REPRODUCTIVE CONFLICT IN BUMBLEBEES AND THE EVOLUTION OF WORKER POLICING

Worker policing (mutual repression of reproduction) in the eusocial Hymenoptera represents a leading example of how coercion can facilitate cooperation. The occurrence of worker policing in “primitively” eusocial species with low mating frequencies, which lack relatedness differences conducive to policing, suggests that separate factors may underlie the origin and maintenance of worker policing. We tested this hypothesis by investigating conflict over male parentage in the primitively eusocial, monandrous bumblebee, Bombus terrestris. Using observations, experiments, and microsatellite genotyping, we found that: (a) worker‐ but not queen‐laid male eggs are nearly all eaten (by queens, reproductive, and nonreproductive workers) soon after being laid, so accounting for low observed frequencies of larval and adult worker‐produced males; (b) queen‐ and worker‐laid male eggs have equal viabilities; (c) workers discriminate between queen‐ and worker‐laid eggs using cues on eggs and egg cells that almost certainly originate from queens. The cooccurrence in B. terrestris of these three key elements of “classical” worker policing as found in the highly eusocial, polyandrous honeybees provides novel support for the hypothesis that worker policing can originate in the absence of relatedness differences maintaining it. Worker policing in B. terrestris almost certainly arose via reproductive competition among workers, that is, as “selfish” policing.     

A test of worker policing theory in an advanced eusocial wasp, Vespula rufa

Mutual policing is an important mechanism for maintaining social harmony in group-living organisms. In some ants, bees, and wasps, workers police male eggs laid by other workers in order to maintain the reproductive primacy of the queen. Kin selection theory predicts that multiple mating by the queen is one factor that can selectively favor worker policing. This is because when the queen is mated to multiple males, workers are more closely related to queen's sons than to the sons of other workers. Here we provide an additional test of worker policing theory in Vespinae wasps. We show that the yellowjacket Vespula rufa is characterized by low mating frequency, and that a significant percentage of the males are workers' sons. This supports theoretical predictions for paternities below 2, and contrasts with other Vespula species, in which paternities are higher and few or no adult males are worker produced, probably due to worker policing, which has been shown in one species, Vespula vulgaris. Behavioral observations support the hypothesis that V. rufa has much reduced worker policing compared to other Vespula. In addition, a significant proportion of worker-laid eggs were policed by the queen.     

Colony stage and not facultative policing explains pattern of worker reproduction in the Saxon wasp

Inclusive fitness theory predicts that in colonies of social Hymenoptera headed by a multiple‐mated queen, workers should benefit from policing eggs laid by other workers. Foster & Ratnieks provided evidence that in the vespine wasp Dolichovespula saxonica, workers police other workers’ eggs only in colonies headed by a multiple‐mated queen, but not in those headed by a single‐mated one. This conclusion, however, was based on a relatively small sample size, and the original study did not control for possible confounding variables such as the seasonal colony progression of the nests. Our aim, therefore, was to reinvestigate whether or not facultative worker policing occurs in D. saxonica. Remarkably, our data show that in the studied Danish population, there was no correlation between worker–worker relatedness and the percentage of worker‐derived males. In addition, we show that variability in cuticular hydrocarbon profiles among the workers did not significantly correlate with relatedness and that workers therefore probably did not have sufficient information on queen mating frequency from the workers’ cuticular hydrocarbon profiles. Hence, there was no evidence that workers facultatively policed other workers’ eggs in response to queen mating frequency. Nevertheless, our data do show that the seasonal progression of the nest and the location in which the males were reared both explain the patterns of worker reproduction found. Overall, our results suggest that the earlier evidence for facultative worker policing in D. saxonica may have been caused by accidental correlations with certain confounding variables, or, alternatively, that there are large interpopulation differences in the expression of worker policing.     

