., 1987; Arechavaleta-Velasco and Hunt, 2003; Hunt et al, 2003). Most guards engage in the activity for about two days, but some individuals persist for six days (Moore et al., 1987; Breed et al., 1989; Arechavaleta-Velasco and Hunt, 2003). Clearly, the duration of guarding behaviour can lead to task specialization, even if there is no genetic influence on the probability of engaging in guarding.

There is a strong correlation between the persistence of individual guards and colony stinging behaviour (Breed et al., 1989). This suggests a causal link between the number of guards and colony level stinging behaviour. More guards mean that the colony is more likely to notice an intruder, decreasing the time to attack. A larger number of guards will also increase the amount of alarm pheromone released when a colony is disturbed, thus increasing the ferocity of the attack. Therefore, genetically-based variance in the number of days an individual spends guarding results in both task specialization in guarding and colony-level variation in stinging. We note that delay in transition from guarding to foraging, or a precocious transition from nursing to guarding could both produce this effect, supporting the hypothesis that genetically-based variance in the age at which behavioural transitions occur plays a pivotal role in the organization of work in honey bee colonies (Page et al., 1991; Huang and Robinson, 1992, 1996; Huang et al., 1994; Fahrbach and Robinson, 1996; Trumbo et al., 1997; Robinson and Huang, 1998; Giray et al., 1999; Schulz and Robinson, 1999; Jassim et al., 2000; Sullivan et al., 2000; Leoncini et al., 2004).

Hunt et al. (1998) identified five potential QTLs (designated stingl -sting5) that are apparently involved in the degree of stinging behaviour exhibited by colonies. The existence of sting-1, sting-2 and sting-3 was subsequently confirmed in independent crosses (Guzman-Novoa et al., 2002; Arechavaleta-Velasco et al., 2003; Arechavaleta-Velasco and Hunt, 2004). Sting-1 influences both the degree of colony-level stinging behaviour (time to first sting) and the probability of being a guard. The tendency to sting is apparently the dominant allele at this locus (Guzman-Novoa et al., 2002), but no candidate gene has been identified. Sting-2 and sting-3 affect the probability of guarding only (Arechavaleta-Velasco et al., 2003), and a further eight loci may also affect this trait (Arechavaleta-Velasco and Hunt, 2004).

The sting-2 locus has a total of 15 predicted genes within an 81 Kb BAC clone containing an STS linked to sting-2. This region does not include the whole confidence interval delimited by the QTL, so may not contain the causal gene(s). However, a large proportion of these predicted genes appear to be transcribed and are unique to honey bees (Lobo et al., 2003). One of these transcripts shows a number of base substitutions between European and African samples, and is therefore a strong candidate to be the gene linked to the QTL marker. If confirmed, this would be the first example of a gene that influences self-sacrificing task specialization being identified via linkage mapping in the honey bee.

The defensive response of a honey bee colony is largely mediated by alarm pheromones (Breed et al., 2004). Thus, in addition to the number and persistence of guards, the amount of alarm pheromone produced by individual workers and the reaction to these pheromones by other bees is likely to be important to the defensiveness of a colony. There is quantitative genetic variation for both the production of alarm pheromones (Collins et al., 1987a) and for responses to it (Collins et al., 1987b), and these two QTLs are not linked. A number of additional QTLs associated with production of the major sting-produced alarm pheromones have been mapped (Hunt et al., 1999), but these await confirmation in independent crosses.

4.3.4 Dance communication

Upon returning from a foraging trip, a successful forager may perform a communication dance that alerts her nestmates to the presence and location of her profitable patch (von Frisch, 1967). The dance is a stylized re-enactment of the foraging trip. During the 'waggle phase' of the dance, the worker strides forward while vigorously vibrating her abdomen from side to side (Tautz et al., 1996). The alignment of the bee's body during the waggle indicates the direction of the profitable patch. If the bee is aligned straight up the comb during the waggle run, the dancer is indicating a patch that is directly in the direction of the sun's current azimuth (the point where the sun is over the horizon). If the waggle run is aligned straight down the comb, the dance indicates a patch that is precisely opposite the sun's azimuth. And a dance orientated at three o'clock indicates a patch at 90° to the current azimuth. We will return to the directional aspects of the bee's dance in the next section.

The duration of the waggle run indicates the distance to the goal; short runs indicate nearby targets and prolonged runs indicate distant ones (von Frisch, 1967). Thus the bees that follow the dance can estimate the distance to the target food source by determining the average duration of the waggle runs. As the target gets closer to the colony, the dance becomes more and more hurried. At some point, the dancer is unable to complete the regular figure-eight that characterizes the true waggle dance and tends to run in excited circles, sometimes wagging her abdomen and sometimes not (Beekman et al., 2005). This form of dance is known as a 'round dance' (von Frisch, 1967). Early studies suggested that the slope of the curves that relate distance to dance tempo (Fig. 8) differ according to ecotype (Boch, 1957; von Frisch, 1967) and species (Lindauer, 1956, 1957; Punchihewa et al., 1985). More recent experiments have not been able to confirm differing species-specific dance forms (Dyer and Seeley, 1991; Sen Sarma et al., 2004).

Furthermore, we now know that the flying bee perceives the distance that she has traveled as the amount of optic flow past her eye as she flies

The Sugar Solution

The Sugar Solution

Curb Sugar Cravings Once And For All With These Powerful Techniques. Sugar sensitive people might be low in specific neurochemicals that help us feel calm, centered, confident, and optimistic. Sugar is a drug that temporarily makes the sugar sensitive feel better, but with damaging consequences.

Get My Free Ebook

Post a comment