Worker policing in the common wasp Vespula vulgaris is not aimed at improving colony hygiene

In insect societies, eggs laid by workers are frequently killed by other workers – a behaviour known as “worker policing”. The traditional explanation of worker policing is that it is a mechanism to resolve intracolony conflict, and maintain the reproductive monopoly of the queen. Recently, Pirk et al. (2004) proposed that worker policing instead is aimed at removing unviable worker-laid eggs and is ultimately just another example of hygienic behaviour. Here we test this hypothesis for the common wasp Vespula vulgaris, a species with highly effective worker policing. We show that worker-laid eggs from queenless colonies have a lower hatch rate (68%) than queen-laid eggs (82%). Analysis of egg laying rates of queens and workers, however, shows that the difference is not big enough to explain the apparent absence of adult worker-derived males in this species. Received 30 January 2006; revised 2 May 2006; accepted 5 May 2006.     

Male parentage in army ants

In most social insects workers do not mate, but have retained the ability to produce haploid eggs that can develop into viable male offspring. Under what circumstances this reproductive potential is realized and how the ensuing worker-queen conflict over male production is resolved, is an area of active research in insect sociobiology. Here we present microsatellite data for 176 males from eight colonies of the African army ant Dorylus (Anomma) molestus. Comparison with worker genotypes and inferred queen genotypes from the same colonies show that workers do not or at best very rarely reproduce in the presence of the queen. Queens of D. (A.) molestus are known to be highly multiply mated. This implies that workers are on average more closely related to queen sons than to other workers' sons, so that our results are consistent with predictions from inclusive fitness theory. It remains unknown, however, whether worker sterility is maintained by active worker policing or by self-restraint.     

THE EVOLUTION OF WORKER STERILITY IN HONEY BEES: AN INVESTIGATION INTO A BEHAVIORAL MUTANT CAUSING FAILURE OF WORKER POLICING

Normally, worker honey bees (Apis mellifera) only lay eggs when their colony is queenless. When a queen is present, worker egg-laying is controlled by mutual “policing” behavior in which any rare worker-laid eggs are eaten by other workers. However, an extremely rare behavioral phenotype arises in which workers develop functional ovaries and lay large numbers of eggs despite the presence of the queen. In this study, microsatellite analysis was used to determine the maternity of drones produced in such a colony under various conditions. One subfamily was found to account for about 90% of drone progeny, with the remainder being laid by other subfamilies or the queen. No evidence of queen policing was found. After a one-month period of extreme worker oviposition in spring, the colony studied reverted to normal behavior and showed no signs of worker oviposition. However, upon removal of the queen, workers commenced oviposition very quickly. Significantly, the subfamily that laid eggs when the queen was present did not contribute to the drone production when the colony was queenless. However, another subfamily contributed a disproportionately large number of drones. The frequency of worker oviposition appears to be determined by opposing selective forces. Individual bees benefit from personal reproduction, whereas other bees and the colony are disadvantaged by it. Thus a behavioral polymorphism can be maintained in the population in which some workers can escape worker policing, with balancing selection at the colony level to detect and eliminate these mutations.     

Worker policing persists in a hopelessly queenless honey bee colony (<Emphasis Type="Italic">Apis mellifera</Emphasis>)

Summary In queenright colonies of Apis mellifera, worker policing normally eliminates worker-laid eggs thereby preventing worker reproduction. However, in queenless colonies that have failed to rear a replacement queen, worker reproduction is normal. Worker policing is switched off, many workers have active ovaries and lay eggs, and the colony rears a last batch of male brood before dying out. Here we report a colony which, when hopelessly queenless, did not stop policing although a high proportion of workers had active ovaries (12.6%) and many eggs were laid. However, all these eggs and also worker-laid eggs transferred from another colony were policed. This unusual pattern was repeated eight weeks later by a second queenless colony made using worker bees from the same mother colony, which strongly suggests genetic determination.Received 19 May 2003; revised 11 September 2003; accepted 23 September 2003.     

Specialization in policing behaviour among workers in the ant Pachycondyla inversa

Most animal societies are non-clonal and thus subject to conflicts. In social insects, conflict over male production can be resolved by worker policing, i.e. eating of worker-laid eggs (WLE) or aggression towards reproductive workers. All workers in a colony have an interest in policing behaviour being expressed, but there can be asymmetries among workers in performing the actual behaviour. Here, we show that workers of the ant Pachycondyla inversa specialize in policing behaviour. In two types of behavioural assays, workers developed their ovaries and laid eggs. In the first experiment, reproductive workers were introduced into queenright colonies. In the second experiment, WLE were introduced. By observing which individuals policed, we found that aggressive policing was highly skewed among workers that had opportunity to police, and that a similar tendency occurred in egg policing. None of the policing workers had active ovaries, so that policing did not incur a direct selfish benefit to the policer. This suggests that policing is subject to polyethism, just like other tasks in the colony. We discuss several hypotheses on the possible causes of this skew in policing tasks. This is the first non-primate example of specialization in policing tasks without direct selfish interests.     

Convergent evolution of worker policing by egg eating in the honeybee and common wasp

Mutual policing, where group members suppress each others' reproduction, is hypothesized to be important in the origin and stabilization of biological complexity. Mutual policing among workers in social insects can reduce within-colony conflict. However, there are few examples. We tested for worker policing in the common wasp Vespula vulgaris. Workers rapidly removed worker-laid eggs but left most queen-laid eggs (four out of 120 worker eggs versus 106 out of 120 queen eggs remained after 1h). Ovary dissection (1150 workers from six colonies) revealed that a small but significant number of workers have active ovaries (4%) equivalent to approximately five to 25 workers per colony. Consistent with effective policing of worker reproduction, microsatellite analysis of males (270 individuals from nine colonies) detected no workers' sons. Worker policing by egg eating has convergently evolved in the common wasp and the honeybee suggesting that worker policing may have broad significance in social evolution. Unlike the honeybee, relatedness patterns in V. vulgaris do not explain selection for policing. Genetic analysis (340 workers in 17 nests) revealed that workers are equally related to the queen's and other workers' sons (worker-worker relatedness was 0.51 +/- 0.04, 95% confidence interval). Worker policing in V. vulgaris may be selected due to the colony-level benefit of conflict suppression.     

Worker reproduction in Formica ants

A potential tragedy of the commons arises in social-insect colonies where workers are fertile if egg-laying workers decrease their contribution to other tasks. We studied worker ovary development and egg laying in relation to kin structure, colony size, and the presence of a queen in nine species (11 populations) of Formica ants. Workers were highly fertile and laid eggs in the presence of a queen in five out of the seven species where egg samples were obtained. Worker fertility correlated neither with colony size nor with kin structure, which suggests that colony-level costs and efficiency of policing precede relatedness as the most important conflict determinant. We conclude that careful quantification of the costs of worker reproduction and policing is essential for inferences about the tragedy of the commons.     

Are workers of Atta leafcutter ants capable of reproduction?

Summary. Workers of most eusocial Hymenoptera can produce sons after queen loss, which (posthumously) benefits the queen and increases worker inclusive fitness. However, the evolutionary loss of worker ovaries has occurred in several lineages, while workers in other taxa may be infertile despite having ovaries. Workers of Atta leafcutter ants only lay trophic eggs in queenright colonies. Although Atta colonies are commonly kept at universities, museums, and zoos, no reports of worker sons in orphaned colonies exist, suggesting that Atta workers are infertile. To explicitly test this, we created eleven orphaned laboratory nests of Atta cephalotes, A. sexdens, and A. colombica, and maintained them for 3–6 months after queen loss. Eight colonies did not produce any brood, but three nests produced 1–4 worker-derived male larvae and pupae. Microsatellite genotyping indicated that these were worker sons. However, all males were tiny (3.5–9 mm long) compared to normal queen sons (16 mm long), and would almost certainly be unable to mate. We also found reproductive eggs, but most of these had no yolk and were thus inviable. We conclude that Atta workers are not completely infertile, but that worker fertility is low compared to the sister genus Acromyrmex, where workers routinely produce normally-size males after queen loss in the laboratory. We hypothesize that worker reproduction in orphaned Atta field colonies is almost never successful because the last workers die before their sons can be raised to adulthood, but that the importance of worker-laid trophic eggs for queen feeding has precluded the evolutionary loss of worker ovaries. Received 17 January 2005; revised 12 September 2005; accepted 3 October 2005.     

Do hornets have zombie workers?

Colonies of the European hornet, Vespa crabro, are typically founded by a single queen mated to a single male. From the resulting colony relatedness pattern we predicted strong worker-queen conflict over male production where both the workers and the queen attempt to produce the colony's males. To test for this conflict, male production was studied in 15 hornet nests using a combination of DNA microsatellite analysis (282 males), worker ovary dissections (500 workers from eight nests) and 50 h of observation (four nests). In contrast to our prediction, the data show that hornet males are queens' sons, that workers never attempt to lay eggs, rarely have activated ovaries, and that there is no direct aggression between the queen and the workers. This contrasts with other data for vespine wasps, which support relatedness predictions. Dolichovespula arenaria has the same kin structure as V. crabro and workers produce males in many colonies. The similarity between these two species makes it difficult to explain why workers do not reproduce in V. crabro. Self-restraint is expected if worker reproduction significantly reduces colony productivity but there is no obvious reason why this should be important to V. crabro but not to D. arenaria. Alternatively, queen control may be important. The absence of expressed queen-worker conflict rules out physical control. Indirect pheromonal control is a possibility and is supported by the occurrence of royal courts and queen pheromone in Vespa but not Dolichovespula. Pheromonal queen control is considered evolutionarily unstable, but could result from a queen-worker arms race over reproductive control in which the queen is ahead. The genetic data also revealed diploid males in one colony, the first example in the vespine wasps, and two colonies with double matrilines, suggesting that occasional usurpation by spring queens occurs.     

Conflict over male parentage in stingless bees

Summary Stingless bees usually have one, singly-mated queen. This can lead to a genetic conflict of interest between the queen and the non-laying workers over who should produce the males. In many stingless bee species workers have developed ovaries and can produce male-destined eggs. In this study we compile the available data on who produces the males in stingless bees. Worker reproduction is common but less frequent than expected from predictions built on relatedness-based preferences of non-laying workers. We tested whether the pattern in worker reproduction can be explained best by queen control, by an arms race between workers and their queen, by the costs of losing workers to reproductive competition, or by phylogenetic constraints. The data are consistent with the view that there is ongoing conflict over male production that is resolved differently depending on the specific dynamics of costs and benefits of worker reproduction. There was also a role for phylogeny; Melipona workers often reproduced while Plebeia and Australian stingless bee workers seldom or never did. The high worker reproduction in Melipona may reflect low costs, because many of the replaced queen-laid eggs would become excess queens.Received 17 April 2003; revised 2 September 2003; accepted 5 September 2003.     

Workers dominate male production in the neotropical bumblebee Bombus wilmattae (Hymenoptera: Apidae)

Background

Cooperation and conflict in social insects are closely linked to the genetic structure of the colony. Kin selection theory predicts conflict over the production of males between the workers and the queen and between the workers themselves, depending on intra-colonial relatedness but also on other factors like colony efficiency, sex ratios, cost of worker reproduction and worker dominance behaviour. In most bumblebee (Bombus) species the queen wins this conflict and often dominates male production. However, most studies in bumblebees have been conducted with only a few selected, mostly single mated species from temperate climate regions. Here we study the genetic colony composition of the facultative polyandrous neotropical bumblebee Bombus wilmattae, to assess the outcome of the queen-worker conflict over male production and to detect potential worker policing.

Results

A total of 120 males from five colonies were genotyped with up to nine microsatellite markers to infer their parentage. Four of the five colonies were queen right at point of time of male sampling, while one had an uncertain queen status. The workers clearly dominated production of males with an average of 84.9% +/- 14.3% of males being worker sons. In the two doubly mated colonies 62.5% and 96.7% of the male offspring originated from workers and both patrilines participated in male production. Inferring the mother genotypes from the male offspring, between four to eight workers participated in the production of males.

Conclusions

In this study we show that the workers clearly win the queen-worker conflict over male production in B. wilmattae, which sets them apart from the temperate bumblebee species studied so far. Workers clearly dominated male production in the singly as well the doubly mated colonies, with up to eight workers producing male offspring in a single colony. Moreover no monopolization of reproduction by single workers occurred.     

Intracolonial conflict in the slave-making ant <Emphasis Type="Italic">Protomognathus americanus</Emphasis>: dominance hierarchies and individual reproductive success

Summary Social group viability results from a trade-off between cooperation and conflict, driven respectively by group and individual interests. Workers of the slave-making ants are known to have a high egg-laying potential, leading to a potential conflict over male production. Queenright and queenless nests of the slave-making ant Protomognathus americanus show a near-linear dominance hierarchy, and dominance rank is correlated with reproductive activity. Genetic and behavioural analysis revealed that the queen, when present in the nest, is behaviourally dominant and monopolises reproduction. In queenless nests, the haploid (male) brood is produced primarily by a single worker. We suggest the dominance hierarchy regulates male production, between the queen and her workers as well as among workers. Comparison of our results to another study allows us to place our data in an ecological context. This slave-making ant species appears to fit the concession model of reproductive skew: where resources (i.e. host nests) are poor, there is strong skew and where resources are richer reproduction is more egalitarian.Received 31 July 2003; revised 7 October 2003; accepted 9 October 2003.     

Worker policing in the German wasp Vespula germanica

In some ants, bees, and wasps, workers kill or "police" maleeggs laid by other workers in order to maintain the reproductiveprimacy of the queen. Kin selection theory predicts that multiplemating by the queen is one factor that can selectively favorworker policing. This is because when the queen is mated tomultiple males, workers are more closely related to the queen'ssons than to the sons of other workers. Earlier work has suggestedthat reproductive patterns in the German wasp Vespula germanicamay contradict this theory, because in some colonies a largefraction of the adult males were inferred to be the workers'sons, despite the effective queen mating frequency being greaterthan 2 (2.4). In the present study, we reexamine the V. germanicacase and show that it does support the theory. First, geneticanalysis confirms that the effective queen mating frequencyis high, 2.9, resulting in workers being more related to thequeen's sons than to other workers' sons. Second, behavioralassays show that worker-laid eggs are effectively killed byother workers, despite worker-laid eggs having the same intrinsicviability as queen-laid ones. Finally, we estimate that approximately58.4% of the male eggs but only 0.44% of the adult males areworker derived in queenright colonies, consistent with workerreproduction being effectively policed.     

Self-restraint and sterility in workers of Acromyrmex and Atta leafcutter ants

Summary. Queens of leafcutter ants (Acromyrmex and Atta) are highly multiply mated, resulting in a potential queenworker and worker-worker conflict over who should produce the males in the colony. We studied whether this conflict is expressed, by determining the amount of reproductive egg-laying by workers in queenright colonies of Acromyrmex echinatior, Acromyrmex octospinosus, Atta cephalotes, and Atta sexdens through ovary dissections. Worker sons are absent or rare in queenright Acromyrmex colonies, but can be produced in orphaned colonies. In Atta, most workers have rudimentary ovaries that never produce eggs, but a few (mostly small and medium workers that form a retinue around the queen) lay many trophic eggs for consumption by the queen. These eggs are large, flaccid, and lacking in yolk compared to queen-laid eggs, and appear to be always inviable. In Acromyrmex, many workers (especially young large workers) lay eggs that are similar in size to queen-laid eggs, but mostly with a reduced amount of yolk. Trophic eggs appear to be an important source of food for larvae in Acromyrmex (especially in Ac. echinatior), but not in Atta. Five (0.8) of 616 dissected Ac. echinatior workers but no Ac. octospinosus workers (n = 552), had ready-to-lay reproductive eggs. Old workers in all four species are incapable of laying eggs due to ovary resorption. We conclude that Atta workers are sterile, while Acromyrmex workers display reproductive self-restraint, possibly representing an earlier stage in the evolution towards worker sterility. Worker selfrestraint in Acromyrmex may be maintained by a queen or worker policing mechanism, but individual cost-benefit explanations may also apply.Received 1 March 2004; revised 28 June 2004; accepted 1 July 2004.     

Egg marking pheromones of anarchistic worker honeybees (Apis mellifera)

In honeybees, worker policing via egg eating enforces functionalworker sterility in colonies with a queen and brood. It is thoughtthat queens mark their eggs with a chemical signal, indicatingthat their eggs are queen-laid. Worker-laid eggs lack this signaland are, therefore, eaten by policing workers. Anarchistic workerhoneybees have been hypothesized to circumvent worker policingby mimicking the queen egg-marking signal. We investigated thisphenomenon by relating chemical profiles of workers and theireggs to egg acceptability. We found that the ability of someworkers (anarchistic workers in queenright colonies and deviantworkers from a queenless colony) to lay more acceptable eggsis due to them producing significant amounts of queen-like estersfrom their Dufour's gland. These esters appear to be transferredto eggs during laying and increase egg survival. However, theseesters cannot be the normal queen egg-marking signal, as theyare generally absent from queen-laid eggs and only increasethe short-term persistence of worker-laid eggs, because only7–30% of anarchistic worker-laid eggs persisted to hatchingversus 91–92% of queen-laid eggs. All workers can producesome esters, but only workers that greatly increase their esterproduction lay more acceptable eggs. The production of estersappears to be a flexible response, as anarchistic workers rearedin queenless colonies did not increase their ester production,while some deviant workers in queenless colonies did increasetheir ester production.     

Why acquiesce? Worker reproductive parasitism in the Eastern honeybee (Apis cerana)

Most societies are vulnerable to rogue individuals that pursue their own interests at the expense of the collective entity. Societies often protect themselves from selfish behaviour by ‘policing’, thereby enforcing the interests of the collective over those of individuals. In insect societies, for example, selfish workers can activate their ovaries and lay eggs, exploiting the collective brood rearing system for individual benefit. Policing, usually in the form of oophagy of worker‐laid eggs, controls selfish behaviour. Importantly, once an effective system of policing has evolved, the incentive for personal reproduction is lost, and ‘reproductive acquiescence’ in which ovary activation is rare or absent is predicted to evolve. Studies of social Hymenoptera have largely supported the prediction of worker ‘acquiescence’; workers of most species where policing is well developed have inactive ovaries. However, the eastern honeybee Apis cerana appears to be an exception. A. cerana colonies are characterized by highly efficient policing, yet about 5% of workers have active ovaries, even when a queen is present. This suggests that the evolution of acquiescence is incomplete in A. cerana. We regularly sampled male eggs and pupae from four A. cerana colonies. Workers had high levels of ovary activation overall (11.7%), and 3.8% of assignable male eggs and 1.1% of assignable male pupae were worker‐laid. We conclude that workers with active ovaries lay their eggs, but these rarely survive to pupation because of intense policing. We then used our findings as well as previously published data on A. cerana and A. mellifera to redo the meta‐analysis on which reproductive acquiescence theory is based. Including data on both species did not affect the relationship between effectiveness of policing and levels of worker reproduction. Their inclusion did, however, seriously weaken the relationship between relatedness among workers and levels of worker reproduction. Our work thus suggests that relatedness among workers does not affect the probability that workers will attempt to reproduce, but that it is coercion by peers that limits worker reproduction.     

Comparative analysis of worker reproduction and policing in eusocial hymenoptera supports relatedness theory

In many bees, wasps, and ants, workers police each other in order to prevent individual workers from selfishly producing their own male offspring. Although several factors can selectively favor worker policing, genetic relatedness is considered to be of special importance. In particular, kin selection theory predicts that worker policing should be more common in species where workers are more related to the queen's sons than to other workers' sons. Here we provide strong novel support for this theory based on a comparative analysis of policing and male parentage in 109 species of ants, bees, and wasps. First, an analysis of behavioral data confirms that worker policing occurs more frequently in species where workers are more related to the queen's sons than to other workers' sons. Second, an analysis of male parentage shows that a significantly higher percentage of the males are workers' sons in species where the workers are more related to other workers' sons. Both conclusions also hold if data are analyzed using phylogenetically independent contrasts. Although our analysis provides strong overall support for the theory that relatedness affects kin conflict over male parentage, there is also significant residual variation. Several factors that may explain this variation are discussed.